Split (1)

train · 5.09k rows

pmcid stringlengths 6 6	title stringlengths 9 374	abstract stringlengths 2 4.62k ⌀	fulltext stringlengths 167 106k	file_path stringlengths 64 64
503391	Shoulder posture and median nerve sliding	Background Patients with upper limb pain often have a slumped sitting position and poor shoulder posture. Pain could be due to poor posture causing mechanical changes (stretch; local pressure) that in turn affect the function of major limb nerves (e.g. median nerve). This study examines (1) whether the individual components of slumped sitting (forward head position, trunk flexion and shoulder protraction) cause median nerve stretch and (2) whether shoulder protraction restricts normal nerve movements. Methods Longitudinal nerve movement was measured using frame-by-frame cross-correlation analysis from high frequency ultrasound images during individual components of slumped sitting. The effects of protraction on nerve movement through the shoulder region were investigated by examining nerve movement in the arm in response to contralateral neck side flexion. Results Neither moving the head forward or trunk flexion caused significant movement of the median nerve. In contrast, 4.3 mm of movement, adding 0.7% strain, occurred in the forearm during shoulder protraction. A delay in movement at the start of protraction and straightening of the nerve trunk provided evidence of unloading with the shoulder flexed and elbow extended and the scapulothoracic joint in neutral. There was a 60% reduction in nerve movement in the arm during contralateral neck side flexion when the shoulder was protracted compared to scapulothoracic neutral. Conclusion Slumped sitting is unlikely to increase nerve strain sufficient to cause changes to nerve function. However, shoulder protraction may place the median nerve at risk of injury, since nerve movement is reduced through the shoulder region when the shoulder is protracted and other joints are moved. Both altered nerve dynamics in response to moving other joints and local changes to blood supply may adversely affect nerve function and increase the risk of developing upper quadrant pain.	Background Non-specific arm pain (NSAP), often called repetitive strain injury, describes the common problem of upper limb pain and functional impairment without objective physical findings. The contributing factors to the development of NSAP are not fully understood but ergonomic guidelines commonly suggest that good upper body posture protects against NSAP (e.g. [ 1 ]). In a study of 485 NSAP patients, shoulder protraction and forward head position were reported in a majority of patients (78% and 71% respectively) [ 2 ]. Poor upper body posture (e.g. rounded shoulders, head forward) has also been reported to increase the incidence of neck and shoulder pain [ 3 ]. The possible mechanisms leading to pain in patients with postural malalignment have not been examined in any detail. The painful symptoms often associated with NSAP suggest a minor neuropathy involving at least in part the median nerve [ 4 ]. Together, shoulder protraction, forward head position and flexion of the trunk form the main components of slumped sitting. The present study examines the effects of each of these components on longitudinal sliding of the median nerve using high frequency ultrasound imaging in asymptomatic normal subjects. In addition, the effects of sustained protraction on nerve sliding through the shoulder region are examined. Methods Ultrasound imaging Longitudinal nerve movement was measured using high frequency ultrasound imaging, as previously described by Dilley et al. [ 5 , 6 ]. A Diasus ultrasound system (Dynamic Imaging, Livingston, Scotland, UK) was used to collect sequences of ultrasound images at 10 frames/second for 50 to 70 seconds, running at 10–22 MHz and using a 26 mm linear transducer. A cross-correlation algorithm was used to determine relative movement between adjacent frames in sequences of images [ 5 ]. The maximum correlation coefficient (r) was calculated for each pixel shift determining the relative movement between frames. To account for probe movement the same method was employed on deep stationary structures (eg. bone or interosseous membrane) and the result subtracted from the nerve excursion values. Subject details Fourteen healthy subjects, 5 male and 9 female, aged 25–38 years (mean = 32 years) were screened to exclude upper limb or cervical spine pathologies, rheumatological or neurological conditions. In each subject the nerve bed length was estimated, from the C6 spinous process to the tip of the index finger (mean = 97.0 cm (SD, 5.9)), and used to normalise the ultrasound transducer position between individuals. All measurements were taken from the right upper limb only. Set-up and procedure The median nerve was imaged in longitudinal section in the forearm during forward head position, trunk flexion and protraction and in the forearm and upper arm during contralateral neck side flexion (CNSF). Each movement was repeated three times, including some reverse trials. Forward head position Each subject (n = 8) was imaged in the proximal forearm whilst positioned upright on a chair fitted with a back and head support, hips and knees at 90° flexion, and the trunk fixed with Velcro strapping. The right upper limb was strapped to a Perspex plate in 90° flexion and 20° abduction at the glenohumeral joint, with the elbow fully extended, 45° forearm supination, and the wrist, hand and fingers in neutral. An active forward head position movement was performed, which included lower cervical spine flexion and upper cervical spine extension. Trunk flexion Each subject (n = 8) was imaged in the proximal forearm whilst positioned upright on a chair, with hips and knees at 90 degrees flexion. The right upper limb was positioned as for the forward head position trials. The subject was taught to actively flex their trunk whilst posteriorly tilting their pelvis. Protraction Each subject (n = 13) was imaged at two locations in the forearm and positioned as for the forward head position. The distal upper arm was imaged in three of the 13 subjects. For each trial the shoulder girdle was passively protracted from neutral (i.e. the scapulothoracic joint in neutral) by sliding the Perspex plate supporting the arm on an adjustable table. In three of these subjects, additional data was also obtained during ultrasound imaging. A potentiometer attached using strong thread to the acromion process allowed measurement of the amount of protraction. Protraction data was captured on to a PC and synchronised offline to the recorded ultrasound sequence. In four subjects, good quality images of the median nerve within the upper arm could be obtained. In the majority of subjects, it was difficult to acquire good quality images because of dense tissue overlying the nerve that reduced the image quality. From these images, nerve trunk bowing was measured in the distal upper arm with the shoulder girdle in the neutral and protracted positions. The maximum deviation of the nerve from a straight line across single ultrasound frames was measured offline in both positions and the difference used as a measure of additional bowing. Repeat trials were averaged. Contralateral neck side flexion Each subject (n = 11) was imaged in the distal forearm and distal upper arm, whilst lying supine with the right upper limb abducted to 90° at the glenohumeral joint. Ninety degree abduction at the glenohumeral joint rather than 90° glenohumeral flexion was used, so that the present data could be related to previous work [ 6 ] which has shown that median nerve movements can be reliably measured with the glenohumeral joint abducted to 90°. The examined limb was fixed to a Perspex plate using Velcro strapping with the elbow extended, forearm supinated and the wrist, hand and fingers in neutral. The head was supported on a movable plate, with the centre of rotation positioned at the C7 spinous process. In each subject, the neck was passively moved to 35° CNSF with (a) the scapulothoracic joint in neutral (i.e. relaxed lying in supine) and (b) in full protraction. This movement was repeated several times. In four subjects, a potentiometer attached to the plate allowed continuous measurement of the angle of CNSF. Joint angle data was captured on to a PC and synchronised offline to the recorded ultrasound sequence. Subject movement measurements For each procedure the range of movement was determined from pictures obtained using a digital camera. Changes in joint angle and distance for each movement were determined from skin surface markers and measured using either CorelDraw (Kodak Digital Science, USA) or "tpsDig" (F. James Rohlf, Department of Ecology and Evolution, State University of New York). Measurements for the individual components of slumped sitting are summarised in figure 1 . The posterior-anterior shift of the acromion was used as a measure of protraction during CNSF. Strain calculations Strain is defined by the difference in the amount of elongation that occurs at two points along a nerve divided by the distance between these two points. In practice strain was determined by using regression lines fitted to plots of nerve movement against the distance along the arm. Note that the strain estimates represent the additional strain produced by the movement rather than the total nerve strain. Statistical analysis Comparisons of nerve movement and strain in scapulothoracic neutral and protraction during CNSF were performed using paired t-tests. Results Median nerve movements in the arm in response to components of slumped posture Forward head position Moving the head forward while maintaining the shoulder and trunk position was tested in 8 subjects. This movement produced no detectable median nerve excursion in the forearm, the average trend being a movement of 0.1 mm (SEM, 0.02) occurring in a proximal direction. The repeat measure variability within subjects was very low, with a standard deviation ranging from 0–0.2 mm (mean = 0.1 mm). The mean change in the angle of lower cervical spine flexion and upper cervical spine extension was 23.6° (SD, 2.8) and 2.9° (SD, 1.9) respectively. Trunk flexion Trunk flexion also produced minimal median nerve excursion with a mean over 8 subjects of 0.1 mm (SEM, 0.1) proximal movement. The mean change in the angle of trunk flexion was 19.7° (SD, 4.7). Shoulder protraction In 13 subjects the median nerve moved in a proximal direction during shoulder protraction with more movement at proximal locations (mean in forearm = 3.5 mm (SEM, 0.3), mean in upper arm = 5.9 mm (SEM, 0.6)) (figure 2 [see additional file 1 for ultrasound sequence of median nerve sliding in the forearm]). The mean extent of scapular anterior translation was 38.3 mm (SD, 13). The additional strain on the median nerve was 0.7% (SEM, 0.3), given by the slope of the regression of nerve movement against distance along the arm (Figure 2 ). Median nerve excursion was measured in 3 subjects with simultaneous measurement of protraction. The results revealed an initial delay of 6.5–33.0 mm (mean = 17.0 mm; equivalent to 15.8–34.0% (mean = 23.7%) of the total protraction) before significant nerve movement occurred (see figure 3 ). After the initial delay, nerve movement was proportional to the extent of protraction. In three of four subjects, nerve bowing was observed in the upper arm with the shoulder girdle in the neutral test position. The maximum nerve course deviation from a straight line with the shoulder girdle in neutral compared to protraction was approximately 0.5 mm in all three subjects over the length of the ultrasound transducer, (26 mm). The nerve straightened during protraction (figure 4 ). Median nerve movement in the arm in response to contralateral neck side flexion (CNSF) with or without shoulder protraction In 11 subjects the median nerve moved in a proximal direction during 35° CNSF when the scapulothoracic joint was in neutral, as reported previously [ 6 ]. The movement increased at the more proximal location (scapulathoracic neutral, mean in upper arm = 2.3 mm (SEM, 0.2) and forearm = 1.5 mm (SEM, 0.2)). With the shoulder protracted, there was a 60% reduction in nerve movement in both upper arm and forearm locations (p < 0.05 for both locations) (mean in upper arm = 0.9 mm (SEM, 0.2) and forearm = 0.6 mm (SEM, 0.1)) (figure 5 ). The mean extent of protraction was 48.0 mm (SEM, 4.3). The additional strain on the median nerve was 0.3% (SEM, 0.1) in scapulothoracic neutral. There was a significant reduction in strain in protraction (0.1% (SEM, 0.1); p < 0.05, paired t-test). Median nerve excursion was measured in 4 subjects with simultaneous measurement of CNSF. The results revealed no obvious delay in the onset of nerve movement in protraction compared to scapulothoracic neutral. Despite less movement in protraction, the pattern of nerve movement mimicked that observed in scapulothoracic neutral (figure 6 ). Nine of 11 subjects reported paraesthesia in the distribution of the median nerve dermatone once the shoulder girdle was sustained in protraction. The onset of symptoms ranged from 1 to 4 minutes. Symptoms disappeared when the shoulder was repositioned in scapulothoracic neutral. Discussion Direct effects of the components of slumped sitting on median nerve movement Nerves are designed to slide and stretch to accommodate joint movement. Using the method of Dilley et al. [ 5 , 6 ], median nerve sliding was examined during the individual components of slumped sitting. Both forward head position and trunk flexion produced only minimal nerve movement in the forearm. The only examined component to produce substantial nerve movement was shoulder protraction. The median nerve strain in the forearm with protraction was 0.7%, which was well below the limits that cause changes to nerve function (reviewed in Grewel et al. [ 7 ]). As the shoulder girdle is protracted there is a delay in nerve movement, which is followed by a steady increase in nerve excursion. During this initial toe region the median nerve appears bowed in the upper arm. The nerve trunk appears to straighten as the range of protraction progresses. It therefore seems that with the upper limb in scapulothoracic neutral and the glenohumeral joint in 90° flexion, the median nerve is unloaded. If this is the case, the strain value of 0.7% will represent the total strain. Effects of protraction on the transmission of median nerve movement through the shoulder region The results for CNSF provide evidence for a possible restriction within the shoulder region during shoulder protraction. With the shoulder protracted there was a 60% reduction in the transmission of nerve movement through the upper limb. Consistent with a reduction in movement, there was also significantly less strain in the forearm. The possibility that the nerve becomes unloaded when the shoulder is protracted is unlikely since it had been found that protraction itself causes some median nerve stretch. In addition, there was no obvious delay in nerve movement in response to CNSF (Figure 6 ). The evidence for a restriction is consistent with previous suggestions that shoulder protraction may cause a neurovascular impingement within the shoulder region resulting in pain [ 2 , 8 ]. This suggestion was further supported by the experience of paraesthesia within the median nerve distribution during sustained protraction in 82% of subjects. These symptoms indicate the presence of a vascular restriction, which in turn affects neural function. Scapular protraction is a complicated movement, often resulting in the combined movement of numerous other structures within the shoulder girdle, including anterior displacement of the head of the humerus. It is therefore difficult to establish the precise cause of a neurovascular entrapment. Shortening of pectoralis minor and the downward displacement of the coracoid process might affect sliding of the cords of the brachial plexus. Alternatively, elevation of the first rib during full protraction (due to its soft tissue attachments with surrounding structures) might reduce the space between the clavicle and the first rib, restricting nerve sliding. Clinical significance The components of slumped sitting (i.e. forward head position, trunk flexion and protraction) are associated with poor posture [ 2 , 9 - 11 ], and are often adopted by office workers. Shoulder protraction is the only component of this posture to tension the median nerve, although the level of nerve strain in the forearm with the shoulder at 90° flexion and elbow extension, is not sufficiently high to result in direct neural injury. Problems are more likely to result from local effects of shoulder protraction on the chords of the brachial plexus. The present study shows that protraction restricts nerve sliding through the shoulder region. Most subjects also experienced paraesthesia when maintaining shoulder protraction plus elbow extension and shoulder abduction. Therefore, sustained shoulder protraction may place the median nerve at enhanced risk of injury and possibly cause a vascular compromise. This may in turn explain for the trend that a high number of NSAP patients have poor shoulder posture. (e.g. [ 2 ]). Conclusions The direct effects of slumped sitting on median nerve strain are not sufficient to alter nerve function. However, shoulder protraction does appear to restrict nerve sliding, and prolonged protraction leads to pareasthesias. Competing interests None declared. Authors' contributions AJ and RL participated in the study design, data collection, analysis and manuscript preparation. AD participated in the study design, analysis and manuscript preparation. BL participated in the study design and manuscript preparation. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Median nerve sliding in the forearm during protraction. Ultrasound sequence of median nerve sliding in the distal forearm during protraction. The subject was imaged with the limb in 90° flexion and 20° abduction at the glenohumeral joint and elbow neutral. The median nerve can be seen to slide in a proximal direction as the shoulder is protracted. In a repeat of the sequence the nerve movement is tracked using cross-correlation analysis (yellow plus sign). The total nerve movement was 4.70 mm. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC503391.xml
517496	Coactivators p300 and PCAF physically and functionally interact with the foamy viral trans-activator	Background Foamy virus Bel1/Tas trans -activators act as key regulators of gene expression and directly bind to Bel1 response elements (BRE) in both the internal and the 5'LTR promoters leading to strong transcriptional trans -activation. Cellular coactivators interacting with Bel1/Tas are unknown to date. Results Transient expression assays, co-immunoprecipitation experiments, pull-down assays, and Western blot analysis were used to demonstrate that the coactivator p300 and histone acetyltransferase PCAF specifically interact with the retroviral trans -activator Bel1/Tas in vivo . Here we show that the Bel1/Tas-mediated trans -activation was enhanced by the coactivator p300, histone acetyltransferases PCAF and SRC-1 based on the crucial internal promoter BRE. The Bel1/Tas-interacting region was mapped to the C/H1 domain of p300 by co-immunoprecipitation and pull-down assays. In contrast, coactivator SRC-1 previously reported to bind to the C-terminal domain of p300 did not directly interact with the Bel1 protein but nevertheless enhanced Bel1/Tas-mediated trans -activation. Cotransfection of Bel1/Tas and p300C with an expression plasmid containing the C/H1domain partially inhibited the p300C-driven trans -activation. Conclusions Our data identify p300 and PCAF as functional partner molecules that directly interact with Bel1/Tas. Since the acetylation activities of the three coactivators reside in or bind to the C-terminal regions of p300, a C/H1 expression plasmid was used as inhibitor. This is the first report of a C/H1 domain-interacting retroviral trans -activator capable of partially blocking the strong Bel1/Tas-mediated activation of the C-terminal region of coactivator p300. The potential mechanisms and functional roles of the three histone and factor acetyltransferases p300, PCAF, and SRC-1 in Bel1/Tas-mediated trans -activation are discussed.	Background In the sequential model of transcriptional regulation including viral trans -activation, coactivators CBP/p300 require concerted action of multiple protein factors provided the nucleosomal structures allow access to the DNA template [ 1 - 3 ]. The factors that interact with coactivators encompass sequence-specific DNA binding activators, non-DNA binding coactivators, and essential components of the basal transcriptional machinery. Two large, closely related human proteins, p300 and CBP, were identified and shown to function as versatile signal integrators of many transcription factors to facilitate transcriptional activation or repression, and, in addition, as connectors of multiple transduction pathways. Both proteins contain several conserved domains that include three Cys/His-rich (C/H1, C/H2, and C/H3) domains, the histone acetyltransferase (HAT), KIX, and Gln-rich (Q) domains among others (Fig. 1 ) [ 2 , 4 , 5 ]. It is mainly due to these domains that a plethora of transcriptional activators interact with p300/CBP. Thus, p300/CBP coactivators act as a physical and functional scaffold or bridge between various cellular or viral trans -activators and the basal transcriptional machinery. Both proteins function by mediating positive or negative cross talk between different signaling pathways and participate in fundamental cellular processes that include embryonic development, cell growth, differentiation, and apoptosis. In addition, they can act as tumor suppressors and, last but not least, directly interact with diverse viral trans -activators to facilitate virus replication or viral activator-mediated transformation [ 6 , 7 ]. Figure 1 Coactivator p300 enhances Bel1/Tas-mediated trans -activation. Simplified schematic drawing of p300 domains and structure of deletion mutants (not drawn to scale). Characterized domains of p300 [2, 5] relevant for this report are highlighted by differently marked boxes: the FLAG epitope upstream of the N-terminus of p300 is marked by shaded boxes in constructs no. 1 through 8, the C/H1 and KIX domains by black and striped boxes, respectively. Numbers above the rectangles indicate the number of amino acids of p300 and its derivatives. The intrinsic HAT domain of p300 is depicted between the C/H2 and C/H3 domains; additional HATs PCAF and SRC-1 reported to associate with the distinct p300 regions [8, 24, 25] are indicated as rectangles above p300. The domain marked "Q" is the glutamine-rich domain of p300. Construct no. 5 has a deletion between the C/H1 and KIX domain (broken line). Constructs no. 9 and 10 contain the GST-C/H1 and GST-ΔC/H1-KIX fusion proteins; partial shadings mark GST regions. Several models have been proposed to explain transcriptional activation. Coordinated recruitment of coactivators by diverse transcriptional activators to specific promoter target sites has been shown by a collective effort of many groups [ 8 ]. According to the models, different coactivators either modify chromatin structure by altering the nucleosomal DNA thereby affecting its accessibility to DNA-binding proteins or, alternatively, form complexes with HAT activities that by acetylation of specific Lysines in histone N-terminal tails weaken interactions between DNA and the histone octamer. Moreover, the HAT activities of some coactivators acetylate non-histone substrates such as viral and cellular trans -activators, for instance p53 [ 9 ]. Diverse viral trans -activator proteins were found to interact with distinct domains of p300/CBP and PCAF. Prominent among them are the early adenoviral E1A antigen [ 10 , 11 ], Epstein-Barr virus protein EBNA-2 [ 12 ], and human T-cell leukemia virus oncoprotein Tax [ 13 , 14 ]. Several other modifications including methylation, phosphorylation, and ubiquitination lead to either diminished or increased DNA binding of the activators that, in turn, will result in either a repression or activation of gene expression. In addition, both coactivators were reported to interact with additional HAT enzymes, namely PCAF and SRC-1 [ 15 , 16 ]. The apparently nonpathogenic primate foamy viruses (PFV) show a wide host range and tissue tropism and have been developed into vectors that efficiently transduce SCID-repopulating cells [ 17 ]. The PFV Bel1/Tas protein has been characterized as a transcriptional trans -activator of the acidic class and is known to directly interact with its responsive elements (BRE) [ 18 , 19 ]. Bel1/Tas is a nuclear protein and acts as the key regulator absolutely required for virus replication. The minimal Bel1-specific DNA target site is 27 base-pairs long and located within the internal promoter (IP.BRE) upstream of the second cap site that is part and parcel of the second PFV transcription unit [ 20 ]. Additional Bel1/Tas DNA target sites in the LTR region of the PFV DNA genome were not analyzed in this study. The acidic trans -activation domain (TAD) was mapped to the C-terminus of Bel1 with little if any protein homology to other FV Bel1/Tas proteins from different species. Previously, we identified the nuclear factor 1 (NF1) as a repressor of Bel1/Tas-mediated trans -activation [ 21 ]. This repression was due to the fact that the specific family members NF1-C and -X interacted with parts of the IP.BRE and its flanking sequences. Since the NF1-mediated repression of the promoter of mouse mammary tumor virus was abrogated by distinct coactivators [ 22 ], we investigated which of the known coactivators and HAT proteins were capable of interacting with the PFV Bel1/Tas activator in the context of the IP.BRE promoter that is absolutely required for virus replication [ 20 ]. Here we report that the Bel1/Tas DNA binding protein functionally interacted with p300 and with the well-known HAT factor PCAF. In addition, SRC-1 enhanced Bel1/Tas trans -activation. This is the first time that these cellular coactivators have been shown to interact with the Bel/Tas1 trans-activator protein. Furthermore, Bel1/Tas binding to the C/H1 domain of p300 and coactivator-driven trans -activation seem to follow a unique pathway. Results Coactivator p300 enhances Bel1/Tas-mediated activation To examine whether p300 (Fig. 1 , first line) enhanced the ability of Bel1 to trans -activate the Bel1 internal promoter (IP.BRE), transient reporter gene assays were performed. The IP.BRE that extends from -1 to -192 of the second cap site of the PFV genome was cloned into the pGL3-pro-luc reporter plasmid [ 21 ]. The results of the luc assays showed that full-length p300FL enhanced Bel1/Tas-mediated activation in an apparently nonlinear fashion (Fig. 2 , upper left panel). To monitor the expression level of the p300 protein, Western blot analysis was carried out in parallel with increasing concentrations of the coactivator at fixed concentrations of the pbel1s expression plasmid that carries the retroviral trans -activator under the control of the CMV-IE promoter. The results of immunoblotting shown in the lower left panel of Figure 2 revealed that Bel1/Tas was expressed at similar levels, as expected, and p300FL expression levels proportional to the input. In parallel experiments, truncated p300 forms, p300N and p300M, were also assayed and yielded moderate levels of enhancement of Bel1/Tas-mediated activation lower than those of p300FL (Fig. 3 ). Unexpectedly, the most extensive level of enhancement of Bel1/Tas-induced activation was reached with the C-terminal region p300C (Fig. 2 , upper right hand panel). Again, Western blot analysis of p300C showed that expression levels of p300C protein increased proportionately to the transfected plasmid DNA while Bel1/Tas expression levels were unchanged (Fig. 2 , lower right panel). The precise boundaries of the three p300 versions used are shown in Figure 1 . The p300 bands marked by arrows are likely due to modified p300 proteins that are known to be modified by phosphorylation, acetylation and sumoylation (2, 5, 26). We next sought to determine whether the enhancing effect of p300 on Bel1/Tas-mediated activation was due to a physical interaction with the Bel1/Tas protein. Figure 2 Coactivator p300 enhances Bel1/Tas-mediated activation . Transient expression assays were performed with pGL3-luc plasmids containing the internal PFV BRE promoter (-1 to -192) after transfection of the pCMV-bel1s expression plasmid alone or separate cotransfection with p300FL and p300C expression plasmids [4]. Normalized luc activities are shown as fold activation in upper panels (for details, see Methods). Expression levels of Bel1/Tas and p300 proteins after cotransfection of 293T cells with increasing p300FL and p300C DNA concentrations and pbel1s DNA of 1.0 μg. Aliquots of the cellular lysates used for luc assays shown in upper panel were in parallel subjected to immunoblot analysis with monoclonal antibody against the FLAG epitope of the p300FL protein (lower left panel), and against the p300C protein (lower right hand panel). Polyclonal antibody against Bel1 was used for Bel1/Tas expression (bottom lanes in lower panels). Figure 3 Reporter gene assays of shortened p300 expression plasmids . Transient expression assays were done as described under Fig. 2 except for that cotransfections were carried out with p300N and p300M plasmids; for boundaries of p300, see Fig. 1. Bel1/Tas interacts with p300 in vivo To examine whether the coactivator p300 interacts with the retroviral activator Bel1/Tas, binding of p300FL to the Bel1/Tas protein was analyzed by immune precipitation. 293T cells were cotransfected with the full-length p300FL and pbel1s expression plasmids and metabolically labeled with [ 35 S]-Methionine and -Cysteine. Cellular lysates were precleared and subjected to immunoprecipitation with an antibody directed against the Bel1/Tas protein except for that in lane 1 (Fig. 4 ). After separation by SDS-PAGE and exposure, the resulting autoradiogram showed that the Bel1-specific antibody had effectively precipitated the p300FL-Bel1/Tas protein complex at both p300FL DNA concentrations of 10 μg (Fig. 4 , lane 3) and 20 μg (lane 6). In the control where pbel1s was omitted the immune precipitation did not reveal any band comparable in size to p300 (lane 4). In the immunoprecipitation shown in lane 1 instead of an antibody against Bel1/Tas, an antibody against p300 was used and showed that the bands marked in Fig. 4 was p300 (lane 1). Figure 4 Direct physical interaction of Bel1/Tas with p300 determined by co-immunoprecipitation . Bel1/Tas binds to p300 in vitro and in vivo . After transfection of 293T cells with expression plasmids pbel1s and p300FL, cells were labeled with [ 35 S]-Methionine plus [ 35 S]-Cysteine. Cellular extracts were precipitated, separated by SDS-PAGE and exposed. Protein p300FL is marked by bold arrowhead (lanes 1, 3, and 6). In the control, pbel1s was omitted (lane 4). Arrowheads indicate the following protein size markers (M, lanes 2 and 5): myosin of apparent molecular mass of 236, phosphorylase b of 97, and BSA of 66 kDa. To check the data obtained, co-immunoprecipitations were performed with non-labeled 293T cells followed by Western blot analysis. Cellular lysates were prepared from 293T cells separately cotransfected with pbel1s and each one of the three truncated p300 expression plasmids, p300N, p300M, or p300C and analyzed as described above. A polyclonal antibody directed against the Bel1/Tas protein was used in the co-immunoprecipitation followed by immunoblotting with a monoclonal antibody directed against the FLAG epitope fused in-frame with the N-terminus of the three truncated p300 versions. The results shown in Figure 5 indicate that p300N protein specifically interacted with Bel1/Tas (lane 1) whereas the middle and C-terminal regions, p300M and p300C, respectively, do not seem to bind to Bel1/Tas under the conditions used (Fig. 5 , lanes 3 and 5). This experiment was repeated several times and yielded the same result. In the controls, pbel1s was omitted in the cotransfections (lanes 2, 4, and 6). Figure 5 Analysis of physical interaction of Bel1/Tas with truncated forms of p300. After cotransfection of 293T cells with 2 μg pbel1s, and, separately with 2 μg each of p300N, p300M, and p300C expression plasmids, cellular extracts were prepared in parallel, divided into equal parts, and analyzed by co-immunoprecipitation and Western blotting. Lysates were precipitated with anti Bel1/Tas antibody and the Western blot was developed with a monoclonal anti-FLAG antiserum. The three arrows in the middle panel mark the correct sizes of the p300N, p300M, and p300C proteins (two blots pasted together). Controls for each p300 plasmid without pbel1s are in lanes 2, 4, and 6. Lower part confirms that the p300N protein specifically binds the Bel1/Tas protein (lane1, marked by arrow); two blots pasted together. Mapping of the p300-Bel1/Tas interaction domain We next determined which N-terminal p300 domain was responsible for the specific interaction with the Bel1/Tas protein. Different truncated versions of p300N (Fig. 1 ) were prepared, cloned, and subjected to separate immunoprecipitations and Western blot analyses as described above for p300N. In addition, two different GST fusion proteins that contained either the C/H1 or KIX domain were bacterially expressed, purified, and analyzed (Fig. 1 ). Two different expression plasmids p300N-C/H1-ΔKIX that lacked the KIX but still expressed the C/H1 domain were capable of binding Bel1/Tas (Fig. 6 , right panel, lanes 2 and 3). In contrast, the results shown in Fig. 6 , lanes 4 and 5, revealed that the plasmid p300-ΔC/H1-ΔKIX that lacks both the C/H1 and KIX domains but retains the short N-terminal region of 196 amino acids did not bind Bel1/Tas. The expression levels of the five constructs were monitored by immunoblot analysis and exhibited the expected bands of p300-derived proteins (Fig. 6 , left panel). Figure 6 Mapping of the p300 Bel1/Tas-interaction domain . The p300 expression plasmids, constructs no. 2, and 5 to 8 shown in Fig. 1 were analyzed by immunoprecipitation as described above for p300N. The right hand panel shows the results of immunoblotting obtained after reaction with the monoclonal antibody against the FLAG epitope; arrows and asterisks mark the Bel1/Tas-interacting protein bands from cellular extracts of pbel1s-cotransfected cells (lanes 1–3); intentionally overexposed to visualize the marked bands in lanes 1 and 2. The left panel presents the Western blots of the five recombinant p300N plasmids used for separate co-immunoprecipitations. Numbers in brackets refer to Fig 1. To unambiguously demonstrate that the C/H1 domain of p300 is the Bel1/Tas-interacting region, we performed pull-down assays with two pGST-C/H1 and pGST-ΔC/H1-KIX fusion proteins (Fig. 1 , constructs no. 9 and 10). The results revealed that purified pGST-C/H1 clearly interacted with Bel1/Tas as shown in Fig. 7 , lane 2 whereas the pGST-ΔC/H1-KIX and a control GST plasmid did not (lanes 1 and 3). Figure 7 Bel1/Tas interacts with GST-C/H1 fusion proteins detected by GST pull-down assays . The recombinant GST-C/H1 and GST-ΔC/H1-KIX proteins were expressed in E. coli BL21 cells and purified by binding to glutathione Sepharose 4B. Each GST-fusion protein bound to glutathione Sepharose 4B was separately mixed with lysates obtained from pbel1s-transfected 293T cells. After incubation and extensive washing with the binding buffer, bound proteins were eluted, separated by SDS-PAGE, and visualized by staining (upper panel); immunoblotting was carried out with an antibody against Bel1/Tas protein (lower panel). To summarize this part, our data show that p300 physically interacted with Bel1/Tas in vivo , and that the C/H1 domain of p300 was responsible for this interaction at least in vitro . Effect of the Bel1-C/H1 domain on Bel1/Tas-mediated activation by p300C To gain more insight into the mechanism of the C/H1-Bel1 complex that affects3 p300C-mediated activation, cotransfections of pbel1s and p300C with expression plasmid p300N-C/H1-ΔKIX were carried out (Fig. 1 , construct 6). Cellular lysates of 293T cells were prepared and luciferase assays performed. Cotransfections of fixed concentrations of the pbel1s, 0.5 μg, with the pC/H1-ΔKIX expression plasmid did not enhance p300C-mediated trans -activation (Fig. 8 , lanes 5, 7, 9, 11, and 13). In contrast, the expression of the C/H1 domain resulted in a partial suppression of the p300C-driven activation at higher C/H1 domain concentrations (Fig. 8 ). Similar degrees of inhibition were obtained when lower Bel1/Tas concentrations were used. Western blot analysis was carried out in parallel with increasing concentrations of the coactivator to ascertain Bel1/Tas expression (data not shown). The inhibition by the C/H1 domain explains why the level of trans -activation of p300N and p300FL did not reach the full extent of p300C-driven activation. Figure 8 Effect of p300-C/H1 domain expression on enhancement of p300C-mediated trans -activation. Transient expression assays with pGL3-luc containing the internal BRE promoter after cotransfection of the p300C and pC/H1 expression plasmids with pbel1s. PCAF interacts with Bel1/Tas We next analyzed whether different HAT-expressing genes such as GCN5, PCAF, and SRC-1 were able to enhance and interact with Bel1/Tas. Transient luciferase gene assays with pGCN5 expression plasmids did not affect Bel1/Tas-mediated activation (data not shown). In contrast, when the HAT PCAF expression plasmid was used for co-expression, an enhancement of Bel1/Tas-induced activation was detectable at a concentration of 0.02 μg PCAF DNA and 0.5 μg pbel1s (Fig. 9 , upper panel). At higher PCAF DNA concentrations, repression of Bel1/Tas-mediated activation was observed. When the levels of PCAF and Bel1/Tas protein expression was determined by Western blot analysis, a decreased level of the Bel1/Tas expression was detected that was likely due to degradation (Fig. 9 , lower panel). The band of the PCAF protein corresponded to the correct size of about 95 kDa (marked by arrow). Figure 9 Functional and physical interaction of Bel1/Tas with PCAF . A , Transient expression assays with pGL3-luc containing the internal BRE promoter after cotransfection of the pCI-FLAG-PCAF with 0.5 μg pbel1s expression plasmid (upper panel). Aliquots were subjected to Western blot analysis (lower panel) We next analyzed the potential interaction between the Bel1/Tas and PCAF proteins by carrying out co-immunoprecipitation with 293T cellular lysates that had been cotransfected with 2.0 μg pbel1s and pCI-PCAF expression plasmids, followed by immunoreaction with an antibody against Bel1 and subsequent immunoblotting with an anti-FLAG antibody to detect PCAF. The results showed that PCAF did indeed interact with the PFV Bel1/Tas activator (Fig. 10 , lane 2). As control, an immunoprecipitation and Western blot analysis was performed with the PCAF plasmid in the absence of pbel1s (Fig. 10 , lane 1). The result showed that a PCAF band was not detectable under the conditions used. Figure 10 Physical interaction of PCAF with Bel1/Tas. Co-immunoprecipitation and immunoblot analysis of pCI-FLAG-PCAF-cotransfected 293T cells with pbel1s DNA (lane 2). The cellular lysates were incubated with an anti Bel1/Tas antibody and treated as described in the legend to Fig. 5 except for that an antibody directed against the FLAG epitope of PCAF was used for immunoblotting. In the control, cotransfection was done without pbel1s (lane 1). To determine if an additional HAT protein, SRC-1, affected Bel1/Tas-mediated activation, transient reporter gene expression assays were carried out after cotransfection of 293T cells with the coactivator SRC-1a and 0.5 μg pbel1s. Surprisingly, a relatively strong enhancement of Bel1/Tas-mediated activation was detected (Fig. 11 , upper panel). The level of expression of both SRC-1a and Bel1/Tas proteins was determined and found to be approximately proportional to the input (Fig. 11 , lower panel). Figure 11 Coactivator SRC-1 enhances Bel1/Tas-mediated trans -activation . Reporter luc gene expression assays with pGL3-luc plasmids containing the internal PFV BRE promoter after cotransfection of pCR3.1-FLAG-SRC-1a with the pbel1s expression plasmid (upper panel). Immunoblotting of 293T cellular extracts cotransfected with the SRC-1a and pbel1s expression plasmids (lower panel). To assess whether SRC-1a was able to physically interact with the Bel1/Tas protein, immunoprecipitations and Western blot analysis were carried out under different conditions. The specificity of SRC-1a was ascertained by using monoclonal antibody directed against SRC-1a in control reaction. However, evidence for an interaction between the SRC-1a protein and Bel1/Tas was not obtained. Discussion It was previously reported that Bel1/Tas is capable of inducing the expression of many cellular genes [ 23 ]. While it is known that Bel1/Tas binds directly and to a large number of DNA target sites [ 18 - 21 ], the mechanism of activation and the identity of the cellular partner molecules of Bel1/Tas remained unknown. As a first step, we have sought to identify the cellular proteins that interact with the PFV retroviral trans -activator and mediate its activating potential. The data presented here show that the coactivators p300 and HAT PCAF physically bound Bel1/Tas in vitro and both enhanced Bel1/Tas-mediated activation whereas SRC-1 enhanced with Bel1/Tas activation without direct binding. According to our data, Bel1/Tas specifically interacted with the C/H1 domain of p300, although we cannot rule out binding to other p300 domains with much lower affinity not detectable under the rather harsh conditions of co-immunoprecipitation used here. When the levels of the relative luciferase activity of p300FL and its three shortened versions are compared, it is noteworthy that p300C reached the highest level of enhancement of Bel1/Tas-mediated activation (Fig. 2 ). Besides its intrinsic HAT activity, p300C contains both the intact C/H3 and Q domains that interact with the HAT enzymes PCAF and SRC-1, respectively [ 10 , 24 , 25 ]. These three HAT enzymes are likely responsible for the large enhancement of the observed trans -activation either directly by acetylation of Bel1/Tas or indirectly by histone acetylation, or both. In contrast to p300C, p300N and p300M do not possess any HAT activities nor do they bind to HAT-containing interaction partners. On the other hand, it is well known that p300 and its three subregions bind a plethora of various partner molecules leading to either activation or repression of transcription. Since the high level of p300C-mediated activation was partially inhibited in cotransfections with C/H1 and pbel1s, competition for Bel1/Tas between the C/H1 and the p300C-terminal interaction partners PCAF and SRC-1 cannot be ruled out so that both direct and indirect mechanisms might be responsible for the relative increase in Bel1/Tas-mediated activation by p300C. When Bel1/Tas binds to the C/H1 domain, the degree of Bel1/Tas acetylation may be much lower, since the protein surface of Bel1/Tas may be occluded and, hence less accessible. Some residual Bel1 acetylation might still occur by endogenous p300 and PCAF. Other factors might play additional roles. The ability of full-length p300 to trans -activate Bel1/Tas was relatively low for two reasons. First, the transfection efficiency of the full-length p300 is very low because of its large plasmid size, and the concentration of endogenous p300 is limiting. Secondly, the activation loop of p300 HAT is not fully activated by auto-acetylation as required for full trans -activation [ 26 ]. Since the relative activation by SRC-1 was not as high as that of p300C, we consider the HAT activity of PCAF as one of the major players of Bel1/Tas-mediated trans -activation. This result is supported by the observed enhancement of Bel1/Tas activation after cotransfection with PCAF that resulted in higher levels of Bel1 acetylation thereby leading to increased binding to the IP.BRE [J. Bodem, personal communication]. Thus, the observed high level of enhancement of p300C might be due to the synergistic effects brought about by formation of ternary p300C-PCAF-Bel1 and binary p300C-SRC complexes, respectively. In these multimeric protein complexes, Bel1/Tas binds p300C indirectly through PCAF. Consistent with the HAT activities of PCAF and p300, we detected acetylated Bel1/Tas in pbel1s-transfected 293T cells using monoclonal antibody against acetyl-Lysine after immunoblotting (our unpublished data). It is intriguing that the distribution of the closely spaced Lysines of Bel1/Tas apparently mimics the correspondingly spaced Lysines in histones. This observation is further complicated by our observation that cotransfection with higher levels of PCAF led to a reduced stability of Bel1/Tas (Fig. 9 ). We assume that PCAF acetylates or even hyper-acetylates the Bel1/Tas protein at closely spaced Lysines in analogy to other activators reported previously [ 9 ]. The decreased stability of acetylated Bel1/Tas might indicate that modified Bel1/Tas is less stable than the unmodified form. This observation adds an additional layer of combinatorial regulation to Bel1/Tas-mediated trans -activation. It is intriguing that some viral trans -activators interact with more than a single p300 domain [ 10 , 14 ]. However, Bel1/Tas might recruit a second interacting region of p300 through binding PCAF (Fig. 10 ) that is known to interact with a p300 domain different from the C/H1 domain (Fig. 1 ) [ 10 ]. The complex nature of p300-Bel1/Tas interactions reported here might serve to strengthen the overall binding affinity between Bel1/Tas and the PCAF interaction domain of p300 within a larger transcriptional complex [ 27 ]. PCAF is known to specifically acetylate distinct Lysine residues of a subset of core histones and thereby regulate the transcriptional activity of many genes depending on the genetic context. It is well documented that acetylation, methylation and other covalent histone modifications are essential signals for the regulation of transcription [ 28 ]. It remains to be seen whether the stronger level of enhancement of p300C is a special if not unique feature of Bel1/Tas activation and due to over-expression or to repressive effects of other p300-interacting protein factors that cannot bind to the truncated p300 protein. Alternatively, many other factors were reported to bind to the C-terminal domains of p300 that also encompass general transcription factors TBP and TFIIB proteins that might also be responsible for the enhancement observed here [ 1 - 3 ]. In search of viral and cellular activators that are comparable with the ability of Bel1/Tas to interact with the C/H1 domain of p300, we found one case. A report indicates that EBNA-2 protein shares many features with Bel1/Tas that include the C-terminal acidic activation domain as well as the abilities to bind both the C/H1 domain and PCAF [ 12 ]. There remain two differences, however. First, EBNA-2 binds to both the C/H1 and the C/H3 domain, and, secondly, PCAF does not coactivate EBNA-2 in strong contrast to Bel1/Tas [ 12 ]. Of note, an additional HAT enzyme, GCN5, did not interact with Bel1/Tas when tested in reporter genes assays indicating that only distinct HAT sets such as those identified in this report specifically interact with Bel1/Tas in trans -activation. The precise roles of the HAT activities of PCAF, p300, and SRC-1 during Bel1/Tas-mediated trans -activation remain to be addressed in future studies. Conclusions Coactivators PCAF and p300 were identified to physically and functionally interact with the spumaviral Bel1/Tas trans -activator. Coactivator SRC-1 was found to strongly enhance Bel1/Tas-mediated trans -activation. The C/H1 domain of p300 was responsible for binding the retroviral activator and found to partially inhibit the p300-driven trans -activation. Methods Antibodies Mouse monoclonal antibodies directed against the FLAG epitopes of p300FL, p300N, p300M, p300C, and PCAF were purchased from Sigma, rabbit polyclonal antibodies against the N-terminal p300FL, the C-terminal p300C, and SRC-1a from Santa Cruz Biotechnology. The polyclonal serum direct against Bel1/Tas was used as described previously [ 23 ]. Typically, 5 μl of each antiserum was used for each immunoprecipitation. Monoclonal antibody against acetyl-Lysine was purchased from Sigma. Plasmids, cells, transfections, and metabolic labeling Plasmids pUC18, pCMVβ-gal, pbel1s [ 29 ], pGL3-pro-IP. BRE (-1 to -192), pCI-FLAG-p300FL, pCI-FLAG-p300N, pCI-FLAG-p300M, pCI-FLAG-p300C [ 4 ], and pCI-FLAG-PCAF [ 29 ] were separately, or in the combinations indicated, transfected into 293T cells using Lipofectamine 2000 (Invitrogen). In general, unless otherwise indicated, 0.1 – 10 μg plasmid DNA were transfected into 293T cells grown in Petri dishes with a diameter of six cm. Full-length pCI-FLAG-300FL plasmids and three different shortened versions (Fig. 1 , constructs no. 1 to 4) were grown in E. coli , DH5α cells [ 4 ]. The PFV internal promoter was constructed by PCR-mediated amplification of the defined promoter fragments as reported previously [ 19 , 21 ]. Recombinant clone p300N-C/H1-ΔKIX* (construct no. 5, Fig. 1 ), was constructed by first generating two PCR products with pCI-p300FL as template using the sense (s) and antisense (as) primers s1: 5'-CTTATGGTTCACCATATACTCAGAATCC-3', as1: AAACTGGAACCATGCCTGCATTTCTCTTATCACC-3', s2: 5'-GAAATGCAGGCATGGTTCCAGTTTCCAT-3', and as2: 5'-GGAAGGAACTGGCCCTGGTTGGAAGGCTGTTG-3' to amplify the sequence from nucleotide 755 to 1275 and nucleotides 1984 to 2297 fused in-frame. The resulting DNA product of 833 nucleotides was cloned into the pCR2.1 Topo vector and designated as pCR2.1-ΔKIX. An SphI-NotI DNA fragment obtained from pCR2.1-ΔKIX and was inserted into pCI-p300N that had been predigested with Sph I and NotI . The borders of construct no. 5 are shown in Fig. 1 and the expressed recombinant protein had the expected size. p300-C/H1-ΔKIX (no. 6) was constructed by inserting the SphI / NotI DNA fragment from pGEX-C/H1 (35) into pCI-p300N digested with SphI / NotI for expression of residues 1 to 424 of p300. Construct no. 7 (Fig. 1 ) was constructed by digesting pCI-p300N with SphI and re-ligating the larger fragment for the expression of residues 1 through 347 of p300. Finally, p300ΔN-ΔC/H1-ΔKIX was constructed from pCI-p300N by digesting with MunI and re-ligated to express residues 1 to 196 of p300. Bacterial plasmids coding for glutathion-S-transferase (GST) fusion proteins pGST-C/H1 (328–424) and pGST-KIX (436–661) (Fig. 1 ) were constructed from pGEX-6p-2GST-p300 [ 30 ] and grown in E. coli , BL21 cells. pCR3.1-FLAG-SRC-1a was grown in E. coli , JM109 cells [ 24 ]. Human 293T or HeLa cells were cultivated in DMEM medium supplemented with 1% penicillin and streptomycin, 1% Glutamine and 10 % fetal calf serum. Plasmid p300FL was transfected into 293T cells and metabolically labeled with L-[ 35 S]-Methionine plus L-[ 35 S]-Cysteine (spec. act. of 37 TBq/mM, PRO-MIX, Amersham) for 6 hr. Cells were harvested and used for immunoprecipitation as described above. The precipitates were analyzed by SDS-PAGE on 12% gels, dried, and exposed on KODAK Biomax MR1 films. Luc reporter gene expression assays Plasmid pCMV-βgal directing β-galactosidase expression from the CMV-IE promoter was used for normalization of transfection efficiency. Luc reporter gene assays were performed and quantified as described [ 31 ] using a Luminoskan TL Plus luminometer (Labsystems, Frankfurt, FRG). pUC18 vector plasmid DNA was used as carrier DNA to equalize the DNA concentration of each transfection. Cells were harvested 18 h after transfection. The results of luc assays were based upon at least triplicate experiments on multiple independent occasions. Expression levels were monitored by Western blot analysis. Co-immunoprecipitation Immunoprecipitation was performed as described previously with minor modifications [ 32 ]. Lysates of subconfluent cotransfected layer of 293T cells were prepared by first washing the cells with PBS, and subsequently lysed in lysis buffer (150 mM NaCl, 20 mM Tris-HCl [pH 7.5], 1 mM phenylmethylsulfonyl-fluoride containing 1% (v/v) Triton X-100. To inhibit unspecific protease activity, protease inhibitors (Biomol) were added to the lysis buffer. Lysates were precleared with protein A-SepharoseCL-4B (Amersham Bioscience AB, Uppsala). Co-immunoprecipitation of p300FL, p300N, p300M, p300C, SRC-1, and PCAF were performed with rabbit anti Bel1 antiserum [ 23 ]. The immune precipitates were retrieved with protein A-SepharoseCL-4B (Pharmacia) and eluted by boiling. To detect the specific precipitate, immunoblotting was performed with specific antibodies against expressed p300, shortened p300 versions, and SRC-1. For the detection of expressed PCAF, Western blot analysis with monoclonal anti-FLAG antibody was carried out. The immunoprecipitates were washed three times with the lysis buffer and analyzed by immunoblotting on 12% gels for expressed p300FL and on 14% gels for shortened p300 forms, PCAF, and SRC-1a. The experiments were repeated at least three times, specially the immunoprecipitations of lysates of p300C-transfected cells. GST pull-down assays The recombinant GST-C/H1 and GST-KIX proteins were expressed in E. coli BL21/DE3 cells after transformation with the corresponding plasmids. Expressed proteins were purified by binding to glutathione Sepharose 4B resin. Each of the GST-fusion proteins bound to glutathione Sepharose 4B were mixed with lysates obtained from pbel1s-transfected 293T cells. After incubation at 4°C overnight in binding buffer [ 30 ] and extensive washing with the binding buffer, bound proteins were eluted, separated by SDS-PAGE, and visualized by staining with Coomassie Brilliant Blue according to Ariumi [ 30 ] or detected by Western blot analysis. Immunoblotting Cells were harvested two days after transfection by lysis in 1% SDS and the protein concentration was determined using the DC protein assay (BioRad). Identical amounts of proteins were separated by SDS-PAGE on 12% gels, blotted, reacted with monoclonal serum directed against the FLAG epitope of the four different FLAG-tagged p300 (Sigma), or polyclonal serum direct against Bel1 [ 23 ], and detected by enhanced chemoluminescence. List of abbreviations used CBP, CREB-binding protein; C/H1, 2, and 3, Cys/His-rich domains of p300; FL, full-length; HAT, histone acetyltransferase; LTR, long terminal repeat; luc, luciferase; PCAF, p300/CBP-associated factor; Q, Glutamine-rich domain of p300; SRC, steroid receptor coactivator; p300N, p300M, and p300C, amino-, middle and C-terminal regions of p300; PFV, primate foamy virus; IP.BRE, internal promoter Bel1 response element; TAD, trans -activation domain; NF1, nuclear factor 1; CMV, cytomegalovirus; GST, glutathione-S-transferase. Authors' contributions HB and WM contributed equally to the manuscript. VO carried out the molecular cloning of p300 derivatives. YN participated in the design of the study. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517496.xml
314475	PLoS Biology in Action	Some examples of how the PLoS Biology content is used and a request for feedback from creative users	In the first month after our launch, the PDF of the “monkey-robot” article by Miguel Nicolelis received tens of thousands of downloads. We are not sure who downloaded the paper because we do not ask people to register at our site. We suspect, however, that its popularity is in part due to the widespread media coverage of the article (from Time magazine to Al-Jazeera and Die Zeit ), which demonstrates a thirst for the original scientific paper and makes a strong argument for open access. We were even more surprised by the popularity of the “issue PDF,” a whopping 72 MB file that contains the entire journal, cover to cover. The inaugural issue PDF was downloaded thousands of times within the first few weeks, even though we had printed 30,000 hard copies and distributed them widely. The Sky's the Limit (with Proper Attribution) But who are our users, and what do they do with the content? We know that PLoS Biology articles are used in a variety of educational settings. For example, the Nicolelis article has already been used for several high-school science projects and by a psychology student who compared the original research paper to its media derivatives. And the paper by Joseph DeRisi and colleagues on the malaria transcriptome has served as the basis for a continuing medical education drug discovery class and has been the topic of several undergraduate classes. Under the terms of the Creative Commons Attribution License, not only can PLoS Biology articles be reproduced and distributed without the need to obtain explicit permission, they can also be used for the publication of derivative works. Two PLoS Biology articles have already been entered in the Internet Encyclopedia . The source of the articles is clearly cited; it is also clear that they have been modified ed by the addition of extra links and information and that they are editable by users of the encyclopedia. Although this is an experiment in freely available and editable information, there will also be opportunities for entrepreneurs to produce derivative works with the type of added value that some users might wish to pay for. Open access provides free access to the research literature, but also provides publishers with new commercial opportunities. Once a significant body of full-text literature is available, it also becomes possible to use it for the development of new tools and resources for text- and data-mining and knowledge discovery. We plan to collaborate with developers of such tools. For their use—and for anybody else who likes their text “marked-up”—we make the XML version of our articles available. These are formatted according to the Journal Publishing DTD (Document Type Definition) from the United States government's National Library of Medicine, which provides a standard for archiving and exchanging XML versions of published documents. Overcoming Obstacles A barrier for many potential users is that all our content—at least for now—is in English. We are delighted to hear that some of it is already being translated into other languages for local use. The feature article on the environmental benefits and risks of genetically modified crops, for example, will be republished (in Spanish) in the Argentinian environmental magazine Gerencia Ambiental . We hope that this will catch on—and urge anyone translating our content to let us know so that we can point others to the various language versions. We merely ask that the translators and their publishers acknowledge the authors and the source, by including a statement such as “this is a translation from the original article by Virginia Gewin published in PLoS Biology , DOI: 10.1371/journal.pbio.0000008 .” Almost any translation is better than none for those excluded by language barriers, but quality control is a concern, and we are keen to collaborate with individuals or organizations who are interested in providing high-quality translations for some or all of our content on a regular basis. Internet connectivity is another obstacle. We know that some of the downloads of the issue PDF were transferred onto CD-ROMs that were copied and distributed in Uganda and Cambodia, areas in which Internet access is often slow and expensive. Other copies of the PDF were being used to create local hardcopies of the journal for communal use. We are happy to support these and related efforts to bring PLoS Biology content to readers by, for example, increasing the range of formats available at our Web site. Let us know what would help. Wanted: More Feedback Besides the encouraging Web statistics, we have heard from many individual users since the launch of PLoS Biology . We'd like to hear from even more. Tell us how you use PLoS Biology : your ideas might inspire others. As a way of building on the work and ideas of others, we have added a page on the PLoS Web site ( www.plos.org/creative_uses ) where we list some of the more creative and unusual uses of PLoS Biology . Let us know what you do with PLoS Biology , or what you'd like to do, and we'll see what we can do to make it possible.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314475.xml
512288	Candidate high myopia loci on chromosomes 18p and 12q do not play a major role in susceptibility to common myopia	Background To determine whether previously reported loci predisposing to nonsyndromic high myopia show linkage to common myopia in pedigrees from two ethnic groups: Ashkenazi Jewish and Amish. We hypothesized that these high myopia loci might exhibit allelic heterogeneity and be responsible for moderate /mild or common myopia. Methods Cycloplegic and manifest refraction were performed on 38 Jewish and 40 Amish families. Individuals with at least -1.00 D in each meridian of both eyes were classified as myopic. Genomic DNA was genotyped with 12 markers on chromosomes 12q21-23 and 18p11.3. Parametric and nonparametric linkage analyses were conducted to determine whether susceptibility alleles at these loci are important in families with less severe, clinical forms of myopia. Results There was no strong evidence of linkage of common myopia to these candidate regions: all two-point and multipoint heterogeneity LOD scores were < 1.0 and non-parametric linkage p-values were > 0.01. However, one Amish family showed slight evidence of linkage (LOD>1.0) on 12q; another 3 Amish families each gave LOD >1.0 on 18p; and 3 Jewish families each gave LOD >1.0 on 12q. Conclusions Significant evidence of linkage (LOD> 3) of myopia was not found on chromosome 18p or 12q loci in these families. These results suggest that these loci do not play a major role in the causation of common myopia in our families studied.	Background Myopia is one of the leading causes of vision loss around the world[ 1 ]. In the United States, myopia affects approximately 25% of adult Americans[ 2 ]. Ethnic diversity appears to distinguish different groups with regard to prevalence. Caucasians have a higher prevalence than African Americans[ 3 ]. Asian populations have the highest prevalence rates with reports ranging from 50–90%[ 1 , 4 , 5 ]. Jewish Caucasians, one of the target populations of the present study, have consistently demonstrated a higher myopia prevalence than the general Caucasian population in both U.S. and European population surveys; Orthodox Jewish males in particular show increased susceptibility[ 6 , 7 ]. Despite many decades of research, little is known about the precise molecular defects and abnormal biochemical pathways that result in myopia. Compelling data from familial aggregation and twin studies indicate that susceptibility to myopia is inherited. Several familial aggregation studies have reported a greater prevalence of myopia in children of myopic parents compared to children of nonmyopic parents [ 8 - 12 ]. Several twin studies have demonstrated a very high heritability (estimates ranging from 60 to 90%) for myopia [ 13 - 15 ]. Other recent genetic studies of families with -6.00 D or more of myopia (termed high or pathological myopia) have reported significant linkage to regions on chromosome 18p11.31, 12q21-23, 17q21-22 and 7q36 [ 16 - 19 ]. The 18p candidate region has been confirmed in an independent study of high myopia [ 20 ]. Mutti et al.[ 21 ] examined the hypothesis that families with milder, juvenile onset myopia might show linkage to these same candidate regions. They found no evidence to support such a role in this more common form of myopia but their study was not highly powered in the presence of heterogeneity. Evidence also exists that myopia may be under environmental influences. The rapid increase in the prevalence of myopia over the last several decades suggests that environmental factors are important [ 22 , 23 ]. Furthermore, studies have shown a positive correlation of specific environmental factors, such as nearwork, with myopia [ 24 , 25 ]. It has been postulated that myopia develops in a person who engages in significant periods of sustained nearwork as an adaptive response to achieve better focus for near images[ 26 ]. Interestingly, Cordain et al.[ 27 ] suggest a positive correlation for myopia with increased consumption of carbohydrates, hyperinsulinemia and type II diabetes. Finally, experimental findings from animal studies show that the refractive state of young chicks will adapt to compensate for refractive errors induced by spectacle lenses[ 28 ]. The combination of genetic and environmental influences on the development of myopia suggests that myopia is a complex disorder and should not be classified as a simple Mendelian trait. Further evidence is shown by studies that have reported correlation coefficients for myopia between offspring and parents and between pairs of siblings to lie between 0.07–0.36 [ 29 - 32 ]. Due to the possible complexity of myopia, population isolates offer many advantages for genome-wide mapping studies[ 33 ]. First, they have reduced genetic complexity. Second, the people in most isolates share a common environment and culture. Differences in diet, exercise, sanitary conditions, and exposure to infectious diseases are minimized. A common language and religion usually promote social cohesion. Therefore, some of the environmental noise surrounding complex diseases that are determined by a combination of nature and nurture may be avoided. To avoid some of the complexity in mapping genes for myopia we have collected refractive measurements and DNA samples from Amish and orthodox Jewish families with myopia. The Old Order Amish are mostly rural farmers and craftsmen. They lead a culturally and technologically distinct lifestyle. They are a genetically well-defined founder population with large families and well-documented genealogies [ 34 , 35 ]. Family history records of the Amish in Lancaster County, Pennsylvania, beginning from 1727 are highly preserved[ 36 ]. Other features of this population include a relatively high standard of living, low migratory tendencies, and no practice of birth control, which facilitate the recruitment of large and extended families. The orthodox Jewish families in this study are all of Ashkenazi descent, a population with known founder effects in other common diseases[ 37 ]. This population also has somewhat larger family sizes than average in the US. In this initial report, we describe the design of our study and show that two regions (18p and 12q) previously reported to be linked to high myopia cannot explain the familial aggregation in these families with mostly moderate to milder forms of myopia. We had hypothesized that allelic heterogeneity might exist at these candidate loci such that in addition to highly penetrant alleles for extreme high myopia, there might also exist other susceptibility alleles of (possibly) lower penetrance that produce milder phenotypic forms of myopia. However, we found no strong evidence in support of this hypothesis. Methods Family screening The study protocol adhered to the tenet of the Declaration of Helsinki and was approved by the University of Pennsylvania and the National Human Genome Research Institute, National Institutes of Health institutional review boards. Informed consent was obtained from the subjects after explanation of the nature and possible consequences of the study. The collection of orthodox Jewish individuals was begun by a mass mailing of 3900 letters to all the known orthodox Jewish families living in Lakewood, New Jersey. Questionnaires were sent with letters explaining the study. If willing to participate, individuals completed and returned questionnaires that included their contact and physician information. Second and third mailings went out to individuals who did not respond – either positively or negatively – to the first mailing. The total number of questionnaires returned was 1,310. All Jewish individuals included in the study were of Ashkenazi heritage. Collection of Amish families was done by an advertisement in an Amish newspaper, referrals from local eye doctors in the Lancaster County community and word of mouth. Criteria for entry into the study included the following: 1) Negative history of systemic or ocular disease which may predispose to myopia, 2) negative history of a premature birth, 3) proband must be affected and must have a family history of myopia in either their parents or children, 4) only one parent (as opposed to both parents) of the proband can be myopic. For the Orthodox Jewish population, an individual's myopic status was obtained from the most recent (within 2 years) measurement of refractive error. If not recent, an individual was given a repeat exam by their local eye doctor or one of the study investigators (D.S.). For the Amish subjects, all participants were examined by a study investigator (D.S.) at the Amish Eye Clinic in Strasburg, PA. Amish participants were brought to the study site because they do not have phone access making it difficult to obtain a past history and records. Cycloplegic refractions were done on all individuals less than 40 years of age with one drop each of 1% cyclogyl, 1% mydriacyl and 2.5% phenylephrine. A manifest refraction was performed if an individual was older than 40 years of age. Classification as myopic required at least -1.00D in each meridian of both eyes. Individuals were classified as nonmyopes if they were over 21 years of age and did not meet the above criteria for myopia. Other individuals were classified as nonmyopic if they were 5–10 years old and had ≥ +3.00D in each meridian, 10–18 years old with ≥ +2.00D in each meridian or 18–21 years old with ≥ +0.50D in each meridian. All other individuals were classified as unknown for the trait. This ascertainment protocol resulted in the collection of 40 Amish families and 38 orthodox Jewish families. Of the 40 Amish families, phenotype data were available on 340 persons (170 individuals were affected and 170 were unaffected) but only 323 DNA samples were available to be genotyped. In the 38 Jewish families, phenotype data were available for 313 persons (177 affected, 122 unaffected and 14 of indeterminate phenotype) and DNA samples were available and genotyped for 290 of these family members. DNA extraction and genotyping Peripheral blood was collected from family members. High molecular weight genomic DNA extraction from the blood samples was performed with a kit (Puregene; Gentra Systems, Inc.; Minneapolis, MN, USA). Polymerase chain reactions were performed in a 17.05 ul volume containing 12–320 ng/ul of DNA; 880 uM each of dATP, dCTP, dGTP, and dTTP; 3 mM MgCl 2 ; 10 mM Tris/HCl (pH8.3); 50 mM KCl; 0.6 uM of each primer; and 7.6 units/ul of Taq polymerase. Standard thermocycling was as follows: 94°C for 30 sec., 55°C for 30 sec. and 72°C extension time for 30 sec. Markers used included D12S85, D12S1706, D12S346, D12S78, D12S79, D12S86, D18S59, D18S481, D18S63, D18S452, D18S53, and D18S474 located in the 18p and 12q regions implicated in high myopia [ 16 , 17 ]. Power studies A simulation study was conducted on the first 44 Ashkenazi Jewish families collected in this study, using the computer program SIMLINK [ 38 , 39 ], to compare the projected power from alternative parametric trait models (five of these families contributed no information about linkage and so were not genotyped and the sixth family was dropped after genotyping because of sample problems that resulted in inadequate linkage information). It was assumed that the myopia trait is controlled by an autosomal dominant bi-allelic locus and the frequency of the high risk allele was varied in different simulations, using both 0.05 and 0.01. The actual observed pedigree structures, trait phenotypes and DNA sample availability were used to simulate the trait locus genotypes and linked and unlinked marker loci were also simulated. A highly polymorphic marker locus (9 equally frequent alleles) was assumed. The power available from these families to detect linkage was evaluated using different models for penetrance and sporadic rates. For each of the 12 models tested, simulations were performed assuming that the underlying proportion of families linked to the same marker locus (α) was 25%, 50%, 75% and 100%. Furthermore, for each model at each specified level of α, simulations were performed for six recombination distances (θ) between the disease and the marker loci (i.e., θ = 0.01, 0.05, 0.1, 0.2, and 0.5); for three maximum penetrances (0.6, 0.7 and 0.8) for gene carriers; and for two phenocopy rates (0.08 and 0.15). LOD scores assuming homogeneity were calculated for each of 100 replicates. The average LOD score over all replicates (ELod) and its standard deviation were calculated for each model simulated. The power of these families to detect a linkage (i.e., to obtain a LOD score ≥ 3.0) was tabulated for the linked marker and the probability of obtaining a LOD score greater than 1.0 when no linkage exists (Type I error) was tabulated for the unlinked marker in all simulations. Linkage analysis The data on 40 Amish and 38 Ashkenazi Jewish families were checked for misspecification of family structures, data entry errors and genotyping errors using the program SIBPAIR[ 40 ]. This program was also used to estimate allele frequencies at marker loci from the unrelated founder individuals in the families. Parametric two-point linkage analysis was performed with the MLINK program of the FASTLINK package [ 41 , 42 ] and the utility programs MAKEPED, Linkage Control Program, and Linkage Report Program from LINKAGE 5.1 [ 43 - 45 ]. Intermarker distances (Kosambi cM) of the microsatellite markers were obtained from the Marshfield database : D12S85-42.78-D12S1706-0.53-D12S346-7.22-D12S78-13.44-D12S79-9.23-D12S86; D18S59-6.94-D18S481-1.36-D18S63-10.4-D18S452-22.54-D18S53-30.08-D18S474. To carefully explore the possibility of linkage of common myopia to these high myopia candidate regions, we utilized 12 different parametric models (Table 1 ). Analyses were performed assuming all combinations of three different frequencies for the myopia susceptibility allele (0.0133, 0.5 and 0.10) and four different sets of genotypic penetrances for the gene carriers and non-gene carriers, respectively: 0.90 and 0.0; 0.80 and 0.0; 0.80 and 0.05; and 0.60 and 0.15. Models 1–4 (Table 1 ) assume an allele frequency for the putative myopia susceptibility allele of 0.0133, which is the same value used by Young et al. in their linkage studies of high myopia [ 16 , 17 ] and close to the value of 0.01 that showed good power in our power simulation (note that a more frequent allele frequency of 0.05 resulted in similar but always lower predicted power in our simulations than the power obtained when an allele frequency value of 0.01 was used; note also that this allele frequency applies only to the linked trait locus, so that if there are multiple loci and environmental factors involved in causing myopia under a heterogeneity model, any single locus might only account for a small proportion of all myopia cases). No sex difference was assumed in any of these models. All persons younger than age 5 were coded as unknown for the trait. This analysis assumed autosomal dominant inheritance of a myopia susceptibility allele. Recombination fractions were assumed to be equal in men and women. The program HOMOG[ 46 ] was used to test for evidence of heterogeneity in the presence of linkage in the two-point parametric linkage analyses. The heterogeneity testing was performed separately in the Jewish and Amish families and also in a joint analysis of LOD scores from the two datasets combined. Multipoint parametric and nonparametric linkage analyses were performed with the GENEHUNTER program[ 47 ]. Because of program memory constraints, one large Amish pedigree was split into three small ones for the GENEHUNTER analysis. The parametric analyses in GENEHUNTER used the same models described above, while allowing for locus heterogeneity. The nonparametric statistic NPL all , which estimates the statistical significance of alleles shared IBD between all affected family members, was calculated also, together with an estimated P value for the Amish and Jewish datasets separately. A nonparametric analysis combining the Amish and Jewish families was then performed by calculating the sum of NPL scores for each family (obtained in the separate Amish and Jewish analyses just described) divided by the square root of the total number of families (N = 78)[ 48 ] to obtain an overall combined NPL score. Table 1 Different parametric models utilized for the linkage analysis Model Allele Frequency Penetrance in DD:Dd susceptibility allele carriers Penetrance in dd normal homozygotes 1 0.0133 0.90 0.00 2 0.0133 0.80 0.00 3 0.0133 0.80 0.05 4 0.0133 0.60 0.15 5 0.05 0.90 0.00 6 0.05 0.80 0.00 7 0.05 0.80 0.05 8 0.05 0.60 0.15 9 0.10 0.90 0.00 10 0.10 0.80 0.00 11 0.10 0.80 0.05 12 0.10 0.60 0.15 Results Power simulation As expected, the estimated average maximum LOD score decreases with the distance between the linked marker locus and the trait locus, and with increasing heterogeneity. However, only minimal changes in projected power for our Ashkenazi families were observed as penetrance, phenocopy rate and disease allele were varied. Projected power was always higher when an allele frequency of 0.01 was used for the susceptibility allele at the trait locus than when an allele frequency of 0.05 was used; however, these differences in power were very small. Table 2 shows a representative sample of the predicted power results for detecting linkage to a marker 5 cM from the trait locus (the average maximum distance that a trait locus would be from our genotyped markers if it fell within the confines of either of these two candidate regions on 18p and 12q) assuming an autosomal dominant susceptibility allele frequency of 0.01. If all families were linked to one locus, these families were predicted to have 100% power to detect linkage to a marker 5 cM away from the trait locus with a LOD of 3 or more (the ELods were all ≥ 14). As less families were linked to the marker locus (i.e., as genetic heterogeneity increased) the power decreased but was still good (≥ 67%) if 50% or more of the families were linked. Even when only 25% of families were linked to the same locus, the expected LOD score was over 1.0 for all models. Of course, these LOD scores were calculated assuming homogeneity, and it is well known that power can be substantially increased when heterogeneity exists if LOD scores are calculated assuming heterogeneity (HLODs) as we have done in this study. So we would expect our actual power to detect linkage to be even higher than our simulations of heterogeneity predict. Observed Type I error rates were compatible with the nominal Type I error levels for all models. Between 0 and 1% of replicates produced a LOD score > 1 at any test map distance for unlinked markers. Table 2 Power and ELods from 100 replicates of simulated data, dominant susceptibility allele frequency of 0.01 PENETRANCE IN DD:Dd SUSC. ALLELE CARRIERS PENETRANCE IN dd NORMAL HOMOZYGOTE % FAMILIES LINKED 100% 75% 50% 25% Power 1 ELod ± s.d. 2 Power 1 ELod ± s.d. 2 Power 1 ELod ± s.d. 2 Power 1 ELod ± s.d. 2 0.8 0.08 100 14.0 ± 0.3 99 8.9 ± 0.3 82 4.7 ± 0.2 18 1.7 ± 0.1 0.8 0.15 100 14.7 ± 0.3 100 9.1 ± 0.3 73 4.8 ± 0.2 14 1.6 ± 0.1 0.7 0.08 100 14.8 ± 0.3 99 8.5 ± 0.3 77 4.8 ± 0.2 17 1.8 ± 0.15 0.7 0.15 100 15.2 ± 0.3 99 8.8 ± 0.3 70 4.5 ± 0.2 14 1.6 ± 0.1 0.6 0.08 100 15.0 ± 0.3 98 8.6 ± 0.25 67 4.3 ± 0.2 14 1.7 ± 0.1 0.6 0.15 100 14.2 ± 0.3 100 8.4 ± 0.3 73 4.3 ± 0.2 5 1.2 ± 0.1 1 Power = 100 X Proportion of replicate samples that yielded a homogeneity LOD score ≥ 3.0 2 ELod = average homogeneity LOD score over all replicates ± its standard deviation Linkage Parametric and nonparametric LOD scores were calculated for 40 Amish pedigrees and 38 Jewish pedigrees. The six markers on chromosome 12q21-q23 spanned 73 cM, and the six markers on chromosome 18p11.2-p11.32 spanned 70 cM. Markers D12S1706, D12S346, D18S59, D18S481 and D18S63 were previously reported by Young et al. [ 16 , 17 ] as showing evidence of linkage to autosomal dominant high myopia. Under all 12 parametric models, the evidence in favor of linkage to these candidate regions was minimal and this evidence varied only slightly as the assumptions of the trait model were changed across the models. Therefore, only the results from model 1 are presented here. Two-point linkage analyses in the Amish and Jewish populations Results of two-point parametric linkage analysis of myopia assuming linkage heterogeneity to the chromosome 12 and 18 markers in 40 Amish families are presented in Table 3 . Statistically significant or suggestive linkage under locus homogeneity was not observed for either chromosome 12 or chromosome 18. Only one marker, D18S474 showed a two-point LOD ≥ 1.0 (LOD = 1.39 at θ = 0.3). Testing for linkage heterogeneity using HLODs in HOMOG did not significantly improve the evidence for linkage to any of these markers. Table 3 Two-point parametric LOD scores for myopia (Model 1) in 40 Amish Families RECOMBINATION FRACTION, θ MARKER 0.0 0.01 0.05 0.1 0.2 0.3 0.4 D12S85 -19.55 -12.97 -9.02 -6.27 -2.92 -1.15 -0.31 D12S1706 -39.91 -29.33 -17.82 -10.79 -3.85 -1.06 -0.18 D12S346 -36.18 -24.54 -13.68 -7.57 -1.99 -0.14 0.08 D12S78 -40.39 -28.61 -17.98 -11.46 -4.57 -1.5 -0.33 D12S79 -49.21 -32.42 -19.78 -12.36 -4.82 -1.5 -0.2 D12S86 -40.95 -32.39 -20.01 -12.52 -5.14 -1.85 -0.43 D18S59 -26.98 -20.07 -11.74 -6.77 -2.2 -0.48 -0.02 D18S481 -29.61 -19.93 -11.02 -5.99 -1.42 -0.04 0.16 D18S63 -32.32 -23.1 -12.11 -5.98 -0.87 0.5 0.35 D18S452 -38.37 -27.85 -16.16 -9.31 -2.76 -0.35 0.15 D18S53 -35.69 -25.42 -15.08 -9.07 -3.22 -0.77 -0.01 D18S474 -26.64 -14.04 -5.89 -2.01 0.94 1.39 0.7 The same markers on chromosome 12q21-q23 and chromosome 18p11.2-p11.32 were analyzed using 38 Ashkenazi Jewish families (Table 4 ). Statistically significant or suggestive linkage was not observed on either chromosome, no homogeneity LOD scores ≥ 1.0 were observed, and testing for linkage in the presence of heterogeneity (HLODs in HOMOG) did not alter this result. Table 4 Two-point parametric LOD scores for myopia (Model 1) in 38 Ashkenazi Jewish families RECOMBINATION FRACTION, θ MARKER 0 0.01 0.05 0.1 0.2 0.3 0.4 D12S85 -7.5 -6.71 -4.62 -2.56 -0.44 0.19 0.17 D12S1706 -36.21 -28.66 -17.67 -10.77 -3.85 -1.04 -0.15 D12S346 -32.76 -24.46 -13.34 -7.11 -1.68 -0.04 0.11 D12S78 -27.99 -20.59 -10.1 -4.65 -0.46 0.43 0.24 D12S79 -38.53 -30.26 -18.47 -11.21 -4.12 -1.24 -0.26 D12S86 -39.59 -31.68 -19.83 -12.58 -5.12 -1.8 -0.45 D18S59 -35.34 -27.25 -16.82 -10.69 -4.4 -1.51 -0.26 D18S481 -32.69 -24 -14.58 -9.19 -3.73 -1.26 -0.24 D18S63 -36.39 -28.55 -17.92 -11.36 -4.65 -1.63 -0.35 D18S452 -34.97 -25.21 -15.47 -9.72 -3.78 -1.22 -0.23 D18S53 -38.03 -26.36 -14.65 -8.62 -3.15 -0.94 -0.14 D18S474 -29.88 -22.79 -14.41 -9.29 -3.98 -1.45 -0.31 Heterogeneity testing using HOMOG in the combined Jewish and Amish families also did not yield any significant evidence of linkage in these two regions, with the maximum HLOD's being 0.39 and 0.95 on chromosomes 12 and 18 respectively. Furthermore, nonparametric two-point NPL scores did not show any significant evidence for linkage in either the Amish or Jewish populations. The observed combined NPL score of 1.37 for D12S1706 approached nominal significance at p = 0.09 but was not close to the significance level of at least p = 0.01 needed to provide confirmation of a prior linkage[ 49 ]. Multipoint linkage analyses in the Amish and Jewish populations Multipoint parametric linkage analyses assuming homogeneity were consistently negative in both the Amish and Jewish datasets. A maximum multipoint parametric HLOD of 0.92 was observed at D18S474 in the Amish population. However, multipoint parametric HLOD scores were essentially zero for the chromosome 12 region in the Amish and for both the chromosome 12 and 18 regions in the Jewish families. The multipoint nonparametric analyses did not show statistically significant evidence for linkage of myopia to either candidate region in the Amish (Figures 1 and 2 ) or the Jewish (Figures 3 and 4 ) families. Only very mild evidence for linkage of myopia in the Amish was observed between markers D18S59 and D18S481 (NPL= 1.54, p = 0.05) (Figure 2 ). Figure 1 Multipoint nonparametric linkage analysis of myopia to chromosome 12q in 40 Amish families Figure 2 Multipoint nonparametric linkage analysis of myopia to chromosome 18p in 40 Amish families Figure 3 Multipoint nonparametric linkage analysis of myopia to chromosome 12q in 38 Ashkenazi Jewish families Figure 4 Multipoint nonparametric linkage analysis of myopia to chromosome 18p in 38 Ashkenazi Jewish families Individual families showing linkage Only one Amish family (3061) showed marginal evidence for linkage (LOD > 1.0) to the region previously reported on chromosome 12 (D12S1706 and D12S346) for both two-point and multipoint parametric analyses (Table 5 ). Three Amish families gave LOD >1.0 at 2 markers on chromosome 18p. In both two-point and multipoint parametric analyses, family 3064 showed mild evidence of linkage (LOD = 1.30) to marker D18S63, and families 3049 and 3053 showed slight evidence of linkage to marker D18S474 (two-point LOD= 1.14 and 1.30, respectively). Simulations using SIMLINK (model 1) showed that for these individual families, the maximum two-point LOD score obtained when a linked marker was simulated at a recombination fraction of 0.0 ranged from 0.99 to 1.4, and the probability of obtaining a LOD over 1.0 for an unlinked marker ranged from <0.01 to 0.036. The nominal significance level corresponding to a LOD score of 1.0 is approximately 0.01. A total of three Jewish families showed marginal evidence for linkage to chromosome 12q markers for both two-point and multipoint analyses, with two-point and multipoint parametric LOD > 1.0. Simulations using SIMLINK (model 1) showed that for these individual families the maximum two-point LOD score obtained when a linked marker was simulated at a recombination fraction of 0.0 ranged from 0.93 to 1.06, and the probability of obtaining a LOD over 1.0 for an unlinked marker ranged from <0.001 to 0.05. Table 5 Families showing slight evidence for linkage of myopia (Model 1) to chromosome 12q or 18p AMISH FAMILIES FAMILY ID TWO-POINT LOD MULTIPOINT Marker Zmax LOD NPL P value 3061 D12S1706 1.04 1.04 * N/A D12S346 1.04 1.04 * N/A D12S78 1.04 1.04 * N/A 3049 D18S474 1.14 1.27 2.19 0.02 3053 D18S474 1.30 1.30 -* N/A 3064 D18S63 1.30 1.30 -* N/A JEWISH FAMILIES FAMILY ID TWO-POINT LOD MULTIPOINT Marker Zmax LOD NPL P value 20 D12S79 1.12 1.12 3.45 0.02 D12S86 1.12 1.12 3.45 0.02 58 D12S1706 1.16 1.16 3.01 0.06 D12S346 1.16 1.16 3.01 0.06 78 D12S85 1.07 1.07 3.38 0.01 * Only single affected parent-offspring pairs were genotyped in these families so the NPL analysis was uninformative in these families. However, the parametric LOD score analysis that uses both affected and unaffected family members was informative for linkage. Discussion The overall results of these preliminary studies do not indicate any strong evidence for linkage of myopia in these families to the candidate regions on chromosomes 12 or 18. Although some families show marginal evidence of linkage to one of these regions, the results could be due to chance. Our negative results for these candidate regions have several possible explanations. First, the diagnostic criteria used in the previous studies [ 9 , 10 ] that implicated these candidate regions were based on limiting the affection status to the sphere component of a plus cylinder refraction. An individual was considered affected with high myopia if the sphere was equal to or greater than -6D regardless of the astigmatic error. Our study required an individual to have -1D in each meridian to be considered affected. Therefore, the criterion for being affected was quite different between the two studies. Second, the study population in our study included moderate and low myopes in addition to a small number of high myopes. None of our families showed strong aggregation of high myopia. Therefore, there were no families in our study recruited exclusively for high myopia and no families that would have been highly powerful for the detection of linkage to a high myopia trait. We utilized this study design to search for allelic heterogeneity with regard to the 18p and 12q loci thinking that one or both loci may predispose to moderate/mild forms of myopia. Thus, the current linkage analysis was done to test the hypothesis that other alleles at the candidate high myopia loci on chromosomes 18 and 12 might contribute to the etiology of moderate/mild myopia. The mild evidence of linkage in a few families indicates that this hypothesis cannot be fully ruled out for a very small proportion of families with mild forms of myopia. However, there is no strong evidence in favor of this hypothesis and strong negative evidence against linkage in most of the families in this study. Previous studies attempting to confirm the high myopia loci on chromosomes 18 and 12 have yielded inconsistent results. Naiglin et al.[ 19 ]collected 23 French families with high myopia (spherical equivalent ≥ -6D) and performed a genome scan with 400 markers. Significant linkage was not found on 18p and 12q. Lam et al.[ 20 ] mapped 15 families with high myopia, ≥ -6.0D, using only 18p markers. Statistically significant (LOD > 3) linkage was not demonstrated although a multipoint LOD over 2.0 was observed, thus giving evidence of confirmation of the 18p candidate region. Mutti et al.[ 21 ] collected 53 families with varying degrees of myopia (affected ≥ -0.75D in each meridian) and genotyped the family members with 18p and 12q loci markers implicated in high myopia. No evidence of linkage to milder forms of myopia was found to the chromosome 18p and 12q loci previously associated with high myopia. Our study, although consistent with the results of Mutti et al.[ 21 ] was significantly different in design. First, Mutti et al. used a heterogeneous population that could decrease the chances of obtaining significant linkage for a minor gene effect from 18p or 12q if substantial ethnic heterogeneity exists. Both the Amish and Ashkenazi populations used in our study are more homogeneous and each sample was analyzed using marker allele frequencies estimated from the sample. Second, their study utilized 221 samples while we genotyped 613 individuals. Our power simulations predicted higher power in the presence of heterogeneity in our Ashkenazi families than was predicted for the Mutti et al. study. Our Amish families were of similar size and structure and so should have similar predicted power as the Ashkenazi families, and our combined analyses of the two data sets should provide much more power than that predicted by the simulations of the Ashkenazi families alone. The combination of the Mutti et al. study with the results presented here strongly suggest that these two candidate regions do not play a large role in the causation of moderate/mild myopia in several populations examined. These studies suggest that myopia is complex and probably caused by the interaction of multiple genes with the environment. Therefore, to understand myopia it is necessary to apply the equation: Genes + Environment=Outcome. The difficulty here is the uncertainty surrounding both terms in the equation; ideally, one set of genetic factors will interact with one set of environmental influences to produce identical outcomes, but it is unknown whether this is always going to be the case. Therefore, to lessen the problem of multiple gene interaction as well as gene-environment interaction confounding the results, strategies to limit this problem should be utilized in the genetic mapping of myopia. The use of isolated populations is one approach to limiting the heterogeneity across populations and is the approach we are using for a genome wide scan in these families. Furthermore, the definition of myopia needs to be standardized so comparisons across studies can be made accurately. Previous studies have utilized different requirements with regard to affection status making cross comparisons difficult. In conclusion, we find little evidence implicating previously described susceptibility loci for high myopia on chromosomes 12 and 18 as being important in the etiology of common, moderate/mild myopia in our two population samples. Competing interests None declared. Authors' contributions Dwight Stambolian, Lauren Reider, Debra Dana, Robert Owens, and Elise Ciner recruited patients for the study. Melissa Schlifka carried out the genotyping in chromosomes 18 and 12. Dwight Stambolian and Joan E. Bailey-Wilson performed the study design and wrote part of the manuscript. Joan Bailey-Wilson oversaw all statistical analyses; Grace Ibay performed statistical analyses and wrote part of the manuscript; Taura Holmes, Betty Doan and Jennifer O'Neill assisted with analyses of the data. All authors read and approved of the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC512288.xml
503385	Identification and characterization of a nontypeable Haemophilus influenzae putative toxin-antitoxin locus	Background Certain strains of an obligate parasite of the human upper respiratory tract, nontypeable Haemophilus influenzae (NTHi), can cause invasive diseases such as septicemia and meningitis, as well as chronic mucosal infections such as otitis media. To do this, the organism must invade and survive within both epithelial and endothelial cells. We have identified a facilitator of NTHi survival inside human cells, v irulence- a ssociated protein D ( vapD Hi , encoded by gene HI0450). Both vapD Hi and a flanking gene, HI0451, exhibit the genetic and physical characteristics of a toxin/antitoxin (TA) locus, with VapD Hi serving as the toxin moiety and HI0451 as the antitoxin. We propose the name VapX Hi for the HI0451 antitoxin protein. Originally identified on plasmids, TA loci have been found on the chromosomes of a number of bacterial pathogens, and have been implicated in the control of translation during stressful conditions. Translation arrest would enhance survival within human cells and facilitate persistent or chronic mucosal infections. Results Isogenic mutants in vapD Hi were attenuated for survival inside human respiratory epithelial cells (NCI-H292) and human brain microvascular endothelial cells (HBMEC), the in vitro models of mucosal infection and the blood-brain barrier, respectively. Transcomplementation with a vapD Hi allele restored wild-type NTHi survival within both cell lines. A PCR survey of 59 H. influenzae strains isolated from various anatomical sites determined the presence of a vapD Hi allele in 100% of strains. Two isoforms of the gene were identified in this population; one that was 91 residues in length, and another that was truncated to 45 amino acids due to an in-frame deletion. The truncated allele failed to transcomplement the NTHi vapD Hi survival defect in HBMEC. Subunits of full-length VapD Hi homodimerized, but subunits of the truncated protein did not. However, truncated protein subunits did interact with full-length subunits, and this interaction resulted in a dominant-negative phenotype. Although Escherichia coli does not contain a homologue of either vapD Hi or vapX Hi , overexpression of the VapD Hi toxin in trans resulted in E. coli cell growth arrest. This arrest could be rescued by providing the VapX Hi antitoxin on a compatible plasmid. Conclusion We conclude that vapD Hi and vapX Hi may constitute a H. influenzae TA locus that functions to enhance NTHi survival within human epithelial and endothelial cells.	Background Culturable Haemophilus influenzae are acquired in the nasopharynx shortly after birth, and are thought to persist throughout life. H. influenzae adheres to and penetrates into and between cultured human respiratory epithelial cells, a mechanism that may contribute to its persistence in chronic obstructive pulmonary disease (COPD) and cystic fibrosis (CF) patients [ 1 , 2 ]. H. influenzae can be found in the respiratory tracts of these patients even after they have undergone antibiotic treatment [ 3 ]. As well, COPD sputum cultures can be sterile, while H. influenzae can still be isolated from the subepithelial matrix [ 4 ]. Finally, we have found in a recent in vivo study that H. influenzae can persist in a human bronchiolar xenograft for at least three weeks [ 5 ]. This suggests that the organism can survive and persist in protected biological compartment(s). The ability of H. influenzae to survive antibiotic treatment and reappear when growth is favorable may be responsible for the reseeding of the middle ear observed in chronic otitis media. Often, middle ear fluid from children presenting with otitis media with effusion are sterile when cultured, but PCR analysis of the fluid has determined the presence of H. influenzae [ 6 ]. Further, RT-PCR studies of this sterile fluid have shown the presence of bacterial mRNA, confirming that the bacteria are alive and persisting in a viable but nonreplicative state [ 7 ]. Persistence was investigated in vitro using a NTHi strain that was susceptible to β-lactam antibiotics. This strain was allowed to invade a human respiratory epithelial cell monolayer for 24 hours, which was subsequently treated with a 4 hour incubation in 10 × MIC concentrations of the β-lactam antibiotics ampicillin, imipenem, cefuroxim, amoxycillin/clavulanic acid, or cephalothin. The antibiotics killed all the extracellular bacteria, but none of the intra- or paracellular bacteria, suggesting that the organism was not replicating inside or between the epithelial cells [ 8 ]. Non-replicating bacteria are not susceptible to the cidal action of β-lactam and aminoglycoside antibiotics. During a study aimed at identifying genes associated with virulence in pathogenic strains of the Gram-negative, strict anaerobe Dichelobacter nodosus , the causative agent of ovine footrot, Katz et al . [ 9 ] reported the discovery of a novel area of the chromosome that hybridized to all virulent strains tested, but to only 23% of the avirulent strains studied. They designated the four genes found on this fragment as vapA - D , for v irulence- a ssociated p roteins. Homologues of these genes appear on the chromosomes and plasmids of a number of pathogenic microorganisms, including Neisseria gonorrhoeae , Helicobacter pylori, Reimerella anatipestifer and Actinobacillus actinomycetemcomitans . The chromosome of H. influenzae strain Rd KW20 (hereinafter referred to as strain Rd) contains vapA , vapBC , and vapD homologues, with one pair, vapBC , in duplicate. The genome organization of the vap genes in H. influenzae differs from that of D. nodosus , in that vapA Hi (HI1250) is preceded by a conserved hypothetical protein, HI1251, and both genes are likely transcribed as an operon. As well, vapD Hi is flanked by a gene encoding another conserved hypothetical protein, HI0451 (which we have named vapX Hi ), again in an apparent operon configuration (8 nucleotides separate HI0450 and HI0451). To determine if the vap homologues played a role in the persistence of NTHi, we chose vapD Hi (HI0450) for further study, since this protein was found in a proteomic survey to be expressed in the soluble fraction of strain Rd [ 10 ]. VapD Hi is 40% identical and 67% similar to the Dichelobacter VapD and belongs to the Cluster of Orthologous Groups (COG) 3309 and Pfam 04605, termed the "N-terminal conserved domain of VapD". Results Mutation of vapD Hi results in attenuated survival in human endothelial cells When the Rd vapD Hi mutant strain was used to invade the in vitro model of the blood-brain barrier, human brain microvascular endothelial cells (HBMEC) in 12-well plates, the amount of gentamicin-resistant bacteria recovered from the monolayer after three hours declined to approximately 60% of wild type levels, an average of 2.2 × 10 3 CFU/ml for the wild-type strain Rd versus 1.3 × 10 3 CFU/ml for the Rd vapD Hi mutant (n = 3 (number of independent assays performed in at least duplicate); P < 0.05, Student's t test). No significant difference was observed between the wild type and vapD Hi mutant in adherence to the human cell monolayers: the average number of cell-associated bacteria (both adherent and invaded) recovered for strain Rd was 1.5 × 10 5 CFU/ml and 1.8 × 10 5 CFU/ml for the Rd vapD Hi mutant (n = 3; P > 0.05). To determine if the phenotype of attenuated survival observed in the Rd vapD Hi mutant was a general phenomenon and not restricted to strain Rd, another isogenic pair was constructed and analyzed using a different strain, R3001. R3001 is a bronchoalveolar lavage isolate from a pediatric cystic fibrosis patient, and is considered invasive since it came from a normally sterile site [ 5 ]. The average number of gentamicin-resistant bacteria recovered from HBMEC monolayers was 1.2 × 10 5 CFU/ml for the parent strain R3001 versus 7.1 × 10 4 CFU/ml for the R3001 vapD Hi mutant (n = 3; P < 0.05). Although the absolute numbers of bacteria recovered were higher with strain R3001 than with Rd (as is often observed with invasive isolates), the attenuation of survival inside the HBMEC monolayer of ≤ 60% observed in the strain with a vapD Hi mutation was maintained. There was no significant difference between the wild-type R3001 and the R3001 vapD Hi mutant in adherence to the monolayer: the average numbers of cell-associated bacteria recovered for strain R3001 were 1.7 × 10 7 CFU/ml versus 1.3 × 10 7 CFU/ml for R3001 vapD Hi (n = 4; P > 0.05). Unlike Rd, strain R3001 carries the high molecular weight (HMW) adhesins, which may account for its more efficient adherence to the HBMEC monolayer [ 11 ]. No significant difference in the growth rates of either of the vapD Hi mutants versus their cognate parent strains were observed, whether grown in bacteriological media (sBHI broth) or on HBMEC or NCI-H292 monolayers (data not shown). Mutation of vapD Hi results in diminished long-term survival inside human respiratory epithelial cells To determine if the vapD Hi mutation would affect the ability of H. influenzae to survive inside human respiratory epithelial cells over a longer period of time, 18-hour invasion assays were performed using NCI-H292 cells. The number of gentamicin-resistant bacteria recovered from the NCI-H292 monolayer after 18 hours for the parent strain Rd was an average of 6.4 × 10 4 CFU/ml versus 3.2 × 10 4 CFU/ml for the Rd vapD Hi mutant (n = 3; P < 0.05). This represents a 50% reduction in survival of the vapD Hi mutant within epithelial cells as compared to the parent strain, more attenuation than was seen for the three hour assays. Transcomplementation of Rd vapD Hi The vapD Hi locus from strain R3001 was cloned into the mobilizable broad host range plasmid pDD515, creating pDD564, and conjugally transferred into the Rd vapD Hi mutant (Table 1 ). The plasmid pDD515 is a derivative of the IncQ plasmid RSF1010 and has an approximate copy number of 12 per chromosome in H. influenzae [ 12 ]. The survival inside HBMEC of strain Rd (pDD515) was within 5% of strain Rd without the vector in identical assays (data not shown). Carrying a vapD Hi locus in trans restored wild-type survival of Rd vapD Hi (pDD564) within HBMEC monolayers. The amount of gentamicin-resistant bacteria recovered from the endothelial cell monolayer after a three hour invasion assay was an average of 8.0 × 10 2 CFU/ml for Rd (pDD515), the vector control, and 7.6 × 10 2 CFU/ml for the mutant strain that carried the wild-type vapD Hi allele in trans , Rd vapD Hi (pDD564) (n = 3; P > 0.05), indicating that there was no significant difference in the survival inside HBMEC monolayers of the wild type strain versus the transcomplemented strain. These data confirm that the phenotype of attenuated survival was due to the interruption in vapD Hi and not to polar effects. The R3001 vapD Hi mutant mirrored the survival defect seen with the Rd vapD Hi mutant, at approximately 60% of wild-type R3001 levels. However, attempts to conjugate the mobilizable broad host range plasmid carrying the vapD Hi locus, pDD564, into strain R3001 for transcomplementation studies failed repeatedly. This clinical isolate likely has a plasmid or an origin of replication of the same incompatibility group incorporated into its chromosome and therefore will not maintain the broad host range plasmid for transcomplementation [ 13 ]. Reverse-transcriptase PCR of bacteria from human cell monolayers To determine if the vapD Hi locus was transcribed during contact with a human cell monolayer, total RNA was isolated from the wild-type strain Rd recovered after 18 hours on HBMEC or NCI-H292 monolayers and was used as the template for RT-PCR. Figure 1 shows the 153 bp band amplified with the vapD Hi -specific primers, with a molecular weight marker in lane A. The template for lane B is Rd RNA from HBMEC endothelial cell monolayers, lane C is its cognate negative control, with no reverse transcriptase added to the RNA prior to PCR amplification. Lane D shows the results using Rd RNA from NCI-H292 epithelial cell monolayers; lane E is its cognate negative control. The vapD Hi locus is transcribed when strain Rd is in contact with either human epithelial or endothelial cell monolayers. PCR survey of vapD Hi In order to estimate the carriage rate of vapD Hi among the highly heterologous NTHi strains, a PCR survey of 59 commensal and disease-associated strains was undertaken (Table 2 ). The vap HI primer set was used. In Rd, these primers amplify a 769 bp PCR product that includes the full-length vapD Hi gene. Purified chromosomal DNA preparations from 53 randomly-chosen NTHi strains and one each of the six capsular serotypes of H. influenzae (types a through f) from the American Type Culture Collection (ATCC) reference strains described in Table 2 were subjected to PCR with the vap HI forward and reverse primers. The NTHi strains included nasopharyngeal, blood, CSF, middle ear, tracheal aspirate, and sputum isolates. A PCR product was amplified in 100% of the strains. All of the ATCC encapsulated reference strains and Rd displayed a full-length vapD Hi allele. Ninety-three percent of the nasopharyngeal strains carried a full-length allele, as did 71% of the blood and CSF isolates, and 50% of the middle ear, tracheal aspirate, and sputum isolates. Overall, only ten strains of the 59 included in the study displayed a truncated gene. Sequencing of the truncated vapD Hi allele To study the truncated vapD Hi in more detail, five out of the ten alleles that represented the smaller isoform of vapD Hi from the PCR survey were sequenced on both strands. It was found that, in all cases, the gene had undergone a deletion event that had left the protein in frame, but missing 46 amino acids from the interior of the protein, resulting in a 45 amino acid protein rather than the full-length 91 amino acid protein (Figure 2 ). This corresponds to the loss of Rd genome coordinates 473123 to 473263. In addition, all of the smaller alleles had an aspartate residue inserted at position #7 as compared to Rd, which has a leucine at that position. The significance of this is unclear, as the full-length R3001 vapD Hi allele, which did transcomplement the Rd vapD Hi mutant, also has an aspartic acid inserted at the same position, resulting in a 91 amino acid protein. Interestingly, the VapD homologues from N. gonorrhoeae , H. pylori and A. actinomycetemcomitans , as well as VapD in D. nodosus , all have an aspartate at that position. Rd appears to be the only H. influenzae strain studied which lacks that particular residue. Full-length VapD Hi homodimerizes Using an E. coli -based protein-protein interaction system that is dependent upon the DNA-binding domain (DBD) of LexA, homodimerization of identical protein subunits can be quantitated [ 14 , 15 ]. In this system, protein-protein interactions result in a LexA dimer that is active as a repressor, and consequently, the beta-galactosidase activity of the reporter strain (SU101) diminishes. Full-length VapD Hi from strain R3001 was ligated to the DBD of LexA in plasmid pDD559 and the clones were analyzed on MacConkey agar with lactose. If there was no homodimerization of the LexA::VapD Hi fusion protein, the colonies appeared red on MacConkey agar, as the native level of beta-galactosidase expression in the reporter strain was not inhibited. If the subunits interacted, the colonies appeared pale on MacConkey, as the engineered LexA operator controlling the lacZ reporter gene had been repressed by a homodimer of the LexA fusion protein. This repression was quantitated by beta-galactosidase activity assays. Each measurement is the mean of at least three experiments performed in triplicate. It was found that VapD Hi interacted strongly with itself. The beta-galactosidase activity of the reporter strain SU101 carrying the vector control (pSR658) was an average of 975 (± 29) Miller units, and the activity of SU101 with the LexA::VapD Hi fusion (pDD559) was an average of 16 (± 1) Miller units, indicating strong interaction. Full-length VapD Hi forms homodimers in vivo . This protein may also form higher-order multimers, since this would result in a number of dimeric forms being available to act as a repressor of lacZ transcription in the reporter strain. Truncated VapD Hi does not homodimerize, but interacts with full-length VapD Hi Homodimerization assays with the small allele revealed that the subunits of the truncated VapD Hi did not interact efficiently. The vector control for the homodimerization assays SU101 (pSR658) yielded 1490 (± 31) Miller units, and SU101 carrying the LexA fusion to the truncated VapD Hi protein from strain R2866 (pDD577) displayed 1357 (± 54) Miller units of beta-galactosidase activity, showing little repression in this system. Since the wild-type VapD Hi subunits homodimerized strongly but the truncated subunits did not, the truncated subunit and the full-length subunit were examined for interaction in heterodimerization assays. In the reporter strain SU202, a LexA operator with a mutated half-site was engineered upstream of the lacZ gene. This strain was then transformed with two compatible plasmids, one that carried a fusion of the truncated VapD Hi with a wild-type LexA DBD, and one that carried the full-length VapD Hi fusion to a mutated LexA DBD that only recognized the mutated LexA operator half-site in SU202 [ 15 ]. If a heterodimer of a truncated subunit and a full-length subunit was formed, a functional LexA repressor could recognize the hybrid operator and repress transcription of lacZ . It was determined that the truncated and full-length subunits could interact with each other. The vector control for the heterodimerization assays (SU202 with pSR658 and pSR659) yielded 1855 ± 196 Miller units, and SU202 carrying the LexA fusion to the truncated VapD Hi protein from strain R2866, pDD577, coupled with the mutated LexA DBD fusion to the full-length VapD Hi protein from strain R3001, pDD561, resulted in 792 ± 19 Miller units of beta-galactosidase activity, showing that the truncated subunit did heterodimerize with the full-length VapD Hi subunit. Truncated VapD Hi does not transcomplement the mutant and has a dominant-negative effect in the wild-type strain To investigate whether the truncated VapD Hi protein could transcomplement a mutation in the full-length gene, the truncated locus from strain R2866 was cloned into the mobilizable broad host range vector pDD515, creating pDD594, and conjugally transferred into the Rd vapD Hi mutant strain. Three-hour survival assays using HBMEC monolayers were performed, and the number of gentamicin-resistant bacteria recovered from the monolayer was an average of 5.6 × 10 2 CFU/ml for Rd (pDD515), the vector control, versus 1.9 × 10 2 CFU/ml for Rd vapD Hi (pDD594) (n = 3; P < 0.005). Survival in HBMEC by the mutant strain did not increase to wild-type levels, as was the case with the Rd vapD Hi mutant transcomplemented with a full-length vapD Hi allele on pDD564. Since subunits of the full-length VapD Hi and the truncated protein interacted, the plasmid pDD594 carrying the truncated allele was conjugally transferred into wild-type strain Rd to determine whether expression of the small protein would interfere with the function of the wild-type VapD Hi protein. The strain was used in three-hour assays of HBMEC monolayers, and it was found that Rd (pDD594) was attenuated in human cell survival as compared to Rd (pDD515), the wild-type strain carrying the vector without an insert. The average number of gentamicin-resistant bacteria recovered from the monolayer for Rd (pDD515), the vector control, was 5.7 × 10 2 CFU/ml versus 1.8 × 10 2 CFU/ml for Rd (pDD594) (n = 3; P < 0.005). The in trans expression of a truncated vapD Hi allele in the wild-type strain Rd resulted in a dominant-negative effect on survival within HBMEC monolayers. Expression of vapD Hi in Escherichia coli DH5α results in cell growth arrest To test the hypothesis that VapD Hi constituted the toxin, and that VapX Hi encoded the antitoxin portion of a TA locus, both proteins were expressed in an E. coli background. E. coli does not contain a homologue of either protein. Initially, the vapD Hi gene, HI0450, was cloned into the pTrcHisA vector (Invitrogen, Carlsbad, CA). This resulted in vapD Hi being under the control of the strong P trc promoter, which is repressed for the most part until induced by IPTG. This plasmid was designated pDD560. Both the vector control, DH5α (pTrcHisA), and DH5α (pDD560) were grown to mid-log phase in LB broth with 100 μg/ml ampicillin and aliquots were spread on LB agar plates with 100 μg/ml ampicillin and 0.1 mM IPTG. Strain DH5α (pTrcHisA) grew on the plates, but DH5α (pDD560) did not, indicating that induction and overexpression of vapD Hi was toxic to E. coli . The putative antitoxin VapX Hi , was then cloned into the pTrcHisA vector. No growth disruption occurred in DH5α with the overexpression of VapX Hi alone. The vapX Hi gene plus the lacI q gene were then subcloned into the broad host range mobilizable plasmid, pDD515, resulting in pDD672. This strategy allowed the DH5α test strain to carry two compatible plasmids, one which encoded the vapD Hi gene (pDD560) and the other expressing the vapX Hi antitoxin gene (pDD672). Both genes were under the control of a P trc promoter and were therefore both repressed until induced by IPTG. When both genes were induced with 0.1 mM IPTG on LB agar plates that contained 100 μg/ml ampicillin and 10 μg/ml chloramphenicol in strain DH5α (pDD560, pDD672), growth was restored. This indicated that the concurrent expression of the vapX Hi antitoxin with the vapD Hi toxin ameliorated the cell growth arrest observed with expression of vapD Hi alone, and that vapX Hi was necessary for this rescue. Discussion Mutation of the vapD Hi allele in strains Rd and R3001 resulted in attenuation of survival within both HBMEC and NCI-H292 monolayers, suggesting that in H. influenzae , the presence of a functional VapD Hi facilitates persistence within epithelial and endothelial cells. Mutants invaded and survived in human cells at ≤ 60% of wild-type levels. Although relatively modest, this level of attenuation has been observed during the mutational analysis of other Haemophilus virulence factors, such as opacity-associated protein A as well as the high molecular weight (HMW) proteins [ 11 , 16 ]. H. influenzae survival within human cells is multifactorial, and our data indicate that VapD Hi contributes to this process. However, strain Rd contains three other vap genes ( vapA Hi , vapB Hi , and vapC Hi ), and it is possible that these Haemophilus vap genes act synergistically, such that multiple mutations may result in a more attenuated survival phenotype. Indeed, a recent study has determined that a chromosomally-located homologue of the VapBC locus acts as a toxin-antitoxin module in the spirochete Leptospira interrogans [ 17 ]. It would be interesting to characterize a Haemophilus strain with mutations in all the vap genes. Neither of the vapD Hi mutants displayed differences in adherence to the monolayers compared to the parent strain, so the defect occurred after binding and affected the organism's ability to persist inside or between cells. Interestingly, the vapD Hi mutants were not attenuated in growth rate when compared to the parent strains, either in bacteriological media or on the surface of human cell monolayers. The observed survival attenuation of the mutants could be transcomplemented with a full-length allele from a clinical isolate, R3001, demonstrating that the phenotype was due to the mutation in vapD Hi and not polar effects. A truncated allele from another clinical isolate, R2866, did not transcomplement the Rd vapD Hi strain, indicating that the truncated protein was not functional in vivo . RT-PCR analysis confirmed that the full-length vapD Hi locus in Rd was transcribed during contact with both epithelial and endothelial cells. VapD Hi has also been identified in the soluble fraction of strain Rd grown in bacteriological media [ 10 ]. It remains to be seen if the transcription of this locus increases upon contact with human cells. Results of a PCR survey on 59 randomly-chosen strains showed that nearly all of the genetically highly heterologous NTHi commensal isolates surveyed (93%) carried a full-length vapD Hi allele on their chromosomes, suggesting that maintenance of a functional vapD Hi gene was beneficial to survival in this niche (Table 2 ). Of the invasive strains isolated from the blood or cerebrospinal fluid, 71% retained a full-length allele. Fifty percent of the isolates from sputum, tracheal aspirates, and the middle ear carried the full-length allele. Finally, all the encapsulated strains tested contained a full-length allele. It must be noted, however, that this analysis was not an exhaustive study, since a limited number of strains were included. Many clinical NTHi strains have previously been shown to express various virulence factors that enhance adherence and invasion into human cells which are not found in the sequenced Rd strain [ 11 , 18 - 20 ]. The strains identified in this PCR study that lack a functional vapD Hi allele probably compensate for its loss with genes that are not found in Rd, and these "extra genes" may well include other TA loci. The calculated molecular mass of the VapD Hi protein in Rd is approximately 10 kilodaltons, and small bacterial proteins often form multimers. Full-length VapD Hi subunits exhibited strong homodimerization in a LexA-based protein-protein interaction system, and this may indicate that the subunits form higher-order multimers such as homotrimers or homotetramers in vivo . However, the protein encoded by the truncated allele did not homodimerize in the same system, further evidence of its loss of function. Interestingly, the truncated subunit did interact with full-length subunits in heterodimerization assays. Further evidence of this heterodimerization in vivo was that the expression of the truncated subunit in the wild-type strain resulted in a dominant-negative effect on survival within HBMEC monolayers, the levels of which mimicked the attenuation observed with the vapD Hi mutation. This was likely due to truncated subunits forming hybrid complexes with full-length subunits and interfering with their structure and/or function, resulting in the observed dominant-negative phenotype. The activities of only a few toxins encoded by TA loci have been elucidated thus far. Two specific targets of plasmid-encoded toxins have been identified: CcdB of the F plasmid and ParE of plasmid RK2 inhibit DNA gyrase, and Kid of plasmid R1 was previously thought to interact with DnaB helicase, but has recently been shown to cleave cellular mRNA [ 21 - 23 ]. The target and function of a toxin from a chromosomally-encoded TA locus ( relBE ) was determined to be cleavage of mRNA in the ribosomal A site [ 24 ]. Strain Rd contains relBE homologues (HI0710 and HI0711) as well as a homologue of higA ( h ost i nhibition of g rowth antidote protein) from plasmid Rts1. The higBA TA locus is unusual in that the toxin gene ( higB ) exists upstream of the antidote protein ( higA ). Interestingly, VapA Hi of strain Rd is 29% identical and 53% similar to HigA. While the data acquired in this study suggests that VapD Hi and VapX Hi form a toxin-antitoxin pair, it is unusual to find homologues of VapX Hi only in H. influenzae and the gonococcal plasmid. H. influenzae strains R2846 and R2866 both have a truncated vapD Hi toxin gene, and possess a vapX Hi which is 100% identical to that gene in strain Rd. Their respective genomes can be searched at . Interestingly, there are complete genome sequences available for two isolates of H. pylori, N. meningitidis and Haemophilus somnus , but only one strain of each has a homologue of vapD Hi . Thus, several features of the vapD Hi / vapX Hi gene pair are unusual for a toxin-antitoxin locus. Conclusions Persistence of NTHi is important in the progression of disease caused by this organism. Many investigators have previously reported the discovery of a number of virulence factors associated with adherence, invasion and survival of NTHi inside human cells [ 1 , 3 , 4 , 16 , 18 , 25 ]. Here we report a locus that is also involved in the pathogenesis of nontypeable H. influenzae . Further studies are required to fully characterize the mechanism of VapD Hi function and to define its role in the modulation of NTHi persistence in human cells. Methods Bacterial strains, media and reagents All H. influenzae strains used are listed in Table 2 . H. influenzae was grown on chocolate agar (36 g Difco GC medium, 10 g hemoglobin, 10 ml Difco Supplement B (Becton Dickinson, Sparks, MD), 5,000 Units bacitracin per liter) or supplemented BHI (sBHI) broth or agar (37 g brain heart infusion media ± 15 g Bacto agar per liter (Remel, Lenexa, KS) with 10 μg/ml β-NAD, 10 μg/ml heme-histidine, and 5 Units/ml bacitracin). Strains containing the TSTE cassette [ 26 ] were grown on media with 15 μg/ml ribostamycin sulfate (CalBioChem, San Diego, CA). Bacteria were diluted for plating with PBS-G (phosphate-buffered saline (pH 7.0) with 0.1% gelatin). Escherichia coli strains used were DH5α, to clone fragments of NTHi DNA; DD12, as the host strain in conjugations; and SU101 or SU202 as the reporter strains for the homodimerization and heterodimerization assays, respectively [ 14 , 15 ]. T4 bacteriophage was obtained from the American Type Culture Collection (ATCC #11303). Antibiotics and other chemicals were from Sigma-Aldrich (St. Louis, MO). Restriction enzymes, deoxyribonucleotides, T4 DNA polymerase, and T4 DNA ligase were from Promega (Madison, WI). Enzymes and other reagents for PCR were from Eppendorf Scientific (Westbury, NY) and Bioline (Canton, MA). Enzymes and reagents for reverse transcriptase PCR (MasterAmp RT-PCR Kit) were from Epicentre Technologies (Madison, WI). Oligonucleotide primers were synthesized by Integrated DNA Technologies (Coralville, IA). DNA sequencing was performed by the University of Missouri DNA Core Facility (Columbia, MO), Davis Sequencing, LLC (Davis, CA), and the DNA Core Facility at Seattle Biomedical Research Institute (Seattle, WA). Plasmids were isolated using the Wizard SV Plus Plasmid Miniprep kit, PCR products and restriction digests were purified using the Wizard PCR Prep kit, and total bacterial RNA was isolated using the SV Total RNA Isolation System (Promega, Madison, WI). Plasmids and conjugations For transcomplementation, the 276 bp R3001 vapD Hi allele, along with 269 bp upstream and 227 bp downstream, was PCR-amplified using primers vap HI forward 5'-TATG TCTAGA CAGTCGCTTCATAAGC-3' and vap HI reverse 5'-CCAT TCTAGA TTTGAGGTTAAATATGG-3'. Both primers included a XbaI site (underlined) and amplified Rd genome coordinates 472803 to 473572. The product was sequenced and cloned both into the XbaI site of pBluescript SK + (creating plasmid pDD562) and into the NheI site of pDD515, a mobilizable broad-host range vector that could be conjugally transferred into and stably maintained in H. influenzae [ 12 ], creating plasmid pDD564 (Table 1 ). Plasmid pDD564 was used for transcomplementation of Rd vapD Hi . The same primers were used to PCR amplify the 135 bp truncated vapD Hi allele from strain R2866, the product of which was cloned into the NheI site of pDD515, creating pDD594. The insert was sequenced, then the plasmid was conjugally transferred into both Rd vapD Hi and wild-type Rd. Conjugations were carried out as previously described [ 12 ]. For allelic exchange, the plasmid pDD563 was constructed, which consisted of pDD562 with an interruption of vapD Hi by an aminoglycoside phosphotransferase gene ( aph (3')II ). Specifically, the 2184 bp BamHI fragment from pTSTE [ 26 ], which had been rendered blunt-ended with mung bean nuclease, was inserted into the BsaBI site of vapD Hi . For the homodimerization assays, the plasmids pDD559 and pDD577 were derived from pSR658 and carried the NTHi strain R3001 vapD Hi allele or the NTHi strain R2866 vapD Hi allele fused in-frame to the wild-type LexA DNA-binding domain, respectively [ 27 ]. For the heterodimerization assays, the plasmid pDD561 derived from pSR659 was constructed, which consisted of the R3001 vapD Hi allele fused in-frame to the mutated LexA DNA-binding domain. Mutation of vapD Hi The vapD Hi genes in Rd and strain R3001 were disrupted by allelic exchange. Briefly, strains Rd and R3001 were made competent using the M-IV media technique [ 28 ] and pDD563 linearized with XmnI was used to transform each strain. Transformants were selected on chocolate agar supplemented with 15 μg/ml ribostamycin sulfate (CalBioChem, San Diego, CA). The insertion in vapD Hi was confirmed by Southern blotting using a digoxygenin-labeled denatured PCR fragment of vapD Hi as the probe. The orientation of the aminoglycoside phosphotransferase cassette was determined by PCR using a primer that originated inside the aph (3')II gene and another that flanked vapD Hi . The resistance gene was found to be transcribed in the opposite orientation of vapD Hi in both strains. Cell culture Human brain microvascular endothelial cells (HBMECs) were a gift from K. S. Kim [ 29 ]. Cells were passaged in collagen-1 coated T-25 flasks and monolayers for invasion assays were grown in 12-well collagen-1 coated BioCoat plates (Becton Dickinson, Bedford, MA). HBMEC media contained 760 ml RPMI 1640 with 25 mM HEPES and 2 mM L-glutamine, 100 ml heat-inactivated fetal calf serum, 10 ml each of 200 mM L-glutamine, 100× MEM non-essential amino acid solution, 100× MEM vitamin solution, 100 mM MEM sodium pyruvate solution (Gibco, Grand Island, NY), and 100 ml heat-inactivated NuSerum V (Becton Dickinson, Bedford, MA) per liter. Media was changed every two days and cells were passaged every 3–5 days. Monolayers were seeded at a density of ~2.0 × 10 5 cells per well and used 48 to 72 hours after seeding. NCI-H292 human respiratory epithelial cells (ATCC catalogue # CRL-1848) were passaged in collagen-1 coated T-25 flasks and monolayers for invasion assays were grown to confluency in 12-well collagen-1 coated BioCoat plates (Becton Dickinson, Bedford, MA). NCI-H292 media consisted of 870 ml RPMI 1640 medium with 25 mM HEPES and 2 mM L-glutamine, 10 ml of 100 mM MEM sodium pyruvate solution, 10 ml of 7.5% w/v sodium bicarbonate solution (Gibco, Grand Island, NY), 10 ml of 450 mg/ml filter-sterilized glucose solution, and 100 ml heat-inactivated fetal calf serum per liter. As above, media was changed every two days and cells were passaged every 3–5 days. Monolayers were seeded at a density of ~2.5 × 10 5 cells per well and used 72 to 96 hours after seeding. Invasion and survival assays Gentamicin-resistance invasion and survival assays were performed on HBMEC and NCI-H292 monolayers as previously described [ 5 ]. Briefly, the inoculum used was 1.0 – 5.0 × 10 6 CFU of H. influenzae in a volume of 1 ml per well of a 12-well plate (an MOI of ≤ 10:1). After a 3 or 18 hour incubation in an atmosphere of 5% CO 2 at 37°C, each monolayer was extensively washed with Dulbecco's PBS and 1.5 ml of media containing 100 μg/ml gentamicin was added to each well. Following a subsequent one hour incubation in the antibiotic, the wells were again washed extensively, harvested with 1% saponin, diluted in PBS-G and plated on chocolate or sBHI agar for viable intracellular CFU/ml. To quantitate total cell-associated bacteria (both intracellular and adherent), wells were also harvested and plated after the first wash and prior to gentamicin addition. Methods for statistical analysis Statistical analyses were performed using the statistical analysis functions of Microsoft Excel (Microsoft Office 1997). For most comparisons of data, the Student's t -test was used and P -values of <0.05 were considered to indicate statistically significant differences. Reverse transcriptase PCR (RT-PCR) Total RNA was isolated using the SV Total RNA isolation system (Promega, Madison, WI) from the wild type strain Rd recovered from the media of 18-hour invasions of either HBMEC (endothelial) or NCI-H292 (epithelial) monolayers. Standard procedures were used, with the modification that two separate DNAseI incubations were performed instead of the single one recommended with the kit. RT-PCR using the MasterAmp RT-PCR kit was then performed as per the manufacturer's instructions (Epicentre Technologies, Madison, WI). Negative controls of no reverse transcriptase added to the RNA followed by traditional PCR using Biolase DNA polymerase (Bioline, Madison, WI) were used to ensure that both RNA preparations were free of contaminating DNA. The primers for RT-PCR, 450 RT for (5'-CAGGCTTATACAGACATTGG-3') and 450 RT rev (5'-TCGTACCGACTGAGAAATCC-3') amplified a 153 bp internal portion of the vapD Hi cDNA. Protein-protein interaction assays The vapD Hi alleles from strains R3001 (full-length) and R2866 (truncated) were amplified by PCR and fused in-frame to the LexA DNA-binding domain (DBD) in pSR658, resulting in pDD559 and pDD577, respectively, and used to transform the reporter strain SU101 for homodimerization assays [ 27 ]. Briefly, strain SU101 carries a lacZ gene controlled by a wild-type LexA operator site [ 14 ]. If a homodimer of two LexA DBD fusions was formed, the complex could bind to the LexA operator region and shut down transcription of lacZ , resulting in diminished levels of beta-galactosidase. The vapD Hi allele from strain R3001 was also fused in-frame to the mutated LexA DNA-binding domain in pSR659, creating pDD561. The compatible plasmids pDD561 (full length vapD Hi ) and pDD577 (truncated vapD Hi ) were both used to transform the reporter strain SU202 for heterodimerization assays. Strain SU202 [ 14 ] also has a lacZ gene controlled by a LexA operator, but this operator site is engineered such that only a mutated LexA DBD subunit (coded on pSR659) can bind to one half-site, while a wild-type LexA DBD subunit (coded on pSR658) can bind to the other half-site. Consequently, only a heterodimer composed of one mutated LexA DBD fusion subunit and one wild-type LexA DBD fusion subunit could bind to this engineered site and decrease transcription of lacZ in SU202. Three independent beta-galactosidase assays were carried out in triplicate as previously described [ 27 ]. Author's contributions DAD conceived of the study, carried out the protein-protein interaction and molecular genetics work, and drafted the manuscript. JJ carried out the intracellular survival experiments. ALS supported the study and participated in its design and coordination. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC503385.xml
515295	Phosphorylated guanine nucleotide exchange factor C3G, induced by pervanadate and Src family kinases localizes to the Golgi and subcortical actin cytoskeleton	Background The guanine nucleotide exchange factor C3G (RapGEF1) along with its effector proteins participates in signaling pathways that regulate eukaryotic cell proliferation, adhesion, apoptosis and embryonic development. It activates Rap1, Rap2 and R-Ras members of the Ras family of GTPases. C3G is activated upon phosphorylation at tyrosine 504 and therefore, determining the localization of phosphorylated C3G would provide an insight into its site of action in the cellular context. Results C3G is phosphorylated in vivo on Y504 upon coexpression with Src or Hck, two members of the Src family tyrosine kinases. Here we have determined the subcellular localization of this protein using antibodies specific to C3G and Tyr 504 phosphorylated C3G (pY504 C3G). While exogenously expressed C3G was present mostly in the cytosol, pY504 C3G formed upon Hck or Src coexpression localized predominantly at the cell membrane and the Golgi complex. Tyrosine 504-phosphorylated C3G showed colocalization with Hck and Src. Treatment of Hck and C3G transfected cells with pervanadate showed an increase in the cytosolic staining of pY504 C3G suggesting that tyrosine phosphatases may be involved in dephosphorylating cytosolic phospho-C3G. Expression of Src family kinases or treatment of cells with pervanadate resulted in an increase in endogenous pY504 C3G, which was localized predominantly at the Golgi and the cell periphery. Endogenous pY504 C3G at the cell periphery colocalized with F-actin suggesting its presence at the subcortical actin cytoskeleton. Disruption of actin cytoskeleton by cytochalasin D abolished phospho-C3G staining at the periphery of the cell without affecting its Golgi localization. Conclusions These findings show that tyrosine kinases involved in phosphorylation of C3G are responsible for regulation of its localization in a cellular context. We have demonstrated the localization of endogenous C3G modified by tyrosine phosphorylation to defined subcellular domains where it may be responsible for restricted activation of signaling pathways.	Background Guanine nucleotide exchange factors (GNEFs) are components of signaling pathways that link transmembrane receptors to intracellular GTPase family members regulating a wide variety of cellular functions such as proliferation, differentiation, adhesion and apoptosis. C3G (RapGEF1) is an ubiquitously expressed GNEF for Ras family proteins that particularly targets Rap1, Rap2 and R-Ras [ 1 - 4 ]. It has been shown to mediate signals received from B and T cell receptor activation, growth factors, cytokines, G protein coupled receptors and also adhesion [ 5 - 15 ]. C3G is present in the cytoplasm in a complex with members of the Crk family of small adapter molecules. In response to stimuli, this complex is recruited to the cell membrane involving association of Crk with phosphotyrosine containing proteins like receptor tyrosine kinases, p130 Cas, IRS-1 and paxillin [ 16 - 18 ]. Following translocation from cytosol to cell membrane, C3G activates downstream signaling. Its activation has been shown to lead to an activation of mitogen activated protein kinase and Jun N-terminal kinase [ 9 , 12 , 19 - 21 ]. Studies involving overexpression of membrane targeted C3G or dominant negative forms have shown that C3G is involved in both growth suppression as well as transformation [ 22 - 24 ]. C3G appears to play an important role in mammalian development because C3G-/- mice die before embryonic day 7.5. These studies have shown that C3G is required for vascular myogenesis and for cell adhesion and spreading [ 25 , 26 ]. The C-terminus of C3G, which shows homology to CDC25, harbors the catalytic domain. The central region of C3G, which spans about 300 residues, has polyproline tracts with the ability to bind to SH3 domains of various proteins like Crk, p130 Cas, Grb2 and Hck [ 1 , 2 , 9 , 18 , 27 ]. No function has particularly been attributed to the N-terminal sequences, which do not show homology to any defined protein sequences. The non-catalytic domain of C3G has been shown to negatively regulate its catalytic activity. Deletion of the N-terminal sequences or its association through its proline sequences to Crk leads to its activation [ 16 ]. Integrin mediated cell adhesion causes tyrosine phosphorylation of C3G [ 28 ]. It has been shown that overexpression of c-Crk1 or stimulation of cells with growth hormone leads to specific phosphorylation of Y504 [ 21 , 29 ]. This modification results in an increase in C3G catalytic activity towards Rap1. Src and JAK have been implicated in Y504 phosphorylation of C3G. More recently we have used site – specific antibodies to show that the activation of Src family kinase Hck, leads to C3G phosphorylation on Y504 suggesting that Src family kinases can directly regulate C3G activity and function [ 27 ]. The effectiveness and precision of intracellular signal transduction depends on protein-protein interactions that regulate enzyme activity as well as subcellular localization. Cell surface receptor activation leads to assembly of adaptor protein complexes at the plasma membrane, which serve to localize guanine nucleotide exchange proteins. Earlier, both endogenous as well as exogenously expressed C3G has been shown to localize to the cytoplasm and not to associate with plasma membrane [ 22 , 30 ]. Since activation of C3G occurs primarily through phosphorylation at Tyr 504 and membrane recruitment, we undertook a detailed study of the subcellular localization of both exogenously expressed and endogenous Y504 phosphorylated C3G (pY504 C3G). Expression of Src family kinases or pervanadate treatment of cells, which mimics stimulation by growth factors, resulted in marked tyrosine phosphorylation of C3G at Y504. Conventional as well as optical sectioning microscopy revealed that pY504 C3G was predominantly located at the Golgi complex and the subcortical actin cytoskeleton unlike non-phosphorylated C3G, which was largely cytosolic. Results Colocalization of C3G with Hck We have recently shown that Hck interacts with and phosphorylates C3G in vivo and we wished to determine if their interaction leads to changes in the subcellular distribution of C3G and whether pY504 C3G locates to specific subcellular domains. Cos-1 cells were transfected with C3G in the presence or absence of Hck and immunostained using anti C3G antibodies. As shown in Fig. 1A in a majority of cells, exogenously expressed C3G showed diffuse cytoplasmic staining that extended up to the plasma membrane. Variation in the level of C3G was observed among the transfected cells with weakly expressing cells showing a more prominent juxtanuclear staining. When cotransfected with Hck, most cells showed prominent staining of C3G at the plasma membrane and a juxtanuclear organelle in addition to the diffuse cytoplasmic staining. This pattern appeared similar to that seen for exogenously expressed Hck and therefore we performed colocalization studies to confirm their distribution. When coexpressed, Hck and C3G are targeted predominantly to the plasma membrane and other intracellular membranous structures (Fig. 1B ). Merged images show that these two proteins colocalize in the subcellular context. Similar patterns of colocalization were observed when C3G and Hck were expressed in HeLa cells (data not shown). Figure 1 Subcellular localization of C3G. (A) Cos-1 cells grown on coverslip were either transfected with C3G or cotransfected with Hck and indirect immunofluorescence staining performed using anti-C3G antibodies and Cy3 conjugated anti rabbit secondaries. (B) Cells transfected with Hck and C3G were stained for both the antigens as described in Materials and Methods. Hck was visualized using FITC conjugated secondaries and C3G by Cy3 conjugated secondaries. The dual panel shows the merged image of an optical section taken using the confocal microscope where the yellow signal generated shows colocalization of the two proteins. Src family kinases phosphorylate C3G and phospho-C3G localizes to the Golgi and plasma membrane To determine the subcellular distribution of phospho-C3G, which is known to be the activated form, specificity of a rabbit polyclonal phosphorylation site-specific antibody (pY504-C3G) was verified by examining its reactivity using cell lysates expressing C3G or Y504F-C3G alone or with Hck. As shown in Fig. 2A , pY504-C3G antibody recognizes only C3G when coexpressed with Hck. Neither the C3G protein expressed in itself nor the Y504F mutant coexpressed with Hck show any reactivity with this antibody suggesting that it reacts only with Y504 phosphorylated C3G. Phosphotyrosine blotting showed that a large number of cellular polypeptides are phosphorylated on tyrosine upon Hck expression, (Fig. 2A , right panel) but except for C3G none of the others show any reactivity with pY504 antibody. Y504F mutant of C3G, which shows low level of phosphorylation on other tyrosine residues is not detected by this antibody indicating its specificity towards Y504 phosphorylated C3G. Unlike Hck, whose expression is restricted to a subclass of hematopoietic cells, C3G is ubiquitously expressed and we wished to determine if other Src family kinases could phosphorylate C3G. We coexpressed C3G with an expression construct for the fusion protein c-Src-GFP and western blotting was performed using pY504-C3G antibody. As shown in Fig. 2B , in vivo, c-Src was also able to induce Y504 phosphorylation of C3G, but not that of the Y504F C3G mutant. Figure 2 Specificity of phosphospecific antibody, and phosphorylation of C3G on Y504 upon coexpression with Src family kinases. Cos-1 cells were transfected with the expression constructs for Hck (A) or c-Src (B) along with C3G as indicated and western blotting of whole cell lysates was performed using the phosphospecific antibody pY504. The blots were reprobed with C3G, Hck and anti pTyr (Panel A) or Src (Panel B) to show their expression in the lysates. pY504 C3G and pTyr was detected by ECL and C3G, Hck and Src by alkaline phosphatase dependent color development. UT indicates untransfected cell lysates. Y504F is a mutant of C3G in which tyrosine 504 is replaced by phenylalanine. We wished to determine whether C3G that colocalizes with Hck was the phosphorylated component and therefore used pY504-C3G antibodies to determine the localization of pY504 C3G in cells expressing Hck and C3G. As shown in Fig. 3A , pY504 C3G showed a staining pattern that exactly matched that of Hck with prominent staining of the plasma membrane, a juxtanuclear organelle and other intracellular membranes (3A). The prominent staining appeared to correspond with the Golgi structure and Hck has earlier been shown to localize to the Golgi [ 31 ]. In cells transfected with C3G and Hck, the pattern of pY504 C3G staining was also compared with that of total C3G as detected by the Flag tag antibody. As shown in Fig. 3B , it was observed by confocal analysis that the tag antibody detects the presence of C3G spread throughout the cytoplasm with some prominence in the juxtanuclear region and plasma membrane. The pattern suggests that the majority of the protein is cytosolic. In contrast staining for phospho-C3G was non-uniform and was particularly prominent at the Golgi and cell membrane. Colocalization of C3G with that of phospho-C3G is seen at the plasma membrane and in the juxtanuclear region. This also suggests that only a proportion of the expressed C3G is phosphorylated at Tyr504. To confirm the presence of pY504 C3G in the Golgi, we coexpressed the viral protein, VSVG-GFP known to localize to the Golgi with Hck and C3G and observed the staining pattern of pY504 C3G and that of GFP. VSVG-GFP locates predominantly at the Golgi, trans-Golgi network and also the endoplasmic reticulum and plasma membrane in a temperature-dependent manner [ 32 ]. As shown in Fig. 3C , the yellow signal generated in the dual image showed colocalization of pY504 C3G with VSVG-GFP suggesting that pY504 C3G was predominantly targeted to the Golgi complex. Unlike C3G, pY504 C3G appeared to be restricted to the plasma membrane and other intracellular membranes with particular concentration in the Golgi. When overexpressed, a large amount of C3G was present in the cytosol and we wished to determine whether cytosolic C3G does not get phosphorylated upon Hck coexpression or whether pY504 C3G in the cytosol is transient due to the action of tyrosine phosphatases. Cos-1 and HeLa cells transfected with Hck and C3G were either left untreated, or, subjected to pervanadate treatment for 10 minutes prior to fixation and stained for pY504 C3G. Pervanadate is a strong inhibitor of tyrosine phosphatases; therefore treatment of cells with pervanadate results in dramatic augmentation of tyrosine phosphorylation on cellular proteins [ 33 ]. As shown in Fig. 3D , pervanadate-treated cells showed an increase in the pY504 C3G staining in the cytoplasm suggesting that it was dephosphorylated by cytosolic tyrosine phosphatases. Figure 3 pY504-C3G colocalizes with Hck and shows predominant Golgi and membrane localization. (A) pY504 C3G colocalizes with Hck. Cos-1 cells transfected with Hck and C3G were stained for pY504 C3G (Cy3) and Hck (FITC) and examined using a confocal microscope. Figure shows an optical section for the individual stains as well as that of the merged (Dual) image. (B) Cos-1 cells transfected with Hck and C3G were dual labeled to detect phospho-C3G (Cy3 staining) and C3G using the Flag tag antibody (FITC staining). Panels show optical sections taken using the confocal microscope. (C) pY504 C3G is localized to the Golgi apparatus. Cos-1 cells were transfected with Hck, C3G and VSVG-GFP expression constructs and stained using pY504 primary antibody and Cy3 conjugated secondary. An optical section taken using the apotome is represented. (D) HeLa or Cos-1 cells transfected with Hck and C3G were left untreated (control) or treated with pervanadate (PV) prior to fixation and stained for pY504. Counter staining with Dapi shows cell nuclei. Phosphorylation of endogenous C3G and its localization to the Golgi and subcortical actin cytoskeleton In the above experiments phosphorylation and localization of C3G was studied using exogenously expressed protein and we wished to determine whether endogenous C3G could be phosphorylated and similarly targeted. Towards this end we checked the phosphorylation of endogenous C3G under conditions of Src and Hck overexpression or upon activation of cellular tyrosine kinases by pervanadate treatment. C3G protein is expressed as a doublet of about 140–150 kDa, which are products of two differentially spliced mRNAs [ 34 ]. Whole cell lysates were prepared from Cos-1 cells and those transfected with Hck or Src and western blotting performed using pY504 antibodies. As shown in Fig. 4A , overexpression of Src or Hck induces tyrosine 504 phosphorylation of endogenous C3G. The same blot was reprobed with C3G, Src and Hck antibodies to show their presence in the lysates. We examined the localization of endogenous phosphorylated C3G after c-Src expression and found that similar to the phosphorylated form of the exogenously expressed C3G, endogenous pY504 C3G was present predominantly at sites of c-Src localization. Intense staining of the Golgi and cell membranes was evident and merged images show colocalization of the two proteins (4B). Cells that did not express Src, did not show any phosphorylated C3G. Figure 4 Phosphorylation of endogenous C3G upon overexpression of Hck and its localization to the Golgi. (A) Cos-1 cells were transfected with expression constructs as indicated and whole cell lysates used in western blotting for pY504-C3G. ECL was used for detection. Blots were reprobed to show expression of C3G and the kinases. (B) Endogenous pY504 C3G colocalizes with c-Src. Cos-1 cells were transfected with the c-Src GFP fusion protein vector and cells stained for pY504-C3G expression (Cy3). c-Src expression was visualized as GFP fluorescence. Images shown are optical sections taken using the apotome. (C) Endogenous pY504-C3G localizes to the Golgi. Cos-1 cells were transfected with Hck along with VSVG-GFP and stained for pY504 C3G. Cells were left untreated (control) or treated with nocodazole (Noc) prior to fixation as described in Methods. Panels show optical section for pY504 by Cy3 and the VSVG-GFP by GFP fluorescence. The localization of endogenous pY504 C3G to the Golgi was examined in Cos-1 cells transfected with Hck and VSVG-GFP. Immunostaining for pY504 C3G was seen predominantly at the cell periphery and the Golgi (Fig. 4C ), which was confirmed by colocalization with VSVG-GFP. The effect of Golgi perturbing drugs on the localization of pY504 C3G was examined by treatment of cells with nocadazole for depolymerization of microtubules and concomitant Golgi fragmentation. Under these conditions pY504 C3G was detected as dispersed vesicles scattered in the cytoplasm and remained colocalized with VSVG-GFP (4C) confirming that endogenous pY504 C3G localized to the Golgi complex. The localization of pY504 C3G formed by the activation of endogenous tyrosine kinases was determined in cells treated with pervanadate, which is known to activate Src family kinases, in addition to inhibiting tyrosine phosphatases [ 35 - 37 ]. Pervanadate treatment results in the dramatic augmentation of phosphorylation of a large number of cellular proteins on tyrosine and therefore mimics activation of signaling pathways by growth factors [ 38 ]. While normal HeLa cells do not show any pY504 C3G, cells treated with pervanadate showed distinct presence of pY504 C3G in whole cell lysates as seen by western blotting (5A). The large number of other cellular proteins phosphorylated on tyrosine (seen upon blotting with antiphosphotyrosine antibodies), as a consequence of pervanadate treatment, do not show reactivity with pY504 antibody. In order to confirm that the signal observed upon pervanadate treatment was specific to phospho Y504-C3G, Cos-1 cells were transfected with either C3G or Y504F mutant of C3G. They were left untreated, or treated for 10 minutes with pervanadate and indirect immunofluorescence performed to observe expression of the wild type or mutant proteins as well as that of phospho-C3G. C3G, and Y504FC3G expression was monitored by staining for Flag and His tags respectively. As observed in Fig. 5B only cells expressing C3G showed intense staining for phosphoC3G while Y504F expressing cells showed no enhanced signal above that of the other non-expressing cells in the field. These results reaffirmed the specificity of the phospho-C3G antibody in detecting only Y504 phosphorylated C3G. Phosphorylated endogenous C3G staining was seen weakly in the non-expressing cells upon pervanadate treatment. Figure 5 Phosphorylation of endogenous C3G upon activation of endogenous tyrosine kinases. (A) Cells were either left untreated (UT) or treated with pervanadate (PV) and western blotting was performed using pY504 antibody. The same blot was reprobed with C3G to show the presence of endogenous C3G in these cells. (B) Cos-1 cells on coverslips were transfected with either C3G or Y504F mutant of C3G and fixed without any treatment (cont.) or after pervanadate treatment (PV). Dual labeling was performed using the tag antibodies (stained with FITC) and pY504 antibody (stained with Cy3). Panels show optical sections obtained by confocal microscopy. (C) Cos-1 and HeLa cells grown on coverslips and transfected with VSVG-GFP were fixed without any treatment (control) or after treatment with pervanadate and stained for pY504 expression. GFP fluorescence was used to visualize the staining pattern of VSVG-GFP protein. Optical sections taken using the apotome are shown. Areas of colocalization are seen from the yellow color generated in the merged images. Indirect immunoflourescence was performed to determine the localization of endogenous pY504 C3G formed by the activation of intracellular tyrosine kinases. Cos-1 and HeLa cells were stimulated by pervanadate and as seen in Fig. 5C , pY504 C3G staining, which is evident only in the treated cells, localized at the cell periphery and the Golgi. Colocalization with VSVG confirmed its localization to the Golgi complex. The staining at the cell periphery appeared to match that of the subcortical actin cytoskeleton. In order to confirm this, we dual stained the cells treated with pervanadate for F-actin and found that the pY504 C3G seen at the cell periphery colocalizes with F-actin suggesting that pY504 C3G is targeted to the subcortical actin cytoskeleton upon activation of endogenous tyrosine kinases by pervanadate (Fig. 6A ). It was also observed that pY504 C3G staining at the cell periphery was particularly prominent in confluent cells compared to cells that were sparsely growing in isolation suggesting that pY504 C3G is particularly enriched along cell-cell junctions. Phospho C3G also shows partial colocalization with filamentous actin known to be associated with the Golgi complex [ 39 ]. Figure 6 pY504C3G localizes to the subcortical actin cytoskeleton . (A) HeLa cells grown on coverslips were left untreated or treated with pervanadate and stained for pY504 expression using Cy3 secondaries. The coverslips were then stained with Oregon-green phalloidin to detect F-actin. (B) C3G phosphorylation requires the activity of Src family kinases and the presence of an intact cytoskeleton. HeLa cells were pretreated with PP2 or cytochalasin D as described in methods prior to pervanadate treatment. Images show the localization of pY504 C3G labeled with Cy3 and F-actin stained with oregon green. Images shown are a single optical section visualized using the apotome. We have observed that pervanadate treatment increased tyrosine phosphorylation of endogenous C3G. Since overexpression of Src as well as Hck results in phosphorylation of C3G, it was of interest to determine whether pervanadate induced tyrosine phosphorylation of C3G was mediated by Src family kinases within the cell. Cells were treated with PP2, a specific SFK inhibitor prior to PV treatment [ 40 ]. As shown in Fig. 6B , pY504 C3G staining of cells stimulated with pervanadate was considerably reduced, but not totally abolished when they were pretreated with PP2 suggesting the possibility of other tyrosine kinase family members activated by pervanadate contributing to C3G phosphorylation. To determine whether the increase in pY504 C3G staining was dependent on the presence of an intact cytoskeleton, we observed its localization in cells treated with cytochalasin D, a reagent that effectively disrupts the actin cytoskeletal network. Under these conditions there is a collapse of cell morphology and F-actin staining shows an irregular distribution at the cell cortex. As observed in Fig. 6B , the staining for pY504 C3G in the subcortical cytoskeleton was largely absent under conditions of moderate disruption of actin organization. Under these conditions, pY504 C3G staining at the Golgi complex, which shows a more dispersed morphology appeared not to be affected. Discussion C3G is involved in a variety of signaling pathways and therefore its dynamic localization under normal and activated situations may be physiologically relevant. In this study we demonstrate the limited subcellular distribution of Y504 phosphorylated C3G, which is predominantly targeted to the Golgi apparatus and the subcortical actin cytoskeleton. This localization has been substantiated by colocalization with a Golgi marker protein and F-actin respectively. Rap1, the substrate of C3G has been localized to the Golgi, lysosomal vesicles and cortical actin cytoskeleton [ 41 ]. But, of the at least eight known exchange factors for Rap1, C3G is the only one that has been linked definitively to the tyrosine kinase signaling pathway. Src family members like Src and Hck have been shown to localize to the plasma membrane and other intracellular membranes with particular concentration in the Golgi [ 31 , 42 ]. When endogenous C3G was phosphorylated by overexpressed Hck or Src, the localization of pY504 C3G matched that of the kinases suggesting that they may be part of the same molecular complexes. This is also evident from the staining pattern of pY504 C3G when C3G is expressed along with Hck, which is distinctly seen in the Golgi and plasma membrane. Since exogenously expressed C3G is predominantly cytosolic, it implies that, at any given time, only a small fraction of it is phosphorylated at Y504 at the plasma membrane and the Golgi. We observed more exogenously expressed pY504 C3G in the cytoplasmic compartment under conditions of inhibition of tyrosine phosphatases suggesting that pY504 C3G may be targeted by cytosolic tyrosine phosphatases. This regulation may help in restricting the activity of C3G to specific compartments. We observe very little endogenous pY504 C3G in the cytosol when HeLa or Cos-1 cells are treated with pervanadate, which not only inactivates tyrosine phosphatases, but also activates tyrosine kinases. It is possible that upon PV treatment endogenous C3G present in the cells is phosphorylated at the sites of location of the activated kinases. Pervanadate treatment has been shown to increase phosphotyrosine staining at the cell periphery indicating activation of kinases present in this subcellular domain [ 43 ]. Recently c-Src and Jak2 have been implicated in the phosphorylation of C3G in response to growth hormone stimulation of NIH 3T3 cells because dominant negative mutants of these kinases inhibit C3G phosphorylation [ 21 ]. It was suggested that this phosphorylation of endogenous C3G by c-Src occurs at Y504 because exogenously expressed Y504F mutant of C3G was not phosphorylated. Using a phosphospecific antibody we have directly shown the phosphorylation of endogenous C3G at Y504 upon overexpression of Hck [ 27 ] and c-Src (this report). c-Src is present in Cos-1 and HeLa cells which lack Hck. c-Src localizes to the cell membrane, focal adhesions and also to the Golgi [ 42 , 44 ]. Since pervanadate is a good activator of Src (36,37), it is possible that pY504 C3G seen in pervanadate treated cells is because of C3G being a Src substrate. Adhesion dependent Src activation leads to Rap-1 activation mediated by Crk and C3G [ 14 ]. Fibroblasts lacking C3G are essentially compromised in adhesion-mediated responses [ 25 , 26 ]. The localization of endogenous pY504 C3G at the subcortical actin cytoskeleton therefore suggests that this may be the site of action of C3G in mediating responses to cell adhesion. Modification of C3G by phosphorylation at defined subcellular domains may be important for restricted activation of C3G mediated signaling functions in the cells. Close structural and functional relationship is known to exist between the structural elements at the cell periphery and the signal transduction machinery. Several tyrosine kinases are known to be located in adherence junctions and the kinetics of phosphorylation and dephosphorylation appears to be controlled by structural molecules at the junctions. Our observation that disruption of actin cytoskeleton results in a loss of pY504 C3G staining at the cell periphery, but not at the Golgi complex reveals an important role for cytoskeletal network in the regulation of C3G. Conclusions The activity of guanine nucleotide exchange factor C3G is known to be regulated by tyrosine phosphorylation and membrane targeting. Using phospho-specific antibodies, we directly demonstrate that expression of Src family kinases or pervanadate treatment of cells induces phosphorylation of C3G on Y504. Unlike C3G, which is mostly cytosolic, pY504C3G locates to the Golgi and subcortical actin cytoskeleton. Demonstration of the localization of the active component of C3G to the Golgi and subcortical cytoskeleton provides evidence for a possible function for C3G at these cellular compartments. Methods Cell culture and treatment of cells HeLa and Cos-1 cells were cultured in DMEM supplemented with 10% FCS. Transfections were performed on cells grown as a monolayer in either 35 mm dishes or glass coverslips using the cationic lipid DHDEAB as described [ 45 ]. Briefly, 1 μl lipid diluted in 50 μl serum free DMEM was mixed with 1 μg DNA in 50 μl serum free DMEM. The mix was kept at room temperature for 30 min to allow complex formation before adding to the cell monolayer. Cells were fed with serum 5 hrs later and harvested 24–30 hrs after transfection. Cells were subjected to pervanadate treatment by the addition of a freshly prepared solution of pervanadate at 50 μM conc. for 10 min prior to harvesting. Pervanadate stock solution (50 mM) was prepared by mixing equal volumes of 100 mM solution of H 2 O 2 , with 100 mM solution of sodium orthovanadate. It was added to the cells within 5 mins of preparation. Golgi disruption was performed by treating the cells with 5 μg/ml of nocadazole for 30 min prior to fixation. To disrupt actin cytoskeleton, cells were treated with 1 μg/ml cytochalasin D for 20 mins. PP2 was added to cells 2 hrs before pervanadate treatment at a concentration of 10 μM to inhibit Src family kinases. Expression constructs Full length human C3G cloned in pcDNA3-FLAG was kindly provided by Dr S Tanaka. Y504F mutant of C3G in which tyrosine 504 is mutated to phenylalanine cloned in a His tagged expression vector was provided by Dr M Matsuda. The wild type rat p59 Hck cDNA was cloned in the pCI plasmid (Promega) and has been described earlier [ 27 ]. Expression plasmid for vesicular stomatitis virus glycoprotein as a GFP fusion protein (VSVG-GFP) was a kind gift from Jennifer Lippincott-Schwartz [ 32 ]. c-Src-GFP expression vector expressing wild type c-Src fused to GFP at C-terminal was from Dr D L Anders [ 46 ]. Wild type human Hck cDNA cloned into pCDNA6 expression vector was a kind gift of Dr Todd Miller [ 47 ]. Western blotting Whole cell lysates were prepared by lysing cells directly in Laemli's sample buffer and subjected to SDS-polyacrylamide gel electrophoresis. After transfer onto nitrocellulose membranes, they were processed for western blotting using the required primary antibodies. Detection was based on either color development using alkaline phosphatase conjugated secondary antibodies or on chemiluminescence using horse radish peroxidase conjugated secondaries. Indirect immunoflourescence and microscopy Cells were processed for immunoflourescence staining as described earlier [ 27 ]. The primary antibodies used were rabbit polyclonal anti-C3G (Santa Cruz Biotechnology), rabbit polyclonal anti pY504-C3G (SC-12926 R from Santa Cruz) and anti-Hck (3E9 monoclonal) made in our laboratory [ 48 ]. Dual labeling for Hck and C3G was performed by incubating the cells serially with C3G antibody, anti rabbit Cy3, monoclonal anti-Hck, and anti-mouse FITC. Cells were incubated with Oregon-green phalloidin after staining for pY504 with Cy3 to visualize F-actin. Cells transfected with vectors encoding GFP fusion proteins (GFP-Src or VSVG) were observed directly by fluorescent microscopy. Dual labeling for the C3G constructs and phospho-C3G was performed using the corresponding monoclonal tag antibodies (detected by FITC) and pY504 antibody (detected by Cy3). C3G was detected using Flag tag antibody (from Sigma) and Y504FC3G by His tag antibody (from Qiagen). Cells were examined using an Olympus microscope equipped with a cool SNAP color CCD camera. Images were captured using Image Pro Plus software. Immunoflurescence staining and colocalization was also observed using a Zeiss Axioplan 2 microscope fitted with an Apotome. The apotome (from Carl Zeiss Microimaging) is a new 3D imaging system for contrast enhancement in fluorescence microscopy. It uses structured illumination to reject signals belonging to regions of the sample that are outside the best focus position of the microscope. Images were captured using the Axiocam (Zeiss) CCD camera and processed using the Axiovision 4 software. Colocalization was also determined by observing the staining patterns using the LSM 510 Meta confocal microscope from Carl Zeiss. Abbreviations GNEF – guanine nucleotide exchange factor SFK – Src family kinase PV – Pervanadate pY504 C3G – Tyrosine 504 phosphorylated C3G VSVG – Vesicular stomatitis virus glycoprotein DMEM – Dulbecco's modified Eagle's medium FITC – Fluorescein isothiocyanate Authors contributions VR designed and carried out the experiments, analysed the data and drafted the manuscript. GS helped with designing the experiments, analyzing the data and writing the manuscript. AR provided technical help for western blotting and indirect immunoflourescence experiments.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC515295.xml
544958	Are zinc-bound metallothionein isoforms (I+II and III) involved in impaired thymulin production and thymic involution during ageing?	Background With advancing age, thymic efficiency shows progressive decline due to thymic involution allowing impaired cell-mediated immunity and the appearance of age-related diseases. The intrinsic cause of thymic involution is still undefined. Chronic inflammation and high glucocorticoids (GCs) may be involved. However, transgenic mice, with increased GC sensitivity and over expression of GC receptors, display delayed age-associated thymic involution. This fact suggests that other substances may affect thymic involution. Among them, both isoforms of metallothioneins (MTs) I+II and III are the major candidates because their increments leads to organ atrophy in constant stress and are induced by IL-6, which increases in ageing. Enhanced MTs in ageing allows constant sequester of zinc ions and no subsequent zinc release leading to low zinc ion bioavailability for thymic efficiency. This sequester is very limited in very old age. Thus, we have investigated the MTmRNA (I+II and III) in the thymus from young, old and very old mice. Methods MTmRNA and IL-6mRNA (RT-PCR) in the thymus from different donors were tested. Concomitantly, TECs proliferation, zinc ion bioavailability (ratio total thymulin/active thymulin), thymulin activity and corticosterone were tested from different donors. Results Both isoforms of MTmRNA and IL-6mRNA increase in old thymus coupled with low zinc ion bioavailability, reduced TECs proliferation, impaired thymulin activity and enhanced plasma corticosterone in comparison with young. Conversely, although the thymus is involuted in very old mice because of no changes in thymus weight in comparison to old mice, reduced MTmRNA, especially MT-I+II isoforms, and low IL6mRNA occur. Concomitantly, good zinc ion bioavailability, maintained TECs proliferation, satisfactory thymulin activity and reduced corticosterone are observed in very old mice. Conclusions The concomitant increments by high IL-6 of both MT isoforms in the thymus from old mice may be involved in thymic involution because provoking low zinc ion bioavailability, which is relevant for thymic efficiency. By contrast, the limited increments of MTs by low IL-6 induce good zinc ion bioavailability and satisfactory thymic efficiency in very old mice. Therefore, abnormal increased MTs may provoke complete thymic involution during ageing and the possible appearance of age-related diseases. If their increments are instead limited by low inflammation, healthy ageing and longevity may be reached.	Introduction The thymus gland is a central lymphoid organ in which bone marrow-derived T cell precursors undergo a complex process of maturation and differentiation leading to migration of positively selected thymocytes to the T cell-dependent peripheral areas [ 1 ]. Although thymocytes proliferation and differentiation persist throughout life, they diminish with ageing. Older thymuses are significantly atrophied and have fewer thymocytes than younger ones. Therefore, the thymus undergoes an age-dependent degenerative process, which allows a progressive loss of thymocytes as well as thymic lymphoid tissue becoming involuted, atrophic and full of fat [ 2 ]. Thymic involution is particularly important in relation to immunosenescence because leading to an impaired T cell-mediated immunity with the subsequent appearance of some age-related diseases [ 3 ]. The loss of thymocytes in ageing is also due, other than to diminished size of thymic cortex, to decreased production of thymic hormonal factors, which are important for thymocytes maturation, differentiation and proliferation [ 4 ]. Thymic hormonal factors, such as thymulin, thymopentin and thymosines, are produced by the Thymic Epithelial Cells (TECs), which number and proliferation decrease in ageing together with thymocytes [ 5 ]. The following one another of thymic negative events during ageing have been attributed to concomitant increments of glucocorticoids (GCs). Specific GCs receptors are present both on thymocytes and TECs leading the thymic cells to undergo apoptosis via Fas [ 6 ]. However, it has been recently reported in transgenic mice with increased GC sensitivity and over expression of GC receptors, a delayed age-associated thymic involution when compared with wild-type mice. These mice display a higher number of thymocytes and, surprisingly, thymic apoptosis is unaffected [ 7 ]. These data suggest that endogenous GCs may not be directly involved in thymic atrophy in ageing or, at least, they may act concomitantly or synergistically with other substances. In this context, some proteins, such as zinc-bound metallothioneins (MT) (isoforms I+II and III), may be involved in age-related thymic involution for the following reasons. Firstly, MT induction is controlled by GCs and pro-inflammatory cytokines (IL-6) [ 8 ], which increase in ageing and inflammation [ 9 ]; and also high IL-6 is involved in thymic dysregulation [ 10 ]. Second, MT increases in ageing and strictly related to high IL-6 and GCs [ 11 ]. Third, high MT are harmful in immunosenescence because they sequester zinc and are unable, within old lymphocytes, in the zinc release [ 11 ], which is in turn pivotal for immune efficiency and in conferring biological activity to thymulin [ 12 ], and thymulin activity, immune efficiency and free zinc ion bioavailability decrease in ageing [ 13 ]. Fourth, concomitant increments of MT-I+II and III during persistent stress like-conditions, as it occurs in ageing [ 14 ], lead to pancreas atrophy in stressed mice [ 15 ]. Although, MT-III isoform may be only present in the brain [ 16 ], its existence also in peripheral organs has been reported [ 17 ]. Following these considerations, we have investigated the presence of MT I+II and III gene expression and zinc content in the thymus from young, old and very old mice. Concomitantly, the IL-6 gene expression and TECs number and proliferation in the thymus from different donors have been evaluated as well as thymulin activity, corticosterone and zinc plasma levels. We have chosen very old mice because the MT (I+II) gene expression is low, like in younger, allowing satisfactory peripheral immune response [ 11 ]. Materials and Methods Mice Balb/c male inbred mice were used at the age of 2–3 months (young = n.10 mice), at the age of 20 months (old = n.10 mice) and at the age of 28–30 months (very old = n.10 mice). Although the maximum thymus sizes as well as peaks in thymocytes maturation and differentiation occur at 2–4 weeks of age in mice [ 18 ], no differences in thymulin activity and TECs number exist among 1 and 2–3 months of age [ 19 ]. Therefore, the choice of young mice at 2–3 months of age for the present study is appropriate. Mice were housed in plastic non-galvanized cages (5–6 mice for cage) and fed with standard pellet food (Nossan, Italy) and tap water ad libitum. Under our housing condition, the life span of Balb/c mice was of 30 months [ 13 ]. Since about 50% of survival occurred at 20 months of age, mice at this age were considered old [ 13 ]. Mice were maintained on a 12-h light/12-h dark cycle from 7:00 a.m. to 7:00 p.m. at constant temperature (20 ± 1°C) and humidity (50 ± 5%). Mice were sacrificed under ether anaesthesia. Heparinized blood samples were collected by cardiac puncture for plasma determinations of corticosterone, thymulin and zinc. Freshly thymuses were frozen in liquid nitrogen for MT-I+II, MT-III and IL-6 mRNA expressions and for testing zinc content. RNA isolation and RT-PCR analysis Total RNA was extracted from frozen thymus using Tri-Reagent according manufacture s protocol (Sigma, USA). 3 μg of RNA sample were reverse transcribed adding Olio d(T) and kept at 70°C for 10 min. dent, Raise inhibitor and MMLV reverse transcriptase were subsequently added and incubated at 37°C for 1 h. Samples were heated at 95°C to inactivate enzymes and stored at 20°C. PCRs were performed using sense and antisense primers as follows: MT-I: 5'-ATGGACCCCAACTGCTCCTGCTCCACC-3', 5'-GGGTGGAACTGTATAGGAAGACGCTGG-3' (259 bp) MT-III:5'-ATGGACCCTGAGACCTGCCCCTGTCCT-3', 5'-GGCCTCTGCCTTGGCCCCCTCTTCACC-3',(183 bp); β-actin: 5'-GGACTCCTATGTGGGTGACGAGG-3', 5'-GGGAGAGCATAGCCCTCGTAGAT-3' (366 bp); IL-6: 5'-ATGAAGTTCCTCTCTGCAAGAGACT-3', 5'-CACTAGGTTTGCCGAGTAGATCTC-3' (615 bp). Conditions for amplification were as follows: for MT-I each cycle consisted of 94°C 0.30 min, 50°C 0.30 min, 72°C 0.30 min with 30 cycles; for MT-III each cycle consisted 94°C 45 sec., 55°C 30 sec., 72°C 1.5 min with 30 cycles; for β-actin each cycle consisted 94°C 1 min, 61°C 1 min, 72°C 1 min with 24 cycles; for IL-6 each cycle consisted 94°C 1 min, 65°C 2 min, 72°C 3 min with 40 cycles. The products of amplification were size-fractionated by 2% agarose gel electrophoresis and visualized by staining with ethidium bromide. Semi-quantitative analysis of the amplified products was performed with an image analyser (Gel-doc 2000 instrument, Bio-Rad, USA). The results were evaluated as a relative unit determined by normalisation of the density of each band to that of the β-actin one. This method reflects MT protein production tested with Ag + saturation method [ 11 ]. Plasma Active Thymulin (AT) and total thymulin (TT) Plasma active zinc-bound thymulin (AT), as extensively described elsewhere [ 19 ], was measured using a bioassay based on the ability to restore the inhibitory effect of azathioprine on rosette formation in spleen cells from young Tx mice. Results were expressed as log -2 of the maximal dilution of tested plasma able to induce this phenomenon [ 19 ]. In order to avoid interference due to zinc, zinc sulphate at final concentration of 200 nM was added up to plasma samples. This fact shows the total amount of thymulin produced (active thymulin+ inactive thymulin) (TT) [ 19 ]. The ratio TT/AT is an index of zinc ion bioavailability because of strict inverse correlation between ratio TT/AT and plasma zinc levels. In particular, ratio >2 = low zinc ion bioavailability; ratio <2 = mild zinc ion bioavailability; ratio = 1 normal zinc ion bioavailability [ 19 ]. Plasma zinc and thymus zinc content Plasma and tissue zinc content were determined in Atomic Absorption Spectrophotometer (AAS) against zinc reference standards (Sigma USA). Plasma zinc was determined after plasma dilution 1:5. Thymus tissue (1 gr) was put in muffle furnace at 550°C overnight. The ash obtained was diluted with 3 ml of 3 N HCl and transferred to a 25 ml volumetric flask and further diluted with 3 ml of 0.36 N HCl. The determination of zinc was then performed at AAS. Plasma Corticosterone Plasma corticosterone level (ng/ml) was determined by RIA rat-corticosterone- 3 H kit (ICN Biomedicals, CA, USA) and referred against a standard curve. The percentage of cross-reaction with other steroid was <0.01. The sensitivity was of 0.05 ng/ml of corticosterone. Immunocytochemistry studies a) TECs characterization Anti pan-cytokeratin IgG1/FITC MoAb (Sigma, USA) diluted 1/25 and anti-keratin MoAb (Sigma, USA) diluted 1/20 were used. For this latter, guinea pig IgG/FITC (Sigma, USA) diluted 1/60 was used as second antibody. These MoAbs are specific to detect TECs (cortical and medullary) [ 20 ]. b) TEC separation and percentage TECs were separated with method described by Kurz et al. [ 20 ]. Briefly, the thymus from young, old and very old mice after 6 h of culture was minced into small fragments and incubated with collagenase (1 mg/ml, Sigma, USA) in PBS for 1 hr at 37°C (1 ml of collagenase solution/thymus). The choice of 6 h of culture is because the maximum thymulin production and TECs number and proliferation occurred at this time of culture in experiments of thymulin kinetic (from 1 h to 12 hrs) from young thymic cultures [ 21 , 22 ]. The suspension was then centrifuged (2 min, 400 g) and the pellet suspended in 1 ml of Dulbecco s modified Eagle medium/Ham s F12 medium (1:1) (DMEM/F12, Gibco, Germany). The cells were subjected to two-steps trypsin (0.1 and 0.25%, respectively) and 0.001% DNase treatment in order to avoid fibroblasts [ 19 ]. After three washes in PBS, the cells were dissociated by cautious triturating through Eppendorf tips and incubated in 3 ml of DMEM/F12 medium for 2–3 h at 37°C in humidified 5% CO 2 -atmosphere in order to make to adhere the cells. The supernatant containing unattached TEC was seeded into another plastic flask containing DMEM/F12 medium supplemented with 10% horse serum and put in culture in humidified 5% CO 2 -atmosphere. The cultures were inspected for morphologically visible fibroblasts (spindle shaped cells). In cases of significant contamination, the cells were washed with PBS and underwent again to trypsinization [ 20 ]. Separated TECs were washed three times in PBS. An aliquota (10 3 ) was resuspended in 1 ml of medium and underwent to TEC percentage analysis. Percentages of separated TECs were counted in 1.000 cells at fluorescence microscope [ 22 ]. Tests were performed after pre-fixation with cold methanol in the slides. Controls were performed without the primary antibodies. c) TECs proliferation After TECs separation, another aliquota (40 × 10 3 ) was resuspended in 4 ml of DMEM/F12 medium for TEC proliferation analysis, which was approached using [ 3 H] thymidine incorporation using 96 microtiter plates (Nunc, Denmark). 40 × 10 3 TECs were put in 40 wells (100 μl/well = 103 TECs/well). 10 wells were used as young; 10 wells as old; 10 wells as very old. Concomitantly, 1 μCi [ 3 H]-thymidine/well (Amersham, UK) was added. The plates were incubated in humidified 5%-CO2 atmosphere for 6 hrs. Automatic harvester collected the samples and the amount of incorporated radioactivity was determined in a liquid scintillation beta-counter (Perkin-Elmer, USA). Statistical analysis Two-tailed Student s t test, and ANOVA test (one-way) evaluated differences between means. Correlations were determined by linear regression analysis by the least square method. Differences were evaluated by analysis of covariance. Differences were significant when p < 0.05. Results MT-(I+II and III) and IL-6 mRNAs and zinc content in the thymus from young, old and very old mice Table 1 shows that MT-I+II and MT-III increase in old mice in comparison with young (p < 0.001). The same increment is also observed in very old mice as compared to young ones (p < 0.01), but at lower levels than old especially for MT-I+II. The increments of both isoforms of MT in old mice are correlated with high gene expression of IL-6 when compared to young mice (p < 0.01). The increments of IL-6 from the thymus of very old mice are lower, but still significant when compared to young (p < 0.05). Conversely, the zinc content within the thymus is very high in old mice as compared to young and very old mice (p < 0.01). Since AAS tests zinc-bound and zinc unbound [ 12 ], this last finding is not so surprising because it suggests that a large amount of zinc ions are bound to MT in the thymus from old mice. As a consequence, free zinc ions are not available for thymic efficiency in old age. Significant positive correlation exists between zinc content and MT-I+IImRNA from the thymus of young, old and very old mice (r = 0.83, p < 0.01). The thymus weight from old and very old mice is strongly reduced in comparison to young mice (p < 0.001), but with no changes between old and very old mice (Table 1 ). Table 1 MT-I+II, MT-III, IL-6 mRNAs and zinc content in the thymus from young, old and very old mice. Mice MT-I+II (MT-I/βactin) IL-6 (IL-6/βactin) MT-III (MT-III/βactin) Zinc content (μg/gr.) Absolute thymus weight (mg) Young 0.18 ± 0.02 0.14 ± 0.03 0.48 ± 0.02 62.3 ± 11.2 30.6 ± 5.0 Old 3.52 ± 0.3* 0.23 ± 0.02 § 1.65 ± 0.03* 107.4 ± 27.5** 13.6 ± 2.0 § Very old 1.29 ± 0.6 + 0.18 ± 0.04+ 1.63 ± 0.02* 77.4 ± 8.7 15.4 ± 2.3 § p < 0.001 when compared to young mice; + p < 0.01 when compared to old mice; p < 0.01 when compared to young and very old mice; § p < 0.01 when compared to young mice; ++ p < 0.05 when compared to old mice Thymic efficiency, plasma zinc and Corticosterone in young, old and very old mice Table 2 shows that thymulin activity is strongly reduced in old mice in comparison with young (p < 0.001). Thymulin activity is instead satisfactory in very old mice when compared to old ones (p < 0.05), even if its plasma value does not reach to that observed in young mice (Table 2 ). These data reflect the number (in percent) and the proliferation of TECs. Both TECs number and proliferation are reduced in old mice when compared to young and very old mice (p < 0.01), even if the TECs proliferation is lower in very old mice than in young ones, but still significant in comparison with old mice (p < 0.05) (Table 2 ). The proliferation data in young mice agree with testing TECs proliferation in pure murine TECs cell line (IT-45RI), as previously shown [ 21 ]. Table 2 Zinc ion bioavailability, corticosterone, thymulin and TECs number and proliferation in young, old and very old mice Mice Thymulin activity (log -2 ) AT/TT (log -2 (zinc ion bioavailability) Corticosterone (ng/ml) Plasma zinc (μg/dl) % TECs TECs proliferation (cpm) Young 5.5 ± 0.5 1.1 ± 0.3 153 ± 18.3 110 ± 11 53 ± 11 750 ± 25 Old 1.07 ± 0.3 3.0 ± 0.3 265 ± 16.6 80 ± 5.7 ++ 21 ± 7 227 ± 34 Very old 2.5 ± 0.3 + 1.0 ± 0.3 180 ± 12.4 + 87 ± 43 ++ 40 ± 12 + 450 ± 39 + p < 0.001 when compared to young mice; + p < 0.05 when compared to old mice; *p < 0.01 when compared to young and very old mice; ++ p < 0.01 when compared to young mice. The ratio total thymulin (TT)/active thymulin (AT) represents the zinc ion bioavailability. More high is the ratio (≥ 2) less zinc ion bioavailability is present, whereas ratio < 2 or equal to 1 means satisfactory or good zinc ion bioavailability, respectively [ 19 ]. The ratio TT/AT is higher in old mice in comparison with young and very old mice (p < 0.01) (Table 2 ). This fact means that a good zinc ion bioavailability exists in very old mice, as in younger ones, despite plasma zinc levels are lower in very old mice than in young ones (p < 0.01) and, at the same time, not different to those observed in old mice (Table 2 ). With regard to plasma Corticosterone, higher values are observed in old mice when compared to young and very old mice (p < 0.01) (Table 2 ). Significant inverse correlation exists between zinc and Corticosterone (r = -0.71, p < 0.01), whereas significant positive correlation exists between thymulin activity and TECs number and proliferation (r = 0.81, p < 0.01; r = 0.79, p < 0.01, respectively) from young, old and very old mice. Discussion Although MT-III isoform may be present exclusively within the brain [ 16 ], some peripheral organs (testis, prostate, epididymis, tongue, ovary, uterus, stomach, heart, pancreas and seminal vesicles) may also express MT-III isoform together with MT-I+II [ 17 ]. We herein present for the first time the concomitant gene expression of MT-I+II and MT-III also in the thymus. Both isoforms of zinc-bound MT (I+II and III)mRNA increase within the thymus of old mice, but with a minor extent of MT-I+II isoform in the thymus from very old mice. Concomitantly, the gene expression of IL-6 is higher in old mice than in young and very old ones. The zinc content within the thymus is enhanced in old mice respect to young and very old mice, whereas some thymic functions (thymulin activity and TECs number and proliferation) are impaired in old mice and preserved in very old mice. These last findings regarding to zinc content and thymic efficiency seems contradictory between old and very old mice. Really, they are not contradictory. Since AAS tests zinc-bound and zinc-unbound [ 13 ], the higher zinc content in the old thymus is largely due to high zinc-bound MTs, which recall zinc from the periphery, via zinc transporters ZnT1–4 [ 23 ], and sequester a lot amount of zinc ions [ 22 ]. The inflammation provokes zinc loss with subsequent impairment of immune response [ 24 ]. Thus, such a recall and sequester by MT are due to the great inflammation by high IL-6 and GCs because zinc ions have not to be lost. But, free zinc ions are subsequently not available for thymic efficiency due to inability of MT in zinc release in constant inflammation [ 14 ]. Conversely, limited recall of zinc ions occurs in very old mice because the inflammation is less deep. As a consequence, the zinc content in the thymus from very old mice is lower and, at the same time, more free zinc ions are available for thymic efficiency by low zinc-bound MT. Anyway, the present data show that abnormal increments of both isoforms of MT are present in the atrophic thymus from old and very old mice, but with less extent of MT-I+II in very old mice. MT-I+II and III are expressed in the brain with a balance between the two isoforms (25). When one isoform increases, the other decreases due to a possible genetic control of MRE region on the chromosome 8, which brings the two isoforms [ 26 ]. Concomitant increments of the two isoforms within the hippocampus from old rats leads to impaired number and functions of synapses coupled with low zinc ion bioavailability [ 27 ]; and synaptic function is zinc-dependent [ 28 ]. The same phenomena occur in age-related neurodegenerative diseases [ 29 ], suggesting a possible role of increased MT-I+II and III in neurodegeneration [ 27 ]. Moreover, the concomitant presence of MT isoforms provokes the atrophy of the pancreas in stressed mice [ 15 ]. Therefore, enhanced MT I+II and III in the old thymus may lead to the thymic involution and atrophy because of an unbalance between the two MT isoforms. The mechanism may be largely due to the constant sequester of zinc ions by MT (I+II and III) with no subsequent zinc release leading to low zinc ion bioavailability for thymic endocrine activity and TECs proliferation [ 13 ]. In this context, IL-6 and glucocorticoids (GCs) may play key roles because IL-6 and GCs affect MTmRNA [ 8 ] and, in turn, abnormal high GC levels are involved in thymic atrophy through the activation of GC receptors on TECs and thymocytes [ 6 ]. A lack of free zinc ions also provokes thymic atrophy [ 30 ]. We have found enhanced IL-6mRNA within the thymus from old mice. Concomitantly, strong increments of plasma corticosterone are observed. Such increments are strictly related to high MTmRNA and low zinc ion bioavailability. These findings suggest that the chronic inflammation, via IL-6 and GCs, allows high MTs induction, low zinc ion bioavailability and subsequent thymic atrophy in old mice. The thymus is obviously atrophic in very old mice. But, the MT-I+IImRNA and IL-6mRNA are lower than old mice as well as reduced corticosterone. High corticosterone provokes zinc loss by urine and faeces [ 31 ]. Corticosterone is low in very old mice (Table 2 ). Thus, very old mice display more free zinc ion bioavailability with subsequent more thymic efficiency and preserved TECs number and proliferation. TECs produce thymulin, a zinc-dependent thymic hormone [ 12 ]. Zinc-bound MTs transfer zinc to thymulin in TECs [ 32 ]. The less MT-I+IImRNA in very old mice may thus allow less sequester of zinc ions. Alternatively, an easier release of zinc by MT for thymic efficiency might occur due to reduced inflammation by low IL-6mRNA and corticosterone. Anyway, the thymus from very old mice is still efficient despite it is involuted because the thymus weight between old and very old mice does not change (Table 1 ). This means that the thymic reconstitution in old age might not be necessary because the age-related loss of thymic efficiency appears to be only quantitative and not qualitative [ 33 ]. Thus, it might be sufficient to maintain inflammatory status and MTs homeostasis below a critical threshold in order to preserve thymic efficiency. Further experiments in thymic output from very old mice and in genes involved in thymocytes maturation and differentiation (Rag 1 and Rag 2) [ 2 ] are in progress in our lab. On the other hand, MT-ImRNA is lower in lymphocytes from human nonagenarians coupled with satisfactory thymic and peripheral immune efficiency and good zinc ion bioavailability [ 11 ]. Moreover, very old mice display still efficient thymic functions during liver regeneration after partial hepatectomy (model of acute and constant inflammation) [ 34 ]. In addition, the presence of involuted thymus in stressed MT transgenic mice [ 22 ] further supports the involvement of high MT in thymic involution during ageing. However, the thymic involution is not irreversible phenomena because zinc treatment in old mice restores thymic efficiency with a re-growth of thymic cortex [ 19 ]. This finding suggests that in ageing the thymus is in quiescent phase, which is less deep in very old age due probably to more zinc ion bioavailability, via MTs homeostasis, and less inflammation. Indeed, zinc also affects cell cycle and, therefore, cellular proliferation [ 35 ]. Thus, the thymus from very old mice is still active and not quiescent. The satisfactory zinc ion bioavailability coupled with the maintenance of TECs proliferation in the thymus from very old mice is in line with this interpretation. This means that in vivo condition TECs from very old mice are still capable to proliferate at the occurrence, for example in presence of external noxae in order to have a sufficient thymulin production for a prompt immune response, as occurring in human centenarians, who are high responder individuals showing a great capacity in remodelling thymulin activity [ 36 ]. In conclusion, concomitant increments of zinc-bound MT-I+II and III within the thymus may lead to thymic involution in ageing because they sequester zinc and are unable in the subsequent zinc release, which is indispensable for thymic efficiency. The cause may be related to chronic inflammation by high IL-6 and GCs. Without excluding a direct role of GCs in thymic involution [ 6 ], GCs may synergistically act with MT isoforms because GCs also affect MTmRNA [ 8 ]. Less inflammation by low IL-6 and GCs in very old mice allows reduced MTmRNA with subsequent satisfactory zinc ion bioavailability and, therefore, preserved thymic efficiency. However, if the thymic involution may be a necessary event in order to avoid autoimmune phenomena in ageing is an intriguing point to be investigated. Indeed, high IL-6 provokes an enlargement of the thymus with the appearance of autoimmune phenomena [ 10 ]. Thus, if on one hand high MTs may be of protection inducing thymic involution in order to escape autoimmune phenomena by high IL-6, on the other hand high MT are harmful because leading to low zinc ion bioavailability for thymic efficiency. However, in this context, the involvement of zinc and MT in the efficiency of extrathymic T-cell pathway [ 13 ] has to be also considered, because this pathway is prominent in ageing and autoimmunity in order to compensate thymic failure [ 37 ]. Very old mice display satisfactory thymic efficiency (present study) and good extrathymic T-cell functions [ 38 ]. Moreover, thyroid autoantibodies are rare in healthy centenarians [ 39 ]. Therefore, the thymic involution, via MTs homeostasis, has to be limited or controlled concomitantly with the appearance of efficient extrathymic T-cell functions in order to reach healthy ageing and longevity. In other words, a correct balance between thymic involution and extrathymic T-cell functions has to exist in ageing. Otherwise, a complete thymic atrophy by abnormal high MT-I+II and III, via high IL-6 may provoke continuous immune dysfunctions (thymic and extrathymic). Altered genetic controls between the two MT isoforms on MRE region of chromosome 8 may be involved, representing an interesting field of investigation in immunosenescence. Works are in progress in our lab.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544958.xml
549529	Perils of paradigm: Complexity, policy design, and the Endocrine Disruptor Screening Program	The Endocrine Disruptor Screening Program (EDSP), mandated by the United States Congress in the Food Quality Protection Act of 1996, attempts to protect public health from adverse endocrine effects of synthetic chemical compounds by establishing a new testing regime. But the complexities and uncertainties of endocrine disruption and its broader regulatory and social context all but ensure the failure of this policy. This article addresses the issues facing EDSP comprehensively and in detail, in order to move beyond the current regulatory paradigm and foster discourse on a positive role for scientists in support of EDSP's end goal: to protect public health.	Introduction The United States Environmental Protection Agency (EPA) created the Endocrine Disruptor Screening Program (EDSP) to regulate endocrine-disrupting chemicals (EDCs) as mandated in the Food Quality Protection Act of 1996 (FQPA) and the Safe Drinking Water Amendments Act of 1996 (SDWAA). Unlike the more easily appreciated effects of toxic chemicals, EDCs interact with the human body indirectly by mimicking, blocking, or otherwise disrupting the normal function of hormones. The prolific study of endocrine disruption has since uncovered many unconventional and worrisome mechanisms, exposures, and effects [ 1 ]. The goal of EDSP is to: [D]evelop a screening program, using appropriate validated test systems and other scientifically relevant information, to determine whether certain substances may have an effect in humans that is similar to an effect produced by a naturally occurring estrogen, or such other endocrine effects as the [EPA] Administrator may designate [ 2 ]. If such an effect is discovered, "the [EPA] Administrator shall, as appropriate, take action under such statutory authority as is available...as is necessary to ensure the protection of public health" [ 2 ]. Unfortunately, due to four complicating factors, EDSP cannot protect public health. The first complication is practical considerations. EPA estimates the universe of potential EDCs numbers more than 87,000 items. Testing this many chemicals would take an unreasonable investment of time and resources, but even scientifically prioritizing among them is highly problematic. The second complication is hazard complexity. Establishing relationships between EDCs and health hazards proves very difficult if not impossible. Endocrine-disrupting action breaks all the rules and assumptions that have guided toxicology through the era of modern chemical regulation. Without these simplifying assumptions, science cannot establish causation efficiently or with sufficient certainty for regulation. The third complication is exposure complexity. Determining exposure levels becomes more important and more difficult because the low-dose effects of many EDCs means that low-dose and transient exposure can be just as or more dangerous than high-dose and prolonged exposure. Assessments typically discount these ill-defined exposures, but we can no longer assume them insignificant. The final complication is regulatory deficiencies. Although FQPA and SDWAA provided new authority to test for endocrine disruption, they provided no new authority for the regulation of EDCs. As a result, multiple government agencies must manage future test-positive EDCs under their jurisdiction using fragmentary and incomplete statutory authorities and different regulatory standards. This introduces significant confusion to the institutional and decision-making aspects of the EDSP regulatory framework. The EDSP policy design represents revision at the margins of U.S. chemical regulatory policy, not a radical revision. EDSP employs the same basic strategy used to regulate carcinogenic pesticides or toxic industrial chemicals – scientifically proving harm prior to regulating a chemical. Two important aspects of this strategy include an epistemological assumption that science has the capacity to 'prove' harm under the relevant scientific and legal standards, and an ethical position that prioritizes profit over human health by placing the burden of proof on public and environmental health advocates. These assumptions remain all but unchallenged in the U.S. context, and thus comprise a paradigm. While this paradigm has faced some critique in the context of carcinogenic pesticides and toxic industrial chemicals, questions of its efficacy remain unresolved. Because EDCs present new and fundamental difficulties for the science underlying the regulatory paradigm, a critical analysis of EDSP provides a more compelling case that the current chemical regulatory paradigm is in need of radical revision. This study investigates the policy design of EDSP and its broader context. The above four complications play varying roles in each stage of the EDSP policy design as discussed below. See figure 1 for a diagram roughly depicting the relationship between the four complications and the policy stages of EDSP. After considering the complications of each policy stage, this study briefly discusses the role of politics in regulation before considering the implications for the conventional chemical regulatory paradigm and a positive role for scientists in support of EDSP's end goal to protect public health. Figure 1 EDSP complications and policy stages relationship The following discussion provides a comprehensive empirical basis for considering alternatives to the status quo. This study aims to integrate the many factors conditioning the failure of EDSP for the purpose of fostering constructive discussion on U.S. regulatory policy concerning EDCs and chemicals more generally. The author does not possess a unique answer to the many and complicated issues surrounding endocrine disruption and the U.S. chemical regulatory paradigm. Given that no simple, well-developed alternatives exist that merit immediate consideration by decision-makers, it stands to reason that more creative and open discussions of EDCs, the chemical regulatory paradigm, and possible roles for the scientific community may provide long-term payoffs in public and environmental health protection well worth our attentions today. Discussion The design of EDSP consists of three main stages: priority setting, screening and testing, and a risk analysis leading to potential regulation. The complications shown in figure 1 and detailed throughout this study undermine each of these stages. Before dealing explicitly with these policy stages, however, some practical considerations deserve note. EPA estimates 87,000 chemicals require testing as potential EDCs, including pesticide chemicals, non-pesticide commercial chemicals, cosmetic ingredients, food additives, nutritional supplements, mixtures, and environmental contaminants [ 3 , 4 ]. This sets a daunting task; no U.S. chemical regulatory program has ever successfully tested so many chemicals. A quote from U.S. Congressman Mike Synar (D-OK), during a committee hearing on the safety of pesticides in foods states the problem dramatically: "Almost 20,000 pesticide products have been under review since 1972 and only 31 have been re-registered. At this rate it will take us to the year 15,520 A.D. to complete. I believe in good science. What I don't believe in is geologic time" [ 5 ]. Other researchers and watchdogs note the failure of other U.S. chemical regulatory programs to effectively gather information or protect public health (e.g. TSCA [ 6 , 7 ], and FQPA [ 8 ]; for a broader critique [ 9 , 10 ]). By applying Congressman Synar's analysis to endocrine-disrupting chemicals, we can expect characterization of all potential EDCs to take 59,000 years. EPA stated: "Testing of all of these chemicals cannot be supported at the same time because, even if EPA and industry had the resources to do so, there are not enough laboratories or other facilities capable of conducting the testing" [ 11 ]. In other words, there is reason and precedent to doubt our ability to accomplish this feat. More importantly, despite our best efforts to mobilize science in support of this difficult task, policy mechanisms allow chemical use and abuse to go forward regardless of scientific results or lack thereof. "Pesticides are registered for use while important health and safety data are still being generated; they may continue to be used after evidence of their hazards is given to EPA; they may be registered through alternative processes that bypass important tests; and they may never be required to be tested for certain kinds of hazards" ([ 9 ], also see [ 12 ]). Additionally, politics often plays a greater role in the decision to regulate than science (see Politics section below). Politics and policy design play significant roles in the modern chemical regulatory regime. Hence, a comprehensive analysis of EDC regulation must take politics and policy design as well as science into account. The appropriate standard by which to judge these disparate policy elements is Congress's mandated end goal: to protect public health. We will return to this issue. Priority setting After EPA sorts chemicals according to statutory considerations, data availability, and qualitative judgment, EPA decides which of the estimated 87,000 chemicals merit consideration first through 'priority setting' [ 13 ]. The sorting of chemicals into the four categories in figure 2 and the setting of priorities within 'Category 2' require functionally equivalent information (Note figure 2 disaggregates the policy stages from the right hand column of figure 1 ). So priority setting, as used in this article, applies to both EDSP activities described as sorting and priority setting (i.e. everything above the dashed line in figure 2 ). The Endocrine Disruptor Screening and Testing Advisory Committee (EDSTAC) defined 'priority' as 'of greatest concern' in their final report, presumably as determined by science [ 14 ]. EPA, however, added statutory criteria to the scientific considerations by necessity: "EPA plans to use three main categories of information to set priorities: exposure-related information, effects-related information, and statutory criteria" [ 15 ]. Setting aside the chemicals Congress mandated EPA test (the statutory criteria), how well can EPA scientifically set priorities among potential EDCs? Figure 2 EDSP policy design EPA simply cannot set priorities based on science alone. Almost no data on potential endocrine disruption exists for the 87,000 prospective EDCs, creating a catch-22 of needing unavailable information to decide how to gather information. EPA wants to prioritize which chemicals to develop data on by examining hazard and exposure data on those chemicals. In the information poor environment of endocrine disruption, EPA has no basis to commence setting priorities. Two methodologies, high-throughput pre-screening (HTPS) and quantitative structure activity relationships (QSAR) both discussed below, have attracted attention and resources due to a general recognition of this problem. But more important than our current lack of data, EDCs operate with a high degree of complexity. Because of system complexity, some uncertainty about endocrine disruption probably cannot be resolved – resulting in some abiding doubt about the significance of a chemical as a potential endocrine disruptor. The scientific community and EPA seem quite cognizant of this complexity, but its relevance for policy and for the end goal of protecting public health deserves careful attention. EPA has not realized their ideal of priority setting based on hazard and exposure information because of the catch-22 mentioned above. As a result, "EPA's proposed approach focuses on human exposure-related factors rather than on a combination of exposure- and [hazard]-related factors" [ 16 ]. While this statement acknowledges some of the difficulty using scientific information for priority setting, it is misleading. EPA's current stated policy prioritizes only pesticide active ingredients and high production volume (HPV) pesticidal inerts not because of scientific criteria (hazard- or exposure-related), but for statutory reasons; Congress specifically mandated testing of these compounds [ 2 , 17 ]. In fact, the concept of setting priorities for potential EDCs based only on exposure-related factors is fundamentally flawed. To set priorities based on exposure factors alone, one must assume greater exposure to a chemical implies greater potential hazard (or some other arbitrary assumption). For EDCs, this assumption is scientifically insupportable. The complexity of low-dose effects (discussed in more detail in the Screening and testing section below) implies that exposure to some EDCs at extremely dilute doses may have a greater effect than massive exposure to that same chemical. Transient or low-concentration EDCs may also pose a greater risk than other high-exposure chemicals. Low-dose effects and other exposure complexities make exposure alone a poor proxy for setting priorities. Since different vulnerabilities and sometimes different health effects manifest at different developmental stages, any exposure-only judgment will run into significant difficulties defining spatial and temporal boundaries for exposure determinations. Some short-lived chemicals may have important endocrine- disrupting effects, but may not show up in EPA's most robust exposure data sources: biological sampling and environmental monitoring [ 15 ]. Further complexities undermine scientific determinations of exposure. Maternal metabolism of fat stores containing bioaccumulated EDCs may lead to practically unidentifiable fetal exposure. Some poorly understood exposure sources, such as flame retardants in clothing and furniture or phthalates leaching from plastics, would be extremely difficult to determine because of the complex social, cultural, and ecological conditions that affect chemical release and exposure. Even conventional exposure determinations, such as ingested pesticides, are fundamentally dependent upon patterns of food consumption. Averaging exposure may obscure vulnerabilities brought on by complicated cultural, social, and economic patterns of food consumption and other subpopulation attributes or behaviors. For example, research has shown significant differences in the exposure of adults and children to certain pesticides via residual contamination of fresh and processed foods [ 18 , 19 ]. FQPA may further obscure exposure determinations through mandates requiring EPA to assess cumulative exposure, including all exposure routes and sources, all chemicals with similar modes of action, and other mixtures of multiple chemicals. The complexity of endocrine disruption undermines old assumptions about the relevance of exposure and prevents scientifically meaningful prioritization on the basis of exposure data alone. Understanding this limitation to some degree, EPA continues to develop and evaluate two methodologies to include health-effects criteria in the prioritization process. The first method is high-throughput pre-screening, or HTPS. This method allows for fast, large-scale testing of chemicals for interactions with estrogen, androgen, and now thyroid receptors. HTPS, unfortunately, has flaws as a means of detecting potential hazard for priority setting. Most basically, HTPS only tests for hormone receptor interactions. The possibility of this leading to a systematic bias against consideration of non-receptor mediated endocrine disruption is significant. Receptor interaction is only one means by which a chemical can disrupt the endocrine system. Interaction with the hormone molecules themselves, stimulation or suppression of hormone production, and disruption of old hormone metabolism can all lead to endocrine-disrupting effects as well. HTPS cannot test for these effects. Other hazard-related shortcomings relevant to HTPS are discussed more thoroughly in the Screening and testing section of this study. An EPA feasibility study cited some of the same issues raised here in declaring HTPS insufficient for regulatory purposes [ 20 ]. A second methodology under development is a computer modeling technique called quantitative structure activity relationships, or QSAR. QSAR simulates the behavior of a chemical based on its structure. EPA would use QSAR to predict chemical binding with estrogen, androgen, and thyroid receptors. The dominant criticism of HTPS applies to QSAR as well – it tests for receptor binding only. However, the use of computer models will incorporate new uncertainties via the selection of system boundaries and functional relationships that may preclude mechanisms and variables relevant to some endocrine-disrupting action. While a modeling effort may yield useful knowledge, as a decision-making tool QSAR is wanting. The inevitable and likely widespread false positive and false negative results will demand a parallel testing procedure to establish QSAR's utility for priority setting. But the drive to develop QSAR derives from an inability to devise an efficient and reliable testing procedure (like HTPS) in the first place. While these limitations may or may not be overcome in time, at present the methodology is not useful for setting priorities. It is instructive, however, to consider the justifications for the development of QSAR. "Systematic toxicity testing, using conventional toxicology methodologies, of single chemicals and chemical mixtures is highly impractical because of the immense numbers of chemicals and chemical mixtures involved and the limited scientific resources" [ 21 ]. QSAR was developed as an attempt to solve the very problems cited in the Policy Design section above. Although models can provide much useful information, they are unlikely to help prioritize EDCs anytime soon. Screening and testing To amass the evidence necessary for regulation, EPA designed two tiers of scientific assays. Tier 1 screening involves short-term assays to detect potential chemical interaction with the endocrine system. Tier 2 testing involves long-term assays to establish such interactions, explore more complicated endpoints, and establish dose-response relationships. If enough data exists, a chemical can go straight to Tier 2 testing. Otherwise chemicals are assigned to Tier 1, where chemicals are prioritized and screened, with all positive results forwarded for Tier 2 testing (see figure 2 for the policy design). "The Tier 2 tests are longer in duration than Tier 1 tests and are designed to encompass critical life stages and processes as well as a broad range of doses, and are intended to be administrated by a relevant route of exposure" [ 16 ]. Although screening and testing are separate EDSP regulatory stages, their vulnerabilities to complexity are similar enough to group them together for purposes of this discussion. Both screening and testing focus on identifying hazard, leaving exposure considerations for the final risk assessment. As such, this discussion addresses only hazard-related complexities and uncertainties. The toxicology of endocrine disruption is inherently complex in the sense that scientists must abandon the simplifying assumptions of standard toxicology. Most notably, we must abandon the assumption of monotonic dose-response relationships, which assume an increased exposure to a substance always leads to an increase in effect. Increasing exposure to some EDCs swamps the endocrine system and prevents or reduces dysfunction (i.e. an inverted U dose-response, e.g. [ 22 , 23 ]), while other EDCs exhibit effects at both high- and low-doses, but not in between (i.e. a U- or J-shaped dose-response, e.g. [ 24 ]); still others may exhibit hormesis, whereby a small dose has a beneficial effect [ 25 , 26 ]. The monotonic assumption allows for statistically significant results using smaller sample sizes exposed to higher doses for shorter periods of time. Linearly scaling these results down to typical exposure levels presumably yields approximate quantitative rates of, for example, disease or cancer. Abandoning this assumption decreases testing efficiency and multiplies the time and other resources necessary to understand the potential hazard posed by a chemical. A quote from University of Washington, Seattle toxicologist David Eaton, states the issue simply: "It's just too expensive ... you'll never be able to characterize [a low-dose effect] to the point where people think it's real" [ 25 ]. Non-monotonic dose-responses may also indicate some unresolvable system complexity. Other standard toxicological assumptions suffer the same fate as monotonic dose-response, for example: the threshold assumption and the assumption that a chemical has a uniform effect. Other factors complicate a scientific determination of hazard. Two chemicals can interact in ways that alter their effects. Some chemicals together inhibit their individual effects, reducing or preventing an adverse effect where one is expected. Others simply add their effects together, and yet others interact synergistically, magnifying the effect either or both would normally have. "Synergistic interactions are the most problematic, because they indicate that the effects of multiple chemicals together can be significantly more powerful than might be predicted simply by adding up their effects one at a time. Regulatory science rarely incorporates any interactions; it is incapable, at present, of coping with synergies" [ 27 ]. Regardless of this incapacity, EPA seems determined to try and deal with this complexity: "EPA recognizes that the science of evaluating mixtures remains complex and unclear, but believes that it should begin to confront the issues raised by them" [ 28 ]. Additionally, scientists have evidenced possible synergism between EDCs and infectious disease agents [ 29 ]. Synergies with nutrients or poor nutrient levels might also prove significant (e.g. lead, [ 30 ]). These interaction effects further aggravate the difficulty of determining hazard. Several studies of the body burden of chemicals in humans evidence high and diverse concentrations of synthetic chemicals, indicating the importance and likelihood of chemical interactions [ 31 - 33 ]. Another serious complication involves the selection of testing endpoints, or dysfunctions possibly caused by EDCs. Some of the dysfunctions already identified through animal studies include cancer susceptibility and birth defects, but also more subtle endpoints like immunological dysfunction, suppression of secondary sex characteristics, decreased fertility, increased aggression, decreased mental capacity and focus, disrupted brain development, etc [ 34 , 35 ]. While scientists can examine some of these endpoints relatively easily in human populations (e.g. cancer incidence), others would be incredibly difficult to observe, measure, or prove with sufficient statistical certainty (e.g. feminization of boys or masculinization of girls). The difficulty of isolating and measuring these more subtle effects makes them impractical as regulatory endpoints. The inability of scientific testing to measure such endpoints, however, does not justify their exclusion from regulatory consideration. Such a policy would (and in fact does) bias the regulation of chemicals by exempting the most complicated chemicals and the most complex health effects from regulatory consideration. Risk analysis and regulation A risk analysis concludes the EDSP policy design. EPA claims it will use its standard human health risk assessment process for EDCs [ 36 ]. Simply put, EPA considers hazard and exposure data and uncertainties to make regulatory decisions (see bottom figure 2 to right of dashed line). For example, an extremely hazardous chemical associated with insignificant exposure probably would not require regulation while a mildly hazardous chemical with widespread and pervasive exposure probably would. Safety factors are built into this process to protect public health. The standard safety factor for pesticides is 100× to compensate for uncertainties such as response differences between humans and the animals studied. FQPA added an additional 10× safety factor to protect children, but EPA's uses this additional safety factor inconsistently [ 8 ]. Because EPA's risk analysis demands an explicit integration of hazard and exposure data, the risk assessment itself is vulnerable to the exposure complexities treated in the Priority setting section and the hazard complexities treated in the Screening and testing section. These complexities include: HAZARD – low-dose effects, mixtures and synergies, and uncertain endpoints; and EXPOSURE – transient and low-concentration exposure to EDCs, maternal metabolism of bioaccumulated EDCs, varying vulnerability and response by developmental stage, poorly understood exposure sources, vulnerable subpopulations, and cultural, social, and economic patterns and spatial and temporal bounds of chemical release and exposure. More science cannot resolve all of these complexities and uncertainties. But these complexities translate into an even more pernicious difficulty, as decisions assumed answerable by science must be made under conditions of scientific ambiguity. Ambiguity is related to complexity as follows: Complexity refers to properties of the system under study, not the study itself. Such properties include interactive effects (like synergies), feedback loops, temporal delays between cause and effect, chaotic or stochastic system behavior, large numbers of intervening variables, and inter-individual variations. Uncertainty refers to limitations of the human analysis of a complex system. While science is often employed 'to reduce uncertainty,' the complexity of a system sets bounds on the prospective certainty of informed scientific judgment. For example, further science can never remove the chaos from a chaotic system or the randomness from a stochastic one. And to illustrate the point, exposure of a fetus to environmental EDC contamination in utero might easily be a chaotic or stochastic system. Ambiguity refers to a situation in which existing scientific data support equally valid but competing interpretations of risk (see [ 37 ] for a discussion of concepts). Ambiguity is the single greatest limitation to the use of endocrine disruption science in policy. While an oft-cited truism holds that decision-makers can make better decisions with reliable information at hand, this hardly tells the whole story. Other considerations play into decision-making, including value tradeoffs, time and resource limitations, and informational constraints, including scientific ambiguity. A National Academy of Sciences report on endocrine disruption substantiates the ambiguity in this field: [I]t became clear as the work of the committee progressed that the same data could be approached from different viewpoints. Those different views led to different judgments among the committee members about the significance of the threat posed by [EDCs]. In [some] cases, the differences do not reflect the need for research but reflect differing judgments about the significance of information. The differences are not confined to the members of this committee but are also reflected in the scientific community at large and in the comments received during review [ 35 ]. The recommendations of this committee amounted to suggestions for more research. While more research might be a good thing, such a suggestion provides no guidance on how to address the more policy-relevant question of how to make decisions with ambiguous scientific information. In a nutshell, more science is not a panacea for all the problems of risk analysis or decision-making. This brings us to making decisions about regulation. If a reasonably certain determination of harm about an EDC could somehow be made, how would one regulate that chemical? This discussion of regulation is only tentative since EPA's Regulatory Activities Workgroup continues to study the issue. Since EPA has not yet validated any screening or testing procedures [ 38 ], regulation under EDSP has so far received little attention. Under section 408p of FQPA, EPA must use "such statutory authority as is available " to protect public health [ 2 ] (emphasis added). In other words, the statutes requiring testing for endocrine disruption provide no new process by which to regulate those chemicals – the standards for regulation remain those of previous regulatory laws. Unfortunately, this complicates enforcement authority for EDCs. Regulation must be authorized under one of four laws: the Toxic Substances Control Act (TSCA), the Federal Food, Drug, and Cosmetic Act (FFDCA), the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), or the Safe Drinking Water Act (SDWA). The agencies with regulatory jurisdiction for EPA's list of 87,000 chemicals are: the Environmental Protection Agency (EPA), the Food and Drug Administration (FDA) in the U.S. Department of Health and Human Services, and the Food Safety and Inspection Service (FSIS) in the U.S. Department of Agriculture. EDSTAC recommended and EPA adopted the following list of chemicals for endocrine disruptor testing: 75,500 commercial chemicals listed under TSCA, 900 pesticide active ingredients, 2,500 pesticide inert ingredients, 5,000 cosmetic ingredients, 3,000 food additives, an unspecified number of nutritional substances, and an unspecified number of natural hormonally active plant residues [ 14 ]. Testing and enforcement authority for this universe of chemicals is fragmentary. For example, EPA has authority under FIFRA and FFDCA to set tolerances for pesticides on food, but enforcement authority falls to FSIS for meat and poultry products, and FDA for other food items. The authority is also incomplete. For example, FDA has authority over the estimated 5,000 cosmetic chemicals, but FDA has no authority to require any information from the manufacturer or to mandate product safety testing. FDA's regulatory authority over cosmetics begins only after a product (possibly without any safety information) enters the market. Additionally, the standards by which to regulate differ. Under TSCA and SDWA the economic costs of regulation must be balanced against the public health threat, but under FIFRA and FFDCA, economics can be considered in only narrowly crafted situations – the standard is largely health-only based. FFDCA and FIFRA as amended by FQPA use the "reasonable certainty of no harm" standard. This standard translates into a 95% certainty that fewer than one in a million additional cancer deaths will occur due to the expected exposure over a lifetime. No translation specific to EDCs of this standard is yet available. TSCA, on the other hand, must prevent "unreasonable risk" of injury to health or the environment. A risk is "unreasonable" if the risks exceed the benefits associated with that activity, including economic benefits. SDWA explicitly requires a consideration of the cost of compliance to state, local, and other water systems when setting safety standards. The estimated 2,500 pesticide inert ingredients may defy regulation due to trade secret norms, poor EPA data quality, and historic government neglect [ 39 , 40 ]. Additionally, no federal statute delegates any authority at all for the testing or regulation of nutritional supplements. Presumably, endocrine-disrupting nutritional supplements could be regulated only voluntarily, but the onus of testing would fall completely upon the executive agency that volunteers to expand its responsibilities. Significant difficulties involving confidential business information, including proprietary chemicals and chemical mixtures may further compromise enforcement capability. The complexity of the institutional and legal landscape (multiple interacting agencies with multiple overlapping mandates and authority) creates substantial regulatory confusion. In this situation, the powers and responsibilities of different government agencies might be interpreted differently by other agencies or by affected parties. Such confusion leaves room for interpretation that may require long delays and intensive court battles to resolve. The history of TSCA indicates that such confusion (for TSCA, 'balancing economic cost' with regulations to protect public health) as well as the menace of legal action can lead to crippling regulatory inaction [ 41 ]. By requiring enforcement under existing statutory authority, FQPA leaves the regulation of the already complicated universe of EDCs to a complicated web of regulatory regimes of questionable efficacy. Politics Finally, the role of politics in regulatory decision-making deserves note. The conventional ideal of regulation under the current paradigm is that good science leads to an informed decision-maker who can then remove or limit a proven hazardous chemical from commerce (or, rarely, prevent its introduction). The complexities of endocrine disruption science, practical considerations, and the regulatory deficiencies discussed above impose limitations on this conventional ideal. Neglecting the role of politics in regulatory decision-making, however, is perhaps this ideal's most significant omission. A variety of academic and government research as well as environmental and public interest group analysis points to the failure of testing regimes to produce significant regulation or protect public health (e.g. [ 7 , 8 , 10 , 12 , 41 , 42 ]). In fact, past regulation often addressed specific chemicals by legislative mandate (e.g. the mandated ban of PCBs in TSCA) or due to media-promoted public awareness and its resultant outcry (e.g. the January 1971 court order essentially forcing EPA Administrator Ruckelshaus to ban DDT). In other words, politics often leads to regulation regardless of scientific considerations. The Alar 'scare' of 1989, when the public and EPA reacted to evidence that contamination of apples might endanger child health, provides a visible recent example of this dynamic. A 60 Minutes show aired on February 26, 1989 [ 43 ] dedicated to the findings of a Natural Resources Defense Council study titled 'Intolerable Risk: Pesticides in our Children's Food' [ 18 ]. The public outcry about Alar (a.k.a. daminozide) led to a drop in apple sales and pushed EPA and Alar's manufacturer, Uniroyal Chemical Company, Inc., to take action [ 44 ]. After announcing the safety of Alar in March 1989, and an intention to take no action before July 1990 [ 45 ], EPA announced a preliminary determination to eventually cancel all registrations of daminozide used on foods in May 1989 [ 46 ]. A little over a week later, Uniroyal announced a voluntary recall of all remaining stocks of Alar, and EPA approved a voluntary cancellation of all Uniroyal's daminozide registrations that November [ 47 ]. In an attempt to avoid the still bitter battle between Alar critics and advocates, the relevant point is not whether Alar is or is not a health hazard, but that politics played a major if not a dominant role in its regulation. The very fact that a bitter argument about the actual risk posed by Alar persists indicates that science does not always play a definitive role in regulatory decision-making [ 44 , 48 , 49 ]. But if politics significantly affects decision-making, what is the role of science? To understand the interplay between science and regulation, we must critically consider the conventional assumptions of the modern U.S. chemical regulatory paradigm. Conclusions The conventional paradigm underlying EDSP and most other U.S. chemical regulation amounts to 'science leads to regulation;' it assumes a scientific determination of harm must and, in fact, does precede regulatory action. In this context, Congress mandated EPA protect public health from EDCs, but only after "develop[ing] a screening program, using appropriate validated test systems and other scientifically relevant information, to determine [harm]" [ 2 ]. Real progress on protecting public health waits on the development of a scientific testing regime, on faith that scientific testing is both necessary and sufficient to protect public health. Practical considerations, hazard complexity, exposure complexity, and regulatory deficiencies all challenge the naivety of this assumption. Rational analysis of these factors leads to the inescapable conclusion that science has limitations within the existing regulatory regime and that other important factors are disregarded by the current paradigm. This criticism does not discount the contributions of science, although it seriously questions the assumption that simply doing more science will protect public health. Additionally, this argument does not promote unconsidered action, although it does stress the need to make decisions in the face of ambiguous information. The paradigm itself, though invisible to most adherents, is quite real. The words and actions of industry and environmental groups, government agency personnel, members of Congress, and other concerned interests, regardless of their side of the debate, indicate near universal buy-in to the 'science leads to regulation' paradigm (see [ 10 ] for discussion). But the complexity, uncertainty, and ambiguity of endocrine disruption and its broader context undermine this paradigm's simple logic. The internal validity of 'science leads to regulation' presumes the capacity of science to prove harm with sufficient certainty to regulate. Exposure complexities, including transient and low-concentration exposure to EDCs, maternal metabolism of bioaccumulated EDCs, varying vulnerability and response by developmental stage, poorly understood exposure sources, vulnerable subpopulations, and cultural, social, and economic patterns and spatial and temporal bounds of chemical release and exposure, place the ability of science to make solid exposure determinations in significant doubt (see Priority setting section). Hazard complexities, including low-dose effects, mixtures and synergies, and uncertain endpoints, contribute further obstacles to an unambiguous scientific determination of harm (see Screening and testing section). Under current policy mechanisms, these complications of endocrine-disruption science will prevent any meaningful regulatory action. Essentially, endocrine disruption is too complex and our science too uncertain; most scientific information regarding EDCs will remain ambiguous, with the available information supporting quite different judgments of risk. Recall this observation by the National Research Council: "In [some] cases, the differences [in scientific judgments of EDC significance] do not reflect the need for research but reflect differing judgments about the significance of information" [ 35 ]. The National Research Council has also recognized the more general prejudice that hinders regulation: "The assumption of the null hypothesis as used in risk analysis [as in the case of regulating chemicals] contains an implicit bias because it places a greater burden of proof on those who would restrict than those who would pursue a hazardous activity, presuming these activities are safe until proven otherwise" [ 50 ]. In other words, the paradigm is biased against regulation, and the complexity and uncertainty of endocrine disruption will further undermine attempts to regulate. The external validity of 'science leads to regulation' must take into account the broader context of potential endocrine disruptor regulation. The regulatory deficiencies addressed in the Risk analysis and regulation section, the practical considerations addressed in the Policy Design section, and the significant role of politics in regulation discussed in the Politics section challenge the arbitrary constraints the paradigm places on non-scientific factors. The roles of actors besides the scientific community and agency scientists can make or break regulation. Furthermore, the legal ambiguity of regulating different chemicals under four statutes with different regulatory standards and fragmentary, incomplete statutory authority guarantees further difficulty with regulation, even if the science could meet the near-impossible burden of proof. The most basic practical considerations, including other U.S. regulatory precedents, policy mechanisms to avoid regulation, and most importantly, the sheer number of potential EDCs, each bring the conventional paradigm into doubt. 'Science leads to regulation' simply leaves too much of the decision-making context out of the picture. Yet discussions of improving chemical regulation primarily deal with the minutiae of scientific testing regimes. For example, an Environmental Defense Fund analysis exposing the utter futility of chemical regulation under TSCA (based on the U.S. Government's own damning analysis) came to the conclusion that the failure of testing in the past means we need to test better in the future [ 7 , 41 ]. Though better testing might improve things, such a suggestion still ignores the hard reality of decision-making. Scientific information often remains ambiguous and consistent with quite different action alternatives. Furthermore, scientific information forever remains only one consideration of the decision maker – economic impact, resource tradeoffs, political considerations, constituent needs, and agency funding are other obvious and equally relevant considerations. The relevance of science for regulatory decision-making lies predominantly outside the current trend of increasingly detailed mechanistic investigation of endocrine disruption. While endocrine disruption science can contribute much to a decision-maker, it cannot provide unambiguous information that 'objectively' determines the correct decision. But science could help guide decision-making under conditions of ambiguity; uncertainty does not entail 'anything goes.' On the contrary, the uncertainty itself might guide decision-making better than any other factual information. The 2001 Intergovernmental Panel on Climate Change reports provide a precedent in the explicit treatment of uncertainty – the reports indicate the relevance of scientific information for decision makers by ranking scientific conclusions by Bayesian confidence estimates (e.g. virtually certain means greater than 99% confidence, very likely 90–99%, etc) [ 51 ]. This empowers decision makers and the public by contextualizing expert knowledge and frees scientists to provide relevant information that has not yet met the rigorous standards of scientific proof and peer review. Some excellent work extrapolating from this precedent and considering its ramifications already exists [ 52 , 53 ]. A franker treatment of uncertainty can improve the relevance of science to decision-making and promote realistic expectations of science and the scientific community. Such progress might turn the spotlight on more significant impediments to regulatory decision-making under conditions of ambiguity, such as the difficulty of testing 87,000 chemicals and the role of politics and values in making regulatory decisions. Scientists can take action within their own communities to support the policy goal of protecting public health by exploiting and improving the role scientific information plays in chemical regulation. The possibility that scientists can empower decision makers and the public by developing community standards and norms explicitly addressing uncertainty is one creative idea, but more are needed. The key to progress is taking an active and creative role in support of the protection of public health. Waiting for a solution in the form of national legislation has failed to inspire significant change over the last several decades. But explicit action within the scientific community that discourages unrealistic expectations of science will support long-term progress on chemical regulatory policy and the protection of public health. List of abbreviations DDT – dichlorodiphenyltrichloroethane EDC(s) – endocrine disrupting chemical(s) EDSP – (U.S.) Endocrine Disruptor Screening Program EDSTAC – Endocrine Disruptor Screening and Testing Advisory Committee EPA – (U.S.) Environmental Protection Agency FDA – (U.S.) Food and Drug Administration FFDCA – (U.S.) Federal Food, Drug, and Cosmetic Act FIFRA – (U.S.) Federal Insecticide, Fungicide, and Rodenticide Act FSIS – (U.S.) Food Safety and Inspection Service FQPA – (U.S.) Food Quality Protection Act HPV – high production volume HTPS – high-throughput pre-screening PCBs – polychlorinated biphenyls QSAR – quantitative structure activity relationships SDWA – (U.S.) Safe Drinking Water Act SDWAA – (U.S.) Safe Drinking Water Amendments Act TSCA – (U.S.) Toxic Substances Control Act Competing interests The author(s) declare that they have no competing interests.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549529.xml
548151	Results of paclitaxel (day 1 and 8) and carboplatin given on every three weeks in advanced (stage III-IV) non-small cell lung cancer	Background Both paclitaxel (P) and carboplatin (C) have significant activity in non-small cell lung cancer (NSCLC). The weekly administration of P is active, dose intense, and has a favorable toxicity profile. We retrospectively reviewed the data of 51 consecutive patients receiving C and day 1 and 8 P chemotherapy (CT) regimen in advanced stage NSCLC to evaluate the efficacy and toxicity. Methods Patients treated in our institutions having pathologically proven NSCLC, no CNS metastases, adequate organ function and performance status (PS) ECOG 0–2 were given P 112.5 mg/m 2 intravenously (IV) over 1 hour on day 1 and 8, followed by C AUC 5 IV over 1 hour, repeated in every three weeks. PC was given for maximum of 6 cycles. Results Median age was 58 (age range 39–77) and 41 patients (80%) were male. PS was 0/1/2 in 29/17/5 patients and stage was IIIA/IIIB/IV in 3/14/34 patients respectively. The median number of cycles administered was 3 (1–6). Seven patients (14%) did not complete the first 3 cycles either due to death, progression, grade 3 hypersensitivity reactions to P or lost to follow up. Best evaluable response was partial response (PR) in 45% and stable disease (SD) in 18%. Twelve patients (24%) received local RT. Thirteen patients (25%) received 2nd line CT at progression. At a median follow-up of 7 months (range, 1–20), 25 (49%) patients died and 35 patients (69%) progressed. Median overall survival (OS) was 11 ± 2 months (95% CI; 6 to 16), 1-year OS ratio was 44%. Median time to progression (TTP) was 6 ± 1 months (95% CI; 4 to 8), 1-year progression free survival (PFS) ratio was 20%. We observed following grade 3 toxicities: asthenia (10%), neuropathy (4%), anorexia (4%), anemia (4%), hypersensitivity to P (2%), nausea/vomiting (2%), diarrhea (2%) and neutropenia (2%). Two patients (4%) died of febrile neutropenia. Doses of CT were reduced or delayed in 12 patients (24%). Conclusions P on day 1 and 8 and C every three weeks is practical and fairly well tolerated outpatient regimen. This regimen seems to be comparably active to regimens given once in every three weeks.	Background Lung cancer is the leading cause of cancer related deaths all around the world. About 80% of all lung cancers are non-small cell lung cancer (NSCLC) and more than 50% of these patients present with locally advanced or metastatic disease. Meta-analysis of several randomized trials have demonstrated a modest survival advantage for treatment with cisplatin-based regimens in patients with advanced stages of NSCLC [ 1 , 2 ]. Furthermore, chemotherapy (CT) also has been shown to ameliorate symptoms and increase quality of life [ 3 ]. Addition of second generation CT regimens with cisplatin or carboplatin plus newer agents, such as taxanes (paclitaxel and docetaxel), gemcitabine, vinorelbine have shown increased response rates and 1-year survival ratios, but overall survivals have not been altered [ 4 - 6 ]. Being the first of the taxane antimicrotubule agents, paclitaxel (P) demonstrated overall response rates of 21–24% and 1-year survival rates of 37–42% in the phase II trials where it was used as a single agent [ 7 , 8 ]. Antiangiogenic effect of P has also been reported (9). Carboplatin (C) has also demonstrated comparable activity but better toxicity profile than cisplatin in the treatment of advanced NSCLC [ 10 , 11 ]. P and C used in combined chemotherapy regimens have significant activity in NSCLC. PC given every three weeks is considered to be one of the standard regimens being used worldwide [ 10 ]. The weekly administration of P is active, dose intense, and has a favorable toxicity profile. To evaluate the efficacy and toxicity of C and day 1 and 8 P in advanced stage NSCLC, we retrospectively reviewed 51 consecutive patients receiving this CT regimen. Methods All patients with stage III or IV NSCLC treated at Medical Oncology Units of Marmara University Hospital, Dr. Lutfi Kirdar Research and Training Hospital, SSK Sureyyapasa Chest and Cardiovascular Diseases Hospital and Gulhane Military Medical Academy Hospital within July 2002 and August 2003 were considered for this protocol. Eligible patients were required to have pathologically proven NSCLC, stage III or IV disease at presentation or progressed after surgery, performance status (PS) ECOG 0–2, objective measurable disease, adequate bone marrow functions (white blood cell count ≥ 3500/mm 3 , hemoglobin ≥ 9 g/dl, and platelet count ≥ 100000/mm 3 ), and adequate liver functions (bilirubin ≤ 1.5 mg/dl and alanine aminotransferase ≤ 2 times upper limit of normal and ≤ 5 times upper limit of normal for the patients with liver metastases) and kidney functions (creatinine ≤ 1.5 mg/dl). No prior chemotherapy or radiotherapy (except to bone metastases for palliation) was allowed. Patients presenting with known central nervous system (CNS) disease and uncontrolled cardiac arrhythmia were excluded from this study and they were treated with other chemotherapy regimens (single agent or combination of platinum and vinorelbine, docetaxel or gemcitabine). Patients were treated with P (112.5 mg/m 2 /day) on days 1 and 8, followed by C (AUC 5/6) on day 1, repeated in every three weeks. Both drugs were diluted in 250 ml of normal saline and given intravenously (IV) over 1 hour. No growth factors were administered. Anti-allergic premedication included IV diphenhydramine 50 mg, IV ranitidine 50 mg, and IV dexamethasone 16 mg 1 hour prior to P administration. Toxicity evaluation and routine physical examination were performed in every 3 weeks. Complete blood count (CBC) was done on days 1 and 8 of each cycle, liver and kidney function tests on every 2 cycles. Cranial computed tomography scans (CT), magnetic resonance imaging (MRI) and bone scans were performed as clinically indicated. Side effects of the treatment were graded according to the National Cancer Institute Common Toxicity Criteria (CTC), version 2.0 [ 12 ]. Colony stimulating factors were not used. Response was evaluated with CT of chest and/or abdomen on every 3 rd cycle and standard World Health Organization (WHO) criteria were used to determine response [ 13 ]. Independent of the stage at presentation, patients having partial response (PR), stable disease (SD) or progressive disease (PD) during CT were consulted for radiotherapy (RT) for either primary treatment or palliation. The treatment was stopped for patients with PD. Patients with CR, PR or SD after 3 cycles continued their treatment. PC was given for maximum of 6 courses to the patients having PR or SD. Patients were irradiated with CT based treatment planning and multiple fields arrangements with custom blocking to all fields and involved hilar and mediastinal lymph nodes up to 40–41.4 Gy. Boost was given to the primary tumor. Total dose of 60–61.2 Gy was administered in 1.8–2 Gy daily fractions for 5 days a week and completed in 6 weeks. Statistical analysis Overall survival (OS) and time to progression (TTP) were assessed from the date of diagnosis to the date of death (any cause) and the date of objective disease progression (death was considered a progression event in patients who died before disease progression), respectively. Survival rates were calculated by using the Kaplan-Meier method [ 14 ]. The pre-specified prognostic value of age (< 60 years vs. ≥ 60 years), gender, PS (0 vs. 1–2), histology (adenocarcinoma vs. squamous cell vs. NSCLC), stage (III vs. IV), smoking history (present vs. absent), and response after third cycle of CT (PR vs. other) were evaluated in univariate and multivariate analyses. Log rank test was used for univariate survival analysis [ 15 ]. The multivariate Cox proportional hazard model was applied to identify the variables that can independently influence survival. Results The data of 51 patients receiving PC treatment were collected retrospectively between July 2002 and November 2003. Median follow-up time was 7 months (range, 1–20). Median age was 58 years (range 39–77) and 45% of patients were 60 year-old or above. Eighty percent were male. PS was 0 in 57% of patients and 67% had presented with stage IV disease. Most frequent metastatic sites were the other lung (17), adrenal (10), liver (7) and bone (7). Eighty-two percent of the patients had smoking history, median of which was 40 pack-years (range, 0–135). Patients' baseline characteristics are presented in Table 1 . The median number of cycles administered was 3 (range, 1–6). Seven patients (14%) did not complete the first 3 cycles either due to death (2), progression (3), grade 3 hypersensitivity reaction to P (1) or lost to follow up (1). Best evaluable response was PR in 45% and SD in 18%. Only 22 (43%) patients continued the treatment after the 3rd course. At the end of treatment of these 22 patients 10 (46%) had PR and 6 (27%) had SD, but the other 6 patients (27%) had PD. No complete remission was seen. Twelve patients (24%) received local RT and 4 of these patients were given low dose gemcitabine (75 mg/m2/week × 5–6 weeks) as radiosensitizing agent. Of these 12 patients 3 presented with stage IIIA and all had PR to PC therapy. But of the 5 patients with IIIB disease who were irradiated only one patient had PR, 3 had SD and another one had PD after the 3 cycles of CT. Four patients with stage IV were offered RT for palliation. Thirteen patients (25%) received 2nd line CT at progression and of those only one patient had PR and another SD to this treatment. For 2nd line CT gemcitabine ± cisplatin or C was used in 10 patients (78%) and the rest received other agents like vinorelbine, C or docetaxel. Details of this data can be seen in Table 2 . At a median follow-up of 7 months 25 (49%) patients died and 35 patients (69%) progressed. Median OS time was 11 ± 2 months (95% CI; 6 to16), 1-year OS ratio was 44% (Figure 1 ). Median TTP was 6 ± 1 months (95% CI; 4 to 8), 1-year progression free survival (PFS) ratio was 20% (Figure 2 ). The most frequent toxicity related symptoms were asthenia (61%), neuropathy (42%) and anorexia (35%). We observed the following grade 3 toxicities: asthenia (10%), neuropathy (4%), anorexia (4%), anemia (4%), hypersensitivity to P (2%), nausea/vomiting (2%), diarrhea (2%) and neutropenia (2%). Two patients (4%) died of febrile neutropenia due to a three day delay in referral to hospital after the onset of fever > 38°, although they were warned about the side effects of the therapy. Doses of CT was reduced or delayed in 12 patients (24%) (Table 3 ). Univariate analysis showed that patients presenting with PS of 0, stage III disease and having PR after the 3rd cycle of PC have statistically higher OS (p = 0.015, p = 0.018 and p = 0.047, respectively)(Table 4 ). PS and stage of the disease at presentation and response to the CT after the 3rd cycle were also statistically significant independent prognostic factors influencing the OS in multivariate Cox regression analysis (p = 0.034, p = 0.049 and 0.021, respectively). Discussion Paclitaxel and carboplatin have been shown to be an effective and well tolerated CT regimen in advanced stage NSCLC [ 10 ]. PC given once in every three weeks is one of the most widely used standard schedules worldwide based on the spectrum of activity and the ease of administration. This regimen results in an objective response rate of 17–25% with a median survival time of 8 months in stage IIIB and IV NSCLC patients. The major toxicities of this regimen are neuropathy and neutropenia [ 10 , 16 ]. Weekly P is a relatively new strategy for lowering toxicity and increasing dose-intensity and possibly efficacy. Alvarez et al. have used weekly P on patients who progressed or remained stable on P administered in every three weeks and reported that it can induce response in 62.5% of patients with low toxicity [ 17 ]. Akerley has also studied weekly P administration on phase I and phase II settings [ 18 - 20 ]. They started with a P dose of 175 mg/m 2 /week × 6 every 8 weeks in the phase II trial, but they had to reduce the dose up to 50% due to primarily neutropenia and neuropathy with extended therapy. Therefore, they recommended 150 mg/m 2 as the weekly dose of P [ 20 ]. Weekly dose of P in combination with cisplatin or C had been administered in NSCLC patients by Belani et al. [ 21 , 22 ]. They used this combination in a multicenter three arm trial in 401 patients with stage IIIB and IV disease [ 21 ]. In that trial P was given 100 mg/m 2 /week for 3 weeks out of 4 week cycles in arms I and II, with C either AUC of 6 on day 1 or AUC of 2 on days 1, 8 and 15 of each of four 4-week cycles. Arm III of this trial consisted of P (150 mg/m 2 ) and C (AUC = 2) given weekly for 6 out of 8 weeks for a total of two cycles. Greater percentage of the patients on arm I received intended CT (30% of P and 55% of C) compared with the other arms (28–29% of P and 21–22% of C). Patients on arm I received more than half of the planned C dose. The main reasons for discontinuation of therapy were progression of disease (31%) and adverse events (15%). Median time to progression and median survival time were significantly higher for arm I than arm II for patients with stage IIIB disease. Performance status of the patients was also statistically related to the survival times. Patients with PS-0/1 had longer median PFS with treatment arm I than arm II and patients with PS-2 had higher median OS with arm I than arm II. Although arm I was the most easily tolerable schedule between the three arms, grade 3 or 4 neutropenia was observed in 22% of the patients included. In this trial treatment arm I had a response rate of 32%, median TTP of 6.9 months, median OS time of 11.3 months and 1-year survival rate of 47%. In our study, response rate was 45%, median TTP was 6 months, median OS time was 11 months and 1-year survival rate was 44%. The majority of our patients comprised of stage IIIB and IV disease, similar to the patient group in Belani's study resulting in similar response rates and survival data [ 21 ]. These results also seem more effective than the regimen given once in every three weeks of the same drugs [ 10 , 16 ]. We used the standard dose of P (225 mg/m 2 ) given in every three weeks and divided into two consecutive weeks. C dose was calculated according to Calvert formulation with an AUC of 5. This is a lower dose than the dose of C being used in other phase III trials in the literature. In our study only 4 patients (8%) had dose reduction of 10% and 16% of patients had treatment delays of 1 week because of side effects. According to this data, 76% of patients have received the total planned doses of the drugs on scheduled date. Two patients (4%) died of febrile neutropenia due to a three day delay in referral to hospital after the onset of fever > 38°, although they were warned about the side effects of the therapy. It is worth mentioning that none of our patients received any colony stimulating factors. We have shown that the response to treatment after the third cycles of CT was one of the independent prognostic factors influencing OS. It has already been reported by Socinski et al. that 4 cycles of CT give the maximum benefit which could be obtained from CT in patients with stage IIIB and IV NSCLC [ 23 ]. Smith et al. also studied 3 cycles versus 6 cycles of CT in the same group of patients and failed to show any survival advantage for longer treatment durations [ 24 ]. In addition, there was an increase in the side effects such as fatigue, nausea and vomiting in the patients receiving six courses. It has been shown that PC combination has relatively mild toxicity profile. Belani et al. observed in their phase I trial that patients who received the PC combination in every three weeks experienced less severe thrombocytopenia than would be expected from C alone. In the view of this finding they suggested that there was a platelet-sparing effect of P on the dose-limiting thrombocytopenia side effect of C [ 25 ]. This phenomenon was also shown by Akerley [ 18 ] and Kearns [ 26 ]. Akerley reported that platelet counts rose by 17000/mL/week with weekly P administration [ 18 ]. Belani also speculated on the mechanism for this platelet protective effect and said that it may involve some alteration of megakaryocytopoiesis or thrombocytopoiesis, which could result in increased levels of endogenous thrombopoietin or other cytokines [ 27 ]. Kearns et al. suggested that prior exposure to P may suppress the inhibition of platelet formation, which is associated with C [ 26 ]. None of our patients experienced thrombocytopenia during our CT treatment with day 1 and 8 P with day 1 C on every three weeks. One of the most frequent side effects during our treatment was neuropathy, but it was usually mild (Grade 1 or 2), only 4% of our patients experiencing grade 3 sensory neuropathy. Grade 3 or 4 neuropathy has been reported to be 10–20% in schedules given every three weeks [ 10 , 16 ]. Belani reported 3–13% of grade 3 or 4 neuropathy, but the incidence was lower for arms 1 (P given weekly and C every four weeks) and 2 (P and C both given weekly), at only 5% and 3%, respectively [ 21 ]. This result for arm 1 is similar to the neuropathy rate in our study. Besides the reduced toxicity, weekly administration of P also increases the drugs' anti-angiogenic and apoptotic effects. The metronomic schedule of P has been studied widely during the last few years. P had been shown to inhibit endothelial cell proliferation, motility, invasiveness, and cord formation both in vitro and in vivo Matrigel assays in a dose dependent manner [ 9 ]. Belani et al. randomized the patients having objective response to weekly P and C regimens into two arms (maintenance and observation arms). Patients were either treated with weekly P (70 mg/m 2 /week × 3 weeks out of four weeks cycles) in maintenance arm or observed until disease progression has occurred. They reported that the maintenance arm was compared to the observation arm and had a median PFS of 38 weeks vs. 29 weeks, median OS of 75 weeks vs. 60 weeks, respectively [ 21 ]. Although there was not a statistically significant difference between the two arms, authors concluded that this might be a result of low number of patients enrolled in the study (only 65 patients in each arm). It is not yet known whether these responses have an anti-angiogenic basis, or whether such responses will translate into a significant prolongation of survival. Although our study is a retrospective analysis, it is one of the few manuscripts on this PC scheduling in NSCLC in the literature. Conclusions Paclitaxel on day 1 and 8 and carboplatin every three weeks is a practical and fairly well tolerated outpatient regimen. This regimen seems to be comparably active to regimens given every three weeks. This schedule needs to be further evaluated by well planned randomized phase III trials where it could be compared to the standard regimens in patients with advanced stage NSCLC. Competing interests The author(s) declare that they have no competing interests. Authors' contributions PFY designed the study, followed the patients, collected the data, performed the statistical analysis and drafted the manuscript. NST followed the patients and helped with the manuscript. MG followed the patients and helped with statistical analysis. NFH, OT, AO, TS, MA followed the patients. RA confirmed the diagnosis. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548151.xml
521681	Normal histone modifications on the inactive X chromosome in ICF and Rett syndrome cells: implications for methyl-CpG binding proteins	Background In mammals, there is evidence suggesting that methyl-CpG binding proteins may play a significant role in histone modification through their association with modification complexes that can deacetylate and/or methylate nucleosomes in the proximity of methylated DNA. We examined this idea for the X chromosome by studying histone modifications on the X chromosome in normal cells and in cells from patients with ICF syndrome ( I mmune deficiency, C entromeric region instability, and F acial anomalies syndrome). In normal cells the inactive X has characteristic silencing type histone modification patterns and the CpG islands of genes subject to X inactivation are hypermethylated. In ICF cells, however, genes subject to X inactivation are hypomethylated on the inactive X due to mutations in the DNA methyltransferase (DNMT3B) genes. Therefore, if DNA methylation is upstream of histone modification, the histones on the inactive X in ICF cells should not be modified to a silent form. In addition, we determined whether a specific methyl-CpG binding protein, MeCP2, is necessary for the inactive X histone modification pattern by studying Rett syndrome cells which are deficient in MeCP2 function. Results We show here that the inactive X in ICF cells, which appears to be hypomethylated at all CpG islands, exhibits normal histone modification patterns. In addition, in Rett cells with no functional MeCP2 methyl-CpG binding protein, the inactive X also exhibits normal histone modification patterns. Conclusions These data suggest that DNA methylation and the associated methyl-DNA binding proteins may not play a critical role in determining histone modification patterns on the mammalian inactive X chromosome at the sites analyzed.	Background Although it has been known for some time that histone modifications play a role in gene expression [ 1 ], it is only in the last several years that the details of these modifications have been more fully described. Acetylation and methylation of histone tails, for example, exhibit characteristic patterns for expressed and repressed genes in all eukaryotes studied [ 2 ]. This generality of histone modification and gene expression holds for eukaryotes with and without DNA methylation, indicating that DNA methylation is not required for histone modification. In organisms with DNA methylation, however, interactions between histone modification and DNA methylation do appear to exist. In Neurospora , histone methylation appears to determine DNA methylation patterns [ 3 , 4 ]. In Arabidopsis , non-CpG DNA methylation also appears to be determined by histone methyltransferases, whereas CpG methylation does not [ 5 , 6 ]. In mammals, there is considerable evidence suggesting that methyl-CpG binding proteins may play a significant role in histone modification through their association with histone deacetylases [ 7 - 11 ]. Mutations in the MeCP2 methyl-DNA binding protein, which are the cause of most Rett syndrome cases [ 12 ], support this model, because human male and female cells with MECP2 mutations exhibit histone hyperacetylation [ 10 ]. Histone hyperacetylation was also observed in mice with Mecp2 mutations [ 13 ]. Thus, DNA methylation is upstream of histone modification in this model of methyl-DNA binding proteins and histone modification. Another possibility is that DNA methyltransferases themselves may target histone deacetylases through a noncatalytic domain, leading to histone modifications that are independent of other methyl-DNA binding proteins [ 14 ]. We are especially interested in the X chromosome with respect to the question of the relationship between DNA methylation and histone modification. The mammalian X chromosome is unusual in that about a thousand gene-associated CpG islands are hypermethylated on the inactive X and hypomethylated on the active X. Except for imprinted loci, methylation patterns at most other regions of the genome are similar between homologs. Histone modification differences known to be associated with either silent or expressed chromatin also distinguish the active and inactive X chromosomes [ 15 - 19 ]. Thus, the mammalian X chromosome inactivation system would appear ideal for testing whether or not a methyl DNA binding protein – histone modification pathway exists for the inactive X chromosome. To examine more fully the possible relationships between DNA methylation and histone modification, we have utilized cell cultures from individuals with a human hypomethylation disease called the ICF syndrome. This disease is clinically characterized by " I mmune deficiency, C entromeric region instability, and F acial anomalies". In most cases, the molecular defects result from mutations in the DNMT3B methyltransferase gene [ 20 - 22 ]. Certain heterochromatic regions are markedly hypomethylated as a result of these mutations, including the CpG islands on the inactive X chromosome that are associated with genes [ 23 ] and LINE-1 elements [ 24 ]. If DNA methylation is upstream of histone modification, the histones on the inactive X should not be modified to a silent form in ICF cells. Our results indicate, however, that these histones do have modifications typical of silenced genes, suggesting that methyl-DNA binding proteins may not be critical with respect to histone modification on the inactive X chromosome. In addition, we examined clonal primary fibroblast cultures from two individuals with Rett syndrome and found that the histone modification pattern of the inactive X is not affected by mutations in MECP2 . This suggests that this specific methyl-DNA binding protein does not have a major role in silencing the inactive X through histone modification. Results Cytological analysis of histone modification DNMT3B mutant cells (ICF syndrome) We examined histone modifications known to be associated with the inactive X chromosome in two ICF cell lines and normal control cells. Specific histone modifications including histone H3 and H4 acetylation, and histone H3 methylation at K4 and K9, were examined using antibodies to stain nuclei and metaphases [ 15 - 17 ]. We also examined histone macroH2A1 staining, which is known to be concentrated on the inactive X at interphase [ 25 ]. One hundred or more interphase nuclei that had an obvious sex chromatin body by DAPI staining were scored for histone modification. For acetylated histone (H3 and H4) and K4-methylated histone H3, the majority (>60%) of nuclei showed a characteristic hole at the sex chromatin body in both normal and ICF cells (Fig. 1A ). For histone H3 methylation at K9, the frequency of positive cells (Fig. 1B ) was lower (about 50%). We often noted a positive signal for methylated K4 histone H3 in an otherwise negative-staining sex chromatin region in both normal and ICF cells (Fig. 1A ). This signal appears to derive from the DXZ4 locus that was previously reported by Chadwick and Willard [ 19 ] as having active-type histone modifications. DXZ4 is a megabase-sized region known to be hypermethylated on the active X and hypomethylated on the inactive X in normal cells [ 26 ]; this locus appears to be modified normally in ICF cells. Surprisingly, we did not observe this signal on metaphase spreads, suggesting that our resolution on these preparations may be much lower than on interphase chromatin. For metaphase spreads (acetylated histone H3 and H4 and histone H3 methylated at K4) we analyzed 50 cells per line, and in the great majority of analyzable metaphases (>80%) a single lightly-labeled chromosome was detected (Fig. 2 ). In some of these cells, the tip of the short arm was labeled, as would be expected for the pseudoautosomal region (Fig. 2 ). The expected patterns of macroH2A1 histone concentration and histone modification on the inactive X were found in cells derived from ICF individuals and in control normal cells (Figs. 1 and 2 and Table 1 ). At the cytological level, therefore, no difference could be found between normal and ICF cells with respect to the histone modifications on the inactive X. MECP2 mutant cells (Rett syndrome) We examined histone modifications and macrohistone association in clones from two individuals heterozygous for a mutation in MECP2 , an X-linked gene that is subject to X inactivation [ 12 ]. MECP2 mutations lead to Rett syndrome, and the protein product codes for a methyl-CpG binding protein known to recruit a transcriptional silencing complex that deacetylates histones. In Rett syndrome and in mice with mutant Mecp2 , histones exhibit hyperacetylation [ 10 , 13 ], as would be expected if this methyl-DNA binding protein functions upstream of histone modification. In one case, we had complementary clones with either the mutant (1786YY) or wild type MECP2 allele (1786QQ) on the inactive X, and in the second case, we had one clone with the wild type MECP2 on the inactive X (1789V) [ 27 ]. For all the histone modifications we examined (H3 and H4 acetylation and H3 K4 and K9 methylation), the cytological patterns on the active and inactive X chromosomes in mutant MeCP2-expressing cells were indistinguishable from those in clones expressing the wild type allele or in other control cultures (Fig. 1 ). These results suggest that MeCP2 does not function in determining these histone modification patterns on the inactive X chromosome. ChIP analysis To verify our cytological histone modification results at the gene level, we searched for promoter polymorphisms at seven X-linked loci ( G6PD , NEMO , MECP2 , SYBL1 , AR , FMR1 , and PGK1 ) in ICF cells so that we could employ allele-specific chromatin immunoprecipitation (ChIP) analysis. We restricted our search for polymorphisms to the promoter region, as several reports have indicated that marked differences in histone modifications between active and inactive alleles are seldom detected at other regions [ 28 - 30 ]. We found useful promoter polymorphisms at two loci, SYBL1 (synaptobrevin-like gene in the pseudoautosomal portion of Xq28) and AR (androgen receptor in Xq12). Previously, one of us (RSH) has reported on a ChIP study at the SYBL1 locus in male ICF cells where the inactive Y allele had reactivated and the histone modification pattern was that of an active gene [ 31 ]. Here we report on ChIP studies at the SYBL1 and AR loci in ICF female cultures using antibodies to histone H3 dimethylated at K4 and to acetylated histone H3. Both loci are subject to X inactivation, and the inactive X alleles remain inactive in ICF cells even though the 5' CpG islands are hypomethylated [ 23 ]. In the case of the SYBL1 inactive X allele, the methylation level is reduced by over 90% with most chromosomes exhibiting no methylation. An XhoI restriction site polymorphism in the untranslated exon 1 of SYBL1 permitted separation of the active and inactive alleles in cloned cells. A CAG repeat number polymorphism in the 5' coding region of the AR gene (1.3 kb downstream of the transcription start site according to reference sequence NM000044) was informative in one ICF sample (PT 4) and in several controls, thus permitting separation of the active and inactive alleles in cloned cells, and in cultures with highly skewed X inactivation. The antibodies were highly specific under the amplification conditions chosen, so that a strong signal was seen for the pull-down experiment with antibody and little or no signal for the "no antibody" control (Fig. 3A ). The fluorescent amplification products from the AR gene were then separated on an automated sequencer according to CAG repeat number. Two major peaks are detected in the input control DNA, corresponding to the active (A) and inactive (I) X alleles, differing in CAG repeat number (Fig. 3B ). An allele was determined to be from the active X by RT-PCR analysis (data not shown). The lesser "shadow band" peaks, labeled S, probably derive from PCR errors. In the methylated K4 H3 and acetylated H3 antibody ChIPs, a single peak predominates in both normal and ICF cells (Figs. 3B and 3C ) that corresponds to the active X allele. Our ChIP analysis of the inactive X at the SYBL1 locus in an ICF female (PT3) also showed normal histone H3 hypoacetylation and K4 H3 hypomethylation in spite of the very low levels of DNA methylation in this CpG island region (Fig. 4 ). These data, therefore, agree with our cytological observations in that only the active X alleles are positive for the histone modifications known to be associated with active genes, though a minor portion of the inactive X allele was found to precipitate with the acetylated H3 antibody in both normal and ICF cells (Fig. 3C and data not shown). Discussion The major observation reported here is that ICF cells, despite being hypomethylated at gene- and L1-associated CpG islands on the inactive X chromosome, exhibit the same histone modification patterns as inactive Xs in normal cells. In addition, we show that cells mutant for MeCP2, a methyl DNA binding protein, also exhibit normal histone modification patterns on the inactive X. These results imply that DNA methylation and/or this methyl DNA binding protein are not critical for determining histone modification patterns on the inactive X chromosomes. Two major questions can be raised about our results: (1) is the sensitivity of the cytological histone modification assay too low to detect active-type histone modifications on the ICF inactive X? and (2) is the extent of methylation on the ICF inactive X greater than is suggested by CpG island and LINE-1 methylation patterns? The cytological results imply that most of the genes on the inactive X in ICF cells are subject to inactivation, a conclusion supported by our allele-specific expression analyses of individual genes, such as AR , in ICF cells ([ 23 ] and data not shown). For genes subject to X inactivation in ICF cells, we expect histone modifications at the gene level to be similar to those detected cytologically at the chromosome level, and this is what we have shown here for the AR gene. For genes that escape X inactivation in ICF cells, we expect their histone modification patterns to be those of expressed genes, and one of us (RSH) has previously reported this to be the case for the SYBL1 gene ([ 31 ] and data not shown). We did not detect these active patterns cytologically, suggesting that there are no large blocks of genes escaping inactivation in ICF cells except at the Xp pseudoautosomal region, which normally contains escaping genes (Fig. 2 ). Methylation levels at inactive X-linked CpG islands in ICF cells are decreased by an average of 89% from normal as determined by bisulfite analyses at the G6PD , FMR1 , and SYBL1 loci, and many of the cloned alleles analyzed were completely unmethylated like active X alleles [ 23 ]. It is possible that DNA methylation at other CpG-rich regions (e.g., Alu and LINE-1 elements) on the X chromosome might be differentially methylated and play a role in the X chromosome histone modification pathway. One of us (RSH) has recently shown that LINE-1 elements are hypermethylated on both active and inactive X chromosomes in normal cells but, surprisingly, they are hypomethylated on the inactive X and hypermethylated on the active X in ICF cells [ 24 ]. These results argue against a role for LINE-1 methylation in histone modification on the inactive X chromosome. A more complete DNA methylation analysis of the ICF and normal inactive Xs needs to be done, however, because other widespread sequences may be hypermethylated on the ICF inactive X that could direct histone modifications. Because we know that promoter methylation is important in gene expression, it seems reasonable that if DNA methylation were directly involved in the histone modification pathway, CpG island methylation would play a critical role. Further support for this idea comes from the fact that histone modifications distinguishing active and inactive X-linked genes are concentrated in promoter regions [ 28 , 32 ]. In fact, Rougeulle et al. [ 32 ] propose that the promoter-restricted histone modification seen at X-linked loci may be unique to monoallelically-expressed genes and provide them with an epigenetic mark. That DNA methylation is not critical to the developmental appearance of histone modifications is further supported by recent murine studies showing that differential histone modification of the Xs in early development precede differential developmental appearance of DNA methylation [ 33 , 34 ]. The fact that DNA methylation does not appear to be critical to the development of histone modifications in X-linked gene expression should not be confused with a more important role for DNA methylation in maintaining repression of X-linked genes. Some years ago we showed that the earliest events in reactivating inactive X-linked genes were hemidemethylation followed by chromatin hypersensitivity, and then transcription factor binding and transcription [ 35 , 36 ]. More extensive studies have recently pointed to a similar conclusion [ 11 , 37 , 38 ]. Thus, DNA methylation appears to play a dominant role in maintaining repression, even though it is a late event in establishing silent chromatin. We can also consider the implication of this work for the proposed role of methyl-CpG binding proteins in the histone modification pathway. Our ICF cell data and the results from murine developmental studies, showing that histone modification of X-linked genes precedes DNA methylation, argue against such a role for the X chromosome. A role for methyl-DNA binding proteins in the histone modification pathway is supported by studies with Rett syndrome cells where a methyl-DNA binding protein, MeCP2, is mutated. In both humans and mice with Rett syndrome mutations, general hyperacetylation of histones was reported, albeit at different sites. In human cell lines, H4 was hyperacetylated preferentially at K16 [ 10 ], while in mouse mutant tissues hyperacetylation was reported specifically at H3K9 [ 10 , 13 ]. In our work, however, we saw no major effect of two different MECP2 mutations on inactive X histone modification. The recent discovery that LINE-1 elements on the inactive X are methylated by a methyltransferase distinct from the one that carries out the same modification on the active X raises the possibility that the inactive X could have its own modification rules [ 24 ]. We must consider, therefore, the possibility that the inactive X chromosome does not utilize methyl-DNA binding proteins in the histone modification pathway. Such a possibility would fit with the failure to detect protein footprints at promoters on the inactive X chromosome, whereas they are readily detectable on the active X [ 36 , 39 - 42 ]. It should be noted that only a small fraction of possible histone modifications have been elucidated at this time, and it is possible that histone modification on the inactive X that depends on methyl-DNA binding protein(s) will be found in the future. Finally, we would like to comment on the implication of this study regarding the inactive X silencing complex. Systems controlling gene expression tend to be multilayered and the X inactivation system is no exception. We know that silencing on the inactive X involves XIST RNA, DNA methylation, histone modification patterns, chromatin sensitivity, and delayed replication. It is our opinion that these factors tend to act in a more or less independent manner, as we have suggested several times in the past [ 23 , 43 - 46 ]. For example, promoter demethylation of inactive X-linked genes, as occurs in ICF cases, does not necessarily lead to reactivation; markedly advanced replication time must also be present for reactivation to take place [ 23 ]. The present study would appear to add further support to this idea. Conclusions The inactive X chromosome in mammalian cells is characterized by a particular set of histone modifications. It has been suggested that methyl-DNA binding proteins may be involved in these modifications through their interactions with histone deacetylases. We have investigated this idea by studying histone modification patterns on the inactive X in ICF and Rett syndrome cells. ICF cells are hypomethylated on the inactive X, in contrast to normal cells, and the Rett cells we studied were deficient in MeCP2, a specific X-linked methyl-DNA binding protein. We found that the histone modification patterns on the inactive X in these mutant cells were indistinguishable from those in normal cells. We conclude that DNA methylation and the associated methyl-DNA binding proteins do not appear to play a critical role in determining histone modification patterns on the mammalian inactive X chromosome, either globally or at the level of the promoter. Methods Cells and cell culture Mutant fibroblast cell cultures included two from female ICF individuals, whose DNMT3B mutations have been previously described [ 20 ], and complementary clonal cultures from an individual (Rett 1) heterozygous for a mutation (1155del132) in the MECP2 gene [ 27 ]. In one clone (1786YY), the mutant gene is on the active X, leading to a culture with nonfunctional MeCP2 protein; the complementary control clone (1786QQ) has functional MeCP2 because the mutant gene is on the inactive X. In another clone (1789V), derived from a Rett individual with the mutation R106W (Rett 2), the active X carried the mutant allele. Normal fibroblast cultures were obtained from commercial sources. For chromatin precipitation studies, the ICF fibroblast clones were immortalized with hTERT, as previously described [ 47 ]. Cells were grown in AmnioMax-C100 (Gibco Invitrogen Corp.; Carlsbad, CA) and harvested in trypsin:EDTA (Gibco Invitrogen Corp.) under standard conditions [ 46 , 48 ]. Cytology For analysis of interphase stages, cells were plated on alcohol-washed 22 mm square cover slips in 35 × 10 mm Petri dishes. On the following day the medium was removed and the cells were washed once with PBS followed by fixation in 95% ethanol:5% acetic acid for 1 min at room temperature. The rest of the procedure followed the "Immunocytochemistry Protocol" of Upstate (Lake Placid, NY). DAPI-stained slides mounted in antifade were examined with a Nikon Microphot FXA microscope and images were captured with a Nikon Coolpix 995 digital camera. The inactive X was recognized under DAPI staining as sex chromatin. Absence of a particular histone modification on the inactive X was seen as a hole or gap at the sex chromatin location. For analysis of metaphase cells, we plated cells in 150 × 25 mm tissue culture dishes and added colcemid (Gibco Invitrogen Corp.) 48 h later (0.1 μg/ml for 2 h). The medium was removed and the cells were washed once with Hanks' balanced salt solution, followed by trypsinization (Gibco Invitrogen Corp.) with slight agitation to collect metaphase cells. Serum was added to stop tryptic action and the cells were recovered by centrifugation, then placed in hypotonic solution (3 mg/ml KCl and 1 mg/ml sodium citrate) at 37°C for 15 min. Cells were collected on to slides using a Cytospin centrifuge and then fixed in 95% ethanol:5% acetic acid for 1 min. The rest of the procedure followed the Upstate protocol mentioned above, followed by DAPI staining, mounting in antifade, and examination with a Nikon Microphot FXA microscope. Antibodies used to detect histone modifications were obtained from Upstate and included: "Anti-acetyl-Histone H4," recognizing acetylated lysines 5, 8, 12, and 16, "Anti-acetyl-Histone H3 (Lys 9)," "Anti-dimethyl-Histone H3 (Lys 9)," "Anti-dimethyl-Histone H3 (Lys 4)," and "Anti-Histone macroH2A1." ChIP studies Chromatin immunoprecipitation was performed using the protocol of Upstate with slight modifications. For each experiment, a near-confluent 75 cm 2 tissue culture flask (about 3 × 10 6 cells) was washed with PBS and treated with 4% formaldehyde (pH > 7) for 10 min at 37°C. Protease inhibitor cocktail (Complete) from Roche Diagnostics (Indianapolis, IN) was used in place of individual inhibitors, and Protein A-Separose 4B (Zymed Laboratories Inc.; South San Francisco, CA) was used to collect immune complexes. After elution of immune complexes, they were heated at 65°C for 4 h to reverse crosslinks, and the DNA was recovered with a "QIAquick" PCR purification kit from Qiagen Inc. (Valencia, CA). Primary amplification of AR DNA was across the polymorphic 5' CAG repeat region as previously described [ 23 ], using 27–35 cycles of PCR amplification with 10% of the immunoprecipitated material or 50 ng of input DNA in a 50 μl reaction volume. Allele-specific analysis was performed by amplifying the primary product with a 5'-FAM-labeled nested primer as previously described [ 23 ], using 6–15 cycles of PCR amplification with 2–10 μl primary product in a 50 μl reaction volume. All amplification conditions were chosen so as to produce visible products by ethidium staining only for antibody-precipitated material, and not for "no antibody" controls. The fluorescent products were then run on an ABI PRISM 310 capillary sequencer (Applied Biosystems; Foster City, CA) to separate alleles differing in CAG repeat number and analyzed using GeneScan software (Applied Biosystems). Allele-specific expression analysis by RT-PCR was performed on DNaseI-treated RNA using a similar procedure [ 23 ]. Analysis of the 5' region of the SYBL1 gene was performed similarly to that of AR , except the allele-specific reaction entailed XhoI digestion of the primary amplification product followed by these products being separated by agarose gel electrophoresis. Conditions for PCR amplification and XhoI digestion were as previously described [ 31 ]. List of abbreviations ac = acetylated ChIP = chromatin immune precipitation DAPI = 4,6-diamidino-2-phenylidole FB = fibroblast FAM = 5-carboxyfluorescein FITC = fluorescein-isothiocyanate ICF = immune deficiency, centromeric region instability, facial anomalies H3 = histone 3 H4 = histone 4 K4 = lysine 4 K9 = lysine 9 LINE-1 = long interspersed nuclear element 1 MeCP2 = methyl-CpG binding protein WT = wild type Xa = active X chromosome Xi = inactive X chromosome Authors' contributions SMG and RSH conceived the study design, supervised and coordinated its progress, and drafted and prepared the final manuscript. KRV and PL carried out the cell culture and cytological studies. TKC carried out the ChIP analyses. JT and UF developed the cloned Rett cell cultures. All authors read and approved the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521681.xml
548145	Molecular mechanisms of severe acute respiratory syndrome (SARS)	Severe acute respiratory syndrome (SARS) is a new infectious disease caused by a novel coronavirus that leads to deleterious pulmonary pathological features. Due to its high morbidity and mortality and widespread occurrence, SARS has evolved as an important respiratory disease which may be encountered everywhere in the world. The virus was identified as the causative agent of SARS due to the efforts of a WHO-led laboratory network. The potential mutability of the SARS-CoV genome may lead to new SARS outbreaks and several regions of the viral genomes open reading frames have been identified which may contribute to the severe virulence of the virus. With regard to the pathogenesis of SARS, several mechanisms involving both direct effects on target cells and indirect effects via the immune system may exist. Vaccination would offer the most attractive approach to prevent new epidemics of SARS, but the development of vaccines is difficult due to missing data on the role of immune system-virus interactions and the potential mutability of the virus. Even in a situation of no new infections, SARS remains a major health hazard, as new epidemics may arise. Therefore, further experimental and clinical research is required to control the disease.	Introduction Severe acute respiratory syndrome (SARS) is the first new infectious disease of this millennium. SARS has originated from Southern China at the end of 2002 and has a high mortality and morbidity. Within a period of six months beginning at the end of 2002, the disease has affected more than 8,000 people and killed nearly 800 [ 1 ]. The disease poses a new threat for respiratory medicine and represents a challenge for antiviral drug development and administration [ 2 , 3 ]. SARS is caused by a novel, SARS-associated coronavirus (SARS-CoV) [ 4 - 6 ] which has been identified by a World Health Organization (WHO)-led global laboratory network. The first cases of SARS were reported from a hospital in Hanoi, Vietnam, by Carlo Urbani, a WHO scientist who himself died from the disease [ 7 ]. After reports from health authorities in Hong Kong on the outbreak of a new form of epidemical atypical pneumonia in public hospitals, the WHO issued a global alert on the disease. During this period, cases of SARS were also reported from China, other Asian countries and even other continents including America (Canada, U.S.A.) and Europe (Germany). Shortly after the initial global alert, the WHO initiated a collaborative multi-center research project on SARS diagnosis, led by eleven principal laboratories in nine countries [ 8 ]. Using modern communication technologies to optimize the analysis of SARS tissue samples, it was soon shown that a novel coronavirus is the causative agent of SARS (SARS-CoV) [ 4 - 6 ]. Due to the death of Carlo Urbani who first identified the new disease, the first isolate of the virus was proposed to be named Urbani strain of SARS-associated coronavirus, but a final terminology has not been proposed so far [ 9 ]. Since Koch's principles have been shown to be fulfilled by the new pathogen [ 10 , 11 ], it is not necessary to call the virus SARS-associated and the general agreement is now to call it SARS coronavirus (SARS-CoV). Parallel to the progress made in the epidemiology and clinical diagnosis which has recently been demonstrated by numerous case reports, clinical studies and definitions [ 1 ], scientists have also revealed basic mechanisms of the underlying causative agent, the SARS coronavirus. As it is crucial for future strategies that SARS is detected in its earliest stages and that therapeutic options are optimized, insights into the molecular mechanism of SARS have to be used to develop new therapeutic strategies and vaccines. While other reviews have focused on the epidemiology, clinical presentation and potential treatment of SARS, the present overview aims to analyze and present the currently available data on molecular mechanisms of SARS. In this respect, it is important to underline that in the present state of no specific drug or vaccine being available, research on molecular mechanism is crucial to identify potential treatment targets. Etiology Prior to the development of therapeutic regimes based on molecular mechanisms of the disease, the causative agent had to be isolated and analysed. Soon after the fast establishment of the international WHO laboratory network, rapid progress was made in the identification process of the causative agent, and it was reported that SARS is most probably caused by a novel strain of the family of coronaviruses [ 4 - 6 ]. These viruses are commonly known to cause respiratory and gastrointestinal diseases of humans and domestic animals [ 12 , 13 ]. The group of coronaviruses is classified as a member of the order nidovirales, which represents a group of enveloped positive-sense RNA viruses consisting of coronaviridae and arteriviridae [ 14 ]. Viruses of this group are known to synthesize a 3' co-terminal set of subgenomic mRNAs in the infected cells [ 15 ]. Origin of the SARS virus Soon after the identification of a new coronavirus as the causative agent of SARS and of a southern Chinese province as the first area of occurrence, animal species of this area have been speculated to be the origin of the SARS-CoV. As analysis of the SARS-CoV genetic sequence revealed large differences to any other currently known coronaviruses in humans or domestic animals [ 16 , 17 ], it was hypothesized that the new virus might originate from wild animals. This hypothesis was supported by a search for coronaviruses in wild animals sold on markets in southern China, which identified the presence of a coronavirus in civet cats. This animal coronavirus was shown to have a sequence identity of more than 99% to the SARS coronavirus [ 18 ] with only a limited number of deletions and mutations between both viruses. SARS-CoV has a deletion of 29 nucleotides relative to the civet cat virus, indicating that if there was direct transmission, it went from the animal to man, because deletions occur probably more easily than insertions. Recent reports indicate that SARS-CoV is distinct from the civet cat virus and it has not been answered so far if the civet cat virus is the origin of the SARS-CoV or if civet cats were also infected from other species [ 19 ]. Therefore, there are no data available on the possibility of horizontal transmission between animals, and the question whether the jump of the virus from an animal to humans was a single accident or may frequently occur in future with the animals as dangerous reservoirs for future SARS epidemics remains unanswered. So far, the SARS-CoV has been reported to be able to infect not only humans but also macaque monkeys [ 11 ], domestic cats, and ferrets [ 20 ]. However, transmission of the virus from the domestic cat to man has not been shown. The ability of the SARS-CoV to infect other animal species could point to potential natural reservoirs of the virus. In this respect, coronaviruses are known to relatively easily jump to other species. I.e., the human coronavirus OC43 shares a high degree of genetic sequence homology to bovine coronavirus (BCoV) and it is commonly assumed that it has jumped from one species to the other [ 21 , 22 ]. In the same way, BCoV has been reported to be able to infect humans and cause diarrhea [ 23 ]. Whereas the precise mechanisms of these species jumps remain unclear, it is most likely that they represent the results of mutations and epidemiological studies of coronavirus infections in wild animals will therefore be crucial for future understanding and control of new SARS outbreaks. SARS virus taxonomy Until the identification of the new SARS-CoV, the coronaviruses have been divided into three subgroups, which differ with respect to their genome [ 24 ]. The first group consists of viruses such as the human coronavirus 229E (HCoV-229E), porcine respiratory coronavirus (PRCV), porcine transmissible gastroenteritis virus (TGEV), feline infectious peritonitis virus (FIPV) and feline enteritis virus (FEV) or the canine coronavirus (CCoV). The second group comprises human coronavirus OC43 (HCoV-OC43), bovine coronavirus (BCoV), and mouse hepatitis virus (MHV), and the third group mainly consists of avian species such as the chicken infectious bronchitis virus (IBV). Whereas the SARS-CoV has been shown to cross-react with some group I coronavirus antibodies [ 6 ], its genetic sequence does not belong to this group. Within the nucleic acid or protein sequence phylogenetic trees of the coronavirus family, the SARS-CoV has first been located at an equal distance from the second and third group, irrespective of which SARS-CoV RNA region is used for analysis [ 6 , 16 , 17 ]. Therefore, the SARS-CoV may represent the first member of a new group of coronaviruses (Figure 1 ). However, the taxonomy is still no clear [ 19 , 25 ], and recent studies that focused on the N-terminal domain of the spike protein and on poorly conserved proteins such as Nsp1, matrix protein, or nucleocapsid, have suggested a relation to group II viruses [ 26 ]. A similar conclusion can be drawn if the polymerase gene is examined, pointing to an early split-off from the coronavirus group 2 lineage [ 27 ]. Figure 1 Coronavirus classification. The family of coronaviruses belongs to the order of nidovirales and consists of three groups so far. It is still debatable whether the new SARS-CoV should be assigned to group II or to a new fourth group. Group I includes human coronavirus 229E (HCoV-229E), transmissible gastroenteritis virus (TEGV), porcine epidemic diarrhea virus (PEDV), canine coronavirus (CCoV), and feline coronavirus (FIPV). Group II viruses include human coronavirus OC43 (HCoV-OC43), murine hepatitis virus (MHV), and bovine coronavirus (BCoV), and group III species are turkey coronavirus (TCoV), and avian infectious bronchitis virus (IBV). Despite the fact that this new virus most likely jumped to humans from wild animal species, it has remarkably well adapted to the human organism as shown by its high person-to-person transmissibility. SARS virus genome structure The structure of the SARS viral RNA is organized in 13–15 open reading frames (ORF) and contains a total of approximately 30,000 nucleotides [ 6 , 16 , 17 ]. Recently, 61 SARS-CoV sequences derived from the early, middle, and late phases of the SARS epidemic together with two viral sequences from palm civets were analyzed [ 28 ]. Genotypes characteristic of each phase were discovered, and it was found that the neutral mutation rate of the viral genome was constant but the amino acid substitution rate of the coding sequences slowed during the course of the epidemic. The spike protein showed the strongest initial responses to positive selection pressures [ 28 ]. Only ORFs exceeding fifty amino acids in translational capacity are considered relevant as they contain the sequences for the structural and functional properties of the virus and are therefore of potential interest for the development for future therapeutic strategies. The comparison of the different SARS-CoV ORFs with those of other coronaviruses reveals a familiar pattern of structural gene arrangement with replicase and protease genes (gene 1a-1b) and the spike (S), envelope (E), membrane (M) and nucleocapsid (N) genes in a typical 5'- to 3' order of appearance [ 29 ]. The proteins encoded by these genes may be targets for novel treatments. Between these well-known genes, a series of ORFs of unknown function was found: There are two ORFs situated between the spike and the envelope genes and three to five ORFs between the membrane and nucleocapsid genes. Comparison of this gene organization with other known coronaviruses does not indicate a closest proximity to group II coronaviruses. Also, the SARS-CoV genomic sequence does not contain a gene for hemagglutinin-esterase (HE) protein, which is present in the majority of group II coronaviruses. Two-thirds of the SARS RNA is organized in the gene 1a-1b. The sequence of this gene is highly conserved among all coronaviruses [ 17 ]. ORFs 1a and 1b encode two polyproteins, pp1a and pp1ab, the latter through a ribosomal frameshifting mechanism. These polyproteins are processed by virus-encoded proteinases, to yield 16 individual proteins. Most potential gene 1a-1b products are fairly well conserved between SARS-CoV and other coronaviruses [ 17 , 29 ]. Many of their functions are unknown but it is suggested that they participate in viral RNA replication, making them potential targets for the development of antiviral compounds. Therefore, research efforts will focus on these proteins. One exception from the overall conservation of SARS-CoV gene 1a-1b is the lack of a sequence coding for PL1 pro , one of the two papain-like proteinases operating on cleavage sites at the N-terminus of the polyproteins (Figure 2 ). The main proteinase (M pro ), also called 3C-like protease (3CL pro ), is responsible for the cleavage of all the remaining proteins encoded by gene 1a-1b [ 29 , 30 ]. Figure 2 SARS-CoV genome organization. The structure of the SARS viral RNA is organized into 13–15 open reading frames (ORFs) and contains an overall amount of approximately 30,000 nucleotides. The sequence can be separated into different elements and genomic and subgenomic mRNAs. SARS virus gene expression Apart from gene 1, coronavirus genes are known to be usually expressed from subgenomic mRNAs. They share a common leader sequence at the 5'-end and initiate at different places in the genome extending toward the 3'-end of the virus genome [ 31 ]. Some ORFs may also be unconventionally translated from a single mRNA. As these uncommon translation mechanisms are not very efficient and the gene products are not very abundant, these ORFs typically encode nonstructural proteins. Whereas the ORFs between the structural protein genes are very heterogeneous among the different coronaviruses and not essential for viral replication, recent studies suggested that deletion of non-essential ORFs may result in a reduced virulence [ 32 ]. In agreement with this, some of these non-essential ORFs of the new SARS-CoV genome may be responsible for the high SARS-CoV virulence. So far, five to eight subgenomic mRNAs were found in SARS-CoV-infected cells [ 17 , 27 ]. Thiel and colleagues performed the first detailed study on mechanisms and enzymes involved in SARS-CoV genome expression (Figure 2 ) [ 29 ]. They determined the sequence of the SARS-CoV isolate Frankfurt 1 and characterized the major RNA elements and protein functions involved in the genome expression by characterizing regulatory mechanisms such as the discontinuous synthesis of eight subgenomic mRNAs, ribosomal frameshifting and post-translational proteolytic processing. Also, the activities of SARS-CoV enzymes such as the helicase or the two cysteine proteinases (PL2 pro and M pro ) were addressed as they are involved in replication, transcription or post-translational polyprotein processing [ 29 ]. In conclusion, research in the area of coronavirus gene expression is important to delineate components which directly affect SARS-CoV virulence. SARS virus structural proteins The structural proteins of the new SARS-CoV are potential targets for new treatment options. The new SARS-CoV only contains the three envelope proteins, spike (S), envelope (E), and membrane (M) but not the hemagglutinin-esterase (HE) protein, which is present in some coronaviruses of the second group. The spike glycoprotein is responsible for the characteristic spikes of the SARS-CoV (Figure 3 ). Intra- and extracellular proteases often cleave the S protein into S1 and S2 domains, with the cleavage process often increasing infectivity of the virus. Molecular modelling has been performed for the S1 and S2 units of the SARS-CoV spike protein [ 33 , 34 ]. The spike proteins of coronaviruses are reported to bind to receptors on their target cells and the domains responsible for receptor-binding are commonly situated in the N-terminal region of S1 [ 35 - 40 ]. The spikes consist of oligomeric structures, that are formed by heptad repeats of the S2 domain which also represent a fusion peptide sequence. This peptide is responsible for the coronavirus fusion activity. Figure 3 SARS-CoV transmission electron microscopy. In the supernatant of SARS-CoV infected cytopathic Vero E6 cells, characteristic virus particles can be found. The diameter of the viruses ranges between 60 nm and 120 nm and the virus shapes are round or oval. There are many protrusions from the envelope which are arranged in order with wide gaps between them. There are also many virus particles in the infected cells present. They often form a virus vesicle with an encircling membrane. A: Higher magnification B: Lower magnification. Scale bars represent 100 nm. Reproduced with permission from Acta Biochimica et Biophysica Sinica 2003, 35(6):587–591 [126]. The SARS-CoV has also been reported to cause the formation of syncytia in vivo, but so far only under the condition of cultured Vero cells [ 6 ]. The SARS-CoV S protein seems to have most of its characteristics in common with the S proteins of other coronaviruses, but it will be important for the understanding of the SARS-CoV pathogenic properties to identify the exact conditions of membrane fusion, i.e. pH dependency and protease sensitivity, which can increase the infectivity. The envelope and membrane proteins are integral membrane proteins and required for virus assembly [ 41 ]. In the case of the murine coronavirus MHV-A59 the coexpression of the E and M proteins but not the S or N proteins is needed for the release of virus-like particles (VLP) [ 42 ]. The nucleocapsid and viral core of the SARS-CoV are likely to be formed by the N protein. An interesting feature of the SARS-CoV and other coronaviruses is the resistance against the gastrointestinal fluids despite the lipid composition of their envelope. It has been reported that the SARS-CoV can survive in diarrheal stool for four days and also, patients with SARS often suffer from gastrointestinal symptoms with the virus to be detected in the stool [ 4 ]. As the molecular basis for the envelope's resistance against acidic environments and gastrointestinal enzymes is unclear, further research has to be carried out in this area which is important for the control of future SARS outbreaks. Evolution of the SARS virus It is unclear when and how novel pathogens such as the SARS-CoV cross the barriers between their natural reservoirs and human populations, leading to the epidemic spread of novel infectious diseases [ 43 ]. As with the SARS-CoV, new pathogens are believed to emerge from animal reservoirs and a variety of molecular mechanisms may contribute to the evolution of the viruses or bacteria. Due to the estimated error frequency of 1 × 10 -4 for RNA-dependent RNA polymerases [ 44 ], RNA viruses such as the SARS-CoV can undergo mutation at a high frequency. The SARS-CoV seems to be relatively genetically stable as the RNA sequences from different SARS patients were quite homogeneous. Even the entire genomic sequences of virus isolates from different continental areas did not differ by more than ten amino acids and it seems that two lineages of the virus can be traced [ 45 ]. This obvious contradiction to the high potential error rate of the RNA-dependent RNA polymerase suggests the presence of some proofreading mechanism connected with this enzyme. In fact, a detailed analysis of the SARS-CoV genome by bioinformatics indicates the presence of an exonuclease activity [ 27 ]. Next to mutations, a further threat of the SARS-CoV is based on the ability of coronaviruses to undergo RNA recombination at a high frequency [ 15 ]. For a variety of other coronaviruses, both recombination and mutation in natural infections have been shown to contribute to the diversification of the coronaviruses. Because of the demonstrated ability of coronaviruses to recombinate, the question whether the SARS-CoV will show a higher frequency of mutations within possible future seasonal changes or in respond to drug treatment is an issue of major concern. It was reported that in the initial phases of the SARS epidemic, the mutation rate was high in the gene for the spike protein, but this stabilized during the middle and final stages of the 2003 epidemy [ 28 ]. Thus, the virus had experienced great pressure to adapt to the new host after crossing the species barrier, but has then been optimized [ 28 ]. Duration of infection Although human coronaviruses are characteristically causing self-limiting short diseases, the question of potential chronic SARS infections is of major importance for a future disease control. If the SARS-CoV is able to cause a chronic persistent infection, chronic carriers may serve as sources for new SARS outbreaks. However, the detection of SARS-CoV in stool of patients for longer periods than 6 weeks after hospital discharge has not been reported so far. Therefore, the danger of chronic carriers may not be relevant. In contrast to common human coronavirus infections with short durations, most animal coronaviruses cause persistent infections. As an example, the feline coronavirus FIPV infects animals which then continue to shed virus for periods reaching up to seven months after infection without carrying disease symptoms [ 46 ]. Also, TGEV and MHV tend to cause chronic infections as these viruses may be found in the airways and small intestine (TGEV) or the nervous system (MHV) several months after infection [ 47 , 48 ]. Although the SARS-CoV has jumped to humans it may still have this property of inducing chronic infections. Thus, SARS-CoV RNA was found in patients' stool specimen more than 30 days after the infections. Clinical picture of SARS The mean incubation period of SARS was estimated to be 6.4 days (95% confidence interval, 5.2 to 7.7). The mean reported time from the onset of clinical symptoms to the hospital admission varied between three and five days [ 49 ]. Main clinical features of the disease are in the initial period common symptoms such as persistent fever, myalgia, chills, dry cough, dizziness, and headache. Further, although less common symptoms are sore throat, sputum production, coryza, vomiting or nausea, and diarrhea [ 50 , 51 ]. Special attention has been paid to the symptom of diarrhea: Watery diarrhea has also been reported in a subgroup of patients one week after the initial symptoms [ 52 ]. The clinical course of the disease seems to follow a bi- or triphasic pattern. In the first phase viral replication and an increasing viral load, fever, myalgia, and other systemic symptoms can be found. These symptoms generally improve after a few days. In the second phase representing an immunopathologic imbalance, major clinical findings are oxygen desaturation, a recurrence of fever, and clinical and radiological progression of acute pneumonia. This second phase is concomitant with a fall in the viral load. The majority of patients is known to respond in the second phase to treatment. However, about 20% of patients may progress to the third and critical phase. This phase is characterized by the development of an acute respiratory distress syndrome (ARDS) commonly necessitating mechanical ventilation. SARS in adults and children Rapid progress has been made in understanding the clinical presentation of SARS in adults and children [ 53 - 56 ]. In comparison to adults, SARS seems to be less aggressive in younger children, with no children in one case series requiring supplementary oxygen [ 57 ] while in adults, systemic infection as well as respiratory infection may be the rule. SARS is much milder with non-specific cold-like symptoms in children younger than 12 years than it is in adolescents and adults [ 58 ]. The reason for the milder clinical presentation of SARS in children is most likely due to differences in developmental stage of the immune system. The course of the disease in teenagers more likely resembles adults in concerning clinical presentation and disease progression [ 58 ]. SARS may also develop severe illness requiring intensive care and assisted ventilation in these adolescent patients. The common presenting features are fever, malaise, coryza, cough, chills or rigor, headache, myalgia, leucopaenia, thrombocytopaenia, lymphopaenia, elevated lactate dehydrogenase levels and mildly prolonged activated partial thromboplastin times [ 59 ]. The radiographic findings are non-specific: However, high-resolution computed chest tomography in clinically suspected cases may prove to be an early diagnostic aid when initial chest radiographs appeared normal. While rapid diagnosis with the first-generation RT-PCR assay was not satisfactory, improved RT-PCR assays may help to diagnose SARS in early stages. In this respect, a sensitivity approaching 80% in the first 3 days of illness when performed on nasopharyngeal aspirates may be achieved. The best treatment strategy for SARS among children still has to be determined while no case fatality has been reported in children. In comparison to the prognosis in adults, there is a relatively good short- to medium-term outcome. However, it is crucial to emphasize that continued monitoring for long-term complications due to the disease or its treatment is of major importance [ 60 ]. Molecular mechanisms of SARS virus pathogenesis Cytocidal mechanisms Coronaviruses are known to exert their effects by cytocidal and immune-mediated mechanisms. In vitro studies using cell culture assays have shown that coronavirus infection commonly results in cytopathic effects such as cellular lysis or apoptosis [ 61 ]. Also, the virus can cause cellular fusion leading to the formation of syncytia. These cytopathic effects are caused by steps of the viral replication such as the mobilisation of vesicles to form the viral replication complex [ 18 ], leading to the disruption of Golgi complexes [ 62 ]. Parallel to results on other coronaviruses, SARS-CoV has been shown to cause cytopathic effects in Vero cells and the formation of syncytia in lung tissues. A further similarity with other coronaviruses seems to be the potential of the SARS-CoV to cause tissue fibrosis [ 63 ]. As molecular mechanism for this fibrosis which has been reported for infections with the coronavirus MHV, the N protein has been demonstrated to induce promoter activity of the prothrombinase gene that correlates with fibrin deposition [ 64 ]. Immune-mediated mechanisms Next to cytocidal effects, also immune-mediated mechanisms of both the innate and adaptive immune system seem to contribute to the pathogenesis of SARS-CoV infections. In this respect, it has been shown that in MHV infection, T cells and cytokines play an important role in development of the disease [ 65 ]. Also, humoral antibodies have been reported to be crucial in infections caused by coronaviruses such as FIPV. Herein, antibodies against the spike protein were shown to be related to the induction of peritonitis [ 66 ]. For SARS-CoV infections, it has been reported that there seems to be an inflammatory cell influx consisting in particular of macrophages in the airways, and a massive release of cytokines during the peak of the infection [ 67 , 68 ]. It is therefore crucial that these immune mechanisms are further analysed on the molecular level as it seems appropriate that not only antiviral but also anti-inflammatory strategies are evaluated for a use in the clinical management of future SARS cases. The pharmacotherapy for SARS with anti-inflammatory steroids is controversial and largely anecdotal [ 69 ]. It was reported that the initial use of pulse methylprednisolone therapy appears to be more efficacious and equally safe when compared with regimens with lower dosage and should therefore be considered as the preferred steroid regimen in the treatment of SARS, pending data from future randomized controlled trials [ 70 ]. A further preliminary, uncontrolled study of patients with SARS, reported that the use of interferon alfacon-1 plus steroids was associated with reduced disease-associated impaired oxygen saturation and more rapid resolution of radiographic lung abnormalities [ 71 ]. Mechanisms of target cell specificity The most obvious gene which is likely to be a key modifier of SARS pathomechanisms is the spike (S) protein gene. As known for other coronaviruses, it does not only affect viral pathogenesis by determining the target cell specificity but also by other mechanisms. In this respect, a single mutation in the S gene of MHV has significant effects on the viral virulence and tissue tropism [ 72 ]. Also, mutations in the S gene led to the emergence of the weakly virulent PRCV from the virulent enteric TGEV [ 73 ]. Further potentially important genes are the 'non-essential' ORFs which show a significant divergence between SARS-CoV and other coronaviruses. In this respect, it was reported that the civet cat coronavirus has a 29-nucleotide deletion leading to a fusion of two non-essential ORFs into one new ORF in the SARS-CoV [ 18 ]. It was shown that deletion mutants of 'non-essential' ORFs of the group 2 mouse hepatitis virus (MHV) leads to a lower virulence without an impact on viral replication [ 74 ]. It has to be established if this also applies to 'non-essential' ORFs of SARS-CoV. Also, other viral gene products such as the M or E proteins may have an impact on the pathogenesis of the disease as they may induce interferon production or apoptosis [ 75 , 76 ]. Molecular targets for antiviral treatment The primary target cells of SARS-CoV infection are respiratory epithelial cells. As the virus can also be detected in stool specimen and patients with SARS often also have gastrointestinal symptoms, epithelial cells of the gastrointestinal tract also seem to be major target cells. Next to these epithelial cells, the SARS-CoV has also been found in macrophages and many other cells as it has been detected in not only in the respiratory tract and stool specimen but also in the blood, liver, kidney and urine [ 6 ]. In this respect, pathological examination did not only show changes in the respiratory tract, but also in splenic lymphoid tissues and lymph nodes. Furthermore, signs of a systemic vasculitis were found which included edema, localized fibrinoid necrosis, and infiltration of monocytes, lymphocytes, and plasma cells into vessel walls in the heart, lung, liver, kidney, adrenal gland, and the stroma of striated muscles. There was also thrombosis present in veins. Systemic toxic changes included necrosis and degeneration of parenchymal cells of the lung, heart, liver, kidney, and adrenal gland [ 77 ]. It may therefore be concluded that SARS can induce a systemic disease and thereby injuring many other organs apart from the respiratory tract. Target cell receptors The SARS-CoV target cell specificity is determined by the spike protein affinity to cellular receptors. In contrast to the all group III coronaviruses and the SARS-CoV for which the receptors have not been finally analyzed, it is known that group I coronaviruses bind to aminopeptidase N (CD13) as receptors [ 78 ], while group II coronavirus such as MHV use carcinoembryonic antigen (CEA) as receptor [ 79 ]. Recently, it was shown that a metallopeptidase, angiotensin-converting enzyme 2 (ACE2), efficiently binds the S1 domain of the SARS-CoV S protein. SARS-CoV replicated efficiently on ACE2-transfected but not mock-transfected 293T cells. Also, anti-ACE2 but not anti-ACE1 antibodies blocked viral replication on Vero E6 cells, indicating that ACE2 is a functional receptor for SARS-CoV [ 80 ] which was also identified by a further study [ 81 ]. Recently, the C-type lectin CD209L (also called L-SIGN) was discovered to be a further human cellular glycoprotein that can serve as an alternative receptor for SARS-CoV [ 82 ]. The interruption of virus-receptor interactions could be a potential target for future therapeutic strategies (Figure 4 ). In this respect, the receptor-binding S1 domain of the SARS-CoV S protein represents a possible target for new SARS antiviral drugs. Also, antibodies against ACE2, but not inhibitors binding to the active site of ACE2 may be useful for the development of therapeutic strategies. Figure 4 Potential target sites for therapeutic strategies. In view of the viral life cycle, there are several potential targets for the development of antiviral drugs. Starting from the binding of the virus to the target cell, the spike protein or receptors such as angiotension-converting enzyme 2 (ACE2), cell entry or the different replication steps may be targeted. After replication, virus assembly and exit mechanisms may also be used for antiviral strategies. VLP, virus-like particles. Virus entry After binding to the receptor, the next molecular step of potential use for the development of anti-SARS drugs is the virus entry into the cells. While most coronaviruses enter their target cells via plasma membrane fusion, a further entry mechanism may be acidic pH-dependent endocytosis [ 83 ]. Focusing on these mechanisms, it will be crucial to gain further knowledge about SARS-CoV fusion activity. As a drug development candidate, a putative fusion peptide has good potential (Figure 4 ). Intracellular replication After the binding to a host cell receptor and entry into the cells, the molecular steps of transcription, translation and protein processing display further potential targets for new therapeutic strategies. In this respect, the RNA-dependent RNA polymerases (SARS-CoV RdRp) may be a potential target for a future anti-SARS therapy. A recent study located its conserved motifs and built a three-dimensional model of the catalytic domain [ 84 ]. The authors suggested that potential anti-SARS-CoV RdRp nucleotide-analog inhibitors should feature a hydrogen-bonding capability for the 2' and 3' groups of the sugar ring and C3' endo sugar puckering. Also, the absence of a hydrophobic binding pocket for non-nucleoside analog inhibitors similar to those observed in hepatitis C virus RdRp and human immunodeficiency virus type 1 reverse transcriptase seems to be crucial [ 84 ]. Also, protease activity is crucial for SARS-CoV RNA replication and protein processing [ 29 , 85 ], and the inhibition of protease function leads to an immediate stop of viral RNA synthesis. Most of the coronaviruses express one major cysteine proteinase, called the main proteinase (M pro ) or the 3C-like proteinase (3CL pro ), and two auxiliary, papain-like proteinases (PL1 pro and PL2 pro ). The latter two are responsible for the cleavage of the viral polyproteins, pp1a and pp1ab, at three sites near the amino-terminus, while the M pro processes these proteins at as many as 11 additional sites. Interestingly, SARS-CoV lacks the PL1 pro [ 16 , 17 ], but it can be assumed that its action is taken over by the PL2 pro [ 29 ]. This is conceivable since operation of the PL2 pro on PL1 pro cleavage sites has been shown in IBV and HCoV [ 86 ]. Roughly at the position of the PL1 pro gene in other coronavirus genomes, SARS-CoV displays a domain within ORF1a that lacks any detectable sequence homology and has therefore been named the SARS-unique domain (SUD) [ 27 ]. It is not known whether the SUD protein is ever expressed in the life cycle of SARS-CoV but if it is, it may be connected to the high pathogenicity of SARS-CoV compared to other human coronaviruses and, therefore, it may constitute an attractive target for therapeutic intervention. Crystal structures have been determined for the M pro s of TGEV [ 87 ], HCoV 229E [ 85 ], and, more recently, SARS-CoV [ 88 ]. They all show a similar overall architecture for the 34 kD enzyme which forms a dimer in the crystals and also at intermediate and high concentrations in solution. The monomer consists of three domains of which the first two are β-barrels with an overall similarity to the 3C proteinases of picornaviruses and to the serine proteinase, chymotrypsin. The third domain is α-helical and was shown to be essential for dimerization [ 85 , 87 , 88 ]. The active site of the enzyme is located in a cleft between domains I and II and comprises a catalytic dyad of Cys...His, rather than the catalytic triad common for cysteine and serine proteinases. Anand et al. [ 85 ] have synthesized a substrate-analogous hexapeptidyl chloromethylketone inhibitor and bound it to TGEV M pro in the crystalline state. The X-ray structure of the complex revealed binding of the P1 glutamine, P2 leucine, and P4 threonine side chains of this compound to the respective subsites in the substrate-binding cleft, in agreement with the pronounced specificity for cleavage by the M pro after the substrate sequence (Thr, Val, Ser)-Xaa-Leu-Gln. The structure also showed the expected covalent attachment of the methyl ketone group at P1 of the inhibitor to the catalytic cysteine of the enzyme. In spite of 40% and 44% sequence identity, respectively, to the M pro s of HCoV 229E and TGEV, the crystal structure of the SARS-CoV M pro revealed some surprises [ 88 ]. Within the dimer, one molecule was in the active conformation seen in the other structures, whereas the other one adopted a catalytically incompetent conformation. This enzyme had been crystallized at a pH value of <6, which in one of the monomers apparently led to the protonation of a histidine residue at the bottom of the S1 specificity pocket. This resulted in major conformational rearrangements leading to the collapse of this binding site for the P1 glutamine residue of the substrate and to a catalytically incompetent conformation of the oxyanion-binding loop. However, when the crystals were equilibrated at higher pH values, their X-ray structures revealed the active conformation for both monomers in the dimer. This pH-dependent activation mechanism allows interesting conclusions to be made for the self-activation of the M pro from the viral polyprotein, which probably involves a pH-dependent step. The same hexapeptidyl chloromethylketone inhibitor used by Anand et al. [ 85 ] in their crystallographic study of the TGEV M pro was employed by Yang et al. [ 88 ] to characterize the interaction of the SARS-CoV enzyme with substrate. This was performed by soaking the inhibitor into crystals grown at the low pH. In spite of the inactive conformation of one of the two monomers in the dimer being preserved, the compound was found to bind to it, but with its P1 glutamine side chain pointing towards bulk solvent rather than into the S1 binding site, because of the collapse of the latter. The binding mode of the inhibitor to the active monomer was also somewhat unusual and is not fully understood at present. On the basis of their crystallographic work, Anand et al. [ 85 ] found that the binding mode of their hexapeptidyl chloromethylketone inhibitor to the TGEV M pro resembled that of AG7088 in complex with its target, the 3C proteinase of human rhinovirus [ 89 ], even though the respective target enzymes displayed large structural differences except in the immediate neighbourhood of the active site. AG7088 is in phase II/III clinical studies as an inhalation treatment for the common cold as caused by human rhinovirus. Anand et al. [ 85 ] therefore proposed that AG7088 should be a good starting point for the design of anti-SARS drugs, and indeed, the manufacturer of AG7088 confirmed only a few days after their proposal had appeared on-line that the compound was effective against SARS coronavirus in cell culture. AG7088 is now the subject of intensive optimization efforts [ 90 ]. Other studies used molecular dynamics simulations of the M pro and screened 29 approved and experimental drugs against a model of the SARS CoV proteinase as well as the experimental structure of the transmissible gastroenteritis virus (TGEV) proteinase [ 91 ]. It was suggested that existing HIV-1 protease inhibitors, L-700,417 for instance, may have high binding affinities and may therefore provide another good starting point for the future design of SARS-CoV proteinase inhibitors [ 92 ]. However, this has to be proved experimentally. Further potential targets are the E and M proteins (Figure 4 ) as they represent the minimum essential components for the assembly of coronaviruses which form the virus-like particles [ 41 , 42 ]. Ultrastructurally, the process of SARS-CoV assembly is most likely localised to the ER-Golgi intermediate compartment [ 93 ]. Together with strategies that may focus on the inhibition of virus assembly, the virus exit through secretory pathways is also of interest for the development of new antiviral compounds. With regard to the multitude of potential epithelial target cells, specific endogenous drug delivery systems may also be of relevance. In this respect, the family of peptide transporters consisting of PEPT1 and PEPT2 which are differentially expressed in potentially infected cells of the respiratory tract [ 94 , 95 ], small intestine [ 96 ], kidneys [ 97 , 98 ], nervous system [ 99 ] and other organs [ 100 ], may serve a target for the rational drug design of antiviral drugs. So far, a variety of antiviral drugs or prodrugs such as valacyclovir [ 101 ], valganciclovir [ 102 ] or the valyl ester of zidovudine [ 103 ] have been shown to be transported via these systems and minimal structure requirements for substrate transport have been determined [ 104 ]. A further tool which may be used to approach antiviral therapies is the technique of small interfering RNAs (siRNAs). SiRNAs are double-stranded RNAs which lead to a sequence-specific degradation of mRNAs [ 105 ]. Recent in vitro studies used six 21-mer siRNAs that were targeted to different sites of the replicase 1A region of SARS-CoV [ 106 ]. Monkey kidney cells (FRhk-4) were infected with the SARS-CoV GZ50 strain and transfected eight hours later with the siRNAs. Three of the six siRNAs led to a marked inhibition of virus cytopathic effects and a reduction of virus copies between 85 and 92 %, indicating that siRNAs may have a potency as antiviral treatment options and that the 1A region displays a promising region to suppress virus replication [ 106 ]. Vaccines against the SARS virus As most patients develop an immunity against the SARS-CoV and survive the infection, the possibility of creating an effective and safe vaccine seems to exist [ 107 ]. There are several options to develop vaccines against the SARS-CoV [ 108 ]. Live-attenuated vaccines Live-attenuated coronavirus vaccines can be generated by deletions in "group-specific genes". The deletions of these genes do not change replication properties but attenuate the virus [ 109 ]. Examples for the use of live-attenuated vaccines to prevent coronavirus infections are live attenuated IBV vaccines which are used in broiler chickens [ 110 ]. For the animal coronavirus infections, live attenuated vaccines have been proven to be significantly more effective than whole killed vaccines, indicating that cell-mediated immunity is a crucial defence mechanism. However, the great threat remains that a vaccine strain can recombine with a circulating wild type strain [ 111 ] and without evidence that recombination and reversion of a live-attenuated SARS-CoV to virulence can not occur, it is unlikely that a live attenuated SARS-CoV vaccines will be developed and used. Whole killed vaccines Whole killed vaccines are generally safe and easy to generate. In fact, this technique has been applied in veterinary medicine to generate vaccines for BoCV and IBV [ 112 ]. Also, an inactivated canine coronavirus vaccine has been produced [ 113 ]. A SARS inactivated vaccine was recently developed using the SARS coronavirus (SARS-CoV) strain F69 treated with formaldehyde and mixed with Al(OH)(3) [ 114 ]. However, killed vaccines may not protect against different strains of coronaviruses, and live attenuated vaccines have been shown to be more effective than whole killed vaccines in preventing coronavirus animal infections [ 115 ]. Recombinant subunit vaccines Using molecular biology techniques to generate large quantities of recombinant viral proteins, recombinant subunit vaccines, e.g. against the spike protein, are expected to be created relatively easy as shown by two recent studies [ 116 , 117 ]. Eight recombinant human single-chain variable region fragments (scFvs) against the S1 domain of spike (S) protein of the SARS-CoV from two nonimmune human antibody libraries were screened and one scFv 80R efficiently neutralized SARS-CoV and inhibited syncytia formation between cells expressing the S protein and those expressing the SARS-CoV receptor angiotensin-converting enzyme 2 (ACE2) [ 117 ]. A recent study used the SARS-CoV spike protein receptor binding domain (aa 318–510) for immunization, which resulted in the induction of effective neutralizing antibodies [ 118 ]. However, recombinant subunit vaccines may have a limited ability to protect against SARS-CoV infections in view of the variations which may arise in the viral genome in future outbreaks. Therefore, the approach of recombinant subunit vaccines may have to be supplemented by further vaccine strategies which focus on cell-mediated immunity. Recombinant vectored vaccines An approach using recombinant vectored vaccines with DNA or a viral vector could be a promising target. The DNA prime and adenovirus or MVA boost approach which is currently analysed for a potential use in the development of HIV vaccines, may also offer a strategy to prevent SARS infections. In this respect, a multi-valent approach which induces both humoral and T cell-mediated host responses seems to be the most attractive strategy. From the field of veterinary medicine, data on this approach are already available: A recombinant fowlpox with the S1 gene of IBV was demonstrated to be relatively protective against IBV [ 119 ]. Also, a DNA vaccine was developed which contains the nucleocapsid protein gene of porcine transmissible gastroenteritis virus (PTGV). This vaccine was shown to initiate both humoral and cell-mediated immune host responses [ 120 ]. Recently, three murine studies demonstrated that DNA vaccines encoding different SARS-CoV antigens are capable of generating humoral and cellular immunity and may potentially be useful for control of infection with SARS-CoV [ 121 - 123 ]. However, it was also shown that immunization with modified vaccinia virus Ankara-based recombinant vaccine against SARS is associated with enhanced hepatitis in ferrets [ 124 ]. Epitope-based vaccines A further strategy is based on the use of epitopes which can be delivered using viral or DNA vectors. Such an epitope-based strategy for coronavirus vaccination has already been reported [ 125 ] and the major advantages is the prevention of a possible vaccine reversion to virulence. A further benefit of this technique is the possibility to eliminate any regions of the viral genomic sequence which be associated with a potential autoimmune effects. The limitation of this approach is mainly based on potential variations. In this respect, epitopes which frequently undergo mutations will not protect against the SARS-CoV infections if used in epitope-based vaccines. If the SARS-CoV evolves as a highly variable virus, it will be crucial to identify highly conserved epitopes of the virus. In summary, the important development of SARS vaccines can be approached using several techniques which should ideally encompass the induction of both humoral and cell-mediated mechanisms. As coronavirus vaccines in animals have partly been reported to cause an enhancement of viral infections [ 66 ], a cautious approach has to be followed. A first study has investigated the ability of adenoviral delivery of a codon-optimised SARS-CoV spike protein S1 fragment, membrane protein, and nucleocapsid protein to induce immunity in rhesus macaques. The immunization with a combination of these three Ad5-SARS-CoV vectors and a booster vaccination on day 28 demonstrated antibody responses against the spike protein S1 fragment. Also T-cell responses against the nucleocapsid protein were found and all vaccinated animals displayed strong neutralising antibody responses in vitro. These results indicated that an adenoviral-based vaccine can induce SARS-CoV-specific immune responses in monkeys. Conclusion In summary, the onset of the SARS epidemic in different continents has led to the formation of a successful laboratory network to identify the molecular mechanisms underlying the SARS infection. Next to the development of early diagnostic tests and effective treatment strategies, it is most important to orchestrate research activities which lead to the development of vaccines and antiviral agents, as there is no established therapy to date. Even now in a situation of only a handful of new cases, SARS remains a major global health hazard which may reappear.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548145.xml
535810	Thromboembolic events and haematological diseases: a case of stroke as clinical onset of a paroxysmal nocturnal haemoglobinuria	Some haematological diseases are associated to an increased risk of thromboembolic events. We report a case of paroxysmal nocturnal haemoglobinuria (PNH) in which a cerebrovascular event represented the first clinical manifestation of disease. PNH is associated to thromboembolic events, generally of venous districts often involving unusual locations such as mesenteric vessels, sagittal veins, inferior vena cava and renal veins. To our knowledge arterial thrombotic episodes are rare and the involvement of arterial cerebral vessels is exceptional. Then, our case points out the importance of investigating about haematological disorders in all patients presenting with a stroke, in which the common predisposing conditions are excluded.	Background Paroxysmal nocturnal haemoglobinuria (PNH) is an acquired clonal disorder of haematopoietic stem cells clinically characterized by acute intravascular haemolytic crisis, in particular nocturnal, often overlapped to chronic haemolysis, and by thrombotic events and bone marrow failure. It is associated with a somatic mutation in the phosphatidylinositol glycan complementation class A (PIG-A) gene, mapped to the X chromosome; the subsequent deficiency of glycosylphosphatidylinositol (GPI) and of GPI-anchored molecules, as the decay accelerating factor (DAF or CD55) and the membrane inhibitor of reactive lysis (MIRL or CD59), causes an increased susceptibility to complement-mediated lysis of erythrocytes, leukocytes and platelets [ 1 ]. The association between PNH and thromboembolic accidents, generally manifesting as thrombotic events of venous vessels sometimes complicated by pulmonary embolism, is well established. Arterial thrombotic episodes, particularly of cerebral vessels are enough rare [ 2 ]. We report a case of PNH presenting with thromboembolic events, both venous (proximal deep venous thrombosis of lower limbs) and arterial (stroke). Case history Clinical summary A 56-year-old woman, with history of peptic ulcer and family history for cerebrovascular disease was referred to our Division of Internal Medicine with asthenia and generalized discomfort. She reported a cerebrovascular accident manifesting as a right brachial and crural hyposthenia ten month ago, almost completely receded at observation time; she also referred recurrent episodes of proximal deep venous thrombosis (DVT) of lower limbs in the last seven months. Pathological findings In order to identify any hypercoagulable state (i.e. inherited or acquired thrombophilia), in view of her personal and familiar history, we tested prothrombin time, as INR, activated partial thromboplastin time, as ratio, fibrinogen, protein C and S, antithrombin III, activated protein C resistance, anti-cardiolipin antibodies IgG and IgM, lupus anticoagulant, plasminogen activator inhibitor type 1, d-dimer, gene polimorphism of clotting factor II and V, gene polimorphism C9774T and G3775A of apolipoprotein B and gene polimorphism C3932T and C4070T of apolipoprotein E resulted all in normal range; while gene polimorphism of tetrahydrofolate reductase and angiotensin converting enzyme revealed heterozigosity for both. Subsequently, homocysteinemia test revealed mild hyperhomocysteinemia. All thrombophilic tests are summarised in table 1 . Table 1 Thrombophilic tests Thrombophilic tests (units of measurement) Results Normal range Protein C (antigen) (%) 99% 60–125 Protein S (antigen) (%) 102% 60–125 Antithrombin (activity) (%) 105% 80–120 Activated protein C resistance (Bertina) 0,90 >0,77 Anti-cardiolipin antibodies IgG (U/GPL) 4 <7 Anti-cardiolipin antibodies IgM (U/MPL) 2 <4 Lupus anticoagulant absent absent Plasminogen activator inhibitor type 1 (ng/dl) 30 4–44 PTHRA20210 gene polimorphism wild type wild type Factor V Leiden gene polimorphism wild type wild type Apolipoprotein B gene polimorphism C9774T and G3775A wild type wild type Apolipoprotein E gene polimorphism C3932T and C4070T wild type wild type Methylene-tetrahydrofolate C677T gene polimorphism heterozigosity wild type Angiotensin converting enzyme deletion gene polimorphism insertion/deletion insertion/insertion Homocysteinemia (μM) 22 5–15 Prothrombin time (INR) 0.95 0.8–1.2 Activated partial thromboplastin time (ratio) 0.92 0.8–1.2 Fibrinogen (mg/dl) 305 220–400 D-dimer (ug/l) 188 0–198 A magnetic resonance imaging scan showed little and multiple ischemic lesions in particular in left cerebral peduncle (fig 1A ), semioval centres (fig 1B ), left pons and midbrain. Moreover, a vascular ultrasound examination ruled out the presence of significant stenosis of arterial cerebral vessels and confirmed proximal DVT and post-thrombotic syndrome of lower limbs. Figure 1 Magnetic resonance imaging scan showing multiple ischemic lesions in left cerebral peduncle (1A) and semioval centres (1B). Other available data showed: red blood cells 2.470.000/mm 3 , hemoglobin 7.9 g/dl, hematocrit 25%, mean corpuscular volume 99,6 fl, mean corpuscular hemoglobin 32 pg, mean corpuscular hemoglobin concentration 32 gr/dl, white blood cells 4,040/mm 3 , platelets 93.000/mm 3 , reticulocytes 5,4%, serum iron 76 μg/dl, erytro-sedimentation rate 1° hour 40 mm, lactate dehydrogenase 944 UI/l, total bilirubin 0,72 mg/dl, indirect bilirubin 0,36 mg/dl, and presence of hemoglobinuria. Coombs' test, cold agglutinins, antinuclear antibodies, anti-extractable nuclear antigens antibodies, anti-mithocondrial antibodies, anti-smooth muscle antibodies were negative. Laboratory data and their range are summarised in table 2 . Table 2 Other laboratory findings Laboratory data (units of measurement) Results Normal range Erytro-sedimentation rate 1° hour (mm) 40 <10 lactate dehydrogenase (UI/l) 944 100–190 total bilirubin (mg/dl) 0,72 0–1 indirect bilirubin (mg/dl) 0,36 0–0,5 antinuclear antibodies absent absent anti-extractable nuclear antigens antibodies absent absent anti-mithocondrial antibodies absent absent anti-smooth muscle antibodies absent absent An abdominal ultrasonography excluded a hypersplenism and/or Kasabath-Merritt syndrome. A peripheral blood smear did not show any finding suggestive for haematological disorders. A bone marrow biopsy showed a slight hyperplasia of erythrocytic bone marrow cell line. These laboratory and morphological findings suggested a non-immune haemolytic anemia. In particular, due to the exclusion of other non-immune haemolityc disorders by means of age and clinical history together with the presence of hemoglobinuria and pancytopenia, we hypothesized paroxysmal nocturnal hemoglobinuria. This diagnosis was confirmed by an immunophenotypic profile of peripheral blood cells, showing a 15% of deficient CD59 erythrocytes, and by the presence of hemosiderinuria. Haematological findings are summarised in table 3 . Table 3 Haematological data Laboratory data (units of measurement) Results Normal range red blood cells (cells/mm 3 ) 2.470.000 4.200.000 – 5.400.000 hemoglobin (g/dl) 7,9 12–16 hematocrit (%) 25 37–45 mean corpuscolar volume (fl) 99,6 81–99 mean corpuscolar hemoglobin (pg) 32 27–31 mean corpuscolar hemoglobin concentration (g/dl) 32 32–36 white blood cells (cells/mm 3 ) 4.040 4.800 – 10.800 Platelets (cells/mm 3 ) 93.000 130.000 – 400.000 Reticulocytes (%) 5,4 <2 haemoblobinuria traces absent Hemosiderinuria present absent Coombs'test negative negative cold agglutinins negative negative Peripheral blood smear Normal Bone marrow biopsy slight hyperplasia of erythrocytic cell line Immunophenotypic profile of peripheral blood cells 15% of deficient CD59 erythrocytes During her hospitalization two haemotrasfusions were necessary in occasion of two concurrent haemolytic crises. Following dismission, in order to prevent further thromboembolic events, the patient began oral anticoagulation therapy with warfarin according with INR value in range of 2–2,5. Moreover she was treated with B12 vitamin and folate supplementation. Discussion The association between haematological diseases and thromboembolic events is well established. In particular high thrombotic risk is recognized in patients with essential thrombocythemia, polycythemia vera, PNH and drepanocytosis [ 3 ]. PNH is associated to venous thrombosis in approximately one third of cases. The most frequently reported locations are unusual such as mesenteric vessels, sagittal veins, inferior vena cava and renal veins. When thrombosis occurs in the pre-hepatic or hepatic veins, the patient develops a Budd-Chiari syndrome [ 4 ]. Arterial thrombosis is rare, even if few cases of cerebral arterial thrombosis [ 5 ] and acute myocardial infarction [ 6 ] are described in the literature. The mechanism whereby PNH causes an hypercoagulable state is not clear. PNH platelets lack the GPI-linked proteins CD55 and CD59, and respond to the deposition of terminal complement components by vesiculations of portions of their plasma membrane, resulting an increased procoagulant property. PNH cells also lack the receptor of the GPI-linked urokinase plasminogen activator, which may result in impaired fibrinolysis [ 4 ]. Also an increase of membrane-derived procoagulant microparticles (phosphatidylserin) stemming from the platelets of PNH patients has been described [ 3 ]. In our case, thrombotic events represented the clinical onset of PNH and involved both venous (DVT) and arterial (stroke) vessels. Neurological manifestations in PNH patients are generally due to cerebral venous thrombosis [ 7 , 8 ], even if a few cases of cerebral arterial episodes, involving large vessels, are described. However, usually cerebral ischaemia in PNH did not occur as presenting sign of the disease nor affect small and middle cerebrovascular arteries [ 5 ]. In our patient the relationship between PNH and thrombotic events is strongly suggested, especially after excluding inherited or acquired thrombophilia and atherosclerotic risk factors. Heterozigosities for gene polimorphism of tetrahydrofolate reductase and angiotensin converting enzyme, detected in our patient, are not associated to an increased risk of stroke, while acquired or inherited hyperhomocysteinemia may be involved [ 9 - 11 ]. Also haemotological findings agree with PNH diagnosis because of the association of thrombosis, anemia and thrombocytopenia. We excluded further causes of non-immune haemolityc anemia (i.e. spherocytosis, enzymatic disorders, microangiopathic anemia) and thrombocytopenia (i.e. disseminated intravascular coagulation, haematological malignancies, systemic erythematosus lupus, primary or secondary antiphospholipid syndrome, hypersplenism). In conclusion, PNH is associated to thromboembolic events, especially in the venous district and should be considered as a possible cause of an hypercoagulable state, in particular when unusual vascular locations are involved. Our case indicates the possibility of arterial thrombotic episodes in a patient with PNH and suggests a thorough evaluation of any haematological disorders in patients presenting with stroke or myocardial infarction, especially in the absence of atherosclerosis risk factors and/or a thrombophilic state.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535810.xml
535804	Testicular seminoma after the complete remission of extragonadal yolk sac tumor : a case report	Background Between 2% and 5% of malignant germ-cell tumors in men arise at extragonadal sites. Of extragonadal germ cell tumors, testicular carcinoma in situ (CIS) are present in 31–42% of cases, and CIS are reported to have low sensitivity to chemotherapy in spite of the various morphology and to have a high likelihood of developing into testicular tumors. A testicular biopsy may thus be highly advisable when evaluating an extragonadal germ cell tumor. Case presentation A 36-year-old man was diagnosed as having an extragonadal non-seminomatous germ cell tumor, that was treated by cisplatin-based chemotherapy, leading to a complete remission. In the meantime, testicular tumors were not detected by means of ultrasonography. About 4 years later, a right testicular tumor was found, and orchiectomy was carried out. Microscopically, the tumor was composed of seminoma. Conclusions We herein report a case of metachronous occurrence of an extragonadal and gonadal germ cell tumor. In the evaluation of an extragonadal germ cell tumor, a histological examination should be included since ultrasonography is not sufficient to detect CIS or minute lesions of the testis.	Background Between 2% and 5% of malignant germ-cell tumors in men arise at extragonadal sites [ 1 ]. Cytogenetically most extragonadal germ-cell tumors (EGGCTs) i.e., the seminomas and non-seminomas, are similar to their testicular counterparts [ 2 , 3 ]. But there are CIS in 31–42% of EGGCT patients' testes [ 4 ]. Ultrastructural studies indicate that CIS originate from rather primitive cells and can develop into different categories of germ cell carcinomas. Furthermore, since CIS are reported to respond poorly to chemotherapies, a metachronous development of testicular cancer will possibly occur, in spite of the various morphologies of testicular cancer [ 5 , 6 ]. In the present case, the etiology of a metachronous appearance of EGGCT and testicular cancer is discussed. Case presentation A 36-year-old man was admitted to our hospital with the chief complaint of right-sided scrotal enlargement. He had previously received treatment for an extragonadal germ cell tumor. At the age of 32, he presented with lumbago. CT showed a retroperitoneal tumor (Figure 1 ), and a transabdominal needle biopsy was carried out. Microscopically, the tumor was composed of a yolk sac tumor (Figure 2A,2B ). We performed two courses of systemic chemotherapy using bleomycin, etoposide, and cisplatin, leading to a partial response. As the tumor size was not seemed to decrease after the two courses of the chemotherapy, retroperitoneal lymph node dissection was performed, but failed to show any residual viable cells. An ultrasonic study did not reveal any testicular tumors. About 4 years after the previous treatment, he presented with scrotal enlargement and tumor markers such as AFP and HCG β were within normal limit. A right orchiectomy was performed on 23 rd July. Pathology showed the resected tumor was a seminoma with CIS (Figure 3A,3B ). No recurrence has been seen since the surgery (Figure 1B ). Conclusions There are reports that approximately 4% of patients with EGGCT develop a metachronous testicular cancer despite the use of cisplatin-based chemotherapy [ 7 ], and the cumulative risk of developing a metachronous testicular cancer 10 years after a diagnosis of EGGCT is 10.3% [ 8 ]. However, there is disagreement over whether EGGCT is a primary disease or metastatic from the burned-out primary testicular lesion. Actually, burned-out tumors have been detected in 76% of cases of EGGCT [ 9 ]. CIS is also found via biopsy in 31–42% cases [ 4 , 10 ]. Testicular CIS is thought to have resistance to systemic chemotherapies and to develop later to metachronous testicular cancer. In the present case, the EGGCT was a non-seminomatous germ cell tumor including a yolk sac component, whereas the testicular cancer was a seminoma. We believe CIS was present at the time of the treatment of EGGCT and testicular CIS is so primitive that it could differentiate into any type of germ cells. It is also possible that these metachronously developing germ cell tumors developed independently. A testicular biopsy could clarify the relationship between these tumor cells and the expansion of the disease. In the present case no biopsy was done, but an ultrasonic examination ruled out the possibility of testicular CIS. Giwercman et al. [ 11 ] emphasized the necessity of histological examinations of the testis upon an evaluation of EGGCT [ 12 - 14 ] and also urged a careful follow-up for patients with EGGCT who do not have simultaneous testicular cancer. On the other hand, there is an opinion that any patients with retroperitoneal masses should undergo scrotal ultrasound. Comiter et al. [ 15 ] showed definite pathological evidence of a burned-out testicular carcinoma in 5 of 6 patients (83%) with presumed extragonadal germ cell tumors and concluded that scrotal ultrasound studies are useful for the evaluation of the palpably normal testes [ 15 ]. Kitahara et al. reviewed the incidence of scrotal echogenic leisions with testicular cancers or burned-out tumors of 22 EGGCT patients and found echogenic changes in 17 patients (77.3%) [ 16 ]. This means that disease was overlooked by ultrasonic examinations in 22.7% of cases. In our case, it is possibile that metachronous testicular cancer oriented in testicular CIS, grew from a burned-out tumor, or was independent of the EGGCT. We should have performed testicular biopsies at the time of the diagnosis of EGGCT and reflect the strategies of treatments of EGGCT. Now we propose a surveillance protocol of EGGCT as Table 1 , concerning with following four points. 1. As we mentioned, the overall risk of development a testicular tumor is not so high(4–10.3%). 2. The side effect of CIS therapy (whether irradaition, orchiectomy or chemotherapy) are significant, especially concerning fertility and androgen production. 3. Testicular tumors early detected by adequate surveillance respond well to treatments. 4. Testicular biopsy is not entitled to detect all the CIS. Competing interests The author(s) declare that they have no competing interests. Authors' contribution IK, MU, HY, KN, TT and ND carried out clinical treatments. TM carried out histopathological studies. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535804.xml
553968	Determining the quality of educational climate across multiple undergraduate teaching sites using the DREEM inventory	Background Our obstetrics and gynaecology undergraduate teaching module allocates 40–50 final year medical students to eight teaching hospital sites in the West Midlands region. Based on student feedback and concerns relating to the impact of new curriculum changes, we wished to objectively assess whether the educational environment perceived by students varied at different teaching hospital centres, and whether the environment was at an acceptable standard. Methods A Dundee Ready Education Environment (DREEM) Questionnaire, a measure of educational environment, was administered to 206 students immediately following completion of the teaching module. Results The overall mean DREEM score was 139/200 (70%). There were no differences in the education climate between the teaching centres. Conclusion Further research on the use of DREEM inventory, with follow up surveys, may be useful for educators to ensure and maintain high quality educational environments despite students being placed at different teaching centres.	Background The undergraduate curriculum at our medical school was redesigned in 1998/99 to bring it in line with recommendations suggested by the General Medical Council (GMC) in Tomorrow's Doctors [ 1 ]. Obstetrics and Gynaecology is taught as a final year module. Around 20–30 students, of a total year group of around 200 students, are allocated to eight teaching hospital sites in the West Midlands region, and remain their for the length of the module (eight weeks). Throughout the placement, all formal lectures take place at the principal Teaching Hospital (Birmingham Women's Hospital). A comprehensive course handbook and web-based multiple choice formative assessment accompany the module, and detail the teaching, practical and assessment objectives for students and clinicians. We have aimed to ensure there are no significant differences in the way the curriculum is delivered between centres. All 200 students sit the final exam in Obstetrics and Gynaecology straight after completing the 8-week course module. Based on previous student feedback reporting differences in educational experiences, together with our concerns relating to the impact of new curriculum changes, we wished to objectively assess whether the educational environment perceived by students varied at different teaching hospital centres, and whether the environment was at an acceptable standard. In particular, was there any potential loss of teaching experience when students were placed away from the principal Teaching Hospital. Thus, the null hypothesis we wished to test was that there was no difference in the learning environment between centres. Several questionnaire-based educational tools are available that set out to 'quantify' the educational environment [ 2 - 4 ]. However, we chose to use the Dundee Ready Education Environment Measure (DREEM) inventory, as more studies had evaluated and validated this method [ 5 , 6 ]. The DREEM inventory consists of 50 questions, each scoring 4, giving a total maximum individual DREEM score of 200. The five domains that comprise the DREEM are depicted in Table 1 . Table 1 Domains of DREEM questionnaire TOPIC Number of questions Maximum DREEM Score Students' Perception of Learning 12 48 Students' Perception of Teachers 11 44 Students' Academic Self-Perceptions 8 32 Students' Perception of Atmosphere 12 48 Students' Social Self-Perceptions 7 28 Total 50 200 Methods The DREEM questionnaire, based on a Likert scale, was administered to the full class of 206 final-year Birmingham University medical students undertaking the exam module in Obstetrics and Gynaecology in 2000. All questionnaires were distributed and returned the same day of the exam, which allowed us to achieve a 100% response rate. Students were told to only comment on their recent 8-weeks experience of Obstetrics and Gynaecology. Statistical analysis was performed using Microsoft Excel and Arcus Quickstat Biomedical Statistical software, and utilised single-sample T test and One-way analysis of variance (ANOVA). Results The year group comprised 42% male and 58% female. The overall mean DREEM score for the study group was 139/200 (95% CL 136.1 to 141.9), or expressed as percentage of the maximal score, 70% (95% CL 68% to 71%). There was no statistically significant difference between the mean scores for the contributory DREEM domains, which were as follows: perception of learning, 34.52/48 (72%); perception of teaching, 32.05/44 (73%); academic self-perception, 19.46/32 (61%); perception of atmosphere, 34.07/48 (71%), and for social self perceptions, 18.90/28 (68%). The DREEM scores for each hospital, with comparison of all contributory elements of the DREEM inventory, are depicted in Table 2 and Figure 1 . Table 2 The DREEM domains and overall score for each hospital HOSPITAL Number of Students LEARNING Mean Score/48 TEACHERS Mean Score/44 ACADEMIC SELF-PERCEPTION Mean Score/32 ATMOSPHERE Mean Score/48 SOCIAL Mean Score/28 OVERALL DREEM Score/200 DREEM percentage for each hospital (total of 206) Percentage of maximum score Percentage of maximum score Percentage of maximum score Percentage of maximum score Percentage of maximum score Percentage of maximum score BWH 53 33.77 31.89 19.77 33.40 19.32 138.15 69% Good Hope 20 33.30 30.10 18.15 33.40 18.20 133.15 67% B'ham Heartlands 26 34.15 32.73 18.92 32.77 19.58 138.15 69% Walsall Manor 20 34.10 34.35 19.90 34.30 19.05 141.70 71% City 32 35.13 28.31 19.41 34.97 17.75 135.56 68% Wolverhampton 22 35.77 32.59 20.41 33.64 18.86 141.27 71% Shrewsbury 13 34.77 32.85 18.69 35.00 19.15 140.46 70% Wordsley 20 35.15 33.60 20.40 35.10 19.30 143.55 72% Mean overall 34.52 72% 32.05 73% 19.46 61% 34.07 71% 18.90 68% 139.00 70% Lower 95% CL 33.83 70% 30.41 69% 18.77 59% 33.33 69% 18.38 66% 136.13 68% Upper 95% CL 35.21 73% 33.69 77% 20.14 63% 34.82 73% 19.42 69% 141.88 71% CL Confidence Limit Figure 1 Graphical representation of the contribution of each DREEM domain to the overall mean DREEM score When converting the raw DREEM score to percentages, two-sided P-value single-sample Student's T test showed no statistically significant difference between hospitals by each DREEM domain, or between each DREEM domain within the same hospital. Greatest variation between hospitals occurred in the Students' Perception of Atmosphere domain, where there were four hospitals beyond the 95% Confidence Limits; this compared to three hospitals beyond 95% Confidence Limits in all other DREEM domains (Table s 2 ). One-Way analysis of variance (ANOVA) yielded F (variance ratio) = 0.5222, P = 0.8111, which indicated no statistically significant differences between hospitals, DREEM domains, or overall DREEM scores (Table 3 ). Table 3 ANOVA analysis between different hospitals for the differing DREEM domains Percentage of maximum score for each DREEM component for each hospital Birmingham Women's Good Hope Birmingham Heartlands Walsall City Wolver-hampton Shrewsbury Wordsley Learning 70% 69% 71% 71% 73% 75% 72% 73% Teachers 72% 68% 74% 78% 64% 74% 75% 76% Academic 62% 57% 59% 62% 61% 64% 58% 64% Atmosphere 70% 70% 68% 71% 73% 70% 73% 73% Social 69% 65% 70% 68% 63% 67% 68% 69% One-way analysis of variance (ANOVA) yielded F (variance ratio) = 0.5222, P = 0.8111. Discussion We have used the Dundee Ready Education Environment Measure (DREEM) in 'diagnosing' the educational environment of eight different teaching centres and making comparative analysis between these centres. The overall mean DREEM score was 139/200, or expressed as a percentage, 70% (95% CL 68–71%). The educational learning environment did not vary between centres. The two lowest scoring contributory domains, academic self-perception (61%) and social self-perceptions (68%), were not statistically significantly different from the other three DREEM domains or overall mean DREEM score. This study has benefited by using an established educational measure and obtaining a 100% response rate. No students had been previously taught at the principal teaching hospital as this was solely used for Obstetrics and Gynaecology teaching. However, some of the students (surveyed to be 16/206, 8%) had previously attended the other seven teaching hospital centres due to prior clinical teaching attachments. Thus, previous experiences may have biased the teaching assessment completed by some students. Furthermore, the DREEM questions are of such a nature that it is likely that the environment of the entire curriculum was being assessed. However, by performing the DREEM survey immediately at the end of the obstetrics and gynaecology module, and emphasising reporting only the last eight weeks experience, we believe this maximised the chance that the DREEM measure assessed only the recent hospital teaching site and minimised any recall bias. Other groups [ 7 ] have highlighted the potential flaws in using means and parametric statistical tests on ordinal data from Likert scales. As there is no firmly established consensus, we adopted to use the Student's T test and ANOVA calculation to fulfil best statistical methodology. The DREEM domains are unlikely to be independent variables, and may be less of an environment test but more of a measure of the overall motivation and learning attitude of the individual. The Course Valuing Inventory (CVI) score is made up of five domains: worthiness of learning experience, emotional awareness, personal development, cognitive enhancement and task drive. A recent study of first year medical students showed a correlation between higher Course Valuing Inventory (CVI) scores, female gender, stronger self-confidence as a learner, greater motivation to learn and higher DREEM scores [ 8 ]. There is no accepted agreement on what is an acceptable DREEM inventory score from published literature. Nevertheless, our DREEM score of 139/200 was higher than other reports. A study of final year medical students in Trinidad reported an overall mean DREEM of 109.9/200 [ 5 ]. A larger scale study, involving students from both final and earlier undergraduate training years, showed a DREEM score of 118/200 in a Nigerian medical school, and 130/200 in a Nepalese medical school [ 9 ]. Our higher score is reassuring, and is perhaps an indicator of better hospital teaching environment, the positive value of using a comprehensive course handbook, and the encouragement of formative self-assessment as guided by the course handbook and web-based package. The non-significant differences between the DREEM domains and between hospitals were significant findings. This was conveyed to our tutors based at the various teaching centres as a positive and encouraging result. In practical terms, this meant that regardless of hospital capacity or student group size, their education delivery and environment was no different to other centres in the student's curriculum. The DREEM inventory may thus be a useful tool for educators to ensure and maintain high quality educational environments and uniformity in educational delivery despite students being placed at different teaching centres. Competing interests The author(s) declare they have no competing interests. Authors' contributions RV and ET carried out the statisitcal analysis, data interpretation, and drafted the manuscript. JKG conceived and coordinated the study, acquired the results, and made revisions of the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553968.xml
514612	Investigating the utility of combining Φ29 whole genome amplification and highly multiplexed single nucleotide polymorphism BeadArray™ genotyping	Background Sustainable DNA resources and reliable high-throughput genotyping methods are required for large-scale, long-term genetic association studies. In the genetic dissection of common disease it is now recognised that thousands of samples and hundreds of thousands of markers, mostly single nucleotide polymorphisms (SNPs), will have to be analysed. In order to achieve these aims, both an ability to boost quantities of archived DNA and to genotype at low costs are highly desirable. We have investigated Φ29 polymerase Multiple Displacement Amplification (MDA)-generated DNA product (MDA product), in combination with highly multiplexed BeadArray™ genotyping technology. As part of a large-scale BeadArray genotyping experiment we made a direct comparison of genotyping data generated from MDA product with that from genomic DNA (gDNA) templates. Results Eighty-six MDA product and the corresponding 86 gDNA samples were genotyped at 345 SNPs and a concordance rate of 98.8% was achieved. The BeadArray sample exclusion rate, blind to sample type, was 10.5% for MDA product compared to 5.8% for gDNA. Conclusions We conclude that the BeadArray technology successfully produces high quality genotyping data from MDA product. The combination of these technologies improves the feasibility and efficiency of mapping common disease susceptibility genes despite limited stocks of gDNA samples.	Background In order to locate the disease variants involved in complex common disease it is now generally accepted that very large sample numbers will be required [ 1 - 4 ]. Not only do the sample collections need to provide high quality gDNA, for the purpose of accurate genotyping, they also need to be sustainable. If, for example, one million SNPs were to be genotyped in a whole genome association scan and only 1 ng were required per SNP genotype, 1 mg of DNA would be required from each clinical sample. Given that the gDNA yield from a typical blood sample of 8 ml is approximately 200 μg, and that the typical yield from a mouth-swab is just 10 μg, there is clearly a short-fall in available quantities unless other means are employed to amplify the DNA resource. Moreover, many existing and effectively irreplaceable DNA sample collections, which have been used in previous studies and are now depleted, may consist of only nanogram quantities of gDNA. At present, the gold standard method for generating gDNA from whole blood samples is through the process of immortalisation by transformation of the peripheral blood lymphocytes with Epstein-Barr Virus (EBV) [ 5 ]. Although this method of transfecting EBV creates an unlimited resource of gDNA, the procedure is costly, lengthy and not applicable to existing collections for which the gDNA has already been extracted. If there was a reliable method to enzymatically amplify the whole genome from nanogram-levels of gDNA and directly from clinical samples to microgram amounts then this would enable the use of archived gDNA in future studies, as well as providing an accelerated route to full use of newly collected clinical samples for high-throughput genotyping. Molecular Staging, Inc. (MSI) (New Haven, CT, USA) have developed a method for whole genome amplification by Φ29 polymerase Multiple Displacement Amplification (MDA). It has been reported by the company that this method can reliably amplify the whole genome from gDNA, whole blood and other clinical samples [ 6 - 8 ]. Each DNA sample should give similar yields of product in all reactions with little dependency on the quantity of starting template [ 6 , 7 ]. Moreover, the MDA reaction should give complete coverage of the genome with little regional bias [ 6 ], which is critical when the product is to be used for high-throughput SNP genotyping. We set up a series of experiments with MSI in order to validate their claims that MDA product from gDNA is a viable alternative template to un-manipulated gDNA in SNP genotyping. Recent studies have been conducted using MDA product from Amersham [ 9 , 10 ], and report the high level of accuracy achieved when these products are genotyped using TaqMan or multiplex, four-colour fluorescent minisequencing with six and 45 SNPs, respectively. However, without DNA resource limitations, a genotyping bottleneck exists mostly as a result of time- and assay set-up costs and hence, in order to achieve large-scale genotyping, highly multiplexed assays are required. In such multiplexed assays, there is greater potential for erosion of genotyping quality, due to reduced substrate integrity. The validation of the use of amplified DNA resources with such highly multiplexed methods is, therefore, essential. The BeadArray genotyping platform of Illumina™ Inc. (San Diego, CA, USA) offers a high-throughput, highly multiplexed and highly automated genotyping service facility [ 11 ]. The BeadArray platform is highly miniaturised, using fibre optic bundles as a substrate for a high-density microarray [ 12 ]. It is the combination of this miniaturisation with an ability to multiplex up to 1,536 SNP assays [ 11 ] that makes BeadArray an attractive potential solution to the genotyping bottleneck. A recent study by Barker and colleagues, with 2,320 SNPs and five samples, found 99.86% concordance between MSI MDA product and gDNA [ 13 ]. However, since only five samples were studied it was not possible to evaluate accurately the efficacy of BeadArray on MDA product template, including estimation of sample exclusion and failure rates. In the present report we have, therefore, studied 86 MDA product samples and 384 SNPs using BeadArray, allowing comparison with the single-plex methods TaqMan ® (Assays-by-Design SM , Applied Biosystems, Foster City, CA, USA) and Invader ® (Third Wave Technologies, Inc., Madison, WI, USA) with gDNA. Results MDA yield We selected and sent to MSI 20 ng of 88 gDNA samples for amplification, from which an average of 200 μg of MDA product was yielded in 100 μl reactions. The yields ranged from 85 μg to 280 μg, with 61% of samples yielding between 100 μg and 200 μg. The HLA-DRB1 genotype of each MDA sample was entirely concordant with the corresponding gDNA template, verifying the identity of each MDA sample and ruling out the possibility of contamination. When 100 ng of 448 gDNA samples were amplified using reagents supplied by MSI in kit form and the amplification carried out in-house, an average of 155 μg of MDA product was yielded in 150 μl reactions. The yields ranged from 31 μg to 260 μg, with 80% of samples yielding between 100 μg and 200 μg. Compatibility of MDA product with TaqMan and Invader Of 88 MDA products and their corresponding gDNAs tested at 95 SNPs using the TaqMan method of genotyping there were no samples that consistently failed to produce any data. This confirmed that, for all samples, amplification had been sufficiently successful for the TaqMan chemistry to perform at most SNPs. Genotype concordance rates between MDA product and gDNA and genotype failure rates are given in Table 1 . These results demonstrate that the use of MDA product as a template for the TaqMan assay produces accurate data comparable to that from gDNA. We observed that, for the majority of TaqMan assays, the clustering of data points was less distinct when MDA product was used as a template, compared to gDNA. Example data from an assay in which deterioration in clustering was observed are shown in Figure 1 . For a single SNP of the 95 tested, the insulin gene ( INS ) -23 Hph I (rs689), an allelic bias was observed in the MDA process, which resulted in the merging of the heterozygote cluster with the homozygote cluster of the major allele, making the correct assignments of genotype impossible, shown in Figure 2 [see additional file 1 ]. MDA product used as a template for the Invader genotyping method at this SNP produced similarly un-useable data. Using gDNA template at this SNP, however, both the TaqMan and Invader methods produced acceptable results, shown in Figure 2 [see additional file 1 ], indicating that the allelic bias was occurring at the MDA stage and not during subsequent genotyping. Interestingly, allelic bias has previously been observed at two other SNPs at INS in PCR reactions designed for the Pyrosequencing method (Pyrosequencing AB, Uppsala, Sweden) [ 14 ]. These results may indicate that INS may be situated in a sequence region that is predisposed to such allelic bias and the INS variable number of tandem repeats polymorphism, only 580 bp 5' to the -23 Hph I SNP, is a candidate for such an effect. For 13 additional SNPs for which Invader genotyping was performed, comparison of genotypes generated from MDA product with those from gDNA are shown in Table 1 . Two SNPs were genotyped by TaqMan on the 448 samples amplified using reagents supplied by MSI in kit form (and the amplification carried out in-house), and on the corresponding gDNAs. The gDNA samples used for amplification had been extracted from whole blood. Genotype concordance rates between MDA product and gDNA and genotype failure rates are given in Table 1 . Validation of BeadArray genotyping technology with gDNA template We commissioned Illumina to conduct a large-scale project using BeadArray genotyping technology involving 3,036 samples (2,950 gDNA samples and 86 MDA products) and 384 SNPs i.e. >1.1 million genotypes. In the first instance, 757 SNP sequences were sent to Illumina for in silico assay design. These SNPs were selected for their relevance to a range of ongoing projects in our laboratory, located at genes of strong functional candidacy and within regions of linkage to type 1 diabetes e.g. the putative IDDM10 locus on chromosome 10p14-11. All SNPs were validated, having been identified either from empirically confirmed SNPs in dbSNP or from our own re-sequencing efforts. Based on ranking from the in silico design criteria [ 15 ], 404 SNPs from 757 (53%) were suggested as most suitable for assay development, from which 384 were chosen. Thirty-nine of these failed to be converted into a viable assay (10.2%), leaving a total of 345 working assays. As well as excluding SNPs that fail to produce robust genotypes, the Illumina protocol excludes samples that do not consistently perform. Of the total number of samples 10.5% were excluded and as a consequence very few data points were missing from the data set, resulting in an apparently low genotype failure rate (Table 2 ). Within the 2,781 successfully genotyped gDNA sample set, 26 were duplicate samples. Of these 52 samples at 345 SNPs, the genotypes of 23 duplicates did not match each other and 19 data points were missing, giving a discordance rate (error rate) of 0.26% (23 of 8,951 data points). As our samples were family-based, a quality control check of misinheritance rates was possible using PedCheck [ 16 ]. Of the 345 SNPs, 20 displayed ≥ 10 misinheritances in the 742 families genotyped. For ten of these SNPs, TaqMan genotyping was attempted in the same samples in order to verify the results. It was possible to design TaqMan assays to only seven of the ten SNPs and, of these, only three produced interpretable data. At these three SNPs the numbers of misinheritances were 41, four and eight, respectively, by TaqMan, compared to 17, 22 and 14, respectively, by BeadArray. The number of SNPs with <10 or <5 misinheritances from the BeadArray experiment is shown in Table 3 , along with our previous year's TaqMan results considering SNPs with allele frequencies >1%. The poor performance of both Illumina and TaqMan at the ten SNPs compared in detail, as well as the lab misinheritance rate for TaqMan (Table 3 ), indicates that the high misinheritance rates observed for some SNPs in the Illumina experiment is not a technology-specific failing. Within the panel of 384 SNPs attempted by Illumina, 17 were controls for which we had already produced genotyping data by either TaqMan or Invader methods, enabling an evaluation of the BeadArray data for concordance. Two of these 17 control SNPs failed to be converted to a working assay, giving 15 SNPs and a maximum of 2,503 samples that were genotyped in common. Excluding failed duplicates noted above, comparison of BeadArray genotypes with existing data revealed a concordance rate of 99.6% (129 discordant in 34,219 comparisons), indicating the compatibility of the non-excluded gDNA samples with BeadArray, and the quality of existing data. Of the 15 control SNPs, 11 had been genotyped using TaqMan and four using Invader. The concordance rates for each platform were 99.7% using TaqMan (104 discordant in 25,203 comparisons) and 99.6% using Invader (25 discordant in 9,016 comparisons) when compared with BeadArray data, showing no significant difference between the two platforms. Compatibility of BeadArray with MDA product Within the BeadArray experiment described above were 86 MDA products and their corresponding gDNA samples. These data were directly compared for sample failure rate and genotype failure rate as shown in Table 2 . BeadArray genotype concordance rate between MDA product and gDNA are given in Table 1 . These results provide evidence for the compatibility of the non-excluded MDA products with BeadArray technology. Evaluation of the Illumina's quality scores revealed no significant difference between the MDA and gDNA samples for any SNP ( t -test P- value >0.05 for every SNP). Discussion In this study we have evaluated the Φ29 polymerase MDA whole genome amplification method from MSI by assessing the compatibility of its product with the established TaqMan and Invader genotyping chemistries and with the highly multiplexed BeadArray genotyping platform. We have also evaluated Illumina's BeadArray genotyping platform for a large-scale experiment using gDNA. At 95 SNPs, comparison of TaqMan genotypes generated from MDA product and gDNA templates revealed a very good concordance rate but a higher failure rate for MDA product compared to gDNA. This would need be estimated in a sample size larger than the current n = 88 in order to be confirmed. This result is comparable to the smaller study by Tranah et al . [ 9 ], in which six SNPs were genotyped by TaqMan on 172 samples, resulting in 100% concordance of pre- and post-MDA DNA. In the present study, the MDA product genotypes were slightly more difficult to assign, owing to more dispersed clusters. This was not observed by Lovmar et al . with fluorescent minisequencing on Amersham MDA products compared to gDNA [ 10 ]. One marker in our study, which may be unusually prone to allelic bias, was impossible to score using MDA product but was acceptable when using gDNA as a template ( INS -23 Hph I, rs689). Compared to the yields indicated in Dean et al. [ 7 ], our average yield from in-house amplification using the reagents in kit form were in the order of five- to six-fold higher. This was probably due to differences in the two protocols: for example, our protocol used an increased reaction volume compared to the protocol used in Dean et al. [ 7 ]. Furthermore, the Dean et al. [ 7 ] protocol omitted the denaturation step, which is now standard practice. One other potential explanation for this variation is possible differences between laboratories in the quantitation of DNA using PicoGreen, the application of which requires a standard reference data set. We cannot at present fully resolve the differences in yields between studies but we can conclude that very large amounts of DNA are synthesized during the Φ29 reaction and that this is an excellent template for genotyping. MDA product should, therefore, be quantified and its concentration on completion of the MDA reaction not assumed to be consistent. Genotype failure rate, concordance rates with gDNA and the nature of genotype clustering showed similar patterns to service-generated MDA. However, a larger number of SNP markers would need to be genotyped on the MDA product using purchased kit reagents in order to verify these figures for in-house amplifications. In the evaluation of MDA product in conjunction with BeadArray technology, the high concordance rate between genotypes obtained from MDA product and gDNA templates is encouraging. A concordance rate of 99.86% has been reported by Barker et al . using 2,320 SNPs and five samples [ 13 ]. However, as our study used 86 samples, we were able to observe differences in genotype failure rate between the different templates, not noted in the previous study [ 11 ]. As with the TaqMan evaluation, BeadArray had a higher genotype failure rate for MDA product compared to gDNA (0.2% for MDA versus 0.06% for gDNA). We did not find any evidence for allele drop-out with MDA compared to gDNA. BeadArray genotyping excluded more MDA samples than gDNA samples (10.5% for MDA versus 5.7% for gDNA) indicating that gDNA is a superior genotyping template for BeadArray technology. This 2-fold exclusion rate for MDA is consistent with the approximately 2 to 3-fold genotype failure rate of MDA typically observed with TaqMan and Invader, compared to gDNA (unpublished data). The performance of MDA product is continuously being monitored in our laboratory. In a study blinded to all genotypers and database administrators, 288 family-based gDNA samples (prepared by the salting out method), were replaced with MDA product and left in continual use in our genotyping pipeline for 12 months. The change went undetected by all users. The failure rate for MDA was 3.34% for 15,921 genotypes, compared to 2.39% for 19,272 gDNA genotypes. Therefore, this improvement in the MDA performance for TaqMan is likely to be applicable to BeadArray, which improves the feasibility of mapping susceptibility loci in complex traits. When using a highly multiplexed, highly automated genotyping platform, slight reductions in the quality of template material are likely to have a greater adverse effect on data than in scenarios in which markers are assessed individually and manual scoring is undertaken. Our results indicate that MDA is an adequate solution for the vast majority of SNP markers, even in this highly multiplexed allelic assay platform. It is noted that 5.8% of markers that passed the Illumina acceptable scoring threshold were in fact showing high misinheritance rates in our family samples. This problem was at the same magnitude as TaqMan for individually genotyped markers. This highlights the importance of checking potential positive results with a second genotyping technology. MDA should allow the continuation of genetic analysis on archived DNA in researchers' freezers worldwide, providing the very necessary increases in sample sizes so urgently required [ 1 , 2 , 17 ]. Conclusions The combination of BeadArray high throughput, multiplex genotyping and amplified DNA (MDA product) successfully produced high quality genotype data thereby improving the feasibility and efficiency of mapping common disease susceptibility genes despite limited stocks of gDNA samples. Methods MDA product preparation For both the validation experiments (MDA product as a template for TaqMan and Invader genotyping) and for the combined experiment (MDA product as a template for BeadArray genotyping) the same MDA samples were tested. We sent to MSI 20 ng (5 μl at 4 ng/μl) of 88 gDNA samples for amplification, which was performed as a service according to the protocol for human gDNA with the omission of the denaturing step [ 7 ]. These gDNA samples had been extracted from cell pellets of EBV derived cell lines using a standard chloroform protocol that produces very high quality and stable gDNA [ 18 ]. The MDA product returned to us was quantified using PicoGreen dsDNA quantitation reagent (Molecular Probes Europe B.V., Leiden, the Netherlands). In order to verify the identity of each MDA-produced sample, genotyping was performed at HLA-DRB1 and comparison made with data generated from the corresponding gDNA. HLA-DRB1 genotyping was performed using the Dynal Auto RELI™ SSO HLA-DRB Test system (Dynal ® Biotech, Wirrel, UK) for each gDNA sample and their MDA products. Although, these samples were amplified by MSI as a service, the reagents are also available from MSI in a kit form for amplification in-house. Following the amplification by MSI we have amplified 448 DNA samples by using the reagents in kit form and 100 ng (25 μl at 4 ng/μl) gDNA template in 100 μl reactions. In the interim, one major change to the MDA protocol had taken place, the inclusion of a denaturation of the DNA template prior to amplification. Previously no denaturation step took place. Two TaqMan markers were tested on these 448 samples and the genotype failure rate calculated. Evaluation of MDA product as a template for TaqMan and Invader genotyping SNP TaqMan assays were carried out for allelic discrimination, 8 ng of DNA (2 μl at 4 ng/μl) used in a 5 μl total reaction volume. TaqMan genotypes from the 88 MDA samples, described above, were compared with TaqMan genotypes generated from their corresponding gDNAs, at 95 SNPs, with a broad range of allele frequencies. The Invader method was used to genotype 13 additional SNPs on the same samples. Comparison was made between data generated from both templates by the measurement of genotype failure and genotype concordance rates. Evaluation of BeadArray genotyping technology and its compatibility with MDA product Of the 384 SNPs selected for genotyping 3,036 samples, 17 were control SNPs for which we had existing genotype data, generated by either TaqMan or Invader methods, with which comparison of genotype failure and genotype concordance rates were made. These 384 SNPs covered a broad range of allele frequencies. Incorporated into this experiment was an assessment of suitability and compatibility of the BeadArray genotyping method with MDA product. This involved 86 of the 88 amplified samples, described above, for which genotyping was attempted at all 384 SNPS. Concordance between genotypes generated from MDA product and gDNA templates, together with the genotype and sample failure rates of each template type were measured. In our laboratory we store genotyping data in a MySQL database on a Sun server. The volume of data expected from Illumina was the equivalent of 6 months' in-house genotyping. We separated phenotypic and pedigree information, which is associated with a sample, from genotype data, which is associated with a DNA plate and well position, with a link table to relate the two. Sample aliases are also supported, so that no recoding of identifiers is required, either to export or import Illumina data [ 19 ]. Authors' contributions RP participated in the design of the study, performed in-house genotyping experiments, and assisted in preparing the manuscript. HER prepared DNA and MDA amplified samples. BJB participated in the design of the study and assisted in the preparation of the manuscript. SN prepared DNA and MDA amplified samples. DS performed in-house genotyping experiments. MS prepared DNA and MDA amplified samples. RCJT participated in the design of the study and assisted in the preparation of the manuscript. AS participated in the design of the study. ACL and LJS performed bioinformatics. NMW managed the data and participated in its interpretation. JAT participated in the design of the study and assisted in the preparation of the manuscript. All authors read and approved the final version of the manuscript. Supplementary Material Additional File 1 Figure 2 TaqMan and Invader fluorescence data plotted for the INS -23 HphI SNP. Both the MDA product plots could not be scored. All plots represent the same individual samples with gDNA plots containing 8 additional samples. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514612.xml
547909	A7DB: a relational database for mutational, physiological and pharmacological data related to the α7 nicotinic acetylcholine receptor	Background Nicotinic acetylcholine receptors (nAChRs) are pentameric proteins that are important drug targets for a variety of diseases including Alzheimer's, schizophrenia and various forms of epilepsy. One of the most intensively studied nAChR subunits in recent years has been α7. This subunit can form functional homomeric pentamers (α7) 5 , which can make interpretation of physiological and structural data much simpler. The growing amount of structural, pharmacological and physiological data for these receptors indicates the need for a dedicated and accurate database to provide a means to access this information in a coherent manner. Description A7DB is a new relational database of manually curated experimental physiological data associated with the α7 nAChR. It aims to store as much of the pharmacology, physiology and structural data pertaining to the α7 nAChR. The data is accessed via web interface that allows a user to search the data in multiple ways: 1) a simple text query 2) an incremental query builder 3) an interactive query builder and 4) a file-based uploadable query. It currently holds more than 460 separately reported experiments on over 85 mutations. Conclusions A7DB will be a useful tool to molecular biologists and bioinformaticians not only working on the α7 receptor family of proteins but also in the more general context of nicotinic receptor modelling. Furthermore it sets a precedent for expansion with the inclusion of all nicotinic receptor families and eventually all cys-loop receptor families.	Background Nicotinic acetylcholine receptors (nAChRs) are the most studied members of the cys-loop family of ligand-gated ion channels (LGICs) which also contains γ-aminobutyric acid (GABA) receptors, glycine receptors and 5-HT 3 receptors [ 1 ]. Distinct subtypes of nAChRs mediate, for example, fast synaptic transmission in the brain and at neuromuscular junctions [ 2 ]. All are believed to be pentameric assemblies of various combinations of different subunits (α, β, γ/ε, δ), some of which exist as multiple isoforms [ 3 ]. nAChRs are important targets for novel analgesics as well as new drugs being devised for Alzheimer's disease and schizophrenia [ 4 , 5 ]. Mutations in nAChRs are associated with certain forms of epilepsy [ 6 , 7 ] and several congenital myasthenias [ 8 ]. There is a substantial and growing body of physiological, pharmacological, genomic structural and modeling data on receptors formed from these subunits. The volume and diversity of these data present severe challenges for their efficient storage and interpretation. Here we describe a relational database, initially relating to the α7 subunit, whose aim is to provide an easy to use, extensible web-based interface to access functional data and relate it back, wherever possible, to structural data. Construction and content The database is currently limited to nAChR α7 subunits. This makes the initial task of populating the database manageable. To allow 3 rd level normalization, the principal information prototype stored in the database is referred to as an "experiment event". An experiment event is a collection of simultaneous measurements and their associated measurement conditions. For example, two drugs tested against the wild type α7 and a mutant under the same experimental conditions would be described by 4 (2 drugs × 2 sequences) experiment events. If the identical measurements were repeated in, say, calcium-free saline, this would yield 8 experiment events. Using this prototype a unique set of experimental conditions is associated with a single set of data. Therefore, the database tables (see Figure 1 ) have been designed around the central experiment table [see Additional file 1 ]. Quite often, several real experiments are performed within any one publication. We wanted to capture all of that data. Thus, all the data for one experiment that was reported have been stored (non-redundantly) as a separate entity. The database is managed by the MySQL relational database management system. PHP scripts query the MySQL system and transform the data into HTML pages for serving to the client. The database currently aims to store molecular information rather than whole-organism genetic information which can be sourced from elsewhere, see for example [ 9 , 10 ]. A typical set of experiments reported in one paper might concern the properties of one mutation at a particular position and its associated changes in physiology and/or pharmacology. All of this data is held in the database. By use of appropriate alignments (either stored within the server or uploaded) the positions of these data can be highlighted on a homology model based on the acetylcholine binding protein AChBP [ 11 ]. The model is viewable by either using the Chime plugin, or by setting the browser to use Rasmol as a helper application. User instructions for configuring different browsers are provided on the server, as well as the list of combinations of operating system and browser that we have tested to date. For any database, data integrity will ultimately determine its usefulness. The initial population of the database has been carried out by the authors, but it is desirable and practical in the long term to have a procedure whereby any laboratory can submit data. To help ensure reliable deposition procedure, each depositor will have a unique identifier (thus making the data accountable). A depositor will submit their data which is then tabulated and presented as a form which the depositor is then asked to confirm as being correct. Only when this confirmation is received is the data committed to the database. Although this will reduce some entry-based mistakes, the database will nonetheless need to be curated to maintain integrity. The database does not attempt to duplicate the functionality of other databases. For example we store a local alignment of α7 subunits taken directly from pfam [ 12 ], but we do not perform any alignments ourselves. Such tools are already available on numerous web-servers and in any case the user may wish to upload their own alignment. Wherever possible, we link to a well-established resource, as is the practice in for example, SWISS-PROT [ 13 ]. Utility and discussion Access to the database is through various routes. The first route is by a very simple search string (e.g. citation = Smith*). The user is then presented with a page informing him of the number of hits and asking what information should be displayed from those hits. The second is a simple incremental search in which the user applies criteria sequentially until he chooses to examine the results. As the query is built, the number of hits returned by the query in its current state is shown. The third search method is also incremental but via the use of a set of tabbed pages which divide the information into intuitive categories such as pharmacology and physiology. Additionally, the user can make choices about alignments used and even upload their own. As the query is built, summary information of its status is also displayed. The fourth method to access the data is provided by a 'fast lane' route which provides more experienced users with a quicker and more direct route to the data of interest. The user formulates and constructs a query offline and then uploads it as a simple ASCII file. A sample query form and details of the format are presented in the supplementary information. The result of a query is presented in tabular form. Only the information requested is actually presented with the option for further or refined searches. In addition to this, the results of any position matches can be displayed via the homologous 3-dimensional structure of the AChBP. This is automatically available if the user selects the alignments available from the server. If the user uploads their own alignment the model will only be produced if AChBP was included. The server streams a Rasmol script or spawns a Chime window depending on browser configuration. In addition to these query routes, the user is also able to browse the contents of the database. The current database has been designed such that it could easily be extended to include other nAChR subunits and then on to other members of the cys-loop ligand-gated ion channel family. We are currently pursuing this aspect. Such expansion will allow trends that link structure to function in this receptor family to be more readily identified. The database is very complementary to some existing resources, in particular the LGICdb [ 14 , 15 ]. This database contains sequence, phylogenetic data and sets of coordinates, but does not attempt to store all aspects of individual experiments as we report here. However, this could be used in conjunction with our database to for example explore equivalent positions in related receptors using sequence and phylogenetic information stored therein. Our approach is somewhat similar to that employed by the ProTherm database [ 16 ] which aims to store thermodynamic data for mutant and wildtype proteins [ 17 ]. Although currently there are no entries common to both databases, it is envisaged that thermodynamic data reported for either the α7 nAChR receptor or related channels will appear in the ProTherm database and can thus be used in conjunction with the A7DB to help illustrate for example how loss of function might be related to protein stability. Furthermore, the similarity in design of these databases should make cross-interrogation possible in the future. We should state that the main difference between A7DB and ProTherm is that instead of thermodynamic data we store pharmacological and physiological data and in that sense it is closer to the voltage-gated potassium channel database [ 18 , 19 ]. Finally, the A7DB is also complementary to the Protein Mutant Database [ 20 ] where mutations across many proteins are stored but the searching and data tabulation is primarily text-based. This might be particularly useful if another mutation had been reported in a different protein that bound a similar ligand, for example in acetylcholine also binds to acetylcholinesterase, which we do not store data for. Conclusions We have shown here how the collation and careful storage of experimental data pertaining to one sub-family (the α7 nAChR sub-family) of the ligand-gated ion channels can be assembled into a useful resource. However, the real power of the database will be when it is combined with machine-learning technologies to explore complex relationships between sequence, structure and function [ 21 ]. We believe that this resource will grow in usefulness as the amount of data increases. For example, during the construction of this database, atomic coordinates were released for the transmembrane region of the nAChR from Torpedo marmorata [ 22 ]. Its development is particularly timely given these recent structural developments which allow three-dimensional information to be correlated with function derived from an ever-increasing body of experimental data. Availability and requirements The database is freely available at: Authors' contributions The project was conceived and technical aspects managed by PCB. The coding and design was done by SB. MSPS and DBS provided critical evaluation of the design and construction process. DBS, LP, AKJ and LB populated the database. All authors have approved the manuscript. Supplementary Material Additional File 1 Supplementary information. This doc file provides a complete description of the table entries and also provides a description of the uploadable query file. Further information about the online help is also provided. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547909.xml
535347	Review on "Atkins Diabetes Revolution: The Groundbreaking Approach to Preventing and Controlling Type 2 Diabetes" by Mary C. Vernon and Jacqueline A. Eberstein	null	Before beginning the review of this book, we had no particular opinion about the role of low carbohydrate diets in diabetes. In order to write a fair and unbiased review, we have done a rather extensive search on the subject. One of the most disturbing findings of our search is the amount of hostility towards low carbohydrate diets that is on the web and in the scientific literature. We found several sites that present no scientific arguments but are, rather, full of ad hominem attacks. This was particularly disturbing in that we are in the midst of a growing epidemic of obesity and diabetes with very alarming figures and projections from all over the world. Any intervention that has the potential for helping curb this dangerous epidemic which claims thousands of lives every day should be looked at with a great deal of objectivity. The low carbohydrate approach, in fact, is not new and was used in England more than a century ago, made popular by William Harvey [ 1 ], an ENT surgeon. He prescribed a low carbohydrate diet for William Banting, an obese carpenter who had been having a great difficulty losing weight. Banting was able to lose weight and as a service, he published in 1863 a small booklet entitled Letter on Corpulence Addressed to the Public [ 2 ], the first book to be published on obesity and one which popularized low carbohydrate diets. He has been called "Father of low carbohydrate diets" and was honored by his name being included in the dictionary as the verb "to bant" meaning "to diet". The low carbohydrate diet also been called a "Harvey-Banting diet" after the names of these pioneer. Since then, it has been in and out of fashion with different versions and names but with the same underlying concept, most recently popularized by the late Dr. Robert C. Atkins. The Atkins Diabetes Revolution [ 3 ] plan is similar to the Atkins weight loss strategy: four levels of carbohydrate restriction are instituted. The induction phase restricts dieters to 20 g of carbohydrate. On the weight loss plan, this is recommended for about 2 weeks. In diabetes this is maintained until glycemic control is attained. In the latter stages, carbohydrates are added as long as weight loss or stability is maintained. For diabetes, carbohydrates are only reintroduced if glycemic control is acceptable. In the later phases, the Atkins Diabetes plan adds a Glycemic Ranking (AGR), derived from the glycemic index, glycemic load and net carbs. Preference is given to whole fruits and berries and juices and dried fruits are low on the list. As in weight loss, exercise is "mandatory." The Atkins Diabetes Revolution book is an attempt by the authors to present the low carbohydrate diet as a preventive and treatment strategy for patients with type 2 diabetes and those with the metabolic syndrome, who are at high risk for developing diabetes and cardiovascular disease. In doing so, the book, which is very well written, and which clearly presents illustrative cases, explains very complex metabolic concept in a very easy to read and understandable format. The first nine chapters explain the different concepts involved in glucose and lipid metabolism and the interplay of the various cardiovascular risk factors that culminate in cardiovascular disease the number one killer of Americans today. Definitions of metabolic syndrome, pre-diabetes, body mass index, waist to hip ratio, central obesity and their relationship to diabetes, heart attacks and strokes, are eloquently presented with a great deal of accuracy yet in a simple format. Most impressive were the case presentations, especially that of reactive hypoglycemia and carbohydrate craving. This response is associated with hyperinsulinemia in the pre-diabetic phase and sometimes puzzles clinicians unless they know to look for it. The second section of the book is devoted to an in-depth discussion of the various macro and micronutrients and their role in diabetes and obesity. Concepts such as the glycemic index and glycemic load are very well illustrated. The last section consists of meal plans and menus of low carbohydrate diet that the book is advocating. The concept of low carbohydrate diet and glycemic control certainly has a pathophysiological merit. First, dietary carbohydrates are the principal source for the initial rise of glucose in the diabetic populations, who generally have a defect in the first phase insulin secretion that is responsible for handling the glucose load [ 4 ]. There is mounting evidence that postprandial hyperglycemia is in itself a risk factor for cardiovascular disease in the diabetic patients [ 5 ]. This evidence comes from large, well-conducted, randomized controlled trials [ 5 , 6 ]. Furthermore, control of postprandial hyperglycemia has been shown to provide cardiovascular benefits, and contribute to the overall decrease of hemoglobin A 1c , something that has been clearly shown to reduce microvascular disease in both type 1 and type 2 diabetes [ 7 , 8 ]. Second, the initial blood glucose rise associated with high carbohydrate load, in the presence of absolute/relative insulin deficiency leads to significant rise in triglycerides and free fatty acids which perpetuate the cycle of insulin resistance [ 9 , 10 ]. So, from a metabolic stand point, low carbohydrate diet makes physiologic sense. However, in the science and practice of medicine, not everything that makes sense turns out to work the way it is supposed to. In looking at the low carbohydrate diet, we must examine the evidence from the studies that were conducted using such diets keeping in mind that weight loss by itself, is beneficial in terms of improving insulin sensitivity and correcting the abnormalities associated with the metabolic syndrome and insulin resistance [ 9 , 10 ]. Also, weight loss has much greater effect on the prevention of type 2 diabetes in pre-diabetic patients than pharmacological interventions [ 9 ]. This fact was well illustrated in the Diabetes Prevention Program, a large multicenter trial sponsored by the National Institute of Health, where pre-diabetic patients on diet and exercise program had a 58% reduction in the development of diabetes, compared to only 34% reduction with the use of metformin [ 11 ]. This landmark study had a population where women and minorities were very well represented [ 11 ]. The fact that weight loss was associated with reduction of type 2 diabetes in high risk populations was illustrated in several other studies including examples from Finland and from China, making it evident that weight loss works for a variety of ethnic populations [ 12 - 15 ]. In two recent randomized controlled trials published in the New England Journal of Medicine [ 16 , 17 ], the effects of low carbohydrate and low fat diets were compared in obese and diabetic patients. Both of these studies showed a substantial decrease of triglycerides in patients on low carbohydrate diet with simultaneous increase in high-density lipoprotein (HDL) over 6 month to 1 year period. The studies did not show a change in the low-density lipoprotein (LDL) values in the low carbohydrate group compared to their baseline, while those on traditional low fat diet had a reduction in LDL levels. Patients on low carbohydrate diet, however, had substantially significant weight loss, almost double that achieved with the traditional diet, in the first 3–6 months. At one year, there was no significant difference in weight loss between the two groups [ 16 - 18 ]. Although participants on the low carbohydrate diet initially tended to have higher rate of side effects such as nausea, muscle cramps and constipation, compliance with diet was similar in both groups. In fact, more participants adhered to the low carbohydrate diet. Although weight loss was similar after one year between groups, the effects on atherogenic dyslipidemia and glycemic control were still more favorable with a low-carbohydrate diet after adjustment for differences in weight loss. Despite the evidence from these randomized controlled trials, published in the prestigious New England Journal of Medicine, there is a significant amount of reluctance in the scientific community to acknowledge the beneficial effects of low carbohydrate diets. These studies, in fact, provide a striking example of this resistance. A commentary in the same issue of the New England Journal of Medicine [ 20 ] states that "In both studies, the reduction in serum triglyceride levels in subjects randomly assigned to the low-carbohydrate diet might have been anticipated as a result of their greater weight loss, although it is true that reduced carbohydrate intake is generally associated with reduced triglyceride levels" [ 20 ]. In this statement, despite the fact that low carbohydrate diet is known to reduce serum triglyceride, the authors suggest otherwise. In another statement, the authors of the commentary state that "the rise in HDL cholesterol in the subjects following the low-carbohydrate diet (a change observed only by Foster et al.) may reflect a change in HDL subfractions that occurs with increased intake of saturated fats, and this change has not been shown to be beneficial. Thus, caution is urged about over-interpretation of this observation as a beneficial result of a low-carbohydrate, high-fat diet" [ 20 ]. Again this statement illustrates the difficulty in acknowledging what a randomized controlled trial has shown. The authors suggest, without any evidence that the rise in HDL cholesterol might have been in the non-beneficial HDL subfraction. In other words, when low carbohydrate diet is shown to decrease triglycerides, a suggestion is made that it might be just secondary to weight loss and when this diet increases HDL, it is also suggested that it could be the non-beneficial HDL. Now, let us examine the evidence provided by the one year follow-up study on the same group of patients where the investigators conclude that "Although weight loss was similar between groups, the effects on atherogenic dyslipidemia and glycemic control were still more favorable with a low-carbohydrate diet after adjustment for differences in weight loss" [ 18 ]. This indicates that the statements made in the commentary [ 20 ], in an attempt to dismiss or downplay the beneficial effects of low carbohydrate diet were simply wrong. Furthermore, the statement made in the commentary regarding the HDL cholesterol, not only lacks objective evidence, but also contradicts the current findings that lowering insulin level by controlled carbohydrates shift HDL production to a much more desirable, lighter HDL 2 subfractions [ 21 , 22 ]. On the other hand, the American Diabetes Association, despite recommending the traditional low fat diet, has recently reduced the recommended carbohydrate contents in the diet, perhaps reflecting a trend towards a reduced carbohydrate diet to follow [ 19 ]. Returning to the Atkins book, despite the fact that the book is very well referenced, certain statements such as "high carbohydrate diet leads to diabetes" are not well substantiated, unless of course such a diet leads to weight gain, which it may. Furthermore, the book does not devote a sufficient amount of space discussing the side effects associated with dieting in general and low carbohydrate diet in particular. This is of concern, since it leaves the reader with the impression that the low carbohydrate diet or dieting, in general, has no negative consequences. Nonetheless, the amount of information the book provides in a simple, yet accurate format will benefit patients with diabetes and their families as well as those who are at risk for developing diabetes and the metabolic syndrome. If, after reading this book, the reader is able to identify that he or she is at risk for diabetes and the metabolic syndrome and takes action that could potentially save his or her life the book will be a valuable contribution. Atkins Diabetes Revolution has a list price of $25.95 and is available at Amazon.com and presumably other sites for half that price. Possibly, a shorter and still more affordable version of the book would be helpful for diabetic patients, their families and for the general reader, to help identify their risk for the disease. As clinicians, we would not be comfortable recommending any diet without first hand experience. The Atkins Diabetes Revolution , however, is sufficiently convincing to make us believe that some form of low carbohydrate intervention is worth investigating and should be considered by practitioners. The highly negative un-scientific response of critics, if anything, encourages us in this direction.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535347.xml
516039	Topical NSAIDs for chronic musculoskeletal pain: systematic review and meta-analysis	A previous systematic review reported that topical NSAIDs were effective in relieving pain in chronic conditions like osteoarthritis and tendinitis. More trials, a better understanding of trial quality and bias, and a reclassification of certain drugs necessitate a new review. Studies were identified by searching electronic databases, and writing to manufacturers. We identified randomised, double blind trials comparing topical NSAID with either placebo or another active treatment, in adults with chronic pain. The primary outcome was a reduction in pain of approximately 50% at two weeks, and secondary outcomes were local and systemic adverse events and adverse event-related withdrawals. Relative benefit and number-needed-to-treat (NNT), and relative harm and number-needed-to-harm (NNH) were calculated, and the effects of trial quality, validity and size, outcome reported, and condition treated, were examined by sensitivity analyses. Twelve new trials were added to 13 trials from a previous review. Fourteen double blind placebo-controlled trials had information from almost 1,500 patients. Topical NSAID was significantly better than placebo with relative benefit 1.9 (95% confidence interval 1.7 to 2.2), NNT 4.6 (95% confidence interval 3.8 to 5.9). Results were not affected by trial quality, validity or size, outcome reported, or condition treated. Three trials with 764 patients comparing a topical with an oral NSAID found no difference in efficacy. Local adverse events (6%), systemic adverse events (3%), or the numbers withdrawing due to an adverse event were the same for topical NSAID and placebo. Topical NSAIDs were effective and safe in treating chronic musculoskeletal conditions for two weeks. Larger and longer trials are necessary to fully elucidate the place of topical NSAIDs in clinical practice.	Background A systematic review of topical NSAIDs reported that they were effective for relieving pain in both acute and chronic conditions [ 1 ]. Number-needed-to-treat (NNT), the number of patients that need to be treated for one to benefit from a particular drug, who would not have benefited from placebo, was used to estimate efficacy. In chronic conditions, NNT for topical NSAIDs at two weeks was 3.1 (2.7 to 3.8). There are three reasons why an updated review of topical NSAIDs in chronic pain is needed. First, we have a better appreciation of factors that can introduce bias [ 2 - 4 ], and would not now accept trials that were not double blind, or were very small. Second, topical salicylate and benzydamine are no longer classed as topical NSAIDs [ 5 ]. Thirdly, there are now more trials. We believed that updating the review would improve efficacy estimates for topical NSAIDs, with a prior intent to determine efficacy for individual drugs. Methods Searching Relevant studies were sought regardless of publication language, type, date or status. Studies included in the previous review were considered for inclusion, and the Cochrane Library, MEDLINE and PreMedline, EMBASE and PubMed, were searched for relevant studies published since the last review, for the years 1996 to April 2003. The search strategy included "application: topical" together with "cream", "gel" etc, together with generic names of NSAIDs, and proprietary preparations of topical treatment in which the principal active ingredient was an NSAID [ 6 , 7 ] ( additional file 1 :search strategy). Reference lists of retrieved articles were also searched. We wrote to 20 pharmaceutical companies in the UK, 66 in continental Europe, and two in North America, known to manufacture topical NSAIDs, asking if they could supply papers. Selection We identified reports of randomised, double blind, active or placebo-controlled trials in which treatments were given to adult patients with moderate to severe chronic pain resulting from musculoskeletal or other painful disorders. We excluded treatments for mouth or eye diseases. At least ten patients had to be randomised to a treatment group and application of treatment had to be at least once daily. Outcomes closest to two weeks (but at least seven days) were extracted. Longer outcomes were also accepted when available. Quality and validity assessment Trial quality was assessed using a validated three-item scale with a maximum quality score of five [ 8 ]. Included studies had to score at least two points, one for randomisation and one for blinding. A sixteen-point scale was used to assess trial validity [ 9 ]. Quality and validity assessments were made independently by at least two reviewers and verified by one other reviewer. Disputes were settled by discussion between all reviewers. Outcomes We defined our own outcome of clinical success, representing approximately a 50% reduction in pain [ 1 ]. This was either the number of patients with a "good" or "excellent" global assessment of treatment, or "none" or "slight" pain on rest or movement (or comparable wording) measured on a categorical scale. A hierarchy of outcomes was used to extract efficacy information [ 1 ], shown below in order of preference: 1) number of patients with a 50% or more reduction in pain 2) patient reported global assessment of treatment 3) pain on movement 4) pain on rest or spontaneous pain 5) physician or investigator global assessment of treatment In addition, the number of patients showing undefined "improvement" was also accepted. All of these outcomes were grouped together as a "success", and categories 1–4 were used as preferred outcomes in the sensitivity analysis. Secondary outcomes were extracted from included papers that reported them. These were the number of patients (i) reporting one or more local adverse event (itching, stinging, rash), (ii) reporting one or more systemic adverse event (iii) withdrawing from trials due to adverse events. Quantitative data synthesis The number of patients randomised into each treatment group (intention to treat) was used in the efficacy analysis. Information was pooled for the number of patients in each trial approximating at least 50% pain relief, or similar measure, for both topical NSAID and control. These were used to calculate NNT with a 95% confidence interval (CI) [ 10 ]. Relative benefit and relative risk estimates with 95% CIs were calculated using the fixed effects model [ 11 ]. A statistically significant benefit of topical NSAID over control was assumed when the lower limit of the 95% confidence interval (CI) of the relative benefit was greater than one. A statistically significant benefit of control over active treatment was assumed when the upper limit of the 95% CI was less than one. Homogeneity of trials was assessed visually [ 12 - 14 ]. Number-needed-to-harm (NNH) and relative risk were calculated in the same way as for NNTs and relative benefit. All calculations were performed using Microsoft Excel X for the Macintosh and RevMan 4.2. In sensitivity analyses the z test was used [ 15 ]. QUOROM guidelines were followed [ 16 ]. Sensitivity analysis Our prior intention was to perform sensitivity analyses on pooled outcomes using the z test [ 15 ] for quality score (2 versus 3 or more), validity score (8 or less versus 9 or more), trial size (less than 40 patients per group versus more than 40 patients per group), reported outcome (higher versus lower preference), drug, and condition treated (knee osteoarthritis versus other musculoskeletal). At least three studies had to be available in each category before information was pooled. Results Study characteristics Ten out of the 20 UK companies, and two out of the 66 continental European companies replied to our request for studies. Only three companies supplied useful material, either published studies or bibliographies. None provided unpublished material. Searches identified 60 target papers, but 35 were excluded; 23 studies failed to meet the inclusion criteria and 12 had no useable data. Twenty-four of these 60 target papers were included in the previous review. We included 13 of those in this review, and excluded 11; seven were not double blind, two compared a salicylate with placebo or oral analgesics, one did not have daily application, and one had insufficient data ( additional file 2 : excluded studies, additional file 3 : QUOROM flow diagram). Twenty-five trials therefore met the selection criteria, 12 of which were additional trials. Fifteen trials had only placebo controls [ 17 - 31 ], seven only active controls [ 32 - 38 ], and three had both placebo and active controls [ 39 - 41 ]. Of the 10 active controlled trials, four compared a topical NSAID with a different topical NSAID, three compared a topical NSAID with a different oral NSAID, and one each compared a topical NSAID with a homeopathic gel, a topical rubefacient, and topical trinitroglycerin (GTN). Details of all included studies with outcomes and quality and validity scores are in additional files 4 (Outcome details of placebo-controlled trials) and additional files 5 (Outcome details of active-controlled trials). Patients were generally over 40 years old, predominantly with musculoskeletal disorders, and with baseline pain of moderate to severe intensity. Fourteen studies examined general musculoskeletal conditions, and eleven examined osteoarthritis (9 studies of the knee, one of finger joints, and one of mixed sites). Five studies in osteoarthritis specified use of a standard scale (ACR, Kellgren and Lawrence, ISK) to assess the severity of disease, four specified that the disease was radiologically confirmed, one specified that patients had "well documented mild osteoarthritis", and one made no statement. Quality scores were high, with 16/18 placebo controlled and 9/10 active controlled trials scoring 3 or more points out of a maximum of 5. Validity scores were also high, with 14/18 placebo controlled and 8/10 active controlled trials scoring 9 or more out of a maximum of 16 ( additional files 4 and 5 ). Placebo controlled trials Dichotomous information was available to pool from 14 placebo controlled trials for efficacy, from 16 placebo controlled trials for local adverse events, 17 placebo controlled trials for systemic adverse events, and from 11 placebo controlled trials for adverse event related withdrawals. Efficacy Fourteen trials (1,502 patients) provided data on efficacy. Topical NSAIDs were significantly better than placebo (Table 1 ). The mean placebo response rate was 26% ranging from 7% to 78%. The mean treatment response rate was 48% ranging from 2% to 90% (Figure 1 ). The NNT was 4.6 (95% CI 3.8 to 5.9) for one patient to experience improvement in chronic musculoskeletal pain at two weeks with topical NSAIDs, compared with placebo. Sensitivity analyses (Table 1 ) showed no significantly greater effect for low quality trials (quality score 2/5) compared with higher quality trials (quality score 3–5/5) (z = 1.69, p = 0.091). There was no significant difference for smaller versus larger trials using 50 patients per group (median group size for topical NSAID was 49) as a cut off (z = 0.40, p = 0.69), for preferred outcomes versus lower preference outcomes (physician determined or general improvement) (z = 1.56, p = 0.12), or for patients with knee osteoarthritis compared with other musculoskeletal conditions (z = 0.99, p = 0.32) (Figure 2 ). The 10 trials with both a quality score of 3/5 or greater and a validity score of 9/16 or greater had an NNT of 4.4 (95% CI 3.6 to 5.6). There were insufficient data to allow comparisons of efficacy between different NSAIDs. Harm All 18 placebo controlled trials (2,032 patients) provided some information on adverse events (Table 2 ). There was no statistically significant difference between topical NSAID and topical placebo for the number of patients experiencing local adverse events (6%), systemic adverse events (3%), or the number withdrawing due to an adverse event (1%). With topical NSAID or topical placebo, local adverse events were usually described as rash, itching or stinging, and were predominantly mild. Active controlled trials Efficacy There was sufficient information to pool results only from the three trials comparing a topical NSAID with an oral NSAID in patients with osteoarthritis of the knee or finger joints. One trial [ 34 ] compared piroxicam 0.5% gel with oral ibuprofen 1200 mg daily, another [ 38 ] compared diclofenac 1% gel with oral ibuprofen 1200 mg daily, and the third [ 41 ] compared eltenac 1% gel with oral diclofenac 100 mg daily. In these trials, with 764 patients, 37% had a successful outcome both with topical NSAID and oral NSAID. There was no statistically significant difference (relative risk 1.1; 95% CI 0.9 to 1.3). The other seven studies used different topical preparations and different comparators in small trials ( additional file 5 : Outcome details of active-controlled trials). Harm Eight of the active controlled trials (1,461 patients) provided some information on adverse events (Table 2 ). In two active controlled trials comparing topical with oral NSAID, local adverse events occurred more frequently (8%) with topical than with oral NSAID (3%). Systemic adverse events and adverse event withdrawals did not differ between topical and oral NSAID. No study documented specific instances of upper gastrointestinal bleeding or symptomatic ulcers. Discussion Patients in these trials all had moderate to severe baseline pain, and for those with osteoarthritis, disease severity was generally mild to moderate. Patients with most severe disease were specifically excluded in several trials because authors regarded topical NSAID to be inappropriate for their treatment. Both the original and this updated review concluded that topical NSAIDs were effective in chronic conditions. However, removing trials of lower quality, and topical agents that are not now regarded as topical NSAIDs, increased (worsened) the NNT from 3.1 (95% CI 2.7 to 3.8) to 4.7 (95% CI 3.8 to 5.9) for the outcome of at least half pain relief at two weeks for all topical NSAIDs compared to placebo. For every four or five patients with chronic pain treated with topical NSAID, one would benefit who would not have done with placebo. Three trials comparing topical with oral NSAID found no difference in efficacy. There are a number of aspects of this review that might question this demonstration of efficacy. The trials spanned several decades and retrospective examination finds fault with them in several respects. Many trials were small, and small size can allow chance effects to influence treatment and placebo event rates [ 4 ]. Different preparations were used, with different formulations, concentrations of active ingredient, and application schedules. Reported outcomes were not consistent, and a hierarchy of outcomes had to be constructed. It was inevitable that there would be some clinical heterogeneity, even when similar patients were treated, and when trials were both randomised and double blind, and of appropriate duration. We addressed these limitations with pre-planned sensitivity analyses. Using studies with higher quality and validity scores, larger size, or higher rather than lower preference outcomes made no difference. Patients treated for knee osteoarthritis derived the same degree of pain relief as those treated for general musculoskeletal conditions. The evidence was that topical NSAIDs were effective whatever strategy was used for sensitivity analysis, improving the robustness of the overall result. A possible criticism might be that there has been selective publication of trials showing topical NSAIDs to be effective, and suppression of trials where there was no difference between topical NSAID and placebo. Funnel plots do not reliably detect publication bias [ 13 , 14 ], so we did not use them or make any adjustment for possible publication bias [ 42 ]. We did approach every company in the world that we could identify as being involved with topical NSAID manufacture or sale for any additional unpublished trials, but no more unpublished material was identified. When unpublished material is found, it often does not change the relevance of a result [ 43 - 45 ]. It is important to emphasise that both active and placebo treatments were rubbed on, making any effect of rubbing equal in both groups. The mean placebo response in the included trials was 26%, compared with the mean response of 48% with topical NSAID. The response with placebo is consistent with that found in acute and chronic pain with a variety of conditions and endpoints [ 46 ]. Local adverse events were reported with equal frequency for topical NSAID and topical placebo in placebo-controlled trials, but more often for topical NSAID than oral NSAID in active controlled trials. There were no differences between topical NSAID and topical placebo, or topical NSAID and oral NSAID, for systemic adverse events, or withdrawals due to adverse events. Studies of short duration will not capture important long-term safety information, and this may be important for ongoing applications of gels, creams or sprays in chronic conditions. There is, however, information that indicates that topical NSAIDs do not cause the gastrointestinal harm found with oral NSAIDs [ 47 ], nor are they associated with increased renal failure [ 48 ]. Clearly there is a body of evidence to support the efficacy of topical NSAIDs in chronic painful musculoskeletal conditions. Despite removing smaller studies that were not double blind, and substituting newer, larger trials of higher quality, topical NSAIDs remained effective, though the NNT was higher (worse) than originally estimated [ 1 ]. More information of high quality is required, to compare the relative efficacy of topical and oral NSAIDs, and between different topical NSAIDs. We are able to compare the evidence for different topical analgesics in chronic musculoskeletal pain (Table 3 ). Systematic reviews of topical salicylate [ 49 ] and capsaicin [ 50 ], tell us what is known about those treatments. As Table 3 shows, topical NSAIDs have been tested in many more studies, and in four times as many patients as these other topical analgesics, and have the lowest (best) NNT. The limitation of this comparison is essentially the same limitation as with all these reviews, that the included trials were too short and too small to be sure of the result. Topical NSAIDs have the best evidence for chronic musculoskeletal pain, supporting the excellent evidence available in acute painful conditions [ 51 ]. Authors' contributions LM was involved with planning the study, searching, data extraction, analysis, and preparing a manuscript; RAM with planning, data extraction, analysis and writing the manuscript; JE with searching, data extraction, and writing; SD with data extraction, analysis, and writing; HJM with planning, analysis and writing. All authors read and approved the final manuscript. Competing interests RAM & HJM have consulted for various pharmaceutical companies. RAM, HJM & JE have received lecture fees from pharmaceutical companies related to analgesics and other healthcare interventions. All authors have received research support from charities, government and industry sources at various times, but no such support was received for this work. No author has any direct stock holding in any pharmaceutical company. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Search strategy for RCTs of topical NSAIDs in chronic pain Click here for file Additional File 2 Excluded studies Click here for file Additional File 3 QUOROM flow diagram Click here for file Additional File 4 Outcome details of placebo-controlled trials Click here for file Additional File 5 Outcome details of active-controlled trials Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516039.xml
516777	Survival of patients with metastatic breast cancer: twenty-year data from two SEER registries	Background Many researchers are interested to know if there are any improvements in recent treatment results for metastatic breast cancer in the community, especially for 10- or 15-year survival. Methods Between 1981 and 1985, 782 and 580 female patients with metastatic breast cancer were extracted respectively from the Connecticut and San Francisco-Oakland registries of the Surveillance, Epidemiology, and End Results (SEER) database. The lognormal statistical method to estimate survival was retrospectively validated since the 15-year cause-specific survival rates could be calculated using the standard life-table actuarial method. Estimated rates were compared to the actuarial data available in 2000. Between 1991 and 1995, further 752 and 632 female patients with metastatic breast cancer were extracted respectively from the Connecticut and San Francisco-Oakland registries. The data were analyzed to estimate the 15-year cause-specific survival rates before the year 2005. Results The 5-year period (1981–1985) was chosen, and patients were followed as a cohort for an additional 3 years. The estimated 15-year cause-specific survival rates were 7.1% (95% confidence interval, CI, 1.8–12.4) and 9.1% (95% CI, 3.8–14.4) by the lognormal model for the two registries of Connecticut and San Francisco-Oakland respectively. Since the SEER database provides follow-up information to the end of the year 2000, actuarial calculation can be performed to confirm (validate) the estimation. The Kaplan-Meier calculation for the 15-year cause-specific survival rates were 8.3% (95% CI, 5.8–10.8) and 7.0% (95% CI, 4.3–9.7) respectively. Using the 1991–1995 5-year period cohort and followed for an additional 3 years, the 15-year cause-specific survival rates were estimated to be 9.1% (95% CI, 3.8–14.4) and 14.7% (95% CI, 9.8–19.6) for the two registries of Connecticut and San Francisco-Oakland respectively. Conclusions For the period 1981–1985, the 15-year cause-specific survival for the Connecticut and the San Francisco-Oakland registries were comparable. For the period 1991–1995, there was not much change in survival for the Connecticut registry patients, but there was an improvement in survival for the San Francisco-Oakland registry patients.	Background Prospective trials have the disadvantages of requiring a long time to complete, and using highly selected patient subgroups in tertiary centers. While one waits for the results to mature, this delays additional research to improve treatment. If there were a method that allowed earlier prediction of the results of prospective trials, advances in cancer treatment could be attained within a shorter time period. There is a parametric lognormal model, proposed by Boag [ 1 - 3 ] that had been retrospectively validated in the literature, and could be used prospectively for clinical trials to predict long-term survival rates several years earlier than would otherwise be possible using the standard life-table/actuarial Kaplan-Meier method of calculation [ 4 ]. The prognosis for metastatic breast cancer is generally poor and therefore it is believed that statistical prediction models for long-term survival rates are not needed. Nevertheless, specific subgroups of metastatic breast cancer patients exist, for which depending on the treatment given, the prognosis is improved so that some patients can survive for some time, particularly for those with limited organs involvement such as involvement with bone and/or skin only. In this situation, for which the present study was relevant, a prediction model, even for metastatic breast cancer, can be useful. Breast cancer, among other cancers, has the highest incidence in women, and many studies are currently in progress to assess treatment regimens. If, even for a subgroup of patients, the 10- and 15-year survival rates can be predicted from follow-up data available only 3 years after a 5-year diagnosis period, this would be a useful means of obtaining study results earlier than would otherwise have been possible. For example, a 15-year survival rate calculated by the Kaplan-Meier method requires at least some patients to have been followed for 15 years. In addition prediction model such as the lognormal model can also be used to review the progress of treatment results for a specific period from a treatment center, and to compare that with another specific period of the same treatment center to evaluate the potential impact for any possible change in treatment policy or guideline. Boag's lognormal model for long-term cancer survival rates has been available for use for some 50 years. When the lognormal model was first proposed in the 1940s, it was difficult to implement because of a lack of computing power, and lack of good quality long-term follow-up data from cancer registries. Since 1970s the model has been used by some authors in breast cancer, cervix uteri cancer, head and neck cancer, intraocular melanoma, choroidal-ciliary body melanoma, and small cell lung cancer [ 5 - 10 ]. Currently, although the computing power is sufficient, good quality follow-up data on a sufficient number of patients are seldom available, and it can be a limitation for its application. Large data registry such as the Surveillance, Epidemiology, and End Results (SEER) data [ 11 ] with good long-term follow-up data available can overcome this potential limitation. Methods Between 1981 and 1985, 782 and 580 female patients of metastatic breast cancer were extracted respectively from the Connecticut and San Francisco-Oakland registries from the SEER database using SEERStat 5.0 software. The two registries were chosen because they are two of the earliest registries, with a large population. The data used were survival time, vital status, cause of death, age at diagnosis, and race. The cause-specific survival was defined as the interval from the date of diagnosis to the date of death from breast cancer or the last follow-up date for censoring purposes, if the patient was alive and still being followed at the time of analysis. The survival time of the uncured group of patients who died of breast cancer had been verified to follow a lognormal distribution previously [ 12 ]. Next, between 1991 and 1995, 752 and 632 female patients of metastatic breast cancer were extracted respectively from the two registries. The data were used to estimate the 15-year cause-specific survival rates before the year 2005. To be comparable, for both the 1981–1985 and 1991–1995 cohorts, the staging system used was the SEER historical system (classified as localized, regional, or distant, based on combined pathologic and clinical data). The choice of 1981–1985 and 1991–1995 has the advantage that the two time periods are not too far apart otherwise there would be too much changes of medical practice. These time periods have a minimum of 5 years follow-up. The overall survival rates (OSR) of the two time periods were calculated using the Kaplan-Meier method. The actual relative survival rates (RSR) were calculated using SEERStat 5.0 software. The modified version of period analysis [ 13 ] was applied using the Hakulinen method [ 14 ] to obtain more up-to-date absolute survival rates (ASR) and relative survival rates (RSR) for comparison purpose with a computer program run by Microsoft Excel software. Validation of the lognormal model The validation of the lognormal model has two phases. Phase 1 tests the goodness of fit to a lognormal distribution of the survival time of those cancer patients who died with their disease present, termed an uncured group with a fraction of 1-C, where C is the cured proportion of patients. The lognormal distribution is similar to the normal distribution in that if the variable in the normal is time t, the variable in the lognormal is the logarithm of t. In other words, the investigators attempt to show that the logarithm of the survival time follows a normal distribution. Phase 2 attempts to show that the lognormal model, using short-term follow-up data, can predict long-term survival rates comparable to those calculated by the Kaplan-Meier life-table method with long-term available. This model can be used to estimate long-term cause-specific survival rates (CSSR) by a maximum likelihood method (e.g., 10-year and 15-year survival rates) from only short-term follow-up data. The maximum likelihood method is used to estimate the CSSR at time τ, and is calculated as [C+(1-C)·Q]·100%, where Q is the integral of the lognormal distribution between the limits time τ and infinity. The lognormal statistical model had been validated in stages III and IV breast cancer in a previous publication that survival rates could be estimated several years earlier than is possible using the standard life-table actuarial method [ 12 ]. The survival time of unsuccessfully treated cases could be represented by a lognormal distribution, the long-term survival rates were predicted by Boag's method using a computer program run by Microsoft Excel. In this parametric lognormal model, the standard deviation S was fixed, and only the two remaining parameters, mean M and proportion cured C, were kept floating when using the maximum likelihood method. Multiple iterations converged to a stable solution for C. A 5-year period of diagnosis could be selected and patients followed as a cohort for an additional 3 years. The current study was for metastatic breast cancer patients treated between 1981 and 1985, with follow-up to the end of year 2000, making the series ideal for validating purposes. For example, for cases diagnosed during the 5-year period, prediction of the 15-year survival rate was made using data at the follow-up cutoff date of December 31, 1988 (i.e., 3 years after 1985). The 15-year survival rate prediction was then validated by Kaplan-Meier life-table calculations using the follow-up data available in 2000. For metastatic breast cancer patients treated between 1991 and 1995, and follow-up to the end of year 2000, prediction of the 15-year survival rate was made using data at the follow-up cutoff date of December 31, 1998 (i.e., 3 years after 1995) before the year 2005. Results From the cohort of 1981–1985 inclusively, 782 patients from the Connecticut registry were followed to the end of 1988. The lognormal model predicted the 15-year CSSR to be 7.1% (95% CI, 1.8–12.4). The 15-year CSSR was 8.3% (95% CI, 5.8–10.8) validated by the Kaplan-Meier calculation using actuarial follow-up data up to the end of year 2000. From the cohort of 1981–1985 inclusively, 580 patients from the San Francisco-Oakland registry were followed to the end of 1988. The lognormal model predicted the 15-year CSSR to be 9.2% (95% CI, 3.9–14.5). The 15-year CSSR was 7.0% (95% CI, 4.3–9.7) validated by the Kaplan-Meier calculation using actuarial follow-up data up to the end of year 2000. Using the same method, the cohort of 1991–1995 inclusively, 752 patients from the Connecticut registry were followed to the end of 1998. The lognormal model predicted the 10-year CSSR to be 12.6% (95% CI, 7.3–17.9). The 10-year CSSR was 11.3% (95% CI, 7.8–14.8) validated by the Kaplan-Meier calculation using actuarial follow-up data up to the end of year 2000. The lognormal model predicted the 15-year CSSR to be 9.1% (95% CI, 3.8–14.4), which cannot be validated before 2005. For the cohort of 1991–1995 inclusively, 632 patients from the San Francisco-Oakland registry were followed to the end of 1998. The lognormal model predicted the 10-year CSSR to be 17.0% (95% CI, 12.1–21.9). The 10-year CSSR was 15.9% (95% CI, 11.4–20.4) validated by the Kaplan-Meier calculation using actuarial follow-up data up to the end of year 2000. The lognormal model predicted the 15-year CSSR to be 14.7% (95% CI, 9.8–19.6), which cannot be validated before 2005. For the period 1991–1995, there was not much change of only about 2% absolute percentage point in the predicted 15-year CSSR for the Connecticut registry, but there was an improvement of about 6% absolute percentage points for the San Francisco-Oakland registry when compared with the period 1981–1985 15-year CSSR, which was validated by the Kaplan-Meier calculation. (Table 1 ) For comparison purpose, the actual OSR and RSR were compared with the ASR and RSR obtained by the period analysis. (Tables 2 and 3 ) It was found that there were more patient survival improvements shown in the actual OSR and RSR for the San Francisco-Oakland registry, but not much for the Connecticut registry. However the period analysis results did not show such improvements. Discussion Lognormal model Rutqvist studied the fit of Boag's lognormal model to the survival times of 8170 breast cancer cases reported to the Swedish Cancer Registry during 1961–1963. The model fitted the 1961–1963 data well for the entire case material and for patients aged less than 70 years. In this registry, the lognormal model did not fit the data for patients aged greater than 70 years, who were more likely to be censored because of coincidental causes of death. Another disadvantage stated by the author was that large number of patients was required to obtain estimates with reasonably small standard errors for breast cancer. With another series of the Norwegian Cancer Registry of 14,000 breast cancer cases, Rutqvist et al . [ 15 ] deduced that lognormal is the best model because other models did not fit the observed survival in all stages, ages, and time periods (two-parameter models, such as exponential or extrapolated actuarial, or three-parameter models, such as sum of two exponential, exponential with shoulder, Weibull). Both the exponential and extrapolated actuarial models assume that the conditional relative survival is lowest immediately after treatment. With the lognormal model, the survival curve has a low initial mortality that rapidly increases to a maximum, with a slow decrease in the mortality after the maximum has occurred. Requirements for using the lognormal model The lognormal model can only predict cause-specific survival, because other coincidental causes of death are too unpredictable (e.g., the rate of stroke). Therefore, overall survival cannot be predicted. The maximum likelihood method is the most accurate method for fitting the lognormal model with the smallest mean squared error. However, there are some requirements for its use. The maximum likelihood method fails to converge to a stable solution using the initial estimates if there is extensive censoring within the data. This occurs if patients are lost to follow-up or die from coincidental non-cancerous causes. The frequency of failure to yield a successful fit for lognormality was greater when one-fourth of cases were designated as lost to follow-up. Gamel et al . established a stable linear algorithm for fitting the lognormal model to survival data. To achieve convergence, some authors have fixed one or two parameters of the lognormal model to pre-selected values to simplify the iterative procedure required for convergence [ 6 ]. Some prognostic factors follow lognormal distribution Prognostic factors in patients with distant metastases at the time of diagnosis were investigated by Rudan et al . [ 16 ], and Chapman et al . [ 17 ], primary tumor size was a significant prognostic factor. Engel et al . [ 18 ] found that the number of metastatic cases and the time to metastasis depended on the tumor diameter at diagnosis. Cell growth is essential for the development of tumors. Tumor size is therefore the most important factor in describing tumor biology. As the tumor size increases, the probability of node-positivity increases. Another study group also found this correlation up to 5 cm [ 19 ]. Tubiana and Koscielny [ 20 ] have found a highly significant correlation between tumor size and the probability of distant metastasis. The distribution of tumor sizes at metastatic spread was lognormal with a median diameter equal to 3.5 cm. The patients were subdivided into 3 groups according to the histological grade. In each subgroup there was a significant correlation between tumor size and the probability of distant spread. The distributions were lognormal and the median size was markedly larger for grade 1 tumors. A number of quantitative postmortem observations regarding the size distribution of metastases have been published [ 21 - 23 ]. These studies revealed a skewed distribution with a high proportion of smaller metastases, and a significant tail extending to the larger metastases, consistent with a lognormal distribution. The more detailed measurements from human liver metastases provided by Yamanami et al . were found to approximate the lognormal distribution reasonably well. A hypothesis was proposed by Kendal [ 24 ] that the time available for the growth of metastases is normally distributed, presumably as a consequence of the summation of multiple independently distributed time intervals from each of the steps and of the Central Limit Theorem. For exponentially growing metastases, the corresponding size distribution would be lognormal; Gompertzian growth would imply a modified (Gompertz-normal) distribution, where larger metastases would occur less frequently as a consequence of a decreased growth rate. These two size distributions were evaluated against 18 human autopsy cases where precise size measurements had been collected from over 3900 macroscopic hematogenous organ metastases. The lognormal distribution provided an approximate agreement. Its main deficiency was a tendency to over-represent metastases greater than 10 mm diameter. These observations supported the hypothesis of normally distributed growth times, and qualified the utility of the lognormal and Gompertz-normal distributions for the size distribution of metastases. Why is the lognormal model applicable to so many organ sites [ 3 , 6 - 10 , 12 , 25 - 36 ] (Table 4 )? Boag's explanation for the lognormal survival time distribution was that if the patient was not cured by treatment, the length of the remaining survival time would be dependent principally on the growth rate of the tumor remnants. Pearlman [ 37 ] estimated the growth rates of breast cancer that recurred in the scar, assuming that the recurrence started from a single cell. He found that the growth rates were approximately lognormally distributed. Likewise, von Fournier et al . [ 38 ] found that the growth rates of breast cancers followed by serial mammography were lognormally distributed. Variation of survival rates over time In order to determine whether current programs for the management of metastatic breast cancer have led to improved patient survival, Debonis et al . [ 39 ] determined the median survival times for five-year intervals of 849 patients admitted to the City of Hope National Medical Center with metastatic breast cancer from 1955 to 1980. Survival times in each of the clinical subsets remained unchanged during the period of observation, regardless of the therapeutic modalities included in the treatment regimens. The study indicates that changes in palliative therapy for metastatic breast cancer during the 25 years of observation have not influenced overall survival. On the contrary, Dickman et al . [ 40 ] studied the survival of cancer patients in Finland during the years 1955–1994. The 5-year RSR for distant metastases breast cancer had increased from 10% for the period 1955–1964 to 22% for 1985–1994. The tumor registry at Yale-New Haven Hospital, which began recording data in 1920, was utilized by Todd et al . [ 41 ] to examine the ultimate outcome of all breast cancer patients who were initially diagnosed at Yale with metastatic breast cancer. The median survival of these patients increased steadily from 21 months in 1920 to 41 months in the decade from 1970 to 1980. The percentage of women actually surviving 5 years increased from 5% in the 1920s to approximately 25% in the 1960s. Despite the use of combination drug programs in the 1970s, the percentage of these patients remaining alive at 5 years remained near 25%. Firm conclusions cannot be made from a retrospective study spanning 60 years, although the trends depicted the lack of continued improvement indicate that the current therapeutic approach to metastatic breast cancer in that period may not result in dramatic improvement in overall survival. Geographical variation of survival rates Farrow et al . [ 42 ] documented substantial geographical variation in patterns of treatment of cancer and other diseases. Because cancer treatment is not uniform nationwide in the States, survival following the diagnosis of cancer might also be expected to vary geographically. Survival data from the nine population-based registries in the SEER Program were analyzed for cancers of the stomach, colon, rectum, lung, breast, uterus, ovary, prostate, and bladder. The patients included all non-Hispanic white patients diagnosed with cancer of one of the selected sites during 1983–1991. Regional variation in crude five-year survival rates across the nine SEER areas was most marked for cancers of the uterus and prostate. For uterine cancer, for example, five-year survival ranged from 73.2% in Connecticut to 84.0% in Hawaii. Less marked variation was observed for cancers of the colon, rectum, and breast. For cancers of the bladder, ovary, stomach, and lung, survival rates five years after diagnosis were relatively invariant across the SEER areas. Maggard et al . [ 43 ] also found that variations in the breast cancer mortality rates exist between states. A nearly 50% increase is observed between the states with the highest and lowest mortality rates. Adjusted analyses demonstrated that stage at presentation is a more important predictor of mortality variation than treatment differences. Goodwin et al . [ 44 ] examined breast cancer incidence, survival, and mortality in the 66 health service areas covered by the SEER program for women aged 65 and older at diagnosis. They found that there was considerable geographic variation in survival from breast cancer among older women, and this contributed to variation in breast cancer mortality. The elevated mortality in the Northeast is apparent only in older women [ 45 ]. For women aged 65 years and older, breast cancer mortality is 26% higher in New England than in the South, while incidence is only 3% higher. Breast cancer mortality for older women by state correlates poorly with incidence (r = 0.28). The above-mentioned results are consistent with that from the present study: the Connecticut registry has lower CSSR than the San Francisco-Oakland registry for the period 1991–1995. The Connecticut cohort has median age at diagnosis of 66 (range 25–103), while the San Francisco-Oakland cohort has lower median age of 63 (range 26–96). It could be argued that new treatments evolved in the recent decade have improved the survival of the breast cancer patients, and younger patients benefit more than the older patients. Apart from treatment offered, changes of survival rates over time or geographical areas can be due to co-morbidities or other characteristics such as race, age, and differences in staging procedures. Conclusions For the period 1981–1985, the 15-year cause-specific survival for the Connecticut and the San Francisco-Oakland registries were comparable. For the period 1991–1995, there was not much change in survival for the Connecticut registry patients, but there was an improvement in survival for the San Francisco-Oakland registry patients. List of abbreviations CSSR: Cause-specific survival. SEER: Surveillance, Epidemiology, and End Results. OSR: Overall survival rate. ASR: Absolute survival rate. RSR: Relative survival rate. Competing interests None declared. Authors contributions PT: Data analysis and writing of manuscript. EY, VVH, GC, GV: Critical appraisal of manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516777.xml
532400	A Global View of Gene Expression in the Aging Kidney	null	Four years ago in Science , Stuart Kim, a Stanford developmental biologist, made the case for laying down the broad strokes of a complex physiological process before defining its mechanisms. “A powerful, top-down, holistic approach,” he wrote, “is to identify all of the components of a particular cellular process, so that one can define the global picture first and then use it as a framework to understand how the individual components of the process fit together.” To get a broad view of gene expression in the aging nematode, Kim's lab turned to DNA microarrays and functional genomics. In a new study, Kim and colleagues apply this same approach to the decidedly more complex problem of human aging and “present a molecular portrait” of the aging kidney. Scientists have identified a wide range of molecular pathways and mechanisms associated with aging. Many have been found in evolutionarily distant organisms, suggesting they have been conserved and could shed light on human aging. Yet other studies suggest that since few animals reach old age in the wild, any aging-related physiological changes aren't likely to impact the fitness of a population and so aren't likely to be conserved. Consequently, aging pathways in worms, for example, would have little bearing on humans. To investigate the molecular pathways associated with human aging, the authors focused on human tissue—in this case, the kidney. Kidneys came from 74 patients, ranging in age from 27 to 92. Samples were extracted from donated kidneys or “meticulously harvested” from kidneys with localized disease to ensure only normal tissue was taken. Two structures that are critical to kidney function (removing toxins from blood) were removed from each sample: the renal cortex, which filters plasma, and the medulla, which alters urine composition to maintain fluid balance. Both deteriorate with age. An extensive clinical history was noted for each sample to account for any potentially confounding medical factors. Transcriptional profiling to study aging in the kidney Kim and colleagues then isolated RNA transcripts from the samples to determine the activity of every gene, broken down by age and kidney section, through microarray analysis. Looking for differences in gene expression across the genome, they identified genes that showed a statistically significant change in expression as a function of age. Of 33,000 known human genes on the microarray, 985 showed age-related changes, most showing increased activity. These changes are truly age-regulated, the authors conclude, since none of the medical factors impacted the observed changes in gene expression. Although cortex and medulla have different cell types and perform different functions, their genetic aging profile was very similar, suggesting a common aging mechanism operates in both structures. In fact, these mechanisms may function broadly, as most of the age-regulated kidney genes were also active in a wide range of human tissues. Other organisms appear to lack these changes, however, prompting the authors to argue that understanding aging in humans will require human subjects. Most importantly, the genetic profile of the tissue samples correlated with the physiological and morphological decline of an aging kidney. An 81-year-old patient with an unusually healthy kidney had a molecular profile typical of someone much younger, while a 78-year-old with a damaged kidney had the profile of a much older person. Using the power of functional genomics, this study has identified a set of genes that can serve as molecular markers for various stages of a deteriorating kidney and predict the relative health of a patient compared to their age group. These gene sets can also serve as probes to shed light on the molecular pathways at work in the aging kidney, and possibly on the process of aging itself.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532400.xml
521695	Chromosome loops arising from intrachromosomal tethering of telomeres occur at high frequency in G1 (non-cycling) mitotic cells: Implications for telomere capture	Background To investigate potential mechanisms for telomere capture the spatial arrangement of telomeres and chromosomes was examined in G1 (non-cycling) mitotic cells with diploid or triploid genomes. This was examined firstly by directly labelling the respective short arm (p) and long arm subtelomeres (q) with different fluorophores and probing cell preparations using a number of subtelomere probe pairs, those for chromosomes 1, 3, 4, 5, 6, 7, 9, 10, 12, 17, 18, and 20. In addition some interstitial probes (CEN15, PML and SNRPN on chromosome 15) and whole chromosome paint probes (e.g. WCP12) were jointly hybridised to investigate the co-localization of interphase chromosome domains and tethered subtelomeres. Cells were prepared by omitting exposure to colcemid and hypotonic treatments. Results In these cells a specific interphase chromosome topology was detected. It was shown that the p and q telomeres of the each chromosome associate frequently (80% pairing) in an intrachromosomal manner, i.e. looped chromosomes with homologues usually widely spaced within the nucleus. This p-q tethering of the telomeres from the one chromosome was observed with large (chromosomes 3, 4, 5), medium sized (6, 7, 9, 10, 12), or small chromosomes (17, 18, 20). When triploid nuclei were probed there were three tetherings of p-q subtelomere signals representing the three widely separated looped chromosome homologues. The separate subtelomere pairings were shown to coincide with separate chromosome domains as defined by the WCP and interstitial probes. The 20% of apparently unpaired subtelomeric signals in diploid nuclei were partially documented to be pairings with the telomeres of other chromosomes. Conclusions A topology for telomeres was detected where looped chromosome homologues were present at G1 interphase. These homologues were spatially arranged with respect to one-another independently of other chromosomes, i.e. there was no chromosome order on different sides of the cell nuclei and no segregation into haploid sets was detected. The normal function of this high frequency of intrachromosomal loops is unknown but a potential role is likely in the genesis of telomere captures whether of the intrachromosomal type or between non-homologues. This intrachromosomal tethering of telomeres cannot be related to telomeric or subtelomeric sequences since these are shared in varying degree with other chromosomes. In our view, these intrachromosomal telomeric tetherings with the resulting looped chromosomes arranged in a regular topology must be important to normal cell function since non-cycling cells in G1 are far from quiescent, are in fact metabolically active, and these cells represent the majority status since only a small proportion of cells are normally dividing.	Background In both plants and animals, during early meiosis in normal cells there is a clustering of all or most of the telomeres of the entire chromosome set to a single region on the nuclear membrane [ 1 - 3 ]. This meiotic looping of chromosomes with clustered ends has been termed the bouquet arrangement which appears synchronously with synapsis of bivalents. The reason why the telomeres attach to the nuclear membrane in meiosis is not dependent on the presence of normal numbers of TTAGGG repeats and, in fact, still occurs in late generation Terc-/- mice without detectable pantelomere repeats [ 5 ]. In plants the meiotic telomere clustering can be inhibited by colchicine [ 6 ] but a polarization still remains within the nuclei such that microtubules and nuclear pores are still arranged in a region that normally would face the telomere cluster on the opposite side of the nuclear membrane [ 3 ]. In mammals the bouquet arrangement seen at early meiosis occurs with some minor differences between males and females [ 4 ] but has disappeared in both by diplotene/dictyotene. For mitosis there is much less data on the position or possible associations of telomeres and subtelomeres. However, the spatial arrangement of chromosomes at mitotic interphase has been studied intensively [ 7 , 8 ] but there are few studies with data on the principles that dictate nuclear organization. Nagele et al. [ 9 , 10 ] using whole chromosome paints on fixed normal diploid human cells described a radial array (rosette) of prometaphase chromosomes where the chromosomes were apparently arranged in two tandemly linked haploid sets. That interphase chromatin formed ring-like shapes was already known [ 11 , 12 ] but Nagele et al. [ 9 , 10 ] proposed that there was a chromosome order in each of the haploid sets in diploid cells during mitosis which was thought to be reversed with respect to one-another. A chromosome order was also described as being present at interphase in non-cycling cells [ 13 ] where the nuclear organization seems to be fundamentally different from that in dividing cells [ 14 , 15 ]. From observations in triploid cells, Nagele et al. [ 10 ] proposed that the three haploid sets were spatially arranged, two with the chromosomes arranged in tandem and the third with a reversed chromosome order. The relationship of subtelomeric regions to these concepts of a chromosome order within a radial chromosome array is less clear. Stout et al . [ 16 ] studied subtelomeric chromosome regions at interphase and showed that compared to interstitial chromosome sites, subtelomeres showed an increased number of somatic pairings. By FISH within living cells, Molenaar et al. [ 17 ] were able to demonstrate that these telomeric associations are dynamic. The rate of telomeric associations apparently depends on the stage of the cell cycle. Nagele et al. [ 18 ] utilising a telomere-specific peptide nucleic acid probe has demonstrated that the prevalence of such telomeric associations is far higher at interphase in non-cycling cells than in their cycling counterparts. In the present study we examine the telomere associations in mitotic interphase in human non-cycling cells of diploid or triploid karyotype. The cell types used were from skin, fetal cartilage, and long-term culture of chorionic villi but colchicine and hyptotonic treatments were avoided during cell harvest because of the potential effect of disrupting any topology present [ 6 , 19 ]. We report a new finding, the detection of looped chromosomes in mitotic G1 by the intrachromosomal tethering of short-arm (p) and long-arm (q) telomeres. This new finding has implications for the understanding of the normal dynamics of chromosome behaviour at interphase but also for the processes involved in telomere capture. Results Fluorescence in situ hybridization (FISH) performed with the various subtelomere probes (Table 1 ) gave discrete signals in all experiments attempted. Figure 1 shows FISH of cells probed with the p-subtelomeres labeled red and q-subtelomeres labeled green for chromosomes 4, 5, 7, 10, 17, and 20, arranged respectively in figure 1A,1B,1C,1D,1E,1F (diploid cell line CG04-0743BBRS) and for chromosomes 18, 12 and 6 in figure 1G,1H,1I (triploid cell line CG01-2042YA). The proportion of p-q associated signals is shown in Table 2 . The frequency of p-q subtelomere tethering ranged from 76–85% in the diploid cells but was a little less in the triploid cells (58–94%). Not all signal pairs were tethered. The percentage of diploid cells with all p-q signals tethered was 46–72% as compared to 33–58% in the triploid cells. This would be expected with more opportunity for interhomologous tethering in the triploid nuclei with an extra chromosome. Table 1 Origin and derivation of the telomere clones used in the study. Clone Chrom Supplier 1186B18 3p Flint 196F04 3q Incyte 36P21 4p Incyte 963K6 4q Flint 189N21 5p Incyte 240G13 5q Incyte 62I11 6p Incyte 57H24 6q Incyte 164D18 7p Incyte 3K23 7q Incyte 43N06 9p Incyte 112N13 9q Incyte 306F07 10p Incyte 137E24 10q Incyte 496A11 12p Flint 221K18 12q Incyte 2111b1 17p ATCC 362K4 17q Flint 52M11 18p ATCC 964M9 18q Flint 1061L1 20p Flint 81F12 20q Incyte Note: The host strain was E. coli DH10B with kanamycin resistance in all cases except the clone 2111b1 (17p probe) which was ampicillin resistant. Table 2 Rate of tethering in non-cycling cells at G1 interphase of p (short arm) to q (long arm) subtelomeric signals in single homologues. Chromosome Genome of cells Numbers and [Percentage] of p-q signal pairs** tethered (95% confidence limits) Numbers and [Percentage] of cells with all p-q signals tethered# (95% confidence limits). 4 Diploid 50/60 [83] (71–91%) 10/14 [72] (42–92%) 5 Diploid 123/148 [83] (76–89%) 26/41 [63] (47–78%) 7 Diploid 86/113 [76] (67–84%) 26/45 [58] (42–72%) 9 Diploid ND* ND* 10 Diploid 92/108 [85] (77–91%) 24/35 [69] (51–83%) 17 Diploid 62/76 [82] (71–90%) 13/28 [46] (28–66%) 20 Diploid 72/90 [80] (70–88%) 14/24 [58] (37–78%) 3 Triploid 25/30 [83] (65–94%) 5/10 [50] (19–81%) 6 Triploid 49/61 [80] (68–89%) 12/22 [55] (32–76%) 12 Triploid 57/82 [70] (58–79%) 8/24 [33] (16–55%) 18 Triploid 65/85 [77] (66–85%) 12/28 [43] (25–63%) ND-not determined. * Cross-hybridization between 9q and 18p prevented analysis. ** p-q subtelomere signal pairs scored, irrespective of cell numbers. #Separate scoring of discrete cells only; i.e. diploid cells with clearcut tethering of both p (short-arm) to q (long-arm) subtelomere signal pairs and triploid cells with all three p-q subtelomere signal pairs tethered. Figure 1 A-I Intrachromosomal tethering of the subtelomeres of each single homologue in diploid and triploid non-cycling interphase nuclei at G1. FISH of diploid (A-F) or triploid interphase nuclei (G-I) from the following cell lines: CG04-0743BBRS (diploid) derived from skin and CG01-2042YA (triploid) derived from CVS. These non-cycling cells were probed with p-subtelomeric probe (labelled with spectrum orange) and q-subtelomeric probe (spectrum green) for (A) chromosome 4; (B) chromosome 5; (C) chromosome 7, (D) chromosome 10, (E) chromosome 17, (F) chromosome 20, (G) chromosome 18, (H) chromosome 12, (I) chromosome 6. The proportion of p-q tethered signals is shown in Table 2. In each case the majority of cells (76%–85%) showed pairing of short-arm and long-arm subtelomeres from single homologues often arranged on opposite sides of the interphase nucleus. Note also that an interphase topology is exhibited such that oval rosettes of chromatin can be seen in the present study in figs 1H, 1I, an elongated rosette in fig 1C, and off-centre rosettes in figs 1A, 1B. The triploid cells were used to test the likelihood that intrachromosomal pairing of subtelomeric signals was occurring rather than the pairing of p and q signals with the q and p signals of the other homologue(s). As can be seen in figure 1G,1H,1I , there were three p-q tethered signal pairs in the triploid interphase nuclei. Figure 2 shows single arm subtelomeric probes from three different chromosomes demonstrating that there is no linkage of positioning (chromosome order) between non-homologues. This is evidence challenging the claims of haploid groups being present at the interphase of non-cycling cells. Figure 2 A-F Chromosome homologues at G1 in nuclei of non-cycling cells are spatially arranged without respect to non-homologues. Same cell line and same cell harvest as the triploid cells probed in Fig 1. Two combinations of three subtelomeric probes (see Table 1 for clones) are shown hybridized to triploid cells. In fig 2A-C the nuclei are probed with three single subtelomere probes from 4p (spectrum orange); 18q (spectrum green) and 6p (both spectrum orange and spectrum green labels, i.e. yellow signal). In figs 2D-F, the nuclei are probed with three subtelomere probes labelled 5p (spectrum orange), 12q (spectrum green), and 20p (spectrum orange and spectrum green, i.e. yellow signal). Note: There was no segregation into haploids sets of chromosomes at G1 interphase. Homologues were regularly arranged without any defined relationship to non-homologous signal groups; i.e. haploid sets of interphase chromosomes distributed to separate nuclear regions do not appear to exist. Note also there is a low frequency of isolated non-homologous associations: between 4p and 6p (Fig 2A); between 6p and 18q (Fig 2C), and between 5p and 20p (Fig 2D). Further confirmation that the p-q tetherings in figure 1 were from single chromosomes is shown in figure 3 . Figure 3A,3B,3C , shows chromosome 15 interstitial loci (diploid cell line CG04-0743BBRS) probed together with the 15 alpha centromeric probe, and a 15q subtelomeric locus. Separate chromosome domains surround the subtelomeric signals (Fig 3A,3B,3C ). Similarly for chromosome 12 (Fig 3D,3E,3F ), using the same diploid cells (CG04-0743BBRS), subtelomeric probe pairs are defined to occur within the two separate chromosome domains by jointly using WCP12. Figure 3 A-F Looped chromosomes in G1 arrested cells: the distribution of tethered subtelomeric signals coincides with chromosome domains. Diploid non-cycling cells harvested after confluence arrest. The diploid cells are from the same cell line as in Fig 1 (i.e. CG04-0743BBRS). Fig 3A-3C shows diploid cells probed for chromosome 15 with CEN15 (larger signal spectrum green); SNRPN at 15q12 (spectrum orange); PML at 15q22 (spectrum orange); and subtelomeric 15q probe (smaller signal spectrum green). Note: The two chromosome 15 domains coincide with and envelop the 15q subtelomeric signals (there is no currently recognized specific 15p subtelomeric sequence and hence no 15p subtelomeric probe). Fig 3D-3F shows diploid cells probed for chromosome 12 with the subtelomeric probes for 12p (labeled with spectrum orange) and 12q (labeled with spectrum green) and the WCP chromosome 12 (the spectrum green smear). Note: The three chromosome 12 domains as defined by the (directly labeled) WCP12 envelop the three tethered subtelomeric probe pairs. This confirms that the telomeric tethering represents looped chromosomes. Discussion Evidence for short-arm and long-arm subtelomeres of the one homologue associating This study shows that the pairings of red/green signals from the subtelomeres of the short-arm and long-arm respectively occur at high frequency in these non-cycling diploid nuclei. In many cases the association is so close that the subtelomeric signals are superimposed (e.g. figure 1A,1F ). The pairs of red/green, p/q signals are from a single chromosome with the two diploid homologues arranged on different sides of the nucleus. This has been shown in this study in several ways. Firstly, it is highly likely that separate looped chromosomes are involved since the paired subtelomeric signals occur with small chromosomes (chromosome 17, 18, 20), intermediate chromosomes (7, 9, 10, 12) or large chromosomes (3, 4, 5) and are observed in triploid as well as diploid cells. Indeed the wide separation of the two subtelomeric signals from pairs of homologues (e.g. fig 1B,1D ) supports the present interpretation that the telomeric tetherings of p-q signal pairs are intrachromosomal and not between homologues. Secondly, when interstitially located probes are used, for example on chromosome 15 (Fig 3A,3B,3C ) in diploid cells, two distinct chromosome domains are seen. Thirdly, when subtelomeric probe pairs are used with a WCP probe for example on chromosome 12 (Fig 3D,3E,3F ) in diploid cells, two distinct chromosome domains are seen that envelop the two tethered pairs of subtelomeric regions. In diploid nuclei the pairs of tethered subtelomeric signals are distributed to two areas and in triploid nuclei (fig. 1G,1H,1I ), the tethered signals are distributed to three areas. If the signal pairings were between the short-arm from one homologue with the long-arm of another it is especially unlikely in the triploid cells that the chromosomes could span the diameter of the nucleus consistently. This is especially unlikely in light of the finding by Nagele et al. [ 10 ] that the nucleus normally exhibits a rosette of (chromosome rich) chromatin with a less dense central core (doughnut shape). If inter-homologous telomeric associations were the explanation for the regular p-q signal pairings then, especially in triploid cells, chromosomal threads would have to be arranged in very complex formations across the chromatin poor cores of rosettes. Finally, there is separate evidence that there are small non-overlapping chromosome territories at interphase in mammalian cells [ 20 , 21 ] where the chromosomes are extended but not entwined. In the present study we have also been able to show the presence of these interphase chromosome domains both with the use of several probes spanning the length of chromosomes (e.g. Fig 3A,3B,3C ) or with chromosome paints (e.g. Fig 3D,3E,3F ). Nagele et al. [ 18 ] showed that there were very few coincident telomeric associations (TA's) in rapidly cycling mitotic cells. However, these authors showed [ 18 ] that in non-cycling cells there was a high rate of double associations, and a lesser frequency of triple and quadruple associations or unassociated telomeres. These authors [ 18 ] concluded that the replicative status of the cells was the prime determinant in the level of telomere associations. The finding of a high intrachromosomal p-q telomere association rate in the present study probably explains the underlying high telomere association rate described by Nagele et al. [ 18 ]. In that study [ 18 ], a universal telomere probe was used so the specificity of the associations, if present, was unrecognisable. In the present study, there was a high (~80% but not saturated) rate of intrachromosomal pairing with only ~20% of telomeres unpaired with their homologous subtelomere. These two studies can be reconciled if the apparently (~20%) unpaired subtelomeres (present study) were actually associated with non-homologous subtelomeres. Fig 2 shows the presence of an underlying low rate of non-homologous telomere tetherings in these G1 arrested cells. Regulation of telomere associations In early meiotic cells the presence of the normal numbers of universal TTAGGG sequences is not required for massed telomere clustering [ 5 ]. A complementary finding was reported by Nagele et al . [ 18 ] who showed that in late passage mitotic cells the number of telomere associations (TA's) did not increase during progression to late passage crisis. This indicates that telomere shortening did not increase the number of TA's. Since the pantelomeric repeats occur at all telomeres, the specific intrachromosomal association presently observed also cannot be due to their presence. Neither can the mechanism of tethering be related to chromosome specific subtelomeric sequences since the two homologues with identical sequences remain separated (Fig 1 ). There clearly are similarities between the looping of chromosomes seen in the present non-cycling mitotic cells and in the chromosome bouquets of early meiosis [ 2 , 3 ]. These two apparently disparate phenomena may be related. If the synapsis of bivalents, unnecessary in mitotic cells, was removed from the meiotic bouquet arrangement mechanism, the intrachromosomal tethering of separated homologues as presently observed is what may be left. This mitotic looping may have been originally present since meiosis is believed to have evolved from mitosis. Chromosome topology at interphase The global organisation of the interphase nucleus has provoked the interests of cell biologists for several decades but detecting the presence of any macromolecular domains has been challenging [ 8 ]. Nagele et al. [ 9 , 10 ] was able to confirm with Feulgen staining and FISH that the chromosomes were arranged in rosettes, a ring of chromatin with partly-condensed chromosomes, which persisted through mitosis and was even maintained in the daughter cells at telophase. Oval rosettes can be seen in the present study in figs 1H,1H , and 3D ; an elongated rosette in fig 1C , and off-centre rosettes in figs 1A,1B , and 2B . Through the use of FISH with chromosome specific alphoid probes and whole chromosome paints, Nagele et al. [ 10 , 13 ] attempted to show that chromosomes in the rosettes appeared to be in an orderly arrangement in both diploid and triploid cells. These authors interpreted this order as specifically positioned haploid sets [ 9 , 10 , 13 ]. The pairing of subtelomere signals in non-cycling cells at interphase, as in the present study, is in some aspects consistent with these prior observations though we do not accept that haploid sets are spatially segregated and we found no evidence for an interphase chromosome order in the non-cycling cells of our cell lines. We have repeated this work with centromeric probes (not shown) and again there was no evidence of haploid groups or of a regular chromosome order though widely spaced homologous centromeric signals are usually observed (with respect to each chromosome considered separately). With respect to telomeric tethering in cycling cells (at G2) no such p-q telomeric tethering pattern is present in our observations of lymphocytes (not shown) and the only associations are of sister chromatids. That most lymphocytes are at G2 can be observed by the doubled signals representing sister chromatids (not shown) which is in contrast to the single (chromatid) signals in the unreplicated G1 nuclei (see fig 1 ). With centromeric and painting probes, Nagele et al. [ 9 , 10 , 13 ] detected the presence of what they believed to be haploid sets of chromosomes in both diploid and triploid cells with the sets on opposite sides of the nucleus. In some cell shapes (e.g. elongated, polymorphic, or lenticular shaped cells) this regular order was obscured but in spherical nuclei it was mostly evident. Whereas there is often a spatial separation of the telomeric signals from the various homologues of the diploid or triploid G1-arrested cells in the present data (see fig 1 ) there was no evidence for a chromosome order or haploid groups in the cell nuclei (fig 2 ). In the explanation of Nagele et al. [ 13 ] the haploid sets these authors proposed represented maternal and paternal chromosome contributions. In the present data each set of identical homologues (two in diploid or three in triploid cells) appear to be arranged without respect to those of other chromosomes (fig 2 ), i.e. the spatial arrangement is not an interchromosomal phenomenon. This means the theoretical haploid sets of chromosomes described by Nagele et al. [ 9 , 10 , 13 ] probably do not exist. Figure 2 illustrates the two experiments performed in the current study to address the possible existence of haploid sets. These comprised examining the chromosome order for the single telomeres 4p (labelled with spectrum orange – Vysis, Downers Grove, Illinois), 18q (spectrum green), and 6p (spectrum green and spectrum orange, i.e. yellow signal) jointly hybridised to the same confluence arrested cells, and in a second experiment: 5p (spectrum orange label), 12q (spectrum green), and 20p (spectrum green and spectrum orange) hybridised to a second slide of triploid cell nuclei. These cells are from the same harvest as those shown to display the interphase topology of p-q intrachromosomal subtelomere tethering. In these latter results, homologous subtelomeres were regularly arranged without any defined relationship to non-homologous signal groups. This demonstrates that there is no interchromosomal order transferable between nuclei and challenges the concept of the presence of haploid sets within these non-cycling cells. In the view of Nagele et al. [ 10 ] the dual odd topology that he observed with (i) homologues arranged on opposite sides of the nuclei (diploid cells) or regularly arranged around the nucleus (triploid cells), and (ii) a chromosome order possibly manifesting as "haploid sets" may just be a relic of fertilization. Whereas, in our view, these intrachromosomal telomeric tetherings with the resulting looped chromatids must be important to normal cell function. Possible relationship of telomere tetherings to telomere captures As reviewed by Ballif et al. [ 23 ] there are two general pathways whereby chromosomes can acquire a new telomere and thus become stabilised. Firstly, by "telomere healing", i.e. the direct addition of telomeric repeats by: (i) telomerase [ 24 ] or by (ii) telomerase-independent recombination-based mechanisms [reviewed in [ 25 ]]. The second pathway is by "telomere capture" in which a chromosome acquires a telomere from another chromosome or chromosome end [reviewed in [ 23 ]]. Telomere captures are observed in two forms, those that are within the one homologue or intrachromosomal telomeric captures or transpositions [ 22 , 23 ], and those between non-homologues [ 26 ]. Ballif et al. [ 23 ] considered four different models for telomeric captures involving the p and q arms of a single homologue (intrachromosomal captures). These telomeric captures where the telomere from one chromosome arm is deleted and replaced by a telomere from the other arm of the homologous chromosome were termed intrachromosomal transpositions of telomeres [ 22 ] because of the uncertainty that simple reciprocal translocation was involved in this type of telomere capture. Ballif et al . [ 23 ] suggested that breakage induced replication (BIR), reviewed in Kolodner et al. [ 28 ], was the most likely model for these intrachromosomal captures based on their observation that there was observed heterozygosity between the duplicated ends on the one chromosome. This mechanism was initially described by Reddel et al. [ 27 ] under the unwieldy name "alternative lengthening of telomeres mechanism". Ballif et al. [ 23 ] suggested that BIR simply copied the sequence from the other end of the same homologue. Furthermore, that obligatory crossing-over during meiosis would mean that heterozygosity between duplicated ends would occur in many cases. The detection in the present study for the first time that in non-cycling mitotic cells in G1 most short-arm and long-arm telomeres from the one chromosome are tethered together is a likely staging point for mitotic events as per the fourth model of telomere capture reviewed in Ballif et al. [ 23 ]. This fourth model is that of the present authors in a prior study [ 22 ]. In the explanation of that fourth model by Ballif et al . [ 23 ], telomere capture was believed to occur by a pre-meiotic interhomologous exchange. The imbalanced chromosome was then generated through normal meiotic recombination. This (model) thus resulted firstly in a balanced translocation, termed telomere transposition by Daniel et al. [ 22 ] since reciprocal translocation was unproven. This translocation relocated the telomeres to the opposite chromosome arm and then by recombination the result was a duplication of one telomere on both chromosome ends and a deletion of the other. For this model to be correct a high frequency of balanced telomeric translocations would have to occur. These would be observed as large pericentric inversions and are rarely reported – see review in Daniel, 1988 [ 30 ]. However, the transposition of telomeres to opposite chromosome ends resulting in large pericentric inversions would not be easily noticed during FISH in many cases. This is in contrast to translocations between non-homologues which are very obvious to an observer in a FISH study. In this connection, for telomere translocations between non-homologous the rate of clinically ascertained balanced translocations has been reported as very high. Flint and Knight [ 26 ] record that for the several types of (non-homologous) telomeric rearrangements: unbalanced translocations account for 54% of cases; deletions for 39%; and duplications for 6%. According to Flint and Knight [ 26 ] in almost all cases unbalanced translocations occur because a parent carries the balanced form. When the data used to draw this conclusion are scrutinised, see De Vries et al. [ 29 ] it includes many rearrangements that are microscopically detectable, i.e. essentially regular reciprocal translocations. Such latter rearrangements are not really "telomere captures", are often familial, and would be expected to be associated with a high rate of balanced carriers. In our experience (Greg Peters and Luke St Heaps – CHW Telomere Study Group) we have not detected a balanced carrier of a telomere capture of either the intrachromosomal type or the interchromosomal type. In our view the issue of the frequency of balanced telomere rarrangements needs to be revisited since telomere captures are technically sub-microscopic telomere rearrangements. This data impinges on the likelihood that BIR is the preferred method of telomere capture [see that view expressed in ref [ 23 ]]. Since with the BIR model immediate recombinants are formed with no balanced carriers, if balanced (telomere capture) carriers are frequently reported, then BIR is ruled out as the common mechanism of telomere capture. This judgement currently cannot be performed without a more rigorous approach to the whole data set and additional assessment of the de novo or alternative origin of telomere rearrangements. Conclusions A topology for telomeres was detected where looped chromosomes were present at G1 interphase in confluence arrested cells. It was shown that the p and q telomeres of each chromosome in G1 cells associate frequently (80% pairing) in an intrachromosomal manner which was confirmed by studying chromosome domains with interstitial probes (chromosome arms) and WCP probes. It was found that homologues were regularly arranged without any defined relationship to non-homologous signal groups; i.e. there was no apparent chromosome order on different sides of the nuclei and no segregation into haploid chromosome sets was detected. The normal function of this high frequency of intrachromosomal telomeric pairings is unknown but a potential role is likely in the genesis of telomere captures whether of the intrachromosomal type or between non-homologues. In our view, these intrachromosomal telomeric tetherings with the resulting looped chromosomes arranged in a regular topology must be important to normal cell function since non-cycling cells in G1 are far from quiescent, are in fact metabolically active, and these cells represent the majority status since only a small proportion of cells are normally dividing. Materials and methods Cell culture Cell lines were retrieved from liquid nitrogen, washed in Dulbecco's phosphate buffer (DPB) and reconstituted in Hams F10 medium. The following lines were used: a diploid skin fibroblast line CG04-0743BBRS with karyotype 46,XX derived from fetal cartilage and a triploid 69,XXX cell line CG01-2042YA of diandric origin derived from mesodermal cells of a chorionic villus biopsy. These were cultured until they reached confluence via contact inhibition. At this stage the cells exhibit a number of swirls of closely packed cells in parallel. They were then severally prepared for trypsin harvest usually 48 hours after the last media change without colcemid/colchicine treatment and without the usual 0.075 M KCl hypotonic treatment. The cells were trypsinised off and fixed three times in 3:1 methanol to glacial acetic acid and were stored at room temperature (R.T.) in fixative for 1–3 days. This period allowed some mild acidic digestion of the chromatin and spreading of the nuclei when slides were prepared. At the end of the storage period, cells were rewashed once with fresh fixative and dropped onto glass slides as per routine techniques, and stored on trays in a low humidity cabinet until used for FISH. Choice of probe and probe label The identity of the probes used in the study is shown in Table 1 . The clones containing the DNA for the subtelomere probes were obtained from three sources: Incyte Genomics (Fremont, California); Dr Jonathan Flint (John Radcliffe Hospital, Oxford, U.K.), via Dr David Mowat, or the ATCC, (Manassas, Virginia). All were grown in Luria Broth (LB) with kanamycin by standard techniques unless specified otherwise (Table 1 ). Plasmid DNA was extracted with QIAGEN midi kits as per the manufacturer's instructions except that DNA elution was achieved at 60°C overnight. Probes were all labelled by nick translation (using VYSIS kit and the fluorophores spectrum orange and spectrum green, Vysis, Downers Grove, Illinois) as per the manufacturer's instructions. Fluorescence in situ hybridization (FISH) Slides were pretreated with a combined Pepsin/Rnase step. This was performed by prewarming RNAse and pepsin to 37°C, 200 μl of RNAse (0.1 mg in saline/sodium citrate – 2xSSC) was dispensed onto each slide, coverslipped and incubated at 37°C for 40 minutes in a humidified chamber. Coverslips were removed and slides washed twice for 5 minutes in 2xSSC at room temperature (RT). Slides were briefly drained and 200 μl of pepsin (0.2% in 0.01 M HCl) was placed on the slides, coverslipped and incubated at 37°C for 30 seconds. Coverslips were removed and slides were washed twice for five minutes in phosphate buffered saline (PBS) at RT. Fixation was with 6% paraformaldehyde in PBS, by dispensing 200 μl/slide, and adding a coverslip for 2 minutes at RT. Slides were then washed twice for five minutes in PBS at RT, dehydrated through 70, 90 and 100% ethanol for 3 minutes/wash at RT, and air dried. Probes in hybridisation mix were stored at -20°C, removed and thawed for 30 minutes; dispensed onto slides, covered with 15 mm diameter coverslips, and sealed with liquid rubber – art cement. Joint denaturation was achieved at 75°C for 5 minutes on a Omnigene hot plate, transferred to a humidified hybridization chamber at 37°C and hybridised overnight. After this the coverslips were removed. Post-hybridization washes were 0.4 SSC/0.3% NP40 at 73°C for 2 minutes then quickly transferred to 2xSSC/0.1% NP40 at RT for 1 minute. Slides were counterstained in DAPI and then rinsed and air dried. When ready, slides were mounted in antifade (2.3% DABCO in 40% glycerol/0.02 M TRIS-HCl) and covered until fluorescence examination. Slides were examined on a Zeiss Axioscop 20 fitted with a Zeiss fluoarc light source and images captured on an Applied Imaging Cytovision station using the false colours that are attributed by the software. Scoring of signal pairings to detect telomere tethering Initially, the subtelomere probes were labelled in Spectrum Orange for all short arms and Spectrum Green for all chromosome long arms. Cells were separately probed with the two subtelomeric probes for a single chromosome at the one time. Probe pairs were used for the subtelomeres of chromosomes 1, 3, 4, 5, 6, 7, 9, 10, 12, 17, 18, and 20. Images were captured for a large number of cell groups for each chromosome and pairings were scored on the captured images. Signals were interpreted as paired if the distance between signals was 10% or less of the greatest diameter of the nucleus (many cells were oval in shape). In addition to the above subtelomeric probe pairs, other probes were used to investigate the frequency of non-homologous tetherings (subtelomeric probes 1p and 9q, see Table 3 ) and the coincidence of the subtelomere tetherings and interphase chromosome domains (Fig 3 ). For the latter experiment the following Vysis probes were used: PML (promyelocytic leukemia locus) mapping to 15q22, SNRPN (small nuclear ribosomal protein locus) mapping to 15q12 – both labelled with spectrum orange; CEN15 (a probe for alpha centromeric sequences specific to chromosome 15) labelled with spectrum green. In addition, the chromosome 12 subtelomeric probes, i.e. 12p (labelled with spectrum orange) and 12q (spectrum green), and the WCP (Vysis whole chromosome painting probe) for chromosome 12 (spectrum green). Table 3 Rate of subtelomeric tetherings of non-homologues in G1 of non-cycling cells. Pairs of telomeres tested for tethering No (%) of signal pairs tethered 95% confidence limits 1p telomere; 9q telomere* 7/109 (6.4) 2.6–12.8% Note: This is a control for chromosome specific p-q subtelomere signal pair tethering (Fig 2, Table 2). For this experiment the 1p subtelomere was labelled with spectrum orange and the 9q subtelomere with spectrum green. Subtelomere tethering between these non-homologues was not increased over chance expectation. The telomere pair* 1p and 9q were chosen because they represent one of the most frequent telomere translocations reported (Lisa Shaffer, personal communication 2004). Up to 10% of signal associations in a two colour matrix can be regarded as random (VYSIS guidelines for interphase FISH scoring). There was no departure from randomness for tethering with respect to these two pairs of non-homologous subtelomeres. Additional experiments to detect a regular chromosome order reflecting the possible existence of haploid sets regularly arranged around the nuclei Two such experiments were performed in the current study (see fig 2 ). These comprised examining the chromosome order for the single subtelomeres (see Table 1 for clones) 4p (labelled in spectrum orange), 18q (spectrum green), and 6p (spectrum green and spectrum orange, i.e. yellow signal) jointly hybridised to the same triploid cells. In a second experiment subtelomeres were labelled as follows: 5p (spectrum orange), 12q (spectrum green), and 20p (spectrum green and spectrum orange, i.e. yellow signal) hybridised to a second slide of diploid/triploid cell nuclei. Authors' contributions AD designed the study, captured and analysed all FISH signals, and drafted the manuscript. LH performed all growing of probes, labelling of probes, and most probe hybridizations.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521695.xml
535353	Oligo kernels for datamining on biological sequences: a case study on prokaryotic translation initiation sites	Background Kernel-based learning algorithms are among the most advanced machine learning methods and have been successfully applied to a variety of sequence classification tasks within the field of bioinformatics. Conventional kernels utilized so far do not provide an easy interpretation of the learnt representations in terms of positional and compositional variability of the underlying biological signals. Results We propose a kernel-based approach to datamining on biological sequences. With our method it is possible to model and analyze positional variability of oligomers of any length in a natural way. On one hand this is achieved by mapping the sequences to an intuitive but high-dimensional feature space, well-suited for interpretation of the learnt models. On the other hand, by means of the kernel trick we can provide a general learning algorithm for that high-dimensional representation because all required statistics can be computed without performing an explicit feature space mapping of the sequences. By introducing a kernel parameter that controls the degree of position-dependency, our feature space representation can be tailored to the characteristics of the biological problem at hand. A regularized learning scheme enables application even to biological problems for which only small sets of example sequences are available. Our approach includes a visualization method for transparent representation of characteristic sequence features. Thereby importance of features can be measured in terms of discriminative strength with respect to classification of the underlying sequences. To demonstrate and validate our concept on a biochemically well-defined case, we analyze E. coli translation initiation sites in order to show that we can find biologically relevant signals. For that case, our results clearly show that the Shine-Dalgarno sequence is the most important signal upstream a start codon. The variability in position and composition we found for that signal is in accordance with previous biological knowledge. We also find evidence for signals downstream of the start codon, previously introduced as transcriptional enhancers. These signals are mainly characterized by occurrences of adenine in a region of about 4 nucleotides next to the start codon. Conclusions We showed that the oligo kernel can provide a valuable tool for the analysis of relevant signals in biological sequences. In the case of translation initiation sites we could clearly deduce the most discriminative motifs and their positional variation from example sequences. Attractive features of our approach are its flexibility with respect to oligomer length and position conservation. By means of these two parameters oligo kernels can easily be adapted to different biological problems.	Background During the last years, a large number of machine learning approaches have been developed to analyze and annotate genomic sequence data. Different concepts of statistical pattern analysis and modelling are successfully used for this purpose. Sophisticated methods like Hidden Markov Models [ 1 ], neural networks [ 2 ] or support vector machines (SVM) [ 3 ] became indispensable concepts and are routinely utilized in computational biology. Kernel-based learning algorithms, like SVM, are among the most advanced machine learning methods. As compared with probabilistic (Markov) models and classical neural networks, SVM provide a well-understood regularization [ 4 ] mechanism which makes learning from few examples in high-dimensional feature spaces possible. In that way, SVM and related methods can effectively cope with the "curse of dimensionality", which has been difficult for the more traditional tools in machine learning. Together with the kernel trick which provides a technical basis for learning in arbitrarily high-dimensional spaces, the principled approach to regularization has been the foundation for a large variety of successful applications of SVMs to realworld pattern-classification tasks [ 3 ]. However, a certain drawback of kernel methods is that, from a user's point of view, they usually behave like black boxes. Once the training phase is done, it is not easy to identify the features which actually determine the quality of the classification. This fact complicates the design of such systems and an interpretation of the learnt representation. Therefore additional feature selection procedures have been proposed to cope with that inherent shortcoming of traditional kernel methods. In that way, Degroeve et al . [ 5 ] proposed to combine SVM based on traditional kernels with techniques for feature selection in order to localize mononucleotide occurrences relevant for the prediction of splice sites. Herein, we introduce oligo kernels which provide a novel approach to datamining on biological sequences, based on the powerful concept of kernel feature spaces [ 6 ]. Our approach has two main advantages compared to previous applications of kernel methods to biological sequence analysis: ( a ) the existing methods provide either position-dependent representations based on mononucleotide occurrences [ 5 , 7 ] or they consider general K -mer occurrences restricted to a completely position-independent representation [ 8 , 9 ]. By contrast, with our method it is possible to adjust the required level of position-dependency to any degree for oligomers of any length. This convenient feature, in turn, facilitates the modelling of positional and compositional variability of complex biological signals. On the other hand, as we shall show in the next section, traditional monomer-based representations and position-independent K -mer representations may also be realized by the oligo kernel. ( b ) Furthermore, our method provides an intuitive visualization approach to present relevant oligomers and their positional variability to the user. For that purpose a suitable measure of discriminative power can be utilized to score different motifs according to their relevance for classification. Oligo kernels can therefore be applied to infer characteristic sequence features which, in turn, can be used to identify functionally important signals. This property qualifies the oligo kernel as a useful datamining tool for the analysis of biological sequences. To demonstrate the usefulness of our approach, we applied oligo kernels to analyze translational initiation sites (TIS) for Escherichia coli K-12 because a reliable set of biochemically verified sites is available for that case. Prediction of TIS is not satisfactorily solved for prokaryotic genes [ 10 ]. For bacteria, there are a number of gene-finding tools that reliably predict the location of genes in a genome under study. Essentially, these methods work by looking for open reading frames of a certain statistically significant minimal length. But while it is obvious how to identify the end position of a predicted gene, it is by no means trivial to determine the corresponding start position as start codons are not unique; they are also used to code for amino acids inside genes. Systematic studies have shown that existing gene-prediction programs perform poorly when it comes to predicting the correct TIS [ 11 , 12 ]. Consequently, many start positions are incorrectly annotated in databases and, due to the concepts used for gene annotation, these errors tend to be propagated to newly annotated genomes. Although our method is based on classification of sequences we do not intend to provide a TIS prediction tool in this paper. On one hand the method provides a general tool for identification and characterization of signals in biological sequences which is not restricted to TIS sequences. On the other hand our datamining approach may be viewed as a prior step for constructing TIS predictors. By means of an easy visual interpretation of the inferred model it can provide new insights which in turn can steer the construction of efficient predictors. We expect that this feature will in particular be useful for analysis of prokaryotic genomes for which only small sets of experimentally verified TIS exist. Oligo kernels Kernel-based learning [ 6 ] provides a powerful framework for many kinds of pattern analysis tasks usually encountered in statistical evaluation of experimental data. Given an input space a kernel is simply a function . It implicitly applies a usually nonlinear transformation to elements x , x' in input space and then computes the inner ("dot") product within a resulting feature space : k ( x , x' ) = Φ ( x )· Φ ( x' ). (1) A kernel can be viewed as a similarity measure for the input space objects x and x' which is defined as an inner product of the feature space objects Φ ( x ) and Φ ( x' ). The concept of kernel-induced feature spaces may first be motivated by the objective to provide an adequate representation for the data, which is more suitable for the classification task at hand than the original input space . Second, the so-called kernel trick suggests to perform learning in that feature space without any explicit computation of the mapping Φ , just using inner products of the feature space representatives. The inner products in turn may be computed by some realization of the above kernel function (1) in a hopefully efficient way. That trick may prove useful in situations where the feature space objects are rather high-dimensional and computation of Φ is costly. As will be shown below, these objects may even be functions, i.e. objects of infinite dimensionality. For construction of a linear classifier in feature space the discriminant for the two-class problem requires a feature space weight vector w ∈ and an scalar offset b . With an explicit feature space mapping of the input elements the primal form of the discriminant is f ( x ) = w · Φ ( x ) + b . (2) The sign of that discriminant can be used to assign x to one of the two classes, i.e. to realize a binary classifier. By means of the kernel trick the above primal form can be replaced by its dual representation, which does not require access to any feature space elements. The basis for constructing a kernel classifier is a training set of n labelled input space examples x 1 , x 2 ,..., x n . Using these examples, we can construct a linear discriminant in feature space without explicit computation of the feature space representatives. Instead we only compute inner products according to the above kernel function weighted by parameters α i in order to get the dual form of the discriminant: where we assumed b = 0 for convenience. The weights α i determine how much individual training examples contribute to the discriminant. Optimal values have to be determined by some suitable learning algorithm which will be outlined below. Unlike most SVM approaches which just propose a kernel and do not care about the primal form of the discriminant, i.e. about interpretability, we here first propose a suitable primal representation and then apply the kernel trick in order to provide a general learning scheme for that discriminant. By means of the primal/dual-concept of oligo functions and oligo kernels we achieve both: learnability and interpretability of the proposed model. For a suitable realization of the primal form, the idea is to represent oligomer ( k -mer) occurrences in terms of smooth functions which preserve positional information about oligomer locations to an adjustable extent. In order to model positional variability of the underlying biological signals we introduce some measure of positional uncertainty . The degree of positional uncertainty is controlled by some smoothing parameter σ which allows the model to be selected from a continuous space of candidates between the two extreme cases of completely position-dependent and position invariant models, respectively. An important advantage of the oligo kernel over previous position-dependent kernels is that, besides the dual representation of the feature space discriminant which is suitable for learning, it also provides a primal representation which is suitable for interpretation of the learnt discriminant. Therefore, classifiers based on oligo kernels not only provide a method for predicting signals in sequences but they may also help to identify relevant K -mers of any length and to analyze their positional variability. The two kinds of representations are detailed in the following. Primal representation: oligo functions For the primal representation of the feature space discriminant (2) we introduce the concept of oligo functions which encode occurrences of oligomers in sequences with an adjustable degree of positional uncertainty. For that purpose, occurrences are not represented by exact ("hard") assignment of the oligomers to their observed positions, but rather by some kind of "fuzzy" assignment according to the assumed positional uncertainty. Therefore oligo functions can be viewed as fuzzy membership functions [ 13 ], up to an arbitrary scaling, which in the case of fuzzy membership functions restricts the function values to the range [0,1]. In that way oligo functions assign K -mer occurrences to positions in a "soft" manner. For a convenient realization we choose these functions to be mixtures of Gaussians with the variance σ 2 of the Gaussians controlling the degree of positional uncertainty. Thus, for an alphabet and a sequence s which contains K -mer ω ∈ at positions S ω = { p 1 , p 2 ,...} we obtain the oligo function with continuous position variable t which does not need to be restricted to a discrete domain so far. The smoothing parameter σ adjusts the width of the Gaussians which are centered on the observed oligomer positions and therefore it determines the degree of position-dependency of the function-based feature space representation. While small values for σ imply peaky functions, large values imply flat functions with the limiting case σ → ∞ preserving no positional information at all. The effect of positional uncertainty is shown in figure 1 where the Gaussian bumps of two distinct oligomer occurrences result in a single bump of the corresponding oligo function. Consequently, for the sequence s the occurrences of all K -mers contained in = { ω 1 , ω 2 ,..., ω m } can be represented by a vector of m oligo functions which yields the final feature space representation of that sequence: Note, that the feature space objects are vector-valued functions, which can be stressed using the following notation: φ s ( t ) = [ μ 1 ( t ), μ 2 ( t ),..., μ m ( t )] T (6) where we used the shortcut . Thus, the feature space representatives are curves in the m -dimensional space of all K -mers. If we discretize the oligo functions according to the actual number of sequence positions, we would be able to assemble feature vectors of finite dimensionality, just stacking the vectors φ s ( t i ) of function values for all positions t i considered. Then, usual vector-based learning algorithms could be applied in order to infer relevant sequence characteristics. Unfortunately, this approach would only be feasible for short oligomers, i.e. for small K , because already for DNA sequences with l positions the number of required vector dimensions is l × 4 k while for protein sequences it is l × 20 K . In our application we used l = 200 positions and therefore we would have got nearly a million dimensions for representation of hexamer occurrences, which is prohibitive for usual vector-based learning schemes. Visualization While oligo functions in general cannot be used directly for learning of a feature space discriminant they are well-suited for the interpretation of a learnt discriminant. The latter fact is an important prerequisite for our datamining approach which provides an intuitive insight of the kernel-based sequence model. Considering the primal form of the feature space discriminant, the weight vector becomes a vector-valued function arising from a linear combination of the feature space representations of the sequences. With the learnt parameters α i we can construct the vector-valued weight function of the discriminant: which is a curve in the m -dimensional space of oligomers. For each of the m components we have a linear combination of oligo functions where the weights α i determine the contribution from each of the n training sequences. Due to its primal representation based on oligo functions, an oligo kernel classifier provides an intuitive insight into its discriminant because high positive (negative) values of the weight function contribute to prediction of the positive (negative) class. So the user can easily inspect the learnt representation which in turn allows him to draw conclusions about the relevance of certain oligomers and their locations in terms of their discriminative power. In order to facilitate the interpretation of the discriminant, one may restrict the analysis to the most discriminative oligomers. This can be done by ranking the component weight functions of w ( t ) = [ w 1 ( t ), w 2 ( t ),..., w m ( t )] T according to their L 2 -norm Because higher norms indicate a more important role in discrimination, selection of the corresponding weight functions helps to keep the focus on the relevant oligomers. A similar criterion has also been suggested for the position-independent spectrum kernel in order to identify discriminative motifs which can be used to distinguish real exons from pseudo exons [ 14 ]. Because the feature space weight vector can be represented as a vector of functions, a suitable visualization of these discriminative weight functions provides complete access to the information which has been extracted from the data. For an overview, all discriminative weight functions w i may be discretized and stored in a matrix which may be visualized as a bitmap image using grey values or color to encode the function values. Using l discrete sequence positions t i , the m × l image matrix can easily be obtained as W = [ w ( t 1 ), w ( t 2 ),..., w ( t l )]. (9) For longer oligomers only a subset of weight functions may be used, which can be selected according to the above L 2 -norm. For a closer look on the role of single oligomers the corresponding weight functions can be visualized as usual 2D function plots. Examples for both kinds of visualization will be shown in the section on results. Dual representation: oligo kernels In the previous section we have derived a feature space representation for our biological sequences, where a certain sequence s i from our dataset is represented by a vector of oligo functions, i.e. a curve in the m -dimensional space of all K -mers. While this feature space representation is well-suited for interpretation of the discriminant, it is impractical for learning the discriminant from sequence data. For that purpose we shall utilize the dual form of the discriminant, based on oligo kernels . By means of the kernel trick the dual representation is well-suited for training and any kernel-based learning algorithm [ 6 ] may be utilized. With the shortcut an inner product of two sequence representations φ i , φ j can be defined as Note that the second integral is a function of the distance d = \| p - q \| between two oligo positions p and q . This function I ( d ) is equivalent to the convolution of two Gaussians with equal variance. Up to some constant factor, the convolution of two Gaussians is a Gaussian whose variance is the sum of the original variances. Therefore the integral can be calculated according to Replacing the second integral in (10), the oligo kernel, i.e. the inner product of two feature space representatives, can be computed according to From the above definition of the oligo kernel, it is easy to see the effect of the smoothing parameter σ . For the limiting case σ → 0 with no positional uncertainty, only oligomers which occur at the same positions in both sequences contribute to the sum. In general it would not be appropriate to represent oligomer occurrences without positional uncertainty, which would imply zero similarity between two sequences if no K -mer appears at exactly the same position in both sequences. Regarding the other extreme σ → ∞ with maximum positional uncertainty, position-dependency of the kernel completely vanishes: all terms of oligomers, occurring in both sequences, contribute equally to the sum, regardless of their distance. Together with the normalization which is introduced below, in the latter position-independent case the oligo kernel becomes identical to the so-called spectrum kernel [ 8 ] which has been proposed for position-independent representation of sequences. Regarding computational complexity of the oligo kernel, for two sequences of length l 1 and l 2 , respectively, the above oligo kernel (12) can be computed by evaluation of at most l 1 × l 2 exponential functions. For a speed-up of the computations these evaluations may be realized by fast table lookups. Fortunately, the maximum number of l 1 × l 2 evaluations is hardly reached in practice, because we only have to compute terms of (12) for oligomers occurring in both sequences. Only in cases where oligomers often occur at many positions in both sequences the complexity is rising towards O ( l 1 × l 2 ). These cases become more unlikely for longer oligomers and therefore, with an efficient implementation, the computational cost rapidly decreases with increasing oligomer length K . Normalization According to the above derivation, the feature space representations may have different norms. Here, for the representation of biological sequences, the norm of a feature space object roughly corresponds to the absolute count of the oligomer occurrences. In order to improve comparability between sequences of different length, the feature vectors should be normalized to unit L 2 -norm. Therefore we compute the normalized oligo kernel according to Training The basis for learning a kernel classifier is a training set of n labelled example sequences = {( s 1 , y 1 ), ( s 2 , y 2 ),...,( s n , y n )} with y i = 1 or y i = -1 for positive and negative examples, respectively. Because we want to propose a general learning scheme which is even applicable to problems with small data sets, a regularized learning scheme is essential in order to prevent overfitting. Many regularized learning schemes are available, among which soft-margin SVMs [ 15 ] are usually best-known. For convenience, we here use regularized least squares classifiers [ 16 ], which have been shown to provide equivalent performance on many classification problems, as compared with SVMs. Implementation of these kernel classifiers is very simple because training can be achieved by solving a system of n linear equations, as shall be outlined below. Given the set containing n labelled sequences, with labels y i ∈ {-1,1} contained as components in n -vector y , we shall now train a kernel classifier based on discriminant (3) using the proposed oligo kernel. With the kernel matrix which contains all possible inner products of the training set examples in feature space, we realize a classifier by minimizing the λ -penalized prediction error with respect to parameter vector α = [ α 1 ,..., α n ] T . By means of the regularization parameter λ > 0 the penalty controls the norm of the feature space discriminant in order to avoid overfitting. Bounding the norm of the discriminant restricts the learning algorithm to put higher weights only on "effective" features which are important for classification. In that way learning is forced to focus on that task-specific information which can actually be drawn from the data. Choosing the parameter vector α to yield a minimum of the above error, an optimal realization can be found by solving the following system of linear equations: ( K + λ n I ) α = y (15) where I is the n × n identity matrix. Because solving the above system in general is of complexity O ( n 3 ) the computational cost of the learning algorithm is rapidly increasing for an increasing training set. In practice we observed that effective routines for matrix inversion enable training with a few thousands of examples. However, on one hand this behavior is a general shortcoming of kernel methods and recently several approximations have been suggested, which can effectively decrease the computational cost [ 17 , 18 ]. With these approximations training with a few 10000 of examples becomes feasible. On the other hand, for our case study on prokaryotic TIS we encountered data sets with at most 3500 exemplary TIS sequences which can even be handled with the above learning scheme. Results Datasets In order to create a reliable dataset we utilized E. coli genes from the EcoGene database [ 19 ] and considered only those entries with biochemically verified N-terminus. For training, validation and testing of classifiers for both the positive and negative examples we chose the sequences according to large windows of 200 nucleotides (nt) length around the candidate start codon, in order to encounter the risk of missing relevant information. The positive examples were 722 sequences covering a range of 100 nt upstream to 99 nt downstream of the annotated TIS. For the negative examples we extracted sequences centered around a codon from the set {ATG, GTG, TTG} and accepted sequences, if the codon was in-frame with one of the appropriate start sites used as positive case, if its distance was < 60 nt and if no in-frame stop codon occurred in between. The rationale for this approach comes from our analysis of predictions for the E. coli genome of those tools integrated into YACOP [ 12 ]: We found that nearly two-thirds of the predictions inconsistent with the annotated TIS were located within a distance less than 50 nt to its respective start codon (data not shown). Therefore we concluded that these false sites are the most difficult candidates for TIS discrimination. We finally obtained a set of 854 negative examples with 576 of them being located downstream and 278 upstream of a TIS. Prediction performance To analyze the TIS sequences by means of the proposed oligo kernel, first of all we tested the predictive power of the feature space representation as it depends on different oligomer lengths. In order to compare their discrimination performance, we trained classifiers using oligo kernels according to K -mers of length K = 1,...,6. For a reliable estimate of the prediction error we performed 50 runs for each oligomer length where each run comprises training, validation for adjusting the hyperparameters σ and λ and final testing on data which have not been used to adjust the parameters of the classifier. The latter was done on a test set which contained one third of the data. From the remaining data two thirds were used for training and one third for validation of the classifiers with respect to hyperparameter values. Thus, for each run the data were randomly partitioned into training, validation and test sets of size 631, 378 and 567, respectively. For validation we varied the smoothing parameter σ ∈ {0.5, 0.75, 1, 1.5, 2} and the regularization parameter λ ∈ {0.1·0.9 i \| i = 0, 1,..., 100} in order to minimize classification error on the validation set. With the optimal hyperparameter values we then trained a classifier on the union of training and validation set. Finally that classifier based on optimal hyperparameters was evaluated on the test set to yield the final classification error. The mean test error over the 50 runs together with the corresponding standard deviation is shown in table 1 . Additionally, table 1 also shows the mean optimal σ over the 50 runs. The table shows that the lowest error rate of 8.9 percent has been achieved for the 3-mer kernel with a mean value 1.25 of the hyperparameter σ which models the positional variability of the oligomers. This result indicates that the best representation does not necessarily require the lowest positional variability. Note that the mean error is monotonically increasing for decreasing oligomer lengths below and for increasing lengths above the best length 3. Therefore we did not try to model the occurrences of K -mers for K > 6. To investigate the effect of an enlarged data set, we chose additional TIS sequences according to GenBank annotations to yield a data set with about four times the number of examples of our original EcoGene-based set. We included all non-hypothetical coding sequences from the E. coli (U00096) dataset, including those from the EcoGene dataset, as positive examples. The negative examples were generated in the same way for the EcoGene-based set. So we obtained an enlarged set of 2980 positive and 3968 negative examples. The enlarged set was randomly partitioned, as described above, with the same proportions of training, validation and test sets. The average performance over 20 runs with different partitions is shown in table 2 . Unfortunately, the increased data set size resulted in a worse performance for all oligomer lengths, although an improvement would have been expected. Therefore we decided not to use the enlarged data set for further analysis because we could not exclude the possibility of a large number of erroneous annotations for TIS which had not been verified experimentally. Visualization For interpretation of the learnt TIS models, we applied the visualization techniques which have been described above. To visualize the feature space discriminant we first generated an overview bitmap image of the matrix W in equation (9) which contains the discretized functions of the primal form weight curve (7) as rows. The image contains the discriminative functions as horizontal lines, with one line for each oligomer. The color of a pixel indicates the level of the corresponding function at each integer position within a window of length 200 nt. Figure 2 shows a 64 × 200 pixel image which was obtained from an average over all optimal W -matrices from the above 50 runs using the trimer kernel. The complete matrix of function values is scaled to yield a unit maximum which is attained by the ATG-function at position 0. In addition, for noise reduction all matrix elements with an absolute value below 0.1 are set to zero. In figure 3 four exemplary weight functions for trimers ATG, GGA, AAA and TTT are depicted as 2D-plots showing more detailed information. The complete set of discriminative functions for K = 3 can be found on the web page [ 20 ]. Oligomer ranking In the following we use the terms monomer, dimer etc. instead of the more specific terms mononucletide dinucleotide etc. In order to identify the most important K -mers for the kernel-based TIS prediction, we computed the L 2 -norm in equation (8) for all oligomer-specific weight functions of the learnt discriminants. The ten most discriminative K -mers are identified using the average norm over the above 50 runs. The resulting rankings for K = 3,...,6 are depicted as bargraphs in figure 4 . The height of the bars is proportional to the average norm of the corresponding K -mer weight function and has been scaled to yield a unit maximum height for each K . For the monomer kernel we found that A is most discriminative followed by G, T, C in decreasing order of the norm. For dimer occurrences the most discriminative oligomer is GG followed by GA, AG, AT, TT and AA, again in decreasing order. For the longer K -mers depicted in figure 4 , from the bargraphs one can identify two major groups of motifs which are most prominent: on one hand the start codon itself is an important signal for TIS prediction. Therefore some motifs which contain ATG are associated with high norms and high positive values of the corresponding weight functions at position 0. This can be observed best for the discriminative weight function of trimer ATG itself (see figure 2 and 3 ). From these motifs and the associated weight functions we can conclude that C or T at position -1 and A at position 3 seem to be characteristic for TIS as they show high positive peaks at the corresponding positions (see also the above web page). From the pentamer and hexamer occurrences one can deduce a preference for AAA or GCT as a second codon. Start codons GTG and TTG are not characteristic for TIS and therefore the discriminative weight functions of these trimers show high negative peaks at position 0 which can be seen from the overview matrix image of figure 2 . On the other hand oligomers contained in AAGGAGA or GAGGAGA have high rank. The corresponding discriminative functions usually put high positive weights on regions located about 10 nt upstream the start codon (see GGA in figure 2 and 3 or GAG and AGG in figure 2 ). Obviously, these weight functions utilize the presence of a Shine-Dalgarno sequence for discrimination. Besides the two prominent groups of motifs related to start codon and Shine-Dalgarno sequence, interestingly, poly-A and poly-T motifs seem to be discriminative, too. As it can be seen from the discriminative functions for AAA and TTT in figure 3 , poly-A motifs mainly occur downstream next to the start codon while poly-T occurrences seem to be characteristic in a region ≈ 20 nt upstream. The corresponding discriminative functions for the ten most important oligomers for lengths 1,...,6 can be found on the before mentioned web page. Performance comparison In order to compare our approach with standard methods for modelling of sequence sites, we utilized inhomogeneous, i.e. position-dependent, Markov models which are widely used for the estimation of positional weight matrices. The probabilistic models can easily be estimated from the data and do not require any hyperparameter optimization. For evaluation we utilized one model for each of the two classes (positive/negative) and assigned a sequence to that class with highest probability of the corresponding model. In the case of zero probabilities for both models, as a tie-breaking rule we assigned the sequence to that model with the smallest number of zero probability positions. Classification was considered correct if the model with the smaller number was associated with the right class. For training we estimated the position-specific oligomer probabilities of the model using two thirds of the data while testing was performed on the remaining third. Prediction performance on the test set was again averaged over 50 runs with different random partitions of the data. For comparison with a position-independent kernel we utilized the spectrum kernel (SK) which can be obtained as a special case of the oligo kernel for σ → ∞. Therefore the spectrum kernel was evaluated in the same way as the above oligo kernels. The methods were compared with the best previous oligo kernel, i.e. the trimer kernel (OK 3 ), and with a combined oligo kernel (OK 1...6 ) which incorporates all K -mers for K = 1,...,6, simply by adding the six different kernels. With respect to the primal representation of the discriminant, adding the kernels means to stack the vectors of oligomer-specific weight functions for different K -mer lengths. Obviously this implies an augmented feature space which might combine the advantages of representations based on short and long oligomers. For each of the six added kernels we chose the smoothing parameter according to the median of the optimal values obtained from the 50 previous runs with single oligo kernels which resulted in values [0.5, 0.5, 1.0, 1.0, 1.0, 1.0] for the length-specific smoothing parameters [ σ 1 ,..., σ 6 ]. As can be seen from table 3 the position-dependent oligo kernels yield the best performance, while the best position-independent spectrum kernel with K -mer length 2 failed to discriminate TIS correctly for nearly half of the data. With an average error rate of 44.6%, performance was not much better than classification by chance, indicating the importance of position information for TIS prediction. With the combined kernel OK 1...6 our method could successfully exploit the combination of different K -mer representations and slightly improved the performance of the trimer kernel. In addition, table 3 shows that 0th and 1st order Markov models, based on monomer (MM 1 ) and dimer (MM 2 ) occurrences, respectively, performed equally well. We found that for the data at hand the 0th order Markov model could not be improved using higher order Markov models. Using Markov models above order 1 the performance even broke down, with a resulting error of ≈ 30%. Therefore in our case the Markov model based representations do not provide an adequate tool for analyzing occurrences of oligomers above length 2. As shown in table 3 even for dimers the oligo kernel provides better discrimination than the 1st order Markov model and therefore it may be preferred over the more simple model. Considering the occurrences of monomers the 0th order Markov model performs slightly better than the corresponding oligo kernel with K = 1. While theoretically the oligo kernel should not be worse, in practice it is difficult to find the optimal hyperparameters for best discrimination. Performing a two-dimensional grid search for smoothing parameter σ and regularization parameter λ , it cannot be expected to find the exact global optimum. On the other hand, for monomer occurrences there seems to be less positional uncertainty, so that the smoothing parameter is not likely to improve the representation anyway. Therefore the oligo kernel does not seem to be the best tool for pure analysis of monomer occurrences. Discussion We have introduced a novel concept of datamining on biological sequences and exemplified its application on the analysis of prokaryotic translational initiation sites. The interpretation of our results makes clear that the most pronounced signal indicating a TIS besides the start codon is the Shine-Dalgarno region [ 21 ]. With our approach we found oligomers contained in AAGGAGA or GAGGAGA to be most discriminative. The corresponding discriminative functions indicate that for characteristic TIS these oligomers are located ≈ 10 nt upstream the start codon. For trimers this can be seen in figure 2 and 3 . These results correspond to known findings both with regard to composition and to localization [ 21 , 22 ] and confirm the validity of our approach. The variation in spacing between the start codon and the Shine-Dalgarno region determined previously [ 21 ], correlates well with the wider peak of the discriminative functions (compare GGA with ATG peak in figure 3 ). In addition we observed some evidence for a downstream box, which was previously identified as an additional element modulating the expression level [ 23 - 25 ]. The evidence is a weak positive maximum of the discriminative function for AAA (see figure 3 ) downstream the TIS. The ranking of discriminative hexamer functions (see figure 4 ) also shows a preference for the codon AAA immediately following the start codon, as proposed in [ 25 ]. In addition also GCT seems to be a characteristic second codon as implied by pentamer and hexamer rankings in figure 4 . The analysis of individual oligo functions (see figure 3 ) makes it possible to identify subtle signals: The slight decay observed in a region ± 40 nt around position 0 in the plots of the discriminative weight functions for ATG and GGA is due to our selection of negative examples (see Datasets). The weak local minima of these functions seen upstream and downstream of position 0 are simply caused by shifted ATG and GGA sites of positive examples occurring also in our negative examples. The positive signal for TAT and CAT at position -1 (see figure 2 ) is due to the prominence of the oligomer ATG at position 0, but also indicates a preference for T and C at -1. This is also indicated by the tetramer ranking in figure 4 which identifies TATG and CATG as discriminative oligomers. Comparing our approach with other methods we found that position-dependency is crucial for an appropriate TIS representation. The position-independent spectrum kernel showed a bad discrimination performance which does not qualify that method for prediction or for analysis of TIS sequences. On the other hand the 0th order Markov model showed a competitive performance as compared with the monomer-based oligo kernel. Thus, for the analysis of monomer occurrences the oligo kernel seems to be "oversized" and we cannot recommend its application for merely that purpose. However, we argue that most biological signals are not well-characterized by monomer occurrences. In the living cell, sequence patterns for realization of relevant signals have to be selective and recognition of the these patterns is usually robust with respect to small variations. Therefore the observed patterns are more complex and show some specific variability. For that reason, in general a collection of longer oligomers, i.e. certain motifs should be more suitable for the modelling of biological signals. The oligo kernel can be used to identify these motifs and provides a tool for analyzing their positional variability. Unlike Markov model based representations, it is not restricted to short oligomers in practice, but may be used for K -mers of any length. By means of an effective regularization, characteristic occurrences of longer oligomers may even be found with small data sets, usually encountered in prokaryotic TIS analysis. As already mentioned above, computational cost is even decreasing for longer oligomers. We made clear that oligo functions are valuable to identify relevant signals in biological sequences. The presented dissection of TIS clearly identifies those areas bearing relevant information. These findings directly influence the design of TIS prediction tools e.g. with respect to the length of up- and downstream regions that have to be analyzed or the selection of discriminative oligomers whose occurrences have to be considered. Attractive features of our approach are its flexibility with respect to oligomer length and position conservation. Both parameters allow an easy adaptation to different biological problems. We conjecture an important role of the oligo kernel in computational biology: In addition to the application presented here, the kernel is well-suited to analyze and model splice sites, transcription factor binding sites or eukaryotic transcription initiation sites. Although we only presented its application to DNA sequences, also sites on protein sequences, like signal peptide cleavage sites are well suited for analysis by means of the oligo kernel. We are currently preparing a web-interface which will allow biologists to perform the analysis, as presented here, on their own sequence data. Conclusions We introduced the oligo kernel for datamining on biological sequences and we showed that it can provide a valuable tool for the identification and analysis of relevant signals. In the case of translation initiation sites (TIS) we could clearly deduce the most discriminative motifs and their positional variation from example sequences. These findings directly influence the design of tools for TIS prediction e.g. with respect to the length of up- and downstream regions that have to be analyzed. Attractive features of our approach are its flexibility with respect to oligomer length and position conservation. By means of these two parameters oligo kernels can easily be adapted to different biological problems. We showed that the position independent spectrum kernel can be viewed as a special case of the oligo kernel and that for the analysis of TIS sequences the incorporation of position information is crucial. In contrast to other position-dependent sequence kernels our approach not only provides learnability of a suitable model but also an easy interpretation of the learnt representation. Authors' contributions PM designed, implemented and tested the oligo kernel and the associated visualization method and he drafted most of the manuscript. RM accounted for biological expertise and substantial parts of the draft. MT contributed in biological expertise and prepared the datasets. BM assisted in coordination and manuscript writing. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535353.xml
539233	DNA-free RNA preparations from mycobacteria	Background To understand mycobacterial pathogenesis analysis of gene expression by quantification of RNA levels becomes increasingly important. However, current preparation methods yield mycobacterial RNA that is contaminated with chromosomal DNA. Results After sonication of RNA samples from Mycobacterium smegmatis genomic DNA is efficiently removed by DNaseI in contrast to untreated samples. Conclusions This procedure eliminates one of the most prevalent error sources in quantification of RNA levels in mycobacteria.	Background Mycobacterium tuberculosis is the leading cause of death from a single infectious disease with approximately 8.8 million new cases and two million deaths per year. To understand the pathogenesis of M. tuberculosis , analysis of gene expression by relative or absolute quantification of RNA levels using microarrays and RT-PCR (batch- and real-time) becomes increasingly important [ 1 ]. Widely used methods to isolate bacterial RNA are acid-phenol extraction or guanidinium isothiocyanate extraction combined with cesium chloride purification or nucleic acid binding resins [ 2 ]. However, the cell wall of mycobacteria is very stable and a very effective permeability barrier, and, therefore, rather refractory to lysis by chaotropic agents and detergents, hampering RNA isolation from these microorganisms [ 3 ]. Since the average half-life of mycobacterial mRNA is in the range of a few minutes, mycobacteria have to be vigorously treated (e.g. bead-beating, freeze-thawing, nitrogen decompression) to quickly isolate RNA [ 4 ]. This causes fragmentation of chromosomal DNA that contaminates RNA preparations, which is one of the most prevalent error sources in quantification of RNA levels in mycobacteria. Several methods have been suggested to circumvent this problem [ 5 , 6 ]. Virtually all RNA isolation protocols use DNaseI, which does not completely remove large amounts of DNA. Our goal was to improve the efficiency of DNaseI digestion by solubilizing chromosomal DNA with sonication prior to DNaseI treatment. Mycobacterium smegmatis is especially refractory to lysis and therefore was chosen as a model organism. Results and Methods M. smegmatis SMR5 [ 7 ] was grown in 10 ml Middlebrook 7H9 liquid medium (Difco Laboratories; supplemented with 0.2% glycerol, 0.05% Tween 80) to an OD 600 of 0.8 and mixed with 5 ml killing buffer (20 mM Tris-HCl, 5 mM MgCl 2 , 20 mM NaN 3 ) [ 8 ]. The cell suspension was incubated on ice for 5 min. Cells were harvested by centrifugation (20 min at 6000 × g and 4°C). 20 mg cells (dry weight) were lysed in FastRNA Blue-Tubes (Bio-101 Inc.) using a FastPrep FP120 bead-beater apparatus (Savant, USA) for 20 sec at level 6.5. The tubes were centrifuged for 10 min at 10000 × g and 4°C. The supernatant was transferred to microcentrifuge tubes containing a nucleic acid binding resin (Nucleospin RNA II; Macherey-Nagel), and further experimental steps were done as described by the manufacturer. A total of 62 μg RNA was eluted in 60 μl of RNase-free water. The RNA was diluted to 50 ng μl -1 into several aliquots. One aliquot containing 10 μg RNA was left untreated. The second aliquot was directly treated with 10U of RNase-free DNaseI (Roche) for 1 h at 37°C, while the third aliquot was sonicated two times for 20 sec with 0.9 sec intervals at 20 % power (Sonopuls HD 2070; Bandelin electronic) prior to DNaseI treatment. Between the two sonication steps the cell suspension was chilled on ice for 5 min. DNaseI was removed by precipitation with polyethylene glycol (PEG) 6000. As a control for the RNA quality, cDNA was synthesized by Omniscript reverse transcriptase and sensiscript reverse transcriptase (OneStep RT-PCR system, QIAGEN) from total RNA (100 ng) for 35 min at 50°C followed by an inactivation step of 15 min at 95°C. The 16Sr RNA was then amplified with the primers 16S-FP (5'-TGCTACAATGGCCGGTACAAA-3') and 16S-RP (5'-GCGATTACTAGCGACGCCGACTT-3') using up to 30 cycles of 1 min at 94°C, 30 sec at 53°C, and 1 min at 72°C before a final extension step of 7 min at 72°C. As a control for DNA contamination, standard PCRs were performed, to which RNA was added after the RT inactivation step. PCR products were analysed at cycles 23, 25, 27 or 30 to check the purity of the RNA. All samples apparently contained 16S rRNA (Fig. 1 ). However, DNA contamination was detected by PCR at cycle 25 in the RNA sample that was not treated with DNaseI. Conventional DNaseI treatment delayed the appearance of a signal in the sample without the RT step until cycle 27. By contrast, no amplification product was obtained in the control sample even after 30 cycles, when the RNA sample was sonicated before DNaseI treatment (Fig. 1 ). Conclusions This result shows that sonication improved DNA degradation by DNaseI most likely by rendering the chromosomal DNA more accessible to enzymatic action. This work describes a simple and efficient procedure to improve the quality of RNA preparations from M. smegmatis and will be of great value for RNA preparations from other microorganisms, including M. tuberculosis . Authors' contributions JS carried out the RNA preparations, the RT-PCR experiments and wrote a draft of the manuscript. JGB performed RNA preparations and RT-PCR experiments. FT participated in coordinating and supervising the study. MN conceived of the study, participated in coordinating and supervising the study, and wrote the final manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539233.xml
514606	Extrinsic factors regulate partial agonist efficacy of strychnine-sensitive glycine receptors	Background Strychnine-sensitive glycine receptors in many adult forebrain regions consist of alpha 2 + beta heteromeric channels. This subunit composition is distinct from the alpha 1 + beta channels found throughout the adult spinal cord. Unfortunately, the pharmacology of forebrain alpha 2 beta receptors are poorly defined compared to 'neonatal' alpha 2 homomeric channels or 'spinal' alpha 1 beta heteromers. In addition, the pharmacologic properties of native alpha 2 beta glycine receptors have been generally distinct from receptors produced by heterologous expression. To identify subtype-specific pharmacologic tools for the forebrain alpha 2 beta receptors, it is important to identify a heterologous expression system that closely resembles these native glycine-gated chloride channels. Results While exploring pharmacological properties of alpha 2 beta glycine receptors compared to alpha 2 -homomers, we found that distinct heterologous expression systems appeared to differentially influence partial agonist pharmacology. The β-amino acid taurine possessed 30–50% efficacy for alpha 2 -containing receptor isoforms when expressed in HEK 293 cells. However, taurine efficacy was dramatically reduced in L-cell fibroblasts. Similar results were obtained for β-alanine. The efficacy of these partial agonists was also strongly reduced by the beta subunit. There were no significant differences in apparent strychnine affinity values calculated from concentration-response data between expression systems or subunit combinations. Nor did relative levels of expression correlate with partial agonist efficacy when compared within or between several different expression systems. Finally, disruption of the tubulin cytoskeleton reduced the efficacy of partial agonists in a subunit-dependent, but system-independent, fashion. Conclusions Our results suggest that different heterologous expression systems can dramatically influence the agonist pharmacology of strychnine-sensitive glycine receptors. In the systems examine here, these effects are independent of both absolute expression level and any system-related alterations in the agonist binding site. We conclude that complex interactions between receptor composition and extrinsic factors may play a significant role in determining strychnine-sensitive glycine receptor partial agonist pharmacology.	Background It has been well established that the amygdala is important in the acquisition and maintenance of fear/anxiety-related behaviors [ 1 ]. Strychnine-sensitive glycine receptors have recently been found in the adult rat basolateral amygdala (BLA) using whole cell and intracellular electrophysiology [ 2 , 3 ]. Reverse transcription polymerase chain reaction on whole BLA tissue and single cells revealed a prominent expression of α 2 mRNA; and these receptors are likely to be α 2 β heteromers due to their low picrotoxin sensitivity [ 4 ]. This finding is consistent with prominent BLA 'general' immunoreactivity for α/β subunit protein but no apparent α 1 -specific protein expression [ 3 ]. A similar enrichment of α 2 /β heteromers is also evident in striatal cholinergic interneurons [ 5 ]. It is quite possible then that the α 2 β strychnine-sensitive glycine receptors present in the adult BLA and other forebrain areas represents a receptor population that could be functionally distinguished from those found in the spinal cord. Because the BLA regulates a number of anxiety- or fear-related behaviors [ 6 ], it is possible that this population of strychnine-sensitive glycine receptors may represent a novel therapeutic target for anxiety disorders. To insure that novel α 2 β compounds possess an appropriate therapeutic index, the pharmacology of these forebrain glycine receptors must be elucidated and extensively compared with the spinal isoform. There have been conflicting reports regarding the details of glycine receptor pharmacology when expressed in heterologous systems. For example, taurine acts as a partial agonists (ca. 50% efficacy compared to glycine) for GlyRα 1 expressed in Xenopus ooctyes [ 7 ] whereas it shows nearly full agonist efficacy for GlyRα 1 expressed in HEK 293 cells [ 8 ]. Compared to GlyRα 1 , taurine efficacy is even weaker for GlyRα 2 (ca. 5–10% efficacy) when expressed in Xenopus oocytes [ 7 ]. However, native GlyRα 2 β receptors expressed by BLA neurons possess >50% efficacy for taurine and almost full efficacy for β-alanine [ 2 ]. While these results might initially be dismissed as expression system-dependent phenomena, brain region-specific effects are also evident in the literature. Taurine has markedly different efficacies at glycine receptors expressed by isolated adult lateral/basolateral amygdala neurons [ 2 ], adult hypothalamic magnocellular neurons [ 9 ], and juvenile spinal cord neurons [ 10 ]. It is therefore possible that the mechanisms regulating brain region-specific effects are related to those governing the divergence among heterologous expression systems. However, such mechanisms have not been systematically investigated, despite their potential usefulness in understanding region-to-region pharmacologic heterogeneity evident for some native receptors. This study utilizes whole-cell patch clamp electrophysiology to examine the influence of distinct heterologous expression systems on the β-amino acid pharmacology of glycine receptors composed of distinct subunit combinations. We have focused on the α 2 and α 2 β receptors since these appear to be the predominate isoforms found in the embryonic and adult forebrain, respectively. Our results provide potentially important insight into the types of mechanisms that may govern brain region-to-brain region variation in glycine receptor pharmacology. Several aspects of this work have appeared in abstract form [ 11 , 12 ]. Results Subunit- and system-dependent effects on glycine pharmacology Given the variation of glycine receptor partial agonist pharmacology in the literature, we specifically sought to identify any role that expression system may play in their pharmacological profiles. First, glycine concentration-response relationships were established for GlyRα 2 , and GluRα 2 /β in HEK-293 cells and in L-cell fibroblasts. Glycine-gated responses for each receptor isoform were elicited in a dose-dependent manner in both cell types (Fig. 1A ). The apparent EC 50 of glycine HEK cells was 221 μM and 269 μM for α 2 (n = 4–6) and α 2 β (n = 7–8), respectively. GlyR subunits expressed in L-cells displayed a similar pharmacological profile. However, the apparent glycine EC 50 of both GlyRα 2 (446 μM, n = 5–7) and GlyRα 2 β (667 μM, n = 4–8) appeared lower than apparent affinities for the same subunits when expressed in L-cells. Two-way ANOVA on the Log (EC 50 ) values (Table 1 ) indicated a significant effect of system (F = 20.01, P < 0.001). However, the presence of the β-subunit did not significantly affect glycine apparent affinity in either system nor was there a significant interaction between system and subunit composition. These results indicate that glycine is less potent for receptors expressed in L cells compared to HEK cells. Figure 1 Glycine receptors expressed show expression-system dependent agonist pharmacology. (A) Glycine has a reduced potency in L-cells compared to HEK cells. Glycine current responses were plotted versus the log concentration of glycine and normalized to a maximal concentration of glycine (3–10 mM). Data are presented as mean ± SEM; 4 ≤ n ≤ 9 cells for each concentration. Concentration response relationships for GlyRα 2 (□; EC 50 = 221 μM) and GlyRα 2 β (○; 269 μM) in HEK cells and in L-cells (GlyRα 2 , ■, 446 μM; GlyRα 2 β, ●; 667 μM) were derived from logistic equation fits to individual cells. (B) β-alanine has both reduced apparent affinity and efficacy for most glycine receptor isoforms transiently expressed in L-cells compared to HEK 293 cells. β-alanine apparent potency in HEK 293 cells was 717 μM for GlyRα 2 (□, n = 6) and 560 μM for GlyRα 2 β (○, n = 7). For L-cells, β-alanine potency for GlyRα 2 (■, n = 5–8) was 1.61 mM and 1.79 mM for GlyRα 2 β (●, n = 7). Current responses were normalized to a maximal concentration of glycine (10 mM). Note the reduced apparent efficacy of α 2 β receptors compared to the α 2 homomeric isoforms. (C) Taurine has both reduced apparent affinity and efficacy to GlyRs transiently expressed in L-cells compared to HEK 293 cells. Taurine concentration-response relationship in HEK 293 for GlyRα 2 (□; 442 μM, n = 5–6) and GlyRα 2 β (○; 1.25 mM, n = 3–5). Taurine concentration-response relationship in L-cells yielded GlyRα 2 (■, n = 4) and GlyRα 2 β (●, n = 7) potencies estimated at ≥ 3 mM. Current responses were plotted versus the log concentration of taurine and normalized to a maximal concentration of glycine. Table 1 Agonist Pharmacology in HEK and L-cells. GlyR α2 GlyR α 2 β System log EC 50 EC 50 (mM) Efficacy a log EC 50 EC 50 (mM) Efficacy Glycine HEK -3.67 ± 0.15 0.22 -- -3.55 ± 0.06 0.28 -- L-cells -3.36 ± 0.05 0.43 -- -3.17 ± 0.04 0.67 -- β-Alanine HEK -3.12 ± 0.05 0.76 0.80 ± 0.06 -3.24 ± 0.03 0.57 0.55 ± 0.07 L-cells -2.71 ± 0.08 b 1.93 0.39 ± 0.08 -2.59 ± 0.13 2.59 0.25 ± 0.05 Taurine HEK -3.13 ± 0.12 0.74 0.48 ± 0.12 -2.67 ± 0.12 2.20 0.32 ± 0.05 L-cells ≥-3.00 ± 0.11 c 1.00 0.06 ± 0.01 ≥-2.97 ± 0.20 c 1.10 0.05 ± 0.01 a – efficacy relative to maximal glycine response b – p < 0.0001 for system but not subunit using two-way ANOVA c – estimate due to low efficacy Subunit- and system-dependent effects on β-Alanine pharmacology With the glycine pharmacological profile established, we next examined the pharmacology of the partial agonists, β-alanine and taurine. The efficacy and potency of these β-amino acids were compared to glycine by normalizing the current response at each concentration to a maximal glycine response in that same cell. In HEK cells (Fig. 1B ), the average β-alanine EC 50 values calculated from individual cells were 770 μM (n = 6) and 570 μM (n = 7) for α 2 and α 2 /β receptors, respectively. In L-cells, β-alanine also elicited currents in a dose-dependent manner. And, like glycine, β-alanine appeared to be less potent in these cells compared to HEK cells. EC 50 values for α 2 and α 2 /β receptors were 2.0 mM (n = 8) and 2.9 mM (n = 7), respectively. Two-way ANOVA on LogEC 50 values from these studies indicate a significant effect of the expression system on β-alanine potency (F = 43.52, P < 0.0001). There was a trend for the presence of the β-subunit to influence potency but this was not significant nor was there any significant interaction between expression system and subunit composition. β-alanine efficacy was also examined in these same experiments by normalizing the maximal β-alanine response as a fraction of a maximal glycine response. In HEK cells, the α 2 and α 2 β isoforms had efficacies of 80 ± 6% and 55 ± 7% of the maximal glycine response, respectively. A similar trend was noted in L-cells with the α 2 and α 2 β isoform with β-alanine efficacies being 39 ± 8% and 25 ± 5% of the maximal glycine response. Two-way ANOVA analysis of these data indicate that both expression system and subunit composition had a significant influence on β-alanine efficacy (F = 27.6, P < 0.0001 and F = 7.9, P < 0.01 respectively). There was no significant interaction between these variables. These data demonstrate that the presence of the β subunit reduced β-alanine efficacy of α 2 -containing receptors and that this efficacy was substantially smaller L-cells compared to HEK cells. Subunit- and system-dependent effects on taurine pharmacology Similar analysis of taurine pharmacology in HEK and L-cells revealed more dramatic effects of system and subunit on this partial agonist (Fig. 1C ). In HEK cells, the apparent EC 50 for taurine was 501 μM for GlyRα 2 (n = 9) and 2 mM for GlyRα 2 β (n = 7). Because of its remarkably low efficacy in L-cells (see below), we can only provide estimates of taurine potency in this expression system. Regardless, apparent taurine affinity for both GlyRα 2 and GlyRα 2 β expressed in L cells were ~3 mM for both isoforms (n = 4 and 7, respectively). We did not compare L cell data with that obtained from HEK cells due to the uncertainty surrounding the fits. However, there was no significant difference in apparent taurine potency between the α 2 and α 2 β receptors expressed in HEK cells (P >> 0.05, t-test). Taurine efficacy was obviously quite different between the two expression systems. In HEK cells, taurine efficacy was 48 ± 12% of glycine for GlyRα 2 and 32 ± 4% of glycine for the GlyRα 2 β isoform. Efficacy for these same receptors was reduced to approximately 6 ± 1% and 5 ± 0.7% of glycine when they were expressed in L-cells in these particular studies. The system difference was significant with two-way ANOVA (F = 17.4, P < 0.001) with no substantial effects of subunit composition or interactions between these variables. Expression level and system-dependent pharmacology The preceding results suggest that there may be a complex interaction between subunit composition and the expression system in which the receptor is produced. Specifically, the system-dependent agonist pharmacology could be related to differences in the relative expression levels between various systems. Expression level has clearly been demonstrated to influence agonist pharmacology for G protein-coupled receptors (e.g. [ 13 ]), where the levels of G-protein bound to receptor and thus the relative levels of high affinity receptor can vary from system to system. However, the influence of expression level on ligand-gated channel function has not been extensively explored (see Discussion). Unfortunately, it is problematic to compare expression levels between HEK and L-cells since the relative efficiency of transfection varied widely between these systems. Indeed, liposome-mediated transfection is remarkably efficient in HEK 293 cells (70–90% of cells based on GFP fluorescence) but only marginally effective in L-cells (10–20% of cells, not shown). To get around these differences in transfection efficiency, we examined the relative expression level of GlyRα 2 protein using western analysis of total lysate derived from the same number of GFP + HEK 293 or L cells from transfected cultures (Fig. 2A ). For this experiment, cells were harvested under native conditions, GFP + cells were counted, and volumes of lysate corresponding to equivalent numbers of GFP + cells was loaded onto the gel. Western blots from two separate experiments demonstrate that transfected HEK 293 cells expressed 4- to 5-fold more GlyRα 2 protein than transfected L-cells. The mean optical density from the two experiments was 83 ± 2 units for HEK cells and 17 ± 2 units for L-cells. Figure 2 Relative expression levels does not influence taurine efficacy. (A) HEK and L-cells were co-transfected with the GlyRα 2 subunit and GFP. Relative expression levels of α 2 -protein were examined using western blot analysis of total lysate from equal numbers of GFP + HEK and L-cells. α 2 protein was 4- to 5-fold greater in GFP + HEK cells than in GFP + L-cells. (B) Maximal glycine conductance across all experiments was significantly lower in L-cells compared to HEK cells, although only by about 2-fold. This may indicate that a significant amount of α 2 protein in HEK cells (A) is present in a non-functional form or not associated with the plasma membrane. (C) Glycine current density vs. taurine efficacy in different cell lines expressing α 2 (open symbols), α 2 β glycine receptors (closed symbols), and in isolated neurons from the adult rat basolateral amygdala. The correlation coefficient between glycine current density and taurine efficacy was 0.14 and was not significantly greater than zero (P >> 0.05). There was also no correlation (R 2 = 0.01 to 0.3) between glycine current density and taurine efficacy when comparing individual cells within each of these systems. Maximal conductance is an independent measure of functional expression and was also larger for both α 2 and α 2 β receptors expressed in HEK cells compared to receptors expressed in L cells (Fig. 2B ). Across all experiments where maximal glycine concentrations were assayed, the conductance of α 2 receptors expressed in L-cells was 65 ± 11 nS and was 114 ± 21 nS in HEK cells. Similarly, L-cells expressed α 2 β receptors at 42 ± 9 nS while HEK cells expressed this isoform at 104 ± 17 nS. Two-way ANOVA using subunit and system as variables revealed a significant effect of system (F = 15.4, P < 0.001) but not subunit, nor was there a significant interaction between variables. Results from both westerns and functional experiments therefore indicate that relative expression levels of glycine receptor were different between HEK and L-cells. To further explore the interaction between expression level and partial agonist efficacy, both current density and taurine efficacy were compared for α 2 and α 2 β glycine receptors in a number of different heterologous systems, as well as for native receptors expressed in rat lateral/basolateral amygdala. In addition to HEK and L-cells, the heterologous systems included mouse 3T3 fibroblasts and MDCK kidney cells. α 2 β receptors expressed in mouse 3T3 fibroblasts had twice the current density (121 ± 34 pA/pF) of the mouse L-cells (59 ± 19 pA/pF) but had a similar taurine efficacy (13 ± 8% of glycine in 3T3 cells versus 8 ± 1% in L-cells). Similarly, α 2 β receptors expressed in HEK293 cells had a current density similar to GlyRs expressed in 3T3 fibroblasts (115 ± 11 pA/pF) but had a taurine efficacy compared to glycine of 48 ± 3%. This efficacy was similar to glycine receptors expressed by acutely isolated adult rat basolateral amygdala neurons (46 ± 5% of glycine) although the current density in this native system was only 57 ± 14 pA/pF. Note that the channels expressed by these neurons are composed primarily of α 2 +β subunits [ 4 ]. Canine kidney MDCK cells expressed the lowest α 2 β current density (15 ± 5 pA/pF); yet the channels expressed by this system had the highest taurine efficacy of any cell tested (101+6%). For α 2 GlyRs, the rank order of glycine receptor density was 3T3 (111 ± 21 pA/pF)> HEK cell (90 ± 11 pA/pF)> L-cell (50 ± 9 pA/pF); while the rank order of taurine efficacy for these same receptors was HEK (74 ± 9%)> 3T3 (23+7%)> L-cells (11 ± 2%). Across all subunit combinations and systems, there was no significant correlation (R 2 = 0.14, P >> 0.05) between taurine efficacy and glycine current density (Fig. 5C). Indeed, no correlation between expression level and taurine efficacy was evident within any given population of cells whether the receptors were expressed in native or heterologous systems. For example, the correlation coefficients for α 2 β receptors between taurine efficacy and glycine current density in individual systems were 0.11, 0.13, 0.19, 0.05, and 0.31 for HEK, L-cells, 3T3 cells, MDCK cells, and amygdala neurons, respectively (P >> 0.05). Thus, while there is clearly a difference in expression level between both the systems as well as between individual cells in a given system, this particular characteristic cannot account for the apparent taurine efficacy. The agonist-binding site is not affected by expression system There are a variety of possible mechanisms to account for the disparities in partial agonist pharmacology between two expression systems. One way to address this is to examine competitive antagonist binding properties in HEK and L-cells. We therefore examined the potency of the glycine receptor competitive antagonist strychnine in both systems. Following a 30 second pretreatment with the antagonist [ 2 ], we co-applied strychnine and an EC 50 concentration of glycine. The strychnine K B was estimated for HEK and L-cells expressing either the GlyRα 2 or GlyRα 2 +β subunits using the Cheng-Prusoff relationship (see Methods). This relationship takes into account the divergent Hill-slope and potencies for glycine found in these two expression systems. Receptors composed of the GlyRα 2 subunit (Fig. 4A ) had very similar K B values when expressed in either HEK (K B = 49 ± 8 nM, n = 11) or L-cells (K B = 38 ± 7 nM, n = 16; Fig. 4C ). The same was true for cells expressing the GlyRα 2 β subunits where strychnine apparent affinity was 32 ± 7 nM (n = 10) in HEK cells and 38 ± 8 nM (n = 10) in L-cells. Two-way ANOVA did not reveal any significant effect of either system or subunit composition. Since strychnine is a competitive antagonist and site-directed mutation studies suggests that strychnine and glycine interact with overlapping regions of the receptor [ 14 - 16 ], our results strongly suggest that functional strychnine affinity, and hence the general structure of the agonist binding pocket, was not substantially influence by expression system. Figure 4 The association of glycine receptors with the tubulin-cytoskeleton may influence partial agonist efficacy. (A) Tubulin depolymerization with colchicine decreased both taurine and β-alanine efficacy of α 2 β glycine receptors expressed in HEK 293 cells. Cells were treated with 100 μM colchicine or γ-lumicolchicine at 37°C for 30 minutes. The graph shows the partial agonist efficacy as a fraction of the maximal glycine response. For taurine (□), colchicine treatment reduced apparent efficacy from 33 ± 6% in control cells (n = 8) to 13 ± 3% in treated cells (n = 10). γ-lumicolchicine, an inactive analogue of colchicine, had no effect on taurine efficacy (29 ± 9%, n = 5, ** – P < 0.01 from ANOVA). For β-alanine (■), efficacy was reduced from 70 ± 7% in vehicle-treated cells (n = 8) or 67 ± 8% in γ-lumicolchicine-treated cells (n = 5) to 49 ± 6% in cholchicine-treated cells (n = 10, * – P < 0.05, ANOVA). (B) Colchicine treatment does not influence partial agonist efficacy of the GlyRα 2 homomeric channels. Taurine (□) efficacy was 34 ± 15% in control GlyRα 2 cells (n = 4) and was 38 ± 10% in colchicine-treated cells (n = 5, P >> 0.05 t-test). Similarly, β-alanine efficacy was 72 ± 12% and 86 ± 8% in the same control and treated cells, respectively (P >> 0.05, t-test). (C) Colchicine treatment decreases β-alanine efficacy in L-cells expressing GlyRα 2 β heteromeric channels (■), but not those expressing GlyRα 2 homomeric channels (□). For the GlyRα 2 β channels, colchicine treatment significantly reduced efficacy from 23 ± 2% (n = 7) to 12 ± 3% (n = 3, P < 0.05 t-test). (D) Gephyrin-like immunoreactivity was detected in both cells lines using 20 and 40 μg of whole cell lysate. * – denotes expected gephyrin mobility (approx. 100 kD). Cytoskeletal components influence glycine receptor pharmacology A third possible mechanism for reduced efficacy in L-cells compared to HEK cells or neurons could be related to intracellular factors that influence channel gating [ 17 ]. This hypothesis was examined by disrupting the cytoskeletal protein tubulin, which has been shown to be important for glycine receptor localization [ 18 ]. Direct application of 100 μM colchicine did not elicit any membrane currents. Furthermore, acute application of 100 μM colchicine and an EC 50 concentration of glycine (300 μM) did not significantly affect glycine-gated currents themselves. Glycine currents were 17.3 ± 2.3 pA/pF while glycine+colchicine currents were 16.7 ± 2.2 pA/pF (p > 0.5, paired two-tail t-test, n = 7). The relative efficacy of β-alanine and taurine was examined in HEK cells expressing α 2 β subunits following 30 min incubation with 100 μM colchicine at 37°C, enough time to allow irreversible tubulin disruption [ 19 ]. As an additional control, γ-lumicolchicine, an inactive analog of colchicine [ 20 ], was also used to treat α 2 β-expressing HEK cells (Fig 4A ). These brief treatments had no obvious effect on the survival of untransfected cells. There was a trend for colchicine treatment to reduce the overall current density at 300 μM glycine, 56.7 ± 9.1 pA/pF in control cells (n = 8), 41.2 ± 2.3 pA/pF in colchicine-treated cells (n = 10), and 50.1 ± 9.9 pA/pF (n = 6) in γ-lumicolchicine-treated cells; however, this was not significant (p > 0.05, ANOVA) and was probably not related to any direct action of colchicine given that the glycine current density was also slightly reduced in α 2 β-expressing cells exposed to γ-lumicolchicine compared to controls. However, the efficacy of both taurine (p < 0.01, One-way ANOVA) and β-alanine (p < 0.05, ANOVA) were significantly decreased by colchicine but not γ-lumicolchicine treatment. Taurine efficacy was 33 ± 6% of glycine in controls, 13 ± 3% following colchicine, and 28 ± 3% following γ-lumicolchicine. Similarly, β-alanine efficacy was 70 ± 7% of glycine in controls, 49 ± 6% following colchicine, and 72 ± 7% following γ-lumicolchicine. Similar treatment of α 2 -expressing HEK cells with colchicine (Fig. 4B ) did not reveal any significant effect on glycine current density (54 ± 14 pA/pF in controls, 60 ± 15 pA/pF in treated), on taurine efficacy (34 ± 16% in controls vs. 38 ± 10% in treated), or on β-alanine efficacy (71 ± 12% in controls vs. 86 ± 8% in treated). Cholchicine treatment also significantly reduced β-alanine efficacy in L-cells expressing GlyRα 2 β (23 ± 2% in controls vs. 12 ± 3% in treated, P < 0.05, t-test) but not in GlyRα 2 -expressing L-cells (Fig. 4C ). We did not attempt to examine taurine in L-cells treated with colchicine given the exceptionally low efficacy of receptors expressed in this cell line. Because the glycine receptor- and tubulin-binding protein gephyrin provides an obvious link between the receptor and the tubulin cytoskeleton, we used western analysis of HEK and L-cell lysates with a gephyrin monoclonal antibody specific for the C-terminus. These experiments revealed that gephyrin-like immunoreactivity was expressed in both expression systems (Fig. 4D ). Notably, a ca. 100 kD band dominated the HEK cell gephyrin immunoreactivity, while multiple bands of varying intensity could be seen in lysate from L-cells. When taken with our colchicine data, differences in glycine receptor pharmacology between α 2 β receptors expressed in HEK and L-cells may be partially due to distinct, system-dependent interactions with distinct isoforms of the cytoskeletal protein gephyrin. Discussion We have expressed several the 'embryonic' (α 2 homomeric) and 'forebrain' (α 2 β heteromeric) isoforms in two distinct expression systems to understand the influence of endogenous and exogenous factors on receptor partial agonist pharmacology. Although the pharmacology of the 'embryonic' GlyRα 2 isoform and the 'adult spinal' isoform (GlyRα 1 β) have been explored more frequently in the literature, the pharmacology of GlyRα 2 β receptors has remained largely unexplored. Despite this, there is strong evidence that the adult 'forebrain' isoforms, specifically in the rat basolateral amygdala, is indeed α 2 β [ 4 ]. The current study indicates a general trend for decreased apparent affinity and reduced relative efficacy of agonists when receptors consist of the α 2 β subunits compared to their homomeric α 2 counter parts. We were particularly surprised to find that receptors expressed in different expression systems possessed markedly different partial agonist efficacies. The remainder of our study focused on identifying extrinsic factors that influence difference in ligand-gated receptor pharmacology in distinct expression systems. While differences in efficiency of cDNA expression/transfection between systems could explain such differences, the expression levels of GlyRα 2 β receptors measured by current density was not correlated with taurine efficacy across several different cell types or within any given system. Importantly, the efficacy of β-alanine and taurine in HEK cells agree with previous findings where cells expressing GlyRα 2 show almost full efficacy for taurine and β-alanine [ 21 ]. Similarly, distinct ligand binding characteristics of receptors expressed in different expression systems seemed to be another possible mechanism governing agonist efficacy or potency. For the glycine receptor, the binding site for the competitive antagonist strychnine is believed to be adjacent to the agonist-binding site, sterically hindering agonist binding. A gross alteration in the agonist binding pocket, particularly one that hindered agonist binding, would most likely affect strychnine binding as well. To examine this, strychnine K B was calculated for GlyRα 2 and GlyRα 2 β isoforms expressed in both HEK and L-cells. In order to decrease the error in estimating K B , a derivation of the Cheng-Prusoff equation was used that takes in account variations in the slopes of the inhibition curves [ 22 ]. Differences in K B were negligible between expression systems, indicating the strychnine binding site, and presumably the agonist binding site, was altogether similar in these different systems. Differences in pharmacology between systems therefore cannot be explained by substantial alterations in the agonist/competitive antagonist binding pocket. Receptor gating is another mechanism by which receptor function may be altered. Cytoskeletal elements have been shown to play a crucial role in neurotransmitter receptor clustering [ 17 ] and may have a role in receptor function as well. For example, cytoskeletal stabilization has been shown to reduce Ca ++ -dependent inactivation of Ca ++ channels in snail ganglia [ 23 ]; and, actin has been shown to modulate several different types of membrane ion channel [ 24 - 26 ]. Cytoskeletal depolymerization has also been found to inhibit the function of GABA A receptors, which share significant sequence homology and functional characteristics with strychnine-sensitive glycine receptors [ 27 ]. And the tubulin-gephyrin-glycine receptor interaction is critical for establishing functional glycinergic synapses [ 28 ]. The current study suggests that cytoskeletal elements may play a functional role in α 2 β glycine receptor pharmacology as well. β-containing glycine receptors are intimately associated with the tubulin-associated protein gephyrin [ 29 ] via the gephyrin binding site that lies within the intracellular domain of this subunit [ 30 ]. In our studies, the efficacy of both taurine and β-alanine were reduced in cells expressing GlyRα 2 β subunits. This despite the finding that gephyrin-like immunoreactivity in L-cells was apparently distinct from that in HEK cells, suggesting that distinct cytoskeletal components in these systems may have profound influence over GlyR α 2 β pharmacology. This is further supported by suggestions that gephyrin may exist in multiple, tissue-specific isoforms with potentially distinct functional roles [ 31 - 33 ]. It should be noted however that colchicine treatment of α 2 β-expressing HEK cells did not suppress partial agonist efficacy to a level that approached that found in L-cells or 3T3-fibroblasts. Given that colchicine had no perceptible effect of α 2 -homomeric channels expressed in HEK cells, our results suggest that additional system-dependent factors may have a more pronounced influence on the partial agonist pharmacology of strychnine-sensitive glycine receptors. Conclusions It is of particular interest that the β subunit appears to play a functional role in the pharmacology of α 2 -containing receptors regardless of expression system. For example, the beta subunit decreased the apparent efficacy of the partial agonists. Since the β-subunit itself does not appreciably interact with the competitive antagonist strychnine [ 34 ], these results are at least consistent with some allosteric interaction between the β-subunit and the agonist binding site present the α subunit. This may indicate that α 2 β glycine receptors in the forebrain may be distinguishable from other receptor isoforms given the appropriate pharmacologic agent. Although cytoskeletal components potentially play some role for these 'forebrain' receptors, there appear to be other 'extrinsic' factors governing expression system-dependent effects on agonist pharmacology. Since it is conceivable that such factors may be differentially distributed between different forebrain regions, the large apparent differences between glycine receptor pharmacology reported by various studies may not necessarily depend upon differential expression of glycine receptor subunits per se. At the very least, our findings suggest that great care should be taken when utilizing different expression systems to develop screens for novel pharmacophores acting on this receptor. Methods Cell culture and transfection HEK 293 (TSA 201; gift from Michael J. Davis, Dept. Med. Physiol., Texas A&M Univ. Health Science Center, College Station, USA), mouse L-cells (NCTC-929; American Type Culture Collection), NIH/3T3 fibroblasts (ATCC), and MDCK cells (gift from Alan Parrish, Dept. Med. Pharmacol. & Toxicol., Texas A&M Univ. Health Sci. Center) were grown in Dulbecco's modified Eagle's medium (DMEM, SIGMA) with 10% fetal bovine serum (HyClone Laboratories, Logan, UT, USA) and 1X penicillin/streptomycin (Life Technologies) on Thermanox cover slips in 35 mm culture dishes. Cells were transfected during log-phase growth (30–50% confluent) using the Superfect reagent (Qiagen, Valencia, CA, USA), according to manufacturers instructions. Rat glycine receptor α 2 and β subunits were cloned previously [ 4 ]. Briefly, a total of 2–3 μg of plasmid constructs was added to 100 μL serum-free media along with 10 μL Superfect reagent, vortexed for several seconds, and incubated for 10 minutes to allow DNA/liposome formation. Standard media (600 μL) was added and this mixture applied the cells and incubated for 2–3 hours at 37°C. The cells were subsequently washed twice with phosphate-buffered saline, fresh media was applied and cells were then incubated for 24–48 hours prior to recording. Cells were co-transfected with green fluorescent protein (pEGFP-C1, Clontech, Pal Alto, CA, USA) to identify transfected cells before recording. Mass ratios of 1:1:5 were used in the GFP:α 2 :β transfections. An equal mass of the cloning vector pCI (Promega Corp) replaced the β subunit when examining α 2 homomeric channels. Electrophysiology Whole cell recordings were performed at room temperature using standard patch-clamp techniques and the axopatch-1D amplifier (Axon Instruments, Inc., Foster City CA, USA) in the voltage clamp mode. Gigaohm seals were formed using patch pipettes made from borosilicate glass (World Precision Instruments, Sarasota FL, USA). For most experiments, the internal solution contained (in mM): CsCl 100, EGTA 11, HEPES 10, CaCl 2 1, Mg-ATP 4, pH 7.2 with methane sulfonic acid; adjusted to 290–295 mmol kg -1 with sucrose. Whole cell capacitance and series resistance was manually compensated after opening the cell. Cells were continuously bath perfused with a HEPES-buffered saline (in mM): NaCl 150, Glucose 10, HEPES 10, KCl 2.5, CaCl 2 2.5 MgCl 2 1.0, pH 7.4, 305–320 mmol kg -1 Data will be analyzed off-line using pClamp software (Axon). Numerical analysis was performed using commercially available software. Independent student's t-test and two-way ANOVA were used for comparisons where appropriate; and statistical significance was based on p < 0.05. Concentration-response curves were generated from fits of data to a standard logistic equation as previously described (McCool & Botting, 2000). To derive K B from functional IC 50 and EC 50 data, the Cheng-Prusoff equation was used: where S = hillslope of agonist curve, A = concentration of agonist, EC 50 = half-maximal agonist concentration, and IC 50 = half-maximal antagonist concentration. Drugs Stocks of glycine, taurine, β-alanine, strychnine (Tocris) and colchicine (SIGMA) were prepared fresh each day. Agonists and antagonists were applied for 4–10 sec from an array of eight HPLC-grade capillary tubes (150 μm i.d.; Hewlett Packard Analytical Direct) placed within 100 μm of the cell of interest. Western analysis Cells were cultured, transfected as stated above except in 10 cm Petri dishes, and harvested by scrapping with 500 μL of lysis buffer (10 mM Tris pH 7.5, 1% SDS). Proteins were quantified using the Bradford assay and loaded onto 8–10% SDS-polyacrylamide gels, separated, and transferred to a nitrocellulose membrane (Hybond C, Amersham). The membrane was blocked overnight TBS (200 mM NaCl, 10 mM Tris pH 7.5) containing 0.2% Tween-20 and 10% low-fat dry milk. Blots were then incubated in TBS/0.1% Tween-20 containing primary GlyR antibody (monoclonal GlyR4a antibody, 1:200, Alexis Biochemicals) and monoclonal anti-gephyrin antibody (1:2000, Transduction Laboratories) for two hours at room temperature. After several washes, the HRP-coupled rabbit anti-mouse secondary antibody (SIGMA) was added for one hour (1:2000). The detection was performed by the ECL method (Amersham). List of abbreviations GlyR – glycine receptor; BLA – lateral/basolateral amygdala; HEK – human embryonic kidney cells; MDCK – Manin-Darby canine kidney cells Author's contributions JF carried out the electrophysiological recordings and participated in the western analysis. BM conceived of the study, participated in its design and coordination, performed the western analysis and some electrophysiology experiments, and drafted the manuscript. All authors read and approved the final manuscript. Figure 3 Glycine receptors expressed in different expression systems have similar 'functional' strychnine binding. Cells were pre-treated for 30 seconds with strychnine alone then exposed to a strychnine admixture with an EC 50 concentration of glycine. (A) Strychnine-mediated inhibition of glycine-gated currents of GlyRα 2 homomers expressed in HEK 293 (□; IC 50 = 78.2 ± 13.5 nM) and L-cells (■; 33.1 ± 6.3 nM). (B) Strychnine-mediated inhibition of glycine-gated currents of GlyRα 2 β heteromers expressed in HEK 293 (○; 37.9 ± 7.9 nM) and L-cells (●; 23.2 ± 4.6 nM). (C) Functional K B values were calculated from IC 50 values for individual cells using the Cheng-Prusoff relationship and glycine affinity/Hillslope data represented in Figure 1. Average K B values are shown for each subunit combination in the two expression systems. There was no significant effect of subunit compositions or expression system.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514606.xml
549515	The epidemiology of college alcohol and gambling policies	Background This article reports the first national assessment of patterns of drinking and gambling-related rulemaking on college campuses (e.g., punitive versus recovery oriented). Analyses relating school policies to known school rates of drinking or gambling identified potentially influential policies. These results can inform and encourage the development of guidelines, or "best practices," upon which schools can base future policy. Methods The college policy information was collected from handbooks, Web sites and supplemental materials of 119 scientifically selected colleges included in the fourth (2001) Harvard School of Public Health College Alcohol Study (CAS). A coding instrument of 40 items measured the scope and focus of school alcohol and gambling policies. This instrument included items to measure the presence of specific policies and establish whether the policies were punitive or rehabilitative. A total of 11 coders followed a process of information extraction, coding and arbitration used successfully in other published studies to codify policy information. Results Although all schools had a student alcohol use policy, only 26 schools (22%) had a gambling policy. Punitive and restrictive alcohol policies were most prevalent; recovery-oriented policies were present at fewer than 30% of schools. Certain alcohol and gambling policies had significant relationships with student binge drinking rates. Conclusions The relative lack of college recovery-oriented policies suggests that schools might be overlooking the value of rehabilitative measures in reducing addictive behaviors among students. Since there are few college gambling-related policies, schools might be missing an opportunity to inform students about the dangers of excessive gambling.	Background Young people are at increased risk for alcohol- and gambling-related problems compared to their older counterparts [ 1 - 3 ]. College and university students are at special risk because going to college often represents the first move away from their family and, as a result, fewer restrictions on their activities. (Because universities are by definition comprised of colleges, all institutions of higher learning henceforth will be referred to as "colleges.") In the United States, each year approximately 1.2 million freshmen enter four-year colleges [ 4 ]. Some of these freshmen enter college actively involved in recovery programs for alcohol abuse or other addictive behaviors (e.g., illicit drug abuse or gambling). Others will begin a program of recovery for addiction problems that started after they enrolled at school. The college years are a time of developmental transition for most students; like other life transitions, the college experience can be associated with increased risk for a variety of psychosocial problems. The problems associated with addictive behaviors on college campuses have been well documented (e.g., academic difficulties, psychosocial problems, traumatic injuries, overdoses, high-risk sexual behavior, and impaired driving) (e.g., Wechsler et al. 2000 [ 5 ], Wechsler et al. 2002 [ 6 ]). Despite a recent increase in college-based preventative measures (e.g., alcohol education programs, advertising restrictions, alcohol-free dormitories, policy controls), research reveals that addiction-related problems continue to plague college campuses. For example, during the past decade, past-year alcohol use and binge drinking rates have remained steady at approximately 81% and 44%, respectively [ 6 ], and alcohol-related problems have been on the rise. Wechsler et al. (2002 [ 6 ]) found that a greater percentage of students who had used alcohol in the past 30 days were involved in police-related incidents in 2001 than in 1993 (6.5% vs. 4.6%); the same was true of alcohol-related injuries (12.8% vs. 9.3%). Wechsler et al. (2002) also identified a significant increase in the rate of students riding in motor vehicles with alcohol-impaired drivers in 2001 compared to 1993 (23.2% vs. 18.4%). These findings highlight the need for college administrators to reconsider current preventative measures and develop and implement more effective methods for preventing and reducing alcohol use. For example, college health programs might be able to limit or reduce alcohol-related harms on college campuses by implementing and enforcing policies that support recovery-oriented and other programs that discourage substance misuse. The creation and implementation of college alcohol and gambling policies is far from an exact science. Currently, there are no standardized scientific guidelines for the creation of school policy directed toward alcohol and other potentially addictive behaviors (e.g., gambling). However, science can contribute to the creation of successful policy. Recognizing the important role that science can play in the development and evaluation of public policy, the federal government recently released draft "regulatory science" guidelines [ 7 ]. The Office of Management and Budget (OMB) intends these guidelines to direct and inform public agencies in the creation and implementation of effective and targeted regulations. Science-based guidelines also could prove useful to policymaking on college campuses; however, as the results of this study will reveal, college administrators do not use empirical evidence to guide the development and implementation of student substance use and gambling regulations. This situation has led to disjunctive policy strategies among U.S. colleges. The purpose of this study is to encourage the development of science guided school policy. To accomplish this goal, we will examine the prevalence and characteristics of alcohol- and gambling-related policies, including policy provisions for student recovery, in a scientifically selected sample of U.S. colleges. We will not include illicit drug policies in this analysis because illicit drug use is illegal for both adults and young people; these illegal behaviors fall under the purview of state and federal law that supersedes college policy. Our intent is to examine college policies that focus on legal activities. Therefore, using college alcohol and gambling policies, binge drinking rates and gambling frequency as evidence, this report describes the epidemiology (e.g., prevalence) and influence of these assorted policies. Filling the Policy Void: A Federal Drug and Alcohol Initiative During 1989, the federal government initiated basic alcohol and substance abuse education requirements. Previously, there was not a regulatory mandate obligating institutions of higher learning to set alcohol or drug use policy or bring students' attention to these rules if they existed. Schools also were not required to disseminate substance use policy information to parents or other interested parties. This situation changed with the passage of the federal Drug-Free Schools and Communities Act (DFSCA) of 1989. The DFSCA applies to all U.S. colleges. The act specifies that "as a condition of receiving funds or any other form of financial assistance under any Federal program, an institution of higher education (IHE) must certify that it has adopted and implemented a drug [and alcohol] prevention program..."[ 8 ]. Thus, any U.S. college that does not maintain a drug and alcohol education program risks losing all of its federal funding. In addition, to fulfil DFSCA requirements and retain funding, schools must provide students with institutional standards of conduct that explicitly prohibit illicit drugs and illegal alcohol use, a description of potential legal and institutional sanctions for substance use violations, a description of health risks posed by drugs and alcohol, and a listing of available treatment options. The Impact of Government Policy on College Campus Substance Use and Abuse is Unknown The overall impact of mandated drug and alcohol programs is still unknown; as we noted before, there is some evidence that risky and addictive behaviors on college campuses are still prevalent despite targeted efforts by administrators to reduce student substance abuse [ 9 , 6 ]. Several studies have suggested that, despite prevention efforts, established norms of excessive drinking behavior and positive student attitudes regarding the effects of alcohol consumption continue to encourage alcohol consumption on college campuses [ 10 , 11 ]. The absence of universal standards governing the content of school policies on addiction might contribute to this problem. Although the DFSCA mandates that schools must make written drug and alcohol policy available to students on an annual basis, administrators at each institution still determine the content of such policy. Thus, the DFSCA mandates policy without establishing standards for content; as a result, administrative tolerance toward alcohol, drugs, and gambling can vary significantly from institution to institution. The Potential Effect of Inconsistent College Policies Inconsistent policy content among institutions can create a problematic state of affairs. Although DFSCA directives aim to increase awareness of the potential dangers of alcohol and drug use among students, numerous studies continue to identify high levels of alcohol abuse on U.S. college campuses in recent years [ 9 , 12 - 14 , 6 ]. Heavy episodic drinking adversely affects not only those students who actively participate, but also those who do not: one study identified non-heavy drinkers on heavy drinking campuses as 3.6 times as likely to experience at least one problem from another student's drinking as non-heavy drinkers on non-heavy drinking campuses [ 15 ]. Even though individual colleges have adopted different strategies for reducing the problems associated with excessive alcohol consumption, the extent and effect of these efforts are largely unknown. One approach, perhaps in response to DFSCA, has been to develop and enforce policies on student substance abuse and recovery. Although recent psychosocial programs attempting to reduce student drinking behaviors have failed to reduce binge drinking [ 6 ], official school policies on substance abuse and recovery hold the potential to reduce students' alcohol use and the multitude of consequential problems associated with drinking excessively. This potential, however, is likely contingent upon policy content: because there are few federal regulations governing the content of alcohol policies at institutions of higher learning, every college develops unique strategies of combating potentially addictive behaviors. To date, no studies have examined the policy content of a representative sample of colleges in the attempt to identify the effects of these policies on levels of alcohol and gambling involvement among students. Policy and Recovery Students who seek help for alcohol or other substance use problems are faced with a multitude of school-provided and external treatment options. Addiction recovery programs are diverse, ranging from formal treatment programs (e.g., inpatient medical treatment and outpatient psychotherapy) to less formal self-help options, (e.g., 12-step fellowships) [ 16 ]. Regardless of the selected type of treatment, attention to recovery from addiction requires significant time and determination, which can disrupt a schedule of college studies. Twelve-step programs, for example, usually involve attending regular, perhaps even daily meetings. Formal treatment programs frequently demand an even greater level of time commitment: in-patient detoxification or other residential care can remove students from the academic environment altogether. Mandatory abstinence, required by most treatment programs, poses an additional hurdle to treatment-seekers. Students, with their busy and often stressful schedules, undoubtedly face additional challenges in participating in recovery activities; academic and administrative policies that accommodate flexible scheduling will likely assist students seeking recovery, and policies that do not might complicate or inhibit students' recovery efforts. College Binge Drinking and School Policy Binge drinking, the consumption of five or more alcoholic drinks (four or more for women) on at least one occasion at one to two week intervals [ 12 ], has been unaffected by prohibitive and punitive college policies. To illustrate, on one college campus that prohibited all alcohol use in its residence halls, there was virtually no difference in the binge drinking rate among students living within areas regulated by the alcohol policy (35%) compared to those living outside the jurisdiction of the alcohol policy (34%) [ 17 ]. Although school policy (or the lack thereof) is not the only factor that affects binge drinking rates – promotions aimed at students, cheap alcohol prices at surrounding establishments and high numbers of on- and off-campus drinking venues have been found to significantly increase student binge drinking [ 18 ] – placing special emphasis on the enforcement of substance abuse policies can garner positive results. For example, Knight (2003)[ 19 ] found that, although the effect of policy was diluted by considerable variation in policy content among public colleges in a state-wide system, increased enforcement (i.e., application of policy consequences) of alcohol policies aimed at combating underage drinking did result in decreased alcohol consumption among students. Beneficial effects resulting from the enforcement of existing rules, however, can be difficult to interpret. For example, in that study, it is unclear whether the enforcement of rules encouraged lower levels of drinking or entry to treatment for intemperate drinking or, alternatively, simply forced problematic drinkers to withdraw from school. College Gambling and School Policy Some research suggests that gambling on college campuses is commonplace. A study of student gambling at six colleges in five different states (i.e., New York, New Jersey, Nevada, Oklahoma, and Texas) showed that of 1,771 surveyed students, 23% reported that they gambled at least weekly (ranging from 11% in Texas to 39% in Nevada) [ 20 ]. In that study, students reported whether they had ever experienced gambling-related problems as identified by the South Oaks Gambling Screen (SOGS) [ 21 ]. Of the total student sample, 5.5% were classified as lifetime probable pathological gamblers. The prevalence of lifetime pathological gamblers among these students ranged from 4% in Nevada to 8% in New York. A recent report [ 22 ] of a four-campus Connecticut college system reported a similar SOGS-based prevalence estimate of probable pathological gamblers (i.e., 5.2%). For comparison, the National Gambling Impact Study Commission (NGISC) considered the adult rates of lifetime pathological gambling from four sources [ 3 ]. The lowest rates were 0.8% for both the University of Michigan [ 23 ] and National Opinion Research Center [ 24 ] studies; the largest (i.e., 1.5% – 1.6%) were from aggregated statistics of previously published research conducted by the National Research Council [ 25 ] and the original analysis of the same studies by the Harvard Medical School [ 26 ]. This meta-analysis included 14 SOGS-based studies of disordered gambling among college students and indicated that the lifetime prevalence of pathological gambling among college students was 5.1% [ 26 ]. An update of this meta-analysis expanded the number of student studies to 19 and increased the prevalence estimate to 5.6% with a 95% confidence interval of 3.5% to 7.6% [ 1 ]. Based on 66 studies of the general household population in various areas (i.e., states), this estimate of the proportion of college students with gambling disorders was three times the adult rate (1.9%). Other research contradicts the findings that college students are at elevated risk for problem gambling compared to the general adult population. For example, a recently published longitudinal study of students at the University of Missouri-Columbia showed markedly lower prevalence rates than the studies summarized above [ 27 ]. In this longitudinal study, no student met the traditional criteria for problem or pathological gambling. Further, the authors note that, "there were too few participants endorsing multiple gambling problems at a single time point to obtain an adequate sample size of affected individuals for most analyses" (Slutske et al. 2003[ 27 ] p. 265). Overall, 3% of these students endorsed a single problem at any point during their lifetime due to gambling; one student endorsed two problems and all of the others reported never having had a problem due to gambling. At the next interview three years later, when most subjects were seniors, the subjects reported more symptoms; but only one subject (i.e., 0.2% of the sample) endorsed enough symptoms to meet the diagnostic criteria of the American Psychiatric Association [ 28 ] for lifetime pathological gambling. This evidence indicates that gambling behavior among students and its adverse consequences fluctuates with time and other factors and that the development of symptoms is not always progressive. Further, the Slutske results show that most adverse effects of student gambling remain sub-clinical, making this pattern more responsive to interventions than longer standing, more entrenched clinical disorders. Taken together, this evidence suggests that comprehensive college gambling policies might have the capacity to reduce the adverse consequences that can be associated with student gambling. Despite the frequency with which college students engage in gambling activities, some evidence suggests that administrators are unaware of the dangers associated with excessive gambling among students; in addition, colleges do not have adequate policies addressing gambling [ 29 ]. This situation prompted Shaffer to suggest that the government convene "a consortium of college presidents to review their existing gambling related policies and problems so that we can take a systematic approach to the education, prevention and treatment of America's young people, who are at higher risk for gambling related disorders than their adult counterparts"[ 30 ]. Although this consortium has not yet been assembled, research confirms that college students continue to view gambling as a legitimate form of entertainment; for example, 42% of a scientifically selected sample having gambled at least once in the last year [ 31 ]. Unlike drug and alcohol education (i.e., DFSCA), there is no federal mandate requiring schools to educate students or parents about the dangers of excessive gambling; combined with the lack of a policy response by administrators, this situation leaves an open door for student-related gambling disorders to emerge unchecked. Assessing the Relationships between College Policies and Student Drinking and Gambling This study is the first to identify patterns of drinking and gambling-related rulemaking on college campuses (e.g., punitive versus recovery oriented). By relating school policies to known school rates of drinking or gambling [ 31 , 6 ] we can identify potentially influential policies. These analyses can encourage and inform the development of guidelines, or "best practices," upon which schools can base future policy. Hypotheses Given the paucity of empirical college-based policy research, this study will fill an important gap in knowledge. To fill this void, this research will test a variety of addiction-related hypotheses that have not yet been examined empirically. Based upon the extant literature, this study will test the following four primary hypotheses: • Because there are few requirements guiding the creation of school substance use and gambling policies, the content and clarity of these policies will be heterogeneous across schools and modes of policy distribution (e.g., handbooks vs. school Web sites); • College alcohol policies currently devote relatively little attention to student recovery; • Due to differences in enforcement, awareness of the dangers of excessive alcohol consumption, educational programs and types of students, schools with either no or only restrictive alcohol use policies will experience higher levels of binge drinking among students than schools with prohibitive and recovery-oriented alcohol policies; • Absent a federal mandate that requires gambling-related regulations or education on college campuses, gambling policies will be less prevalent than alcohol use policies. Methods Procedure: Sample, Policy Eligibility and Policy Selection The purpose of this study is to identify and assess alcohol and gambling policies among U.S. colleges. To ensure a representative national sample of colleges, we examined the scientifically selected sample of public and private American colleges that was used in a recent series of Harvard studies (e.g., Wechsler 2002 [ 6 ]). The detailed methods by which the previous study identified the sample are available elsewhere [ 6 , 12 - 14 , 31 ]. The potential sample consisted of 120 scientifically selected schools located throughout the nation; one school ceased operation before the start of the study, so 119 schools were eligible to be included in the final sample. We received human subjects approval for this study through the Harvard Medical School Office for Research Subject Protection. On February 14, 2003, the Human Subjects Committee at Harvard Medical School granted an exemption for the study entitled: United States College and University Addiction and Recovery Policies. The study qualified for exemption under 46 CFR §102(f) and the assurance identification number is M1240-01. At the beginning of the project, we submitted an e-mail request for a hard copy of their student handbook to each school's admissions office. Each e-mail specified that we were interested in collecting school alcohol and gambling policies and requested that our inquiry be forwarded to the most appropriate school official. We gathered e-mail addresses for admissions offices from each school's official Web site. Using each school's main telephone number to initiate contact, investigators contacted schools that did not respond within thirty days to our e-mail request and verbally requested a handbook and any other existing alcohol and gambling policy materials. Typically, the person answering the call referred us to admissions offices, deans' offices, or student services offices for further assistance; we identified ourselves as calling from Harvard Medical School only when asked. Policy Eligibility and Identification Eligibility Criteria To be eligible for inclusion in this study, each college policy had to meet the following five eligibility criteria: 1. the policy had to prohibit, govern, or otherwise attempt to regulate alcohol use or gambling among students at a U.S. college or university; 2. the policy had to be in effect (i.e., in the current handbook, Web site or supplementary materials); 3. the policy had to be readily available to the public, either in electronic or hard copy; 4. the policy had to be written in English; 5. the policy had to be available for review by project investigators no later than July 31, 2003. Identifying Policy Our primary source of alcohol and gambling policies was each school's student handbook. (For the purpose of this study, "student handbook" refers to the institution's primary informational document made available to current and prospective students.) The student handbook is a centralized forum for regulatory information and is a primary source of official school policies for students and parents, as well as the public. In addition, the concept of a student handbook is widespread, making handbooks a common information source across many schools. Many institutions distribute student handbooks to all incoming freshmen; therefore, most students are familiar and comfortable with accessing the handbook. Student handbooks also are widely available to the public. When available, we used electronic versions (i.e., pdf or html) of each school's handbook; otherwise, we used a hard copy. Some schools, particularly large universities with many departments and/or divisions, did not have a single handbook that they distributed to all students. In these cases, we retrieved the school's policies from other official documents (e.g., code of conduct, policy manual, judicial procedures manual). Many schools also posted policy information (i.e., separate from the handbook) on their Web sites; we analyzed this information as a secondary source. We conducted an exhaustive search of each school's Web site using each site's integrated search engine and used keywords such as "alcohol," "drinking," "alcohol policy," "gambling," "wagering," "betting," "gambling policy," "substance use policy," "college (university) regulations," and "college (university) policies" to identify relevant sections of each Web site. Several sites did not include a search function; in such cases, we conducted a comprehensive visual search of the site. We also examined supplemental materials provided by schools (e.g., policy manuals, brochures, pamphlets, etc.) for comparison against handbooks and Web-based materials. We conducted a visual search of all hard-copy handbooks and supplemental policy materials and extracted all relevant information from these sources. We systematically archived all of the Web-based and other electronic regulatory sources (e.g., pdf- and text-based student handbooks and policy manuals, html pages, etc.) from each school on a computer. We filed hard copy materials, such as student handbooks and policy manuals, by school and kept these documents on site. Policy Coding Procedure and Instrument Investigators developed a coding instrument by studying alcohol and gambling policies from a variety of U.S. schools outside the current sample and identifying the underlying characteristics of the policies. These characteristics were reduced to 40 items that reflected the scope and focus of school alcohol and gambling policies. The items were converted into a coding instrument that included 25 variables for alcohol policy and 15 variables for gambling policy. This instrument included items to measure the presence of specific policies and establish whether the policies were punitive or rehabilitative. All variables used a nominal scale that included common characteristics of each school policy; response choices varied slightly with the focus of each variable. All of the variables were arranged on a six-page coding form. To simplify coding and allow for within-school comparisons between different formats of policy dissemination (e.g. school handbook vs. school Web site), we separated each school's policy materials into three categories: (1) student handbooks (electronic or paper); (2) Web-based materials; and (3) supplementary materials (paper); a potential 357 documents required coding (three coding categories for each of 119 schools in sample = 357 potential documents). However, because not every school had documents available in all three coding categories, the final document count was 164. Specifically, at the end of our data collection process, we had collected 73 student handbooks, 70 Web-based policies, and 21 supplementary documents. We assigned 11 coders the job of evaluating each school's alcohol and gambling policies. Each coder read a selection of policies and extracted relevant information in accordance with the coding form. The coding process proceeded as follows: 1. Each policy document was assigned to two of eleven eligible DOA coders randomly. Each assigned coder independently abstracted information from each assigned policy document and recorded this information on separate coding forms. 2. For each document, one member of the research team, designated as the "arbiter," compared the two coding forms and marked discrepant items. 3. The arbiter returned the marked coding forms to their respective coders and requested that coders reconsider their answers to the items in question. Upon reconsideration, coders were free to change their answers or keep their original answers. 4. Coders resubmitted their recoded documents to the arbiter who compared the discrepant items again. Discrepancies that remained were noted and resolved by the arbiter. 5. Once all discrepancies had been resolved, the policy assessments on the coding forms were entered into an SPSS database using a procedure that screened entries for out-of-range values and discrepancies in branching among items. 6. We assessed data entry reliability by selecting 10% of the cases in our database and rechecking each data entry point. Of the 680 items entered in these 17 randomly selected cases, there were no observed data entry errors. Shaffer and his associates have used a similar process of information extraction, coding and arbitration successfully in other published studies [ 1 , 32 ]. Results Our analysis of college alcohol and gambling policies generated several types of results. First, we describe the results of our coding procedure, the final sample of schools and available policy information. Next, we examine the policy evidence across information sources by analyzing the consistency between the information provided by handbooks and Web materials. We then present the prevalence of individual policy items and the results of a factor analysis that explored the underlying dimensions of the policy variables. Finally, we analyze the relationships between policies and student drinking and gambling rates using information collected in the most recent Harvard School of Public Health College Alcohol Study (CAS) [ 6 ]. Inter-Coder Concordance We assessed inter-coder reliability by comparing the total number of discrepant coded items to the total number of coded items. As described previously, each policy was assigned to two of eleven eligible DOA coders randomly. The participation of eleven coders yielded 55 possible coding-pair combinations; each of these pairings coded at least one policy. Specifically, the number of policies coded by each coder-pair ranged from a minimum of one (n = 6) to a maximum of six (n = 3). Coders had up to two opportunities to code each document: (a) an initial round of coding; and (b) a second round of coding to reconsider any discrepant items identified by the arbiter after the initial round of coding. The arbiter made the final coding decision on 345 out of a total of 4,100 possible items. The coding process yielded a study-wide inter-coder reliability rate of 91.6%. College Sample After thirty days had passed from our initial e-mail request, 46 of 119 schools had responded by sending hard copy materials. Eighteen of these 46 colleges sent materials completely unrelated to our request for school alcohol policies (e.g., applications for admission, school newsletters). Fourteen schools sent student handbooks, and another 14 schools sent other alcohol and/or gambling related (i.e., non-handbook) materials. Seventy-three schools did not respond to our request within thirty days. Subsequent to our follow-up telephone requests, we received student handbooks and supplemental materials from an additional 22 schools. This recruitment procedure resulted in 50 schools actively providing policy information for this study; for the remainder, policy information was obtained through other investigative procedures as described earlier (e.g., Web sites). Policy Sample This study sought information on alcohol and gambling policy from a representative sample of 119 colleges across the U.S. We utilized three distinct common sources of information on school alcohol policy: student handbooks, school Web sites (non-handbook related) and supplementary materials (e.g., policy manuals, pamphlets). We collected a total of 164 policy-related documents from three sources: 73 policy documents from handbooks, 70 documents from school Web sites, and 21 from supplementary materials. Table 1 presents the sources of alcohol policy information for the schools in our sample. Forty schools presented their full alcohol policy in their handbook, 31 on their Web site, 2 in supplementary materials, and 44 through a combination of handbook, Web site and supplements. We were unable to locate any policy information for two schools in our sample; these schools did not respond to our requests for information. Table 1 Sources of school alcohol policy information Number of Schools Handbook Policies Web site Policies Supplemental Materials No materials 40 31 2 25 5 11 3 2 Total = 119 Total = 73 70 21 2 Policy across Information Sources We aggregated and analyzed policy information across sources because a preliminary examination revealed content differences among handbooks, Web sites [ 33 ] and supplementary materials. Aggregating information across sources provides the most extensive view of each college's policy strategy because it considers all modes of policy distribution. This strategy yields the most comprehensive policy search and identifies more policy mentions than is possible by examining only one policy source. To implement this strategy, we first constructed a new database that included data for schools with a handbook, a Web site or both (n = 115). Next we created a single record of policy mentions for the 28 schools with both handbooks and Web materials by aggregating policies across sources. This database assimilated the unique handbook and Web variables into a single set of "recompiled" variables, reflecting the total number of policies attributable to either the handbook or the Web. To compare the "added value" of school Web sites (i.e., policy information presented on the Web that was not presented in the handbook), we summed the policies reflected by the recompiled variables and then subtracted the policies contained in the handbook-only variables. Of 263 total policy items present, we collected 198 (75%) policy items from student handbooks and 65 (25%) additional policy items from school Web sites that were not available in handbooks. To determine the added contribution of supplemental materials (i.e., policy information presented in the supplements that was unavailable elsewhere), we created another set of recompiled variables following the previously outlined procedure. These variables reflected the total number of policies identified for the three schools with all three types of sources (i.e., handbooks, Web materials and supplements). We summed the policies reflected by the recompiled variables and then subtracted the policies contributed by handbooks and Web sources; this procedure revealed that supplemental materials contributed 4 of 30 (13.3%) policy items. Although school Web sites provide a substantial amount of alcohol policy information that is not contained in the primary document customarily provided to students (i.e., handbook), the overall added contribution of the school Web site in presenting policy information varied among schools. For example, one school's Web site contained an additional eight alcohol regulations that were not included in the handbook; however, several schools' sites contained no additional information. In addition, the type of information that was presented only on Web sites also varied: while most information pertained mainly to secondary alcohol policies (e.g., school-sponsored events and drinking regulations for drinking-aged students), some schools chose to present vital alcohol policy information (e.g., stating that all drinking is prohibited for students <21) on their Web site only (n = 2). Thus, although handbooks and Web sites are both important sources of alcohol policy information and supplements contribute little additional information, consistency across sources varies. The following analyses assess the agreement of information found in multiple sources. Handbook-Web Concordance As mentioned earlier, of the 117 colleges for which we had information, all 117 (100%) had a written policy on student alcohol consumption and 26 schools (22%) had a student gambling policy. Because all schools had a written alcohol policy (and relatively few schools had a gambling policy), the following analyses focus on alcohol policies. Determining concordance between handbook and Web sources is important because administrators might be unaware of inconsistencies between official school documents. In addition, contradictory information can mislead students and potential applicants. We assessed the concordance between sources of college alcohol policy materials by determining the level of agreement (i.e., presence or absence of policy information) between handbooks and Web materials; that is, we compared the content of each type of document to identify differences in the presentation of each school's policy information between sources. We did not extend this particular analysis to include supplemental materials because, as we noted before, only a small number of schools (n = 3) had all three types of sources. Twenty-eight schools had both a handbook and Web materials; for each of these 28 schools we determined the absence or presence of the 25 alcohol policy variables in each source. We predetermined that a concordance rate of 85% would indicate a high level of agreement between documents. To be considered in agreement, complementary information had to be found in (or absent from) both sources; in cases where this requirement was not satisfied, the policies were considered in disagreement. Using these criteria, we determined that three policy variables (i.e., 12% of the policy variables) were mentioned often and were present in both handbooks and on Web sites, and consequently, showed high agreement. Either type of information resource seldom mentioned ten policy variables (i.e., 40% of the policy variables), therefore, also exhibiting high agreement. The remaining 12 policy variables (i.e., 48% of the policy variables) were often mentioned, but not consistently by both sources, indicating low agreement. Table 2 presents the three "high agreement" alcohol policies that were mentioned consistently in both handbooks and Web materials. Variables that fell into this category generally measured broad school policies (i.e., the existence of an alcohol policy). As Table 2 illustrates, schools consistently made these types of alcohol policies available to the public in both print and electronic form, making this information highly accessible. Table 2 "High Agreement" Alcohol Policies, Often Mentioned in Both School Handbooks and Web Materials (N = 28) Policy % of schools, HB only % of schools, Web site only % of schools, HB and Web site % of schools, no mention in HB or Web site Is there an alcohol policy? 0 0 100.0 0 Alcohol is prohibited on campus for students <21 7.1 7.1 85.7 0 Alcohol is allowed at sanctioned events for students ≥ 21 0 0 87.5 12.5 Policy variables that were rarely mentioned in handbooks and Web materials appear in Table 3 . These variables primarily measured on- and off- campus alcohol consumption restrictions and school recovery polices regarding student alcohol use. These policies are in "high agreement," because they were seldom mentioned: as Table 3 demonstrates, this information was missing from handbooks and Web sites in nearly all cases. Table 3 "High Agreement" Policy Variables, Rarely Mentioned in School Handbooks and Web Materials (N = 28) Policy % of schools, HB only % of schools, Web site only % of schools, HB and Web site % of schools, no mention in HB or Web site Alcohol is prohibited off-campus for students ≥ 21 0 0 0 100.0 Alcohol quantity limits at off-campus events 0 0 3.6 96.4 Policy on container restrictions at off-campus events 3.6 0 0 96.4 Policy on leave of absence for recovery 3.6 0 0 96.4 Policy allowing students to participate in recovery while living in dorm 3.6 3.6 0 92.9 Attendance restrictions for hosted events 0 10.7 3.6 85.7 Policy on students with an alcohol problem upon entering school 7.1 3.6 0 89.3 Policy on students who develop an alcohol problem while in school 10.7 3.6 0 85.7 Policy on students already in recovery upon entry to school 10.7 3.6 0 85.7 Policy on students who enter recovery while in school 10.7 3.6 0 85.7 Table 4 presents variables that were mentioned occasionally (i.e., concordance <85%) in handbooks or Web materials. The policies in this category primarily address consumption and event restrictions and student recovery. Table 4 illustrates that we observed considerable inconsistencies in schools' methods of distribution of these types of policies. Table 4 "Low Agreement" Policy Variables, Mentioned Inconsistently in School Handbooks and Web Materials (N = 28) Policy % of schools, HB only % of schools, Web site only % of schools, HB and Web site % of schools, no mention in HB or Web site Alcohol is prohibited on-campus for students ≥ 21 0 7.1 10.7 82.1 Off-campus alcohol restrictions in place for students ≥ 21 10.7 7.1 3.6 78.6 School policy is to defer to local laws on alcohol consumption 14.3 10.7 3.6 71.4 Policy on alcohol quantity limits at events 17.9 10.7 3.6 67.9 Attendance restrictions for school sanctioned events 21.4 10.7 3.6 64.3 Campus operates an alcohol recovery program 32.1 7.1 3.6 57.1 Policy on alcohol-free campus housing 25.0 25.0 3.6 46.4 Document makes clear other ways by which the school makes students aware of the official alcohol policy 25.0 17.9 21.4 35.7 Policy on container restrictions on campus 21.4 21.4 21.4 35.7 Campus makes referrals to off-campus recovery programs 25.0 21.4 28.6 25.0 On-campus alcohol restrictions in place for students ≥ 21 13.0 26.1 56.5 4.3 Policy on alcohol at on-campus sanctioned events for students ≥ 21 17.9 21.4 57.1 3.6 Identifying the Underlying Dimensions of College Policy As noted earlier, the coding process revealed that all 117 colleges (i.e., 100% of the schools for which information was available) in this sample had a written policy on student alcohol consumption, but only 26 (22%) had a published policy that addressed gambling. The small number of schools with gambling policies precludes confident analysis of the dimensional composition of our gambling variables; therefore, we applied the factor analysis that follows only to alcohol policy variables. Three policy variables represented a multi-dimensional measurement strategy to yield detailed policy information. Consequently, we collapsed these three redundant policy items into the primary or gate items from which they originated (e.g., "alcohol is prohibited on-campus for students ≥ 21" and "on-campus alcohol restrictions in place for students ≥ 21" became "policy on alcohol use on-campus for students ≥ 21). This resulted in 22 alcohol policy variables in all remaining analyses. These dependent variables all measured different aspects of school alcohol policies (e.g., policy presence, content, and target). To empirically examine the underlying dimensions reflected by our variables, we conducted an exploratory factor analysis. This procedure employed an initial factor extraction (i.e., component matrix) and then an orthogonal rotation to simple structure. We selected the Varimax rotation to maximize the variance of loadings within factors and minimize the covariance across factors. The orthogonal solution identified eight policy clusters with Eigenvalues greater than 1.0 that explained 72.36% of the total variation. This explained variance lies within the 50–75% useful range suggested by Overall and Klett (1972)[ 34 ]. Consequently, we concluded that our factor analysis provided a valid identification of the policy clusters that underlie college alcohol and gambling regulations. Table 5 presents the structure of the interrelationships among policies. To facilitate interpretation, the table reports only factor loadings ≥ 0.50 (i.e., policies with loadings in this range correlate .50 or greater with a composite measure of the overall dimension). Table 5 Orthogonal Factor Structure and Items Loading ≥ 0.50 on Each Factor Variable Factor Loading % Variance Explained Factor 1 – School policy and the law 4.83 Does the school alcohol policy defer in full to local law? .88 Factor 2 – Prohibition policies 5.43 Does the policy state that alcohol is prohibited for students <21? .85 Does the policy state that alcohol is restricted on-campus for students ≥ 21? .76 Factor 3 – Policies for legal-aged drinkers 6.65 Is alcohol is allowed at sanctioned on-campus events for students ≥ 21? .75 Is alcohol prohibited off-campus for students ≥ 21? .74 Factor 4 – Limits and restrictions – on-campus 9.13 Does the policy state whether the school offers alcohol-free campus housing? .77 Does the policy address alcohol container restrictions on campus? .74 Does the policy address alcohol quantity limits (i.e., total alcohol available) at on-campus sanctioned events? .57 Factor 5 – Events policies 7.20 Are there attendance restrictions for off-campus sanctioned events? .81 Are there attendance restrictions for on-campus sanctioned events? .78 Are there restrictions on off-campus alcohol use for students ≥ 21? .56 Factor 6 – Limits and restrictions – off-campus 13.50 Is there policy on alcohol quantity limits (i.e., total alcohol available) at off-campus sanctioned events? .81 Is there policy on alcohol container restrictions off-campus? .79 Is there policy that permits students a leave of absence to participate in a recovery program? .50 Factor 7 – Recovery recognition policies 20.59 Is there policy on students who enter alcohol recovery while attending? .90 Is there policy on students who are in alcohol recovery upon entry? .88 Is there policy on students who have an alcohol problem upon entry? .87 Is there policy on students who develop an alcohol problem after entry? .79 Is there policy that permits students in an alcohol recovery program to live in a dormitory on campus? .73 Factor 8 – Recovery facilitation 5.03 Does the campus makes referrals to an off-campus recovery program for students with alcohol use disorders? .74 Does the campus operate a recovery program for students with alcohol use disorders? .73 The eight factors are ordered according to the number of policies measuring the overall domain; that is, factors containing general, or "blanket," policy items are listed first, followed by factors containing more specific policy items. Factor 1 (i.e., School policy and the law ) contained only one item that loaded ≥ 0.50 and explained 4.83% of the total variance. Many schools deferred to local law entirely and did not publish other policies that were unique to the school. Factor 1 identifies this deference policy as a unique dimension. Factor 2 (i.e., Prohibition policies ) expands on deference to local law and presents additional school policies that prohibit alcohol for underage and legal age students. These items accounted for 5.43% of the total variance. Factor 3 includes policies that extend restrictions to include drinking by students of legal age (i.e., Policies for legal-aged drinkers ). Items in Factor 3 explained 6.65% of the total variance. Factors 1, 2 and 3 include alcohol policies that focus on the legal status (i.e., legal age) of students; in addition, for those of legal age, these policies range from no school specific policies to prohibitions for students who are old enough to drink legally. Factors 4 through 6 include policy variables directed to alcohol use on-campus and off-campus. Factor 4 (i.e., Limits and restrictions – on campus ) provides specific guidance about where on-campus students can drink and how much alcohol is available (i.e., housing and container and quantity restrictions); these items accounted for 9.13% of the total variance. Factor 5 policies (i.e., Events policies ) accounted for 7.20% of the total variance and focus primarily on restrictions for on- and off-campus events. Factor 6 (i.e., Limits and restrictions – off campus) accounted for 13.50% of the total variance and includes policies that regulate off-campus activities (i.e., alcohol quantities, containers and leaves of absence). Different from the first 6 factors, factors 7 and 8 focus on student recovery. Factor 7 policies (i.e., Recovery recognition policies ) recognize that students can have alcohol related problems that require recovery, and that these problems can exist before entering college or develop during college; these items accounted for 20.59% of the total variance in the data. Finally, Factor 8 (i.e., Recovery facilitation ) accounted for 5.03% of the total variance and includes policies that reflect how the school participates in the recovery process (i.e., triage or treatment). One item (i.e., "how does the campus inform students of the official school policy") failed to load ≥ 0.50 on any factor and was excluded from the final analysis. Alcohol Policy Prevalence College alcohol policies varied widely. Table 6 summarizes the prevalence of alcohol-related policy and the mean prevalence of alcohol policies within each factor. The prevalence of alcohol policies ranged from 100% (i.e., the presence of an alcohol use policy) to 1.7% (i.e., policy that permits a leave of absence to participate in a recovery program). The mean prevalence for the eight policy factors ranged from 92.3% (i.e., Prohibition policies ) to 5.4% (i.e., Limits and restrictions - off-campus ). Table 6 Prevalence of College Alcohol Policies and Policy Attributes Policy & Policy Attributes Prevalence % (N) School has a written policy on alcohol use 100 (117) Policy states that alcohol is prohibited for students <21 97.4 (114) Policy on alcohol use on-campus for students ≥ 21 87.2 (102) Policy addresses alcohol at sanctioned on-campus events for students ≥ 21 68.4 (80) Campus makes referrals to an off-campus recovery program for students with alcohol use disorders ♠ 57.3 (67) Policy addresses alcohol container restrictions on-campus 50.4 (59) Policy makes clear how the campus informs students of the official school alcohol policy 43.6 (51) Campus operates a recovery program for students with alcohol use disorders ♠ 28.2 (33) School alcohol policy defers in full to local law 26.5 (31) Policy addresses alcohol quantity limits (i.e., total alcohol available) at on-campus sanctioned events 26.5 (31) School offers alcohol-free campus housing 22.2 (26) Restrictions on off-campus alcohol use for students ≥ 21 21.4 (25) Attendance restrictions for on-campus sanctioned events 20.5 (24) Policy on students who develop an alcohol problem after entry ♠ 10.3 (12) Policy on alcohol quantity limits (i.e., total alcohol available) at off-campus sanctioned events 8.5 (10) Policy on students who enter alcohol recovery while attending ♠ 7.7 (9) Attendance restrictions for off-campus sanctioned events 7.7 (9) Policy on students who are in alcohol recovery upon entry ♠ 6.0 (7) Policy on alcohol container restrictions off-campus 6.0 (7) Policy permits students in an alcohol recovery program to live in a dormitory on campus ♠ 5.1 (6) Policy on students who have an alcohol problem upon entry ♠ 3.4 (4) Alcohol is prohibited off-campus for students ≥ 21 3.4 (4) Policy permits students a leave of absence to participate in a recovery program ♠ 1.7 (2) Policy Factor Mean Prevalence (%) Factor 2 – Prohibition policies 92.3 Factor 8 – Recovery facilitation ♦ 42.8 Factor 3 – Policies for legal-aged drinkers 35.9 Factor 4 – Limits and restrictions – on-campus 33.0 Factor 1 – School policy and the law 26.5 Factor 5 – Events policies 16.5 Factor 7 – Recovery recognition policies ♦ 6.5 Factor 6 – Limits and restrictions – off-campus 5.4 ♠ Recovery-oriented policy; ♦ Recovery-oriented factor Policy, Binge Drinking and Gambling We conducted several analyses to determine the nature of relationships between student alcohol consumption, gambling behavior and policy content. As the factor analysis above illustrates and the relative prevalence of policies confirms, college alcohol-related policies are primarily intended to prevent, reduce or restrict alcohol use among students on college campuses. To test the relationships between alcohol policies and student drinking behaviour, we compared the mean binge drinking rates of students at schools with and without each policy variable. We obtained the mean binge drinking rates of the schools in our sample from the dataset used in Wechsler et al.'s Harvard School of Public Health College Alcohol Study (CAS)[ 6 ]. Because this is one of the first studies of college policies, we sought to identify as many potential relationships between policy, drinking and gambling as possible; therefore we set a liberal alpha level (α = .1) for this analysis. A one-way analysis of variance (ANOVA) revealed that four of the 22 policy variables had significant relationships with binge drinking rates at the colleges in our sample (see Table 7 ). Table 7 Mean Binge Drinking Rates (%) and Alcohol Policy Variables Policy variable Schools with no policy mention (N) Schools with no policy restrictions (N) Schools with restrictions policy (N) Schools with prohibition policy (N) F df Policy on alcohol use on-campus for students ≥ 21 n/a (0) 39% (15) 47% (77) 36% (24) 7.07* 2,113 Policy on alcohol at on-campus sanctioned events for students ≥ 21 44% (37) 46% (64) n/a (0) 33% (15) 5.25* 2,113 Policy on students already in recovery upon entry to school 43% (109) 56% (6) 57% (1) n/a (0) 2.89* 2,113 No mention (N) Alcohol free housing available (N) Policy on alcohol-free campus housing 42% (90) 49% (26) 4.01** 1,114 * = p < .1; = p < .01; * = p < .001 Schools that had either no policy restrictions or a prohibition policy for on-campus alcohol use by students ≥ 21 had lower mean past-month student binge drinking rates (39% and 36%, respectively) compared to schools that employed an intermediate level of restrictive policies (47%) (F = 7.07, df = 2,113, p < .001). Schools that allowed or did not mention alcohol use at on-campus sanctioned events for students ≥ 21 had a higher mean binge drinking rate (46% and 44%, respectively) than schools that prohibited legal drinking at events (33%) (F = 5.25, df = 2,113, p < .001). Schools that offered alcohol-free housing had a higher mean student binge drinking rate of 49% compared to schools that did not mention alcohol-free housing which had a binge rate of 42% (F = 4.01, df = 1,114, p < .01). Two other alcohol policy variables evidenced significant relationships with student binge drinking behaviors, but these policies lacked widespread implementation at a large number of schools. Schools that specifically allowed a leave of absence for a student to participate in recovery activities (n = 2) evidenced a higher mean binge drinking rate (69%) than schools that did not mention such a policy (43%, n = 114). Keeping the small number of schools in mind, it is worth noting that schools that prohibited off-campus alcohol consumption for students ≥ 21 (n = 4) had a lower mean binge rate (10%) than schools without this provision (45%, n = 112). There were not enough schools with gambling policies to permit a detailed analysis of the relationship between policies and student gambling behavior; therefore instead of conducting an analysis of the relationship between gambling behavior and individual policy variables, we only were able to assess gambling behavior based on whether schools had a gambling policy. Using prevalence data from LaBrie et al.'s (2003) [ 31 ] recent study of student gambling behavior, we determined that no significant difference in mean past-year student gambling behavior existed between schools with a written policy on gambling (i.e., prohibitive or restrictive) and schools with no mention of gambling policy (i.e., approximately 40% regardless of policy presence). Unanticipated Policy Effects: Alcohol Policy can Influence Gambling Participation Long ago, Pigou [ 35 ] noted that public policies can have unanticipated effects; policy intended to regulate one set of behaviors can influence other patterns of behavior. To test the relationships between alcohol policies and student gambling behavior, we compared the mean past-year gambling rate at schools with each alcohol policy to schools without the policy. As before, we used a liberal alpha (α = .1) to identify all potential relationships. A one-way analysis of variance (ANOVA) revealed that four of the 22 alcohol policy variables had significant relationships with the proportion of students who gambled in the past-year school year (see Table 8 ). Table 8 Mean Past-Year Student Gambling Participation Rate (%) and Alcohol Policy Variables Policy variable Schools with no policy mention (N) Schools with no policy restrictions (N) Schools with restrictions policy (N) Schools with prohibition policy (N) F df Policy on alcohol use on-campus for students ≥ 21 n/a 46% (15) 40% (77) 41% (24) 2.64* 2,113 Policy on alcohol at on-campus sanctioned events for students ≥ 21 44% (37) 40% (64) n/a (0) 35% (15) 5.17* 2,113 No mention (N) Limits (N) Policy on alcohol quantity limits at on-campus events 42% (85) 37% (31) 5.39* 1,114 Policy on off-campus alcohol restrictions for students ≥ 21 42% (91) 38% (25) 2.94* 1,114 * = p < .08; = p < .01; * = p < .001 Schools that restricted or prohibited on-campus alcohol use for students over 21 evidenced similar mean past-year student gambling participation rates (i.e., 40% and 41%, respectively), and schools with no restrictive policy evidenced a higher student gambling participation rate (46%) (F = 2.64, df = 2,113, p < .08). Schools that did not mention or allowed alcohol at on-campus events for legal drinkers exhibited a higher mean gambling participation rate (i.e., 44% and 40%, respectively) than schools that prohibited alcohol at on-campus events (35%) (F = 5.17, df = 2,113, p < .001). Similarly, schools that did not limit the quantity of alcohol available at events showed higher past-year gambling participation among students (42%) compared to schools with no such provision (37%) (F = 5.39, df = 1,114, p < .001). Schools that had off-campus alcohol restrictions for legal-aged drinkers had a gambling rate of 38% while schools that did not had a mean gambling rate of 42% (F = 2.94, df = 1,114, p < .1). Two other alcohol policy variables evidenced significant relationships with student gambling behaviors, but these policies were not widely implemented throughout the sample. Schools that did not expressly prohibit alcohol consumption for underage drinkers (n = 3) evidenced a mean past-year student gambling rate of 51%, while schools that prohibited underage drinking (n = 116) had a gambling rate of 41% (F = 3.19, df = 1,114, p < .1). Schools that banned all alcohol consumption, whether on- or off-campus (n = 4) evidenced a lower gambling rate (30%) than schools that allowed at least some drinking (41%, n = 112) (F = 5.80, df = 1,114, p < .05). In addition to the direct relationships between policy variables and binge drinking and gambling, three alcohol policy variables evidenced unexpected interaction or intensification effects when gambling policies also were present. Schools that had both a policy prohibiting or restricting gambling activity among students and a policy prohibiting on-campus legal-aged drinking (n = 9) had a mean binge drinking rate of 29% – much lower than schools with just an alcohol policy (40%, n = 15), a gambling policy (47%, n = 17) or neither (45%, n = 75) (F = 3.15, df = 1,112, p < .1). Schools with both a gambling policy and a policy prohibiting alcohol at on-campus events (n = 6) had a significantly lower mean binge rate (22%) than schools with just an alcohol policy (40%, n = 9), a gambling policy (46%, n = 20) or neither (45%, n = 81) (F = 5.88, df = 1,112, p < .05). Schools that prohibited alcohol at on-campus events also evidenced significantly lower past-year student gambling rates than schools without such prohibitions: 30% versus 38% or higher (F = 8.63, df = 1,112, p < .05). Discussion Using written (i.e., hard copy) and Web based sources, this study examined the nature and extent of alcohol and gambling-related policies among a representative sample of U.S. colleges. Every school in this representative sample had at least one alcohol use policy; however, few schools (i.e., 26 of 117; 22%) had at least one gambling policy available. The relative rarity of gambling-related policies on college campuses represents a lost opportunity by school administrators to (a) prevent or limit disordered gambling among students and (b) facilitate recovery for students in need of gambling treatment. A recent study showed that, while not as prevalent as previously thought, gambling on college campuses is still quite common, with 42% of students having gambled in the past year and 2.6% gambling weekly or more [ 31 ]. The frequency with which gambling occurs on college campuses could be indicative of lingering misconceptions about gambling outcomes among student populations. For example, research has shown that gamblers are largely unaware of the probabilities associated with various forms of gambling (e.g., Ladouceur et. al. 1996[ 36 ]; Rogers 1998[ 37 ]); this circumstance leaves gamblers susceptible to cognitive errors [ 38 ]; [ 39 ]; [ 40 ] and suggests that gambling behavior is largely driven by social factors and injunctive norms (i.e., the tendency to engage in gambling as a function of personal perceptions of society's acceptance of gambling) [ 41 , 42 , 11 ]. By failing to implement comprehensive restrictive and recovery-based gambling policies and neglecting to educate students about the probabilities associated with gambling as well as the dangers of excessive gambling, school administrators are overlooking an important and potentially destructive problem that faces many of today's students. Future policy-based education and recovery initiatives might be able to effectively reduce student gambling behaviors; however, given the current dearth of gambling policies, we cannot determine whether school policies can effectively reduce at-risk gambling behaviors on college campuses. All the schools in our sample recognized the need for some type of alcohol policy; however, the presence of more targeted policies varied considerably. Although this variation might reflect different policymaking strategies across institutions, it also could result from a variety of other influences, including the lack of federal standards guiding the creation of alcohol policy on college campuses. The absence of policy guidelines leaves administrators with a wide range of options about how to best address student substance use and abuse. Some administrators might prefer to match policy to their perception of local needs, while others might welcome policy guidelines. Both of these circumstances encourage additional research designed to help guide administrators to identify effective policies (e.g., "best practices"). Alcohol policies ranged from comprehensive restrictions and prohibition to liberal acceptance of student alcohol use. Policies encouraging recovery among students with alcohol use disorders were decidedly absent from our sample. For example, only 57.3% of schools expressed in writing that they made referrals to alcohol recovery services; all other recovery-oriented policy provisions were in place at fewer than 30% of schools, with two-thirds of these policies in effect at fewer than 10% of schools. Examination of the mean prevalence of the eight policy factors provides additional support for this finding. The mean prevalence of recovery recognition policies was 6.5% (i.e., factor 7). Although recovery facilitation policies were more common, with a mean prevalence of 42.8% (i.e., factor 8), this rate simply reflects that many schools report making referrals to outside treatment facilities. In contrast, the prevalence of prohibition policies was 92.3% (i.e., factor 1). This limited of consideration for student recovery and emphasis on punitive and prohibitive measures might reflect an underlying institutional bias against accommodating students with special needs and an unsupportive atmosphere for those who are at most risk for developing alcohol problems. The higher prevalence of policies that direct referrals to outside treatment resources might indicate an eagerness among administrators to export students with addiction problems to non-school affiliated assistance. Alternatively, placing little emphasis on recovery might simply represent a lack of understanding about addictive behaviors and addiction recovery among school administrators. Even though the results of this study indicate that schools with recovery policies can evidence higher rates of binge drinking, this rate might reflect a pre-existing campus problem (i.e., policy as a consequence of behavior) rather than be a result of recovery policy implementation. Future policy research needs to examine whether a better balance between punishment and treatment policies will yield improved student health services, less substance misuse and, consequently, a better campus life for all students. As hypothesized, there were considerable and important differences between the information that was available in handbooks and on the Web. This observation is not surprising because there are few requirements guiding the creation of school substance use and gambling policies and no standards requiring consistency among sources. Also, the factor analysis illustrated that alcohol policies currently concentrate primarily upon prevention and punishment, and devote relatively limited attention to student recovery. Finally, as we expected, schools with no policy restrictions on alcohol consumption or restrictive alcohol policies often experienced higher levels of binge drinking among students than schools with prohibitive alcohol policies. Nevertheless, alcohol policies were associated inconsistently with student binge drinking rates; few policy variables exerted influence on patterns of student drinking. In some cases, policies designed to reduce student alcohol consumption showed an opposite effect. Absence of Shared Standards and Model Policies The results of this policy analysis suggest diverse and perhaps ambivalent tolerance toward alcohol use among U.S. colleges that has led to a deep underlying problem: there is a lack of model policy guidelines to assist colleges in (a) preventing addictive behaviors among students and (b) providing assistance to students already struggling with addiction. Optimally, if available, such guidelines would provide school administrators with "best practice" model policies that address the many aspects of addiction among students (e.g., permitted and prohibited substance use on campus, legal matters, parental notification, treatment options, financial issues, academic issues, etc.) and would outline strategies for implementing and enforcing such policies. In the absence of an explicit regulatory framework (i.e., the current policy environment), schools are left to regulate addictive behaviors based upon the local attitudes and expectations of communities and implicit moral values held by school administrators. This circumstance can lead colleges to install policies intended to have immediate and drastic results (i.e., prevention and punishment policies) while neglecting policies that would promote recovery and provide lasting benefit to students. The absence of evidence-based policy leaves school administrators in a position of promulgating policy that might have effects that are contrary to their intentions. For example, the Boston Globe reported on October 30, 2002 that, "Forty-seven Massachusetts colleges will sign onto a statewide campaign today designed to punish students who abuse alcohol and to cut down on the rate of binge drinking on campuses by providing alcohol-awareness training for students, athletes, and Greek system members"[ 43 ]. This type of policy imposes blind restrictions and punitive measures on students without considering the underlying factors motivating student drinking (e.g., campus environment) or the need for specialized programs (e.g., recovery programs) to assist in reducing student drinking. Thus, even when colleges have existing policies that address alcohol, drug use or gambling, it is likely that the number of policies related to students involved with or seeking recovery programs is much more limited. It is interesting to consider that colleges routinely provide remedial courses in expository writing to help students cope with the demands of academics but resort to exclusion or punishment for students who fail to adequately cope with an academic environment where 44% of students binge drink [ 6 ]. Inconsistent Policy Presentation The prevalence of alcohol-related problems at U.S. colleges highlights the need for comprehensive student policies addressing all aspects of alcohol use. Further, for policy to be effective students need to be aware of and understand school requirements; this requires that students know how to access policy information. Because schools currently are not required to disseminate alcohol policies in any particular form or through any specific medium, the availability of existing policy information can vary substantially among schools. Though most commonly printed in student handbooks, substance use policies also appear in policy manuals, pamphlets and, increasingly, on school Web sites. One recent study of college Web-based alcohol policy information [ 33 ] found that 50 of 52 schools included alcohol policy information on their Web site; unfortunately, schools rarely consolidated this information onto a single Web page for easy viewing. The result was that alcohol policy information was often incomplete and/or difficult to access [ 33 ]. Although the Internet provides an excellent opportunity for schools to reach technology-savvy students and parents, individuals seeking policy information on the Web are likely to face disorganization and user-unfriendly designs. Thus, while students and parents might expect to access alcohol or gambling policies by turning to a college's student handbook or Web site, the results of this study suggest that their success in locating the desired information will vary considerably. Adoption of generally accepted standards for policy dissemination could increase student awareness, and consequently, compliance with college rules. Tables 2 through 4 illustrate that schools present different types of alcohol policy information in their handbooks and Web sites; sometimes this information agrees across sources, and often it does not. For some policy variables, this discrepancy is problematic. For example, most schools (82.1%) did not report on the presence or absence of alcohol-free campus housing at all. It is difficult to assess whether an absence of policy indicates a tacit acceptance of an activity (i.e., legal-aged alcohol consumption). More likely, however, it indicates that at these schools some drinking is permissible among students over age 21. For other policies, schools' lack of consistency in reporting uniform policy creates different problems. For example, as Table 4 illustrates, handbooks and Web sites mislead people inquiring about on-campus alcohol-free housing 25–50% of the time, at schools that have such a policy, depending upon the information source (i.e., handbook or Web site). Many schools (46.4%) did not mention alcohol-free housing at all; it is unclear whether these schools have alcohol-free on-campus housing, or simply lack an explicit policy addressing the matter. This study reveals that to gain a complete understanding of the components of a school's alcohol policy, it is necessary to consult both the handbook and the school Web site whenever possible. However, schools do not alert students and parents to this fact, and it is unreasonable for school administrators to expect inquiring persons to conduct complex searches for information that is considered to be freely available in the public domain. Policy Content: Prohibition, Punishment and Recovery The results of the factor analysis provide a stark portrait of the current composition of college policies on potentially addictive behaviors. This analysis reveals that six of eight factors contained prohibitive and/or punitive policy variables. Although the factor analysis merely revealed the underlying psychometric properties of our instrument, the prevalence of specific policy variables across schools provides additional support for this finding. Whereas the results suggest that prohibition-oriented policies have been effective in reducing binge drinking and gambling under certain conditions, the relative scarcity of recovery-oriented policies prevents us from properly comparing the efficacy of these two strategies. Policy, Binge Drinking and Gambling The analysis of policy variables and binge drinking rates revealed a variety of relationships among policy variables and student drinking behaviors. Specifically, four policy variables related to student binge drinking (see Table 7 ), and four alcohol policy variables related to student gambling behaviors (see Table 8 ). Interpretation of some of these relationships is relatively straightforward. For example, schools that prohibited alcohol at on-campus events experienced less binge drinking than other schools. This finding suggests that most students will abide by school policy. However, other relationships are more difficult to interpret: schools that restricted legal-age drinking had higher binge drinking rates than schools that did not make restrictions for students ≥ 21. Perhaps, when forbidden, students find alcohol to be more desirable. Alternatively, students of legal age might feel constrained by prohibitions and, as a result, drink more often to excess than they would if the opportunity to drink was commonplace. This pattern of drinking was commonly observed during the Volstead Act (i.e., national prohibition from 1920–1933). Many people did not drink because it was illegal, but those who did tended to drink to excess [ 44 ]. Just as with the Volstead Act, determining the real effect of school alcohol policies is difficult because many other factors (e.g., how long the policy has been in effect, state or local culture, etc.) influence drinking. The relationship between policy and drinking can be counterintuitive. For example, schools that offered alcohol-free campus housing evidenced significantly higher mean student binge drinking than schools that did not mention alcohol-free housing (49% vs. 42%, respectively). Observers might expect schools promoting alcohol-free dormitories to evidence lower binge drinking; alternatively, schools with "dry" housing might better recognize alcohol related problems on their campus and set policy intended to counter these problems among their students. Accordingly, the 49% binge rate observed in this study could reflect a significant improvement in the rate of binge drinking for these schools. Without longitudinal data, however, this analysis is beyond the scope of this study. Four schools in the sample indicated that they prohibited all alcohol use by students both on- and off-campus. These schools had a significantly lower mean binge drinking rate than other schools (10% vs. 45% respectively), suggesting that prohibition discouraged drinking among the majority of the student body; alternatively, these schools might attract students less interested in drinking. Three of the schools were religiously affiliated, and the fourth admitted primarily African-American students; both minority status and religiosity are cultural factors that have been shown to be associated with decreased substance abuse [ 45 , 46 ]. Unanticipated Policy Effects The interaction effects observed among alcohol policy variables and the presence of gambling policy on binge drinking behavior and past-year student gambling behavior presents an interesting and unanticipated finding. Because schools that have prohibitive alcohol policies and prohibitive gambling policies evidence lower mean rates of binge drinking among students than other schools, restrictive policies seem to have the intended effect of countering potentially destructive behaviors among students. However, other conditions such as cultural factors also play an important role in determining student behaviors. For example, students that choose to attend schools with rigorous policy provisions might be intrinsically more likely to refrain from excessive alcohol consumption for ethical or religious reasons. Further, five of the six schools in our sample that had both (a) a policy prohibiting alcohol at on-campus events and (b) a prohibitive gambling policy also had small enrollment (i.e., below the 50 th percentile among schools in our sample); four of these six schools were state-operated. Underlying characteristics of students who seek out small state schools might be associated with the reduced binge rates reported among these institutions. These results suggest that, despite the role for shared policy guidelines, schools will benefit from analyzing the composition of their student body and tailoring new and modified alcohol policies to students' specific characteristics. Although competing explanations prevent the establishment of concrete cause-and-effect relationships in this study, the observed interaction effects between alcohol and gambling policies provide significant impetus for future research into effective policy strategy on college campuses. Implications From an observer's perspective, it appears that the many policy inconsistencies – and policy presentation inconsistencies – observed in this study reflect reactive policymaking strategies that are not guided by empirical evidence. The evidence suggests that effective school alcohol and drug policies, and student awareness of these policies, are important for many reasons. For example, college-aged individuals in recovery are extremely vulnerable to relapse; in addition to the generally high rate of relapse during the first year of recovery [ 47 , 48 ], this circumstance exists in part because of their age and the prevalence of drug and alcohol abuse among their peers. Students need to be aware of the potential health and treatment options that are available to them on-campus if relapse occurs. Young people have not had alcohol use disorders as long as their adult counterparts because of their age; similarly, college students recovering from an alcohol use disorder have not been healing for very long. Relapse can generate harmful financial, academic and other consequences that can impart severe restrictions upon students' actions, both on and off campus. Consequently, it is valuable for students struggling with addiction to be able to access specific school policies before they enter a college or university; students who develop an alcohol problem after they are enrolled in college also need access to this information, as well as policies governing recovery-seeking. Although it is possible that policies can reduce alcohol abuse and dependence, comprehensive policies governing alcohol consumption among students hold the greatest potential to reduce pre-morbid and sub-clinical alcohol use on college campuses. Addiction models generally propose that while sub-clinical alcohol use and gambling can ultimately lead to a pathological state, pre-morbid subjects also can move away from pathology and maintain controlled behavior or abstinence [ 49 - 51 ]. This can occur through the influence of social setting attributes, including policy directives. Nevertheless, the high rate of drinking and binge drinking among college students has continued despite evidence that schools have devoted increased attention to promoting alcohol awareness and prevention recently (e.g., Wechsler et al. 2002 [ 6 ]). Various factors, such as a lack of agreement among school policies, unbalanced policies (e.g., policies that focus on punishment but not recovery), or a failure on the part of school administrators to enforce stated rules could undermine a cohesive alcohol policy and contribute to continued student drinking in the wake of school reform. It is especially important for school administrators to address risky drinking behaviors among students who do not currently have a drinking problem. Research has shown that addictive disorders originate with risk factors that always include exposure to potential objects of addiction [ 27 , 50 , 52 ]. More specifically, repeated object exposure (i.e., alcohol consumption) can combine with an individual's underlying psychosocial and neurobiological vulnerabilities, resulting in desirable subjective shifts and the potential for developing an addictive disorder [ 53 ]. While each person has unique underlying vulnerabilities that make them more or less likely to develop an alcohol use disorder, reduction of opportunities to develop such a disorder (e.g., in the form of focused regulations) will benefit all students. With ever-increasing numbers of students entering four-year colleges in the U.S., clear explanations of institutional expectations and requirements regarding substance and behavioral addictions is an essential component of reducing such behaviors among students. For example, McDonnell (1994)[ 54 ] suggests that "alcohol education should be seen as part of the education of character" (p. 45). DFSCA provisions have been effective in stimulating alcohol education and policy development on college campuses; however, currently, some schools might not be providing reliable and accurate information about addictive behaviors, as evidenced by the inconsistent nature of the alcohol and gambling policies observed in this study. In the absence of a similar federal mandate requiring gambling education on college campuses, other drastic federal measures (e.g., the proposed ban on all collegiate sports betting in Nevada[ 55 ]) have been proposed to reduce student gambling. Such proposals are problematic because these efforts restrict the freedom of responsible gamblers and place policy pressure on states instead of schools, thereby diverting focus from those in need of effective regulation (i.e., college campuses). Scientific guidelines toward school regulation, similar to those proposed for federal regulation by OMB, will provide school administrators a solid foundation for creating comprehensive alcohol and gambling policies. It is the responsibility of administrators to make the first steps toward scientific selection of school policies and take substantive and definitive measures to increase addiction awareness and recovery among students. Caveats Several methodological limitations to this study should be noted. First, this study relied solely upon written policy materials. Although we conducted an exhaustive search of the Internet and requested hard materials from schools, it is possible that we failed to identify some publicly available policy materials that could have provided additional information about schools in our sample. In addition, we identified our policy variables by examining existing policies and identifying relevant regulatory components; however, others attempting to replicate this study might identify and measure different aspects of policy and, consequently, obtain different results. Further, this study considered official alcohol policy content across a sample of U.S. colleges; however, assessing official policy provides no data on whether or how schools enforce their stated policies. It is possible, and in many cases even likely, that schools rely on informal rules and established precedents to govern alcohol use or gambling violations. Due to the complex nature of the relationships between school administrators, students, parents and legal authorities, many schools likely assess alcohol or gambling violations on a case-by-case basis; consequently, practice policy might be very different from the formal rules described by official school documentation. Opportunities for Future Research This study has shown that there are many opportunities and perhaps an obligation to scientifically investigate the complex relationships between college alcohol and gambling policies and addictive behaviors. Future research, for example, should focus on policy enforcement and informal policies adopted by school administrators to provide a better understanding of current college practices regarding alcohol use and abuse. In addition, as schools begin to reevaluate and amend their substance abuse policies, longitudinal research could provide insight into the effects of policy revision on student behaviors. In addition, to advance our understanding of policy effects on intemperate patterns of behavior, future college-based research will need to examine the influence of policies on more specific behaviors. New research should deconstruct the macro indices (e.g., average rates) of binge drinking to determine whether policies can impact college violence, crime, driving under the influence, etc. College alcohol and gambling policy data holds important potential for future research. For example, for students with alcohol or gambling related disorders attending colleges with policies that interfere with treatment or fail to support recovery, the rate of relapse is likely to be higher than under a more treatment favorable public policy context. Similarly, research can demonstrate that policies can influence the likelihood of early identification and intervention: under some unsympathetic regimes, students with addiction problems will not come forward for assistance or adhere to a prescribed treatment program. Finally, new research needs to show that without supportive policies to guide college staff responses to treatment seeking students, this population will miss both the college experience and the opportunity to build a healthy foundation for their future. Conclusions In this study, we analyzed a representative sample of U.S. colleges to determine the attentiveness of school policies toward students with addictive disorders and their recovery. The results encourage the development and implementation of reporting tools (e.g., a rating system) that could prove valuable as both a resource for parents of at-risk students and a vehicle to raise public awareness. Identifying trends in collegiate policymaking as well as distinguishing strong and weak policies will allow us to begin to develop evidence based guidelines, or "best practices," upon which schools can base the development of future policy. Competing interests The author(s) declare that they have no competing interests. Authors' contributions HS conceived the study, was its principal architect, and was responsible for its overall conduct and exposition. AD was responsible for the data acquisition, adjudicated the reviewer ratings, and contributed to the drafting of the manuscript. RL was responsible for the statistical design, execution, and exposition. RK provided direct assistance to AD and participated in the preparation of the manuscript. DL contributed to the study structure, data analyses, and manuscript preparation. All the authors were members of the team that critically reviewed and coded the college policies. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549515.xml
544569	An evaluation of the effects of exogenous ethephon, an ethylene releasing compound, on photosynthesis of mustard (Brassica juncea) cultivars that differ in photosynthetic capacity	Background The stimulatory effect of CO 2 on ethylene evolution in plants is known, but the extent to which ethylene controls photosynthesis is not clear. Studies on the effects of ethylene on CO 2 metabolism have shown conflicting results. Increase or inhibition of photosynthesis by ethylene has been reported. To understand the physiological processes responsible for ethylene-mediated changes in photosynthesis, stomatal and mesophyll effects on photosynthesis and ethylene biosynthesis in response to ethephon treatment in mustard ( Brassica juncea ) cultivars differing in photosynthetic capacity were studied. Results The effects of ethephon on photosynthetic rate (P N ), stomatal conductance (g S ), carbonic anhydrase (CA) activity, 1-aminocyclopropane carboxylic acid synthase (ACS) activity and ethylene evolution were similar in both the cultivars. Increasing ethephon concentration up to 1.5 mM increased P N , g S and CA maximally, whereas 3.0 mM ethephon proved inhibitory. ACS activity and ethylene evolution increased with increasing concentrations of ethephon. The corresponding changes in g s and CA activity suggest that the changes in photosynthesis in response to ethephon were triggered by altered stomatal and mesophyll processes. Stomatal conductance changed in parallel with changes in mesophyll photosynthetic properties. In both the cultivars ACS activity and ethylene increased up to 3.0 mM ethephon, but 1.5 mM ethephon caused maximum effects on photosynthetic parameters. Conclusion These results suggest that ethephon affects foliar gas exchange responses. The changes in photosynthesis in response to ethephon were due to stomatal and mesophyll effects. The changes in g S were a response maintaining stable intercellular CO 2 concentration (C i ) under the given treatment in both the cultivars. Also, the high photosynthetic capacity cultivar, Varuna responded less to ethephon than the low photosynthetic capacity cultivar, RH30. The photosynthetic capacity of RH30 increased with the increase in ethylene evolution due to 1.5 mM ethephon application.	Background Photosynthesis is controlled by several intrinsic and extrinsic factors. Of these, plant hormones have received considerable attention in the past in photosynthetic responses of plants. Ethylene is a phytohormone that influences every aspect of plant growth and development [ 1 ]. It is synthesized by the activity of 1-aminocyclopropane carboxylic acid synthase (ACS). The response of plants to ethylene depends on the sensitivity of plants to the gas. Conflicting results on the effects of ethylene-releasing compounds on net photosynthetic rate (P N ) have been reported. It has been shown to increase P N [ 2 - 7 ] or decrease it [ 8 , 9 ], but no definite reason has been assigned for this. It has been shown that the increase in P N with ethylene-releasing compounds was due to the increase in chlorophyll per unit leaf area [ 10 ] or by greater light interception [ 11 ]. In my earlier report it has been shown that alteration in photosynthesis was due to the changes in ACS activity [ 12 ]. The goal of this work was to compare stomatal and mesophyll effects on P N in response to ethephon treatment. For that, P N , stomatal conductance (g S ) and carbonic anhydrase (CA) activity were recorded. To find a possible relationship of ethylene-mediated changes in foliar gas exchange parameters, activity of ACS and ethylene evolution were also determined. The work was carried out in two cultivars of mustard previously shown to have different photosynthetic capacity [ 12 ]. Results The effects of ethephon on P N , g S and CA were found significant in both the cultivars (Figures 1 , 2 ). Ethephon at 1.5 mM increased the characteristics maximally, increasing P N by 31.8 and 41.8%, g S by 15.0 and 17.1% and CA by 84.6 and 71.4% in Varuna and RH30, respectively. Higher concentration of ethephon (3.0 mM) decreased the characteristics in both the cultivars. The ratio of intercellular to ambient CO 2 concentration (C i /C a ) was constant. Ethephon application significantly affected ACS activity and ethylene evolution, and were greatest with 3.0 mM ethephon (Table 1 ). Low photosynthetic capacity cultivar, RH30 was more responsive to ethephon than the high photosynthetic capacity cultivar, Varuna. Application of 1.5 mM ethephon increased ethylene by 52.6% in Varuna and 75.0% in RH30. Increase in ethylene with 1.5 mM ethephon was associated with the increase in P N , g S and CA. Ethylene evolution with 3.0 mM ethephon proved inhibitory for photosynthetic parameters. Discussion Maximum rates of photosynthesis were found with 1.5 mM ethephon. The increase in P N due to ethephon has been reported [ 4 - 6 ]. Increased g s and CA values in both the cultivars showed stomatal and mesophyll effects on photosynthesis. Mesophyll effects are characterized as a product of CO 2 binding capacity and the electron transport capacity. The carboxylation capacity determines the mesophyll effects [ 13 , 14 ]. Increase in CA activity at the site of CO 2 fixation exhibited the enhanced carboxylation reaction [ 15 - 17 ]. The changes in stomatal conductance due to ethephon were to maintain stable intercellular CO 2 concentration (C i ) under the given treatment. Thus, stomatal and mesophyll processes contributed to the increase in P N in response to ethephon. The ethephon-induced effects on photosynthetic parameters were mediated by ethylene evolved due to ethephon treatment. Taylor and Gunderson [ 18 ] showed a relationship between ethylene-enhanced g S and ethylene-enhanced P N . Higher concentration of ethephon (3.0 mM) decreased the P N and g S . Such condition of inhibition of P N by ethylene-releasing compound has been observed by Kays and Pallas [ 8 ] and Rajala and Peltonen-Sainio [ 9 ]. In all these studies ethylene has been attributed to the changes in P N due to its effect on g S . Mattoo and White [ 19 ] reported that ethylene affected CO 2 assimilation and the plant responded depending on the tissue concentration. On the similar lines, Dhawan et al . [ 20 ], Kao and Yang [ 21 ] and Grodzinski et al . [ 22 ] reasoned that decrease in CO 2 regulated P N and was related to ethylene evolution. In the present study, low photosynthetic cultivar, RH30 responded more to ethephon than the high photosynthetic cultivar, Varuna. In control plants, lesser ethylene evolution in RH30 than Varuna was responsible for lesser P N . As the ethylene evolution increased with ethephon application, the capacity of RH30 for P N also increased resulting in higher per cent increase in P N than the Varuna. An increase of 75% ethylene in RH30 due to 1.5 mM ethephon increased P N by 41.8%, whereas 52.6% increase in ethylene in Varuna due to the same treatment increased P N by 31.8%. Earlier strong positive correlation between ACS activity and P N has been shown [ 12 ]. It therefore, appears possible that the threshold value for ethylene with 1.5 mM ethephon was comparable to that which elicits the ethylene-mediated hormonal responses, which differ with the cultivars inherent capacity of physiological processes. It is that there is some requirement of ethylene for optimum response. Low and high concentration represent the two ends of an optimum curve, promoting at low concentration and inhibiting at high. Conclusions This study shows that ethephon affects P N in both high and low photosynthetic capacity cultivars, Varuna and RH30. In both the cultivars, changes in P N were due to stomatal and mesophyll effects. Ethephon-induced P N was attributed to ethylene evolution. The high photosynthetic capacity cultivar, Varuna responded less to ethephon than the low photosynthetic capacity cultivar, RH30. The low P N of RH30 was due to low level of ethylene. The low photosynthetic capacity of RH30 could be enhanced to give higher P N through increase in ethylene evolution. However, for both the cultivars there is a range of physiologically active concentration of ethylene beyond which it exerts inhibitory effects. Methods Two cultivars of mustard ( Brassica juncea L. Czern & Coss.), namely Varuna (high photosynthetic capacity) and RH30 (low photosynthetic capacity) were grown from seeds in 10 m 2 field plots in complete randomized design with five replications. At seedling establishment a plant population of 12 plants m -2 was maintained and recommended plant cultivation procedures were adopted. A uniform recommended soil application of 18 g N, 3 g P and 3 g K m -2 was given at the time of sowing so as the nutrients were non-limiting. At 30 d after sowing, 0, 0.75, 1.5 and 3.0 mM ethephon (2-chloroethyl phosphonic acid) was sprayed with a hand sprayer. Ethephon is a direct ethylene source when applied to plants and elicits response identical to those induced by ethylene gas [ 23 , 24 ]. Since ethephon on hydrolysis releases ethylene and phosphorus, therefore equivalent amount of phosphorus present in 3.0 mM ethephon was given to all treatments including control to nullify the effects of phosphorus. At 45 d after sowing (15 d after ethephon treatment) P N , g S , C i , CA activity, ACS activity and ethylene evolution were determined. Measurement of photosynthetic parameters P N , g S and C i were measured using infrared gas analyzer (LiCOR 6200, Lincoln, NE) on fully expanded upper most leaves at saturating light intensity on four plants from each replicate. The atmospheric conditions during the experiment between 1100–1200 h were: photosynthetic active radiation about 1050 μmol m -2 s -1 , relative humidity 64% and temperature 23°C, atmospheric CO 2 concentration 360 μmol mol -1 . Measurement of carbonic anhydrase activity The leaves used for photosynthesis measurement were selected for CA activity determination. CA was measured by the method of Dwivedi and Randhava [ 25 ]. Leaves were cut into small pieces in 10 mL of 0.2 M cystein at 4°C. The solution adhering to the leaf surface was removed and immediately transferred to a tube having 4 mL phosphate buffer (pH 6.8). A 4 mL of 0.2 M sodium bicarbonate in 0.002 M sodium hydroxide and 0.2 mL of 0.002% bromothymol blue was added to the tube. The tubes were kept at 4°C for 20 min after shaking. Liberated CO 2 during the catalytic action of enzyme on sodium bicarbonate was estimated by titrating the reaction mixture against 0.05 N hydrochloric acid. Measurement of ACS activity and ethylene evolution Activity of ACS was measured adopting the methods of Avni et al . [ 26 ] and Woeste et al . [ 27 ]. Leaf tissue was grind in 100 mM N-2 hydroxyethylenepiperazine N-2 ethanesulfonic acid buffer (pH 8.0) containing 4 mM dithiothreitol, 2.5 mM pyridoxal phosphate and 25% polyvinylpolypyrrolidone. The preparation was homogenized and centrifuged at 12000 g for 15 min. One mL of the supernatant was placed in a 30 mL tube and 0.1 mL of 5 mM S-adenosyl methionine (AdoMet) was added. This was incubated for 1 h at 22°C. The 1-aminocyclopropane carboxylic acid formed was determined by its conversion to ethylene by the addition of 0.1 mL of 20 mM HgCl 2 followed by 0.1 mL of 1:1 mixture of saturated NaOH/NaCl and incubated on ice for 10 min, and ethylene evolution was measured on a gas chromatograph. For control set AdoMet was not added. For ethylene evolution 5 mL of gas phase was removed with a syringe and ethylene was measured on a gas chromatograph (GC 5700, Nucon, New Delhi) equipped with 1.8 m Porapack N (80/100 mesh) column, a flame ionization detector and an integrator. Nitrogen was used as carrier gas. The flow rates of nitrogen, hydrogen and oxygen were 0.5, 0.5 and 5 mL s -1 , respectively. The oven temperature was 100°C and detector was at 150°C. Ethylene identification was based on the retention time and quantified comparing with the peaks from standard ethylene concentrations. Data analysis Data were analyzed statistically and standard error of the mean value was calculated. Analysis of variance was performed to identify the significant differences among treatments at P < 0.05 [ 28 ]. Abbrevations Adomet – S-adenosyl methionine; ACS – 1-aminocyclopropane carboxylic acid synthase; CA – carbonic anhydrase; C i – intercellular CO 2 concentration; g S – stomatal conductance; P N – net photosynthetic rate	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544569.xml
524504	Comparison of Misoprostol and Dinoprostone for elective induction of labour in nulliparous women at full term: A randomized prospective study	Background The objective of this randomized prospective study was to compare the efficacy of 50 mcg vaginal misoprostol and 3 mg dinoprostone, administered every nine hours for a maximum of three doses, for elective induction of labor in a specific cohort of nulliparous women with an unfavorable cervix and more than 40 weeks of gestation. Material and Methods One hundred and sixty-three pregnant women with more than 285 days of gestation were recruited and analyzed. The main outcome measures were time from induction to delivery and incidence of vaginal delivery within 12 and 24 hours. Admission rate to the neonatal intensive care unit within 24 hours post delivery was a secondary outcome. Results The induction-delivery interval was significantly lower in the misoprostol group than in the dinoprostone group (11.9 h vs. 15.5 h, p < 0.001). With misoprostol, more women delivered within 12 hours (57.5% vs. 32.5%, p < 0.01) and 24 hours (98.7% vs. 91.4%, p < 0.05), spontaneous rupture of the membranes occurred more frequently (38.8% vs. 20.5%, p < 0.05), there was less need for oxytocin augmentation (65.8% vs. 81.5%, p < 0.05) and fewer additional doses were required (7.5% vs. 22%, p < 0.05). Although not statistically significant, a lower Caesarean section (CS) rate was observed with misoprostol (7.5% vs. 13.3%, p > 0.05) but with the disadvantage of higher abnormal fetal heart rate (FHR) tracings (22.5% vs. 12%, p > 0.05). From the misoprostol group more neonates were admitted to the intensive neonatal unit, than from the dinoprostone group (13.5% vs. 4.8%, p > 0.05). One woman had an unexplained stillbirth following the administration of one dose of dinoprostone. Conclusions Vaginal misoprostol, compared with dinoprostone in the regimens used, is more effective in elective inductions of labor beyond 40 weeks of gestation. Nevertheless, this is at the expense of more abnormal FHR tracings and more admissions to the neonatal unit, indicating that the faster approach is not necessarily the better approach to childbirth.	Introduction Induction of labor is carried out for maternal and fetal indications. One of the most common indications is prolonged pregnancy [ 1 ]. Recent studies have suggested that by continuing pregnancy beyond 41 weeks, there is a statistically significant higher perinatal morbidity and mortality as well as an increased risk to the mother [ 2 , 3 ]. Thus, there is a growing body of evidence suggesting the elective induction of labor at 41 weeks of gestation instead of expectant management [ 4 - 6 ]. Prostaglandin analogues, dinoprostone (PGE 2 ) and misoprostol (PGE 1 ), are widely used in "induction of labor" practice for ripening the cervix and stimulating uterine contractions in order to achieve vaginal delivery. Although dinoprostone has been approved by the FDA for cervical ripening in women at or near term, misoprostol is not currently approved for such use by the FDA, although it has the advantages of lower cost, no need for refrigeration and probably higher efficacy. Several studies have demonstrated a higher efficacy of vaginally administered misoprostol compared to vaginal dinoprostone for both cervical ripening and labor induction [ 7 - 19 ]. The Cochrane Pregnancy and Childbirth Group, having reviewed 45 randomized studies, concluded that vaginal misoprostol (25 to 100 mcg) was more effective than oxytocin or dinoprostone at the usual recommended doses for induction, but with increased rates of uterine hyperstimulation both with and without associated fetal heart rate (FHR) changes, as well as meconium stained fluid [ 9 ]. Most of the studies included in the previous meta-analysis used the 50 mcg dose for misoprostol at a maximum interval of six hours between the repeated doses, always resulting in higher rates of hyperstimulation. However, it is difficult to interpret previously published studies comparing misoprostol with dinoprostone for induction of labor since the majority of them are not double-blinded [ 9 ] and they have included both complicated and uncomplicated pregnancies, multiparous women with nulliparous as well as a wide gestational age (GA) range (37–42 weeks). Moreover, to reduce the risk of side effects, one can either decrease the dose of the drug [ 7 , 19 ] or prolong the dosage interval [ 14 , 15 , 19 ]. In addition, Alexander et al. have recently shown that in prolonged pregnancies it was not the induction per se that would increase the risk for caesarean section (CS), but patients related risk factors such as nulliparity and undilated cervix and the use of epidural analgesia [ 20 ]. This study was undertaken to compare the efficacy of vaginal misoprostol (50 mcg) with that of vaginal dinoprostone (3 mg) when both are administered at an interval of nine hours between repeated doses in a well-homogenized cohort of full term pregnancies (nulliparous women with an unfavorable cervix and without pregnancy complications). Material and Methods Between March 1, 2001 and July 16, 2003, 163 women were recruited for the study: 80 women in the misoprostol group and 83 women in the dinoprostone group. All of the women were recruited at Ioannina University Hospital, a tertiary referral center for high-risk pregnancies, with about 1600 deliveries a year. The Ethics Committee of the University of Ioannina approved the study and all participants gave their written informed consent after they had been made aware of the purpose of the study. Although the main indication was prolonged pregnancy, some of the inductions were performed at the patient's request after consultation at 40 weeks of gestation, (without any medical indications) and only if they had not delivered by the 285 th day of gestation. A sequence from a computerized random number generator was used for the allocation of patients to each group. The vaginal administration of prostaglandins was performed by one of the resident doctors on duty, who was not involved in managing these women in labor or delivery. The study was double blind, since the patients were not aware of which type of medication was used, and the deliveries were then performed by two gynecologists blinded to the induction regimen utilized. Inclusion criteria were: 1) age>18 years old 2) nulliparity, 3) accurate dating of gestation, including crown rump length (CRL) measurements in the first trimester of pregnancy, 4) singleton viable pregnancy, 5) gestational age ≥ 285 days, 6) cephalic presentation, 7) unfavorable cervical status defined as a Bishop score (BS) of ≤5, 8) intact membranes, 9) reactive non-stress test (NST). Exclusion criteria were: 1) known contraindications to receiving prostaglandins, 2) placenta previa, 3) prior uterine surgery and 4) any antenatal complications. GA was estimated by ultrasound biometry (via CRL measurements in the first trimester of pregnancy) in cases where there were more than 3 days difference from that obtained from the last menstrual period (LMP) [ 21 ]. Uterine tachysystole was defined as more than five contractions per 10 minutes, uterine hypertonus as when one contraction lasted more than 2 minutes and hyperstimulation syndrome as the presence of non-reassuring FHR tracing combined with either tachysystole or hypertonus. Non-reassuring FHR patterns were defined as persistent or recurring episodes of severe variable decelerations, late decelerations, prolonged fetal bradycardia or a combination of decreased beat-to-beat variability and a decelerative pattern [ 22 ]. A NST to ensure the well-being of the fetus was performed for each patient at the time of recruitment and admission to the hospital (at least 285 days of gestation) and one hour before the application of the prostaglandin. After the reassessment of the cervical BS, either 50 mcg misoprostol, or 3 mg dinoprostone was administered in the posterior vaginal fornix at 23:00 hours. The NST was repeated (duration of two hours) after 1 h and 5 h. If the woman was in active labor, the membranes spontaneously ruptured or the FHR not reassuring, the patient was transferred to the labor room. Otherwise, a second BS evaluation was carried out the next morning at 08:00 am (after 9 hours). If the cervix was favorable, (BS ≥ 5), the patient was admitted to the labor ward where oxytocin augmentation was carried out if the uterine contractions were unsatisfactory and amniotomy was performed when appropriate. If the cervix was still unfavorable, a second dose of misoprostol or dinoprostone was given and the same evaluation steps as described above were followed. After a total of 18 h had elapsed, non-responders were given a third dose of prostaglandin. When the third dose was insufficient for initiating spontaneous labor, a trial of labor was offered with oxytocin infusion and if no progress was achieved within 6 hours (based on digital assessment of the BS), the patient underwent a CS. The outcome measures were divided into "obstetrical" and "neonatal". The primary outcome measures were time from induction to delivery and incidence of vaginal delivery within 12 and 24 hours; the secondary outcomes were the CS rate, the need for oxytocin augmentation, the incidence of meconium stained amniotic fluid, the incidence of uterine tachysystole, abnormal FHR tracings, maternal morbidity, the admission to neonatal intensive care within 24 hours and neonatal arterial cord ph, base deficit. Statistical analysis was performed using SPSS version 11 software. The Chi square test and Fisher's exact test were used to analyze nominal variables in the form of frequency tables. Normally distributed (Kolmogorov-Smirnov Test with Lilliefors correction) metric variables were tested by the T-test for independent samples, while non-normally distributed metric variables were analyzed by the Mann-Whitney U test. All tests were two-tailed with a confidence level of 95% (p < 0.05). Values are expressed as mean ± standard error (SEM). Results The two groups were comparable in terms of patients' age (28.1 years vs.27.5, p > 0.05) and indication for induction (prolonged pregnancy 81.2% vs.78.3%, p > 0.05; social 18.8% vs. 21.7 %,) in the misoprostol and dinoprostone groups, respectively. Gestational age (286 days, range:285–292) and the preinduction BS (2.7 ± 0.1) in the misoprostol group were also comparable to the dinoprostone group (286 days, range:285–293) and (2.9 ± 0.1), respectively. Obstetrical outcome The induction-delivery interval was significantly shorter (11.9 h vs. 15.6 h, p < 0.001) in the misoprostol group, with even less need for a second or third dose (7.5% vs. 22%, p < 0.05) compared to dinoprostone. With misoprostol, more women delivered within 12 h (57.5% vs. 32.5%, p < 0.01) and almost all of the women delivered within 24 h (98.8% vs. 91.6%, p < 0.05). In addition, spontaneous rupture of the membranes occurred more often after the administration of misoprostol (p < 0.05) and there was a reduced need for oxytocin augmentation in labor: 65.8% vs. 80.7% with dinoprostone (p < 0.05). However, uterine tachysystole (p < 0.05)) and meconium stained amniotic fluid (p > 0.05) occurred more often in the misoprostol group as did abnormal heart rate tracing (22.5% vs.12%, p > 0.05) (Table 1 ). Table 1 Obstetrical Outcomes Misoprostol n = 80 (%) Dinoprostone n = 83 (%) Statistical significance Time from induction to delivery (h ± SEM) 11.9 ± 0.6 15.6 ± 0.7 p < 0.001 Delivery < 12 h 46 (57.5%) 27 (32.5%) p < 0.01 Delivery < 24 h 79 (98.8%) 76 (91.6%) p < 0.05 Number of doses Single dose 74 (92.4%) 65 (78.3%) p < 0.05 Second dose 6 (7.5%) 17 (20.5%) Third dose 0 (0%) 1 (1.2%) Required oxytocin augmentation 53(65.8%) 67(80.7%) p < 0.05 Spontaneous rupture of membranes 31 (38.8%) 17 (20.5%) p < 0.05 Meconium stained AF 15 (18.8%) 7 (9.6%) NS Abnormal FHR 18 (22.5%) 10 (12%) NS Uterine Tachysystole 10 (12.6%) 3 (3.6%) p < 0.05 Uterine Hyperstimulation 2 (2.5%) 1 (1.2%) NS FHR = fetal heart rate. AF = amniotic fluid NS = not significant (p > 0.05) SEM = standard error of the mean In both groups, the majority of women had vaginal delivery, 92.5% with misoprostol, and 86.7% with dinoprostone. There was no statistically significant difference between the two groups with regard to the CS rate (Table 2 ). There were no uterine ruptures or other major maternal complications resulting from the use of either of the prostaglandins. There was only one wound infection with dinoprostone, one woman in each group had delayed discharge due to persistent pyrexia and two women in the dinoprostone group required uterine packing (insertion of tampons within the uterine cavity) due to postpartum bleeding. Table 2 Mode of delivery and indications for Caesarean section Misoprostol n = 80 (%) Dinoprostone n = 83 (%) Statistical significance Vaginal 74 (92.5%) 72 (86.7%) NS 1 Spontaneous vaginal 46 (57.5%) 52 (62.6%) NS Vacuum assisted vaginal 28 (35.0%) 20 (24.1%) NS Caesarean section 6 (7.5%) 11 (13.3%) NS Nonreassuring FHR 2 4 (5.0%) 6 (7.2%) NS Failed induction 0 (0.0%) 1 (1.2%) NS Lack of labor progress 1 (1.3%) 2 (2.4%) NS Cephalopelvic disproportion 1 (1.3%) 2 (2.4%) NS 1 NS = not significant 2 FHR = fetal heart rate. Neonatal outcome More neonates in the misoprostol group had first minute Apgar scores lower than 7 (12.6% vs. 6.1%, p > 0.05), or needed neonatal resuscitation (11.4% vs. 9.9%, p > 0.05) but none of the babies had birth asphyxia [ 23 ]. The mean cord pH and the base deficit were comparable in the two groups. No neonate had meconium aspiration syndrome. Two neonates in the dinoprostone group had clavicle fracture (Table 3 ). There was no statistically significant difference in the number of neonates admitted to neonatal intensive care within 24 hours after delivery, between the misoprostol and dinoprostone groups (6.3% vs. 3.6% p > 0.05) (Table 4 ). Table 3 Neonatal Outcomes Misoprostol n = 80 (%) Dinoprostone n = 83 (%) Statistical significance Birth weight (g) 1 3275 ± 430 3373 ± 390 NS Perinatal death 0 1(1.2%) NS Neonatal resuscitation 9 (11.3%) 9 (10.8%) NS O 2 Supplementation 1 2 Ambou ventilation 7 6 Intubation in labor room 1 1 Apgar score < 7 1 min 10 (12.5%) 5 (6.0%) NS 5 min 1 (1.3%) 0 Cord blood pH (arterial) 1 7.28 ± 0.05 7.27 ± 0.05 NS Base deficit 1 5.0 ± 2.3 5.7 ± 3.2 NS 7.01 < cord pH < 7.20 3 (3.8%) 4 (4.8%) NS 10 < base deficit < 16 1 (1.3%) 2 (2.5%) NS Hyperbilirubinemia 2 9 (11.3%) 5 (6.0%) NS Birth trauma 3 0 2 (2.5%) NS 1 Values expressed as mean ± SD 2 Excluding pathological causes of icterus 3 Both were clavicle fractures Table 4 Admission to Neonatal Intensive Care Unit N (%) DA 1 Delivery 2 Indication Diagnosis HDs 3 Within 24 hours Misoprostol† 5 (6.3%) 01 VVD Rule out infection Elevated CRP 4 WBC 5 07 01 VVD Respiratory distress Respiratory infection 11 01 CS Rule out asphyxia Infection 10 01 SVD Respiratory distress Work up for infection 03 01 VVD Respiratory distress Work up for infection 04 Dinoprostone† 3 (3.6%) 01 VVD Respiratory distress Atelectasis 06 01 SVD Rule out asphyxia Infection 10 01 SVD Respiratory distress Respiratory infection 10 After 24 hours Misoprostol† 6 (7.2%) 02 VVD Rule out infection WBC in CSF 6 10 07 VVD Hyperbilirubinemia Urinary infection 07 04 SVD Hyperbilirubinemia Icterus 04 08 SVD Infection Respiratory infection 10 03 SVD Feeding difficulty WBC in CSF 20 05 CS Hyperbilirubinemia Icterus 04 Dinoprostone† 1 (1.2%) 17 SVD Fever WBC in CSF 15 1 DA = day of admittance 2 SVD = Spontaneous Vaginal Delivery, VVD = vacuum assisted vaginal delivery, CS = Caesarean section 3 HDs = hospitalization days 4 CRP = C-reactive protein 5 WBC = white blood cell 6 CSF = cerebrospinal fluid † not significant (p > 0.05) A 28-year-old woman at 41 weeks of gestation had an unexplained stillbirth after receiving a single dose of dinoprostone. Seven hours later she had a cardiotocogram without abnormal FHR patterns and regular contractions of the uterus were evident. We decided to move her to the labor ward and within half an hour of entry, no cardiac activity of the fetus was found. During these 30 minutes FHR monitoring had been discontinued, as it was not included in the study design. We decided to let her attempt vaginal delivery. An amniotomy was performed and the amniotic fluid was found to be clear and a vaginal delivery was achieved within 6 hours. Direct examination of the fetus, the placenta and the umbilical cord (UC) showed only a thin UC with excess twisting around its axis. The anatomopathology examination of the fetus revealed no abnormality except a microscopically decreased Wharton's jelly. Discussion Nowadays, induction of labor is more widely used than ever before [ 24 , 25 ]. Recent studies have shown that this increase is mainly due to a rise of inductions for marginal or elective reasons. The common indications are elective induction and postdate pregnancy often applied to gestations of 40 to 41 weeks [ 1 , 25 ]. Mongelli et al. have also shown that for the detection of post-maturity there is no advantage in using menstrual dates when ultrasound biometry is available [ 26 ]. Women may experience distress when labor has not started by the expected date [ 27 ] and obstetricians have to withstand pressure from these patients as well as the temptation to use prostaglandins earlier. Appropriate evaluation of the pregnancy and consultation with such patients will lead to the correct selection of those who will benefit most from a labor induction, thus eliminating the risk of post-maturity to the fetus without inducing fetal distress during labor. To the best of our knowledge, the present study is the only one that compares misoprostol and dinoprostone in such well-homogenized groups. All of the women were nulliparous with intact membranes and at more than forty weeks' gestation with no antenatal complications and all had an unfavorable cervix. In these carefully selected patients, misoprostol at the dose used not only shortened the time between induction and delivery (11.9 vs. 15.6 h), but it also was significantly more effective than dinoprostone. The positive point was that this result was achieved with a very low CS rate even in the dinoprostone group, (7.5%, and 13.3%), respectively. A difference of 5% in favor of misoprostol, although not statistically significant, might have clinical importance in terms of patient health and cost effectiveness. Although in the recent large meta-analysis [ 9 ] published by the Cochrane Library, the CS rates were inconsistent, they tended to be lower with misoprostol; an earlier study by Sanchez-Ramos et al. found a statistically significant difference in favor of misoprostol [ 28 ]. In addition, our results for this GA window are reassuring with regard to concerns that have been raised from previous retrospective studies reporting an increased risk of Caesarean delivery in nulliparous women when elective inductions are performed [ 29 , 30 ]. Even though misoprostol improves the kinetics of labor during induction in a more efficient way than dinoprostone, concerns persist with respect to intrapartum fetal "wellbeing". In order to avoid uterine hyperstimulation and abnormal FHR tracings, we used for first time in the literature, a 9 h interval between the prostaglandin doses. Although we indeed achieved a low rate of uterine hyperstimulation syndrome (2.5% with misoprostol and 1.2% with dinoprostone, respectively), we still noticed a trend towards a high rate of abnormal FHR tracings during induction with misoprostol. Our findings, in accordance with the previous Cochrane metanalysis [ 9 ], showed that with misoprostol there was an increased probability of meconium staining of amniotic fluid as well as of uterine tachysystole and of abnormal FHR tracings. In the misoprostol group, the majority of women also underwent either a CS or a vacuum operative delivery due to non-reassuring FHR. If neonatal outcomes such as neonatal resuscitation, low Apgar score in the first minute and admittance to the neonatal unit within the first 24 hours (none of the above were statistically significant but they were more frequent with misoprostol) are taken into account, misoprostol may increase these complications in labor. Thus, although our sample size cannot determine safety, misoprostol use is associated with a higher chance of admittance to the neonatal unit within 24 hours even in the absence of asphyxia. This evidence indicates that the faster approach to childbirth is not necessarily the better one. Attempting an explanation to the aforementioned side effects of misoprostol use and taking into account other reports [ 9 , 31 , 32 ], it appears that the increase in clinically relevant adverse effects is not only misoprostol related but it may be dose dependent. Lyons et al. have recently shown in term pregnant rats that a higher dose of misoprostol is needed to induce PGE2 secretion in the cervix than in the myometrium, and furthermore that EP3 receptors (prostaglandin E2 receptors) are differentially expressed in the myometrium (increased) than in the cervix (unaltered) in response to misoprostol [ 33 ]. The above findings indicate that misoprostol not only acts better on the myometrium than on the cervix, but an even higher dose is needed in order to ripen the cervix. Thus, it seems reasonable that increasing the interval between repeated misoprostol doses should reduce the risk of an asynchrony between a well or even hyper-stimulated uterus and a still not efficiently ripened cervix. Misoprostol probably has a large inter-patient variability in terms of pharmacokinetics, but it is also probable that the 50 mcg dosage may induce asynchrony between immature cervix effacement and uterine contractions, resulting in a more rapid but also more "stressful" labor. Based on these findings, we would propose, in future, a slight modification of the misoprostol protocol used in this study. An initial lower dose of misoprostol (20–25 mcg), followed by 50 mcg should be considered in trying to achieve priming of the cervix without inducing such high uterine contractility and neonatal complications. Indeed, in a recent study comparing 25 mcg misoprostol with 1 mg dinoprostone administered vaginally every four hours, the admission rate to neonatal intensive unit was significantly lower in the misoprostol group [ 34 ]. It still has to be mentioned that in many of our participants, the vertex was not engaged in the pelvic inlet on the day of admittance and this should have been included as an independent risk factor in the initial study design. The exact cause of the stillbirth in the dinoprostone group remains unclear, emphasizing thus, the need for continuous FHR monitoring during labor induction if regular uterine contractions persist [ 35 , 36 ]. Conclusions To conclude, 50 mcg misoprostol at a 9 h interval is more highly effective in promoting cervical ripening and in inducing labor, compared to dinoprostone. However, certain aspects concerning fetal well being during labor induction remain questionable. Larger prospective studies comparing elective induction to expectant management after a completed 40-week gestation (on the basis of early ultrasound biometry) might reveal a subgroup of women, such as nulliparous with an unfavorable cervix, who might benefit from an elective induction, preferably with a 25 mcg misoprostol initial dose. Authors' contributions E.P: conceived of the study, participated in the sequence alignment, performed the statistical analysis and drafted the manuscript. N.P: conceived of the study, and performed the labor inductions A.D: performed the neonatal examination and follow up S.A: performed the neonatal examination and participated in the neonatal data analysis C.V: was the midwife involved in women allocation, medications preparation and labour data registration T.S: conceived of the study and data analysis E.P: conceived of the study and coordinated the study K.Z: conceived of the study, performed the labor inductions and coordinated the study	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524504.xml
544582	Potential action of androstenedione on the proliferation and apoptosis of stromal endometrial cells	Background Hyperandrogenic conditions have been associated with a high prevalence of endometrial pathologies related to cell survival. However, the action of androgens on proliferation and apoptosis in endometrial cells is poorly understood. Therefore, the aim of the present study was to evaluate the effect of androstenedione on cell proliferation, cell death and expression of estrogen receptor (ER) isoforms and proteins related to apoptosis in endometrial cells using two in vitro experimental approaches. Methods The endometrial tissue was obtained from 20 eumenorrheic women [28.7 (25 – 35) years] during the early secretory phase. We analyzed cell proliferation (immunohistochemistry of Ki-67 and spectrophotometric assay); apoptosis (DNA fragmentation (TUNEL) and Annexin V-FITC binding); ER-alpha, ER-beta bcl-2 and bax mRNA abundance (RT-PCR) in explants and isolated endometrial epithelial (EEC) and stromal cells (ESC) incubated with androstenedione 1 micro mol/l (A4) or A4 plus hydroxyflutamide 10 micro mol/l (F) for 24 h. Results In explants, A4 induced an increase of cell proliferation and a decrease on apoptosis in the stromal compartment (p < 0.05). In isolated ESC, proliferation augmented with A4 (p < 0.05), whereas, no significant modifications in the expression of ER-alpha, ER-beta bcl-2 and bax nor in the apoptotic index were observed. In EEC, A4 increase the ER-beta mRNA abundance (p < 0.05) and a decrease of the bcl-2/bax ratio (p < 0.05), without an increase in the apoptotic index. Hydroxyflutamide reverted the effect of androstenedione on the parameters described. Conclusions These results indicate that androstenedione may modulate cell survival, expression of ER-beta and proteins related to apoptosis, suggesting a potential mechanism that associates the effect of hyperandrogenemia on the endometrial tissue.	Background Polycystic ovary syndrome (PCOS) is a complex endocrine-metabolic disorder, associated to hyperandrogenism, menstrual disturbances and in many cases to insulin resistance [ 1 , 2 ]. It has been observed that in some of PCOS women, the endometrium is thicker than that of normal cycling women [ 3 ] and a higher prevalence of endometrial hyperplasia and adenocarcinoma have also been described in these women [ 4 - 6 ]. The latter may indicate that the mechanisms that regulate the process of cell survival may be disrupted in the endometrium of PCOS women. Recently, we have shown that the expression of proteins involved in the regulation of apoptosis in PCOS endometria were altered [ 7 ]. Besides, we and other investigators demonstrated an elevated expression of the estrogen receptor (ER) and its co-activators in endometria of women bearing this syndrome [ 7 - 9 ]. Nevertheless, in those studies it was difficult to establish the exact contribution of androgens as a regulatory steroid of endometria of PCOS women, since multiple factors could be affecting their endometrial function, including hyperinsulinemia and the possible contribution of the unopposed effect of estrogens [ 10 , 11 ]. Therefore, in vitro experimental models such as tissue and cell culture may constitute interesting approaches to determine the potential role of androgens in the regulation of endometrial cell survival. Early reports in isolated endometrial stromal cells (ESC) have shown that androgens can induce decidualization and inhibition of the expression of ER and progesterone receptors [ 12 , 13 ]. Moreover, in endometrial epithelial cells (EEC), androgens altered the expression of proteins related to uterine receptivity [ 8 ] and induced a decrease in the proliferation capacity of those cells [ 14 ]. Cell proliferation and apoptosis of the endometrium are importantly regulated by the expression of ER [ 15 ], which exists in two major subtypes, estrogen receptors alpha (ERα) and estrogen receptor beta (ERβ). The two isoforms of ER derive from separate genes, with different ligand binding affinities and the response of the tissue to estrogens will depend upon their relative concentrations [ 16 ]. On the other hand, in some tissues including the human endometrium, the control of apoptosis has been associated to proteins related to the Bcl-2 family, like Bcl-2 that promotes cell survival and Bax which is an inducer of apoptosis [ 17 , 18 ]. Evenmore, other proteins are involved in the machinery of cell death like caspases which are associated with the cleavage and thus, breakdown of cell structure [ 19 ]. In regard to this issue, we have demonstrated that the expression of bcl-2 and bax is increased in the stromal compartment of the endometrium of PCOS women, but we could not observe an increase in the apoptotic index in the endometria of these women [ 7 ]. Previous reports have shown that androstenedione is an important androgen detected in the endometrial tissue [ 20 , 21 ]. Therefore, based on the amount of androgens normally found in endometria and their potential importance in alterations of the endometrial cell survival in PCOS women, the objective of the present study was to evaluate the effect of androstenedione on cell proliferation, apoptosis and the expression of ER isoforms and proteins related to apoptosis using two in vitro experimental approaches. Methods Subjects Endometrial tissue was obtained with pipelle suction curette from the corpus of the uteri of 20 women with regular menstrual cycles, aged 28.7 (25 – 35) years, at the time of bilateral tubal ligation at the San Borja-Arriarán Clinical Hospital, National Health Service, Santiago, Chile. The tissue was obtained during the early secretory phase since the cells obtained from this phase maintain a high degree of proliferation capacity [ 22 ]. None of these women had taken oral contraceptives or other medications for at least 6 months before starting the study. Women who had evidences of PCOS, endometriosis and/or endometrial hyperplasia were excluded. This investigation was approved by the Institutional Ethics Committee of the San Borja-Arriarán Clinical Hospital and an informed written consent was obtained from all subjects. Culture System Explants Human endometrium was cut into slices (20 to 50 mg wet weight) and incubated in 1 mL of Hank's media supplemented with 2 mmol/L glutamax-I (BRL, Life Technology, Bethesda, MD, USA), insulin-transferrin-selenium (ITS) solution (BRL), 0.1% w/v bovine serum albumin (BSA; Sigma Chemical Co., St Louis, MO, USA), 26 mmol/L of NaHCO 3 , 25 mmol/L of HEPES aminoacids solution, 100 IU/ml of penicillin, and 5 mg/mL of streptomycin (Sigma). Incubation was performed during 6 h at 37°C in 5% CO 2 /air in humidified atmosphere in the absence or presence of androstenedione 10 -7 to 10 -5 M (Sigma) or androstenedione 10 -6 M plus hydroxyflutamide 10 -5 M (Sigma), the latter added 30 min before androstenedione. After incubation, one piece from both basal and treated tissue explants were frozen in liquid N 2 and maintained at -70°C for RT-PCR protocols. Another piece was fixed in 4% buffered formaldehyde for 24 h, embedded in paraffin, and cut into 5 μm thick sections before in situ analysis of apoptosis and immunohistochemistry. Cells The cells were separated and purified according to previously described methods [ 23 ]. Briefly, the tissue was cut into small pieces and suspended in Dulbecco's modified Eagle medium (DMEM) (GibcoBRL), collagenase (370 IU/100 mg tissue) (Worthington, Biochemical Corp. Freehold, NJ, USA) and DNAse (14 KU/100 mg tissue) (Sigma) during 1 h at 37°C. Epithelial cells, predominantly from glands, were separated from ESC by decantation and the supernatant containing the ESC was filtered, centrifuged and the cell pellet washed in DMEM. Stromal cells were incubated in appropriate cell culture media (ESC media) (DMEM/MCDB-105 (3:1 v/v), 2% charcoal stripped fetal bovine serum (FBS) (GibcoBRL), insulin-transferrin-selenium (ITS) solution, 2 mmol/L glutamax-I (GibcoBRL), 0.25 μg/mL ascorbic acid (GibcoBRL), 0.25 μg/mL fungizone (GibcoBRL), 100 IU/mL penicillin and 5 mg/mL streptomycin at 37°C in 5% CO 2 /air in humidified atmosphere until confluence. The glands cells were washed in DMEM and incubated for 1 h (30 min in each side of the culture flask), and then cultured in EEC culture media (DMEM/MCDB-105 (3:1, v/v), 10% charcoal stripped FBS, 2 mmol/L glutamax-I, 0.25 μg/mL ascorbic acid, 0.5 mg% insulin (Sigma), 1 μg% transferrin (Sigma), antibiotic and fungizone), similarly to ESC. In both cell cultures, the media were changed every 3 days. Upon reaching confluence, ESC and EEC were passaged by treatment with 0.5 g/L tripsin-0.2 g/L EDTA solution (GibcoBRL). The purity of cell cultures was greater than 90% for ESC and EEC, evaluated by immunocytochemistry of vimentin and cytokeratin, respectively. Epithelial and stromal cell culture Cells, EEC and ESC, were grown in appropriate medium in 6-well or 96-well plates at 2.5 × 10 5 cells/well or 0.1 × 10 5 cells/well, respectively. When cells were subconfluent (48 – 72 h of culture), the media were changed to Hank's media and incubated for 24 h. Then, the cells were incubated in fresh Hank's media at 37°C in 5% CO 2 /air in humidified atmosphere in the absence or presence of androstenedione 10 -6 M or androstenedione 10 -6 M plus hydroxyflutamide 10 -5 M. The latter was added 15 min before androstenedione. The culture was carried out for 24 h to evaluate the early effect of androstenedione on cell survival. The concentration of androstenedione used in the present investigation was established in dose-response experiments and are in agreement with those previously reported [ 14 ]. Immunohistochemical detection Sections (4 to 6 μm) of human endometrial tissue obtained from cultured explants were deparaffinized in xylene and hydrated gradually through graded alcohols. Endogenous peroxidase activity was prevented by incubating the samples in 3% hydrogen peroxide for 5 min. The sections were incubated in 10 mM sodium citrate buffer (pH 6.0) at 95°C for 20 min. Nonspecific antibody binding was prevented with 2% PBS-BSA for 1 h. Primary antibody of Ki-67 (1:200; Novocastra Laboratories, Newcastle, UK) was applied to the samples and incubated overnight at 4°C; the antibody for caspase-3 (1: 100; R&D System, Inc., Minneapolis, MN) was incubated for 1 h at 37°C and the antibody for androgen receptor (AR) was incubated overnight at 4°C (1: 75; Santa Cruz, CA, USA). The second antibody used in both cases was a biotinylated rabbit antimouse immunoglobulin. The reaction was developed by the streptavidin-peroxidase system and 3,3'diaminobenzidine was used as the chromogen; counterstaining was carried out with hematoxylin. The slides were evaluated in a Nikon optical microscope (Nikon Inc., Melville, NY, USA). The immunohistochemical evaluation was determined as the percentage of positive stained cells. In all cases, at least 500 cells were evaluated by three independent observers. In Situ 3'-End Labeling of DNA in Apoptotic cells (TUNEL) Programmed cell death was detected using TdT-mediated dUTP nick end labeling as previously described (Promega, Madison, WI, USA) [ 18 ]. Briefly, paraffin sections were dewaxed with xylene and rehydrated for 3'-end labeling. Tissue sections were incubated with proteinase K (20 μg/ml) at room temperature for 10 min and washed with PBS for 5 min. Then, incubated for 1 h at 37°C with the nucleotide mix labelled with fluorescein and terminal deoxynucleotidyl transferase enzyme and counterstained with propidium iodide. The number of apoptotic cells was quantified by at least counting 1000 cells in a fluorescence microscope by three independent observers. The number of apoptotic cells was expressed as the percentage of positive cells. Cell proliferation CellTiter 96 AQueous Cell Proliferation Assay (Promega) was used to perform cell proliferation, following manufacturer's instructions. Briefly, EEC or ESC were plated in 96-well until sub-confluence; then, the cells were cultured with androstendione 10 -6 M alone or androstenedione 10 -6 M with hydroxyflutamide 10 -5 M, as described above. Twenty μL of a mix of tetrazolium compound (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl) 2H-tetrazolium; MTS) and an electron coupling reagent (phenazine methosulfate; PMS) were added to each well and incubated for 2 h at 37°C. The quantity of the soluble formazan product was measured at 450-nm absorbance in an ELISA reader (Sigma) and expressed as optical density units (OD). Detection of apoptosis by Annexin V Apoptosis was performed using Annexin V-FITC Apoptosis Detection Kit (Oncogene Reasearch Products, Boston, MA, USA). Briefly, after EEC and ESC achieved sub-confluence in 6-well plates, the cells were incubated with androstenedione 10 -6 M alone or androstenedione 10 -6 M with hydroxyflutamide 10 -5 M, as described above. The cells were dissociated by 0.25 g/L trypsin-0.1 g/L EDTA treatment, gently re-suspended in cold binding buffer to approximately 1 × 10 6 cells, and incubated with Annexin V-FITC, as indicated by the manufacturer. The cells were counter-stained with propidium iodide, analyzed and counted by two independent observers in at least 1000 cells in each experimental condition using a fluorescent and optical microscope (Nikon, UFX-NX model. Nikon Inc., Melville, NY, USA). The results are expressed as percentage of apoptotic Annexin V-FITC positive cells respect to total cells counted. RNA isolation and semiquantitative Reverse Transcription-Polymerase Chain Reaction Total RNA was isolated using TRIzol Reagent (GibcoBRL) from endometrial tissue and ESC culture according to the manufacturer's instructions. Total RNA was then reverse transcribed, and cDNA was subjected to polymerase chain reaction (PCR) using specific primers for ERα and ERβ[ 24 ], bcl-2 and bax cDNA (NIDig 4558699) [ 7 , 18 ]. β-actin was used as an internal control. Semiquantitative RT-PCRs were achieved in the exponential linear zone amplification for each gene studied. Two μg of total RNA were used for reverse transcription in a total volume of 20 μl using the Revertaid H Minus M-Mulv Reverse (Fermentas Hannover, MD, USA). The PCR condition for ERα and ERβ was 2 mM MgCl 2 , 0.15 mM of dNTP, 0.63 U of Taq DNA polymerase (Fermentas Hannover, MD, USA), and 0.4 μM of each primer; for bcl-2 and bax was 3 mM MgCl 2 , 0.25 mM of dNTP, 0.63 U of Taq DNA polymerase, and 0.4 μM of each primer. The PCR amplification was carried out in the Thermocycler, model PTC-100 (MJ Research Inc, Watertown, MA), as previously reported [ 23 ]. The PCR products were electrophoretically resolved on 1% agarose gel and stained with ethidium bromide. The bands were evaluated using an image analyser (Kodak Electrophoresis Documentation and Analysis System [EDAS 290] and Kodak 1D Image Analysis software, Rochester, NY, USA), and normalized relative to β-actin PCR product and expressed as arbitrary units (AU). To confirm the specificity of RT-PCR products, the fragments were purified with CONCERT Rapid PCR Purification System (GibcoBRL) and sequenced using an ABI PRISM310 automated Sequencer (Perkin Elmer; Norwalk, CT, USA). Statistical evaluation The results are expressed as the percentage of changes obtained in the treated respect to the basal condition. Comparisons between groups were performed by ANOVA following Dunnett test. The significance level was set at 5%. Results are expressed as mean ± standard error of the mean (SEM). Results Explants culture In all samples, the expression of the AR was present in the nucleus of EEC and ESC cells (data not shown). Moreover, in EEC showed a positive staining in the cytoplasm. The nucleolar antigen Ki-67 was detected in the nucleus of EEC and ESC. After six hours of incubation with androstenedione at different concentrations (10 -7 M – 10 -5 M), the percentage of positive cells increased significantly only in the presence of androstenedione 10 -6 M and on the stromal compartment (basal: 9,9 ± 2.5%; treated: 21.2 ± 6.4%; p < 0.05) (Figures 1A and 1D ). Therefore, the following experiments were performed at a concentration of androstenedione 10 -6 M. Figure 1 Effect of androstendione on endometrial cell proliferation and apoptosis of human endometria. Basal condition (left panel) and androstenedione-treated explants (right panel). The nucleolar antigen Ki-67, evaluated by immunohistochemistry, was detected in the nucleus of both cell compartments (A, D), indicative of cell proliferation. The nucleus of positive cells for TUNEL, determined by TdT-mediated dUTP nick end labeling, were stained in yellow and counterstained with propidium iodide (B, E), showing DNA fragmentation. The positive staining for caspase-3, determined by immunohistochemistry in paraffin wax sections of endometria, was found in the cell cytoplasm of both compartments (C, G). Negative controls (inserts) for inmunohistochemistry was performed with non-immune rabbit serum in the place of the respective primary antibody and for TUNEL, by the replacement of TdT enzyme. Arrowheads indicate positive staining of the respective proteins. Magnification in all panels, ×400. On the other hand, TUNEL positive cells were lower in ESC after androstenedione treatment (basal: 21.2 ± 5.0%; treated:10.3 ± 3.7%; p < 0.05), (Figures 1B and 1E ). The expression of caspase-3 did not change with the androstenedione treatment in both compartments (Figures 1C and 1F ). The addition of hydroxyflutamide plus androstenedione did not modify the degree of proliferation or cell death in tissue explants. Isolated endometrial cells Basal values of cell proliferation were 0.39 ± 0.09 OD and 0.64 ± 0.16 OD for EEC and ESC, respectively, whereas, the percentage of annexin V positive cells was 16.3 ± 4.6% for EEC and 12.7 ± 4.3% for ESC. The addition of androstenedione significantly increased cell proliferation in ESC cultures (Table 1 ; p < 0.05), and this effect was reverted by the addition of hydroxyflutamide (0.63 ± 0.16 OD vs 0.70 ± 0.19 OD); no changes were observed in the EEC subpopulation (Table 1 ). Table 1 Effect of androstenedione in cell proliferation and apoptotic index in endometrial epithelial cells (EEC) and endometrial stromal cells (ESC) in vitro. Cell Proliferation (%) Apoptotic index (%) EEC ESC EEC ESC Basal 100 100 100 100 Androstenedione (10 -6 M) 91.6 ± 6.9 135.4 ± 1.2* 165.0 ± 59.0 120.0 ± 11.0 Androstenedione (10 -6 M) plus hydroxyflutamide (10 -5 M) 99.1 ± 10.0 109.2 ± 4.2 111.2 ± 6.3 106.7 ± 7.4 p < 0.05. Values are calculated as percentage of basal and are expressed as mean ± SEM. The percentage of cells with positive signs of apoptosis was similar between the basal and the treated conditions, independently of the cell type analyzed. The addition of hydroxyflutamide plus androstenedione to both cell cultures did not modify the degree of cell death (Table 1 ). Effect of androstenedione on the abundance of messenger RNA for bcl-2 and bax A similar mRNA abundance for bcl-2 and bax was obtained in tissue explants without treatment (bcl-2; 0.96 ± 0.12 AU; bax: 0.99 ± 0.17 AU). Table 2 shows the effect of androstenedione on the mRNA abundance of bcl-2 and bax in endometrial explants. For bcl-2, the level of its mRNA decreased with androstenedione treatment (p < 0.05) and hydroxyflutamide inhibited this effect, whereas, the mRNA abundance of bax did not change with the treatment. Despite these results, a similar bcl-2/bax ratio was obtained (basal: 1.06 ± 0.21; treated: 0.99 ± 0.12). Table 2 Effect of androstenedione in the mRNA abundance for bcl-2 and bax in endometrial tissue explants. bcl-2 (%) bax (%) Basal 100 100 Androstenedione (10 -6 M) 73.6 ± 3.9 90.0 ± 20.9 Androstenedione (10 -6 M) plus hydroxyflutamide (10 -5 M) 86.7 ± 14.5 113.3 ± 37.1 *p < 0.05. Values are calculated as percentage of basal and are expressed as mean ± SEM. In isolated cells, the basal mRNA abundance for bcl-2 in EEC was 0.47 ± 0.07 AU and for bax 0.39 ± 0.25 AU and in ESC, basal expression of bcl-2 mRNA was 0.87 ± 0.16 AU and 1.09 ± 0.10 AU for bax. Androstenedione induced a decrease of 58% of bcl-2 mRNA expression (p < 0.05) and a 30% increase of bax mRNA (p < 0.05) in EEC (Figure 2 ); therefore, the ratio bcl-2/bax was significantly lower compared to the basal condition (p < 0.05). In ESC, no significant differences on mRNA expression for bcl-2 and bax were found between the basal and androgen-treated conditions. No important changes in mRNAs expression were observed when hydroxyflutamide was added to both cell culture systems. Figure 2 Polymerase chain reaction (PCR) amplification from reverse-transcribed cDNA from endometrial epithelial cells (EEC) under the stimulation with androstenedione 10 -6 M, using primers for bcl-2, bax and β-actin. Results represent six experiments performed in duplicate. Normalized yield for bcl-2 and bax PCR fragments relative to β-actin. PCR products from different experiments are shown as percentage respect to basal. The values are expressed as mean ± SEM. a p < 0.05 between basal vs androstenedione. Effect of androstenedione on messenger RNA abundance of steroid receptors In the endometrial explant cultures, the abundance of mRNA for ERα was similar between the basal condition and the tissue treated with androstenedione (0.44 ± 0.12 AU; 0.41 ± 0.07 AU, respectively). In contrast, androstenedione treatment induced a decrease in ERβ mRNA abundance (basal: 0.91 ± 0.11 AU ; treated: 0.75 ± 0.08 AU; p < 0.05). However, the ratio between the level of ER isoforms did not change. No significant modifications were observed when hydroxyflutamide was added to the cultures. On the other hand, basal expression of ERα mRNA was 0.31 ± 0.05 AU and 0.39 ± 0.08 AU for ERβ in EEC; in ESC was 0.82 ± 0.23 AU for ERα and 0.86 ± 0.28 AU for ERβ. Androstenedione tended to modify mRNA abundance of ERα in EEC and ESC, a 30% and 25% diminution was obtained, respectively (p = 0.07). In EEC, gene expression of ERβ increased 48% with androstenedione (0.39 ± 0.08 AU vs 0.56 ± 0.18, p < 0.05) (Figure 3 ), with no modification in ESC. Therefore, the ratio between the expression of ER isoforms decreased 70% in EEC. Hydroxyflutamide reverted the effect of androstenedione on gene expression of ERβ and on the ratio ERα/ERβ. Figure 3 Polymerase chain reaction (PCR) amplification from reverse-transcribed cDNA from endometrial epithelial cells (EEC) under the stimulation with androstenedione 10 -6 M, using primers for ERα, ERβ and β-actin. Results represent six experiments performed in duplicate. Normalized yield for ERα and ERβ PCR fragments relative to β-actin. PCR products from different experiments are shown as percentage respect to basal. The values are expressed as mean ± SEM. a P < 0.05 between basal vs androstenedione. Discussion The present investigation represents an interesting approach that associates the potential effect of androgens on endometrial cell survival. By means of two in vitro models, we could observe that androstenedione can modulate the proliferation and apoptosis of the stromal compartment and modify the mRNA abundance of proteins related to apoptosis and β-isoform of ER in EEC. Thereby, in explant cultures, androstenedione stimulated cell proliferation in stroma. One possible explanation to this finding may be related to the fact that androgens can induce the expression of the receptor of epidermal growth factor in the stromal compartment, as reported previously [ 25 ]. In turn, the increase in the proliferation rate may occur through an indirect effect of the growth factor on stromal cells. Also, in the tissue explants model we observed that androstenedione induced a diminution in the apoptosis degree; although, the abundance of bcl-2 gene decreased. This observation may be in contrast to the mechanism that regulates apoptosis in cells; however, the ratio bcl-2/bax did not change and, concomitantly, we did not observe important differences in the expression of caspase-3. This is a relevant point because according to previous reports the control of cell death is principally associated with an unbalance in the expression of proteins related to the bcl-2 family, mainly bcl-2 and bax [ 17 , 18 ]. These findings may suggest that alternative apoptotic pathways can be also operating in the endometrial cells. On the other hand, it is well known that in endometrial tissue, estrogen actions mediated through their own receptors have been related to cell survival and progression of proliferation. Furthermore, during the menstrual cycle, it has been demonstrated the presence of the isoforms α and β of ER in endometrial tissue [ 26 , 27 ]. Moreover, it has been postulated that the expression of ERα may be associated to bcl-2 gene expression [ 28 ]. Studies in other reproductive tissues have also suggested that ERβ could be involved in the inhibition of cell proliferation [ 29 , 30 ]. Therefore, it acquires great relevance the relationship between α and β isoforms of ER, considering that both isoforms could have an antagonistic action ligand-dependent, in accordance to the relative expression of each isoform in different tissues [ 16 , 31 ]. In the tissue explant model used in the present study, in contrast to a previous report [ 8 ], we were unable to demonstrate an increase of ERα. Nevertheless, we observed a decrease of the β-isoform of ER under the effect of androstenedione, which could be associated to the increase of stromal cell proliferation detected in the explants. Numerous clinical and in vitro studies have suggested that the imbalanced of ERα/ERβ is a common feature and could be critical in the progression of estrogen-dependent tumors. It seems that ERβ is an important modulator of the mitogenic action estrogen and it confers protection against the ERα hyperproliferation [ 32 ]. Moreover, in prostate carcinoma cells the expression of ERβ has been associated with triggers of apoptotic pathway, similar to observed in models of ovarian cancer cells [ 33 , 34 ]. Therefore, the observations in our models open an important field in a possible relationship between androgens effects with control of ERβ-mediated cell proliferation When isolated cells were evaluated, some differences were obtained in the parameters studied compared to the explant cultures, which highlight the importance of the relationship between the different cell compartments in the regulation of the cell survival [ 35 ]. In ESC, androstenedione stimulated cell proliferation with no changes in the apoptosis degree nor on the expression of the genes bcl-2 and bax. This observation suggests that ESC exhibit an independent capacity to respond to androstenedione, whose action may be mediated by the androgen receptor. The latter is based mainly on the effect of the competitive inhibitor of the androgen receptor, hydroxyflutamide, which reverts the effect of androstenedione in the two models. In EEC, androstenedione induced a diminution in the ratio between the mRNA abundance of bcl-2 and bax without an evident increase in the apoptosis degree. Previous studies in other models, such as breast cancer cell lines, have demonstrated that androgens can induce a decrease on the expression of bcl-2 and also an atrophy of the mammary ephitelium [ 36 ]. The present observations suggest that the mechanisms of control of cell death in EEC are different from those of ESC, indicating that ESC may be responsible in providing molecular and physical interactions that can inhibit the early changes in the balance of apoptosis control genes in EEC. In contrast to our results, a previous study showed that in EEC cultures, androstenedione produces a fall in cell proliferation [ 14 ]. In fact, we did not observe this phenomenon, most likely due to technical differences in which the treatment was performed for 24 h in the present study with the aim to evaluate the early expression of genes related to apoptosis. Moreover, in EEC we observed that androstenedione up-regulated the ERβ expression and hence, a decrease in ER /ERβ ratio was obtained. The meaning of this effect on cell survival is difficult to evaluate in our experimental model, since no estradiol was added to the culture system. Furthermore, it is unlikely that androstenedione may produce estrogen by P 450 aromatase activity, because normal endometrial cells present a very low expression of this enzyme, as previously demonstrated [ 37 ]. However, in EEC the decrease in ERα/ERβ may be related to an unbalance of the bcl-2/bax ratio although this hypothesis needs further studies. Taking together our results, we can speculate that in the presence of androgens the regulatory mechanisms of cell survival in endometrial cells would be associated to the interaction capacity elicited by the different cellular components of the tissue. In fact, we postulate that androgens induce an increase of proliferation in stroma, probably related to growth factors and that these signals interact with epithelial cells, promoting an inhibition on the expression of genes related to cell death. In isolated cells, the mechanisms that allow these interactions between cell compartments are lost and the cells act according to their proper feature. Our findings on the effect of androstendione on cell proliferation and apoptosis of ESC allows us to suggest a potential regulatory action of androgen in the physiology of the endometrium and its implications in the genesis of endometrial pathologies frequently found in women with hyperandrogenism. Even more, we have observed that endometrial androstenedione concentration in women with PCOS are three times higher than in normal women during the proliferative phase [unpublished data). Therefore, these in vitro models are important approaches to understand the potential role of androstenedione in the PCOS endometria. It is well established that women with PCOS exhibit a high prevalence of hyperplasia and endometrial cancer, which is associated with disturbances in the regulation of cell survival [ 4 - 6 ]. According to our results, androgens may be involved in these endometrial alterations. However, we cannot ruled out the possible action of hyperinsulinaemia, common feature observed in PCOS women, which has been associated to increase the potential for neoplastic change through theirs effects on sex hormone binding-globulin (SHBG), insulin-like growth factor-1 and estrogen concentrations [ 6 , 38 ]. Conclusions In summary, our results indicate that androstenedione may modulate cell survival, the expression of β-ER isoforms and proteins related to apoptosis. These observations are closely related to the control of endometrial cell function and may help to understand the possible effect of the hyperandrogemia on endometrial tissue. Authors' contributions MM conceived and designed the study, carried out the experimental protocols, and drafted the manuscript. SQ, MC and KB performed the RT-PCR and scored of immunohistochemistry, MA carried out the score of immunohistochemistry studies. CJ reviewed and supported in the drafting the manuscript and MV conceived the study as principal investigator and participated in drafting the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544582.xml
532399	A Fly Enzyme for Motor Control	null	Long-range communication requires special technology. Although we usually think of nerve cells communicating over distances measured in millimeters, they must also stretch over centimeters and even meters to enable movement and sensation throughout our bodies. In a disorder known as spastic paraplegia, people experience stiffness and loss of function in their legs because the far ends of the longest nerves in the spinal cord degenerate. One form of this disease, autosomal dominant hereditary spastic paraplegia (AD-HSP), is most frequently caused by mutations in a gene that codes for the enzyme Spastin. Until recently, the function of Spastin has only been inferred by its similarity to another protein, Katanin, which chops up microtubules. The microtubules in neurons, as in all types of cells, provide a network to transport materials from one place in the cell to another. For neurons to communicate with each other and with muscles, neurotransmitters packaged at one end of the cell make the journey through the cell's long, narrow axon on the backs of microtubules. At the other end of the cell, the neurotransmitters reach small cellular projections (boutons) where they are released as a signal to the receiving cell. Since long spinal cord axons are most often affected by AD-HSP, researchers have suggested that transport through axons might be culpable. Is there a link between Spastin, microtubule severing, and AD-HSP? To address this question, Nina Tang Sherwood and colleagues studied the function of Spastin in the fruitfly Drosophila by identifying the spastin gene in this species and manipulating its expression. They found that indeed Drosophila Spastin in neurons regulates microtubule networks. Overexpressing spastin caused collapse of the embryonic central nervous system and also eliminated the microtubule network, as expected based on the related Katanin protein's microtubule-severing activity. But—surprisingly—knocking out spastin did not yield the opposite result. spastin -null flies had fewer microtubule bundles, particularly at the far ends of the neurons. They also exhibited smaller and more numerous boutons that were unusually clustered together. On the basis of their results, the authors speculate that Spastin cuts microtubules to a manageable size. Too much Spastin chops the microtubules into useless fragments, but too little Spastin may leave microtubule polymers too large to be efficiently moved into newly forming boutons. With an intermediate amount, microtubule pieces are the right size for transport throughout the neuron. Without Spastin, normal motor function ceases. In spastin -null flies, neurotransmitter release is impaired and flying is impossible. Flies even tend to drag their hind legs. This weakness in the legs is just one compelling parallel to human AD-HSP. The severity of symptoms in people is highly variable, similar to the variability in phenotypes exhibited by Drosophila with intermediate spastin gene mutations. The authors caution, however, that they do not prove that the fly phenotypes observed arise through the same mechanisms that cause human AD-HSP. The utility of Drosophila as a model for human AD-HSP has yet to be demonstrated, but the importance of Spastin in regulating neuronal microtubule networks in vivo is no longer in doubt.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532399.xml
554785	Permissive nicotine regulation as a complement to traditional tobacco control	Background Cigarette smoking takes a staggering toll on human health and attracts considerable public health attention, yet real solutions seem distant. The 2004 Family Smoking Prevention and Tobacco Control Act (US Senate bill S2461) would have given the US Food and Drug Administration limited authority to regulate cigarettes to "protect the public health." However, such legislation is unlikely to substantially reduce smoking or related deaths. Discussion The past 500 years of tobacco control efforts demonstrate that nicotine prohibition is a practical impossibility for numerous reasons, state revenue being one of the most ominous. The FDA already has regulatory authority over pharmaceutical grade nicotine products, and requires pharmacists to dispense the most addictive of these only with prescriptions. Meanwhile, every corner store can sell far more addictive and dangerous cigarettes to any adult. The FDA could immediately increase competition between cigarettes and clean nicotine products by approving available nicotine products for over-the-counter sales to adults. Similarly permissive regulation of cigarettes and addictive nicotine products will reduce tobacco use and improve smokers' health, but increase nicotine use in the population. Fortunately, restricted youth access and accurate labeling of nicotine's absolute risks will dissuade many non-smokers from experimenting with it, while accurate depiction of its risks relative to cigarette smoking will encourage many smokers to switch. The FDA could take a series of small steps that might ultimately replace a large proportion of cigarette smoking with equally addictive nicotine products, without risking serious public health setbacks. Vaccine, methadone, and injury prevention policies establish relevant public health precedents. Summary Cigarettes, or an equally addictive alternative, will be a permanent and common product in most societies. Regulations restricting only the safest addictive nicotine products are hard to justify. Addictive nicotine compliments other tobacco control strategies. Modern tobacco control policies are applicable to addictive nicotine. Controlled trials and test market studies are urgently needed to evaluate addictive nicotine as an alternative to smoking. Meanwhile, legislators should preserve the Food and Drug Administration's option to permit non-prescription sales of addictive nicotine.	Background Cigarette smoking is a source of worldwide misery [ 1 , 2 ] and revenue for corporations and governments. Landmark tobacco regulatory efforts in the United States include the 1992 Synar Amendment, requiring States to establish and enforce prohibitions on the sale and distribution of tobacco products to persons under 18 years of age [ 3 ]. The 1998 Master Settlement Agreement (MSA) prohibited numerous marketing practices, especially those likely to entice youth [ 4 ]. The six corporate signatories also promised an endless series of payments to the settling States, currently $8 billion annually. The United States has endorsed the World Health Organizations' 2003 Framework Convention on Tobacco Control (FCTC) [ 5 ]. The FCTC enumerates well-studied, politically safe measures such as taxation; limiting youth access; regulating the content, packaging, advertising, and sales of tobacco products; and educating the public about risks. Fire-safe cigarettes [ 6 ] are now required in New York. Several states and many municipalities now require smoke-free workplaces. The recently defeated [ 7 ] 2004 Family Smoking Prevention and Tobacco Control Act, S2461 [ 8 ], would have established Food and Drug Administration (FDA) regulation of cigarettes for the first time. Although average citizens could easily have taken "FDA regulation" to suggest the extensive power that the FDA holds over pharmaceutical products[ 9 ], tobacco regulations would have been weaker. The defeated Act sought publication of brand-specific ingredient information; forbade adding children's favorite flavorings to cigarettes, further restricted teen sales and advertising; and allowed the FDA to regulate cigarettes' nicotine delivery. The FDA would not have been able to ban or eliminate nicotine from cigarettes, nor could the FDA ban a class of tobacco products. Although modestly effective at preventing and ending tobacco use, such measures will not achieve the Healthy People 2010 (HP2010) goal of 12% adult smoking prevalence [ 10 ] on schedule [ 11 ], if ever. While cigarettes should never again afflict more than a quarter of the United States population, tobacco regulation will abate the current level of carnage only very slowly. A complimentary strategy called "harm reduction" proposes that smokers could improve their health by frequently substituting less hazardous tobacco products, such as 'smokeless cigarettes,' chewing tobacco, and newer smokeless tobacco products for cigarettes [ 12 - 14 ]. 'Smokeless cigarettes' are neither smokeless nor cigarettes, but complicated devices that release a mixture of chemicals from a heated tobacco substrate and can emit more carbon monoxide than a cigarette [ 15 - 18 ]. For individuals who are determined to inhale nicotine, these devices are probably no worse than cigarettes. However, for individuals who would have quit instead of switching to a smokeless cigarette, the health cost may be large. The Institute of Medicine has called for extensive research to quantify harms and patterns of use of these "potential reduced exposure products" (PREP) [ 19 ]. In contrast, chewing tobacco is clearly safer than smoking cigarettes [ 20 , 21 ], exposing users to as little as 2% of the risks of smoking [ 22 ]. The newest smokeless tobacco products are discrete enough to use in almost any social setting. Smokeless tobacco products are much less popular than smoking in the United States for a variety of reasons, including restrictions on advertising. However, the most important limitation of smokeless tobacco may be absorption through the nicotine through systemic veins rather than the pulmonary vasculature. All current pharmaceutical nicotine products share this limitation. Some tobacco control advocates suggest a third strategy: national policies encouraging competition between safer nicotine products and cigarettes [ 20 , 23 - 34 ]. Typically, this involves a "level playing field" with similar regulations for pharmaceutical grade nicotine delivery systems and cigarettes. Advocates cite evidence that health risk correlates with exposure to tobacco's combustion and curing by-products [ 35 - 38 ], with nicotine replacement products posing the least risk [ 39 , 40 ]. There are different ways to level this playing field. At one extreme, the FDA could increase regulation of cigarettes to match that of nicotine products. This is politically challenging, as the FDA discovered in the 1990's [ 41 ], The Family Smoking Prevention and Tobacco Control Act divided both the tobacco industry and the public health community. Altria, corporate parent of Philip Morris, was the only tobacco company to endorse this legislation. Calling a similar bill the "Marlboro monopoly act" [ 42 ], critics suspect that Altria hoped to discourage competition, for instance by encumbering smaller producers with regulations, by prohibiting relative safety claims sought by chewing tobacco manufacturers [ 43 ], or by discouraging new products with stringent new standards [ 44 , 45 ]. Other critics believe that the Act safeguarded certain cigarette marketing and legal defense strategies [ 9 ], or paved the way for the company's unproven 'smokeless cigarettes,' potentially ushering in another generation of suffering. At another extreme, the FDA could regulate other nicotine products more permissively, as cigarettes have been regulated. The FDA has that authority, but would face criticism for facilitating a common addiction. For instance, the FDA could approve non-prescription nicotine nasal sprays and slow acting inhalers for indefinite use by adults. These nicotine replacement therapies have excellent safety records and low potential for addicting non-smokers, although a few ex-smokers become addicted. The widespread misconception that nicotine prohibition is practical and desirable makes it politically difficult for the FDA to pursue this path. Discussion Problems with nicotine prohibition A recent Lancet editorial called for the criminalization of tobacco products [ 46 ]. However, a litany of centuries-old problems with prohibition provides a strong argument for permissive nicotine regulation as an alternative or essential prerequisite to prohibition. Many people will pay high prices to a coalition of suppliers and governments to obtain tobacco cigarettes that efficiently deliver nicotine, a drug perceived to provide some benefits, in spite of substantial immediate and long-term risks of smoking. While comprehensive public smoking restrictions are possible, and some companies may exit the tobacco trade, traditional tobacco control and litigation may not provide additional large public health gains in the United States, and prohibition is a practical impossibility. The first problem is that historically, smokers accept personal and public hazards that make the dangers we associate with tobacco look quaint. After Christopher Columbus failed to control his crew's tobacco use, monarchs from England to China tried to contain the weed by execution, disfigurement, exile, and onerous taxation [ 47 ]. Tobacco trade drained monarchs' wealth, compromising national security, and smoking accidents incinerated whole cities. Tobacco use spread anyway. The second problem is that risk-tolerant smokers are no small group. Smoking persists where Mormon and Islamic prohibitions discourage it [ 48 , 49 ]. If 70% of smokers in the USA want to quit [ 50 ]. then 30% of smokers, more than 6% of the adult population, do not want to quit. The third problem is a common genetic predisposition to nicotine addiction [ 51 ]. The modifiable risk factors we associate with smoking initiation and persistence – parent behavior, peer pressure, role models, advertising, accessibility, repeated exposure, and perceived norms – were irrelevant to Christopher Columbus' crew and the first tobacco users across Eurasia. Risk factor modification will not alter genetically predisposed users' fascination with their first tobacco products. The fourth problem is that so much money is involved. When state and local taxes raised New York City cigarette prices to US$7 per pack, tax revenues and black marketing both increased [ 52 ]. During the invasion of Iraq, stressed US soldiers reportedly paid up to US$50 per pack [ 53 ]. The black market for tobacco in Colorado prisons may achieve a 45,000% markup [ 54 ]. This industry makes large profits selling a simple product. If liability claims bankrupt corrupt companies, new suppliers will fill the void and try to avoid predecessor's mistakes. Black markets will undermine the benefits of high taxation or prohibition. The fifth problem is that government taxation compromises tobacco control efforts. Historically, States willingly trade citizens' health for wealth. Seventeenth century monarchs who opposed smoking relented as tobacco tax revenue accumulated. The States demonstrated the same perverse values with petitions to protect their MSA payments when an Illinois court threatened Altria with a US$12 billion bond [ 55 ]. The States will predictably protect the MSA corporate signatories from new competitors, for instance with taxes targeting generic brands. Tax revenues diminish prospects for even slowing the growth of tobacco sales in Africa, Eurasia, and South America, and prohibition in those areas is currently impossible. The sixth problem is the lack of a popular mandate for prohibition. Tobacco control advocates in the United States are rightfully pleased with smoking's declining prevalence, reduced teen smoking, rising cigarette prices, spreading restrictions on public smoking, and fire-safe cigarette initiatives. However, none of this demonstrates popular support for prohibition. Citizens will balk at limiting the supposedly personal choice to smoke in private. Midwesterners have already rejected modest cigarette tax hikes [ 56 ]. Furthermore, most taxes and MSA payments subsidize programs that benefit non-smokers [ 57 ]. Prohibition would eliminate that subsidy and require non-smokers to pay for enforcement, a very unlikely prospect. The seventh problem is that nicotine use could have a favorable risk-benefit profile for some informed users. Nicotine causes a mild euphoria without intoxication, in contrast to more tightly regulated drugs: it does not destroy relationships as intoxicating drugs routinely do. Many smokers may use nicotine to treat various problems [ 58 - 60 ] including depression [ 61 ], attention deficits [ 62 ], other mental illness [ 63 ], symptomatic systemic diseases [ 64 ], or to control weight [ 65 ]. Nicotine users may perform some tasks better, especially those involving vigilance and rapid visual cue processing [ 66 , 67 ]. There is uncertainty regarding many of these benefits [ 68 ]. Nevertheless, expected benefits are politically hard to withhold. If some of these benefits are real, nicotine prohibition may not even be desirable. Nicotine accounts for very few of the long-term hazards of smoking. A smokeless tobacco proponent has likened nicotine's risks to the risks of consuming caffeine [ 54 ]. Fetal exposure causes placental constriction and reduced birth size, alters brain development in disturbing ways, and may increase susceptibility to later nicotine addiction [ 69 - 76 ]. Nicotine might contribute to sudden infant death syndrome [ 69 , 77 - 79 ], destruction of connective tissue [ 80 ], modulation of immune function [ 81 ], prevention of apoptosis [ 82 , 83 ], and alcohol or other substance abuse [ 84 - 86 ]. Fatal nicotine poisoning is quite unlikely [ 87 - 89 ]. Given that smokeless tobacco users experience only about 2% of the risks of smoking [ 22 ], and that inhaled nicotine is similarly benign in animal models [ 90 - 92 ], it is very unlikely that inhaled nicotine could account for even one-tenth of the harms of smoking. In summary, however desirable tobacco prohibition may be, it is hopelessly impractical – unless smokers, governments, and producers have an equally satisfying alternative. Stifling innovation Legal and regulatory pressures have prevented the development of meaningful alternatives to cigarettes. Directed to approve drugs as "safe and effective" for specific indications, and to ignore tobacco, the FDA has had little reason to approve chronic, addictive nicotine. Although the FDA approves more dangerous drugs for specified indications, without an indication there is no benefit to weigh against any nicotine risk. Consequently the FDA requires prescriptions for slow acting inhalers and nasal spray because of a small risk of inconsequential addiction in ex-smokers. Unfortunately, seemingly risk averse regulation of nicotine forces public health policies to rely on difficult smoking prevention, cessation, use restrictions, and treatment. The net effect is a risky public health policy, and very slow development of new nicotine delivery systems [ 32 ]. Another unfortunate side effect has been that companies put tobacco into any device that they do not want regulated, particularly, 'smokeless cigarettes,' even if the device would be safer without tobacco. In permissive nicotine regulation, the FDA's missing indication is to improve nicotine addicts' safety. As prescription methadone substitutes for illegal, immediately incapacitating heroin, non-prescription nicotine could substitute for legal, slowly injurious cigarettes. Tort threats have also delayed product improvement efforts within the industry. In 1963 an industry lawyer, anticipating condemnation in the 1964 Surgeon General's report, suggested competing on safe nicotine delivery, but was overruled [ 93 ]. Product liability and regulatory issues pushed tobacco companies into pointless projects including filters, light cigarettes, "smokeless cigarettes", nicotine-free cigarettes, and fabricating a controversy over the health risks of smoking. Haddon matrix In the mid 1960's, motor vehicle accidents were also taking a terrific health toll in the United States. While alternatives to driving existed, no one expected private automobiles to disappear. Instead, systematic efforts transformed traffic safety. Haddon's matrix illustrated how complimentary strategies could work together to reduce the morbidity and mortality of driving [ 94 ]. Table 1 shows a matrix with two axes representing time and objects. Most injury control efforts fall into one of the nine cells, although some cells are empty. Before an accident, typical preventive efforts remove or modify unsafe drivers, vehicles, and road conditions. During an accident a vehicle's design and environmental safeguards may prevent or limit injuries. After an accident, prompt medical attention limits the morbidity and mortality of the injuries that still occur. Table 1 Haddon's Matrix for Reducing Traffic Accident Injuries Driver/Passenger Vehicle Environment Before Accident Licensing Stops, tickets, arrests Drug, alcohol screening Physician advice Running lights Antilock brakes High traction tires Vehicle inspections Impact avoidance Road design Road maintenance Road lighting, marking Traffic law Police patrols During Accident ---- Bumpers Seat belts Airbags Crumple zones Fuel containment Deformable barriers Fences After Accident ---- Fuel containment Telecommunications Emergency stabilization Emergency transport Trauma centers For tobacco control, smoking replaces accidents as the focal event (Table 2 – anticipated strategies are italicized ) [ 95 ]. As with accidents, cells contain only partially effective interventions. Many people experiment with tobacco in spite of efforts to prevent tobacco use. During smoking, smokers quit infrequently in spite of warnings and medications. More effective smoking cessation products may appear [ 96 , 97 ]. but some nicotine use will persist. In the center cell, hazardous cigarettes easily dominate all available nicotine replacement products [ 98 ]. Available and anticipated harm reduction products may be safer, but addictive nicotine would be safer still. After injury, detection of disease may be slow, and treatments for tobacco-related diseases are seldom curative. The matrix highlights the problems of incomplete prevention, unsafe nicotine sources, and poor treatment options, and demonstrates why permissive nicotine regulation could benefit public health: when prevention fails, results are bad. Making nicotine addiction as safe as possible would make prevention failures less disastrous. Table 2 Haddon's matrix adapted to tobacco control Potential user Nicotine delivery device Environment Before Addiction Addiction education Disease education Counter-advertising Warning labels Labeling regulations Youth sales restrictions Taxation Black market policing Prohibition During Use Disease education Cessation advertising Bupropion Nortiptylline Varenicline [105], Rimonabant [97] Filters Low tar cigarettes Smokeless tobacco Smokeless cigarettes Content regulation Nicotine replacement Addictive inhalers Taxation Prohibition After Injury Disease awareness Cessation advice Nicotine replacement Addictive inhalers Disease screening Disease treatment Cures Proposal For the reasons outlined above, the health benefits of addictive pharmaceutical grade nicotine products would likely outweigh the harms. If so, the FDA could improve public health by regulating nicotine much as the government has regulated cigarettes. A permissive nicotine regulatory policy would allow sales of pharmaceutical grade nicotine delivery systems to adults without a prescription. The FDA could suggest warnings appropriate for classes of products and delivery system constituents. Manufacturers would accurately label contents, and would be legally liable for undisclosed harms caused by the delivery system and constituents other than nicotine, as with any pharmaceutical product. The FDA could prohibit inherently risky delivery systems, and would undertake a full, traditional review of nicotine systems that give users a faster or higher peak arterial level of nicotine than cigarettes. The FDA would also review additives intended to provide antidepressant or other effects familiar to smokers [ 99 , 100 ]. Youth marketing and access would be illegal, as with cigarettes. The Drug Enforcement Agency would have no jurisdiction over non-intoxicating nicotine products, just as it has no interest in tobacco. Most public and workplace restrictions on cigarette smoking would be irrelevant to other nicotine delivery systems. States could tax cigarettes and nicotine at different rates, in theory recovering expenses related to each product. This tax policy would discourage cigarette use, encourage switching to nicotine, and maintain some State revenue as cigarette sales decline. A satisfying, safe, legal, and affordable alternative to cigarettes would discourage black markets. The FDA and Federal Trade Commission could permit advertising of nicotine as an alternative to smoking, and monitor relative harm claims. Nicotine manufacturers would pay for monitoring of adverse health effects from their products until remaining health questions are answered. The FDA could implement a permissive policy in a series of simple, informative steps. First, the FDA could immediately approve non-prescription sales of the existing nicotine spray and inhaler, and similar competitors, with restraints consistent with the Synar amendment, the MSA, and the FCTC. In particular, it should require disclosure of known nicotine risks, such as addiction, invite comparisons to smoking risks, and forbid marketing to minors. Data collected during this first step will inform subsequent steps. Economists have shown that the price of nicotine gum and patches affects demand for cigarettes [ 101 ]. This means that some smokers will substitute less expensive but very slow acting nicotine replacement products for cigarettes. Non-prescription access to more addictive and competitively priced nicotine should increase this substitution, especially if addicts can purchase a one day supply of about 20 mg. Clinical trials and post-marketing studies of non-prescription use of the spray and inhaler would begin to answer questions about nicotine and pregnancy, substance abuse, and heart disease. Surveillance studies would determine whether clean nicotine options lead more people to smoke, to smoke longer, or to use nicotine during pregnancy. While the Institute of Medicine recommends years or decades of study to quantify the risks of "smokeless cigarettes", the most important remaining uncertainties about nicotine might be answered in a few years. Data collected during this first step should document the relative safety of addiction to pharmaceutical nicotine versus cigarettes, the market share of each product, and the increase in total nicotine use. Analyses using these data will predict the consequences of more permissive regulation [ 102 ]. If experience with the current nicotine spray and inhaler is reassuring, the FDA can take a second step, to approve clearly addictive nicotine inhalers for non-prescription use. These fast acting inhalers will deliver nicotine in a powder [ 103 ] or aerosol to the alveoli, as pulmonary inhalers deliver steroids and beta agonists. Fast acting nicotine inhalers would be subject to the same marketing requirements and surveillance described above. Several trends could develop that encourage smokers to switch to addictive nicotine inhalers, while limiting recruitment of non-smokers. Smokers will appreciate the healthier alternative to cigarettes, especially as they develop smoking related illnesses, while warning labels will discourage casual experimentation by non-smokers. Smoke-free workplace regulations would spread with less controversy, as fast acting inhalers give smokers a reasonable alternative. Employers might voluntarily forbid smoking to reduce health care and workman's compensation costs; reduce workplace fires; and increase productivity by eliminating both smoking breaks and withdrawal symptoms. Physicians might strongly encourage smokers to switch for their own health, primarily, but also for the health of family members. The government might permit more aggressive advertising of nicotine inhalers to further undermine cigarette smoking. Governments will probably find nicotine taxation irresistible, especially if their tobacco revenue declines. Nicotine taxation will raise the cost of experimenting with inhalers beyond some teenager's means. Communities might press smokers to switch so that fire departments, health care providers, and research funding agencies could shift resources to many other pressing problems. Ultimately, palatable inhaled nicotine products could finally allow governments to ban tobacco cigarettes. The public health benefits of these policy shifts could be substantial. We can estimate the long term relative public health burdens of different policies as the product of the risks posed by a delivery system, relative to cigarettes, and the fraction of the adult population using it. Using this formula, the current burden is more than 20% use × 100% risk = 0.2. Healthy People 2010 calls for 12% use × 100% risk = 0.12, about half of the current burden. If nicotine accounted for a surprising 10% of smoking risks, eliminating smoking by quintupling the prevalence of inhaled nicotine use – universal addiction – would cause 100% use × 10% risk = 0.1, a slight improvement over HP2010. A slightly less pessimistic scenario would be that all past and present smokers become inhaled nicotine addicts, but their risk is only twice that of smokeless tobacco users. This scenario is no more than 70% use × 4% risk = 0.028, about a quarter of the HP2010 burden. The most likely scenario is that nicotine risks are about 2% of smoking risks, and that policy changes will slowly shift the probabilities that smokers will attempt and succeed in quitting, that smokers will switch to nicotine indefinitely or use both nicotine and cigarettes, that ex-smokers relapse to nicotine use, and that non-smokers begin using nicotine. If the risks of inhaled nicotine are this low and if at least one smoker can switch to addictive inhaled nicotine for every 50 ex-smokers and non-smokers who start using nicotine, then public health will benefit from permissive nicotine regulation. One public health risk is that more people might begin or continue smoking cigarettes in the belief that inhaled nicotine will provide an easy escape path, but suffer irreparable harm before switching. We should quantify these probabilities through surveillance of the public's behavior and health consequences following each step toward more permissive nicotine regulation. Although S2461 was defeated, the continuing carnage will motivate similar legislative proposals. Tobacco policies must preserve the possibility of permissive nicotine regulation. Bad legislation could solidify the position of cigarettes, delay safer products, prohibit disclosure of relative harms, or otherwise interfere with market forces that ought to benefit nicotine addicts [ 104 ]. The FDA must retain the option to unleash real competition against a disastrous status quo in the tobacco industry. Summary Nicotine use will remain common indefinitely. Pharmaceutical grade nicotine is the safest known substance that could replace cigarettes, but inherently addictive products are required to compliment traditional tobacco control policies effectively. The FDA could apply restrictions, similar to those on cigarettes, to an increasingly addictive portfolio of nicotine products, requiring honest portrayals of absolute and relative risks, and expect a significant reduction in cigarette smoking and related illness with modest recruitment of new nicotine addicts at each step. A brief and focused research effort could solidify the already substantial justification for such permissive nicotine regulation. Meanwhile, federal legislation must not obstruct free market ideals of competitive innovation and informed consumption. Competing interests The author(s) declares that he has no competing interests. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554785.xml
524510	Production of hemo- and immunoregulatory cytokines by erythroblast antigen+ and glycophorin A+ cells from human bone marrow	Background Erythroid nuclear cells (ENC) of the bone marrow (BM) have not previously been considered as important producers of wide spectrum of haemo- and immunoregulatory cytokines. The aim of the current work was to confirm the production of the main hemo- and immunoregulatory cytokines in human ENC from BM. Results We used native human BM ENC in our experiments. We for the first time have shown, that the unstimulated erythroblasts (Gl A + or AG-EB + ) produced a wide spectrum of immunoregulatory cytokines. Human BM ENC produce cytokines such as interleukn (IL)-1β, IL-2, IL-4, IL-6, interferon (IFN)-γ, transforming growth factor (TGF)-β1, tumor necrosis factor (TNF)-α and IL-10. They can be sub-divided into glycophorin A positive (Gl A + ) and erythroblast antigen positive (AG-EB + ) cells. To study potential differences in cytokine expression between these subsets, ENC were isolated and purified using specific antibodies to Gl A and AG-EB and the separated cells were cultivated for 24 hours. The cytokine contents of the supernatant were measured by electrochemiluminescence immunoassay. Quantitative differences in TGF-β1 and TNF-α production were found between Gl A + and AG-EB + BM ENC. Furthermore, in vitro addition of erythropoietin (EPO) reduced IFN-γ and IL-2 production specifically by the AG-EB + ENC. Thus, Gl A + and AG-EB + ENC produce IL-1β, IL-2, IL-4, IL-6, IFN-γ, TGF-β1 and TNF-α. Gl A + ENC also produce IL-10. Conclusion Cytokine production by erythroid nuclear cells suggests that these cells might be involved in regulating the proliferation and differentiation of hematopoietic and immunocompetent cells in human BM.	Background Haematopoesis is regulated by lymphoid and non-lymphoid cells through a complex network of paracrine and autocrine mechanisms involving cytokines, growth factors and their receptors. However, although stromal [ 1 - 6 ], endothelial [ 7 - 10 ], megakaryocytic [ 11 , 12 ] and osteogenic cells [ 13 - 16 ] and lymphocytes are known to express cytokines, erythroid nuclear cells (ENC), the major cell population of the BM, are not considered as important producers of hemo- and immunoregulatory cytokines. Experiments on ENC isolated from mouse spleen undergoing erythroid hyperplasia, and on cells separated from single erythroid colonies, have revealed that mRNAs for cytokines such as IL-1α, IL-1β, IL-4, IL-6, granulocyte-macrophage colony stimulating factor (GM-CSF), TGF-β1 and IFN-γ are present [ 17 , 18 ]. Production of GM-CSF and IFN-γ has also been reported [ 17 ]. Similarly, IL-2 and IL-3 mRNAs were found after erythroid cells were treated with erythropoietin (EPO) [ 17 - 20 ]. Human fetal liver erythroid cells have also been shown to produce such cytokines as IL-1, IL-2, IL-4, IL-6, IL-10, TNF-α, IFN-γ and TGF-β1 [ 20 ]. Stopka and co-authors have demonstrated the ability of burst-forming unit erythroid cells (BFU-E) to express and produce EPO [ 21 ]. Also it has been shown that erythroid cells express and produce vascular-endothelial growth factor (VEGF)-A and placental growth factor (PlGF) both in vitro and in vivo. Production of these proteins varies during differentiation and is increased 100-fold by PlGF and 3-fold by VEGF-A [ 22 ]. Macrophage colony stimulating factor (M-CSF), fibroblast growth factor (FGF)-2, VEGF-A, hepatocyte (H)GF, insulin-like (I)GF-1, IL-1β, trombopoietin (TPO), TNF-α, IFN-γ, FAS-L, and macrophage inflammatory protein (MIP)-1α mRNAs are also expressed in BFU-E isolated from BM, and VEGF-A and TGF-β1 are produced [ 23 , 24 ]. All these data allow us to consider erythroid cells as cytokine producers involved in regulating hemo- and immunopoiesis. The aim of the current work was to confirm previous observations by studying the production of the main hemo- and immunoregulatory cytokines, IL-1β, IL-2, IL-4, IL-6, IL-10, TNF-α, IFN-γ and TGF-β1, in human ENC from BM. Furthermore, we wished to determine whether different subsets of ENC (AG-EB+ and Gl A+ erythroid cells from human BM) might account for altered expression levels. Expression of both these markers is specific to erythroid cells; they are absent from the membranes of other blood-forming cells. AG-EB is expressed in a proportion of BFU-E cells (up to 4% of AG-EB + BFU-E cells) and on colony-forming units – erythroid (CFU-E) (up to 36% of AG-EB + CFU-E cells) [ 25 ]. Its expression on erythroblasts is further increased (up to 90% in AG-EB + erythroblasts). Later in development, AG-EB expression decreases, and only 4% of reticulocytes carry AG-EB [ 25 ]. In contrast, Gl A is expressed in a minor subpopulation of the CFU-E cells (less than 4%). On the erythroblasts this expression is increased and reach up to 100% in reticulocytes [ 26 , 27 ]. Sheme of AG-EB and Gl A expression during erythroid maturation present in Table 1 . Table 1 Expression of AG-EB and Gl A during erythroid maturation. The percentage of positive cells detected by flow cytometric analysis [25, 26, 27]. % of positive cells BFU-E CFU-E Proerythroblasts → erythroblasts reticulocytes AG-EB 0 – 4 36 93–95 4 Gl A 0 0 – 4 Increase with maturation 100 In this article we show which human BM ENC produce the cytokines IL-1β, IL-2, IL-4, IL-6, IFN-γ, TGF-β1, TNF-α and IL-10. Results Erythroid cells were separated according to the presence of the surface markers AG-EB and Gl A. BM erythroid cells carrying AG-EB produce the cytokines IL-1β, IL-2, IL-4, TNF-α, IFN-γ and TGF-β1 (figure 1 ). Those carrying Gl A also produce these cytokines and in addition might be secrete IL-10 (figure 1 ). Interestingly, erythroid cells carrying AG-EB release more TGF-β1 (p < 0.05) than Gl A + cells, whereas Gl A + secrete more TNF-α (p < 0.05). Figure 1 Cytokine production by BM erythroid nuclear cells carrying Erythroblast Antigen and Glycophorin A. □ – BM AG-EB + erythroid nuclear cells; ■ – BM Gl A + erythroid nuclear cells. All populations were cultivated for 24 h at concentrations of 10 6 cells per ml. Results present as (Mean ± SEM). n/d – not detected. * – P < 0.05. To the best of our knowledge, this difference between BM AG-EB + and Gl A + ENC has not previously been reported. To verify that the cytokines were truly expressed by these erythroid cells, we isolated erythoid colonies by cultivating single BM cells in MethoCult H4433 medium. As shown in Table 2 , the cytokines produced were comparable between ENC derived from erythroid colonies and ENC enriched from the BM. Table 2 Concentration of cytokines in conditioned medium derived from erythroid cells. Concentration of cytokines (pg/ml) in conditioned medium derived from enriched population of BM erythroid cells (1 × 10 6 /ml) and erythroid cells isolated from erythroid colonies (1 × 10 6 /ml). Results were obtained from at least three independent experiments (Mean ± SEM). Enriched erythroid nuclear cell population Erythroid nuclear cells derived from erythroid colonies IL-1β 78.64 ± 44.84 10.8 ± 0.39 IL-2 255.36 ± 88 237.73 ± 237.73 IL-4 62.66 ± 45.28 738.26 ± 640.15 IL-6 741.08 ± 395.95 324 ± 56.14 IL-10 121.26 ± 53.22 73.4 ± 73.4 TNF-α 2.59 ± 1.41 0.0 IFN-γ 766.35 ± 319.45 752.88 ± 492.65 TGF-β1 524.88 ± 372.59 479.25 ± 267.75 EPO regulates the proliferation and differentiation of erythroid cells [ 28 - 30 ]. Therefore, we assessed the effect of EPO on our different ENC cell populations and found that the production of some hemo- and immunoregulatory molecules was altered (figure 2 ). For instance, EPO treatment of AG-EB + cells led to a 99% decrease in IFN-γ production and a 50% decrease in IL-2 production, and the production of IL-1β and TNF-α was enhanced. However, EPO treatment had no discernible effect on the cytokine profile of Gl A + cells. Figure 2 Influence of EPO on cytokine production by BM AG-EB + erythroid nuclear cells. □ – BM AG-EB + erythroid nuclear cells; ■ – BM AG-EB + erythroid nuclear cells after treatment with EPO. BM AG-EB + erythroid nuclear cells (10 6 cells/ml) were treated with EPO (2 U/ml) for 24 h. Results present as (Mean ± SEM). n/d – not detected. * – P < 0.05. Discussion We have shown that the native BM ENC can produce a wide spectrum of cytokines, which are capable of both stimulating and inhibiting erythropoiesis. Cytokine production has been confirmed using ENC isolated from erythroid colonies cloned from a single BM cell. Earlier we demonstrated cytokine production by ENC derived from fetal tissues [ 19 , 20 ]. However, there was a great difference between the profiles of cytokines released by BM and fetal ENC. TGF-β1 is known to inhibit the proliferation of erythroid precursors and to provoke or accelerate the terminal differentiation of mature erythroid cells [ 31 - 33 ]. The fetal AG-EB + and Gl A + erythroid cells produced TGF-β1 in concentrations less than 5 pg/ml and we suggested that this could be related to the increased proliferative activity of erythroid cells in the fetal liver [ 20 ]. We propose that the high level of TGF-β1 production by BM erythroid cells is probably related to the need to restrain erythroid cell proliferation in healthy adult individuals. This hypothesis were declared also in review [ 24 ]. It is notable that BM erythroid cells produce not only inhibitors but also high levels of stimulators of erythropoiesis such as IL-4 and IL-6, especially in the presence of EPO [ 34 , 35 ]. The production of these hemoregulatory molecules suggests that erythroid cells have a capacity for self-regulation. Interestingly, IL-2, which inhibits erythropoiesis [ 36 ], is produced by these cells in smaller quantities or, like IL-10, is not produced at all. Besides the differences in production of some cytokines by BM erythroid cells, there are parallels with cytokine production from fetal erythroid cells. A similar situation can be observed when TNF-α production by BM erythroid cells is compared with those from fetal liver [ 20 ]. In both cases AG-EB + cells produce significantly less TNF-α than Gl A + (figure 1 ). TNF-α is a known inhibitor of precursor cell proliferation and an inducer of cell death in a proportion of these cells [ 37 , 38 ]. Therefore it is plausible that erythroid cells use this cytokine for autocrine self-regulation, by maintaining a negative feedback loop. If this is so, then relatively small increases in the number of more differentiated Gl A + cells would lead to a decreased precursor proliferation rate, due to the elevated level of TNF-α. Negative feedback can also be maintained through IFN-γ, which is produced at the same level by both Gl A + and AG-EB + cells. However, IFN-γ initiates apoptosis only in BFU-E [ 39 , 40 ], while in CFU-E cells it increases the expression of Bcl-x and protects these cells from elimination. Thus, IFN-γ enhances the differentiation of erythroid precursors [ 41 , 42 ]. We found no statistical significant differences between AG-EB + and Gl A + cells in the production of IL-1β, IL-2, IL-4, IL-6 and IL-10. The capacity of erythroid cells to change their cytokine expression profiles quantitatively and qualitatively under different conditions is an important indicator of their active role in the regulation of hemo- and immunopoiesis. Similar data were previously obtained in mouse cells and are presented here for human BM erythroid cells. Treatment with EPO leads to significant changes in IL-2 and IFN-γ production. Notably, IFN-γ and IL-2 secretion by AG-EB + cells is lowered in the presence of EPO, while Gl A + cells show no such change. This difference could be related to EPO-R expression in AG-EB + cells, which are mainly CFU-E and erythroblasts [ 43 - 45 ]. Expression of this receptor ceases in more differentiated early erythroblasts. Indeed, Gl A + cells do not express EPO-R, since late erythroblasts and more differentiated forms of erythroid cells are prevalent in this population [ 28 , 43 , 46 ]. The EPO-induced changes in cytokine production mainly concern inhibitors of erythroid precursor proliferation, such as IL-2 and IFN-γ. Thus, one mechanism by which erythroid cell proliferation is stimulated could be a decrease in proliferation inhibitor production by these cells as a result of EPO treatment. Conclusion Erythroid nuclear cells appear to be active producers of hemo- and immunoregulatory cytokines, involved in regulating the proliferation and differentiation of hematopoietic and immunocompetent cells in human BM. Changes in the cytokines produced by erythroid cells in response to EPO suggests the ability of these cells to respond to microenvironmental changes by altering the cytokine production profile. Methods BM cells This work was approved by the local ethics committee. After obtaining informed consent, sternum BM samples were collected from 8 normal healthy volunteers. All samples used in this study exhibited normal myelograms. The enriched population of erythroid cells was isolated from BM mononuclear cells by depleting of adherent cells and granulocytes as described previously [ 19 ]. Isolation of erythroid cell population carrying a surface erythroid antigen and Glycophorin A We used indirect panning to obtain cells expressing either the erythroid antigen (AG-EB) or Glycophorin A (Gl A) [ 19 ]. Monoclonal mouse anti-red blood cell Glycophorin A (MAS518, Harlan-Sera Lab, England), and monoclonal HAE-9 antibodies against AG-EB (kindly provided by Prof. Mechetner, Russia) were used [ 25 ]. Affinity isolated polyclonal rabbit anti-mouse immunoglobulin (Biosan, Russia) was used as a secondary reagent to coat Petri dishes during the panning procedure. To avoid non-specific binding to FcRs, cells were blocked with aggregated normal human IgG and then incubated with anti-Gl A or HAE-9 antibodies for 50 min at 4°C. Cells were transferred to Petri dishes covered with rabbit anti-mouse antibodies for panning for 50 min at 4°C. Subsequently, immobilised cells were collected, cultivated for 24 h in RPMI-1640 with 10% horse serum (Sigma, USA) with and without EPO [ 19 , 47 ]. Samples were collected after cultivation. Cultivation of erythroid cells with EPO 1 × 10 6 cells were placed in culture medium containing 2 U/ml Epo (Boehringer Mannheim, Germany) for a 24-h incubation. CFU-E Cells from the enriched erythroid population were maintained as described above [ 19 , 20 ] at 2·10 5 /ml and then cultivated in MethoCult H4433 medium (Stemcell Tech. Inc. Vancouver, B.C.) for 14 days. Colonies were visually characterised as CFU-E using an inverted microscope. After cultivation, the colonies were harvested and washed twice to remove methyl cellulose. The resulting cellular suspension (1 × 10 6 /ml) was cultivated for 24 h in RPMI-1640 with 10% horse serum, then supernatants were collected. Samples were stored at -20°C. Cells purity Cell purity was assessed using Nocht-Maksimov staining of smears [ 48 ] and staining of haemoglobin with benzidin [ 49 ]. All cells (Gl A + , AG-EB + and CFU-E cells) were characterised as erythroblasts and were haemoglobin positive. Electrochemiluminescence (ECL) method for quantitative determination of cytokines Quantitative determination of cytokines was performed by the electrochemiluminescent immunoassay [ 50 - 52 ] using an ORIGEN Analyzer (IGEN Inc., USA) according to the manufacturer's protocol. Calibration curves ranged from 10 to 10,000 pg/ml. Assay sensitivity was 2.8 pg/ml, 2 pg/ml, 2 pg/ml and 6 pg/ml for IL-1β, IL-2, IL-4 and IL-6, respectively. Assay sensitivities were 2 pg/ml, 1 pg/ml, 1 pg/ml and 2 pg/ml for IL-10, TNF-α, IFN-γ, and TGF-β1, respectively. The ruthenylated and biotinylated antibodies were diluted to working concentrations (μg/ml) 1:1, 1:1, 1:2 and 2:1 for IL-1β, IL-2, IL-4 and IL-6, respectively. The biotinylated and ruthenylated antibodies were diluted to working concentrations (μg/ml) 2:2, 1:1, 2:1 and 4:2 for IL-10, TNF-α, IFN-γ and TGF-β1, respectively. For TGF-β1 detection, samples were treated with HCl to obtain intact and easily detectable proteins as described [ 53 ]. Antibodies All polyclonal and monoclonal antibodies were purchased from R&D Systems (Abington, UK). The following antibodies were used: against recombinant IL-1β, IL-2, IL-4, IL-6, IL-10, TNF-α, IFN-γ, TGF-β1 polyclonal and monoclonal: anti-hIL-1β #AB-201-NA, anti-hIL-1β #MAB201, anti-hIL-2 #AB-202-NA, anti-hIL-2 #MAB202, anti-hIL-4 #AB-204-NA, anti-hIL-4 #MAB204, anti-hIL-6 #AB-206-NA, anti-hIL-6 #MAB206, anti-hIL-10 #AB-217-NA, anti-hIL-10 #MAB217, anti-hTNF-α #AB-210-NA, anti-hTNF-α #MAB210, anti-hIFN-γ #AB-285-NA, anti-hIFN-γ #MAB285, anti-hTGF-β1 #AF-101-NA, anti-hTGF-β1 #MAB240. No antibody used displayed cross-reactivity with other cytokines. Cytokines Recombinant cytokines were purchased from R&D Systems (Abington, UK): rhIL-1β #201-IL, rhIL-2 #202-IL, rhIL-4 #204-IL, rhIL-6 #206-IL, rhTNF-α #210-TA, rhIFN-γ #285-IF, rhTGF-β1 #240-B-002; and PeproTech, Inc. (Rocky Hill, NJ): rhIL-10 #200-10 was used for calibration curves. All the recombinant cytokines were diluted to 2 μg/ml in PBS, pH 7.4, supplemented with 0.05% NaN 3 and 0.1% BSA (aliquots were kept at -70°C before use). Statistical analysis Results are presented as means ± SEM. Statistical significance was determined using the nonparametric Mann-Whitney U-test. P < 0.05 was considered to be significant. Authors' contributions IBK, NJD, ANZ and MIL carried out the selection of donors, sternum BM sampling and analysis of myelograms. TVI, ANS and SVK carried out all of the experiments and drafted the manuscript. SVS participated in design and coordination of the research and revision of the manuscript. ANS participated in the preparation and revision of the manuscript. VAK conceived of the study. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524510.xml
521497	Skipping the co-expression problem: the new 2A "CHYSEL" technology	The rapid progress in the field of genomics is increasing our knowledge of multi-gene diseases. However, any realistic hope of gene therapy treatment for those diseases needs first to address the problem of co-ordinately co-expressing several transgenes. Currently, the use of internal ribosomal entry sites (IRESs) is the strategy chosen by many researchers to ensure co-expression. The large sizes of the IRESs (~0.5 kb), and the difficulties of ensuring a well-balanced co-expression, have prompted several researchers to imitate a co-expression strategy used by many viruses: to express several proteins as a polyprotein. A small peptide of 18 amino acids (2A) from the foot-and-mouth disease virus (FMDV) is being used to avoid the need of proteinases to process the polyprotein. FMDV 2A is introduced as a linker between two proteins to allow autonomous intra-ribosomal self-processing of polyproteins. Recent reports have shown that this sequence is compatible with different sub-cellular targeting signals and can be used to co-express up to four proteins from a single retroviral vector. This short peptide provides a tool to allow the co-expression of multiple proteins from a single vector, a useful technology for those working with heteromultimeric proteins, biochemical pathways or combined/synergistic phenomena.	Introduction For the last 20 years, the gene therapy field has centred many of its efforts on finding ways to deliver a therapeutic gene to certain target cells in order to produce a therapeutic result. It was soon clear that it was necessary to deliver at least two genes, because a reporter/marker gene was needed in order to track the expression of the therapeutic gene (normally not easy to detect). There has been a large increase in vector development during these years, with the appearance of many new viral and non-viral vectors. However, since the late 1980s, few improvements have been made 'inside' those vectors. The linkage of the two genes of interest (therapeutic and reporter) has remained the same. The different strategies known for co-expression were reported during the 1980s -splicing, multiple promoters, fusions, reinitiation and IRESs-, at the same time that the first gene therapy experiments were being performed (for a review [ 1 ]). During the 1990s, nearly all those strategies were abandoned in favour of the IRESs. In bicistronic mRNAs bearing an IRES sequence, the first cistron is translated by scanning ribosomes that enter via the 5' end. The cloning of an IRES sequence downstream of the first cistron, allows the internal entry of ribosomes that translate the second cistron. As each cistron is translated from a different translational initiation event, both translations are uncoupled, and the proteins are not obtained in an equimolecular proportion ("imbalance") leading to a large excess of the first protein. The drive to co-express more than two genes, opening the door to therapies for muti-gene deficiencies, was halted by the inability of vector technology to guarantee a reliable co-expression. Nevertheless, IRESs were the first strategy that met with some success, and several polycistronic vectors able to co-express up to 4 genes were developed during the 1990's [ 2 ]. However, two main problems blocked the successful use of large and complex polycistronic vectors: the large size and imbalance of most IRESs which makes it very difficult to predict the level of expression of the downstream cistron [ 3 ]. This commentary discusses several recent publications that use self-processing polyproteins as a novel strategy for co-ordinated co-expression of several genes. Discussion Although gene therapy has employed the viruses as vectors, the co-expression strategies previously described have not taken advantage of the dominant ways in which viruses achieve co-expression in cells. It is the polyprotein strategy that many viruses use to co-express most of their proteins, or even all of them (as in picornaviruses). Not surprisingly, this strategy is indeed used by cells, although not very often, in particular for the co-ordinated secretion of different proteins and peptides. Recently, several groups have been trying to utilize this co-expression strategy. One of the possibilities is to introduce the target site for a cellular proteinase between two cistrons cloned in frame forming a single open reading frame (ORF; [ 4 ]). In this way the polyprotein is synthesized as a fusion protein that post-translationally is proteolytically cleaved to yield the discrete proteins of interest. Unfortunately, this strategy has several practical difficulties: (1) the polyprotein to be cleaved must reside, or at least pass through, the same compartment as the proteinase, (2) the cleavage is not always independent of the context, (3) the cleavage may be incomplete and unpredictable, (4) efficient cleavage will only be produced in cells actively expressing the proteinase, and (5) the post-translational cleavage is not compatible with all possible sub-cellular targetings. In many ways, a co-translational strategy such as reinitiation, which would be independent of cellular or viral factors, would be desirable. In reinitiation, ribosomes first translate an upstream cistron, although highly inefficiently, then resume translation of the second, downstream, cistron. Reinitiation was indeed tried in the 1980s, but the large imbalance makes it unsuitable for co-expression of even two genes (reviewed in [ 1 ]). The foot-and-mouth disease virus (FMDV) 2A sequence as a co-expression tool Picornaviruses, the same family of viruses to first provide the IRESs, encode all their proteins in a long single ORF that is cleaved post-translationally by viral proteinases. However, it was described in the 1980's that at one position, the polyprotein of some picornaviruses (such as FMDV) underwent a rapid co-translational self-processing. It was soon realised that the key was a small 18aa peptide (2A) that directed its own separation from the growing polyprotein. During the last decade, this mechanism has been studied in detail, resulting in a simple model: the small 2A peptide, during its translation, interacts with the exit tunnel of the ribosome to induce the "skipping" of the last peptide bond at the C-terminus of 2A. The crucial point is that the ribosome is able to continue translating the downstream gene, after releasing the first protein fused in its C-terminus to 2A (reviewed in [ 5 ]). This type of sequence has been termed CHYSEL ( c is-acting h ydrolase e lement). From a biotechnological standpoint, all that is needed is to clone the coding sequence of 2A, followed by the codon for the first amino acid of the next FMDV protein (2B), in frame between the two genes one wishes to co-express. The synthesis of the peptide bond between the last amino acid (Gly) of 2A and the first (Pro) of 2B is skipped, producing an upstream protein with a C-terminal tail of 18aa (2A) and a downstream protein with a Pro at the N-terminus. The extra sequences have minimal effect on the activity of most proteins and none on their stability. In fact, the 2A peptide has been used as an efficient tag for immunoprecipitation and Western blotting, although commercial antibodies are not yet available. Interestingly, additional CHYSEL sequences have been found in viruses other than FMDV (for a review of these "2A-like" sequences, see [ 5 ]). Broad applicability of 2A The initial publications using this strategy have shown that 2A skipping can be used in the typical viral vectors used for gene therapy (retrovirus and adeno-associated virus) to reliably co-express many reporter proteins (neomycin phosphotransferase, NEO; puromycin N-acetyl transferase, PAC; green fluoresecent protein, GFP, etc) and therapeutic proteins ( Herpes simplex virus-1 thymidine kinase, HSV1TK; interleukin-12, IL-12; viral antigens, etc.) in transient transduced or stable cells lines and in animals. A full list of publications using 2A is available on the web [ 6 ]. Several publications in the past few months have shown the potential of this new co-expression strategy [ 7 - 9 ]. Co-ordinating the co-expression of all your genes Up to four genes have been successfully co-expressed from plasmids and retrovirus using several copies of the FMDV 2A or other 2A-like sequences (to avoid direct repeats in retroviruses) [ 8 , 9 ]. Not only was co-expression effective, its co-ordination was also apparent [ 7 , 9 ] (Fig. 1 ), and the imbalance in the level of the proteins expressed was low (determined to 1.2 [ 8 ]). These properties allowed polycistronic vectors bearing pac in the last position to easily generate stable cell lines co-expressing two upstream genes [ 7 , 9 ]. Figure 1 Co-ordinated co-expression to different compartments in HeLa cells. A single ORF was designed with the fluorescent genes eyfp and ecfp plus the puromycin resistant gene pac [9]. These genes were cloned flanking FMDV 2A sequences. An internal signal-anchor from the human β-1,4 galactosyltransferase (GT) was fused to the 5' end of the ecfp for Golgi targeting. During its translation, the self-processing of this polyprotein produced EYFP-2A that diffused to the cytoplasm and nucleus (due to its small size), while GT-EYFP-2A was co-translationally targeted to the Golgi apparatus by the GT signal (some protein also stays in the endoplasmic reticulum, due to the continuous cycling between these compartments). Two fields are shown, in both cases the cell on the left shows a high level of expression of both proteins that were expressed at lower levels in the cell on the right, illustrating the co-ordination obtained with the 2A co-expression strategy. PAC was able to confer resistance to puromycin. Images were taken 48 hours post-transfection. Bar represents 10 μm. Putting your proteins where they should be The CHYSEL strategy of co-expression is also compatible with the most disparate sub-cellular localisations [ 7 - 10 ]. Proteins processed by 2A from polyproteins were targeted to the cytosol, nucleus, mitochondria, endoplasmic reticulum, Golgi apparatus, plasma membrane (both, by transmembrane proteins and by cytosolic attachment due to myristoylation) and the extra-cellular compartment. Post-translationally targeted cytosolic proteins as well as co-translationally secreted and transmembrane proteins type I, II and III, have been successfully co-translated. Only one combination of co-translational signals was not correctly targeted [ 9 ]. Designing complex polyproteins for multi-gene deficiency The results reported in reference [ 8 ] should be particularly interesting for researchers in the gene therapy field. They provide a good example of the potential of the 2A co-expression strategy, introducing up to four genes in a single vector. Furthermore, they show the utility of this strategy to reconstruct a very delicate heteromultimeric protein complex on the cell surface (T-cell receptor:CD3 complex, TCR:CD3; Fig. 2A ). It is known that all six subunits are necessary for the efficient formation of the TCR:CD3 complex and just two retroviral vectors were sufficient to reconstruct it in transfected 293T or infected 3T3 cells: one encoding both subunits of the T-cell receptor and the other the four subunits of the CD3 complex (Fig. 2B ). Figure 2 Self-processing polyproteins to reconstruct the TCR:CD3 complex. (A) Schematic diagram of the TCR:CD3 complex spanning the cytoplasmic membrane. The T-cell receptor (TCR) is formed by two subunits and the other four proteins assemble in three dimers to form the CD3 complex. The square boxes in the cytoplasmic sequences of the CD3 subunits represent the immunoreceptor tyrosine-based activation motifs (ITAMs). (B) To express the TCR:CD3 complex in cells, two retroviral vectors were designed to carry the two ORFs drawn here [8]. In the retrovirus encoding the four CD3 subunits, three different 2A sequences were used to avoid deletions due to direct repetitions. Lethally irradiated CD3ε ΔP/ΔP × CD3ζ -/- mice (lacking all four CD3 subunits) were transplanted with bone marrow from wt C57BL/6 mice or CD3ε ΔP/ΔP × CD3ζ -/- mice transduced with a retrovirus encoding the four CD3 subunits, and in both cases TCR surface expression was detected and the T cells proliferated normally after immune stimulation. Bone marrow from CD3ε ΔP/ΔP × CD3ζ -/- mice without CD3 transduction did not restore T-cell development. T cells were also reconstituted in sub-lethally irradiated RAG-1 -/- mice (lacking mature T and B lymphocytes) in which bone marrow from CD3ε ΔP/ΔP mice (lacking CDε and with a severe inhibition of CD3γ and CD3δ), transduced with a retrovirus encoding these three subunits (via two 2A sequences), was used for a transplant into the RAG-1 -/- mice. The same experiment using three vectors encoding the CD3 subunits separately was unsuccessful. Conclusions The development of FMDV 2A as a cloning tool is an example of how dangerous pathogenic viruses can be harnessed by biotechnology for human benefit. Their molecular "tricks" (as IRES or CHYSEL sequences) are gradually becoming part of the biotechnologists' toolbox. The development of the polycistronic vectors here discussed is a big step forward, a decade and a half after the launching of the very first gene therapy trial with the aim of introducing in blood cells just a single therapeutic gene, adenosine deaminase (ADA), and the NEO marker [ 11 ]. These results represent a considerable advance in the correction of diseases that involve heteromultimeric proteins, several enzymes involved in a biochemical pathway or various proteins for combined/synergistic effects. 2A is not a magic tool that is going to solve all our problems, but it will help to pave the way for gene therapy. Competing interests None declared.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521497.xml
543574	Lifetime total and beverage specific - alcohol intake and prostate cancer risk: a case-control study	Background We investigated lifetime alcohol consumption and prostate cancer risk in a case-control study conducted in Buffalo, NY (1998–2001). Methods The study included 88 men, aged 45 to 85 years with incident, histologically-confirmed prostate cancer and 272 controls. We conducted extensive in-person interviews regarding lifetime alcohol consumption and other epidemiologic data. Results Prostate cancer risk was not associated with lifetime intake of total and beverage specific ethanol. In addition we found no association with number of drinks per day (average drinks per day over the lifetime) or drinks per drinking day (average drinks per day on drinking days only over the lifetime). However, we observed an inverse association with the total number of drinking years. Men in the lowest tertile of total drinking years had a two-fold prostate cancer risk than men in the highest tertile (OR 2.16, 95% CI 0.98–4.78, p for trend <0.05). Conclusion Our results suggest that alcohol intake distribution across lifetime may play a more important role in prostate cancer etiology than total lifetime consumption.	Background Prostate cancer is the most frequently diagnosed malignancy and the second leading cause of cancer death among men in the Western countries [ 1 ]. Notwithstanding the importance of this malignancy, little is understood about its cause. To date the only well established risk factors are age, family history of disease, race and country of residence [ 2 ], while the body of the evidence about the role of alcohol intake is still controversial. Since alcohol consumption is a common lifestyle factor and potentially modifiable, the finding of an association with prostate cancer could have an important impact on public health. Among the population-based case-control studies, those carried out by Heyes et al. [ 3 ] and Sharpe et al. [ 4 ] found an increased risk of prostate cancer associated with alcohol consumption. Risk increased with increasing frequency of alcohol consumption [ 3 ] and among those who drank regularly over a longer period [ 4 ]. Sesso et al., in their prospective cohort study, confirmed the finding of a higher risk associated with alcohol consumption [ 5 ]. However, numerous studies published since 1998 have not found an association between alcohol intake and prostate cancer [ 6 - 17 ]. In a review by Breslow and Weed, only 6 of 32 studies reported a positive association between alcohol use and prostate cancer [ 18 ]; however, they noted that many of the studies had biases that could have attenuated the risk estimates. Although prostate cancer is known to have a long latency period, lifetime alcohol consumption was not addressed in the studies carried out until the late 1990s, and rarely in the more recent studies [ 18 ]. Furthermore, investigators focusing on this topic have considered lifetime alcohol consumption as the average total amount of alcohol consumed over the lifetime, rarely taking into account such characteristics as number of drinks consumed on a typical drinking day or other descriptions of drinking pattern. The distribution of an equivalent volume of alcohol across multiple drinking occasions rather than a single occasion (e.g., one drink per day vs. seven drinks on single day) is likely to have different physiologic effects and impact on cancer risk. Likewise, an examination of average total lifetime alcohol intake does not address the possibility that, although the total lifetime volume may not differ, the duration of intake may, thus effectively resulting in a higher dose over a shorter time period. Alcohol may act as a carcinogen itself and may also modulate risk from other carcinogen exposures. It has been implicated in risk of cancer at a number of sites [ 19 , 20 ]. In the present case-control study we examined the association between lifetime alcohol intake, duration of alcohol use, and drinks per day with risk of prostate cancer in western New York. Methods We conducted a case-control study of prostate cancer and hormones and alcohol intake (the PROMEN STUDY) in Erie and Niagara Counties, NY, USA, between December 1998 and April 2001. The methods for this study have been previously described in detail [ 21 ]. Participants provided informed consent; the Institutional Review Board of the University at Buffalo, School of Medicine and Biomedical Science, and each of the participating hospitals approved the procedures for the protection of human subjects recruited for the study. Cases were men aged 45 to 85 years with incident, primary, histologically confirmed prostate cancer. Men with a previous history of cancer (except non-melanoma skin cancer), or already on hormonal or chemotherapy treatment (current or in the 6 months prior to diagnosis), as well as those affected by chronic or acute liver diseases, were excluded. Cases aged 35–65 years were also required to have a driver's license, because we used driver's license records to identify age matched controls. During the study period, 504 men were identified with incident prostate cancer. Of these, 336 men did not meet the eligibility criteria; we invited the remaining 163 patients to participate in the PROMEN study. After being contacted, 50 men refused to participate resulting in a participation rate of 70%. Ninety-six patients had complete data for the variables of interest. Controls aged between 35 and 65 years were selected from a list of individuals holding a New York State driver's license and residing in Erie and Niagara Counties. Those aged 65 and over were selected from the rolls of the Health Care Financial Administration. As with cases, men on hormonal treatment (current or in the 6 months prior the diagnosis), or diagnosed with metabolic or endocrine disease were excluded, as well as participants with a previous story of cancer other than non-melanoma skin cancer. Since it is well known that latent prostatic carcinoma has a high prevalence in men over 50 [ 22 , 23 ], we evaluated prostate specific antigen (PSA) in the blood samples obtained from controls. Controls found to have a PSA value higher than 4 ng/ml were excluded from the control group, in accordance with the criterion established by the American Cancer Society Prostate Cancer Detection Project [ 24 ] until the completion of further diagnostic procedures to clarify their true case-control status. During the study period, 1373 potential controls were contacted. One hundred and seventy nine of these individuals were deceased or were too ill to participate, 293 did not meet the eligibility criteria and we were not able to contact 272 persons. We identified eight prostate cancer cases as a result of PSA determination in subjects who initially were recruited as controls. Three hundred and seventeen of the remaining 513 subjects (60%) were enrolled and interviewed: 304 had complete data for analysis. Extensive data on demographics, smoking history, alcohol consumption, and other study variables were collected by trained interviewers during in-person computer-assisted interviews [ 25 ] and with self-administered questionnaires. Height, weight, waist and hip circumferences were measured by trained technicians using a standardized protocol. Body mass index (BMI) was calculated as weight in kilograms divided by square of the height in meters (kg/m 2 ). Waist to hip ratio (WHR) was calculated as waist circumference divided by hip circumference. Alcohol intake Detailed information on alcohol consumption throughout the lifetime was collected using the Cognitive Lifetime Drinking History [ 26 , 27 ]. Prior to the interview, participants completed a lifetime events calendar on which they recorded the date and their age when significant events in their life occurred. The calendar was used during the interview to help them remember what they were doing during specified periods of their lives and whether drinking alcohol was involved. Participants reported the age when they started drinking alcohol regularly (at least once a month for six months) and when their drinking changed over the years. When changes were reported, participants were asked whether they continued regular drinking; if not, they were asked if they ever resumed regular drinking. Using this information, we defined intervals during each participant's life when drinking patterns were relatively homogeneous and computed the total number of drinking years and the total number of abstinent years. Lists of alcoholic beverages, beer, wine, wine coolers, and liquor, and models of glasses and bottles were used to help respondents recall what beverages they drank over their lifetimes; their usual drink size of each beverage; and whether drink size changed over their lifetimes. This provides information used to: (1) calculate absolute alcohol intakes and (2) tailor the computer-assisted interview to the each respondent's drinking history. Patterns of drinking were ascertained for intervals during which respondents drank weekly or more often by asking how often respondents drank on Fridays, Saturdays, Sundays, and weekdays, and how many drinks they usually had on each. For intervals during which respondents drank less often than weekly, they were asked standard quantity and frequency questions. Quantity and frequency for times when they drank more than usual were assessed for all intervals, as was the frequency of intoxication; the proportion of drinks they had with meals/snack/without eating; and the proportion of drinks from beer, wine, wine coolers, and liquor. Drinks per interval was estimated by multiplying quantity by frequency for days of the week and more than usual and adding. Drinks per interval was translated into ounces of ethanol per interval based on the proportion of drinks represented by specific beverages, respondents' beverage-specific drink size in ounces, and factors representing the average percent per ounce of absolute alcohol for a given beverage to estimates of drinks per interval. Factors used were 0.048, 0.12, 0.04 and 0.40, for beer, wine, wine cooler and hard liquor, respectively. These estimates were summed across drinking intervals to yield lifetime totals. We considered several variables in these analyses: total number of years alcohol was consumed, number of drinks per day during the drinking years (total number of drinks/total number of days in drinking years), number of drinks per drinking day (total number of drinks/total number of days on which alcohol was consumed in drinking years), total lifetime ounces of ethanol and beverage-specific total lifetime ounces of ethanol. Because few participants consumed wine coolers, wine and wine coolers were combined. A drink was defined as 12 ounces of beer, 5 ounces of wine, and 1.5 ounces of liquor. Statistical analysis Statistical analyses were conducted using SPSS for Windows version 11.0. Differences between cases and controls in demographic characteristics and alcohol consumption were assessed using t-tests for continuous variables and χ 2 for categorical variables. Lifetime abstainers, defined as those subjects who had less than 12 drinks in any one year over their lifetimes, were excluded from our analyses. The biological and social differences between lifetime abstainers and both former and current drinkers [ 28 , 29 ] and the very low number of these subjects in our sample (5 cases and 11 controls) represent the reasons for their exclusion from our analyses. Our final sample size for analysis included 88 cases and 272 controls. In analyses of risk associated with lifetime alcohol intake, tertiles of total and beverage specific ounces and total drinking years were computed based on the distribution in the controls. For the beverage specific analyses, non-drinkers were those respondents not consuming that particular alcoholic beverage. For risk associated with drinks per day and drinks per drinking day, we categorized consumption as two drinks or less per day and greater than two drinks per day. Odds ratios (OR) and 95% confidence intervals (CI) for risk of prostate cancer associated with alcohol consumption were computed using unconditional logistic regression adjusting for age, cigarette smoking status, education, body mass index (BMI), and waist to hip ratio (WHRATIO). The beverage specific analyses were further mutually adjusted for the other beverages. Results Characteristics of the participants in the PROMEN study are shown in Table 1 . Compared to cases, controls were slightly more educated (13.0 vs. 12.3 years) and more likely to be Caucasian (93.0% vs. 67%). No statistically significant differences between cases and controls were observed for age, body mass index, waist to hip ratio, smoking or drinking status. Table 1 Characteristics of prostate cancer cases and controls, PROMEN Study Cases (n = 88) Controls (n = 272) Mean (SD a ) Age, years 69.3 (8.4) 70.0 (6.3) Education, years 12.3 (2.7) b 13.0 (2.8) Body mass index, kg/m 2 29.2 (5.2) 28.6 (4.6) Waist to hip ratio Percent Race White 67.0 c 93.4 Non white 33.0 6.6 Smoking status d Never 23.8 28.3 Former 61.4 61.8 Current 14.8 9.9 Drinking status e Non-current drinkers 36.4 23.5 Current drinkers 63.6 76.5 a standard deviation; b p < 0.05, t-tests for differences in means between cases and controls; c p < 0.001, χ 2 for differences in categorical variables between cases and controls; d smoking status at the time of diagnosis in cases or interview in controls; e drinking status in the 12–24 months prior to diagnosis or interview, non-current drinkers stopped drinking at least 12–24 months prior to interview Means and standard deviations for aspects of lifetime alcohol consumption for the sample overall and by current drinking status are shown in Table 2 . Among drinkers overall and current drinkers, cases drank for fewer years than did controls (38.2 vs. 43.7 years, p < 0.05 and 41.3 vs. 46.8 years, p < 0.05, overall and current drinkers, respectively) and, consequently, had greater numbers of years abstaining. Few differences in lifetime total and beverage-specific ounces consumed, drinks per day, or drinks per drinking day were observed between cases and controls for drinkers overall or current drinkers. However, although not statistically significant, we observed several differences in alcohol consumption between cases and controls who were former drinkers. Among former drinkers, cases consumed more total ethanol, beer and liquor, more drinks per day and more drinks per drinking day, but consumed less ethanol from wine and wine coolers compared to controls. Table 2 Selected lifetime alcohol consumption characteristics among prostate cancer cases and controls, PROMEN study All drinkers (n = 360) Former drinkers (n = 96) Current drinkers (n = 264) Cases (n = 88) Controls (n = 272) Cases (n = 32) Controls (n = 64) Cases (n = 56) Controls (n = 208) Mean (SD) Mean (SD) Mean (SD) Total drinking years 38.2 a (16.5) 43.7 (14.9) 32.9 (18.5) 33.8 (17.2) 41.3 a (14.5) 46.8 (12.7) Total abstaining years 11.4 a (15.0) 6.6 (12.5) 19.8 (16.4) 18.2 (15.3) 6.6 a (11.9) 3.0 (8.8) Drinks per day 2.6 (7.3) 1.6 (3.4) 4.7 (11.6) 2.5 (5.8) 1.3 (1.7) 1.3 (2.2) Drinks per drinking day 4.5 (7.3) 3.6 (4.3) 6.8 (11.3) 5.0 (6.3) 3.2 (2.5) 3.2 (3.4) Total lifetime ethanol, ounces 12904.7 (18681.0) 11735.3 (12904.7) 19051.0 (26382.6) 13498.8 (21019.7) 9392.6 (11187.8) 11192.7 (16880.9) Total lifetime ethanol from beer, ounces 6282.5 (11321.0) 6024.3 (9250.0) 7771.0 (15173.8) 5992.6 (12284.7) 5431.9 (8422.3) 6034.1 (8129.3) Total lifetime ethanol from liquor, ounces 5654.2 (14571.6) 4067.2 (12815.8) 10307.0 (22051.6) 5233.7 (11480.5) 2995.5 (6480.2) 3708.3 (13204.5) Total lifetime ethanol from wine/wine coolers, ounces 953.1 (2715.6) 1634.6 (4168.8) 958.9 (3588.4) 2271.0 (6154.0) 949.8 (2099.5) 1438.8 (3326.0) a p < 0.05, t-tests for differences in means between cases and controls Odds ratios and 95% confidence intervals for the risk of prostate cancer associated with lifetime alcohol consumption are shown in Table 3 . We observed no associations with risk with lifetime ounces of total ethanol, beer, wine, or liquor. Risk associated with total drinking years, years of abstaining (ever/never), drinking status, drinks per day, and drinks per drinking day are shown in Table 4 . Compared to the highest tertile of total drinking years, men in the lowest tertile had a marginally significant increased risk (OR 2.16, 95% CI 0.98–4.78, p for trend <0.05) and, similarly, men reporting ever abstaining compared to those who never abstained had increased prostate cancer risk (OR 1.79, 95% CI 1.05–3.03). No associations with risk were observed for former vs. current drinkers, drinks per day, or drinks per drinking day. Table 3 Odds ratios (OR) a and 95% confidence intervals (CI) for risk of prostate cancer associated with lifetime alcohol consumption Cases (n = 88) Controls (n = 272) Odds Ratios (95% CI) Total lifetime ethanol, ounces ≤2647.62 29 90 1.00 2647.62 – 11048.28 34 90 1.20 (0.65–2.23) >11048.28 25 92 0.83 (0.43–1.60) Total lifetime ethanol from beer, ounces b ≤1941.78 42 120 1.00 1941.78 – 6237.30 25 75 1.16 (0.62–2.16) >6237.30 21 77 0.89 (0.46–1.72) Total lifetime ethanol from liquor, ounces b ≤932.23 51 152 1.00 932.23 – 3976.79 15 59 0.71 (0.35–1.44) >3976.79 22 61 0.91 (0.47–1.76) Total lifetime ethanol from wine and wine cooler, ounces b ≤511.66 67 177 1.00 511.66 – 2283.00 10 47 0.76 (0.35–1.65) >2283.00 11 48 0.60 (0.27–1.30) a Adjusted for race, age (years), smoke, education (years), BMI, WHRATIO; b further mutually adjusted for other beverages Table 4 Odds ratios (OR) a and 95% confidence intervals (CI) for risk of prostate cancer associated with lifetime alcohol consumption: duration, drinking status, drinks per day, and drinks per drinking day. Cases (n = 88) Controls (n = 272) Odds Ratios (95% CI) Total drinking years >53 14 80 1.00 42 – 53 27 94 1.44 (0.66–3.14) ≤42 47 92 2.16 b (0.98–4.78) Ever abstained from drinking never abstained 39 173 1.00 ever abstained 49 99 1.79 b (1.05–3.03) Drinking status c current drinkers 56 208 1.00 former drinkers 32 64 1.40 (0.77–2.53) Drinks per day ≤2 62 218 1.00 >2 26 54 1.38 (0.76–2.51) Drinks per drinking day ≤2 24 106 1.00 >2 64 166 1.57 (0.88–2.79) a Adjusted for race age, smoke, education (years), BMI, WHRATIO; b p for trend <0.05; c drinking status in the 12–24 months prior to diagnosis or interview. Former drinkers stopped drinking at least 12–24 months prior to interview. Discussion The assessment of lifetime alcohol consumption in cancer etiology has been predominantly expressed through the calculation of either total lifetime volume or an average volume per specified time period across the lifetime. Few investigations have evaluated lifetime drinking patterns in relation to prostate cancer risk. While methodological difficulties challenge the evaluation of drinking patterns, our results suggest that failure to take into account aspects of drinking pattern such as the relative duration and dose of consumption may reduce our ability to clearly elucidate the role alcohol may be playing in cancer development. Although we observed no associations with risk for total lifetime alcohol intake or when alcohol was expressed as average drinks per day or even average drinks per drinking day, our results suggest that the impact may differ when the same volume of alcohol consumption takes place in fewer drinking years over a lifetime. Furthermore, it is notable that alcohol consumption was much higher among the cases compared with controls who were former drinkers. As alcohol consumption has been positively related to many causes of morbidity, a proportion of these men may have stopped drinking in response to poor health. Whether pre-existing morbid conditions or heavier drinking is related to subsequent development of prostate cancer remains to be clarified. Our study has several strengths and limitations. A limitation of our study is the small sample size, especially for cases. However, because one of the original aims of the study was an examination of hormones and prostate cancer, both cases and controls were carefully identified. To eliminate the effect on hormone levels by treatment, cases were enrolled in the study prior to starting chemotherapy or hormone therapy; thus increasing the difficulty of case ascertainment. On the other hand, the exclusion of controls with high circulating PSA levels helped to reduce misclassification and to ensure that the control group was free from prostate cancer. The data used in the present analysis were collected as a part of an in-person interview, and the questionnaire about lifetime alcohol consumption was very detailed allowing us to compute both the dose and frequency aspects of alcohol consumption. Given the difficulties involved in measuring alcohol consumption, studies utilizing data collected before diagnosis would appear more likely to lead to valid inferences. Recently, Dennis in his meta-analysis [ 30 ] pointed out that in many of the published cohort studies alcohol consumption was assessed only at a baseline, often many years before the diagnosis, with no subsequent assessment to quantify changes in drinking pattern. While retrospective assessment of lifelong alcohol consumption at diagnosis may appear to be more likely to lead to recall bias, such an assessment may also be more likely to capture relevant attributes of exposure, such as overall duration of alcohol use and timing of potentially important changes in use, such as quitting. These differences are not always taken into account [ 30 ]. The plausibility of alcohol as a risk factor for prostate cancer relates to evidence that alcohol may act as a carcinogen or may modulate risk from other known carcinogens through generation of free radicals, affecting the metabolism of detoxification enzymes, impairment of immune system and depression of DNA repair enzymes [ 31 ]. It remains unclear to what extent alcohol could affect the early phases of cancer development. Some studies suggest that the critical period of exposure may be as early as adolescence as the development of prostate gland begins prenatally, continuing until the end of puberty [ 31 ]. If alcohol contributes to cancer promotion, duration and relative intensity of exposure during a specified period of time, instead of the total amount of the agent itself over the entire life time course may be important. Conclusions Further studies focusing on lifetime exposure and more specifically on patterns of consumption may help in prevention of a disease with considerable public health impact. List of Abbreviations BMI body mass index PSA prostate-specific antigen WHRATIO waist-to-hip ratio Competing interests The authors declare that they have no competing interests. Authors' contributions MB performed statistical analyses, interpreted the results and drafted the manuscript SEM performed statistical analyses, interpreted the results and revised the manuscript HJS interpreted the results and revised the manuscript SS interpreted the results and revised the manuscript BJF performed data analysis, interpreted the results and revised the manuscript SDP interpreted the results and revised the manuscript GC interpreted the results and revised the manuscript JLF interpreted the results and revised the manuscript MT revised the manuscript MR defined the exposure variables and revised the manuscript TN defined the exposure variables and revised the manuscript PM designed and implemented the study, interpreted the study results, revised the manuscript	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543574.xml
546418	Ectopic expression of PTTG1/securin promotes tumorigenesis in human embryonic kidney cells	Background Pituitary tumor transforming gene1 (PTTG1) is a novel oncogene that is expressed in most tumors. It encodes a protein that is primarily involved in the regulation of sister chromatid separation during cell division. The oncogenic potential of PTTG1 has been well characterized in the mouse, particularly mouse fibroblast (NIH3T3) cells, in which it induces cell proliferation, promotes tumor formation and angiogenesis. Human tumorigenesis is a complex and a multistep process often requiring concordant expression of a number of genes. Also due to differences between rodent and human cell biology it is difficult to extrapolate results from mouse models to humans. To determine if PTTG1 functions similarly as an oncogene in humans, we have characterized its effects on human embryonic kidney (HEK293) cells. Results We report that introduction of human PTTG1 into HEK293 cells through transfection with PTTG1 cDNA resulted in increased cell proliferation, anchorage-independent growth in soft agar, and formation of tumors after subcutaneous injection of nu/nu mice. Pathologic analysis revealed that these tumors were poorly differentiated. Both analysis of HEK293 cells transiently transfected with PTTG1 cDNA and analysis of tumors developed on injection of HEK293 cells that had been stably transfected with PTTG1 cDNA indicated significantly higher levels of secretion and expression of bFGF, VEGF and IL-8 compared to HEK293 cells transfected with pcDNA3.1 vector or uninvolved tissues collected from the mice. Mutation of the proline-rich motifs at the C-terminal of PTTG1 abolished its oncogenic properties. Mice injected with this mutated PTTG1 either did not form tumors or formed very small tumors. Taken together our results suggest that PTTG1 is a human oncogene that possesses the ability to promote tumorigenesis in human cells at least in part through the regulation of expression or secretion of bFGF, VEGF and IL-8. Conclusions Our results demonstrate that PTTG1 is a potent human oncogene and has the ability to induce cellular transformation of human cells. Overexpression of PTTG1 in HEK293 cells leads to an increase in the secretion and expression of bFGF, VEGF and IL-8. Mutation of C-terminal proline-rich motifs abrogates the oncogenic function of PTTG1. To our knowledge, this is the first study demonstrating the importance of PTTG1 in human tumorigenesis.	Background Pituitary tumor transforming gene 1 (PTTG1), a recently characterized oncogene, was initially identified on analysis of a rat pituitary tumor [ 1 ]; subsequently, a human homologue of PTTG1 was cloned by us and others [ 2 - 4 ]. Three members (PTTG1, PTTG2 and PTTG3) of the PTTG family, which exhibit differential expression in normal and tumor cells have been reported [ 5 ], although only PTTG1 has been studied in detail. PTTG1 is located on chromosome 5q35.1 [ 6 ], a locus associated with recurrent lung cancer and myelogenous leukemias [ 7 ]. Moreover, it has been shown to be expressed highly in various tumors, and cell lines derived from such tumors, including tumors of the pituitary, thyroid, colon, ovary, testicles, and breast [ 8 - 12 ]. In normal tissues, its expression is low or undetectable except in testis [ 4 , 1 ]. Recent studies have indicated that elevated expression of PTTG1 in some tumors may serve as a prognostic marker for tumor invasiveness and metastasis [ 13 ]. A clue to its function was gained from its structural similarity with the yeast securin, which led to its identification as a human securin [ 14 ] and suggested that it may play a role in regulation of sister chromatid separation. It appears, however, to have multiple effects in cells with enhanced expression being associated with an increase in the expression of the c-myc oncogene [ 15 ], an increase in the expression of p53 [ 16 , 17 ], an increase in the secretion and expression of basic growth factor (bFGF) [ 18 ] and an increase in the secretion and expression of vascular endothelial growth factor [VEGF) [ 18 , 19 ]. To date, the evidence for the oncogenic function of PTTG1 has been obtained by overexpression of PTTG1 in mouse fibroblast cells (NIH3T3) followed by assessment of its ability to induce cellular transformation in vitro (colony formation in soft agar) and tumor formation in nude mice [ 2 , 4 ]. Due to the biological differences between human and rodent cells, however, care must be taken in extrapolating results obtained using rodent cells to human cells. There are now a number of examples in which it has been demonstrated that although overexpression of an oncogene can induce transformation of primary rodent cells [ 20 ], it fails to induce transformation of the same cell type derived from humans. Usually this failure is attributable to the requirement for co-expression of another gene or oncogenic cooperation of another gene [ 21 - 26 ]. Similarly, much of the evidence concerning the mechanisms by which PTTG1 may affect the phenotype of the cell has been obtained using transfected NIH3T3 cells. It is known that the secretion of growth factors and cytokines by tumor cells, and the cells that infiltrate and surround the tumor mass play an essential role in the regulation of tumor growth and metastasis [ 27 ]. Both bFGF and VEGF have been implicated in tumorigenesis and the expression and secretion of these molecules has been demonstrated on transfection of NIH3T3 cells with PTTG1 cDNA [ 18 , 19 ], but this has not been confirmed on transfection of human cells. The effects of expression of PTTG1 on another cytokine that is known to play a key role in tumorigenesis, interleukin-8, (IL-8), have not yet been analyzed. The purposes of this study were, therefore, three-fold. Firstly, to determine whether PTTG1 can induce cellular transformation of normal human cells; secondly, to determine if PTTG1 is sufficient in itself to induce transformation; and, thirdly, to characterize the changes in secretion and expression of key metastatic, angiogenic and chemokine factors (bFGF, VEGF and IL-8) associated with PTTG1-mediated transformation in human cells. For these studies, we selected the human embryonic kidney (HEK293) cell line as our model. The transformation of these cells by human adenovirus type 5 prevents their senescence [ 28 ]. These cells have been reported to have a moderate tumorigenic potential [ 29 ] and have been used as a cellular model for normal human cells to study the oncogenic potential of a number of genes [ 29 - 31 ]. Mice xenografted with these cells do not develop tumors even after three months of injection [ 31 ]. Results Generation of HEK293 cells stably expressing PTTG1 and mPTTG1 HEK293 cells were transfected with pcDNA3.1-PTTG1, pcDNA3.1-mPTTG1 or pcDNA3.1 vector. After G418 selection, 10 clones from each of pcDNA3.1, pcDNA3.1-PTTG1 or pcDNA3.1-mPTTG1 transfected cells were picked, cultured and expanded. The PTTG1 protein expression of these transfectants was detected by western blot analysis using PTTG1 antiserum. Two representative clones from PTTG1 transfected (named HEKPTTG1-1 and HEKPTTG1-3) and mPTTG1 transfected (named HEKmPTTG1-2 and HEKmPTTG1-4), and one clone from pcDNA3.1 vector (named HEKpcDNA3.1) was selected for further studies. Selection of clones was based on the level of expression of PTTG1 protein. Fig. 1 shows the protein expression of these clones. Transfection of cells with the pcDNA3.1 vector resulted in expression of very low level of PTTG1 protein. The clones of the pcDNA3.1-PTTG1 and pcDNA3.1-mPTTG1 transfected cells that exhibited approximately equivalent levels of expression of PTTG1 and mPTTG1 proteins were processed to establish stable cell lines. Figure 1 Western blot analysis of HEK293 cells transfected with pcDNA3.1, pcDNA3.1-PTTG1 or pcDNA3.1-mPTTG1. a: HEKpcDNA3.1, b: HEKPTTG1-1, c: HEKPTTG1-3, d: HEKmPTTG1-2, and, e: HEKmPTTG1-4 cells. Stable transfection of PTTG1 induces cell proliferation and transformation of HEK293 cells overexpressing PTTG1 Previously, we have shown that over expression of PTTG1 in mouse fibroblast NIH3T3 cells results in an increase in cell proliferation [ 4 ]. To determine if over expression of PTTG1 in HEK293 cells produces similar effects, we estimated the proliferation at 24, 48, 72 and 96 hours after plating of stably transfected HEK293 cells expressing high levels of PTTG1 or mPTTG1 protein. Both clones of PTTG1-transfected cells (HEKPTTG1-1 and HEKPTTG1-3) exhibited significantly greater proliferation than the cells transfected with vector only at all time points tested, and the time course of proliferation was very similar in both clones, increasing by 20–30% after 24 hours, 30–40% after 48 hours. This level of increase in cell proliferation was retained at least up to 96 hours (Fig. 2 ). Somewhat surprisingly, the proliferation of the cells expressing the mutated PTTG1 was equivalent to that of the cells expressing the wild-type PTTG1 and was significantly higher than that of the cells transfected with vector only (Fig. 2 ). These experiments indicate that over expression of PTTG1 does induce a significant proliferative effect in HEK293 cells; however, at least under the conditions used, mutation of the proline-rich motifs of PTTG1 does not affect this response. Figure 2 Cell proliferation of HEK293 cells stably transfected with pcDNA3.1, PTTG1 or m-PTTG1. The 5 × 10 3 cells were plated/well. The results are expressed as % of control (HEK293 cells stably transfected with pcDNA3.1 control vector). Error bars represent ± SEM (n = 4) of three independent experiments. Overexpression of PTTG1 induces cellular transformation As anchorage-independent growth is considered to be an in-vitro test for angiogenesis we assayed the effects of transfection with PTTG1 on the ability of the HEK293 cells to form colonies in soft agar. As shown in Fig. 3 , over expression of PTTG1 in HEK293 cells resulted in a higher incidence of colony formation than that observed on transfection with the vector only. The cells transfected with vector only formed few colonies and these were of small size during 14 days of culture, whereas both the cell lines expressing wild type PTTG1 formed a significantly higher number of colonies, which were of a large size. The incidence of colony formation was 2% for HEKpcDNA3.1 cell line but was 19% for the HEKPTTG1-1 cell line and 30% for the HEKPTTG1-3 cell line. In this case, mutation of the proline-rich motifs of PTTG1 resulted in a significant reduction in the number of colonies formed with the incidence of colony formation for the HEKmPTTG1-2 and HEKmPTTG1-4 cell lines being similar to the vector-only transfected cells (HEKpcDNA3.1). These results suggest that over expression of PTTG1 in HEK293 cells induces cellular transformation, and mutation of proline-rich motifs does not effect the cell proliferation but abrogates the cellular transformation ability of PTTG1. Figure 3 Colony formation of HEK293 cells stably transfected with pcDNA3.1, pcDNA3.1-PTTG1 or pcDNA3.1-m-PTTG1 a: HEKpcDNA3.1, b: HEKPTTG1-1, c: HEKPTTG1-3, d: HEKmPTTG1-2 and e: HEKmPTTG1-4. The data shown is representative of three independent experiments. PTTG1 induces tumor formation in nude mice injected with HEK293 cells stably expressing PTTG1 protein To determine whether PTTG1 promotes tumor formation in nude mice, we subcutaneously injected nude mice with HEK293 cells expressing PTTG1 or mPTTG1. Three out of four mice injected with the HEKPTTG1-1 or HEKPTTG1-3 cell lines developed large size tumors within four weeks of injection (Fig. 4 ). Pathologic analysis of the tumors revealed that they were poorly differentiated (Fig. 5 ). Mice injected with the HEKmPTTG1-2 cell line did develop tumors, but the tumors were of a small size. None of the mice injected with the other cell line-expressing mutant PTTG1 (HEKmPTTG1-4) or the vector-only cell line (HEKpcDNA3.1) developed tumors within the time frame of this experiment (Fig. 4 ). The tumor volumes, measured at the end of experiment (six weeks after injection of cells), were 150–1320 mm 3 for HEKPTTG1-1, 72–1404 mm 3 for HEKPTTG1-3 and 8.8–12.6 mm 3 for HEKmPTTG1-2 (Table 1 ). These results clearly demonstrate that PTTG1 gene is a potent oncogene. Moreover, they demonstrate that PTTG1 possesses the ability to enhance the tumorigenic potential of immortal human cells and that it does not require the ectopic co-expression of other oncogene(s) to achieve its tumorigenic function. Figure 4 Tumor development in nu/nu mice on injection of HEK293 cells stably transfected with pcDNA3.1, pcDNA3.1PTTG or pcDNA3.1mPTTG1 cells. Each mouse was injected with 1 × 10 6 cells. After 6 weeks of injection, mice were photographed and sacrificed, tumors and other tissues were collected and tumor volume was measured a: Mouse injected with HEKpcDNA3.1 cells, b: mouse injected with HEKPTTG1-1 cells, c: mouse injected with HEKPTTG1-3 cells and d: mouse injected with HEKmPTTG1-2. Arrows indicate the tumors. Figure 5 Histolopathological analysis of the tumors excised from animals injected with HEK293 cells expressing PTTG1 or m-PTTG1. Tumors were fixed, sectioned and stained for H & E. A: Tumor from animal injected with PTTG1-1, B: Tumor from animal injected with HEKPTTG1-3. Table 1 Tumor formation induced by PTTG1 expressing HEK293 cells in nude mice Stable Cells Animals with tumor Tumor Volume HEKpcDNA3.1 0/4 NA HEKPTTG1-1 3/4 150–1320 mm 3 HEKPTTG1-3 3/4 72–1404 mm 3 HEKmPTTG1-4 2/4 8.8–12.6 mm 3 HEKmPTTG1-2 0/4 NA PTTG1 stimulates expression and secretion of bFGF, VEGF and IL-8 Local invasive growth is a key feature of primary malignant tumors. A correlation between the levels of expression of PTTG1 with increased tumor invasiveness and with the degree of malignancy has been demonstrated in pituitary and colorectal tumors [ 9 , 35 ]. The specific mechanisms by which PTTG1 facilitates the invasive behaviors of tumor cells remain obscure, however. Recently, Ishikawa et al [ 18 ] and McCabe et al [ 19 ] have shown that transfection of NIH3T3 cells with PTTG1 cDNA results in an increase in secretion and expression of both bFGF and VEGF. A direct correlation between high IL-8 expression and tumor metastases has been shown in a number of cancers [ 36 - 38 ], and IL-8 also has been reported to possess mitogenic [ 39 ] and angiogenic effects [ 40 ]. We therefore measured the levels of bFGF, VEGF and IL-8 in HEK293 cells transiently transfected with pcDNA3.1 or pcDNA3.1-PTTG1 cDNA and in tumors developed on injection of nude mice with HEK293 cells that constitutively express PTTG1. As shown in Fig. 6A , the levels of bFGF, VEGF and IL-8 were comparatively higher in conditioned medium of cells transfected with pcDNA3.1-PTTG1 cDNA than from cells transfected with pcDNA3.1 vector only. Cells transfected with pcDNA3.1-PTTG1 showed a 2-fold increase in bFGF, a 3.5-fold increase in VEGF and a 2-fold increase in IL-8 levels compared to cells transfected with pcDNA3.1 vector only. Measurement of the mRNA levels of these factors by RT/PCR showed significantly higher levels in cells transfected with pcDNA3.1-PTTG1 cDNA compared to cells transfected with pcDNA3.1 vector (Fig. 6B ). To determine, if over expression of PTTG1 results in increase in levels of bFGF, VEGF and IL-8 in vivo, we measured the levels of bFGF, VEGF and IL-8 proteins in lysates from tumors developed on injection of nude mice with HEK293 cells stably transfected with PTTG1. As shown in Fig. 7 , the levels of bFGF, VEGF and IL-8 were significantly higher in three out of four tumors compared to normal tissues (kidney, liver, lung and heart) collected from the same animals. Since the size of the tumors that developed on injection of cells expressing mutated PTTG1 (HEKmPTTG1-2) were small, we were unable to analyze the bFGF, VEGF and IL-8 levels in these tumors. Measurement of mRNA for bFGF, VEGF and IL-8 revealed significantly higher levels of expression in tumors compared to normal tissues and tumors developed on injection of HEKmPTTG1-2 cells. As expected, levels of PTTG and mPTTG were higher in tumors as compared to normal tissues (Fig. 8 ). bFGF levels were found to be comparatively higher in heart which is consistent with other investigators [ 41 ] Taken together our results suggest that over expression of PTTG1 in HEK293 cells results in an increase in secretion and expression of bFGF, VEGF and IL-8 in vitro and in vivo, suggesting that increase in secretion and expression of bFGF, VEGF and IL-8 by PTTG1 may be one of the mechanisms by which PTTG1 achieves its oncogenic function and increases tumor angiogenesis. Figure 6 Introduction of PTTG in HEK293 cells induces secretion and expression of bFGF, VEGF and IL-8. HEK293 cells were transiently transfected with pcDNA3.1 or pcDNA3.1-PTTG1 cDNA. The culture medium was collected and lyophilized. bFGF, VEGF and IL-8 secreted in culture medium were measured by ELISA. A: Amount of bFGF, VEGF and IL-8 in culture medium. Vector: cells transfected with pcDNA3.1 vector DNA; PTTG1: Cells transfected with pcDNA3.1-PTTG1 cDNA. B: Expression of bFGF, VEGF and IL-8 mRNA in cells. Lane 1: pcDNA3.1 transfected cells and Lane 2: pcDNA3.1-PTTG1 transfected cells. GAPDH was used as an internal control. The data are representative of two independent experiments. Figure 7 Analysis of bFGF, VEGF and IL-8 expression in tumors and other tissues. Tumors and other tissues were excised from the animals injected with HEK293 cells stably transfected with PTTG1 (clone 1 and clone3) and homogenized. bFGF, VEGF and IL-8 in the homogenates were analyzed by ELISA. Each analysis was performed in triplicate tissue and was normalized to mg of protein. Error bars represent ± SEM of three independent experiments. Figure 8 Analysis of expression of PTTG bFGF, VEGF and IL-8 mRNAs from tumors and other tissues collected from mice subcutaneously injected with HEK293 cells stably expressing PTTG1. A: RT-PCR analysis and, B: Western blot analysis. T1: tumor 1, T2: tumor 2, M: HEKmPTTG1-2, H: heart, K: Kidney, Li: liver, Lu: lung. GAPDH and β-actin were used as control to examine equal loading. The gels are representative of two independent experiments. Discussion Oncogenic function of PTTG1 was established by its overexpression in mouse fibroblast cell line (NIH 3T3) followed by assessment of its ability to induce cellular transformation and tumor formation in nude mice [ 2 , 4 ]. However, the differences in biology between the rodent cells and human cells have brought the validity of this model into question. There are a number of instances in which an oncogene has been shown to induce transformation in rodent cells but has failed to induce transformation of same types of cells obtained from humans. To test the ability of PTTG1 to induce transformation in human cells, we selected the human embryonic kidney-293 (HEK293) cell line as our model. Our data clearly demonstrate that over expression of PTTG1 in HEK293 results in an increase in cell proliferation induces cellular transformation in-vitro (increase in anchorage-independent growth) and promotes tumor formation in nude mice. Cells transfected with pcDNA3.1 vector did not form colonies in soft agar or develop into tumors on implantation in nude mice, confirming that HEK293 cells do not possess a tumorigenic phenotype. Furthermore our data suggest that overexpression of PTTG1 in these cells results in a greater propensity for tumor development, a shortened latency period and an enhanced growth rate compared with pcDNA3.1-transfected control HEK293 cells. A second issue that we were able to address using the HEK293 cell model was the question of the ability of PTTG1 to induce transformation of normal human cells, i.e., whether it acts alone or in cooperation with another oncogene to achieve its tumorigenic function. It has been reported that a single oncogene may not be sufficient for induction of transformation but requires co-expression, or oncogenic cooperation, of another oncogene(s) to induce tumorigenesis in normal primary human cells [ 20 , 21 , 24 , 25 , 42 , 43 ]. Our results clearly show that PTTG possess the ability to induce cellular transformation in vitro and promotes tumor formation in nude mice. Since, HEK293 cells are transformed with adenovirus type 5 thus making them different from the normal primary cells, therefore it remains unclear if PTTG is sufficient by itself to initiate the tumorigenesis of primary human cells. However, the data clearly demonstrate that PTTG1 overexpression in these cells accelerates their tumorigenic capacity in comparison to that of unmodified cells. PTTG1 contains several-proline rich motifs (PXXP); two of these that are located in the C-terminal domain have been reported to be potential binding sites for SH3-domians [ 44 ]. In our study we confirm that mutation of these C-terminal proline-rich motifs abrogates the tumorigenicity of PTTG1 in human cells. Such loss of tumorigenicity on mutation of PTTG1 could be due to a loss of expression. Our western blot analysis of the stable cell lines (HEKmPTTG1-2 and HEKmPTTG1-4) that constitutively express mutated PTTG1 protein showed high levels of expression of mPTTG1 protein (Fig. 1 ), suggesting that the loss of tumorigenic function of mPTTG1 protein is not due to loss of expression but due to the loss of its ability to induce cellular transformation. These results are consistent with other investigators for rodent cells [ 2 ] and confirm the importance of C-terminal proline-rich motifs to mediate the oncogenic function of PTTG1. The molecular mechanisms by which PTTG1 achieves its tumorigenic function remain unclear. PTTG1 has been reported to induce expression of the c-myc oncogene [ 15 ], bFGF [ 18 ] and VEGF [ 19 ]. bFGF is a broad spectrum and pleiotropic mitogen for growth and differentiation affecting various mammalian cells and organ systems and a large number of cells lines [ 45 , 46 ]; besides stimulating wound healing, tissue repair and hematopoiesis [ 47 ], bFGF induces cell migration and proliferation [ 48 ] and acts as an agiogenic factor that induces migration, proliferation and differentiation of endothelial cells [ 49 ]. In addition, it has been reported to modulate the invasion of tumor cells through surrounding tissue to form new capillary cord structures by regulating the activities of extracellular molecules including collagenase, proteinases and integrins [ 49 ]. Regulation of secretion and expression of bFGF by PTTG1 in NIH 3T3 cells has been shown [ 18 ]. Consistent with these reports, our results demonstrate a significant increase in secretion and expression of bFGF in HEK293 cells on transient transfection with PTTG1 cDNA (Fig. 6 ) as well as in tumors developed by injection of stable cell lines that constitutively express PTTG1 both at protein and mRNAs levels (Figs. 6 , 7 , 8 ). VEGF is a potent stimulant of the vascularization of tumors and is one of the most specific markers of tumor vasculature observed to date [ 50 , 51 ]. VEGF is a multifunctional cytokine acting as a potent permeability agent, an endothelial cell chemotactic agent, an endothelial cell survival factor and an endothelial cell proliferation factor [ 52 ]. The expression and secretion of VEGF has been shown to be a crucial rate-limiting step during tumor progression [ 53 ]. Our results demonstrate a significant increase in secretion and expression of VEGF in HEK293 cells on transfection with PTTG1 and also from tumors excised from animals injected with HEK293 cells that stably express PTTG1 (Figs. 6 , 7 , 8 ). A direct correlation between high IL-8 expression and metastases in melanoma [ 36 ], ovarian cancer [ 37 ], prostrate cancer [ 38 ] and pancreatic cancer [ 51 ] has been reported. To determine if overexpression of PTTG1 induces change in secretion and expression of IL-8, we measured its levels in HEK293 cells on transfection with PTTG1 cDNA and in tumors developed on injection of HEK293 cells transfected with PTTG1. Our results demonstrate for the first time that overexpression of PTTG1 induces IL-8 expression in vitro and also in tumors in vivo (Figs. 6 , 7 , 8 ). Methods Generation of cell lines constitutively expressing PTTG1 The human embryonic kidney cell line (HEK293), which had been transfected by exposing these cells to sheared fragments of adenovirus type 5 DNA [ 28 ] was purchased from ATCC (American Type Culture Collection; Rockville, MD) and cultured according to the instructions provided. The cells were transfected with pcDNA3.1 vector, pcDNA3.1-PTTG1 or pcDNA3.1-mPTTG1 to generate stable clones that constitutively express human wild-type PTTG1 or mutated PTTG1 (mPTTG1) protein as described previously [ 4 ]. The mPTTG1, which carries a double amino acid change within the SH3 binding domain of PTTG1 (P 163 to A 163 , P 170 to A 170 and P 172 to A 172 , and P 173 L 173 ), was generated by site-directed mutagenesis using the Quick-change mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. Mutation of these amino acids has been reported to abrogate the tumorigenic function of PTTG1 and to block the secretion and expression of bFGF in mouse NIH3T3 cells [ 2 ]. The primers used for this site-directed mutagenesis were 5'-GATGCTCTCCGCACTCTGGGAATCCAATCTG-3' and 5'-TTCACAAGTTGAGGGGCGCCCAGCTGAAACAG-3'. The transfected cells were then selected in neomycin G418 (500 μg/ml) and the clones that expressed high levels of PTTG1 protein or mPTTG1 protein were selected. One clone from pcDNA3.1 transfected cells (HEKpcDNA3.1) two clones from pcDNA3.1-PTTG1 transfected cells (HEKPTTG1-1 and HEKPTTG1-3) and two clones from pcDNA3.1-mPTTG1 transfected cells (HEKmPTTG1-2 and HEKmPTTG1-4) were propagated into cell lines. Cell proliferation assay Cell proliferation was assayed using the CellTiter 96 non-radioactive cell proliferation assay kit (Promega, Madison, WI) according to the manufacturer's instructions and as described previously [ 4 ]. Briefly, cells growing in log phase were trypsinized and seeded in 96-well plates (5,000 cells/well in a final volume of 100 μl) in replicates of 4 and incubated at 37°C in 5% CO 2 and 95% air. After incubation for 24 h, 48 h, 72 h or 96 h, 20 μl of dye solution from the kit was added to each well and incubated at 37°C for an additional 2 h. The quantity of formazon product was measured by its absorbance at 490 nm using a 96-well plate reader (Molecular Devices, Sunnyvale, CA). Each experiment was repeated at least three times. Soft agar colony formation (anchorage-independent cell growth) assay Anchorage-independent cell growth was determined by analyzing the formation of colonies in soft agar. Cells (10 4 ) from each cell line were suspended in 0.3% agar in DMEM containing 10% fetal bovine serum and plated on solidified agar (0.7%) in 35 mm dishes. After 14 days of culture, colonies formed were counted and photographed as described previously [ 4 ]. In vivo tumor growth assay Cells growing in log phase were harvested by trypsinization and washed twice with PBS. The cells were resuspended in PBS to a final concentration of 5 × 10 6 /ml. The cells (1 × 10 6 cells in 200 μl PBS/site) were injected subcutaneously (s.c.) into both flanks of 5- to 6-week old female nu/nu mice (4 mice/group) (Charles River Laboratory, Wilmington, MA). All procedures were carried out following the protocol approved by The University of Louisville Institutional Animal Care and Use Committee. Four weeks after injection, the mice were scarified, and the tumors and other tissues harvested. The skin and connective tissues were dissected from the tumors, and the tumor volume was calculated from measurements of length × width × height. The tissues were divided into two parts, one part being fixed in 10% buffered formalin and the other stored in liquid nitrogen. For histopathologic analysis, 5 μm sections were cut from paraffin-embedded tissues, and mounted on slides. Sections were stained with H&E [ 52 ], and processed for histopathologic evaluations. Western blot analysis Cells growing in log phase were lysed in chilled lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% NP-40, 1 mM Na 3 VO 4 , and 1 mM NaF) supplemented with Complete Mini Protease Inhibitor tablets (Roche Molecular Biochemicals, Indianapolis, IN). Equal amounts of protein extracts (40 μg) were resolved on 12% SDS-PAGE gel, and transferred onto a nitrocellulose membrane (Amersham, Piscataway, NJ). Blots were probed with PTTG1 antiserum at a dilution of 1:1,500 as described previously [ 53 ]. Immunoreactive proteins were visualized using the Enhanced Chemiluminescent Detection System (Amersham) according to the instructions provided. ELISA analysis of bFGF, VEGF and IL-8 The levels of bFGF, VEGF and IL-8 in tissue culture supernatants and tissue homogenates were measured using commercially available ELISA kits from BD Biosciences (Minneapolis, MN). To measure bFGF, VEGF and IL-8 in the culture supernatants, HEK293 cells were transiently transfected with pcDNA3.1 or pcDNA3.1-PTTG1 cDNA using Fugene6 as the transfectant reagent as described previously [ 54 ]. After 24 h of transfection, the medium was replaced with serum free DMEM medium. Twenty-four later, the medium was collected and concentrated 5-fold (1.0 ml to 200 μl) using a speedVac system (Savant, Holbrook, NY). To measure bFGF, VEGF and IL-8 in tumor and other tissues, tissues were homogenized in 50 mM Tris (pH 7.4), 0.25% Triton X-100, 5 mM EDTA and 0.1% NP40 supplemented with Complete Mini Protease Inhibitor tablets (Roche Molecular Biochemicals, Indianapolis, IN) using a polytron homogenizer. Homogenates were centrifuged to remove particulate matter and then diluted with the diluent provided in the ELISA kit. The concentration of bFGF, VEGF and IL-8 in a sample was determined by interpolation from a standard curve. All measurements were normalized to protein concentration and performed in triplicate. Semiquantitative reverse transcriptase/polymerase chain reaction (RT/PCR) Total RNA from tumors and other tissues was purified using Trizol reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. The RNA pellets were resuspended in RNase-free water, and the contaminating DNA was removed from the preparations with DNaseI. The yield of total RNA was measured using a spectrophotometer, and the quality was assessed by electrophoresis through a 1% agarose gel. First strand cDNA was synthesized using the iScript™ cDNA synthesis kit (BioRad, Hercules, CA). PCR primers (Table 2 ) were designed, based on the human PTTG1, bFGF, VEGF and IL-8 cDNA sequences. The PCR conditions for each gene are listed in Table 1 . GAPDH amplification was used as an internal control. Ten μl from a total of 50 μl PCR reaction mix was applied to a 2% agarose gel and after electrophoresis; the gel was stained with ethidium bromide to visualize PCR products. The densitometric values for the PCR-amplified products were quantified using BioRad software and normalized against the GAPDH values. Table 2 Primer sequences and PCR conditions for the amplification of PTTG1, bFGF, VEGF, IL-8, m-PTTG1 and GAPDH. Sense Primer Antisense Primer PCR Conditions PTTG ATGGCTACTCTGATCTAT AAAATCTATGTCACAGCAAAC 95°C 5 min, 95°C 30 s, 54°C 30 s, 72°C 30 s. 28 cycles. bFGF TTCTTCCTGCGCATCCACCC CTCTTAGCAGACATTGGAAG 95°C 5 min, 95°C 30 s, 56°C 30 s, 72°C 30 s. 26 cycles. VEGF GAATCATCACGAAGTGGTGA AACGCGAGTCTGTGTTTTTG 95°C 5 min, 95°C 30 s, 56°C 30 s, 72°C 30 s. 28 cycles. IL-8 ACCACCGGAAGGAACCATCT GAATTCTCAGCCCTCTTCAA 95°C 5 min, 95°C 30 s, 58°C 30 s, 72°C 30 s. 28 cycles. GAPDH TGATGACATCAAGAAGGTGGT TCCTTGGAGGCCATGTGGGCC 95°C 5 min, 95°C 30 s, 54°C 30 s, 72°C 30 s. 26 cycles. mPTTG GATGCTCTCCGCACTCTGGGAATCCAATCTG TTCACAAGTTGAGGGGCGCCCAGCTGAAACAG 95°C 5 min, 95°C 30 s, 54°C 30 s, 72°C 30 s. 35 cycles. Conclusion In summary, our results demonstrate that PTTG1 is a potent human oncogene and has the ability to induce cellular transformation of human cells. Overexpression of PTTG1 in HEK293 cells leads to an increase in the secretion and expression of bFGF, VEGF and IL-8. Mutation of C-terminal proline-rich motifs abrogates the oncogenic function of PTTG1. To our knowledge, this is the first study demonstrating the importance of PTTG1 in human tumorigenesis. Authors' contributions TH carried out the in-vivo studies, RT-PCR analysis, western blot analysis and drafted the manuscript. MTM generated stable clones for HEKmPTTG1-2 and 4, RT-PCR and ELISA analysis. SSK participated in study design and coordination and generated stable cell lines for HEKPTTG1-1 and 3. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546418.xml
544555	A Shot in the Arm for AIDS Vaccine Research	Why haven't we developed an HIV vaccine yet? And will the new roadmap from the Global HIV/AIDS Vaccine Enterprise help our efforts towards vaccine development? Ho addresses these crucial questions	The scientific strategic plan of the Global HIV/AIDS Vaccine Enterprise, published in this month's PLoS Medicine , is a clear and cogent document describing how major funders and stakeholders in HIV vaccine development should move forward in a collaborative fashion [ 1 ]. There is no doubt that this roadmap will be regarded as a useful instrument to bring greater cohesion and coordination to the field. The individuals who championed this effort should be commended for providing a great service to the scientific community. It is an excellent start to a continuing dialogue of utmost importance. The Challenge Why is it that we still do not have a protective vaccine against HIV 22 years after its initial identification? Many possible explanations come to mind. In the natural course of HIV infection, the virus wins 99% of the time, showing that specific immunity in an infected person is unable to completely clear the virus. We have also known for over a decade that primary HIV isolates are relatively resistant to antibody neutralization, probably because of a “protective shield” on the viral envelope glycoproteins, consisting of variable loop sequences and extensive N-linked glycosylations. Another explanation is the extreme plasticity of HIV that allows new viral variants to evade immune recognition in the same way that they escape from drugs. Moreover, superinfection by a second viral strain has been documented in a number of individuals who have already mounted immune responses to the initial HIV infection. Yet another problem is that the AIDS research community has yet to uncover the correlates of immune protection in vivo. Lastly, proven vaccine approaches from the past have either failed (whole killed virus and subunit vaccines) or faced seemingly insurmountable regulatory hurdles (live attenuated virus vaccine). Given these daunting obstacles, why have so many continued in the long struggle to develop an HIV vaccine? The answer must lie, in part, in the noble cause at hand. Yet there are also some encouraging clinical and experimental observations ( Figure 1 ). Rare patients do control HIV infection spontaneously. Certain people remain virus-negative despite repeated exposures. That superinfection is not more commonly found supports the notion of immune control. Vaccine-mediated protection against simian immunodeficiency virus is indeed possible using live viruses attenuated by specific mutations or by pharmacological interventions. Finally, and perhaps most importantly, HIV transmission by sex in the natural setting is typically inefficient (and thus easier to block), unlike most experimental challenge systems employed in monkey studies to date. Collectively, these findings provide a ray of hope to push on. Figure 1 Rays of Hope: Clinical and Experimental Observations Suggesting That an HIV Vaccine Is Feasible (Illustration: Giovanni Maki) The Enterprise The scientific strategic plan of the Enterprise is spot-on in identifying the major roadblocks in HIV vaccine development, as well as in establishing the key scientific priorities as we see them today [ 1 ]. It rightly recommends the formation of a growing alliance of organizations to foster a better collaborative spirit that could lead to, among other things, stronger political support and increased funding. The proposed greater coordination and management, sharing of information, technologies, and reagents, and harmonization of standards, assays, and approaches could only add to our overall efforts. One might ask, however, whether there are potential downsides to the plan. In the name of continuing this important dialogue, I would like to offer one general concern. Arguably, the reason for the lack of an effective HIV vaccine today is rooted in the basic problems posed by the virus itself. What we need foremost are new scientific solutions, although a prim and proper “process and structure” in our approach will be helpful. The needed breakthroughs to develop a vaccine will likely emerge from the creativity of scientists doing fundamental research that is free of preconceived biases. It is my contention that great new ideas are as likely to come from curiosity-driven basic studies as from the mission-oriented approach that is represented by the new proposal. Therefore, the leadership of the Enterprise must safeguard against the kind of “group think” that is so pervasive in large collaborative endeavors of this nature. The views of a small number of researchers, no matter how smart or accomplished, must not supersede the collective wisdom of the scientific community at large. No doubt important contributions will be made by scientists working outside of the Enterprise. Measures should be taken to ensure that their views and approaches, even if deemed unconventional, are not stifled by the newly established system. Likewise, their research support should not be compromised because the creation of the Enterprise concentrates the funding into the hands of a relatively small number of designated scientists. To me this is a serious risk given the current “flat funding” at the National Institutes of Health. The Future The authors of the “The Global HIV/AIDS Vaccine Enterprise: Scientific Strategic Plan” have laid out a timely and insightful plan to address perhaps the greatest public-health need of the millennium. This document and its later revisions will serve as useful guideposts for the AIDS vaccine development effort for years to come. To be successful in this mission, our research community will ultimately need a specific “scientific blueprint” for making an HIV vaccine. That day will come only after we get another shot in the arm, infusing us with new knowledge and know-how. Is there any doubt that we need to redouble our investment in basic research on the challenges posed by HIV?	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544555.xml
521483	The Internet and HIV study: design and methods	Background The Internet provides a new meeting ground, especially for gay men, that did not exist in the early 1990s. Several studies have found increased levels of high risk sexual behaviour and sexually transmissible infections (STI) among gay men who seek sex on the Internet, although the underlying processes are not fully understood. Research funded by the UK Medical Research Council (2002–2004) provided the opportunity to consider whether the Internet represents a new sexual risk environment for gay and bisexual men living in London. Methods The objectives of the Internet and HIV study are to: (i) measure the extent to which gay men living in London seek sexual partners on the Internet; (ii) compare the characteristics of London gay men who do and do not seek sex on the Internet; (iii) examine whether sex with Internet-partners is less safe than with other sexual partners; (iv) compare use of the Internet with other venues where men meet sexual partners; (v) establish whether gay men use the Internet to actively seek partners for unprotected anal intercourse; (vi) determine the potential for using the Internet for HIV prevention. These objectives have been explored using quantitative and qualitative research methods in four samples of London gay men recruited and interviewed both online and offline. The four samples were: (i) gay men recruited through Internet chat rooms and profiles; (ii) HIV positive gay men attending an NHS hospital outpatients clinic; (iii) gay men seeking an HIV test in an NHS HIV testing or sexual health clinic; (iv) gay men recruited in the community. Results Quantitative data were collected by means of confidential, anonymous self-administered questionnaires (n>4000) completed on-line by the Internet sample. Qualitative data were collected by means of one-to-one interviews (n = 128) conducted either face-to-face or on-line. Conclusion The strength of the Internet and HIV study is its methodological plurality, drawing on both qualitative and quantitative research among online and offline samples, as well as taking advantage of recent advances in web survey design. The study's findings will help us better understand the role of the Internet in relation to gay men's sexual practice	Background Several studies have found elevated levels of high risk sexual behaviour among people who seek – and meet – sexual partners through the Internet [ 1 - 7 ] In a study conducted in a public HIV testing clinic in Denver, Colorado, for example, people who sought sex on the Internet were more likely to have had a sexually transmitted infection (STI) or report sexual exposure to a person with HIV than those who did not seek sex on the Internet [ 2 ]. They were also more likely to be male, gay and to report anal sex. While the study concluded that gay men were more likely than other participants to use the Internet to seek – and meet – sexual partners, it could not establish whether the excess sexual risk actually occurred with partners whom the men had met through the Internet. A San Francisco study also found that gay men were more likely than heterosexual men and women to use the Internet to meet sexual partners [ 5 ]. Around one-third of the gay men interviewed at a public STI clinic had used the Internet to meet a sexual partner compared with one-in-ten heterosexual men and women. The gay men in the study said that their online partners – men they met through the Internet – were more likely to be casual (ie a one night stand) than their offline partners – men they met elsewhere. Similar reports have also emerged from European studies. An investigation of gay men in London gyms found that in the year 2000 over one-third of those with Internet access had used it to look for sex [ 1 ]. When surveyed three years later, this figure had increased to nearly half [ 8 ]. HIV positive men were more likely to use the Internet to look for sex than HIV negative or never-tested men. Seeking sex on the Internet was associated with a recent STI diagnosis and high-risk sexual behaviour, ie unprotected anal intercourse (UAI) with a person of unknown or discordant HIV status [ 1 ]. This presents a risk for HIV transmission. In addition, HIV-positive Internet-sex seekers were more likely to report UAI with another HIV-positive man than those who did not seek sex in this way. This raises the possibility that HIV positive men use the Internet to meet other positive men for unprotected anal intercourse. While this does not present a risk of HIV transmission to an uninfected person, it may lead to co-infection with an STI or with another, potentially drug-resistant strain of HIV [ 9 ]. As in the U.S. studies, the London study could not establish whether the excess risk for HIV and STD occurred with sexual partners whom the men had actually met through the Internet. The association between seeking sex on the Internet and high risk sexual behaviour raises a number of important, as yet unanswered questions. Does the excess risk for HIV and STI occur with sexual partners whom men actually meet through the Internet? Does the association reflect the fact that high risk men are selectively using the Internet to look for sex? Or does the Internet in some way facilitate high risk behaviour? Is the Internet attracting a constituency of men who have little contact with the established gay scene or health promotion agencies [ 10 ]? For example, men who do not use bars and clubs or bisexual men? In other words, does the Internet represent an emerging sexual risk environment for gay men? If so, what are the underlying processes? Answers to these questions are essential if we are to use the Internet effectively for HIV prevention and sexual health promotion. To address these questions we have undertaken research funded for two years (2002–2004) by the UK Medical Research Council and the Department of Health. This has been conducted by researchers at City University London in collaboration with colleagues at the MRC Social and Public Health Sciences Unit Glasgow and University College London (see appendix 1). Research question The core research question is: Does the Internet represent a new sexual risk environment for gay/bisexual men living in London?" (referred to as "gay men" throughout the rest of this paper). And if so, what are the underlying processes? The research focuses on gay men living in London since the incidence and prevalence of HIV infection among gay men in London is higher than elsewhere in the UK [ 11 - 13 ]. The objectives of the research are to • measure the extent to which London gay men seek sexual partners on the Internet • compare the characteristics of gay men who do and do not seek sex on the Internet • examine whether sex with Internet-partners is less safe than with other sexual partners • compare use of the Internet with other venues such as saunas and backrooms • establish whether gay men use the Internet to actively seek partners for unprotected anal intercourse • determine the potential for using the Internet for HIV prevention Methods These objectives have been explored using both quantitative and qualitative research methods. While quantitative research methods can provide data on a range of outcomes they can rarely offer insight into underlying processes. Qualitative research, on the other hand, illuminates our understanding of human behaviour but is limited in the extent to which findings can be generalized. By adopting methodological plurality and employing both quantitative and qualitative methods, we will be able to build on the strengths of both approaches within an integrated research programme In desribing the design and methods of the Internet and HIV study we focus on sampling strategy, recruitment into the study, data collection and data analysis. Sampling Research among hard-to-reach groups such as gay men is usually based on convenience rather than probability samples [ 14 ]. For example, behavioural research among gay men in the UK has primarily been conducted among men recruited in bars, clubs, GUM clinics [ 15 ], gay pride events [ 16 ] and gyms [ 17 ]. While probability sampling would undoubtedly provide a more robust foundation for statistical analysis [ 18 ] such an approach is extremely, if not prohibitively expensive. Convenience samples have the advantage of being affordable and also provide the opportunity to focus on men with characteristics which may be of particular interest, eg men who report high risk behaviour [ 14 ]. The disadvantage of course is that such samples may introduce selection bias. This bias can be partially overcome by including samples from more than one source allowing for triangulation of data. Consequently, we recruited gay men from four different sources – one online, three offline. Each sample has specific features which are pertinent to the research question. HIV positive gay men HIV positive men are central to any research concerning HIV risk, transmission and prevention [ 19 - 21 ]. Furthermore, HIV positive men are more likely to use the Internet to seek sexual partners than other men [ 1 , 8 ]. Consequently it was decided to over-sample HIV positive men for the study to permit an in-depth examination of sexual risk behaviour and use of the Internet in this group of men. In the UK everyone who is diagnosed HIV positive is offered free treatment and care within the National Health Service (NHS) through a hospital outpatients clinic. Consequently an NHS clinic sample will be broadly representative of all those living with diagnosed HIV. Gay men seeking an HIV test Most people seeking an HIV test have been at risk of HIV infection. Furthermore, HIV negative gay men with a history of multiple repeat testing report elevated levels of high risk sexual behaviour [ 22 ]. This group therefore merits inclusion in an investigation of the Internet as an emerging sexual risk environment. In the UK, the NHS offers free voluntary counselling and testing for HIV either in dedicated HIV testing clinics or in general sexual health clinics. Since there is no charge for this service, these clinics attract a broad cross section of people. Gay men in the community Surveying gay men in the community allows us to examine the extent to which they use the Internet for seeking sexual partners and the associated risks. Previous research has shown that gay men surveyed in central London gyms are broadly representative of men "on the scene" in London ie men who go to gay bars, clubs and other venues [ 23 ]. However, whereas questionnaires distributed in London bars and clubs have to be short because of the limited time available for completion, we have found that in gym-based surveys respondents are willing to complete questionnaires that take up to 15 minutes to answer. This allows for a detailed investigation of sexual behaviour. The Internet Men who use Internet chat rooms and profiles to seek sex with other men are clearly of central importance to this research project. The Internet may attract a constituency of men who would not otherwise be included in behavioural surveys among gay men ie men who do not go to gay bars, clubs or other venues and men who have not been tested for HIV [ 10 ]. Comparing the characteristics of men recruited online with the community and clinic samples will throw these differences, where they exist, into sharp focus. Recruitment Based on previous research conducted among gay men in both clinic and community settings [ 1 , 8 , 22 , 23 ] we estimated that, for the quantitative arm of the study, we would needed to recruit 400–500 men in each of the four samples. This would provide sufficient power at a 5% level of significance to compare the characteristics of men who do and do not seek sex on the Internet, to compare use of the Internet with other venues such as saunas and backrooms, and to examine whether sex with Internet-partners is less safe than with other partners. For the qualitative arm of the study, we recruited at least 20 men from each of the four samples to allow us to derive accounts of Internet dating and sexual practice from a diverse range of men, selected purposively according to age, education, employment, HIV status and use of the Internet for seeking sex. HIV positive gay men Men diagnosed with HIV infection attending an outpatient treatments clinic at the Royal Free Hampstead NHS Trust hospital, London over an 8 month period (October 2002–May 2003) were invited to participate in the research. Patients with a limited command of English were ineligible for the study as were those who were too ill to complete a questionnaire. Eligible patients were approached in the clinic's waiting area by a trained member of the research team who discussed the project with them. Patients were provided with written information about the research, contact details of the research team as well as helpline numbers. Once they had provided written consent, respondents were asked to complete a pen-and-paper questionnaire in the clinic and return it in a sealed envelope to the team member (further information about the questionnaire in Research Methods below). Some patients were only in the waiting area for a short time so there wasn't an opportunity to invite them to take part in the study. Over the eight month survey period (October 2002–May 2003), 1001 individual male patients attended the clinic of whom 939 were deemed eligible for the study. Of those who were eligible, 864 were asked to complete a questionnaire and 620 did so. The response rate was 72% of men who were offered a questionnaire and 66% of all eligible men who attended the clinic. Of the 620 men who completed a questionnaire, 542 described their sexual orientation as gay or bisexual or had had sex with another man in the previous year. Of these, 523 men provided sufficient information to be included in the quantitative sample (table 1 ). Table 1 Number of London gay/bisexual men who participated in the Internet and HIV study 2002–2003 Quantitative arm Qualitative arm Recruitment site HIV positive HIV negative Never-tested Total HIV positive HIV negative Never-tested Total HIV treatment clinic 523 - - 523 20 - - 20 HIV testing & sexual health clinics* 15 389 - 404 - 20 - 20 Community 2002 138 592 184 914 } 1 18 4 23 Community 2003 88 361 94 543 } Internet* 2002 142 680 396 1218 } 17 35 13 65 Internet*** 2003 67 315 197 579 } * An additional 435 heterosexual men and 450 heterosexual women completed a questionnaire in the HIV testing clinic for the quantitative arm * * Fifteen men received an HIV positive diagnosis when they returned for their test result * * * An additional 3279 gay/bisexual men living in the UK but outside London completed the questionnaire online in 2002, 1944 in 2003 for the quantitative arm HIV positive gay men who completed a questionnaire were asked on the last page if they would be willing to have an in-depth face-to-face interview, one-to-one, with a qualitative researcher working on the project (MD) (further information about the one-to-one interviews in Research Methods below). If they agreed, the researcher contacted them to arrange a time for the interview. This could be in the researcher's office at City University London, in the hospital or at the respondent's home. In this way 20 HIV positive gay men were recruited for one-to-one interviews as part of the qualitative arm of the study (table 1 ). Gay men seeking an HIV test People seeking an HIV test at the same-day HIV testing clinic, Royal Free Hampstead NHS Trust Hospital over a 13 month period (October 2002–November 2003) were invited to take part in the study. Based on previous research in the HIV testing clinic, we estimated that this would generate a sample of approximately 500 gay men, 500 heterosexual men and 500 heterosexual women [ 22 ]. Although the focus of the study was gay men, data were also collected from heterosexual men and women to provide a valuable comparison. Using a similar strategy to the outpatient HIV treatments clinic, everyone seeking an HIV test was asked to participate in the research. Those who agreed completed a detailed, self-administered pen-and-paper questionnaire, after providing written consent, while they were waiting for their pre-test counselling. People with limited command of English were deemed ineligible for the study as were those who were too ill, too young (under 18 years) or too anxious to complete a questionnaire. Those attending the testing clinic on more than one occasion during the survey period were asked to only complete the questionnaire once. Over the 13 month survey period (October 2002–November 2003), 1889 individuals came to the Royal Free clinic for an HIV test of whom 1753 were eligible for the study. Of those who were eligible, 1640 were asked to complete a questionnaire and 1230 did so. The response rate was 75% of people who were offered a questionnaire and 70% of all eligible persons who attended the clinic. Of the 1230 people who completed a questionnaire, 345 described their sexual orientation as gay or bisexual, 435 as heterosexual male and 450 as heterosexual female; 334 gay/bisexual men provided sufficient information to be included in the quantitative sample. Gay men who completed a questionnaire were asked if they would be willing to have an in-depth, face-to-face interview with the qualitative researcher (MD). If they agreed, the researcher contacted them to arrange a time for the interview as described above. In this way, 16 gay men seeking an HIV test at the Royal Free were recruited for one-to-one interviews for the qualitative arm. Heterosexual men and women who completed the questionnaire were not asked to have a one-to-one interview. Previous research suggested that the majority of clinic attenders would test HIV negative while approximately 6% of gay men were expected to test HIV positive. It was considered inappropriate to ask men who had just been diagnosed HIV positive if they would be willing to have a face-to-face interview. Consequently, the qualitative sample from the HIV testing clinic solely comprised men who had tested negative. Anyone who volunteered for an interview who subsequently received a positive test result was not included in the qualitative sample. After 6 months it became apparent that while the response rate at the Royal Free HIV testing clinic was high the number of gay men seeking a test was not as large as had been expected. We decided therefore to extend recruitment to men seeking an HIV test at a sexual health clinic specifically for gay men at Barts and the London NHS Trust hospital, London. All men attending this clinic are gay. Using a similar strategy to the one developed at the Royal Free, everyone attending the gay men's sexual health clinic at Barts and the London hospital between June and November 2003 was invited to participate in the research. Those who agreed completed a self administered questionnaire while they were waiting in the clinic for their appointment. Although all men were asked to complete a questionnaire, only men seeking an HIV test were eligible for inclusion in the Internet and HIV quantitative sample. Over the 6 month survey period at the Barts and the London sexual health clinic (June – November 2003), 211 gay men came for an appointment of whom 209 were eligible for the study. Of those who were eligible, 198 were asked to complete a questionnaire and 156 did so. The response rate was 79% of men who were offered a questionnaire and 75% of all eligible men attending the clinic. Of the 156 gay men who completed a questionnaire at Barts and the London, 70 were seeking an HIV test and provided sufficient information to included in the quantitative sample for the study. These men were also asked if they were willing to have a one-to-one interview with the qualitative researcher. Four men were recruited in this way for the qualitative arm of the study. Combining the data from the two clinics, 404 gay men seeking an HIV test were recruited for the quantitative arm of the study between October 2002–November 2003, of whom 20 agreed to have a one-to-one interview for the qualitative arm (table 1 ). In addition 450 heterosexual women and 435 heterosexual men also completed a questionnaire for the quantitative arm. Gay men in the community Previous research has shown that gyms in central London provide a suitable environment for undertaking detailed behavioural research among gay men at risk of HIV infection [ 1 , 8 , 17 , 24 , 25 ]. In both 2002 and 2003, all men using any one of 7 central London gyms during a one-week period between January-March were invited to take part in the study. All these gyms have a substantial gay male membership. One gym was exclusively gay whereas the others estimated that gay men comprised 40–90% of their male membership [ 26 ]. All men using the gyms during the survey period were asked to complete a self-administered pen-and-paper questionnaire after providing written consent. A filter question on sexual orientation distinguished gay or bisexual men from straight men. Only gay/bisexual men were requested to answer questions on the Internet and sex. Men could complete the questionnaire in the gym or at home. Respondents returned completed questionnaires to collection boxes in the gym or by post to the research team. In the mixed gyms the number of questionnaires handed out to gay men was estimated by multiplying the total number of questionnaires distributed in the gym by the proportion of male members who were gay (according to the managers' estimates). To calculate the response rates we divided the number of questionnaires returned by gay men (as indicated on the questionnaire) by the estimated number of questionnaires handed out to gay men, as described above. In 2002, 921 gay men completed the questionnaire while in 2003, 550 men did so. Those men who provided information on their HIV status (2002, n = 914; 2003, n = 543) were included in the quantitative sample (table 1 ). The estimated response rate each year was 50%–60%. In both years, gay men who completed a questionnaire were asked if they would be willing to have an in-depth, face-to-face interview with the qualitative researcher (MD). If they agreed, the researcher contacted them to arrange a time for the interview as described above. Some interviewees were recruited through snowballing. In this way, 23 men were recruited from the gyms for the qualitative arm of the study (table 1 ). The Internet In both 2002 and 2003, men using UK chat rooms or personal profiles on gaydar ( )or gay.com ( ) were invited to take part in the study. Gaydar and gay.com are two of the UK's most popular websites for gay men (personal communication H Badenhorst, M Watson). Over a four week period in May-June each year, a series of pop-ups and banners advertised the research project in UK chatrooms and on profiles pages. Clicking on a popup or banner took men to the homepage of the online questionnaire. Men who agreed to complete the questionnaire then did so after providing informed consent – all online. For technical reasons, it was not possible to restrict banner advertising or pop-ups to London chatrooms or personal profiles alone. Instead, the advertising was restricted to UK chatrooms or profiles. Consequently, anyone entering a UK chatroom or profile during the survey period had the opportunity of completing the online questionnaire even though the target group was London gay men. In 2002, 1250 London men completed the online questionnaire while in 2003, 595 did so. Those men who provided information on their HIV status were included in the quantitative sample (2002, n = 1218; 2003, n = 579) (table 1 ). A further 3279 men living in the UK but outside London completed the questionnaire in 2002; 1944 in 2003. The decline in the number of respondents in 2003 compared with 2002 reflects a general pattern seen by gaydar and gay.com in other online surveys in the UK (personal communication H Badenhurst, M Watson). Estimating a response rate for the online quantitative sample is problematic [ 27 , 28 ]. It is impossible to gauge what proportion of chatroom and profile users saw the banners and pop-ups advertising the online survey. Nor do we know what percentage of those seeing the pop-ups and banners went on to complete the questionnaire. Based on estimates provided by and on the number of people using their Internet chatrooms and profiles during the survey periods, it is likely that less than one percent of all users completed the questionnaire. This level of response is standard for online surveys. This highlights the importance of not relying solely on respondents recruited through the Internet for research of this kind. Gay men living in London who completed the online questionnaire were asked if they would be willing to have a one-to-one interview with the qualitative researcher (MD). Those who agreed were asked to send an email to the researcher who then contacted them, also by email, to provide further information about the study and to arrange a time for the interview as described above. Interviews were either conducted online or face-to-face (see Research methods below). Of the London men who completed a questionnaire online in 2002 or 2003, 65 went on to have a one-to-one interview as part of the qualitative arm (table 1 ); 30 men were interviewed face-to-face while 35 were interviewed online Data collection Quantitative data The questionnaires sought detailed information on the men's socio-demographic characteristics (age, ethnicity, employment, education), sexual orientation, HIV test history (date and result of last test), history of STIs, access to and use of the Internet, seeking sex on the Internet, use of other venues (eg back rooms, saunas), as well as sexual risk behaviour in the previous 3 months, differentiating between Internet- and other sex partners. Unprotected anal intercourse (UAI) in the previous 3 months was categorized according to type (regular or casual) and HIV status of partner. Detailed information was collected about sexual behaviour with partners met through the Internet and with partners met elsewhere in order to compare the level and nature of risk with Internet and other sexual partners. Data were also collected on potential confounding factors such as recreational drug use, alcohol consumption, relationships, attitudes towards new treatments for HIV and mental health. Standard and validated questionnaire items were used extensively (copies of the questionnaires are available from JE). Core questions, included in the questionnaires for all four samples, were worded identically to ensure direct comparability between the different groups. In addition, some questions specific to each group were included eg, detailed questions on HIV test history and reasons for testing for people seeking an HIV test; questions on HIV medication, CD4 and viral load for HV positive men. The questionnaires were piloted both online and offline among gay men at the developmental stage of the study and revised in the light of any feedback and comments. People recruited offline, in clinics or gyms, completed a pen-and-paper questionnaire. All questionnaires were confidential and anonymous. They contained no information that would allow an individual respondent to be identified. For the HIV positive men, information on their most recent viral load and CD4 count was abstracted from hospital records and linked to their questionnaire without breaching confidentiality. The HIV test result of those seeking an HIV test was linked to their questionnaire, again without breaching confidentiality. Once the questionnaire and clinic data had been linked in the database individual identifiers (eg hospital numbers) were removed and destroyed to ensure anonymity. Men recruited through Internet chatrooms and profiles completed the questionnaire online. The online questionnaire was constructed by a computer programmer at gaydar ( ) working in close collaboration with the research team. Because of their technical expertise and capacity, gaydar hosted the questionnaire on their server. All questionnaires were confidential and anonymous. Identifiers such as IP addresses were removed from the questionnaires completed online before the data were downloaded to a database. Within each sample men were asked (a) to complete only one questionnaire and (b) whether they had been in any of the other samples. For example, men who completed the questionnaire online were asked whether they had also completed a questionnaire in the gyms or clinics, etc. Qualitative data Qualitative data were collected by means of one-to-one, in-depth interviews conducted face-to-face (n = 93) or on-line (n = 35) Face-to-face interviews The interviewees were volunteers from the quantitative samples recruited in the HIV treatments clinic, HIV testing clinic, gyms or online as described above. They provided written consent for the one-to-one interview which lasted between 50 and 90 minutes and was audio-taped for transcription. The interviews were generally conducted in the research office at City University London but occasionally at the interviewee's home or in the clinic. Confidentiality was provided in two ways; first, the interviewee's contact details were not linked to the interview transcript and secondly personal identifiers such as the person's name, where he was born, lived or worked were removed from the transcripts. Once all the data had been transcribed and entered into the database, individual identifiers (eg interviewees' contact details) were destroyed. Online interviews The interviewees were volunteers from the online quantitative samples. They provided informed consent by email. Interviews were prearranged and conducted in a private room on or which only the interviewer and interviewee could enter. The interviews were synchronous and conducted entirely through text generated as online chat. Each interview lasted between 50 and 90 minutes at the end of which the interview text was copied and pasted into a Word document. Confidentiality was provided in the same manner as the face-to-face interviews.. Interviews were conducted in two phases. Phase one (n = 24 interviews) focused on how gay men used the Internet for sexual partnering while phase two (n = 104) focused on risk behaviour related to the Internet. The interviews for phase one were based on a topic guide comprising questions about age, residence, schooling, employment, relationship status and HIV testing. Social and sexual lifestyles were explored in depth paying particular attention to the role of the Internet. The topic guide included seeking sexual partners, preferences for different ways of meeting partners (including the Internet), Internet experience and skills, learning how to use the Internet for sexual partnering, online communication skills, other uses of the Internet and related media, a recent sexual episode and its relationship with the Internet. The topic guide for phase two comprised social background; sexual lifestyle and role of the Internet; risk episodes including a description of anal sex with a condom and without a condom with Internet and non-Internet partners; HIV testing (where relevant); general discussion of Internet experience and web profiles; safer sex (rules, negotiated safety, serostatus of sexual partner, disclosure, slip-ups, problems, pleasure and rationality); and sources of knowledge and skills about risk reduction (health carers, media, school, common sense). For HIV positive men the topic guide also covered their experience of HIV treatment and care (eg current diagnosis, treatments, treatment effectiveness, clinical markers) and the relationship between treatments and risk (eg prospects for new drugs, role of viral load in risk taking, reinfection); Ethics The research protocol was approved by the following ethics committees: Royal Free Hospital and Medical School Local Research Ethics Committee, the East London and The City Research Ethics Committee and City University London Research Ethics Committee. Sample characteristics and data analysis Over 4000 London gay men were recruited for the quantitative arm of the study in 2002 and 2003 from the HIV treatments and testing clinics, gyms and through the Internet. Of these, a subset of 128 men were interviewed one-to-one for the qualitative arm (table 1 ) Quantitative data Data from the pen-and-paper questionnaires were coded, entered into a database and verified. Data collected online were downloaded directly into a database. The background characteristics of the different samples in the quantitative arm are presented in table 2 . Table 2 Background characteristics of the men in the quantitative sample, by recruitment site HIV treatment clinic HIV testing clinics Community Internet 2002–2003 (n = 523) 2002–2003 (n = 404) 2002 (n = 914) 2003 (n = 543) 2002 (n = 1218) 2003 (n = 579) n % n % n % n % n % n % Age (median; range) 38 23–70 32 17–73 35 17–79 36 18–75 33 18–70 32 18–75 Ethnicity (white) 467 89.6 342 84.7 821 90.4 484 89.6 1117 91.7 526 89.3 Employed 324 62.8 317 91.4 771 85.3 450 83.2 1007 82.7 477 80.2 Higher education 328 65.7 261 72.9 761 83.8 425 79.7 810 66.5 377 63.9 Sexual orientation "gay" 489 93.5 354 87.6 869 95.1 523 96.3 1084 89.0 524 88.1 In a relationship with a man 285 55.3 223 55.9 469 51.6 300 55.4 531 43.6 254 42.7 HIV positive 523 100.0 15 3.7 138 15.1 88 16.2 142 11.7 67 11.3 Treatments optimism 1 201 40.9 92 26.1 175 20.3 120 23.3 164 14.5 79 14.2 Treatments optimism 2 141 28.7 85 24.6 176 20.7 117 22.9 158 15.5 94 17.0 Uses recreational drugs 277 60.6 200 60.4 467 53.2 334 64.4 497 40.8 212 38.4 Felt depressed 274 55.2 85 49.6 389 44.5 253 48.6 552 45.3 274 48.7 Had suicidal thoughts 96 21.9 46 14.3 104 12.8 65 12.8 210 17.3 105 18.7 Has access to the Internet 443 86.3 367 90.8 841 92.6 499 93.1 1193 97.9 588 98.8 Uses the Internet to seek sex 223 43.6 186 46.0 400 44.4 280 52.0 1040 85.4 544 91.4 Has sex with men only 453 96.2 352 87.6 853 93.5 509 93.7 1089 89.4 537 90.3 STI in previous 12 months 13 27.5 79 19.8 203 22.5 127 23.5 295 24.2 128 21.7 Non-concordant UAI 116 22.2 141 34.9 199 22.1 119 22.1 391 32.1 196 32.9 Concordant UAI only 58 11.1 36 8.9 135 15.0 80 14.8 175 14.4 87 14.6 Completed clinic survey 1 - - - - - - 16 3.0 - - 26 4.7 Completed gym survey 2 25 5.0 14 3.9 - - - - 24 2.0 16 2.9 Completed 2002 online survey 3 27 5.4 26 7.3 - - 54 10.2 - - - - Treatments optimism 1: Men who agreed with the statement "I am less worried about HIV now that treatments have improved" Treatments optimism 2: Men who agreed with the statement "I believe new treatments make people with HIV less infectious" Non-concordant UAI: unprotected anal intercourse in the previous 3 months with someone of unknown or discordant status Concordant UAI only: unprotected anal intercourse in the previous 3 months only with someone of the same HIV status. as a percentage of sexually active men (some men had not had sex in the previous year) 1 Number (%) of men completing a 2003 questionnaire in the gyms or online who said they had also completed a questionnaire in the HIV testing clinic or treatments clinic 2 Number (%) of men completing a questionnaire online or in the clinics who said they'd also completed a questionnaire in a gym 3 Number (%) of men completing a questionnaire in the gyms (2003) or in the clinics who said they'd also completed the 2002 online questionnaire In all samples, the majority of men were relatively young, white and identified as gay. There were however, differences between the samples in these and other characteristics. For example, men who completed a questionnaire online were less likely to describe themselves as gay than men who completed a questionnaire in a gym (88% v 96% in 2003). On the other hand, men surveyed online were more likely to have used the Internet to look for sex (approximately 90%) than men who surveyed in the clinics or gyms (40–50%). The differences between the samples will be explored further in future analyses and subject to formal statistical testing. For the most part the four samples were independent; only a minority of men said they had completed a questionnaire in more than one recruitment site (eg in a clinic and in a gym); range 2–10%. Data analysis, using standard statistical packages, will allow us to examine; whether sex with Internet partners is of higher risk than with other men (within-person analysis) and the extent to which Internet sex seekers were specifically looking for unprotected anal intercourse either online or offline. The characteristics of those who have and have not used the Internet to find a sexual partner will be compared to explore whether selected groups of men (eg bisexual or those at high risk) seek sex on the Internet. The importance of the Internet in relation to other venues will be examined. The samples, ranging in size from 404 to 1218, are sufficiently large to allow us to detect statistically significant differences between and within the different groups at the 5% level of significance. All analyses will be conducted for HIV positive, negative and never-tested men separately and comparisons will be made within and between the four samples. Comparison of those recruited through the Internet with the community and clinic samples will indicate whether the Internet attracts men who would otherwise be hard to reach for health promotion and HIV prevention. The data collected online from men living in the UK but outside London will provide opportunities for further analysis. For example their characteristics and behaviours can be compared with those of the London men surveyed online as well as with other men living outside London surveyed in the community (eg in Scotland). The data collected online will also allow us to examine patterns of Internet use in a sample.that covers the whole of the UK. A PhD studentship, funded separately by the UK Economic and Social Research Council (ESRC), will allow us to examine methodological issues around using the Internet for data collection and research [ 29 ], such as mode effect, motivation for participating in online surveys, number of fields completed online and offline, drop outs and non-probability sampling (see appendix 2). The ESRC-funded PhD will utilize all the data collected from men living in the UK who completed an online questionnaire in 2002 or 2003. In addition, data collected in the HIV testing clinic will also allow us to explore Internet sex-seeking and risk behaviours among heterosexual men and women. Qualitative data The interview transcripts from both the face-to-face and online interviews were coded and analysed using Nvivo. The background characteristics of the men interviewed for the qualitative arm of the study are summarized in table 3 . Table 3 Background characteristics of the men in the qualitative sample, by recruitment site HIV treatment clinic HIV testing clinics Community Internet Interviewed....... Face-to-face (n = 20) Face-to-face (n = 20) Face-to-face (n = 23) Online (n = 35) Face-to-face (n = 30) n n n n n Age (median; range) 38 31–59 40 25–66 35 24–51 32 20–63 39 21–60 Employed 9 18 18 30 23 Higher education 15 15 17 20 20 Has regular sexual partner 8 8 12 14 12 HIV status HIV positive 20 - 1 6 11 HIV negative - 20 18 20 15 Never-tested - - 4 9 4 Seeks sex through the Internet Rarely, never or used to 13 14 11 6 0 Seeks sex both on and offline 3 4 5 10 14 Seeks sex mostly through the Internet 4 2 7 19 16 Interviewees were recruited for qualitative interviews once they had completed a behavioural questionnaire for the quantitative arm of the study. Additional purposive criteria were adopted to ensure the sample included men from a range of age groups; of different HIV status (HIV positive, HIV negative and never tested); of varying educational attainment; both employed and unemployed; and who reported differential use of the Internet for sexual purposes. Online, as well as face-to-face (FTF) interviews were conducted to encourage the participation of gay men who used the Internet to look for sex. This combination of matching the quantitative arm in terms of recruitment sites, together with additional purposive criteria and two interviewing methods resulted in 128 qualitative interviews with a diverse range of gay/bisexual men. All interviewees lived in the greater London area. For the qualitative arm of the study, we addressed analytic bias using guidance provided by Barbour [ 30 ], Kvale [ 31 ] and Popay & Williams [ 32 ] for whom methodological rigour is best addressed through the process of qualitative research. For example, the internal logic of the study; emphasis on the quality of the craft of research together with communicative and pragmatic forms of validation [ 31 ]; "privileging subjective meaning" [ 32 ]. In our study we addressed quality in three main ways: • Iteration • Team analysis • Transparency Iteration We conducted a mapping phase to provide an empirical context for our thinking about the Internet-related sexuality of gay men and to orient further research in terms of the sampling and topic guides. We also adopted the practice of constant comparison derived from grounded theory as another form of iteration. Team analysis The research team was involved in a cycle of reflection on the data as they were generated and analysed. In the first meeting the team reviewed a transcript to identify possible themes for analysis. Notes and the transcription of the meeting contributed to the formulation of themes. At an intermediate stage the team was asked to apply the thematic framework to other transcripts. This helped to assess the utility of the framework and elaborate possible new themes. A second meeting was then held where the team reviewed the coding scheme for the qualitative research and its application to the entire data set. This process was conducted twice – first for data generated during phase 1, then for data collected in phase 2. Transparency We documented how the data were generated, catalogued and analysed to make it open to observation. The one-to-one interviews will be analysed for recurring and contradictory themes relating to the core research questions. Phase one will help orientate the qualitative research around the Internet and risk. Phase two will provide information about the range of sexual experiences linked to the Internet as well as how people communicate about sex and risk through the Internet. The qualitative interviews will allow for greater interrogation than is permitted by self-completed questionnaire alone. A number of issues will be examined including personal experiences of using the Internet for sex, partner selection, the context in which unsafe sex occurs, the emergence of social and sexual networks via the Internet and the value or meaning placed on unsafe sex with casual partners met in different settings including the Internet. Discussion Does the Internet represent a new sexual risk environment for gay men? To answer this question, we have employed a range of quantitative and qualitative research methods in online and offline samples of London gay men. The data we collect will allow us to explore in depth the association between seeking sex on the Internet and high risk behaviour and also consider the underlying processes. One of the strengths of the Internet and HIV study is its methodological plurality. Detailed behavioural data have been collected in the quantitative arm of the study from a large number of men recruited in community and clinic settings as well as through the Internet. Questionnaires were completed online by the Internet sample. On the other hand, qualitative interviews will allow for a greater understanding of the context in which unsafe sex occurs and the processes that underlie the behavioural patterns seen in the quantitative analysis. By using both quantitative and qualitative methods, we hope to garner the best that both approaches can offer within an integrated research programme Sampling men online has allowed us to develop innovative research methods for both the quantitative and qualitative arms of the study by taking advantage of recent advances in web-based data collection. The online questionnaire provided opportunities for innovation with respect to its design, format of the questions and data entry. Conducting one-to-one interviews online has opened up new ways of undertaking qualitative research. Ours is one of a small number of studies which have explored this approach to qualitative interviewing [ 33 ]. In addition, the ESRC-funded PhD examining the Internet and research methodology will provide an opportunity to consider the advantages and disadvantages of conducting research online [ 29 ]. The Internet undoubtedly offers enormous potential for HIV prevention and sexual health promotion [ 34 , 35 ]. Our research will reveal whether the Internet reaches a group of men who have little contact with the established gay scene or health promotion agencies. If this is the case, the Internet could provide access to an otherwise hard-to-reach group of men. Exciting as these new opportunities are, however, we still know relatively little about the efficacy of online sexual health promotion and HIV prevention [ 36 ]. Data generated by both the quantitative and qualitative arms of our study will provide a better understanding of the social and sexual networks created through the Internet. Who is using the Internet to look for sex and how are they using it? In this way the potential for using the Internet for sexual health promotion and HIV prevention can be established Competing interests None declared. Authors' contribution JE, GH and LS conceived the study; all authors participated in its design; JE was responsible for overall project management; GB was responsible for conducting the quantitative arm of study; MD was responsible for the qualitative arm; JE drafted the manuscript with input from GB and MD. All authors read, revised and approved the final manuscript. Appendix 1 The Internet and HIV study: MRC-funded research City University London, Institute of Health Sciences, St Bartholomew School of Nursing and Midwifery Professor Jonathan Elford, principal investigator Graham Bolding, research fellow (quantitative arm) Mark Davis, research fellow (qualitative arm) MRC Social and Public Health Sciences Research Unit, Glasgow Professor Graham Hart, co-investigator Royal Free and University College Medical School, London Professor Lorraine Sherr, co-investigator Collaborators , , Royal Free Hampstead NHS Trust Hospital London, Barts and the London NHS Trust Hospital, Camden Primary Care Trust, London, Central London gyms. The research was funded from 2002–2004 by the UK Medical Research Council and the Department of Health as part of its AIDS Epidemiological Research Programme (G0100159. The Internet and HIV: an examination of high risk sexual behaviour among London gay men who seek sex on the Internet) Further information about the project can be found on our website at Appendix 2 The Internet and research methodology: ESRC-funded PhD studentship City University London, School of Social Sciences and Institute of Health Sciences, St Bartholomew School of Nursing and Midwifery Alison Evans PhD candidate School of Social Sciences and Institute of Health Sciences, St Bartholomew School of Nursing and Midwifery Professor Dick Wiggins first supervisor, School of Social Sciences, Department of Sociology Professor Jonathan Elford second supervisor, Institute of Health Sciences, St Bartholomew School of Nursing and Midwifery Advisory Board City University London Graham Bolding, Mark Davis, Professor Jonathan Elford, Professor Roger Jowell, Professor Dick Wiggins MRC Social and Public Health Sciences Research Unit, Glasgow Professor Graham Hart Royal Free and University College Medical School, London Professor Lorraine Sherr Further information can be found at Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521483.xml
553997	Preventing complicated transseptal puncture with intracardiac echocardiography: case report	Background Recently, intracardiac echocardiography emerged as a useful tool in the electrophysiology laboratories for guiding transseptal left heart catheterizations, for avoiding thromboembolic and mechanical complications and assessing the ablation lesions characteristics. Although the value of ICE is well known, it is not a universal tool for achieving uncomplicated access to the left atrium. We present a case in which ICE led to interruption of a transseptal procedure because several risk factors for mechanical complications were revealed. Case presentation A case of a patient with paroxysmal atrial fibrillation and atrial flutter, and distorted intracardiac anatomy is presented. Intracardiac echocardiography showed a small oval fossa abouting to an enlarged aorta anteriorly. A very small distance from the interatrial septum to the left atrial free wall was seen. The latter two conditions were predisposing to a complicated transseptal puncture. According to fluoroscopy the transseptal needle had a correct position, but the intracardiac echo image showed that it was actually pointing towards the aortic root and most importantly, that it was virtually impossible to stabilize it in the fossa itself. Based on intracardiac echo findings a decision was made to limit the procedure only to ablation of the cavotricuspid isthmus and not to proceed further so as to avoid complications. Conclusion This case report illustrates the usefulness of the intracardiac echocardiography in preventing serious or even fatal complications in transseptal procedures when the cardiac anatomy is unusual or distorted. It also helps to understand the possible mechanisms of mechanical complications in cases where fluoroscopic images are apparently normal.	Background Since the advent of ICE in the electrophysiology practice it proved its value in guiding transseptal procedures with providing an extra safety margin for the patients. The possibility to visualize the oval fossa, the LA free wall and the aortic root helps in preventing mechanical complications. ICE can visualize also intracardiac thrombus and spontaneous echocontrast, which is helpful in avoiding thromboembolic complications. Although the value of ICE is well known, it is rather hard to admit that it is a universal tool for achieving uncomplicated access to the left atrium. The aim of this case presentation is to show that ICE can lead to interruption of a transseptal procedure due to the presence of risk factors for mechanical complications when the fluoroscopic image is seemingly satisfying. Case presentation A seventy-year-old male patient with paroxysmal atrial fibrillation and atrial flutter, and concomitant arterial hypertension was referred to our institution for LA circumferential ablation. The preprocedural TEE described an aneurysm of the IAS (See Additional file 1 : TEE.mpeg for the TEE), however the left heart catheterization was deemed feasible. ICE performed at the electrophysiology laboratory (catheter Ultra ICE 9900, 9 MHz, EP Technologies, Boston Scientific Corp., San Jose, CA, USA connected to a console ClearView Ultra, Boston Scientific Corp., San Jose, CA, USA) as a routine part of the transseptal puncture showed more detailed picture of the intracardiac anatomy – dilated aortic root, enlarged right atrium displacing the IAS towards the LA free wall, and small oval fossa with flapping motion of the fossa ovalis membrane. The distance between the IAS and the LA free wall during the atrial diastole was small (Fig. 1 ). During the atrial contraction the LA cavity appeared almost obliterated and the fossa ovalis membrane virtually touched the LA free wall, thus making the transseptal puncture potentially dangerous (Fig. 2 ) (See Additional file 2 : ICE1.mpeg for the unusual intracardiac anatomy). The fossa itself was abouting anteriorly to the non-coronary aortic sinus of Valsalva and although the fluoroscopic position of the transseptal needle was apparently correct, ICE showed that it was always sliding to the junction between the IAS and the aorta and that the needle tip was pointing towards the aortic root. Several unsuccessful attempts were made to change its direction towards the oval fossa. That was also deemed a condition predisposing to complicated transseptal puncture (Fig. 3 ) (See Additional file 3 : ICE2.mpeg for the sliding of the transseptal needle towards the aorta). During the procedure two echocardiographers, including the one, which had performed the TEE, re-evaluated the videotaped TEE and agreement was achieved that the initial interpretation of the TEE needed correction. Although the patient was symptomatic and the arrhythmia episodes could not be suppressed effectively with antiarrhythmic drugs, after considering all pros and cons we decided not to perform the left atrial procedure and thus to avoid serious and potentially fatal complications. Only a cavotricuspid isthmus ablation was done. Figure 1 ICE during atrial diastole before the jump of the transseptal needle into the oval fossa. In this patient the LA cavity has a crescent-like shape at this time point of the cardiac cycle. Ao – non-coronary sinus of the aorta; CT – terminal crest; FO – oval fossa; LA – left atrium; LAFW – left atrial free wall; RA – right atrium; TSN – transseptal needle. Figure 2 ICE during atrial systole before the jump of the transseptal needle into the oval fossa. The LA cavity is virtually missing at this time point of the cardiac cycle. Ao – non-coronary sinus of the aorta; CT – terminal crest; FO – oval fossa; RA – right atrium; RV – right ventricle. Figure 3 ICE after the jump of the transseptal needle into the oval fossa. The needle tip points towards the aorta. Ao – non-coronary sinus of the aorta; LA – left atrium; RA – right atrium; TSN – transseptal needle. Discussion In the last years ICE emerged as a useful adjunctive tool in the field of interventional electrophysiology. It serves not only scientific purposes but practical issues as well. Its value for achieving successful and uncomplicated transseptal access to the LA cavity is well known [ 1 - 4 ]. However, it has not been used consistently for elucidating the possible mechanisms of mechanical complications during transseptal left heart procedures. ICE for transseptal puncture In this case ICE showed that the aneurysm of the IAS observed during the TEE was actually the angulated continuity between the enlarged aortic root and the IAS. The echocardiographic orientations of ICE are sometimes clearly off axis in comparison to standard transesophageal echocardiographic views. Nevertheless, to our opinion ICE is superior in providing more detailed picture of the neighboring cardiac structures. Although useful in guiding transseptal catheterizations, TEE does not always provide complete avoidance of complications even in patients with normal hearts [ 5 , 6 ]. Furthermore, the oval fossa itself was small but its membrane nevertheless showed bidirectional flapping motion. This suggests that ICE and TEE images are indeed no equivalent. As the fossa was abouting the aortic root this flapping motion was located in the angle between the two structures and had probably given rise to the false impression of an aneurysm. Additionally, very small distance between the oval fossa and the LA free wall was also observed. In such circumstances the risk of puncturing the left atrial free wall is prominent even if the puncture of the oval fossa itself is uncomplicated [ 2 , 3 ]. Usually ICE-guided redirection of the transseptal needle is all that is needed for achieving uncomplicated access to the LA in such cases. In the case presented there was virtually no distance between the IAS and the LA free wall during the atrial systole, so such a maneuver would not be of help. Also it was not possible to achieve a stable position of the transseptal needle in the fossa itself. It always slided and pointed to the aorta. This means that even if the distance from the oval fossa to the LA free wall was large enough the puncture would be complicated. To our opinion this is one possible explanation for the rare instances of inadvertent puncturing of the aorta when the fluoroscopic images appear to be completely satisfying. ICE for avoidance of other complications EP procedures are relatively safe procedures and have low complication rate. One of the most frequent complications is related to cardiac wall perforation with consequent pericardial effusion and tamponade. ICE, especially the one with phased-array transducer (deeper penetration) allows continuous monitoring of the pericardial space during EP procedures. This permits prompt detection of a pericardial effusion and an immediate guidance of a therapeutic puncture. Phased-array transducers are equipped with Doppler capabilities allowing assessment of the pulmonary venous flow pattern after pulmonary vein ablation on top of diameter measurements to exclude pulmonary vein stenosis, which is the most important complication of this procedure. Monitoring microbubble formation during RF energy application allows prevention of pulmonary vein stenosis [ 7 ]. ICE also allows detection of intracardiac thrombi during the procedures, especially during left-sided ablations [ 8 ]. Conclusion We strongly believe that this case is a good illustration of the usefulness of the ICE in the electrophysiology field and further enhances its value as a tool for avoiding complications when the intracardiac anatomy is unusual or distorted. As the number of transseptal procedures in the electrophysiology laboratories all over the world is steeply growing ICE definitely has the potential to become a routine at least in those institutions with large volume of left atrial procedures. List of abbreviations EP – electrophysiologic IAS – interatrial septum ICE – intracardiac echocardiography LA – left atrium; left atrial TEE – transesophageal echocardiography Supplementary Material Additional File 1 Preprocedural transesophageal echocardiography. The irregular oval-shaped structure at the center of the screen is the aortic root. At a certain moment one can see at its upper part the ostium and the most proximal part of the right coronary artery. Below is situated the left atrium and to the left – the right atrium. The oval fossa is in between. Click here for file Additional File 2 Intracardiac echocardiography, showing distorted intracardiac anatomy. The oval shape in the center of the screen is the non-coronary sinus of Valsalva. Below is the right atrium at the bottom of which the transseptal needle is clearly visible. The prominent muscular structure in the left-hand part of the image is the terminal crest. The membrane of the oval fossa, adjacent to the right-hand part of the non-coronary aortic sinus shows bidirectional flapping motion. During the atrial contraction the cavity of the left atrium virtually disappears. Click here for file Additional File 3 Intracardiac echocardiography showing sliding of the needle towards the aorta. The transseptal needle is already in the oval fossa with its tip pointing to the aorta. This is especially clearly visible after a premature beat. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553997.xml
555464	An ant colony optimisation algorithm for the 2D and 3D hydrophobic polar protein folding problem	Background The protein folding problem is a fundamental problems in computational molecular biology and biochemical physics. Various optimisation methods have been applied to formulations of the ab-initio folding problem that are based on reduced models of protein structure, including Monte Carlo methods, Evolutionary Algorithms, Tabu Search and hybrid approaches. In our work, we have introduced an ant colony optimisation (ACO) algorithm to address the non-deterministic polynomial-time hard ( NP -hard) combinatorial problem of predicting a protein's conformation from its amino acid sequence under a widely studied, conceptually simple model – the 2-dimensional (2D) and 3-dimensional (3D) hydrophobic-polar (HP) model. Results We present an improvement of our previous ACO algorithm for the 2D HP model and its extension to the 3D HP model. We show that this new algorithm, dubbed ACO-HPPFP-3, performs better than previous state-of-the-art algorithms on sequences whose native conformations do not contain structural nuclei (parts of the native fold that predominantly consist of local interactions) at the ends, but rather in the middle of the sequence, and that it generally finds a more diverse set of native conformations. Conclusions The application of ACO to this bioinformatics problem compares favourably with specialised, state-of-the-art methods for the 2D and 3D HP protein folding problem; our empirical results indicate that our rather simple ACO algorithm scales worse with sequence length but usually finds a more diverse ensemble of native states. Therefore the development of ACO algorithms for more complex and realistic models of protein structure holds significant promise.	Background Ant Colony Optimisation (ACO) is a population-based stochastic search method for solving a wide range of combinatorial optimisation problems. ACO is based on the concept of stigmergy – indirect communication between members of a population through interaction with the environment. An example of stigmergy is the communication of ants during the foraging process: ants indirectly communicate with each other by depositing pheromone trails on the ground and thereby influencing the decision processes of other ants. This simple form of communication between individual ants gives rise to complex behaviours and capabilities of the colony as a whole. From the computational point of view, ACO is an iterative construction search method in which a population of simple agents ('ants') repeatedly constructs candidate solutions to a given problem; this construction process is probabilistically guided by heuristic information on the given problem instance as well as by a shared memory containing experience gathered by the ants in previous iterations ('pheromone trails'). Following the seminal work by Dorigo et al. [ 1 , 2 ], ACO algorithms have been successfully applied to a broad range of hard combinatorial problems, including the traveling salesman problem, the graph colouring problem, the quadratic assignment problem and vehicle routing problems (see, e.g., [ 3 - 5 ]). The research presented in this paper builds on an ACO algorithm first proposed in [ 6 ] (and later improved in [ 7 ]) for ab-initio protein folding under a widely studied abstract model – the hydrophobic polar (HP) model. In particular, we extend our previous ACO algorithm to the 3D HP model and improve its performance by modifying the subsidiary local search procedure. The protein folding problem is one of the most challenging problems in computational biology, molecular biology, biochemistry and physics. Even under simplified lattice models, the protein folding problem is non-deterministic polynomial-time hard ( NP -hard) [ 8 ]. The ab-initio protein folding problem can be broken down into three sub-problems: 1) design of a model (with a desired level of accuracy); 2) definition of an energy function that can effectively discriminate between native and non-native states; and 3) design of a search algorithm that can efficiently find minimal-energy conformations. A number of search (or sampling) methods have been proposed in the literature to solve the protein folding problem, including Monte Carlo algorithms, Evolutionary Algorithms, Tabu Search and hybrid approaches. ACO, which has been very successfully applied to other combinatorial problems, appears to be a very attractive computational method for solving the protein folding problem, since it combines aspects of chain-growth and permutation-based search with ideas closely related to reinforcement learning. These concepts and ideas apply rather naturally to protein folding: By folding from multiple initial folding points, guided by the energy function and experience from previous iterations of the algorithm, an ensemble of promising, low-energy complete conformations is obtained. These conformations are further improved by a subsidiary local search procedure and then evaluated to update the accumulated pheromone values that are used to bias the generation of conformations in future iterations of the algorithm. In this paper, we ask and address the following questions: Is ACO a competitive method for solving the ab-initio protein folding problem under the 2D and 3D HP models? How does its performance scale with sequence length? What is the role of the parameters of the ACO algorithm for the efficiency of the optimisation process? Which classes of structures (if any) are solved more efficiently by ACO than by any other known algorithms? Finally, it should be noted that our ACO algorithm for this problem is based on very simple design choices, in particular with respect to the solution components reinforced in the pheromone matrix and of the subsidiary local search procedure. We discuss which of the many design choices underlying our algorithm should be reconsidered in order to achieve further performance improvements. The hydrophobic polar model Due to the complexity of the protein folding problem, simplified models such as Dill's hydrophobic-polar (HP) model have become one of the major tools for studying protein structure [ 9 ]. The HP model is based on the observation that the hydrophobic force is the main force determining the unique native conformation (and hence the functional state) of small globular proteins [ 9 , 10 ]. In the HP model, the primary amino acid sequence of a protein (which can be represented as a string over a twenty-letter alphabet) is abstracted to a sequence of hydrophobic (H) and polar (P) residues that is represented as a string over the letters H and P. The conformations of such an HP sequence are restricted to self-avoiding walks on a lattice. For the 2D HP model, a 2-dimensional square lattice is typically used, and the 3D HP model is generally based on a 3-dimensional cubic lattice. An example of a protein conformation under the 2D HP model is shown in Figure 1 . The energy of a conformation is defined as the number of topological contacts between hydrophobic amino acids that are not neighbours in the given sequence. More specifically, a conformation c with exactly n such H-H contacts has energy E ( c ) = n ·(-1); for example, the 2D HP conformation shown in Figure 1 has energy -9. The HP Protein Folding Problem can be formally defined as follows: Given an HP sequence s = s 1 s 2 ... s n , find an energy-minimising conformation of s , i.e., find c * ∈ C ( s ) such that E ( c *) = min{ E ( c ) \| c ∈ C }, where C ( s ) is the set of all valid conformations for s . It has been proved recently that this problem and several variations of it are NP -hard [ 8 ]. Existing 2D and 3D HP protein folding algorithms A number of well-known heuristic optimisation methods have been applied to the 2D and 3D HP Protein Folding Problem, including Evolutionary Algorithms (EAs) [ 11 - 15 ] and Monte Carlo (MC) algorithms [ 16 - 22 ]. The latter have been found to be particularly robust and effective for finding high-quality solutions to the HP Protein Folding Problem [ 18 ]. Besides general optimisation methods, there are other heuristic methods that rely on specific heuristics that are based on intuitions or assumptions about the folding process, such as co-operativity of folding or the existence of a hydrophobic core. Co-operativity is believed to arise from local conformational choices that result in a globally optimal state without exhaustive search [ 23 ]. Among these methods are the hydrophobic zipper method (HZ) [ 23 ], the contact interactions method (CI) [ 24 ], the core-directed chain growth method (CG) [ 25 ], and the constraint-based hydrophobic core construction method (CHCC) [ 26 ]. The hydrophobic zipper (HZ) strategy developed by Dill et al. is based on the hypothesis that once a hydrophobic contact is formed it cannot be broken, and other contacts are formed in accordance with already folded parts of the chain (co-operativity of folding) [ 23 ]. The contact interactions (CI) algorithm by Toma and Toma [ 24 ] combines the idea of HZ with a Monte Carlo search procedure that assigns different conformational freedom to the different residues in the chain, and thus allows previously formed contacts to be modified according to their computed mobilities. The core-directed chain growth method (CG) by Beutler and Dill [ 25 ] biases construction towards finding a good hydrophobic core by using a specifically designed heuristic function and by approximating the hydrophobic core with a square (in 2D) or a cube (in 3D). The constraint-based hydrophobic core construction method (CHCC) by Yue and Dill [ 26 ] is complete, i.e., always guaranteed to find a global optimum; it attempts to find the hydrophobic core with the minimal possible surface area by systematically introducing geometric constraints and by pruning branches of a conformational search tree. A similar, but more efficient complete constraint satisfaction search method has been proposed by Backofen et al. [ 27 ] for the more complex face-centred cubic lattice. An early application of Evolutionary Algorithms to protein structure prediction was presented by Unger and Moult [ 14 , 15 ]. Their non-standard EA incorporates characteristics of Monte Carlo methods. Currently among the best known algorithms for the HP Protein Folding problem are various Monte Carlo algorithms, including the 'pruned-enriched Rosenbluth method' (PERM) of Grassberger et al. [ 16 , 18 ]. PERM is a biased chain growth algorithm that evaluates partial conformations and employs pruning and enrichment strategies to explore promising partial solutions. Other methods for solving protein folding problems include the dynamic Monte Carlo algorithm by Ramakrishnan et al. [ 21 ], which introduced long-range moves involving disconnection of the chain, and the evolutionary Monte Carlo (EMC) algorithm by Liang and Wong [ 19 ], which works with a population of individuals that each perform Monte Carlo optimisation; a variant of EMC also reinforces certain secondary structures (alpha-helices and beta-sheets). Finally, Chikenji et al. introduced the multi-self-overlap ensemble (MSOE) Monte Carlo method [ 17 ], which considers overlapping chain configurations. Other Monte Carlo methods that have been particularly useful in off-lattice protein folding include generalised ensemble methods, such as umbrella sampling [ 28 ] (with replica exchange sampling [ 29 , 30 ] being the most common variant) and multi-canonical (entropic) sampling [ 30 , 31 ]. Replica exchange Monte Carlo (parallel tempering) has also been applied to the off-lattice HP model [ 32 ]. Currently, when applied to the square and cubic lattice HP model, none of these algorithms appears to completely dominate the others in terms of solution quality and run-time. Our ACO algorithm for the 2D and 3D HP protein folding problem In previous work, we have applied ACO to the 2D HP Protein Folding Problem [ 6 , 7 ]; in the following, we briefly summarise the main features of our ACO algorithm and the improvements introduced in this work. Details on our ACO framework and the new ACO-HPPFP-3 algorithm developed in the context of this work are given in the 'Methods' section. As usual, the ants in our ACO algorithm iteratively undergo three phases: the construction phase – during which each ant constructs a candidate solution by sequentially growing a conformation of the given HP sequence, starting from a folding point that is chosen uniformly at random among all sequence positions; the local search phase – when ants further optimise protein conformations folded during the construction phase; and the pheromone update phase – when ants update the pheromone matrix (representing the collective global memory of the colony) based on the energies of the conformations obtained after the construction and the local search phases. A general outline of ACO is shown in Figure 2 . The solution components used during the construction process, the local search phase and the pheromone update are local structure motifs (or relative folding directions) straight (S), left (L), right (R) in 2D, and straight (S), left (L), right (R), up (U), down (D) in 3D, which for each amino acid indicate its position on the 2D or 3D lattice relative to its direct predecessors in the given sequence (see Figure 3 ). In 3D, the relative folding directions are defined as in [ 33 ]: A local coordinate system is associated with every sequence position, such that S corresponds to the direction of the x axis, L to the direction of the y axis, and U to the direction of the z axis. Each local motif corresponds to a relative rotation of this coordinate system (for the forward construction: S = no rotation, L = 90° counter-clockwise around the z axis, R = 90° clockwise around the z axis, U = 90° clockwise around the y axis, D = 90° counter-clockwise around the y axis). Since conformations are rotationally invariant, the position of the first two amino acids can be fixed without loss of generality. Hence, we represent candidate conformations for a protein sequence of length n by a sequence of local structure motifs of length n - 2. For example, the conformation of Sequence S1-1 shown in Figure 1 corresponds to the motif sequence LSLLRRLRLLSLRRLLSL. During the construction phase, ants fold a protein from an initial folding point by probabilistically adding one amino acid at a time based on the two sources of information: pheromone matrix values τ (which represent previous search experience and reinforce certain structural motifs) and heuristic function values η (which reflect current energy of the considered structural motif); details of this process are given in the 'Methods' section. The relative importance of τ and η is determined by parameters α and β , respectively, whose settings are detailed in the 'Discussion' section. Similar to other ACO algorithms known from the literature, our algorithm for the HP Protein Folding Problem incorporates a local search phase that takes the initially built protein conformation and attempts to optimise its energy further, using probabilistic long-range moves that are described in detail in the 'Methods' section. Finally, the pheromone update procedure is based on two mechanisms: Uniform pheromone evaporation is modelled by decreasing all pheromone levels by a constant factor ρ (where 0 < ρ ≤ 1), and pheromone reinforcement is achieved by increasing the pheromone levels associated with the local folding motifs used in a fraction of the best conformations (in terms of energy values) obtained during the preceding construction and local search phase. Furthermore, to prevent search stagnation when all of the pheromone is accumulated on very few structural motifs, we introduce an additional renormalisation mechanism for the pheromone levels (controlled by a parameter θ where 0 ≤ θ < 1; details are given in the 'Methods' section). Different from our previous ACO algorithms for the HP Protein Folding Problem, our new algorithm, ACO-HPPFP-3, supports the 3D HP cubic lattice model in addition to the 2D HP square lattice model. Furthermore, it uses a different iterative improvement strategy, a modified long-range move operator and a less restrictive termination criterion in its local search phase. ACO-HPPFP-3 was used in all ACO experiments described in the following. Results To compare ACO-HPPFP-3 with algorithms for the 2D and 3D HP Protein Folding Problem described in the literature, we tested it on a number of standard benchmark instances as well as on two newly created data sets, one of which was obtained by randomly generating amino acid sequences with hydrophobicity value characteristic of globular proteins, while the other consists of biological sequences that were translated into HP strings using a standard hydrophobicity scale. (These new data sets will be described in more detail later in this section.) Results for standard benchmark instances The 21 standard benchmark instances for 2D- and 3D-HP protein folding shown in Table 1 have been widely used in the literature [ 6 , 12 , 14 - 17 , 19 , 25 ]. Experiments on these standard benchmark instances were conducted by performing a number of independent runs for each problem instance (in 2D: 500 runs for sequence length n ≤ 50, 100 runs for 50 < n ≤ 64, and 20 runs for n > 64; in 3D: 100 runs for each sequence). Unless explicitly indicated otherwise, we used the following parameter settings for all experiments: α : = 1, β : = 2, ρ : = 0.8 and θ : = 0.05. Furthermore, all pheromone values were initialised to 1/3 in 2D and to 1/5 in 3D, and a population of 100 ants was used, 50% of which were allowed to perform local search. The local search procedure was terminated when no improvement in energy had been obtained after between 1 000 (for n ≤ 50) and 10 000 (for n > 50) scans through the protein sequence. We used an elitist pheromone updating scheme in which only the best 1% of all ants was allowed to perform pheromone updates. The probability of keeping the previous direction when feasible during the long-range mutation move was set to 0.5 (see 'Methods' section). These settings were determined in a series of experiments in which we studied the influence of different parameter settings and will be further discussed later. All experiments were performed on PCs with 2.4 GHz Pentium IV CPUs, 256 Kb cache and 1 MB RAM, running Redhat Linux (our reference machine), and run-time was measured in terms of CPU time. Most studies of EA and MC methods in the literature, including [ 12 , 14 , 15 , 19 ], report the number of valid conformations scanned during the search. This makes a performance comparison difficult, since run-time spent for backtracking and the checking of partial or infeasible conformations, which may vary substantially between different algorithms, is not accounted for. We therefore compared ACO to the best-performing algorithm from the literature for which performance data in terms of CPU time is available – PERM [ 18 ] (we used the most recent implementation, which was kindly provided by P. Grassberger). We note that the most efficient PERM variant for the HP Protein Folding Problem uses an additional penalty of 0.2 for H-P contacts [ 34 ]. Since this corresponds to an energy function different from that of the standard HP model underlying our ACO algorithm as well as other algorithms developed in literature, we used the best performing variant of PERM [ 18 ] based on the standard energy function in our experiments. It may be noted that the chain growth process in PERM can start from the N - or C -terminus of the given HP sequence, and in many cases, this results in substantial differences in the performance of the algorithm. To capture this effect, we always ran PERM in both directions, and in addition to the respective average run-times, t 1 and t 2 , we report the expected time for solving a given problem instance when performing both runs concurrently, t exp = 2·(1/ t 1 + 1/ t 2 ) -1 . For all runs of PERM, the following parameter settings were used: inverse temperature γ : = 26 and q : = 0.2. The results obtained on standard 2D benchmark instances (see Table 2 ) indicate that ACO-HPPFP-3 is competitive with the EA and MC methods described in the literature; it works very well on sequences of sizes up to 64 amino acids and produces high quality suboptimal configurations for the longest sequences considered here (85 and 100 amino acids). On average, ACO requires less CPU time than PERM for finding best known conformations for Sequence S1-8; but PERM performs better for Sequences S1-6 and S1-7 as well as for the longer sequences of 85 to 100 residues (Sequence S1-9 to S1-11). Sequence S1-8 has a very symmetrical optimal state (see Figure 4 ), which – as argued in [ 18 ] – would be difficult to find for any chain growing algorithm. All algorithms from the literature which we are aware of have problems folding this sequence; ACO-HPPFP-3, on the other hand, is able to handle this instance quite well, since a number of ants folding from different starting points in conjunction with a local search procedure that involves large-scale mutations originating from different sequence positions can produce good partial folds for various parts of the chain. In comparison with other algorithms for the 2D HP Protein Folding Problem considered here (EA, EMC, MSOE), ACO-HPPFP-3 generally shows very good performance on standard benchmark instances. In case of the 3D HP Protein Folding Problem (see Table 3 ), the majority of algorithms for which we were able to find performance results in the literature use heuristics that are highly specialised for this problem. Unlike HZ, CG and CI, ACO-HPPFP-3 finds optimal (or best known) solution qualities for all sequences. However, PERM (when folding from the N -terminus) and CHCC consistently outperform ACO-HPPFP-3 on these standard 3D HP benchmark instances, and CG reaches best known solution qualities substantially faster in many cases. We note that for Sequence S2-3 and S2-7, PERM'S performance is greatly dependent on the folding direction. Result for new biological and random data sets To thoroughly test the performance of ACO-HPPFP-3, we created two new data sets of random and biological sequences of length ≈ 30 and ≈ 50 amino acids (ten sequences for each length; for details, see additional data file 1 ). Random sequences were generated based on the observation that most globular proteins have a fairly uniform amino acid profile, and the percent of hydrophobic residues of majority of globular proteins falls in the range of 40–50% [ 35 ]. Thus, the probability of generating character H at each position of a sequence was chosen to be 0.45, and in the remaining cases ( i.e., with probability 0.55), we generated a P. For the biological test-sets, ten sequences were taken from the PDBSELECT data set with homology < 25% from the Protein Data Bank (PDB) in order to obtain a non-redundant representative set of proteins. These protein sequences were translated into HP strings using the hydrophobicity scale classification of RASMOL [ 36 ], according to which the following amino acids were considered hydrophobic: Ala , Leu , Val , Ile , Pro , Phe , Met , Trp , Gly and Tyr . Non-standard amino acid symbols, such as X and Z, were skipped in this translation. Figures 5 and 6 illustrate the performance of ACO-HPPFP-3 vs PERM in terms of mean CPU time over 10 runs per instance and algorithm; for practical reasons, each run was restricted to 1 CPU hour on our reference machine, and the lowest energies obtained in these runs (listed in additional data file 1 ) are not necessarily optimal. As can be seen from these results, in 2D, ACO-HPPFP-3 performs roughly comparably to PERM (PERM'S t exp was calculated as described in the previous subsection): ACO-HPPFP-3 reaches the same energies as PERM, but on some instances, particularly of length 50, requires more run-time. In 3D, ACO-HPPFP-3 generally requires a comparable amount of run-time on sequences of length 30 and outperforms PERM on one random sequences of length 30, but performs noticeably worse on sequences of length 50 and in some cases does not reach the same energy. We also generated longer sequences of length 75; for these, ACO-HPPFP-3 failed to reach the minimal energy values obtained by PERM in a number of cases. The run-times for both algorithms are reported in detail in Additional file 1 ; we note that on some sequences, the performance of PERM depends significantly on the direction of folding. Interestingly, there is no significant difference in performance between the biological and random test-sets for either PERM or ACO-HPPFP-3. In summary, the performance of ACO-HPPFP-3 is comparable with that of PERM (the best known algorithm for the 2D and 3D HP Protein Folding Problem) on biological and random sequences of length 30–50, but worse on longer sequences. This scaling effect is significantly more pronounced in 3D than in 2D. We note that neither ACO-HPPFP-3 nor PERM were optimised for short sequences ( n ≤ 30), but by using parameter settings different from the ones specified earlier, the performance of both algorithms can be significantly improved in this case. Characteristic performance differences between ACO and PERM To further investigate the conditions under which ACO performs well compared to PERM, we visually examined native conformations found by both algorithms, paying special attention to conformations for which one of the two algorithms does not perform well (see Figures 7 and 9 ). Based on our observations, we hypothesised that PERM usually performs well on sequences that have a structural nucleus in the native conformation at one of the ends of the sequence (particularly the end from which PERM starts folding the sequence); on the other hand, it has trouble folding sequences whose native conformations have structural nuclei in the middle of the sequence. In comparison, ACO is not significantly affected by the location of the structural nucleus (or multiple nuclei) in the sequence, since it uses construction from different folding points as well as the long-range mutation moves in local search, which can initiate refolding from arbitrary sequence positions. Here, we use the term 'structural nucleus' to refer to a predominantly locally folded part of the chain that can be relatively easily folded sequentially based on local sequence information [ 37 ]. For most sequences considered in this study, we observed a single structural nucleus, which is not surprising, given their relatively short length; however, it is generally believed that longer sequences have multiple folding nuclei [ 37 ]. The left side of Figure 7 shows an example of a relatively short biological sequence (B50-7, 45 amino acids) with a unique native hydrophobic core in the 2D HP model. (This is rare for HP sequences, which usually have a high ground state and hydrophobic core degeneracy: According to our observations, of the 11 standard benchmark instances in 2D, only Sequences S1-1, S1-3, S1-4 have a unique hydrophobic core; in 3D, none of the sequences studied here have a unique hydrophobic core.) This sequence has no structural nuclei at its ends; instead, the two ends interact with each other. ACO-HPPFP-3 outperforms PERM by a factor of 2 on this sequence in terms of CPU time: using a cut-off time of 1 CPU hour per run, PERM found the optimum with energy -17 in an average run-time of 284.06 CPU seconds ( t 1 = 271 sec, t 2 = 299 sec), while using the same cut-off time and machine, ACO-HPPFP-3 found the optimum in an average run-time of 130 CPU seconds. We also designed two additional sequences, D-1 and D-2, of length 50 and 60, respectively, that have a unique native state in which both ends of the sequence interact with each other (see Figure 8 ). Sequence D-1 also has a structural nucleus near its C -terminus. When testing the performance of PERM and ACO-HPPFP-3 on these sequences, we found that on D-1, ACO-HPPFP-3 requires a mean run-time of 236 CPU seconds, compared to t 1 = 3 795, t 2 = 1, t exp = 2 CPU seconds for PERM (values are based on 100 successful runs). When this sequence was reversed, PERM started folding the sequence from the structural nucleus, and its mean run-time dropped to 1 CPU second. A result similar to that for sequence B50-7 was obtained for Sequence D-2, which has no structural nuclei at the ends, but a native state in which the ends interact with each other. Here, ACO-HPPFP-3 was found to require a mean run-time of 951 CPU seconds (again, mean run-times were obtained from 100 successful runs), compared to t 1 = 9 257, t 2 = 19 356, t exp = 12 525 CPU seconds for PERM; as expected, in this case, reversing the folding order of the sequence did not cause a decrease in PERM'S run-time. We also analysed native conformations of sequences on which PERM outperforms ACO and observed that the end from which PERM starts folding is relatively compact and forms a structural nucleus in the resulting conformation. An example of a conformation with the structural nucleus at the beginning of the sequence (near the N -terminus, i.e., residue 1) is shown in the right panel of Figure 7 . For this biological sequence (B50-5, 53 amino acids), PERM finds an optimal conformation with an energy of -22 in t 1 = 5, t 2 = 118, t exp = 9 CPU seconds, while the average run-time for ACO-HPPFP-3 is 820 CPU seconds. Our ACO algorithm generally performs worse than PERM on sequences that have structural nuclei at the ends, because it tends to spend substantial amounts of time compacting local regions in the interior of the sequence, while PERM folds more systematically from one end. These observations also hold in 3D, as seen from two random sequences folded in 3D (see Figure 9 ). To further investigate our hypothesis, we studied differences between the distributions of native conformations found by ACO-HPPFP-3 and PERM, respectively. For this purpose, we introduced the notion of relative H-H contact order, which captures arrangement of H residues in the core of the folded protein, and thus determines the topology of the conformation (the closely related concept of contact order was first defined in [ 38 ]). Relative H-H contact order is defined as follows: where l is the number of H-H contacts, n is the number of H residues in the sequence, and i and j are interacting H residues that are not neighbours in the chain. Intuitively, CO H-H specifies the average sequence separation between H-H residues in contact per H in the sequence. Figure 10 shows cumulative frequency distributions of relative H-H contact order values for sets of native conformations of a 2D (the left panel) and 3D (the right panel) standard benchmark instance, respectively, found by ACO-HPPFP-3 and PERM over 500 independent runs, each of which was terminated as soon as a native conformation had been found. These results show that the ACO algorithm finds a set of native conformations with a wider range of H-H contact order values than PERM; in particular, ACO-HPPFP-3 finds conformations with high relative H-H contact oder as compared to PERM (more distant parts of the chain interact; for example, relative CO H-H = 0.324 for Sequence S1-7 in 2D and relative CO H-H = 0.75 for Sequence S2-5 in 3D are not found by PERM; similar results were obtained for other sequences), which further supports our hypothesis that both, in 2D and 3D, PERM is biased toward a more restricted set of native conformations. We performed analogous experiments for the case where PERM is allowed to keep certain statistics from one run to another as in [ 18 ] (runs are no longer independent) and found no significant differences in the set of conformations obtained (data not shown). To further examine the topological differences between ensembles of native conformations found by the two algorithms, we also looked at the hydrophobic solvent accessible area (defined as SA H-H : = ∑ h E h , where E h is the number of unoccupied lattice sites around H residue h ), the number of H-H contacts, and the H-H contact order as a function of the length of the sequence prefix (starting from the N -terminus of the sequence – where PERM starts folding). In this analysis, we calculated the properties of interest mentioned above for the native conformations found in 100 independent runs by ACO-HPPFP-3 and PERM, and plotted the mean values of the respective quantities as functions of the sequence prefix length (see Figures 11 , 12 and 13 ). As seen in Figure 11 , ACO-HPPFP-3 is less greedy than PERM, both in 2D (left side) and in 3D (right side), and it tends to leave more lattice sites around H residues accessible for future contacts with other H residues that appear later in the chain. This is also reflected in the mean number of H-H contacts formed when folding prefixes of increasing length; ACO-HPPFP-3 tends to form fewer H-H contacts than PERM for short and medium size prefixes (see Figure 12 ). By examining the dependence of absolute H-H contact order (defined as , the average sequence separation per H-contact) on prefix length, we furthermore observed that different from PERM, ACO-HPPFP-3 realises the bulk of its local H-H interactions in the middle of the given sequence (see Figure 13 ). This further confirms that ACO is capable of finding native conformations with structural folding nuclei that are not located at or near the end of a given protein sequence. The results illustrated in Figures 11 , 12 and 13 are typical for all 2D and 3D HP instances we studied. Discussion Although conceptually rather simple, our ACO algorithm is based on a number of distinct components and mechanisms. A natural question to ask is whether and to which extent each of these contributes to the performance reported in the previous section. A closely related questions concerns the impact of parameter settings on the performance of ACO-HPPFP-3; further details concerning parameters can be found in the 'Methods' section. To address these questions, we conducted several series of experiments. In this context, we primarily used three standard test sequences: Sequence S1-7 of length 60 and Sequence S1-8 of length 64 (long sequences) in 2D, as well as Sequence S2-5 of length 48 in 3D (all standard benchmark sequences for 3D are 48 amino acids in length); these sequences were chosen because the CPU time required to find the best known solutions was sufficiently small to perform a large number of runs (100–200 per instance). Following the methodology of Hoos and Stützle [ 39 ], we measured run-time distributions (RTDs) of our ACO algorithm, which represent the (empirical) probability distribution over the run-time required to reach (or exceed) a given solution quality; the solution qualities used here are the known optima or best known energies for the respective sequences. Pheromone values and heuristic information Two important components of any ACO algorithm are the heuristic function, which indicates the desirability of using particular solution components during the construction phase, and the pheromone values, which represent information learned over multiple iterations of the algorithm. Three parameters control the influence of the pheromone information versus heuristic information on the construction of candidate solutions: the relative weight of the pheromone information, α ; the relative weight of the heuristic information, β ; and the pheromone persistence, ρ (see also 'Methods' section). In the first experiment, we investigated the impact of pheromone ( α ) and heuristic information ( β ), and their relative importance for the performance of our ACO algorithm. As can be seen from the results shown in Figure 14 , both the pheromone values and the heuristic information are important in 2D and 3D; when ignoring either of them ( α : = 0 or β : = 0, respectively), the algorithm performs worse, particularly for longer 2D sequences ( n > 50; for short 2D sequences with n ≤ 50, the pheromone matrix does not appear to play a significant role, since sequences are generally easily solved by the subsidiary local search procedure alone). The optimal settings for α and β for most problem instances seem to be around α = 1 and β = 2, as shown in Figure 14 . It should be noted that in 3D, pheromone information appears to be less important than in 2D, which suggests that the specific solution components used in our algorithms are somewhat less meaningful in 3D. The goal of the next experiment was to further explore the role of experience accumulated over previous iterations in the form of pheromone values. To this end, we varied the pheromone persistence, ρ , while keeping other parameters constant. The results shown in Figure 15 show that in 2D, it is important to utilise past experience ( i.e., to choose ρ > 0), but also to weaken its impact over time ( i.e., to use ρ < 1). At the same time, closer examination revealed that for ρ > 0, attrition, i.e., the construction of inextensible partial conformations, is a major problem, which is a result of the accumulation of pheromone from multiple conformations. This is why the backtracking mechanism described in the 'Methods' section is extremely important for the performance of our algorithm in 2D. In 3D, for the previously stated reasons and because of the fact that the attrition problem is much less severe, the impact of the persistence parameter is generally smaller than in 2D. Ant colony size and length of local search phase During the initial empirical evaluation of our algorithm, we observed that ant colony size ( i.e., the number of ants used in each iteration) and the duration of local search (expressed as a number of non-improving search steps we are willing to consider before terminating the local search procedure) are correlated and significantly affect its performance. To further investigate this phenomenon, we conducted additional experiments in which we fixed the ant colony size and varied the maximal number of non-improving steps during local search, and vice versa. In this series of experiments, different colony sizes were considered, from a single ant up to a population of 5 000 ants, and the number of non-improving steps in local search was varied from 100 to 10 000. The results, shown in Figure 16 , indicate that there is an optimal colony size of about 100 ants for both, 2D and 3D; ACO-HPPFP-3 is quite robust with respect to colony size, but performance decreases for very small or very large colony sizes. Intuitively, this is the case because the use of a population of ants provides diversification to the search process, which enables it to explore different regions of the underlying search space; very small populations provide insufficient diversification, and the search stagnates easily, while for very large populations, the additional time required for running the search phases for each ant on the same sequential machine is not amortised any longer by increased efficiency of the overall search process. Our results also indicate that the performance of ACO-HPPFP-3 is more sensitive to the number of non-improving steps than to ant colony size. The optimal value for the maximum number of non-improving steps tolerated (per ant) before the local search phase terminates was found to be around 1 000 for short 2D sequences ( n ≤ 50) and around 10 000 for long 2D sequences ( n > 50); the latter value also appeared to be optimal for all 3D sequences considered here. This observation follows the intuition that more degrees of freedom, as present for longer sequences and in higher dimensions, require more time for local optimisation, since for any conformation, improving neighbours tend to be rarer and hence harder to find. Selectivity and persistence of local search As described in the 'Methods' section, our ACO algorithm uses selective local search, i.e., local search is only performed on a certain fraction of the lowest energy conformations. We observed that ACO-HPPFP-3 is fairly robust with respect to the fraction of conformations to which local search is applied; good performance was obtained for local search selectivity values between 5% and 50%, but performance was found to deteriorate when local search is performed by all ants. Intuitively, similar to colony size, local search selectivity has an impact on search diversification. If too few ants perform local search, insufficient diversification is achieved, which typically leads to premature stagnation of the search process. On the other hand, if local search is performed by too many ants, the resulting substantial overhead in run-time can no longer be amortised by increased search efficiency. Similarly to selective local search, pheromone update was performed only by a certain fraction of so-called 'elitist ants' whose solution quality after the local search phase is highest within the population. As in the case of local search selectivity, ACO-HPPFP-3 shows robustly high performance for elitist fractions between 1% and 50% (results are not shown here), but performance deteriorates markedly when all ants in the colony are allowed to update the pheromone matrix. In a final experiment, we studied the impact of the persistence of local search, i.e., of the probability of retaining (feasible) previous relative directions during long-range mutation moves. As can be seen in Figure 17 , good performance is generally obtained for values between 0.3 and 0.7. Both extreme cases, = 0, which corresponds to an extremely H-contact greedy mutation operator, and = 1, in which refolding always follows previous directions when feasible, result in a substantial decrease in performance. When = 0, the decrease of performance in 3D is smaller than in 2D, because there is no severe attrition as in 2D, where greedy placement of H residues leads to early formation of very compact partial conformations, which often cannot be extended into valid complete conformations. The performance decrease for high values is due to insufficient ability of the chain to fold into a new conformation that accommodates well the local change in structure which triggered the refolding. Conclusions In this work, we have shown that ant colony optimisation (ACO) can be applied in a rather straight-forward way to the 2D and 3D HP Protein Folding Problems. Even though our ACO-HPPFP-3 algorithm is based on very simple structure components (single relative directions) and a simple subsidiary local search procedure (iterative first improvement), it performs fairly well compared to other algorithms and specialised heuristics on the benchmark instances considered here, particularly in 2D. The only non-specialised algorithm that typically performs better than our ACO algorithm, both in 2D and 3D, is PERM. We observed that, particularly in 3D, the run-time required by ACO-HPPFP-3 for finding minimum (or best known) energy conformations scales worse with sequence length than PERM. However, our results show that our ACO algorithm finds a different ensemble of native conformations compared to PERM, and has less difficulty folding sequences whose native states contain structural nuclei located in the middle rather than at the ends of a given sequence, as well as sequences with structures in which the ends interact. We found that two major components of ACO-HPPFP-3 – the pheromone values, which capture experience accumulated over multiple iterations of the search process and from multiple conformations, as well as the heuristic information that provides myopic guidance to the folding process – play a significant role for longer 2D sequences and, to a lesser extent, for 3D sequences, which suggests that in 3D, it may be preferable to associate pheromone values with more complex solution components. We also found that the subsidiary local search procedure is crucial for the performance of the algorithm; in particular, to ensure that high-quality conformations are obtained, it is very important to allow the local search procedure to run sufficiently long. In an earlier version of our algorithm [ 7 ], we used substantially more stringent termination criteria, which forced us to additionally use non-greedy local search (probabilistic iterative improvement, which accepts worsening steps) in addition to the greedy local search procedure used here. The results presented in this study indicate that by using a new and simpler local search procedure, ACO-HPPFP-3 achieves better performance; this is probably due to the fact that the new local search procedure is based on a type of long-range move that leads to a larger effective search neighbourhood. We have shown that all components of our ACO algorithms contribute to its performance. In particular, its performance is affected by the following components and parameters (listed in the order of decreasing impact): pheromone values, termination criterion for local search, persistence of long-range moves, ant colony size, pheromone persistence, heuristic function, selectivity of local search, and selectivity of pheromone update ( i.e., fraction of elitist ants). Because of its ability to find more balanced ensembles of minimum (or close to minimum) energy conformations, our new ACO algorithm can greatly facilitate investigations of the topology and location of structural nuclei, which we plan to undertake in future work. Finally, while HP protein folding problems are of considerable interest because of their conceptual simplicity, ultimately, most applications of protein folding algorithms require the use of more realistic models of protein structure. Since it does not rely on heuristics and properties that are specific to the HP model and yet performs very well on this restrictive, but not entirely unrealistic abstract model, we believe that relatively straight-forward extensions of our ACO algorithm to more complex and realistic models of protein structure hold significant promise. Methods Our new ACO algorithm, ACO-HPPFP-3, iterates construction, local search, and pheromone update phases until a termination condition is satisfied; in the context of this work, we mostly terminated the algorithm upon reaching a given energy threshold. In the following, we describe the three search phases in detail. Construction phase, pheromone and heuristic values During the construction phase of ACO-HPPFP-3, each ant first determines a starting point within the given protein sequence; this is done by uniform random choice. From this starting point, the sequence is folded in both directions, adding one residue at a time. Each ant performs probabilistic chain-growth construction of the protein conformation, where in every step, the structure is extended either to the left or to the right, such that the ratio of unfolded residues at each end of the protein remains (roughly) unchanged. Here, we assume that folding is performed in 3D (the 2D case is handled analogously by considering three relative directions { S, L, R } instead of five { S, L, R, U, D }, see also [ 6 ]). The relative directions in which the conformation is extended in each construction step are determined probabilistically based on a heuristic function η i , d and pheromone values τ i , d , according to the formula: The pheromone values τ i , d indicate the desirability of using the local structure motif with relative direction d ∈ { S, L, R, U, D } at sequence position i . Initially, all τ i , d are equal, such that local structure motifs are chosen in an unbiased way; but throughout the search process, the pheromone values are updated to bias folding towards the use of local motifs that occur in low-energy structures (the updating mechanism will be described in more detail later). The heuristic values η i , d are based on the energy function E . They are defined according to the Boltzman distribution as η i , d : = , where γ is a parameter called the inverse temperature (as in [ 18 ]), and h i , d is the number of new H-H contacts achieved by placing amino acid i at the position specified by direction d . During construction, it may happen that the chain cannot be extended without running into itself. This situation is called attrition , and our algorithm overcomes it as follows: First, starting at the end at which attrition occurred, half of the sequence that has been folded up to this point is unfolded. Then, this segment of the chain is refolded; the first residue ( i.e., the last one that was unfolded) is placed such that its relative direction differs from what it had been when attrition occurred, while all of the subsequent residues are folded in a feasible direction that is chosen uniformly at random. This backtracking mechanism is particularly important for longer protein sequences in 2D, where infeasible conformations are frequently encountered during the construction phase. Local search The local search phase is based on a long-range mutation move that has been designed to avoid infeasible conformations. It also has a number of important advantages over the more commonly used point mutation moves or Monte Carlo moves ( i.e., the end, crankshaft and corner moves [ 40 ]): It is easy to implement; it decreases the number of infeasible conformations encountered, even when the protein is very compact (at high densities); it considers a larger neighbourhood that subsumes the single point mutation neighbourhood; and it has some validity in terms of the physical processes taking place during the protein folding process. Similar attempts have been previously undertaken, but these involved disconnection of the chain [ 21 ]. From studies of protein folding dynamics, it is known that proteins display a broad range of motions that range from localised motions to slow large-scale movements [ 37 ]. Inspired by this complex process, we designed a long-range mutation move that starts by selecting a residue whose relative direction is randomly mutated and then adapts the rest of the chain by probabilistically changing relative directions starting from this initial position [ 7 ]. During this adaptation, for each residue, with a probability (0 ≤ ≤ 1) its previous relative direction, if it is still feasible, is left unchanged, and otherwise ( i.e., with probability 1 - , or if the previous direction has become infeasible), a different relative direction is chosen, where the probability for each direction d is proportional to the corresponding heuristic value η i , d . Formally, this can be written as follows: where P [ d i : = ] is the probability of choosing direction as the relative direction d i at sequence position i . Unlike in our previous implementation [ 7 ], the local search phase of our new ACO algorithm is a simple iterative first improvement procedure that is based on this long-range mutation move. The outline of this local search procedure is shown in Figure 18 . Iterative first improvement accepts a new conformation generated via long-range mutation only if the solution quality of a new conformation c ' improves over the current solution quality (energy) of c. This search process is greedy in the sense that it does not allow worsening steps, and it is terminated when no improving steps have been found after a specific number of scans through the chain (this number is a parameter of the algorithm). Since this local search procedure has a relatively high time-complexity, in each iteration of ACO-HPPFP-3 it is only applied to a certain fraction of the highest-quality conformations constructed by the ants in the preceding construction phase. Update of the pheromone values After each construction and local search phase pheromones are updated according to τ i , d := ρ · τ i , d , (4) where 0 < ρ ≤ 1 is the pheromone persistence, a parameter that determines how much of the information gathered in previous iterations is retained. Subsequently, selected ants with low-energy conformations update the pheromone values according to τ i , d := τ i , d + Δ i , d , c , (5) where Δ i , d , c is the relative solution quality of the given ant's candidate conformation c if that conformation contains local structure motif d at sequence position i , and zero otherwise. As a further mechanism for preventing search stagnation, we use an additional renormalisation of the pheromone values that is conceptually similar to the method used in MAX - MIN Ant System [ 41 ]: After the standard pheromone updates according to Equations 3 and 4, all τ values are normalised such that ∑ d ∈{ S , L , R , U , D } τ i , d = 1 for every residue i ; additionally, whenever for a given sequence position i the minimal normalised pheromone value (min d ∈{ S , L , R , U , D } τ i , d )/(∑ d ∈{ S , L , R , U , Dr } τ i , d ) falls below a threshold θ (which is a parameter of the algorithm), the minimal τ i , d value is set to θ , while the maximal τ i , d value is decreased by θ - min d ∈{ S , L , R , U , D } τ i , d . (If there is more than one minimal τ i , d value, all of these are increased to θ , and if there is more than one maximal τ i , d value, one of them is chosen uniformly at random.) This guarantees that the probability of selecting an arbitrary local structure motif for the corresponding sequence position does not become arbitrarily small, and hence ensures the probabilistic approximate completeness of our algorithm (see [ 42 ]). Implementation details and availability ACO-HPPFP-3 has been implemented in C++ and compiled using gcc (version 3.3.3) for the Linux operating system; a Linux executable is available from . Authors' contributions Both authors contributed to the development of ideas, design of experiments, analysis and interpretation of results, and the writing of the paper. AS implemented the proposed method and performed the computational experiments. Supplementary Material Additional File 1 Additional information on biological and randomly generated HP sequences. This file (in .pdf format) contains tables providing additional information on our new test sets of biological and randomly generated HP sequences and the results from our computational experiment with ACO and PERM. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555464.xml
522820	Radiation enteropathy and leucocyte-endothelial cell reactions in a refined small bowel model	Background Leucocyte recruitment and inflammation are key features of high dose radiation-induced tissue injury. The inflammatory response in the gut may be more pronounced following radiotherapy due to its high bacterial load in comparison to the response in other organs. We designed a model to enable us to study the effects of radiation on leucocyte-endothelium interactions and on intestinal microflora in the murine ileum. This model enables us to study specifically the local effects of radiation therapy. Method A midline laparotomy was performed in male C57/Bl6 mice and a five-centimetre segment of ileum is irradiated using the chamber. Leucocyte responses (rolling and adhesion) were then analysed in ileal venules 2 – 48 hours after high dose irradiation, made possible by an inverted approach using intravital fluorescence microscopy. Furthermore, intestinal microflora, myeloperoxidase (MPO) and cell histology were analysed. Results The highest and most reproducible increase in leucocyte rolling was exhibited 2 hours after high dose irradiation whereas leucocyte adhesion was greatest after 16 hours. Radiation reduced the intestinal microflora count compared to sham animals with a significant decrease in the aerobic count after 2 hours of radiation. Further, the total aerobic counts, Enterobacteriaceae and Lactobacillus decreased significantly after 16 hours. In the radiation groups, the bacterial count showed a progressive increase from 2 to 24 hours after radiation. Conclusion This study presents a refinement of a previous method of examining mechanisms of radiation enteropathy, and a new approach at investigating radiation induced leucocyte responses in the ileal microcirculation. Radiation induced maximum leucocyte rolling at 2 hours and adhesion peaked at 16 hours. It also reduces the microflora count, which then starts to increase steadily afterwards. This model may be instrumental in developing strategies against pathological recruitment of leucocytes and changes in intestinal microflora in the small bowel after radiotherapy.	Background Radiotherapy is widely used in treating different types of cancer and is an effective therapeutic modality against abdominal and pelvic cancers. Gastrointestinal tract damage by radiotherapy limits its efficacy in cancer treatment. The small bowel is highly radiosensitive and very mobile and is thus an important dose-limiting organ during radiation therapy for abdominal and pelvic cancer [ 1 ]. Radiation induces an inflammatory response in target and surrounding tissues, which is characterised by accumulation of plasma proteins and leucocytes. Leucocyte recruitment is a multi-step process, which includes leucocyte rolling, activation and firm adhesion to the endothelium [ 2 ]. Leucocyte rolling reduces the velocity and allows time for leucocytes to detect chemotactic signals on the endothelial surface. It is now widely accepted that leucocyte rolling is a precondition for the subsequent firm adhesion and extravascular accumulation [ 3 , 4 ]. With intravital microscopy, several studies have demonstrated that the selectin family of adhesion molecules predominantly mediates leucocyte rolling and that stationary adhesion is supported by the β 2 -integrins [ 5 , 6 ]. Several animal models exist in order to investigate radiation-induced leucocyte endothelium cell responses, which can broadly be divided into two groups. Topical radiation (abdominal and pelvic) [ 7 - 9 ] and more recently, segmental radiation of an isolated short segment of small intestine where different responses to radiation are examined [ 1 , 10 ]. We refined and developed the latter model, incorporating a platform where mice can be placed allowing exposure of the required segment of intestine for irradiation. This allows us to expose exteriorised intestinal sections to tailored high dose radiation, greatly minimizing scattering effects and thereby consequently avoiding surrounding tissue damage. The purpose of this study was to refine a small bowel radiation model in order to study the time responses in terms of leucocyte rolling, adhesion, myeloperoxidase (MPO) levels, histology and intestinal floral changes in response to high dose radiation of the ileum, where the exact biologically effective dose could be calculated. Methods Animals Male C57Bl/6J mice weighing 22–26 g were kept under standard laboratory conditions maintained on a 12 hour light and 12 hour dark cycle and were allowed free access to animal chow and tap water ad libitum . All experimental procedures were performed in accordance with legislation on the protection of animals and were reviewed and approved by the Lund University Ethic's Committee for Animal Experimentation. Anesthetic and surgical preparation The mice were anesthesized with 7.5 mg Ketamine hydrochloride (Hoffman-La Roche, Basel Switzerland) and 2.5 mg Xylazine (Janssen Pharmaceutica, Beerse, Belgium) per 100 g body weight by intraperitoneal ( i.p .) injection. The animals were placed in supine position on a heating pad (37°C) for maintenance of body temperature. A small midline incision (1.0–1.5 cm) was performed and a 5 cm segment of ileum located 5 cm from the ileocaecal valve was exteriorised and marked with 5-0 non-absorbable sutures. Any other visible prolapsed abdominal content was replaced back into the abdomen and the animal was placed on the specially designed frame/chamber (Figure 1 ), with the loop of intestine fixed between two perspex sheets. The exposed ileum was subjected to a single dose of high dose radiation of 19 Gy and thereafter replaced in the abdomen and the incision closed with a polypropylene suture. At the appropriate time a polyethylene catheter (PE-10 with an internal diameter of 0.28 mm) was placed into the internal jugular vein for administration of fluorescent markers. Leucocyte-endothelium interactions were then observed using an inverted intravital fluorescence microscopy (IIVM) at different time points (2–48 hours). Figure 1 The chamber for segmental intestinal radiation exposure. Experimental protocol The animals were divided into two different groups; Radiation & Surgery group (R+) and Sham Radiation & Surgery group (R-) which served as negative controls. The irradiated groups of mice ( n = 6/time point) were exposed to 19 Gy of radiation and leucocyte-endothelium interactions were measured 2, 6, 16, 24 and 48 hours after induction of radiation using an IIVM (each group n = 6, likewise the R- groups). At the end of the procedure samples were collected for histology, and measurement of intestinal microflora, MPO and systemic leukocyte counts. Radiation The irradiations were undertaken using a clinical linear accelerator (Varian Clinac 2100C). The exteriorised mouse intestine was positioned between Perspex slabs to accomplish sufficient secondary radiation scatter and thereby a reproducible and homogenous dose distribution. The absorbed dose was verified with independent measurements and was found to be within 5% throughout the intended volume using this technique. Using an asymmetrically half blocked 6 MV beam and extra lead shielding (Figure 1 ), the treatment field perfectly fitted the exteriorised intestine while the remaining body was kept outside the radiation beam. An absorbed dose of 19 Gy was delivered to the intestine as this dose causes consistent structural, cellular, and molecular changes [ 11 ]. The absorbed dose rate was 3.2 Gy/minute and consequently the irradiation time for each animal was approximately 6 minutes. During irradiation the intestine in the chamber is protected from large temperature variations and trauma by perspex sheets. The exposure time from surgery, through irradiation to wound closure is kept at a minimum, taking approximately 15 minutes, thus keeping stress and trauma levels low. Intravital microscopy Observations of the intestinal microcirculation were made using an inverted Olympus microscope (IX70, Olympus Optical Co. GmbH, Hamburg Germany) equipped with different lenses (x10/NA 0.25 and x40/NA 0.60). The microscopic images were televised using a charge-coupled device video-camera (FK 6990 Cohu, Pieper GmbH, Schwerte, Germany) and recorded on videotape (Sony SVT-S3000P S-VHS recorder) for subsequent off-line analysis. To prevent drying during microscopic observations the intestinal segment was placed on a saline moistened cotton gauze and thereafter positioned under the microscope. After a 5-min equilibration period, quantitative measurements were taken. Analysis of leucocyte-endothelium interactions (rolling and adhesion) was made in venules (inner diameter 15–30 μm) with stable resting blood flow. Blood perfusion within individual microvessels was studied after contrast enhancement by i.v . administration of fluorescein isothiocyanate (FITC)-labelled dextran (MW 150000), (0.05 ml, 5 mg/ml, Sigma Chemical Co. St. Louis, MO, U.S.A.). In vivo labelling of leucocytes with rhodamine-6G (0.1 ml, 0.5 mg/ml, Sigma Chemical Co. St. Louis, MO, U.S.A.) enabled quantitative analysis of leucocyte flow behavior in the ileum microcirculation. Due to its relatively higher molecular weight FITC-dextran stains/labels the intravascular plasmatic phase of the blood under epi-illumination with blue light (excitation wavelength 490 nm; emission wavelength 510 nm) whereas the lower molecular weight of Rhodamine 6G allows for labelling of leucocytes and platelets using green fluorescent light (excitation wavelength 530 nm; emission wavelength 560 nm). Quantification of microcirculatory parameters was performed off-line by frame-to-frame analysis of the videotaped images. Leucocyte rolling was determined by counting the number of leucocytes passing a reference point in the venule per 20 sec and is expressed as cells/min. Firm adhesion was measured by counting the number of cells adhering to the venular endothelium (200–300 μm long segments) and remained stationary for 20 sec and is given as cells/mm venule length. Blood flow velocities were analysed by means of a video assisted computer image analysis programme, CapImage software (Zeintl, Heidelberg, Germany). The staining of the plasmatic phase by FITC-dextran gives an indirect enhancement of red blood cells which appear dark in the illuminated surrounding plasma. The CapImage uses the FITC-dextran image to calculate the red blood cell velocity. The velocity was calculated as a mean value from 5–8 measurements per venule and is expressed as mm/sec. Venular wall shear rate was determined based on the Newtonian definition: wall shear rate = 8 [(red blood cell velocity/1.6)/venular diameter] as described previously [ 12 ]. MPO measurement The enzyme myeloperoxidase (MPO) is abundant in neutrophil leucocytes and has been found to be a reliable marker for the detection of neutrophil accumulation in inflamed tissue. To determine tissue MPO content, radiated ileal tissue was collected, weighed, homogenized in 10 ml 0.5% hexadecyltrimethylammonium bromide, and freeze thawed, after which the MPO activity of the supernatant was assessed. The enzyme activity was determined spectrophotometrically as the MPO-catalysed change in absorbance occurring in the redox reaction of H 2 O 2 (460 nm, 25°C). Values are expressed as MPO units per g tissue. Histological study Samples from the irradiated small intestine were placed in 4% phosphate buffered formaldehyde. Paraffin-embedded samples were sliced and studied under light microscopy after staining with hematoxylin and eosin. At least 3 slides were studied from each specimen in a blinded fashion. Intestinal microflora Tissue samples from the irradiated small intestine were first placed in 5 ml of sterile transport medium [ 13 ]. Samples were then placed in an ultrasonic bath (Millipore, Sweden) for 5 minutes and then rotated on Chiltern (Terma-Glas, Gothenberg, Sweden) for 2 minutes. After a conventional dilution procedure, viable counts were obtained from Brain Heart Infusion (BHI) that was incubated aerobically and anaerobically at 37°C for 72 hours (aerobic and anaerobic bacterial count, respectively), and from Rogosa agar (Oxoid, Hampshire, England) that was incubated anaerobically at 37°C for 72 hours (lactobacilli counts). Viable counts were also obtained from violet red-bile-glucose agar (VRBD) (Oxoid, Hampshire, England) that was incubated aerobically at 37°C for 24 hours (Enterobacteriaceae counts) and from BHI agar containing gram-negative anaerobic supplement (Oxoid, Hampshire, England) that was incubated anaerobically at 37°C for 72 hours (gram negative anaerobic bacterial counts). Systemic leucocyte counts 20 μl blood was mixed with Turk's solution (0.2 mg gentian violet in 1 ml glacial acetic acid, 6.25 % v/v) in a 1:10 dilution. Leucocytes were counted and differentiated as polymorphonuclear (PMNL) or mononuclear (MNL) cells in a Burker chamber. Statistical analysis Statistical evaluations were performed using the Kruskal-Wallis one way analysis of variance on ranks for unpaired samples (Dunn's post hoc test was used). For bacterial microflora in comparing 2 groups we used Mann-Whitney Rank sum test, and for the comparison of the different time points within the radiated groups we used One Way ANOVA followed by multiple comparisons versus control group (Dunnett's method). The results are presented as mean values ± SEM. Differences were considered to be significant at P < 0.05. Results Radiation-induced leucocyte-endothelium interactions in the ileum Intravital microscopic studies in post-capillary venules of the distal ileum in sham operated mice (controls) revealed only occasional interactions between leucocytes and the microvascular endothelium, i.e . the number of rolling and adherent leucocytes was 2.4 ± 1.2 cells/min and 1.7 ± 1.7 cells/mm, respectively. In contrast, radiation (19 Gy) evoked a marked time-dependent leucocyte response, i.e . a significant increase in both leucocyte rolling and firm adhesion over time (Figures 2 and 3 , P < 0.05, vs . controls, n = 5–10). We observed that leucocyte rolling peaked two hours after radiation (38 ± 7 cells/min, (Figure 2 ), P < 0.05 vs . sham, n = 6–10), whereas leucocyte adhesion was maximum after 16 hours showing a marked response of 59 ± 14 cells/mm (Figure 3 , P < 0.05 vs . sham, n = 6). Interestingly, both the leucocyte rolling and adhesion responses to radiation returned to baseline levels 48 hours after radiation (Figures 2 and 3 , P > 0.05, vs . sham, n = 5–10). There was no difference in the hemodynamic parameters between the different experimental groups (Table 1 ) and also no significance difference could be seen in the systemic leucocyte counts. Figure 2 Venular leucocyte rolling in the mouse ileum at different time points after radiation. Data represents mean ± SEM. Figure 3 Venular leucocyte adhesion in the mouse ileum at different time points after radiation. Data represents mean ± SEM. Table 1 Hemodynamic parameters in ileal venules Diameter (μm) Red blood cell velocity (mm s -1 ) Wall shear rate (s -1 ) Sham 2 hrs 25.5 ± 2.1 1.57 ± 0.21 314 ± 44 Sham 16 hrs 26.2 ± 1.2 1.61 ± 0.14 309 ± 19 Radiation 2 hrs 22.2 ± 1.6 0.93 ± 0.08 212 ± 44 Radiation 6 hrs 24.5 ± 2.4 1.53 ± 0.12 319 ± 56 Radiation 16 hrs 21.7 ± 2.5 0.93 ± 0.22 222 ± 74 Radiation 24 hrs 20 ± 1.9 1.26 ± 0.16 408 ± 48 Radiation 48 hrs 25.5 ± 2.3 1.32 ± 0.11 264 ± 57 Radiation was directed to the ileum and leukocytes responses were measured after 2–48 hours. Sham-operated controls underwent identical procedures except undergoing radiation. Responses measured at 2 and 16 hours (n = 5,6). Blood flow velocities were measured off-line by frame-to-frame analysis of the videotaped images. Data are mean ± SEM. Histological changes following radiotherapy At 2 hours we could not observe any marked differences in the number of inflammatory cell types compared to the controls (Figure 4 ). At 6 hours we found quite a number of apoptotic epithelial cells, and a few inflammatory cells – mainly neutrophil granulocytes in the lamina propria. Both the granulocytes and the apoptotic cells increased in numbers at 16 hours. An increase in the inflammatory infiltrate was also observed in the smooth muscle layer (muscularis propria). 24 hours after radiation the muscularis mucosae was oedematous and infiltrated by granulocytes; there was clearly visible lymph vessel ectasia and apoptosis mainly in the deeper parts of the crypts. Forty-eight hours after radiation a vast increase of goblet and apoptotic cells was seen in the whole length of the epithelium and crypts. On the other hand there was a reduction in lymph vessel ectasia, oedema and in the number of inflammatory cells present (Figure 5 ). Figure 4 A Cross section of intestinal wall 2 hrs after irradiation. No marked differences in the number of inflammatory cell types compared to the controls. B Cross section of intestinal wall 48 hrs after irradiation. A vast increase of goblet and apoptotic cells was seen in the whole length of the epithelium and crypts. There was a reduction in lymph vessel ectasia, oedema and in the number of inflammatory cells present compared to earlier time points. Intestinal microflora Compared to the sham groups; the aerobic, Enterobacteriaceae , Lactobacillus and anaerobic counts had decreased two hours after radiation (Figure 5 ), the same groups, with the exception of the anaerobic count were significantly decreased sixteen hours after radiation (Figure 6 ). There were no significant differences between the experimental groups twenty four hours after radiation compared to the sham group (Figure 7 ). When assessing the trends within the various radiated bacterial groups compared to the 24 hour levels we found significant decreases in the aerobic count at 2 hours; in the anaerobic count at 2 and 6 hours; and in the Enterobacteriaceae at 2, 6 and 16 hours. There were no significant changes in the Lactobacillus count at the different time points within the radiated groups (Figure 8 ). Figure 5 Ileum bacterial microflora in sham and 2 hours after radiation groups. * denotes p < 0.05 compared to sham group. Figure 6 Ileum bacterial microflora in sham and 16 hours after radiation groups. * denotes p < 0.05 compared to sham group. Figure 7 Ileum bacterial microflora in sham and 24 hours after radiation groups. No significant difference between the experimental groups. Figure 8 Ileum bacterial microflora in the radiated groups at different time points. * denotes p < 0.05 compared to 24 hours radiated group. MPO measurement There were no differences in MPO measurements in the experimental groups. Discussion The frequent use of radiotherapy for abdominal and pelvic malignancies results in an increased risk of radiation enteritis [ 14 ]. The dose of radiation that can be applied in clinical practice is usually limited by the need to restrict the number and severity of side effects in normal tissues surrounding a tumour, which are unavoidably exposed to radiation [ 8 ]. Intestinal radiation toxicity (radiation enteropathy) is characterised by mucosal barrier breakdown and inflammation, followed by development of progressive vascular sclerosis and intestinal wall fibrosis. The process is accompanied by sustained over expression of inflammatory and fibrogenic cytokines [ 15 , 16 ]. An early inflammatory response, beginning a few hours after irradiation, characterised by leucocyte infiltration into the irradiated organs is regarded as one of the main determinants of radiation-induced organ damage [ 17 , 18 ]. The development of an inflammatory response involves sequential leucocyte-endothelial cell interactions. Different families of cell adhesion molecules have been shown to participate in the process of leucocyte recruitment [ 19 ]. There are three major families of adhesion molecules involved in the leucocyte recruitment process, the selectins, the integrins and the immunoglobulin supergene families [ 20 ]. The present study has concentrated on the acute effects of radiation injury on leucocyte rolling and adhesion at specific time points after radiation. We found that radiation evoked a marked time dependent leucocyte response with a significant increase in both leucocyte rolling and firm adhesion over time. Leucocyte rolling peaked 2 hours after radiation whereas leucocyte adhesion was highest after 16 hours showing a marked response. Interestingly, both the leucocyte rolling and adhesion responses to radiation were back to baseline 48 hours after radiation. An intravital microscopic study of radiation-induced leucocyte-endothelial cell interaction using abdominal radiation and a dose of 20 Gy revealed an increased leucocyte rolling in mesenteric venules 2 hours after radiation, with a marked increase in leucocyte adhesion and emigration noted at 6 hours [ 18 ]. In another study of radiation-induced inflammatory damage, abdominal irradiation was administered using 4 and 10 Gy respectively [ 8 ]. Here an increase in leucocyte rolling was observed 2 hours after radiation, which then returned to basal levels at 6 and 24 hours respectively. An increase in leucocyte adhesion was also observed 2 hours after irradiation, which was then sustained during the 24 hour observation period [ 8 ]. In our study we showed the maximum effect on rolling after 2 hours and adhesion after 16 hours and the return to basal levels 48 hours after radiation. We used a single high dose radiation of 19 Gy directly to an exteriorised segment of ileum. This dose was chosen because it has been shown to give a good correlation or dose response relationship of histopathological changes (e.g. mucosal ulceration, vascular sclerosis) to incidence of clinical complications and cellular evidence of injury [ 1 ]. When comparing our results (ileal venule measurements) to those from other tissue, namely from the pial venules of cerebral microvasculature of the rat after 20 Gy irradiation [ 21 ], we find that the results follow a similar time course. Assuming that the radiation dose distribution is similar in all experiments mentioned above, the differences in peak times for leucocyte rolling and adhesion may probably be due to differences in radiation dose/duration, the extent of trauma, the effect of anaesthesia, the mode and duration of experiments. Endogenous bacterial flora produces nutrients (e.g. short-chain fatty acids) for the mucosa; prevents overgrowth of potentially pathogenic micro-organisms; stimulates the immune system especially the gut-associated lymphoid tissue; helps eliminate toxins from the lumen and participates in intestinal regulation, motility and blood flow [ 22 ]. Radiation on the other hand influences and alters the mucosal microflora, and this in combination with barrier dysfunction leads to a translocation of microbes through the mucosa into blood circulation [ 23 ]. Our experiment shows that radiation affects the intestinal microflora. Two hours after radiation the aerobic, anaerobic, Enterobacteriaceae and Lactobacillus counts were decreased and after 16 hours the aerobic, Enterobacteriaceae and Lactobacillus counts were still decreased in the radiated groups compared to sham controls. Twenty-four hours after radiation there was no significant difference between the experimental groups. Comparing the results of the irradiated groups alone, one observes an increase in bacterial count over time after radiation. It seems that radiation decreases the bacterial count at early time points with no difference in total bacterial count at late time points. This total count does not reflect the difference in bacterial species within each group, and thus, further investigations are needed to study the imbalances that occur. One study has shown that microorganisms such as Escherichia, Proteus, Clostridium, normally absent in healthy animals, appear in the intestines of guinea pigs subjected to irradiation. At the same time lactobacilli and bifidobacteria sharply decrease in number [ 24 ]. Bacterial overgrowth and intestinal pseudo-obstruction may succeed abdominal radiotherapy and the impaired motility emerges as a causal factor for gastrointestinal colonization with gram-negative bacilli. Abnormal motility and gram-negative bacilli in the gut may be essential in the pathogenesis of late radiation enteropathy [ 25 ]. Changes in intestinal microflora therefore most probably affect the course of the development of radiation enteropathy. Acute intestinal symptoms during pelvic radiotherapy may not depend only on mucosal damage [ 26 ]. Post-radiation gut structural damage occurs early and parallels functional changes of the intestinal mucosa, including increased epithelial permeability (shown both in vivo and ex vivo), activation of secretory pathways, decreased nutrient absorption, diarrhoea, and weight loss [ 27 ]. The microfloral changes, which we have shown, could play an important role in the structural and functional intestinal changes after radiation, particularly in the presence of intestinal mucosal changes and increased intestinal permeability. Patients with carcinoma of the uterine cervix or endometrium receiving postoperative radiation therapy have a significant decrease in intestinal microflora after the first radiation exposure, whereas at the end of radiotherapy all bacteria have increased and reached basal values except Enterococcus faecium 1, lactobacilli and total anaerobes. In some patients an overgrowth of some Clostridium spp. (potential pathogens) associated with clinical symptoms, was observed. Patients receiving radiotherapy may thus benefit from the intake of oral bacteriotherapy [ 28 ]. The importance of investigating the effects of radiation on the different bacterial species within the total count is therefore of significance for the modulation of treatment regimes. The histological changes following radiation are both time and dose dependent [ 29 , 30 ]. Soon after radiotherapy we observed an increase in inflammatory cell-infiltrate, apoptosis, mucin producing goblet cells and oedema, representing the morphological expression of an unspecific reactive process with a supposed protective function. Variations of these changes have been previously observed in the clinical situation. The vast increase in goblet cells that we observed may resemble that seen in necrotising enterocolitis. A resemblance to chronic idiopathic inflammatory bowel disease, eosinophilic colitis and microscopic colitis can also be seen if the mild crypt distortion or withering that occurs with radiation injury is confused with proper crypt architectural distortion of inflammatory disease. Isolated crypts due to nuclear regenerative changes may also mimic the microadenomas of familial adenomatosis polyposis [ 30 ]. Histological changes in the pre-existing normal mucosa following preoperative radiotherapy need to be appreciated by the histopathologist if we are to avoid erroneous concurrent diagnosis [ 30 ]. Furthermore, a correct assessment of the effects of new treatment regimes or prophylaxis is based on a sound histological judgment. No differences MPO values could be seen between the controls and the radiated groups. This is probably because it is a crude method of measurement and thus may not be sensitive enough to detect early changes of inflammation. This study therefore presents a refinement of previous methods of examining effects of radiation enteropathy, and a new approach at investigating radiation induced leucocyte responses in the ileal microcirculation. This new model may be instrumental in developing strategies against pathological recruitment of leucocytes and changes in intestinal microflora in the small bowel. Competing interests None declared. Authors' contributions LBJ designed the study and participated in construction of the chamber. Performed experimental studies and drafted the manuscript. AAR performed experimental studies and drafted the manuscript. DA participated in the design of the study and construction of the chamber. Performed experimental studies, drafted the manuscript and performed the statistical analysis. LW participated in the radiological design of the study, construction of the chamber and the implementation of radiotherapy. SB participated in the radiological design of the study, chamber and the implementation of radiotherapy. CT participated in the implementation of radiotherapy. NO carried out bacteriological studies. VC performed the histological analysis. HT assisted with issues related to intravital microscopy. BJ conceived of the design, participation in construction of the chamber, co-ordination of the study as well as supervision and draft of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522820.xml
544541	Accessing the Microscopic World	The Exploratorium in San Francisco offers museum visitors the opportunity to use and manipulate state-of-the-art microscopes to visualize an array of living specimens	The Exploratorium, based in San Francisco, is a “hands on” science museum filled with interactive science and art exhibits, as well as a laboratory for the research and development of innovations in science education. In the summer of 2004, the Exploratorium launched the most ambitious microscope facility ever created for general public use. This initial phase of the project gives visitors the ability to image living specimens, as well as control the instruments themselves. Visitors can select among various specimens, move over them, change the magnification and focus, and, where appropriate, change the lighting to illuminate through the specimen, or use reflected light and fluorescence to dramatically change how it looks. They can image and explore tiny zebrafish embryos from the first stages of development to two-day-old fry with beating hearts and circulating blood cells, as well as a host of other organisms and cells from crawling amoebas to human blood cells. Below the surface, all living things share common features. The primary goals of this facility are to open a door on the wonder of the microscopic world to a diverse range of museum visitors and allow them to explore it, and to allow them to make connections to science and biomedical research. By empowering visitors with the instruments to explore this unfamiliar universe, the Exploratorium seeks to recreate some of the excitement and wonder that the earliest researchers found as they discovered another world all around them ( Box 1 ). Box 1. History of Light Microscope The light microscope falls amongst the greatest inventions of human history. Images from it in the 17th century literally revolutionized our understanding of life, providing first-hand evidence of a previously unseen or unsuspected world of organisms and cells all around us. This knowledge profoundly shaped our view of life, and of our placement in the universe. Robert Hooke used a primitive early microscope to see the walls between cells in a piece of cork (essentially discovering the cellular nature of all life), and Anton van Leeuwenhoek's simple scope revealed a previously unknown world of microorganisms living inside his own mouth. Swimming sperm were observed in semen, changing our fundamental understanding of conception. Since that time our world has become populated by marvelously beautiful and intriguing images and movies created by scientists using precise lighting and optics. Most recently, computer-controlled image-capturing techniques and digital technologies capture events and processes too small, slow, or fast for our unaided eyes to see. Van Leeuwenhoek's first microscopes were probably about as powerful as a simple water-drop scope (see Box 2 ); he used his simple devices to make observations that weren't confirmed for over a hundred years. It's quite possible that early microscope-makers were inspired to make small domed lenses by observing the magnifying properties of a rounded drop of water. Further experimentation with the sizes and shapes of lenses eventually led to much greater magnifications. Beyond the Armored Microscope Bringing a first-hand, high-quality microscope experience to a wide range of visitors and students has been a major challenge for many science museums and classrooms. In the expanded, frenetic classroom of the science museum exhibition floor, microscopes themselves are too delicate and precise for operation without assistance and supervision: the optics can be easily damaged, all but the most robust of specimens are easily destroyed, and the user interface with lighting, positioning, and focus compose a world unfamiliar to most novice users. Over the years, we had collected our share of failed and broken microscopes in various projects. For these reasons most microscopes in museums have been armored, stripped-down, single-magnification instruments. Problems related to the operation of the microscope hardware, however, represent only some of the challenges created by bringing uninitiated visitors to the microscopic world. The experience has to be repeatable. It has to be engaging. It has to be both simple and complex. It needs to work for individuals and groups. It needs to support investigation as well as provide essential information. An array of research scientists joined us as we began work to redefine and renew our existing presentation on biology, most notably Christian Sardet, a cell biologist with a passion for visualization imagery. During the summer and fall of 1999, we engaged in a series of conversations directed at defining and developing a publicly accessible microscope imaging station, and the more we explored, the more technically feasible the project became. At the same time, research-grade microscopes with high-quality optics underwent a revolution of their own. Many manufacturers planned to introduce fully automated, computer-controllable microscopes in their next models, and the prospects of having an off-the-shelf remotely controllable microscope seemed to solve a major hardware problem. It seemed as if software programs might be used to address many other user interface issues. In 2000, we applied for and received funding from the National Institutes of Health (a Science Education Partnership Award from the National Center for Research Resources) and from the David and Lucile Packard Foundation to support the development of a microscope imaging facility. Seven months later we obtained the first microscope and its associated imaging equipment. But the software to control the microscopes proved dauntingly problematic. Programs that worked well in the laboratory proved difficult to adapt for our educational uses. Instead, we wrote our own programs to control stage movement (for specimen positioning), focus, specimen selection, magnification, and lighting. Each of these controls needed to be selectable and limited so that visitors might operate the microscope within a defined range. On a parallel track, with the generous help of numerous biomedical researchers and their laboratories across the United States, we explored specimens for their visitor attraction and interest. Not surprisingly, we found that familiar structures and organisms provided the best entry points. A zebrafish embryo with a beating heart, circulating blood, and twitching tail movements rated more popular than zebrafish or sea urchin embryos at earlier stages of development. The tiny transparent roundworm Caenorhabditis elegans attracted attention with its earthworm-like movements. Overall, these biomedically relevant specimens provided a treasure trove of potential educational activities for visitors. Given a range of wonderfully attractive specimens and potential activities, we then aimed to create an educational experience that combined the observable features in the specimens, supporting information, and further activities to be chosen as desired by the visitor. To achieve this, we adopted a multimedia approach ( Figure 1 ). The incorporation of multimedia into the user–microscope interface required a complex piece of technology: the melding of live video imagery from the microscope with an interactive multimedia touchscreen, where selected microscope controls and specimen information changed as needed. To create this information presentation device, which we fondly refer to as the Usercart, we decided to use two side-by-side monitors with physical control devices for specimen position, focus, and variable magnification. Subsequent evaluation shows that the Usercarts work well; visitors get it! Figure 1 The Microscope Imaging Station Plasma screens, visitor microscope control consoles, and interactive media presentations provide visitors with access to live specimens and biomedical information Visitors can pick and choose specimens of their choice, refine their selection, change magnification, and engage in suggested activities at their own pace with integrated appropriate information. Individuals and groups engage with the microscope imagery and companion information for relatively long periods of time, ranging from several minutes to upwards of 20 minutes, a long time for a museum exhibit. To expand the use of the microscope imagery beyond that of one or two visitors, we incorporated large plasma-screen presentations in to the Microscope Imaging Station facility. What Else Is New? In addition to using some of the latest devices in microscope technologies, the Imaging Station also provides a window on revolutionary research techniques. For example, in 1994, Columbia University's Martin Chalfie inserted a gene for a fluorescent jellyfish protein into a bacterium ( Escherichia coli ) and a roundworm ( C. elegans ) and found that the genetically modified organisms emitted an eerie green glow under certain conditions. Other scientists built on this technique to create a powerful tool that makes hidden structures and processes easier to study. The coupling of vital staining with the green fluorescent protein gene allows scientists to observe events inside developing cells and detect the presence of diseased structures and environmental toxins with extreme sensitivity. This newly created sensitivity has sparked new insights and discoveries, re-revolutionizing the capabilities of the light microscope. At the Imaging Station, visitors can take advantage of this technique to observe specific types of cells—such as the brain cells, sex cells, or muscle cells of roundworms. Visitors can also peer inside the developing embryo of a tropical zebrafish whose circulatory cells have been made visible by the protein made from the transplanted green fluorescent protein gene. In the immediate future, visitors will be able to closely examine living human blood cells and fruit flies that researchers use to study the genetics of a wide range of human disorders. Over the next year, we intend to add a major component on mouse stem cells and the process of differentiation. At the Imaging Station, visitors have access to seamless video footage of events that take place in a fraction of a second or occur slowly over weeks or months ( Box 1 ). Using the latest technologies available, we have built operating software specifically for the Imaging Station that gives visitors the same kind of control that professional researchers have over their own work. The straightforward user interface is integrated with explanatory graphics to provide control and orientation simultaneously. Finally, images are brought to users on large, high-resolution video screens. But the best may be yet to come, because the Imaging Station's capabilities are continually expanding, and visitors will ultimately be able to observe specimens by logging on to the Station's Web site ( http://www.exploratorium.edu/imaging_station ). Supporting Information Video S1 A Medley of Time-Lapse Videos Collected at the Microscope Imaging Station Many of the cells and organisms found in these movies are available via visitor-accessible microscopes on a daily basis. At the station, visitor-directed, microscope-based observations form the basis for informal education on basic and biomedically relevant research. Video microscopy and production by Kristina Yu. (5.4 KB MOV). Click here for additional data file. Box 2. How to Make a Simple Microscope from a Drop of Water Materials. You'll need the clear plastic part of a CD case; if you can't find one, any thin sheet of rigid, clear plastic will work. Remove any paper or packaging so that you can see through the plastic. (Take the CD out, too!) What to do. (1) Put the object you want to examine—a bug, a tiny leaf or plant, the small lettering on a coin or newspaper—on a flat surface. (2) Put one large drop of water on the center of the clear plastic. (An eyedropper makes this easy.) (3) Pick up the plastic and hold it horizontally so that the water drop is directly over the object. Move the plastic slowly back and forth to center the object, and up and down to focus. The water will form a curved shape that magnifies the object!	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544541.xml
544596	Magnetic techniques for the isolation and purification of proteins and peptides	Isolation and separation of specific molecules is used in almost all areas of biosciences and biotechnology. Diverse procedures can be used to achieve this goal. Recently, increased attention has been paid to the development and application of magnetic separation techniques, which employ small magnetic particles. The purpose of this review paper is to summarize various methodologies, strategies and materials which can be used for the isolation and purification of target proteins and peptides with the help of magnetic field. An extensive list of realised purification procedures documents the efficiency of magnetic separation techniques.	Introduction Isolation, separation and purification of various types of proteins and peptides, as well as of other specific molecules, is used in almost all branches of biosciences and biotechnologies. Separation science and technology is thus very important area necessary for further developments in bio-oriented research and technology. New separation techniques, capable of treating dilute solutions or solutions containing only minute amounts of target molecules in the presence of vast amounts of accompanying compounds in both small and large-scale processes, even in the presence of particulate matter, are necessary. In the area of biosciences and biotechnology the isolation of proteins and peptides is usually performed using variety of chromatography, electrophoretic, ultrafiltration, precipitation and other procedures, affinity chromatography being one of the most important techniques. Affinity ligand techniques represent currently the most powerful tool available to the downstream processing both in term of their selectivity and recovery. The strength of column affinity chromatography has been shown in thousands of successful applications, especially in the laboratory scale. However, the disadvantage of all standard column liquid chromatography procedures is the impossibility of the standard column systems to cope with the samples containing particulate material so they are not suitable for work in early stages of the isolation/purification process where suspended solid and fouling components are present in the sample. In this case magnetic affinity, ion-exchange, hydrophobic or adsorption batch separation processes, applications of magnetically stabilized fluidized beds or magnetically modified two-phase systems have shown their usefulness. The basic principle of batch magnetic separation is very simple. Magnetic carriers bearing an immobilized affinity or hydrophobic ligand or ion-exchange groups, or magnetic biopolymer particles having affinity to the isolated structure, are mixed with a sample containing target compound(s). Samples may be crude cell lysates, whole blood, plasma, ascites fluid, milk, whey, urine, cultivation media, wastes from food and fermentation industry and many others. Following an incubation period when the target compound(s) bind to the magnetic particles the whole magnetic complex is easily and rapidly removed from the sample using an appropriate magnetic separator. After washing out the contaminants, the isolated target compound(s) can be eluted and used for further work. Magnetic separation techniques have several advantages in comparison with standard separation procedures. This process is usually very simple, with only a few handling steps. All the steps of the purification procedure can take place in one single test tube or another vessel. There is no need for expensive liquid chromatography systems, centrifuges, filters or other equipment. The separation process can be performed directly in crude samples containing suspended solid material. In some cases (e.g., isolation of intracellular proteins) it is even possible to integrate the disintegration and separation steps and thus shorten the total separation time [ 1 ]. Due to the magnetic properties of magnetic adsorbents (and diamagnetic properties of majority of the contaminating molecules and particles present in the treated sample), they can be relatively easily and selectively removed from the sample. In fact, magnetic separation is the only feasible method for recovery of small magnetic particles (diameter ca 0.1 – 1 μm) in the presence of biological debris and other fouling material of similar size. Moreover, the power and efficiency of magnetic separation procedures is especially useful at large-scale operations. The magnetic separation techniques are also the basis of various automated procedures, especially magnetic-particle based immunoassay systems for the determination of a variety of analytes, among them proteins and peptides. Several automated systems for the separation of proteins or nucleic acids have become available recently. Magnetic separation is usually very gentle to the target proteins or peptides. Even large protein complexes that tend to be broken up by traditional column chromatography techniques may remain intact when using the very gentle magnetic separation procedure [ 2 ]. Both the reduced shearing forces and the higher protein concentration throughout the isolation process positively influence the separation process. Separation of target proteins using standard chromatography techniques often leads to the large volume of diluted protein solution. In this case appropriate magnetic particles can be used for their concentration instead of ultrafiltration, precipitation etc. [ 3 ]. The purpose of this review is to summarize various methodologies and strategies which can be employed for the isolation and purification of target proteins and peptides with the help of magnetic materials. An extensive list of realised purification procedures documents the efficiency of magnetic separation techniques. All these information will help the scientists to select the optimal magnetic material and the purification procedure. Necessary materials and equipment The basic equipment for laboratory experiments is very simple. Magnetic carriers with immobilized affinity or hydrophobic ligands, magnetic particles prepared from a biopolymer exhibiting affinity for the target compound(s) or magnetic ion-exchangers are usually used to perform the isolation procedure. Magnetic separators of different types can be used for magnetic separations, but many times cheap strong permanent magnets are equally efficient, especially in preliminary experiments. Magnetic carriers and adsorbents can be either prepared in the laboratory, or commercially available ones can be used. Such carriers are usually available in the form of magnetic particles prepared from various synthetic polymers, biopolymers or porous glass, or magnetic particles based on the inorganic magnetic materials such as surface modified magnetite can be used. Many of the particles behave like superparamagnetic ones responding to an external magnetic field, but not interacting themselves in the absence of magnetic field. This is important due to the fact that magnetic particles can be easily resuspended and remain in suspension for a long time. In most cases, the diameter of the particles differs from ca 50 nm to approx. 10 μm. However, also larger magnetic affinity particles, with the diameters up to millimetre range, have been successfully used [ 4 ]. Magnetic particles having the diameter larger than ca 1 μm can be easily separated using simple magnetic separators, while separation of smaller particles (magnetic colloids with the particle size ranging between tens and hundreds of nanometers) may require the usage of high gradient magnetic separators. Commercially available magnetic particles can be obtained from a variety of companies. In most cases polystyrene is used as a polymer matrix, but carriers based on cellulose, agarose, silica, porous glass or silanized magnetic particles are also available. Examples of magnetic particles used (or usable) for proteins and peptides separation can be found elsewhere [ 5 - 7 ]. Particles with immobilised affinity ligands are available for magnetic affinity adsorption. Streptavidin, antibodies, protein A and Protein G are used most often in the course of protein and peptides isolation. Magnetic particles with above mentioned immobilised ligands can also serve as generic solid phases to which native or modified affinity ligands can be immobilised (e.g., antibodies in the case of immobilised protein A, protein G or secondary antibodies, biotinylated molecules in the case of immobilised streptavidin). Also some other affinity ligands (e.g., nitrilotriacetic acid, glutathione, trypsin, trypsin inhibitor, gelatine etc.) are already immobilised to commercially available carriers. To immobilise other ligands of interest to both commercial and laboratory made magnetic particles standard procedures used in affinity chromatography can be employed. Usually functional groups available on the surface of magnetic particles such as -COOH, -OH or -NH 2 are used for immobilisation, in some cases magnetic particles are available already in the activated form (e.g., tosylactivated, epoxyactivated etc). In the laboratory magnetite (or similar magnetic materials such as maghemite or ferrites) particles can be surface modified by silanization. This process modifies the surface of the inorganic particles so that appropriate functional groups become available, which enable easy immobilisation of affinity ligands [ 8 ]. In exceptional cases enzyme activity can be decreased as a result of usage of magnetic particles with exposed iron oxides. In this case encapsulated microspheres, having an outer layer of pure polymer, will be safer. Biopolymers such as agarose, chitosan, kappa carrageenan and alginate can be easily prepared in a magnetic form. In the simplest way the biopolymer solution is mixed with magnetic particles and after bulk gel formation the magnetic gel formed is mechanically broken into fine particles [ 9 ]. Alternatively biopolymer solution containing dispersed magnetite is dropped into a mixed hardening solution [ 4 ] or water-in-oil suspension technique is used to prepare spherical particles [ 10 ]. Basically the same procedures can be used to prepare magnetic particles from synthetic polymers such as polyacrylamide, poly(vinylalcohol) and many others [ 11 ]. In another approach used standard affinity or ion-exchange chromatography material was post-magnetised by interaction of the sorbent with water-based ferrofluid. Magnetic particles accumulated within the pores of chromatography adsorbent thus modifying this material into magnetic form [ 12 , 13 ]. Alternatively magnetic Sepharose or other agarose gels were prepared by simple contact with freshly precipitated or finely powdered magnetite [ 12 , 14 ]. Magnetoliposomes (magnetic derivatives of standard liposomes), either in the original form or after immobilization of specific proteins, have the potential for the separation of antiphospholipid antibodies [ 15 ], IgG antibodies [ 16 ] and other proteins of interest [ 17 ]. Recently also non-spherical magnetic structures, such as magnetic nanorods have been tested as possible adsorbent material for specific separation of target proteins [ 18 ]. Magnetic separators are necessary to separate the magnetic particles from the system. In the simplest approach, a small permanent magnet can be used, but various magnetic separators employing strong rare-earth magnets can be obtained at reasonable prices. Commercial laboratory scale batch magnetic separators are usually made from magnets embedded in disinfectant-proof material. The racks are constructed for separations in Eppendorf micro-tubes, standard test tubes or centrifugation cuvettes, some of them have a removable magnetic plate to facilitate easy washing of separated magnetic particles. Other types of separators enable separations from the wells of microtitration plates and the flat magnetic separators are useful for separation from larger volumes of suspensions (up to approx. 500 – 1000 ml). Examples of typical batch magnetic separators are shown in Fig. 1 . Figure 1 Examples of batch magnetic separators applicable for magnetic separation of proteins and peptides. A: Dynal MPC -S for six microtubes (Dynal, Norway); B: Dynal MPC – 1 for one test tube (Dynal, Norway); C: Dynal MPC – L for six test tubes (Dynal, Norway); D: magnetic separator for six Eppendorf tubes (New England BioLabs, USA); E: MagneSphere Technology Magnetic Separation Stand, two position (Promega, USA); F: MagnaBot Large Volume Magnetic Separation Device (Promega, USA); G: MagneSphere Technology Magnetic Separation Stand, twelve-position (Promega, USA); H: Dynal MPC – 96 S for 96-well microtitre plates (Dynal, Norway); I: MagnaBot 96 Magnetic Separation Device for 96-well microtitre plates (Promega, USA); J: BioMag Solo-Sep Microcentrifuge Tube Separator (Polysciences, USA); K: BioMag Flask Separator (Polysciences, USA); L: MagneSil Magnetic Separation Unit (Promega, USA); M: MCB 1200 processing system for 12 microtubes based on MixSep process (Sigris Research, USA); N: PickPen magnetic tool (Bio-Nobile, Finland). Reproduced with the permission of the above mentioned companies; the photos were taken from their www pages. Flow-through magnetic separators are usually more expensive, and high gradient magnetic separators (HGMS) are the typical examples. Laboratory scale HGMS is composed from a column packed with fine magnetic grade stainless steel wool or small steel balls which is placed between the poles of an appropriate magnet. The suspension is pumped through the column, and magnetic particles are retained within the matrix. After removal the column from the magnetic field, the particles are retrieved by flow and usually by gentle vibration of the column. For work in dense suspensions, open gradient magnetic separators may be useful. A very simple experimental set-up for the separation of magnetic affinity adsorbents from litre volumes of suspensions was described [ 19 ]. Currently many projects require the analysis of a high number of individual proteins or variants. Therefore, methods are required that allows multiparallel processing of different proteins. There are several multiple systems for high throughput nucleic acid and proteins preparation commercially available. The most often used approach for proteins isolation is based on the isolation and assay of 6xHis-tagged recombinant proteins using magnetic beads with Ni-nitriloacetic acid ligand [ 20 ]. The commercially available platforms can be obtained from several companies such as Qiagen, USA (BioRobot and BioSprint series), Tecan, Japan (Te-MagS) or Thermo Electron Corporation, USA (KingFisher). Basic principles of magnetic separations of proteins and peptides Magnetic separations of proteins and peptides are usually convenient and rapid. Nevertheless, several hints may be helpful to obtain good results. Proteins and peptides in the free form can be directly isolated from different sources. Membrane bound proteins have to be usually solubilized using appropriate detergents. When nuclei are broken during sample preparation, DNA released into the lysate make the sample very viscous. This DNA may be sheared by repeated passage up and down through a 21 gauge hypodermic syringe needle before isolation of a target protein. Alternatively, DNase can be added to enzymatically digest the DNA. Magnetic beads in many cases exhibit low non-specific binding of non-target molecules present in different samples. Certain samples may still require preclearing to remove molecules which have high non-specific binding activity. If preclearing is needed, the sample can be mixed with magnetic beads not coated with the affinity ligand. In the case of immunomagnetic separation, magnetic beads coated with secondary antibody or with irrelevant antibodies have been used. The non-specific binding can also be minimised by adding a non-ionic detergent both in the sample and in the washing buffers after isolation of the target. In general, magnetic affinity separations can be performed in two different modes. In the direct method, an appropriate affinity ligand is directly coupled to the magnetic particles or biopolymer exhibiting the affinity towards target compound(s) is used in the course of preparation of magnetic affinity particles. These particles are added to the sample and target compounds then bind to them. In the indirect method the free affinity ligand (in most cases an appropriate antibody) is added to the solution or suspension to enable the interaction with the target compound. The resulting complex is then captured by appropriate magnetic particles. In case antibodies are used as free affinity ligands, magnetic particles with immobilised secondary antibodies, protein A or protein G are used for capturing of the complex. Alternatively the free affinity ligands can be biotinylated and magnetic particles with immobilised streptavidin or avidin are used to capture the complexes formed. In both methods, magnetic particles with isolated target compound(s) are magnetically separated and then a series of washing steps is performed to remove majority of contaminating compounds and particles. The target compounds are then usually eluted, but for specific applications (especially in molecular biology, bioanalytical chemistry or environmental chemistry) they can be used still attached to the particles, such as in the case of polymerase chain reaction, magnetic ELISA etc. The two methods perform equally well, but, in general, the direct technique is more controllable. The indirect procedure may perform better if affinity ligands have poor affinity for the target compound. In most cases, magnetic batch adsorption is used to perform the separation step. This approach represents the simplest procedure available, enabling to perform the whole separation in one test-tube or flask. If larger magnetic particles (with diameters above ca 1 μm) are used, simple magnetic separators can be employed. In case magnetic colloids (diameters ranging between tens and hundreds of nanometres) are used as affinity adsorbents, high-gradient magnetic separators have usually to be used to remove the magnetic particles from the system. Alternatively magnetically stabilised fluidised beds (MSFB), which enable a continuous separation process, can be used. The use of MSFB is an alternative to conventional column operation, such as packed-bed or fluidised bed, especially for large-scale purification of biological products. Magnetic stabilisation enables the expansion of a packed bed without mixing of solid particles. High column efficiency, low pressure drop and elimination of clogging can be reached [ 21 , 22 ]. Also non-magnetic chromatographic adsorbents can be stabilized in magnetically stabilized fluidized beds if sufficient amount of magnetically susceptible particles is also present. The minimum amount of magnetic particles necessary to stabilize the bed is a function of various parameters including the size and density of both particles, the magnetic field strength, and the fluidization velocity. A variety of commercially available affinity, ion-exchange, and adsorptive supports can be used in the bed for continuous separations [ 23 ]. Biocompatible two phase systems, composed for example from dextran and polyethylene glycol, are often used for isolation of biologically active compounds, subcellular organelles and cells. One of the disadvantages of this system is the slow separation of the phases when large amounts of proteins and cellular components are present. The separation of the phases can be accelerated by the addition of fine magnetic particles or ferrofluids to the system followed by the application of a magnetic field. This method seems to be useful when the two phases have very similar densities, the volumetric ratio between the phases is very high or low, or the systems are viscous. Magnetically enhanced phase separation usually increases the speed of phase separation by a factor of about 10 in well-behaved systems, but it may increase by a factor of many thousands in difficult systems. The addition of ferrofluids and/or iron oxide particles was shown to have usually no influence on enzyme partioning or enzyme activity [ 24 , 25 ]. Proteins and peptides isolated using magnetic techniques have to be usually eluted from the magnetic separation materials. In most cases bound proteins and peptides can be submitted to standard elution methods such as the change of pH, change of ionic strength, use of polarity reducing agents (e.g., dioxane or ethyleneglycol) or the use of deforming eluents containing chaotropic salts. Affinity elution (e.g., elution of glycoproteins from lectin coated magnetic beads by the addition of free sugar) may be both a very efficient and gentle procedure. Examples of magnetic separations of proteins and peptides Magnetic affinity and ion-exchange separations have been successfully used in various areas, such as molecular biology, biochemistry, immunochemistry, enzymology, analytical chemistry, environmental chemistry etc [ 26 - 29 ]. Tables 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 show some selected applications of these techniques for proteins and peptides isolation. Table 1 Examples of proteinases and peptidases purified by magnetic techniques Purified enzyme Source Magnetic carrier Affinity ligand Further details Reference Aminopeptidase Arabidopsis Amine-terminated magnetic beads N-1-Naphthylphthalamic acid KCl gradient elution [54] Angiotensin-converting enzyme Pig lung membranes Dynabeads Polyclonal antibodies [57] Bromelain Commercial preparation Polyacrylic acid – iron oxide magnetic nanoparticles Elution with KCl solution [59] Caspase (recombinant, histidine-tailed) Human cells Magnetic agarose Ni-NTA Elution with SDS-PAGE buffer [60] Chymotrypsin Commercial preparation Magnetic chitosan beads Elution with N-acetyl-D-tryptophan [62] NIa-protease (recombinant, histidine-tagged) Plum Pox Virus Magnetic core and nickel-silica composite matrix Ni 2+ Elution with imidazole containing buffer [36] Proteinases Commercial sources Magnetic cross-linked erythrocytes Elution with low pH buffer [46] Proteinase, bacterial (Savinase) Bacillus clausii Silanized magnetite particles Bacitracin [84] Trypsin Porcine pancreatin Silanized magnetite particles p-Aminobenzamidine Elution with low pH buffer [50] Porcine pancreatin Magnetic polymer particles Soybean trypsin inhibitor Elution with low pH solution [86] Commercial preparation Silanized ferrite powder Soybean trypsin inhibitor [87] Commercial preparation Magnetic κ-carrageenan particles Soybean trypsin inhibitor Elution with low pH solution, MSFB [88][89] Commercial preparation Magnetic polyacrylamide beads Soybean trypsin inhibitor Magnetically stabilized fluidized beds [90] Commercial preparation Magnetic chitosan particles Aprotinin Elution with low pH solution [91] Commercial preparation Magnetic cross-linked erythrocytes Elution with low pH buffer; separation from large volume sample [19] Urokinase Crude urokinase preparation Magnetic dextran, agarose, polyvinyl alcohol, polyhydroxyethyl methacrylate microspheres p-Aminobenzamide, L-arginine methyl ester, guanidine hexanoic acid or guanidine acetic acid [93] Table 2 Purification of lysozyme by magnetic techniques Purified enzyme Source Magnetic carrier Affinity ligand Further details Reference Lysozyme Hen egg white Magnetic chitin Elution with 0.01 M HCl [71] Hen egg white Magnetic acetylated chitosan Elution with 0.01 M HCl [9] Commercial preparation Magnetic poly(2-hydroxyethyl methacrylate) Cibacron Blue F3GA Elution with 1 M KSCN [72] Magnetic chitosan beads Magnetically stabilized fluidized bed [73] Ornithodoros moubata Magnetic chitin Elution with alkaline, high salt buffer [74] Commercial preparation Magnetic cross-linked polyvinylalcohol Cibacron blue 3GA Elution with high salt buffer [52] Magnetite – polyacrylic acid nanoparticles Ion-exchange separation [75] Magnetic cross-linked polyvinylalcohol beads Adsorption study [76] Commercial preparation Magnetic agarose beads Cibacron blue 3GA Magnetically stabilized fluidized bed [77] Magnetic chitosan Cibacron blue 3GA Study of adsorption properties [78] Commercial preparation Ferrofluid modified sawdust Elution with 0.5 M NaCl [79] Commercial preparation Nano-sized magnetic particles Elution with NaH 2 PO 4 and NaSCN [80] Lysozyme (recombinant, histidine-tailed) T4 BioMag, amine terminated Iminodiacetic acid charged with Cu 2+ Elution with low pH buffer [81] Table 3 Examples of polysaccharide and disaccharide hydrolases purified by magnetic techniques Purified enzyme Source Magnetic carrier Affinity ligand Further details Reference α-Amylases Porcine pancreas, Bacillus subtilis , wheat germ Magnetic alginate beads Elution with 1 M maltose [4] Bacillus amyloliquefaciens , porcine pancreas Magnetic alginate microbeads Elution with 1 M maltose [10] β-Amylase Sweet potato Magnetic alginate beads Elution with 1 M maltose [55] β-Galactosidase Escherichia coli homogenate Silanized magnetite p-Aminophenyl-β-D-thiogalactopyranoside Elution with borate buffer, pH 10 [58] β-Galactosidase (fusion protein comprising the DNA-binding lac repressor) Bacterial lysate Magnetic beads DNA containing Escherichia coli lac operator Elution with lactose analogue [64] Glucoamylase Aspergillus niger Magnetic alginate beads Elution with 1 M maltose [55] Pectinase Commercial preparation Magnetic alginate beads [82] Pullulanase Bacillus acidopullulyticus Magnetic alginate beads Elution with 1 M maltose [55] Table 4 Examples of other enzymes purified by magnetic techniques Purified enzyme Source Magnetic carrier Affinity ligand Further details Reference Alcohol dehydrogenase Yeast homogenate Magnetic cross-linked polyvinylalcohol Cibacron blue 3GA Elution with high salt buffer [52] Saccharomyces cerevisiae extract PEG with bound Cibacron blue Magnetic two-phase system [53] Aldolase (recombinant, histidine tagged) Pea Magnetic core and nickel-silica composite matrix Ni 2+ Elution with imidazole containing buffer [36] AngioI-TEM-β-lactamase Escherichia coli cells extracts Magnetic agarose beads Iminodiacetic acid charged with Zn 2+ Elution with low pH buffer [56] Asparaginase Escherichia coli homogenate Magnetic polyacrylamide gel particles D-Asparagine Elution with D-asparagine solution [58] Carbonic anhydrase Model mixture Magnetic agarose beads Sulfanilimide Elution with high salt buffer [14] Catalase Bovine liver, commercial preparation Magnetic poly(EGDMA-MAH) beads Fe 3+ Elution with NaSCN solution [61] Cytochrome c Horse, Candida krusei Amine terminated iron oxide particles Iminodiacetic acid charged with Cu 2+ Binding studies [63] Commercial preparation Au@magnetic particles MALDI MS analysis [31] Horse heart Magnetic agarose beads Iminodiacetic acid charged with Cu 2+ Elution with EDTA containing buffer [56] Bovine heart Magnetic ion-exchange particles Protein binding studies [12] Glucose-6-phosphate dehydrogenase Ferrofluid modified Sepharose 4B ADP [65] Saccharomyces cerevisiae extract PEG with bound Cibacron blue Magnetic two-phase system [53] Hexokinase Escherichia coli homogenate PEG with bound Cibacron blue Magnetic two-phase system [53] Lactate dehydrogenase Beef heart Ferrofluid modified Sepharose 4B AMP Elution with 1 mM NADH [13] Porcine muscle Magnetic agarose beads Reactive Red 120 Column elution with NaCl gradient [66] Lactoperoxidase Sweet whey Magnetic cation exchanger HGMS [67,68] Luciferase (histidine-tagged) Escherichia coli homogenate MagneHis™ system Ni 2+ [69,70] Phosphatase, alkaline Human placenta Dynabeads M-450 Specific antibody Activity of bound enzyme measured [83] Phosphatase, alkaline (fusion protein comprising the DNA-binding lac repressor) Bacterial lysate Magnetic beads DNA containing Escherichia coli lac operator Elution with lactose analogue [64] Phosphofructokinase Saccharomyces cerevisiae extract PEG with bound Cibacron blue Magnetic two-phase system [53] 6-Phosphogluconate dehydrogenase Ferrofluid modified Sepharose 4B ADP Elution with 1 mM NADP [13] Thioredoxin (recombinant, histidine-tagged) Escherichia coli Magnetic agarose Ni-NTA Elution with imidazole containing buffer [20] tRNA methionyl synthetase (recombinant, histidine-tagged) Escherichia coli MagneHis™ system Ni 2+ Rapid detection and quantitation of isolated protein [85] Uricase (recombinant, histidine-tailed) Bacillus Ion-chelating magnetic agarose beads Ni 2+ Elution by cleavage with proteinase K [92] Table 5 Examples of antibodies purified by magnetic techniques Purified antibody Source Magnetic carrier Affinity ligand Further details Reference Anti-BODIPY-fluorescein antibodies Magnetoliposomes BODIPY-fluorescein [94] Anti-DNA antibody Systemic lupus erythematosus patient plasma Magnetic poly(2-hydroxyethyl-methacrylate) beads DNA Desorption with 1 M NaSCN solution [95] Anti-human chorionic gonadotropin antibody Murine ascites supernatants Magnetic cellulose beads Human chorionic gonadotropin [96] Antibody (from rat) Sample from affinity chromatography Dynabeads M-280 Sheep anti-rabbit IgG Antibody concentration [3] Antibody Rabbit serum Dynabeads M-280 Sheep anti-rabbit IgG Elution with 0.5 M acetic acid [97] Monoclonal antibodies Mouse hybridoma culture broth Magnetite particles Protein A [98] Anti-bovine serum albumin antibodies Thermosensitive magnetic microspheres Bovine serum albumin Immobilization by the carbodiimide method [99] Immunoglobulin G, human Commercial preparation Magnetic poly(ethylene glycol dimethacrylate-N-methacryloly-L-histidine-methylester) beads Elution with 1 M NaCl [100] Immunoglobulin G Blood serum Carboxyl-terminated magnetic particles MproteinAG [101] IgE antibodies Allergic patients sera Magnetoliposomes Antigenic proteins [16] Murine anti-fibroblast growth factor receptor 1 IgM Ascites Polystyrene magnetic beads Rat anti-mouse IgM monoclonal antibody [102] Table 6 Examples of DNA/RNA/oligonucleotide/aptamer binding proteins purified by magnetic techniques Purified protein Source Magnetic carrier Affinity ligand Further details Reference CUG binding proteins Human myoblasts or fibroblasts Dynabeads M-280 streptavidin Biotinylated(CUG) 10 Elution with 1 M NaCl [103] Transcription factor τ Saccharomyces cerevisiae Dynabeads M-280 streptavidin Biotinylated tRNA Glu gene fragment Elution with high salt buffer [104,105] DNA-binding proteins Crude tissue extract Magnetic phospho cellulose particles [106] DNA-binding proteins Escherichia coli Magnetic phospho cellulose particles [107] DNA-binding proteins HeLa nuclear extracts Dynabeads M-280 streptavidin Biotin-labelled DNA fragment Elution with 2 M NaCl [108] Vaccinia virus early transcription factor Vaccinia virions Dynabeads M-280 streptavidin Biotinylated double-stranded DNA Elution with high salt buffer [109] Ecdysteroid receptor Drosophila melanogaster nuclear extract Dynabeads M-280 streptavidin Biotinylated double-stranded oligonucleotide Elution with 0.4 M KCl [110] NanR protein (recombinant) Escherichia coli μMACS streptavidin MicroBeads Biotin-labelled DNA fragment Elution with 1 M NaCl [111] p27 Rabbit hepatocytes Dynabeads M-280 streptavidin Guanine-rich single-stranded DNA Elution with NaCl solution [112] Pigpen protein Endothelial cells Magnetic streptavidin beads Biotinylated aptamer Elution with 1 M NaCl [113] RNA binding proteins Saccharomyces cerevisiae μMACS streptavidin MicroBeads Biotin-labelled RNA probe Elution with 1 M NaCl [114] Single-stranded telomere binding protein (sTBP) Nuclei from vertebrate tissues Dynabeads M-280 streptavidin Biotinylated single stranded TTAGGG n repeats Elution with high salt buffer [115] Transcription proteins Human myeloid cells Dynabeads M-280 streptavidin Biotinylated serum inducible element (hSIE) Elution with high salt buffer [116] Transcription factor γRF-1 Human monocytes and epidermal cells Dynabeads M-280 streptavidin Biotinylated DNA containing γRF-1 sequences Elution with 0.6 M KCl [117] Protein factor MS2 Murine skeletal myotubes Dynabeads Double-stranded DNA Elution with 100 mM sodium acetate, pH 4.2 [118] Guide RNA binding protein Trypanosoma brucei mitochondria Dynabeads M-450 goat anti-mouse IgG Monoclonal antibody Elution with low pH buffer cont. SDS [119] RNA binding proteins Pollen grains Streptavidin MagneSphere particles Biotinylated oligonucleotides Elution with SDS buffer [120] DNA binding protein Schistosoma mansoni Dynabeads M-280 streptavidin Biotinylated DNA Elution with sodium acetate buffer [121] ssDNA binding proteins Transfected mouse fibroblasts Dynabeads anti-rabbit IgG Rabbit antibody Indirect method [122] Tenascin-C Glioblastoma cells Dynabeads streptavidin Biotinylated aptamer Elution with high salt buffer [123] Thermostable brain factor (ThBF) Rat brain Streptavidin magnetic particles Biotinylated oligonucleotides Elution with 0.7 M KCl [124] TTF1 protein Escherichia coli lysate Dynabeads M-280 streptavidin Biotinylated aptamer Elution with DNase [125] Table 7 Purification of albumin and haemoglobin by magnetic techniques Purified protein Source Magnetic carrier Affinity ligand Further details Reference Albumin, bovine serum Commercial preparation Magnetic agar beads Cibacron blue3GA Adsorption experiments [126] Commercial preparation Magnetic cross-linked polyvinylalcohol Cibacron blue3GA Adsorption experiments [76,127] Magnetic chitosan microspheres Cibacron blue3GA [78] Commercial preparation Magnetic poly(glycidyl methacrylate-triallyl isocyanurate-divinylbenzene) particles Anion exchange separation [128] Commercial preparation Magnetic poly(ethylene glycol dimethacrylate-co-N-methacryloyl-(L)-histidine methyl ester) microbeads Cu 2+ Elution with 1.0 M NaSCN [129] Albumin, human serum Commercial preparation Magnetic poly(2-hydroxyethylmethacrylate) beads Iminodiacetic acid charged with Cu 2+ Elution with 1.0 M NaSCN [130] Human plasma Magnetic poly(2-hydroxyethyl methacrylate) beads Cibacron blue F3GA Elution with 0.5 M NaSCN [131] Commercial preparation Magnetic particles covered with thermosensitive polymer - Desorption by decreasing temperature [132,133] Albumin, human serum (recombinant, FLAG tagged) Yeast cells Magnetic glass beads Anti-FLAG antibody Elution with EDTA containing buffer [1] Glycated haemoglobin Human blood Magnetic poly(vinyl alcohol) beads m-Aminophenyl-boronic acid Elution with sorbitol [138] Haemoglobin Bovine, commercial preparation Amine terminated iron oxide particles Iminodiacetic acid charged with Cu 2+ Elution with imidazole containing buffer [63] Haemoglobin A1c Human blood Magnetic particles isolated from Magnetospirillum magneticum AMB-1 m -Aminophenyl-boronic acid used for affinity immunoassay [150] Table 8 Examples of other proteins purified by magnetic techniques Purified protein Source Magnetic carrier Affinity ligand Further details Reference Aprotinin Bovine pancreatic powder Magnetic chitosan particles Trypsin Elution with low pH buffer [134] Concanavalin A Jack bean extract Magnetic particles Dextran [68,135] Solanum tuberosum lectin Potato tuber Magnetic chitosan Elution with low pH buffer [136] Green fluorescent protein (histidine tagged) Magnetic nanoparticles Ni-NTA Elution with imidazole containing buffer [137] SIRT2 protein (recombinant, histidine tailed) Human Magnetic agarose beads Ni-NTA Elution with imidazole containing buffer [139] Elongation factor (recombinant, histidine tailed) Caenorhabditis elegans Magnetic agarose beads Ni-NTA Elution with imidazole containing buffer [140] Protein A Recombinant Escherichia coli Magnetic Eudragit Human IgG Magnetic two-phase system [141] Tumor necrosis factor (TNF) Dynabeads M-280 Mouse monoclonal antibody Solid phase immunoassay [142] Anti-MUC1 diabody fragment Recombinant Escherichia coli Magnetic agarose beads Specific peptide [143] MHC class II molecules MDCK cells Dynabeads M-450 rat anti-mouse IgG1 Specific antibodies Elution with SDS-PAGE buffer [144] Lamin B 3 Xenopus egg extracts Dynabeads Specific antibodies Elution with 6 M urea [44] 6x-His-tagged proteins Human fibroblasts Magnetic agarose beads Ni-NTA Elution with imidazole containing buffer [145] Estrogen receptor Adipose tissue Dynabeads M-280 streptavidin Biotinylated monoclonal mouse anti-human estrogen receptor antibody Indirect method [146] Thiol-reactive chromatin restriction fragments Mouse fibroblasts Mercurated agarose magnetic beads p-Hydroxymercuribenzoate Elution with 0.5 M NaCl and 20 mM dithiothreitol [147] L1 coat protein Human papillomavirus Magnetic polyglutaraldehyde particles Iminodiacetic acid charged with Cu 2+ Elution with imidazole containing buffer [41] Insulin receptor Rat muscle or liver extract Dynabeads M-450 Anti-P5 antibody SDS PAGE analysis [148] Stat3 DER cells Dynabeads Biotinylated tyrosine phosphorylated peptides SDS PAGE analysis [149] Transferrin receptor Human Dynabeads M-450 sheep anti-mouse IgG1 Anti-human transferrin receptor monoclonal antibody SDS analysis [151] Prion protein PrP Sc Brain extract Dynabeads M-280 tosyl activated Plasminogen SDS analysis [39,40] Biotinylated proteins from extracellular matrix Bipolaris sorokiniana Dynabeads Streptavidin SDS analysis [152] Cryoprotectin Leaves of cold-acclimated cabbage ( Brassica oleracea ) Dynabeads-protein A Specific antibody [153] Prostate specific antigen Serum from a prostate cancer suffering patient Streptavidin-coated magnetic beads Biotinylated monoclonal antibody Elution with low pH solution [154,155] Estrogen receptor In vitro translation Magnetic beads Antibody Elution with SDS buffer [156] VHDL receptor Helicoverpa zea Streptavidin-coated magnetic beads VHDL-biotin ligand [157] Fructosyllysine-specific binding protein U937 cells Dynabeads M-280 tosylactivated Poly-L-lysine-glucose conjugate Two proteins isolated [158] Ubiquitin (histidine tagged) Nickel-gold nanorods Elution with acidic buffer [18] Table 9 Examples of peptides purified by magnetic techniques Purified peptide Source Magnetic carrier Affinity ligand Further details Reference Biotinylated peptides Model mixtures Dynabeads M-280 streptavidin Streptavidin Used in MALDI-TOF mass analysis [159] (His) 6 -Ala-Tyr-Gly Synthetic peptide Dynabeads M-280 tosylactivated Aminocaproic nitrilotriacetic acid charged with Ni 2+ Elution with imidazole solution [160] Synthetic pentapeptides against fructose-1,6-biphosphate aldolase Synthetic mixture Streptavidin-coated magnetic beads Biotin labelled fructose-1,6-biphosphate aldolase of T. brucei Pentapeptides were anchored on polystyrene beads [161] Tryptic digest products of cytochrome c Trypsin digested cytochrome c Au@magnetic particles - Ion-exchange separation followed by MALDI MS analysis [31] Glutathione Gold and iron oxide nanocomposites [162] Nisin Z Lactobacillus lactis EDC activated magnetic beads Anti-nisin antibody Elution with 6 M urea [163] In the case of proteins and peptides purifications, no simple strategy for magnetic affinity separations exists. Various affinity ligands have been immobilised on magnetic particles, or magnetic particles have been prepared from biopolymers exhibiting the affinity for target enzymes or lectins. Immunomagnetic particles, i.e. magnetic particles with immobilised specific antibodies against the target structures, have been used for the isolation of various antigens, both molecules and cells [ 5 ] and can thus be used for the separation of specific proteins. Magnetic separation procedures can be employed in several ways. Preparative isolation of the target protein or peptide is usually necessary if further detailed study is intended. In other cases, however, the magnetic separation can be directly followed (after elution with an appropriate buffer) with SDS electrophoresis. Magnetically separated proteins and peptides can also be used for further mass spectroscopy characterization [ 30 , 31 ]. The basic principles of magnetic separations can be used in the course of protein or peptide determination using various types of solid phase immunoassays. Usually immunomagnetic particles directly capture the target analyte, or magnetic particles with immobilised streptavidin are used to capture the complex of biotinylated primary antibody and the analyte. The separated analyte is then determined (usually without elution) using an appropriate method. A combination of magnetic separation with affinity capillary electrophoresis is also possible [ 32 ]. Enzyme isolation is usually performed using immobilised inhibitors, cofactors, dyes or other suitable ligands, or magnetic beads prepared from affinity biopolymers can be used (see Tables 1 , 2 , 3 , 4 ). Genetic engineering enables the construction of gene fusions resulting in fusion proteins having the combined properties of the original gene products. To date, a large number of different gene fusion systems, involving fusion partners that range in size from one amino acid to whole proteins, capable of selective interaction with a ligand immobilized onto magnetic particles or chromatography matrices, have been described. In such systems, different types of interactions, such as enzyme-substrate, receptor-target protein, polyhistidines-metal ion, and antibody-antigen, have been utilized. The conditions for purification differ from system to system and the environment tolerated by the target protein is an important factor for deciding which affinity fusion partner to choose. In addition, other factors, including protein localization, costs for the affinity matrix and buffers, and the possibilities of removing the fusion partner by site-specific cleavage, should also be considered [ 33 , 34 ]. As an example, isolation of recombinant oligohistidine-tagged proteins is based on the application of metal chelate magnetic adsorbents [ 35 , 36 ]. This method has been used successfully for the purification of proteins expressed in bacterial, mammalian, and insect systems. Antibodies from ascites, serum and tissue culture supernatants can be efficiently isolated using magnetic particles with immobilized Protein A, Protein G or anti-immunoglobulin antibodies. Protein A, isolated from Staphylococcus aureus , binds the Fc region of IgG of most mammalian species with high affinity, leaving antigen specific sites free. Protein G, isolated from Streptococcus sp., reacts with a larger number of IgG isotypes. It has a higher binding affinity to immunoglobulins than Protein A, however, it also interacts with the Fab regions of IgG, although the affinity is ten times lower than for the Fc region [ 37 ]. Antiphospholipid antibodies were successfully isolated using magnetoliposomes [ 15 ]. Aptamers are DNA or RNA molecules that have been selected from random pools based on their ability to bind other molecules. Aptamers binding proteins can be immobilised to magnetic particles and used for isolation of target proteins. DNA/RNA binding proteins (e.g., promoters, gene regulatory proteins and transcription factors) are often short-lived and in low abundance. A rapid and sensitive method, based on the immobilization of biotinylated DNA/RNA fragments containing the specific binding sequence to the magnetic streptavidin particles, can be used. The bound DNA/RNA binding proteins are usually eluted with high salt buffer or change of pH [ 38 ]. Other types of proteins were isolated using specific affinity-based procedures. For example, plasminogen immobilized on magnetic particles was used to separate scrapie and bovine spongiform encephalopathy associated prion protein PrP Sc from its conformer which is a cellular protein called PrP C . In fact, plasminogen represents the first endogenous factor discriminating between normal and pathological prion protein. This unexpected property may be exploited for diagnostic purposes [ 39 , 40 ]. Magnetic separation was also successfully used for the recovery of proteins expressed in the form of inclusion bodies, involving at first chemical extraction from the host cells, then adsorptive capture of the target protein onto small magnetic adsorbents, followed by rapid collection of the product-loaded supports with the aid of high gradient magnetic fields [ 41 ]. A new approach for analytical ion-exchange separation of native proteins and proteins enzymatic digest products has been described recently [ 31 ]. Magnetite particles were covered with a gold layer and then stabilized with ionic agents. These charged stabilizers present at the surface of the gold particles are capable of attracting oppositely charged species from a sample solution through electrostatic interactions. Au@magnetic particles having negatively charged surfaces are suitable probes for selectively trapping positively charged proteins and peptides from aqueous solutions. The species trapped by the isolated particles were then characterized by matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS) after a simple washing. Magnetic solid phase extraction (MSPE) enables to preconcentrate target analytes from larger volumes of solutions or suspensions using relatively small amount of magnetic specific adsorbent. Up to now this procedure was used for preconcentration of low-molecular weight xenobiotics [ 42 , 43 ] but using suitable magnetic adsorbents the MSPE could be used to preconcetrate target proteins and peptides as well. Sometimes the removal of certain proteins will reveal functions involving the depleted proteins or will help in the course of subsequent protein isolation. As an example, Dynabeads have been used to remove involved proteins from Xenopus egg extracts for analyses of the cell mitosis mechanisms [ 44 , 45 ]. Rapid removal of contaminating proteolytic enzymes from the crude samples could increase yields of sensitive proteins due to the limitation of their proteolysis [ 46 ]. A combination of mechanical cell disintegration and magnetic batch affinity adsorption was used to simplify the isolation of intracellular proteins. Magnetic glass beads were used because of their hardness and rigidity [ 1 ]. An example of quite different protein purification strategy can also be mentioned. Proteins associated with the endocytic vesicles of Dictyostelium discoideum were separated after magnetic isolation of the vesicles that was accomplished by feeding the amoebae with dextran-stabilized iron oxide particles. The cells were broken, the labelled vesicles were magnetically separated and then disrupted to release proteins which were resolved by SDS-PAGE. After „in-gel“ digestion with endoproteinase Lys-C or Asp-N the generated peptides were used for amino acid sequencing. This strategy allowed the identification of the major protein constituents of the vesicles [ 47 ]. Analogous procedure was used for the separation and study of peroxisomes proteins when at first peroxisomes were separated using magnetic beads with immobilized specific antibodies and then the protein content of the separated peroxisomes was analysed [ 48 ]. Conclusions Standard liquid column chromatography is currently the most often used technique for the isolation and purification of target proteins and peptides. Magnetic separation techniques are relatively new and still under development. Magnetic affinity particles are currently used mainly in molecular biology (especially for nucleic acids separation), cell biology and microbiology (separation of target cells) and as parts of the procedures for the determination of selected analytes using magnetic ELISA and related techniques (especially determination of clinical markers and environmental contaminants). Up to now separations in small scale prevail and thus the full potential of these techniques has not been fully exploited. It can be expected that further development will be focused at least on two areas. The first one will be focused on the laboratory scale application of magnetic affinity separation techniques in biochemistry and related areas (rapid isolation of a variety of both low- and high-molecular weight substances of various origin directly from crude samples thus reducing the number of purification steps) and in biochemical analysis (application of immunomagnetic particles for separation of target proteins from the mixture followed by their detection using ELISA and related principles). Such a type of analysis will enable to construct portable assay systems enabling e.g. near-patient analysis of various protein disease markers. New methodologies, such as the application of chip and microfluidics technologies, may result in the development of magnetic separation processes capable of magnetic separation and detection of extremely small amount of target biologically active compounds [ 49 ]. In the second area, larger-scale (industrial) systems are believed to be developed and used for the isolation of biologically active compounds directly from crude culture media, wastes from food industry etc., integrating three classical steps (clarification, concentration and initial purification) into a single unit operation [ 50 ]. It is not expected that extremely large amounts of low cost products will be isolated using magnetic techniques, but the attention should be focused onto the isolation of minor, but highly valuable components present in raw materials. Of course, prices of magnetic carriers have to be lowered and special types of low-cost, biotechnology applicable magnetic carriers and adsorbents prepared by simple and cheap procedures have to become available. The existence of inexpensive and effective magnetic separators enabling large-scale operations is necessary, as well. In the near future quite new separation strategies can appear. A novel magnetic separation method, which utilizes the magneto-Archimedes levitation, has been described recently and applied to separation of biological materials. By using the feature that the stable levitation position under a magnetic field depends on the density and magnetic susceptibility of materials, it was possible to separate biological materials such as haemoglobin, fibrinogen, cholesterol, and so on. So far, the difference of magnetic properties was not utilized for the separation of biological materials. Magneto-Archimedes separation may be another way for biological materials separation [ 51 ]. It can be expected that magnetic separations will be used regularly both in biochemical laboratories and biotechnology industry in the near future.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544596.xml
551616	Construction and analysis of tag single nucleotide polymorphism maps for six human-mouse orthologous candidate genes in type 1 diabetes	Background One strategy to help identify susceptibility genes for complex, multifactorial diseases is to map disease loci in a representative animal model of the disorder. The nonobese diabetic (NOD) mouse is a model for human type 1 diabetes. Linkage and congenic strain analyses have identified several NOD mouse Idd (insulin dependent diabetes) loci, which have been mapped to small chromosome intervals, for which the orthologous regions in the human genome can be identified. Here, we have conducted re-sequencing and association analysis of six orthologous genes identified in NOD Idd loci: NRAMP1/SLC11A1 (orthologous to Nramp1/Slc11a1 in Idd5.2 ), FRAP1 (orthologous to Frap1 in Idd9.2 ), 4-1BB/CD137/TNFRSF9 (orthologous to 4-1bb/Cd137/Tnrfrsf9 in Idd9.3 ), CD101/IGSF2 (orthologous to Cd101/Igsf2 in Idd10 ), B2M (orthologous to B2m in Idd13 ) and VAV3 (orthologous to Vav3 in Idd18 ). Results Re-sequencing of a total of 110 kb of DNA from 32 or 96 type 1 diabetes cases yielded 220 single nucleotide polymorphisms (SNPs). Sixty-five SNPs, including 54 informative tag SNPs, and a microsatellite were selected and genotyped in up to 1,632 type 1 diabetes families and 1,709 cases and 1,829 controls. Conclusion None of the candidate regions showed evidence of association with type 1 diabetes ( P values > 0.2), indicating that common variation in these key candidate genes does not play a major role in type 1 diabetes susceptibility in the European ancestry populations studied.	Background Type 1 diabetes is a common, multifactorial disease believed to be caused in a proportion of cases by an autoimmune destruction of pancreatic β-cells by an inflammatory infiltrate comprising T lymphocytes, dendritic cells and macrophages. This process results from a complex interaction between genetic and environmental risk factors. Genetically, it is under the control of the major histocompatibility complex (MHC) [ 1 ] and many other genes of smaller effect and mostly unknown identity. A murine model of type 1 diabetes, the NOD mouse, spontaneously develops an autoimmune-mediated diabetes that has many similarities to the human disease. It is likely that components of the pathophysiology and genetic predisposition are conserved across species, and indeed two loci have already been shown to affect type 1 diabetes susceptibility in both species, namely the immunoregulatory MHC HLA class II and CTLA-4 genes. The other causative gene(s) in the known Idd regions controlling type 1 diabetes susceptibility in the NOD mouse could also determine susceptibility in humans, even though this depends on the frequency of susceptibility alleles in human populations, which affects statistical power, and that the correct candidate gene has been chosen from the Idd interval. These Idd intervals might contain many genes, including several involved in the immune response [ 2 ]. Nevertheless, in contrast to studies in humans based on linkage, the localisation of a type 1 diabetes locus to a specific chromosome region in the mouse genome using congenic strain breeding defines with certainty a set of genes, one or more of which is definitely a susceptibility gene [ 3 , 4 ]. The central importance of T cell development and function in type 1 diabetes is evident from the susceptibility genes identified so far. The MHC class II genes are important etiologically in two rat models of type 1 diabetes, the Biobreeding (BB) and KDP strains [ 5 , 6 ], the NOD mouse strain [ 3 ] and in humans [ 1 ], with their essential function not only in T cell activation and expansion but also in T cell repertoire formation in the thymus and clonal deletion of autoreactive cells. The BB rat type 1 diabetes susceptibility locus Ian4/Iddm1 [ 7 ] affects T lymphocyte development whereas the Cblb (KDP rat) [ 8 ] and CTLA4 [ 9 ] (in humans and NOD mice) susceptibility genes highlight the importance of the regulation of T cell activation, expansion and homeostasis in the periphery, and perhaps in the thymus as well. In our selection of candidate genes within NOD congenic intervals, we have, therefore, biased our choice towards immune-related genes such as Il2 [ 2 ], Cd101 [ 10 ] and Nramp1/Slc11a1 [ 11 ]. From each of Idd5.2 [ 11 ], Idd9.2 [ 12 ], Idd9.3 [ 12 ], Idd10 [ 10 ], Idd13 [ 13 ] and Idd18 [ 14 ] we chose immune-associated functional candidate genes to study in human type 1 diabetes: Nramp1 / Slc11a1 from Idd5.2 [ 11 ]; Frap1 from Idd9.2 (unpublished); 4-1bb/Cd137/Tnfrsf9 from Idd9.3 [ 12 ] (unpublished); Cd101/Igsf2 from Idd10 [ 10 ]; B2m from Idd13 [ 15 ] and Vav3 from Idd18 (note that very recent congenic strain mapping results indicate that the Idd18 interval contains only one gene with known immunological function, namely the VAV3 gene, and this will be published elsewhere). Table 1 summarises the main features of the six human candidate genes. Table 1 NOD mouse Idd loci, location of their human orthologous regions, and selected functional candidate genes. Idd Mouse chromosome Interval size (Mb) Number of genes Functional candidate genes in mouse Idd intervals Location of human orthologous region Human orthologue genes Known gene function and previously reported disease associations Idd5.2 1 1.7 47 Idd5.2: Nramp1/Slc11a1 2q35 NRAMP1/SLC11A1 Endosomal/lysosomal acidification and associated with protection from infectious disease and susceptibility to autoimmune disease Idd9.2 4 1.1 13 Idd9.2: Frap1 1p32 FRAP1 FKBP12-rapamycin associated protein of mTOR. Candidate tumour suppressor gene, whose function in apoptosis is influenced by allelic variation Idd9.3 4 1.2 13 Idd9.3: 4-1bb 1p36 4-1BB Role in enhancing and regulating CD4 + , CD8 + T cells and dendritic cells Idd10 3 0.95 7 Idd10: Cd101 1p12 CD101 Co-stimulatory receptor of T cells Idd13 2 6 cM > 50* Idd13: B2m 15q21 B2M Required for antigen presentation by MHC class I molecules and the development of diabetes in NOD mice Idd18 3 0.7 2 Idd18: Vav3 1p13-p21 VAV3 Guanine nucleotide exchange factor involved in signalling of T and B cell receptors Estimated number of genes. A version of this table is provided in Additional file 1 with the supporting published references (see Additional file 1 ). Results and discussion A tag SNP approach to test for association was adopted for all genes, except for 4-1BB [ 16 ], in order to achieve cost-savings in genotyping. A multi-locus test was used to evaluate the association between type 1 diabetes and the tag SNPs due to linkage disequilibrium (LD) with one or more causal variants [ 17 ]. Coding and untranslated regions of NRAMP1 (MIM 600266), FRAP1 (MIM 601231), 4-1BB (MIM 602250), CD101 (MIM 604516), B2M (MIM 109700) and VAV3 (MIM 605541) were re-sequenced in 32 or 96 randomly chosen UK white patients with type 1 diabetes to identify SNPs and for the selection of tag SNPs. As LD between 4-1BB SNPs was weak, eight out of nine common SNPs were genotyped (minor allele frequency, MAF ≥ 0.03; one SNP could not be genotyped due to assay technical difficulties) and analysed using single-locus tests. A total of 110 kb of re-sequenced regions yielded 220 SNPs, including six deletion/insertion polymorphisms (DIPs) (see Table 2 and Additional files 2 , 3 , 4 , 5 , 6 and 7 ). No coding changes or obvious candidates for variants that could change the function or expression of 4-1BB, FRAP1 , or B2M were observed. A synonymous change was detected in exon 3 of NRAMP1 (MAF = 0.32) and a non-synonymous SNP (nsSNP) in exon 15 (MAF = 0.02), causing a conservative amino acid change: Asp543Asn (DIL5202/ss23142243). Interestingly, as in the case of its mouse orthologue [ 10 ], several nsSNPs were discovered in exons 3, 4, 5, and 8 of CD101 (see Additional file 5 ). Re-sequencing of the three alternative transcripts of VAV3 , called VAV3 (27 exons), VAV3β (unique exon 1 and exons 4 to 27) and VAV3.1 (unique exon 18 and exons 19 to 27) yielded six exonic SNPs (see Additional file 7 ). Two SNPs, Pro611Ser (MAF = 0.13) and Gln613His (MAF = 0.13) are located in the SH3 domain of the VAV3 protein and, therefore, could result in VAV3 having altered protein interactions. In order to facilitate the computation of the selection of tag SNPs, VAV3 was divided into three sections as suggested by the pattern of LD across the gene. Table 2 Summary of the re-sequencing study. Gene size, number of exons, amount of re-sequenced DNA for each gene (including 5' and 3' regions of gene), sequencing panel, and number of SNPs identified. Locus Genomic size (kb) n exons Re-sequenced region (kb) n cases re-sequenced n SNPs NRAMP1 13.58 15 (7,13) 12.13 32 20 4-1BB 20.50 8 13.66 96 23 FRAP1 60.98 58 30.88 32 55 CD101 34.61 10 15.90 96 31 B2M 6.61 4 9.33 32 13 VAV3 393.70 27 (25,10)* 27.69 96 78 Total 529.98 122 109.59 - 220 * Number of exons in splice variants. n, number. Two common nsSNPs (MAF ≥ 0.05; DIL1521/rs7528153 and DIL3809/ss23142432) from VAV3 and a microsatellite from NRAMP1 were genotyped a priori in the whole family collection (step 1 and 2) and a single nsSNP from CD101 in step 1 families only (DIL3794/rs3754112). The nsSNP DIL3810/ss23142433 in VAV3 was not tested because it was in quite strong LD with DIL3809/ss23142432 (R 2 = 0.64), so that only DIL3809/ss23142432 was genotyped. Note that in our tag approach, the two VAV3 nsSNPs (DIL1521/rs7528153 and DIL3809/ss23142432) were chosen deliberately as tag SNPs. In a pragmatic, phased genotyping strategy, in step 1, the multi-locus test P values for association between type 1 diabetes and candidate gene tag SNPs all exceeded 0.2, as did the single-locus test P values for 4-1BB SNPs. Consequently, we did not proceed to genotype in step 2 samples for any of the candidate genes (Table 3 and 4 ). Note that none of the nsSNPs of VAV3 and CD101 or the microsatellite of NRAMP1 showed evidence of association (Table 5 ). Allele A3 of the NRAMP1 microsatellite promoter (GT) n has previously shown linkage and association with autoimmune disease, and allele A2 with infectious disease susceptibility [ 18 - 20 ]. The relative risks of allele A3 and genotype A3/A3 in our type 1 diabetes samples was 0.96 (95% CI = 0.94 – 1.17) and 0.90 (95% CI = 0.70 – 1.16), respectively. Table 3 Study design. Lengths of re-sequenced genomic regions, and number of tag SNPs or single SNPs genotyped in a pragmatic two-step genotyping design for NRAMP1 , 4-1BB , FRAP1 , CD101 , B2M , and VAV3 . Locus Re-sequenced region (kb) n common SNPs* n tag SNPs Genotyping strategy (step 1 → step 2) NRAMP1 12.13 12 4 Case-control → Family set 1+2 4-1BB 13.66 8 DIL4279/ss23142250 Family set 1 DIL4277/rs226476 DIL4569/rs226478 DIL4274/ss23142263 DIL4570/ss23142264 DIL4571/rs679563 DIL4273/ss23142265 DIL4272/ss23142270 FRAP1 30.88 21 6 Family set 1 CD101 15.9 18 8 Family set 1 B2M 9.33 10 8 Case-control → Family set 1 VAV3 27.69 19 (block 1) 18 (block 2) 15 (block 3) 7 (block 1) 11 (block 2) 10 (block 3) Family set 1 For the selection of tag SNPs, minor allele frequencies of 0.03 were used for 4-1BB , CD101 and FRAP1 , and 0.05 for NRAMP1 , B2M and VAV3 . Note that the numbers of attempted and actual genotypes are given in Additional file 8 . n, number. Table 4 Disease association results. Multi-locus test P values, lengths of re-sequenced genomic regions, and number of tag SNPs or single SNPs genotyped in a two-step genotyping design for NRAMP1 , 4-1BB , FRAP1 , CD101 , B2M , and VAV3 . Locus Multilocus test P value/ Single-locus TDT P value Case-control Combined test P value Family set 1 Family set 1 + 2 NRAMP1 - 0.56 0.20 0.68 4-1BB 0.71 - - - 0.88 - - - 0.52 - - - 0.35 - - - 0.53 - - - 0.29 - - - 0.95 - - - 0.24 - - - FRAP1 0.44 - - - CD101 0.68 - - - B2M 0.90 - 0.11 0.75 VAV3 0.26 (block 1) - - - 0.80 (block 2) - - - 0.86 (block 3) - - - Table 5 Association analysis of non-synonymous SNPs. SNPs with allele frequencies above 0.05 and the NRAMP1 (GT) n microsatellite in up to 1,476 families with at least one affected offspring. N, number; T, number of transmissions; NT, number of untransmitted alleles; %T, percentage transmission of minor allele from heterozygous parents to type 1 diabetes offspring (obtained by transmission/disequilibrium test (TDT)); GTRR, genotype relative risk; P , probability value (two-sided). Locus Marker ID Amino acid change/ alleles Minor allele frequency N families T NT %T P TDT P GTRR VAV3 DIL1521 Thr293Ser/T>A 0.27 1 476 834 840 49.64 0.77 0.90 DIL3809 Pro611Ser/G>A 0.13 1 476 417 429 49.29 0.68 0.84 CD101 DIL3794 Asn225Ser/A>G 0.32 652 517 515 49.9 0.95 0.96 NRAMP1 (GT) n - - 1 476 - - - - 0.36 With regards to our association study in humans, intronic and potential regulatory regions were not sequenced in the candidate genes since these cover large genomic regions, which will have to wait for much more extensive polymorphism maps [ 21 ]. For example, for VAV3 , which spans almost 400 kb, less than 10% of the genomic region of VAV3 was re-sequenced to identify SNPs. The general importance of intronic and intergenic regulatory sequences as candidates for disease susceptibility is well recognised. Hence, potential unidentified causal variants in introns or flanking regions of the genes may have been missed, and remain a target for future analyses. Despite finding no evidence of association, it remains possible that there exists a common disease variant in one or more of the six candidate genes tested, which either has an effect smaller than would be detected with this study or is in much weaker LD with the tag SNPs than any other SNP known to us [ 22 ]. Finally, the possibility of one or more rare disease variants in a locus needs to be considered [ 23 ]. The best candidates for rare disease variants in the six genes studied here were thus genotyped in an expanded case-control collection of up to 3,704 type 1 diabetes cases and 3,930 controls: DIL5202/ss23142243 causes a non-conservative change in NRAMP (Asp543Asn, MAF = 0.02) and DIL3799/ss23142349 in CD101 (Val839Ile; MAF = 0.03). For both SNPs, P values above 0.05 were obtained ( P = 0.19 for DIL5202/ss23142243 and P = 0.80 for DIL3799/ss23142349), therefore, making it less likely that these rare variants contribute to susceptibility to type 1 diabetes. Nevertheless, causal variants with MAFs less than 0.01 [ 24 ] may well remain undetected in our re-sequencing panels of 32 or 96 case DNAs. However, the re-sequencing of several hundred cases and controls is beyond the scope of the present study in which we have investigated variants with MAF ≥ 0.03. Conclusion Taken together, these data make an association between type 1 diabetes and common variation in coding and untranslated regions of the six functional candidate genes in the investigated human-mouse orthologue regions less likely. Several possibilities may account for this. A gene (or several genes) in an Idd interval may account for disease susceptibility in the NOD mouse, but the human orthologous region may lack this susceptibility variant. The scenario, in which candidate genes in the NOD Idd interval may not necessarily be harbouring a functional, causal variant in their human orthologue genes, was discussed previously [ 25 ]. It is also possible that the selected candidate gene in the Idd interval may not be the gene causing susceptibility to disease. The tag SNP maps described here will be useful for association studies of other diseases. They will be integrated into future SNP maps encompassing the entire orthologous regions and all regulatory sequences and genes encoded within them. Methods Subjects All family members were white and of European ancestral origin. The type 1 diabetes families comprised two parents and a least one affected child. The 748 type 1 diabetes families used in 'step 1' were as described previously [ 26 ]: 472 UK Warren 1 multiplex and 276 multiplex Human Biological Data Interchange families ascertained in the U.S.A. The case-control DNA set for the tag SNP approach consisted of 1,709 Caucasian type 1 diabetes cases, which were recruited from across Britain in the Juvenile Diabetes Research Foundation/Wellcome Trust funded UK Genetic Resource Investigating Diabetes (GRID) study [ 27 ], and 1,829 population-based controls from the 1958 British Birth Cohort (BBC) [ 28 ]. The mean age-at-onset of the cases, with almost all under 16 years of age at diagnosis, is 7.5 years (with a standard deviation of 4 years). The 1958 BBC controls are part of an ongoing longitudinal study and the subjects are British citizens born in a particular week in 1958. In order to test association for type 1 diabetes susceptibility and the rare variants in CD101 and NRAMP1 , DIL3799/ss23142349 and DIL5202/ss23142243, a total of 3,704 type 1 diabetes cases and 3,930 controls were used. For 'step 2' genotyping of NRAMP1 , the 748 type 1 diabetes families described above were used in addition to 343 multiplex/simplex families from the UK, 159 Norwegian simplex families, 322 Romanian simplex families, and 60 multiplex families from the USA totalling the combined DNA sets to 1,632 type 1 diabetes families, as described previously [ 26 ]. Sequencing Nested PCR products from DNA from 96 or 32 type 1 diabetes patients were sequenced using an Applied Biosystems (ABI) 3700 capillary sequencer (Foster City, CA), and SNPs identified using the Staden Package [ 29 ]. Genotyping SNPs were genotyped using the Invader ® assay (Third Wave Technologies, Inc. Madison WI) [ 30 ] and TaqMan MGB chemistry (ABI) [ 31 ]. The NRAMP1 microsatellite was genotyped on an ABI3700 sequencer using fluorescent primers as previously described [ 32 ]. Full details of primers and probes used for genotyping are available upon request. All genotyping data was double-scored independently. Annotation Annotation of NRAMP1 (European Molecular Biology Laboratory [EMBL] accession numbers D50402, D50403, BC041787, L32185, BC033754), FRAP1 (UO88966), 4-1BB (UO3387), CD101 (Z33642), B2M (BC032589) and VAV3 (AF118887, VAV3 ; AF118886, VAV3β ; AF118887, VAV3.1 ) was performed by importing Ensembl information into a temporary ACeDB database as described in Burren et al. [ 33 ]. After confirmation of gene structures by BLAST analysis, these were re-extracted in GFF format and submitted to a local Gbrowse database (National Center for Biotechnology Information build 34) (DIL annotations viewable at T1DBase [ 34 ]. Statistical analysis The program for the selection of tag SNPs [ 17 ] and association analysis used here are implemented in the Stata statistical system and may be downloaded from our website [ 35 ]. All genotyping data were in Hardy-Weinberg equilibrium ( P > 0.05). Authors' contributions LMM and DJS contributed equally to this work by performing the genetic studies and writing the manuscript. AV, FP, RP, CL, JH, HF, CM, KMH, GC carried out the genetic studies and collated data. JDC performed the statistical analysis and participated in the design of the study. LJS participated in the sequence analysis. NW participated in design and collated data. KSR, CG, C I-T, DAS, DPS and LBP participated in the study design and coordination. JAT, LSW and RCT helped to draft the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 2 SNPs, including two deletion/insertion polymorphisms, identified in NRAMP1 . Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Minor allele frequencies are based on the sequencing panel of 32 type 1 diabetes subjects. R 2 values for non-typed SNPs. UTR, untranslated region. Click here for file Additional File 5 SNPs identified in CD101 . Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Minor allele frequencies are based on the sequencing panel of 96 type 1 diabetes subjects. R 2 values for non-typed SNPs. Note that DIL3969 has an allelic R 2 < 0.80 due to technical difficulties with the assay. UTR, untranslated region. Click here for file Additional File 7 SNPs, including four in/dels identified in VAV3 . Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Minor allele frequencies are based on the sequencing panel of 96 type 1 diabetes subjects. R 2 values for non-typed SNPs. Note that DIL6496 and DIL6488 in block 1, and DIL1526 have allelic R 2 values < 0.80, which was due to technical difficulties with those assays. UTR, untranslated region. Click here for file Additional File 1 Supporting published references for Table 1 . NOD mouse Idd loci, the location of their human orthologous regions, and selected functional candidate gene within the Idd interval. Estimated number of genes. Click here for file Additional File 3 SNPs identified in 4-1BB . Minor allele frequencies are based on the sequencing panel of 96 type 1 diabetes subjects. Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Note that DIL4247/rs6694557 could not be genotyped due to assay technical difficulties. UTR, untranslated region. Click here for file Additional File 4 SNPs identified in FRAP1 . Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Minor allele frequencies are based on the sequencing panel of 32 type 1 diabetes subjects. R 2 values for non-typed SNPs. UTR, untranslated region. Click here for file Additional File 6 SNPs identified in B2M . Novel SNPs are denoted by "ss" numbers and previously published SNPs are denoted by "rs" numbers. Minor allele frequencies are based on the sequencing panel of 32 type 1 diabetes subjects. R 2 values for non-typed SNPs. UTR, untranslated region. Click here for file Additional File 8 Genotyping counts Numbers of attempted subjects for genotyping and of subjects with genotypes. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551616.xml
522808	Intracellular shuttling of a Drosophila APC tumour suppressor homolog	Background The Adenomatous polyposis coli (APC) tumour suppressor is found in multiple discrete subcellular locations, which may reflect sites of distinct functions. In Drosophila epithelial cells, the predominant APC relative (E-APC) is concentrated at the apicolateral adherens junctions. Genetic analysis indicates that this junctional association is critical for the function of E-APC in Wnt signalling and in cellular adhesion. Here, we ask whether the junctional association of E-APC is stable, or whether E-APC shuttles between the plasma membrane and the cytoplasm. Results We generated a Drosophila strain that expresses E-APC (dAPC2) tagged with green fluorescent protein (GFP-E-APC) and we analysed its junctional association with fluorescence recovery after photobleaching (FRAP) experiments in live embryos. This revealed that the junctional association of GFP-E-APC in epithelial cells is highly dynamic, and is far less stable than that of the structural components of the adherens junctions, E-cadherin, α-catenin and Armadillo. The shuttling of GFP-E-APC to and from the plasma membrane is unaltered in mutants of Drosophila glycogen synthase kinase 3 (GSK3), which mimic constitutive Wingless signalling. However, the stability of E-APC is greatly reduced in these mutants, explaining their apparent delocalisation from the plasma membrane as previously observed. Finally, we show that GFP-E-APC forms dynamic patches at the apical plasma membrane of late embryonic epidermal cells that form denticles, and that it shuttles up and down the axons of the optic lobe. Conclusions We conclude that E-APC is a highly mobile protein that shuttles constitutively between distinct subcellular locations.	Background The Adenomatous polyposis coli (APC) protein is an important tumour suppressor in the colonic epithelium [ 1 ]. A key function of this highly conserved protein is to antagonize Wnt signalling, by constitutively downregulating the transcriptional activity of β-catenin/Armadillo, a key effector of the Wnt signalling pathway [ 2 ]. Loss of this function is thought to be critical in the initiation of colorectal tumorigenesis as it causes a transcriptional switch in the intestinal epithelium towards actively dividing crypt progenitor cells [ 3 - 5 ]. APC proteins are highly conserved among vertebrates and flies, and flies encode two APC proteins with overlapping roles in Wnt signalling during development [ 6 , 7 ]. However, APC proteins have additional functions in connection with the actin and microtubule cytoskeletons that appear to be separate from their function in controlling Wnt signalling [ 8 , 9 ]. One of these functions is a role of APC in facilitating cellular adhesion, as indicated by studies in Drosophila tissues [ 10 ] and in mammalian colorectal cancer cells [ 11 ]. This function in cellular adhesion is likely to be conferred by the subcellular pool of APC protein that is associated with adherens junctions (AJs) in Drosophila [ 12 , 13 ] and in polarised mammalian cells [ 14 ]. The mechanism by which APC facilitates cellular adhesion is unknown. In order to explore this mechanism, we asked whether Drosophila E-APC (also called dAPC2) might have a structural role at AJs. If so, E-APC would be expected to be stably associated with AJs, similarly to the structural components of the adhesive complex. As in mammalian epithelia [ 15 , 16 ], the main functional components of this complex in Drosophila are the transmembrane protein E-cadherin, a calcium-dependent trans-membrane adhesion molecule, and the catenins (Armadillo and α-catenin) that link E-cadherin to the actin cytoskeleton at the cytoplasmic side [ 17 - 22 ]. We thus conducted photobleaching experiments on live embryos expressing E-APC or structural AJ components tagged with green fluorescent protein (GFP) [ 23 - 25 ], to compare their relative mobility. These experiments revealed that GFP-E-APC is less stably associated with AJs than their structural components. We also found that GFP-E-APC is remarkably mobile in neurons. Results and discussion We used the GAL4 system to express GFP-E-APC throughout the embryo, and found that its subcellular distribution is very similar to that of endogenous E-APC in fixed embryos. In particular, GFP-E-APC is concentrated underneath the plasma membrane in apicolateral regions of embryonic epithelial cells (Fig. 1a ; Fig. 2b ). These regions form the zonula adherens (ZA) which contains the AJs [ 26 ]; they can also be visualised with antibody staining against α-catenin [ 27 ] (Fig. 1b ), and we observe a remarkably close coincidence of GFP-E-APC and α-catenin. Similar results were obtained by Akong at el. who examined the same GFP-E-APC transgene in the embryo [ 7 ], and who also showed that GFP-E-APC is distributed similarly as endogenous E-APC in larval neuroblasts [ 28 ]. Figure 1 Association of GFP-E-APC with AJs of embryonic epithelial cells. Side view of epidermis of a ~6 hour old embryo expression ubiquitous GFP-E-APC, stained with antibody against ( a ) GFP and ( b ) α-catenin to mark the apicolateral AJs; note the co-incidence of GFP-E-APC and α-catenin staining (arrows). Figure 2 FRAP protocol for the analysis of live Drosophila embryos expressing GFP-E-APC. ( a ) Sketch of the microscopy and data analysis used to determine the mobility of GFP-E-APC in epithelial cells of live embryos. ( b ) Consecutive face-on views of a live ~6 hours old embryo (stage 11) expressing GFP-E-APC, with squares marking specific sections of the AJs at cell interfaces; the set of yellow squares in the right-hand image illustrate the cell shape changes that took place during the 5 minutes between the two optical sections shown. White bar in this and subsequent figures, 5 μm. Next, we conducted fluorescence recovery after photobleaching (FRAP) experiments in live embryos, to examine how stably GFP-E-APC is associated with adherens junctions. We bleached the fluorescence in a defined square centred over the junctional region of an epithelial cell with a short laser pulse, and examined the recovery of the fluorescence within this square over time (Fig. 2a ). This revealed a relatively fast rate of recovery of most of the fluorescence within a few minutes (Fig. 3a ) [see additional file 1 ]. Quantitative analysis shows that nearly 80% of the initial fluorescence is recovered within ~220 seconds (with a half-time value of ~60 seconds) (Fig. 4a,4e ). This indicates that the association of E-APC with the ZA is dynamic rather than stable. The fluorescence recovery observed in these FRAP experiments could be due to movement of E-APC between the cytoplasm and the plasma membrane, but also to sideways movement along the plasma membrane (see also below). Figure 3 FRAP of GFP-E-APC and GFP-tagged AJ proteins in early embryonic epithelial cells. Face-on views of live ~6 hours old embryos (stage 11) expressing ( a ) GFP-E-APC, ( b ) E-cadherin-GFP, ( c ) Armadillo-GFP, ( d ) α-catenin-GFP, with white squares marking sections of cell interfaces that were bleached, and red squares marking unbleached control sections. Pre-bleaching images are shown on the left; subsequent images on the right show recovery of fluorescence within white squares 15, 75 and 300 seconds after bleaching [see additional file 1]. Figure 4 Quantitative evaluations of FRAP experiments. Plots of the relative fluorescence in white squares compared to grey squares in Fig. 3 as a function of time; error bars indicate the range of values from >12 different experiments. ( a-d ) FRAP of individual GFP-tagged proteins in wild-type embryos, as indicated; ( e ) combined data of ( a-d ); ( f ) Comparison of FRAP of GFP-E-APC in wild-type and sgg mutant embryos (note that we were unable to distinguish between null mutant and paternally rescued embryos; however, the rescue activity of the paternal allele is minimal as both types of embryos are highly abnormal). We also conducted FRAP experiments with structural AJ components, namely E-cadherin-GFP, Armadillo-GFP and α-catenin-GFP. In these cases, we can only recover a small fraction of the initial fluorescence within the time frame of the experiment (Fig. 4a,4b,4c,4d,4e ; note that these experiments cannot be extended beyond ~6 minutes, due to the extensive cell shape changes during this developmental stage). Furthermore, the rate of recovery is slower than that observed with GFP-E-APC, with estimated half-times of >3 minutes (α-catenin-GFP and of E-cadherin-GFP; Fig. 4b,4d,4e ). This also appears to be true for Armadillo-GFP (Fig. 4c,4e ), though we cannot estimate its half-time of recovery with confidence, given that its fluorescence levels are considerably lower than that of the other GFP-tagged protein examined in this study. We conclude that E-APC is significantly more mobile than the structural AJ components. This suggests that E-APC shuttles either within the cortex, along the zonula adherens, or that it shuttles from the cytoplasm to the plasma membrane (as previously proposed; [ 29 ]). Interestingly, the observed rates of recovery of GFP-E-APC were much slower than the estimated rate of free diffusion (e.g. [ 30 ]; the rate of recovery of GFP alone was <10 seconds, i.e. too fast to be measured by our experimental setup). This suggests that the movements of GFP-E-APC are primarily determined by the kinetics of its binding to ligands. One of these could be Axin which associates with E-APC in Drosophila cells to from large dot-like structures [ 31 ]. Similarly, Axin associates with APC in mammalian cells to form large molecular weight protein complexes [ 32 ]. Our observations argue against a structural role of E-APC in cellular adhesion. However, they are consistent with a catalytic role of E-APC in facilitating cellular adhesion, for example by maintaining the junctional pool of Armadillo [ 10 , 29 , 33 ]. In support of this, recent evidence suggests that there is rapid exchange of β-catenin within the junctional cadherin complex, and that APC is required for this process [ 34 ]. In late embryonic stages, GFP-E-APC forms striking patches underneath the apical plasma membrane of epidermal cells that are in the process of forming denticle extrusions (Fig. 5 ). These striking 'pre-denticle' patches are also seen in embryos stained with antibody against E-APC, and overlap with actin patches [ 25 ]. They may thus represent an actin-dependent association of E-APC as that seen in the cortex of earlier epithelial cells and at the ring canals between nurse cells within the egg chambers [ 25 , 33 ]. FRAP experiments revealed that the presence of GFP-E-APC in these pre-denticle patches is also dynamic, with an estimated half-time of fluorescence recovery of 200–300 seconds (Fig. 5 ). Again, E-APC is therefore unlikely to have a structural role in these patches. Figure 5 FRAP of GFP-E-APC patches in late embryonic epithelial cells. Face-on views of ~17 hours old embryo (stage 17) expressing GFP-E-APC, showing patches of GFP-E-APC at the apical plasma membrane of epidermal cells forming denticles (pre-bleach and subsequent images labelled as in Fig. 3). Note the fast recovery of the fluorescence in these patches after photobleaching. It has been reported that E-APC and Armadillo are required for anchoring mitotic spindles in the cortex of dividing blastoderm cells in the early Drosophila embryo [ 25 ]. We cannot measure the kinetics of GFP-E-APC association with the cortex in these early embryonic cells, because of insufficient expression levels at this stage. However, assuming that these kinetics do not change radically during embryonic development, our observations from the later embryos (Fig. 3 , 4 , 5 ) suggest that E-APC has a catalytic role in capturing microtubules in the cellular cortex, rather than providing a structural tether [ 25 ]. We also expressed GFP-E-APC in eye imaginal discs, to examine its subcellular distribution within a larval epithelial sheet. We thus noticed striking puncta of green fluorescence within the axons of the optic stalk that connects these discs to the larval brain (Fig. 6 ). These puncta resemble the E-APC/Axin-GFP dots that we observe in embryonic cells [ 31 ] and in these axons (not shown), and also the E-APC/Armadillo dots that Peifer and colleagues observed in early embryos [ 25 ]. They may thus represent the Axin destruction complex [ 31 ]. We performed FRAP experiments, bleaching a 6 μm wide strip perpendicularly across the axons and monitoring the recovery of the fluorescence into the bleached section. This revealed that the GFP-E-APC puncta are remarkably dynamic: they re-appear within the bleached area within a minute, with an estimated half-time of ~100 seconds (Fig. 6 ) [see additional file 2 ]. Many of these puncta seem to re-appear from other focal planes, so we cannot be absolutely certain that they represent movement of existing puncta. However, some of the puncta re-appearing in the bleached area can clearly be traced as moving puncta within the same focal plane (e.g. see isolated axon, left-hand side of [ additional file 2 ]). The movement of these GFP-E-APC puncta may be due to tracking (e.g. along microtubules), although we cannot see uni-directionality of movement (i.e. the movement appears to be up and down the axons). The movement we observe in these axons is somewhat reminiscent of that observed with a GFP-tagged truncation of Xenopus APC that misses its C-terminus (thus resembling the overall structure of E-APC): this truncation, despite lacking the putative microtubule-interacting domain within its C-terminus, forms large puncta that can track along microtubules in Xenopus tissue culture cells [ 35 ]. Figure 6 FRAP of GFP-E-APC in the larval optical stalk. Optical sections through the optical stalk of a third instar larva, before and after photobleaching, showing fluorescent puncta in individual axons, and the reappareance of these puncta (arrows) from both sides of the bleached areas within minutes [see additional file 2]. Width of bleached strip, 6 μm. GSK3 is inhibited during Wnt signalling [ 2 ], and GSK3 mutants in Drosophila ( shaggy / zeste white3 , or sgg , mutants) therefore mimic constitutive and sustained Wingless signalling [ 36 ]. The normal level of Wingless signalling in the embryonic epidermis does not appear to change the subcellular distribution of bulk E-APC protein [ 12 ], although it does cause a re-location of Axin-GFP/E-APC complexes to the plasma membrane [ 31 ]. However, a reduction of cortical E-APC has been observed in early sgg mutant embryos [ 12 , 25 ]. Likewise, in older sgg mutant embryos, the levels of membrane-associated GFP-E-APC are also noticeably reduced (Fig. 7a,7b,7c ). However, this does not appear to be due to a change in mobility of GFP-E-APC since the kinetics of fluorescence recovery between wild-type and sgg null mutant embryos were comparable (Fig. 4f ). Instead, it is due to a reduction of the overall E-APC protein levels in these mutants: Western blot analysis of 2–16 hour old embryos revealed that the total levels of GFP-E-APC protein were much lower in sgg mutant embryos compared to the wild type (Fig. 7d ). The same is true for endogenous E-APC whose levels are also substantially reduced in sgg mutants (Fig. 7e ). This indicates that sgg is required for the stability of E-APC protein, and it suggests that sustained Wingless signalling may destabilise E-APC. Similarly, phosphorylation by GSK3 is required for the stability of mammalian Axin, a functional binding partner of APC [ 37 ], and the levels of Drosophila Axin in embryos are also reduced after prolonged Wingless signalling [ 38 ]. Destabilisation of the main components of the Axin complex (Axin and APC) during Wnt signalling may be a positive feedback mechanism resulting in the amplification of the signalling level. Figure 7 Destabilisation of E-APC in sgg mutant embryos. ( a, b ) Face-on and ( c ) side views of ~14 hours old embryos (stage 16), fixed and co-stained with antibodies against GFP and α-catenin as indicated, revealing junctional association of GFP-E-APC in ( a ) wild-type and ( b, c ) sgg mutant embryos (similar in sgg null and paternally rescued embryos, see also Fig. 4f). ( d, e ) Western blots of hand-picked 10–16 hours old wild-type and sgg mutant embryos (~100 embryos per lane), probed with antibodies against ( d ) GFP or ( e ) E-APC, and α-tubulin as internal controls. Note that the levels of GFP-E-APC and of endogenous E-APC are much reduced in sgg compared to wild-type embryos ( sgg mutants represent a 1:1 mixture of sgg null and paternally rescued embryos). The lower bands in upper panels ( d, e ) correspond to breakdown products of GFP-E-APC and E-APC, respectively; their occurrence varies somewhat between preparations. The subcellular distribution of E-APC and its accumulation at the adherens junctions is unchanged in other mutants of the Wingless signalling pathway (including wg , axin , dsh and signalling-defective arm mutants; [ 31 ]; F. Hamada, X. Yu and M. B., unpublished observations). We thus did not expect any of these mutants to affect the shuttling behaviour of GFP-E-APC to and from the plasma membrane. In support of this, preliminary FRAP experiments indicated that the kinetics of fluorescence recovery are unaffected in dsh null mutant embryos (not shown). Taken together with our results from the sgg mutants, this suggests that the kinetic association of GFP-E-APC with the plasma membrane is unaffected by Wingless signalling. Conclusion Our FRAP experiments provided evidence that E-APC is a cytoplasmic shuttling protein whose association with the adherens junctions is highly dynamic. The speed of its shuttling to and from the plasma membrane appears to be constitutive and does not require GSK3 activity. The dynamic association of E-APC with the plasma membrane is consistent with a catalytic role of E-APC, and argues against a structural or tethering role in the cell cortex. Methods Fly strains Fly lines transformed with UAS.GFP-E-APC (full length E-APC tagged with GFP at its N-terminal end, inserted into pUAST [ 39 ]) were generated by R. Rosin-Arbesfeld (see also [ 7 , 28 ]). The GAL4 driver lines arm.GAL4 and GMR.GAL4 (FlyBase) were used to express GFP-E-APC throughout the embryonic epidermis [ 31 ] and in the larval eye disc, respectively. All fly strains were cultured at 25°C. zw3 M11-1 and dsh v26 mutant embryos lacking maternal and zygotic gene function were generated as described [ 40 ]. We did not detect any differences in the subcellular localisation of GFP-E-APC or α-catenin between zygotic null and paternally rescued sgg mutants (identified with an RFP-marked X chromosome [ 41 ]). For Western blot analysis, 10–16 hours old wild-type and sgg mutant embryos were hand-picked (from timed egg collections) under the dissecting microscope, and separated into GFP-positive and GFP-negative embryos; unfertilised embryos were discarded. Analysis of fixed embryos and Western blots Antibody staining of fixed embryos and analysis by confocal microscopy were described previously [ 12 ]. The following primary and secondary antibodies were used: rabbit anti-E-APC [ 12 ], rabbit anti-GFP [ 14 ], rat anti-α-catenin [ 42 ]; goat anti-rat IgG Alexa Fluor 568, goat anti-rabbit IgG Alexa Fluor 488 (Molecular Probes). The following primary and secondary antibodies were used for Western blotting: rabbit anti-E-APC [ 12 ]; mouse monoclonal anti-GFP IgG2a (Santa Cruz Biotechnology); mouse anti-α-tubulin (clone B-5-1-2, Sigma), as internal control; goat anti-mouse and anti-rabbit HRP IgG (Santa Cruz Biotechnology). The enhanced chemiluminescence (ECL) Western blotting system (Amersham) was used for detection [ 43 ]. Live imaging of embryos For live imaging, embryos were dechorionated in 50% bleach for 1–2 minutes and washed. Embryos were transferred to a moistened black filter (Schleicher and Schüll). Embryos were adhered to coverslips with heptane glue, made by mixing heptane and clear sellotape (Sellotape Ltd). Embryos were mounted in Voltalef oil (10S). For short term imaging (<30 minutes), embryos were mounted on a glass slide with small coverslips as supports. For longer term imaging, e.g. for bleaching of pre-denticle patches, embryos were mounted in oil and placed on Bio-foil gas permeable membrane (Sartorius Ltd) mounted on a perspex frame [ 44 ]. Photobleaching of live embryos FRAP experiments were performed using a Bio-Rad Radiance confocal microscope with a 40× NA 1.3 objective lense. Imaging was performed with a 488 nm argon laser at 5% laser power and the following confocal settings: iris at 4 mm, 50% gain, zoom 10, scan speed 500 lps, box size 512 × 512 pixels. These conditions were found to give minimal photobleaching over the observed time. For each FRAP experiment, a pre-bleach image was recorded by selecting a focal plane and taking a Z-series, consisting of 3 0.5 μm steps either side of the desired focal plane (from -1.5 μm to +1.5 μm). The LaserSharp software was used to define several regions of interest (ROIs) for bleaching. A maximum of one bleach ROI was placed in any cell and several cells were always left unbleached. Typically, 3–5 ROIs were bleached in one field of view on one embryo. These regions were bleached at 100% laser power (scanning at 500 lps). 10 bleach scans were found to produce the best results for all constructs. After bleaching, a Z-series was recorded every 15 seconds for 5 minutes. At the time of these experiments, the LaserSharp software did not contain a function for performing this type of 4D bleaching experiment. This problem was overcome by manually switching between imaging and bleaching settings and manually saving pre bleach images and starting the time course. As a result of this, there was usually a 30–60 second delay between the pre-bleach image and the post-bleach images. Data analysis Data sets were analysed with the Bio-Rad LaserPix software. For each time point, the total pixel intensity distribution was compared to the pre-bleach image to select the corresponding region. The two images were then compared by eye to confirm that they did correspond to the same focal plane. The coordinates for the bleach ROIs were used to accurately locate the bleach spots on the pre bleach image, and the mean fluorescence intensity for each ROI was calculated. Several equivalent sized ROIs were also placed on unbleached cells to measure any change in fluorescence due to photobleaching or movement. To track movement of the cells, an acetate sheet was placed over the computer monitor and each ROI was marked on it as well as the shapes of the cells surrounding it. By aligning the sheet with the appropriate cell shapes, the ROI could be appropriately positioned for each time point. This process was used to position each ROI on the appropriate image for each time point. Once all ROIs had been placed on the image, the mean fluorescence intensities were calculated for each ROI, and their positions were saved on a copy of the image (See Fig 2.1 ). Data was exported to Microsoft Excel for analysis. Relative fluorescence was calculated for each bleach area by dividing fluorescence at time (t) by pre-bleach fluorescence. The change in fluorescence was plotted on a graph with Excel. For each construct tested, the data from multiple bleach experiments from multiple embryos were averaged to give the approximate rate of recovery. Data sets were discarded for any of the following reasons. First, if movement of the embryo in the Z axis took the sample outside the range of the Z-series in any time point. Second, if movement in the X/Y axis was sufficient to move significant numbers of the bleach boxes outside of the observed region. Third, if an ROI ever left the field of view, all data points for that ROI was discarded. Fourth, all data sets were discarded if the intensities of the control ROIs changed dramatically at any point in the experiment, or showed a large general increase or decrease. Pre-denticle structures were bleached in a similar manner to junctional E-APC described above. Live imaging of the larval optic stalk GFP-E-APC was expressed in eye imaginal discs by the GAL4 system, using the driver line GMR.GAL4 (described in FlyBase). Eye discs and brains were dissected from crawling third instar larvae in PBS. Eye discs were teased away from the brain and inverted to reveal the optic stalk. Whole disc/brains were mounted in a drop of PBS under a cover slip, supported by two smaller cover slips. Each disc was observed for no more than 30 minutes. Photobleaching of the larval optic stalk FRAP experiments were performed using a Bio-Rad Radiance confocal microscope and Bio-Rad LaserSharp software, using the 100× NA 1.4 objective lens. A narrow strip was bleached across the whole field of view by adjusting the size of the scanning area. These experiments were performed before a FRAP program was available for LaserSharp so bleaching was performed manually, leading to somewhat variable intervals between each stage of the experiment. The region was bleached with the 488 nm line of an argon laser for approximately 20 scans. Time courses were recorded after each bleaching experiment for 5 minutes. Authors' contributions A.C. developed and conducted most of the FRAP experiments; J.M. completed the analysis of GFP-E-APC in live and fixed sgg mutant embryos, including the Western blots; M.B. directed the study, helped with the microscopy and drafted the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 1 FRAP of GFP-E-APC in early embryonic epithelial cells. Example of a FRAP experiment of GFP-E-APC, as described in Figure 3. Click here for file Additional File 2 FRAP of GFP-E-APC in the larval optical stalk. Example of a FRAP experiment of GFP-E-APC, as described in Figure 6. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522808.xml
552322	Review of "Understanding the Human Machine, A Primer for Bioengineering" by Max E. Valentinuzzi	null	The book "Understanding the Human Machine, A Primer for Bioengineering" embraces various aspects of biomedical engineering as an essential resource book for physical scientists, engineers and biomedical students. The book interfaces physiologic systems with engineering principles to capture the important concepts from a plethora of facts in field of biomedical sciences. Humors, exercise, history of biomedical discovery and elements of humility continuously emerge throughout the book; thereby, enhancing the self-taught approach for the new comers. The Introduction reflects the author's philosophy and advice for the readers from his accumulated wisdom and experience in teaching bioengineering. In Chapter 2, the author used the system approach to intertwine cardiovascular, renal, respiratory, gastrointestinal, endocrine, nervous and muscular systems. The organization is timely and effective in the era of systems biology and the emerging research towards predictive and preventive medicine. The system approach in Chapter 2 paves the way for detecting signal output from the physiologic systems. Biosensors are illustrated to measure electric signals from the visual, auditory and olfactory systems as well as the heart, brain, and muscle. The integration of electric circuits using the classic examples such as the Wheatstone bridge and operational amplifiers provides the frame work for signal acquisition and processing. Finally, the author completed the book with a touch of bioinformatics in the post genomic era. In general, the book is comprehensive and succinct, readable and interesting. The presented material spans from the fundamental principles to the applications of bioengineering. Certain areas will improve the quality of this book for the readership of undergraduates and new comers to bioengineering. Some figures are difficult to understand as a result of small captions (eg. figs 2.11 and 3.19). Others are old and lack clear labeling (eg. Figs. 2.52 and 3.15). Tissue engineering deserves a paragraph. Analogous to nanotechnology, a link to a tissue engineering website would be helpful for the readers. A paragraph on molecular engineering involving proteins, DNA and RNA molecules would enhance the balance between classic bioengineering and the emerging fields. The readers would be intrigued if the book ends with a paragraph introducing Bio-MEMS (Micro Electrical Mechanical Systems), bionanotechnology, molecular imaging, and surface chemistry. Overall, "Understanding the Human Machine, A Primer for Bioengineering" is a very useful book that highlights the fusion between organ systems and engineering principles. I would recommend this book as an introduction to bioengineering and a reference for students from physical science, engineering and life science.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552322.xml
544540	Biology by Numbers—Introducing Quantitation into Life Science Education	An online educational module introduces students to concepts of quantitation and numerical simulations in developmental biology	Driven by the massive datasets that are generated by “omics” research, the molecular life sciences are entering a new phase. This phase is characterised by a shift in focus from individual genes and their products to networks and whole systems [ 1 , 2 , 3 ]. For a thorough analysis of the behaviour of networks and their underlying principles, quantitative tools are often necessary. Numerical simulations can, for example, be used to explore the behaviour of a network when the values of different parameters are varied, and, in turn, mathematical analysis can help to understand a particular biological phenomenon [ 2 ]. The successful application of quantitative tools in the molecular life sciences requires a good understanding of these tools and sufficient knowledge of the biological system under study. This can be achieved by collaboration between quantitatively trained scientists such as physicists on the one hand and biologists on the other. However, cultural differences hamper such collaboration [ 1 ]: even at the undergraduate level, students in the different disciplines speak very different languages [ 4 ]. A more productive approach is therefore to prepare students better for the quantitative nature of the molecular life sciences by integrating quantitative thinking and biology in the life science curriculum. This can be achieved in various ways. For example, a curriculum could be developed in which mathematics, the physical sciences, and biology are introduced together [ 4 ]. However, we recommend that quantitative thinking also be included throughout the curriculum in the biology courses themselves, covering topics such as cell biology, developmental biology, and biochemistry. We consider this important because it will help to show students how quantitative tools can be used to address various cutting edge questions in biology. A Modelling Module in Developmental Biology As an example of the integration of quantitative teaching and cutting edge biology, we have implemented an educational module in which numerical simulations are used in an existing course on developmental biology ( http://mbedu.fbt.eitn.wau.nl/demo_plos/ ). Some of the features of this module and the thinking that led to its development are quite general, and so we present the module here as a case study in the hope that this might inspire and guide others to create similar resources. First, we wanted to illustrate to students the value of using numerical simulations to study a developmental process. Therefore, a pattern-forming mechanism was selected that can initially be rather hard to understand: the generation of the morphogen gradient formed by the extracellular signalling molecule decapentaplegic (Dpp) early during Drosophila embryogenesis [ 5 ]. The generation of this gradient results from the fact that key proteins are synthesized in different embryonic regions, from the formation of complexes of these proteins, and from the different diffusion rates of these complexes and their components, as well as from the specific degradation of some components. Students are guided through the creation of a model for Dpp gradient formation based on a set of experimental data. At several stages, students can perform simulations in a separate simulation environment. Students use simulations, for example, to check whether a number of core interactions is sufficient to yield the most important characteristics of the wild-type gradient. Second, we designed the simulation environment in such a way that biology students with their existing mathematical background can build quantitative models and run numerical simulations themselves. In this environment ( Figure 1 ), students do not have to program anything, or set up differential equations, themselves. Instead, they indicate which processes occur at the molecular level, and the program then shows how each of these processes is translated into a term in a differential equation. In Figure 2 , for example, the program adds a diffusion term to a differential equation if the student indicates that diffusion occurs. Besides setting up the equations in this way, students specify the initial localisations and concentrations of the different proteins, as well as the constants that are used in the differential equations. Figure 1 A Simulation That Students Can Perform After several minutes, Dpp forms one peak in the centre of the dorsal region, as in the wild type. The various elements of the quantitative model can be entered under “protein conc. changes”, “initial localizations”, “values of constants”, and “initial concentrations”. The numerical simulation itself shows the dynamic behaviour of the designed quantitative model. Figure 2 Illustration of How Students Can Set Up Differential Equations If a student indicates that Dpp diffusion occurs, a diffusion term is added to the differential equation that describes the changes in Dpp concentration. Third, we wanted to make sure that students would use the simulation environment effectively. Therefore, a clear goal is formulated when students enter the simulation environment. For example, they are asked to make a model that generates a Dpp gradient that fulfills a number of specific criteria, or simulates certain mutants. After running a simulation, students can view feedback that helps them draw conclusions or consider the next step to be taken. If a student's model, for example, generates a gradient that is too shallow, the student has to indicate which change in the model he expects to be useful for generating a steeper gradient. The student then receives an intuitive explanation of the usefulness of the given suggestion. If an increase in the synthesis of one of the proteins, Short gastrulation, is proposed, for example, feedback is given that this could indeed be useful, since there would then be more Short gastrulation available to transport Dpp, such that the gradient can become steeper. In this way, the student is stimulated to carefully consider each step and is provided with sufficient support to decide which is a useful step to follow. In addition, with this type of feedback, explanations are given that relate quantitative changes in the model to qualitative changes in its behaviour, which should increase the student's understanding of the behaviour of the biological model. We consider it important that students, while using the module, are not distracted too much by quantitative issues from the actual biological principles and facts. These have to be mastered in order to obtain a strong biological background. If students want to learn more advanced quantitative skills, they can still follow courses that are specifically aimed at this aspect. The Future Quantitative analysis is already gaining importance in molecular life sciences. Therefore, it is desirable that curriculum changes are implemented in the short term. This poses challenges to faculties, especially to those whose members do not have much, if any, experience with the application of quantitative tools in their own research. Therefore, it may be useful to initially focus on the development of learning materials that are rather self-contained, such that their application requires relatively little competence in quantitative analysis from the teaching staff. If these materials are openly available they can be incorporated rapidly into existing courses, such that even the current generation of students may be better prepared to integrate quantitative thinking and biology in their future research.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544540.xml
554975	The Vpr protein from HIV-1: distinct roles along the viral life cycle	The genomes of human and simian immunodeficiency viruses (HIV and SIV) encode the gag, pol and env genes and contain at least six supplementary open reading frames termed tat , rev , nef , vif , vpr , vpx and vpu . While the tat and rev genes encode regulatory proteins absolutely required for virus replication, nef , vif , vpr, vpx and vpu encode for small proteins referred to "auxiliary" (or "accessory"), since their expression is usually dispensable for virus growth in many in vitro systems. However, these auxiliary proteins are essential for viral replication and pathogenesis in vivo . The two vpr - and vpx -related genes are found only in members of the HIV-2/SIVsm/SIVmac group, whereas primate lentiviruses from other lineages (HIV-1, SIVcpz, SIVagm, SIVmnd and SIVsyk) contain a single vpr gene. In this review, we will mainly focus on vpr from HIV-1 and discuss the most recent developments in our understanding of Vpr functions and its role during the virus replication cycle.	Introduction The viral protein R (Vpr) of HIV-1 is a small basic protein (14 kDa) of 96 amino acids, and is well conserved in HIV-1, HIV-2 and SIV [ 1 ]. The role of Vpr in the pathogenesis of AIDS is undeniable, but its real functions during the natural course of infection are still subject to debate. The Vpr role in the pathophysiology of AIDS has been investigated in rhesus monkeys experimentally infected with SIVmac, and it was initially shown that monkeys infected with a vpr null SIV mutant decreased virus replication and delayed disease progression [ 2 , 3 ]. Moreover, monkeys infected with a SIV that did not express the vpr and vpx genes displayed a very low virus burden and did not develop immunodeficiency disease [ 4 , 5 ]. Regarding these in vivo phenotypic effects, numerous laboratories have dissected the role of Vpr in various in vitro , in vivo and ex vivo systems to explore the contribution of this protein in the different steps of the virus life cycle. Despite its small size, Vpr has been shown to play multiple functions during virus replication, including an effect on the accuracy of the reverse-transcription process, the nuclear import of the viral DNA as a component of the pre-integration complex (PIC), cell cycle progression, regulation of apoptosis, and the transactivation of the HIV-LTR as well as host cell genes (Fig. 1 ). Furthermore, Vpr is found in virions, in cells, and exists as free molecules found in the sera and the cerebrospinal fluid of AIDS patients, indicating that it may exert its biological functions through different manners. Figure 1 Schematic view of the early steps of the HIV-1 infection of a target cell. The functional events in which the Vpr protein is involved are highlighted. Vpr has been shown to play multiple functions during the virus life cycle, including an effect on the accuracy of the reverse-transcription process, the nuclear import of the viral DNA as a component of the pre-integration complex, cell cycle progression, regulation of apoptosis, and the transactivation of the HIV-LTR as well as host cell genes. Structure of the HIV-1 Vpr protein Because the full length protein aggregated in aqueous solution, the overall structure of Vpr has been difficult to access [ 6 ], and preliminary strategies used two distinct synthetic peptides corresponding to Vpr (1–51) and (52–96) fragments for NMR and circular dichroism studies [ 6 - 9 ]. As previously predicted [ 10 ], the structure of the Vpr(1–51) fragment has a long motif of α helix turn-α helix type encompassing the Asp17-Ile46 region, and ends with a γ turn [ 8 ]. The Vpr(52–96) fragment contains an α-helix encompassing the 53–78 region that is rich in leucine residues [ 7 ]. One side of the helix offers a stretch of hydrophobic residues that can form a leucine-zipper like motif [ 11 ]. This structure may account for the formation of Vpr dimers [ 7 , 12 , 13 ] and/or for the interaction with cellular partners [ 14 ]. Finally, NMR analysis of a soluble full length Vpr (1–96) polypeptide was recently performed, and gave access to the tertiary structure of the protein (Fig. 2 ), confirming the amphipathic nature of the three α-helices of HIV-1 Vpr. The helices are connected by loops and are folded around a hydrophobic core [ 15 ] surrounded by a flexible N-terminal domain and a C-terminal arginine-rich region that are negatively and positively charged, respectively. Four conserved prolines (positions 5, 10, 14 and 35) which present cis/trans isomerization are found in the N-terminal domain [ 16 ]. It was reported that the cellular peptidyl-propyl isomerase cyclophilin A was able to interact with Vpr via prolines in position 14 and 35, which insured the correct folding of the viral protein [ 17 ]. The carboxy-terminus of Vpr contains six arginines between residues 73 and 96. This domain shows similarity with those of arginine-rich protein transduction domains (PTD), and may explain the transducing properties of Vpr, including its ability to cross the cell membrane lipid bilayer [ 6 , 18 - 20 ]. Figure 2 Three-dimensional structure of the HIV-1 Vpr protein (from [15]). The three α-helices (17–33, 38–50, 55–77) are colored in pink, blue and orange, respectively; the loops and flexible domains are in green. We can the Trp54 residue localized between the second and the third a-helix, and that is likely accessible for protein-protein interaction with UNG2 [54]. Vpr is packaged into virus particles Vpr is expressed at a late stage of the virus life cycle, but it is present during the early steps of infection of target cells since it is packaged into virions released from the producing cells. The incorporation of Vpr occurs through a direct interaction with the carboxy-terminal p6 Gag region of the gag -encoded Pr55 Gag precursor [ 21 - 24 ]. While the integrity of the α-helices of Vpr is required for efficient packaging into virions [ 25 ], a leucine-rich motif found in the p6 Gag region of the Pr55 Gag precursor is directly involved in the interaction with Vpr [ 23 , 26 ]. After assembly and proteolytic cleavage of Pr55 Gag in matrix, capsid, nucleocapsid (NCp7), and p6 mature proteins, Vpr is recruited into the conical core of the virus particle [ 27 , 28 ] where it is tightly associated with the viral RNA [ 29 , 30 ]. Interestingly, Vpr displays a higher avidity for NCp7 than for the mature p6 protein [ 23 , 24 , 31 ]. Since p6 is excluded from the virion core [ 27 , 28 ], Vpr could switch from the p6 Gag region of the precursor to the mature NCp7 protein to gain access to the core of the infectious virus particle budding at the cell surface. It seems that Vpr is less avid for the fully processed p6 protein than for the p6 Gag region in the context of the p55 Gag precursor. Because of this differential avidity, Vpr is recruited into to the core of the particle where it could interact with nucleic acids, NCp7 [ 24 , 31 ] and/or the matrix protein [ 32 ]. Since it was estimated that Vpr is efficiently incorporated with a Vpr/Gag ratio of ~1:7 [ 33 ], that may represent 275 molecules of Vpr per virion. The incorporation of Vpr has been also used as a unique tool to target cargoes (i.e., cellular and viral proteins, drugs) into viral particles [ 34 , 35 ]. This property was extensively used to study the respective functions of integrase (IN) and reverse transcriptase (RT) during virus replication by expressing Vpr-IN and Vpr-RT fusions in trans in virus-producing cells [ 36 - 38 ]. This strategy of trans -complementation also allowed the analysis of mutant of IN without altering assembly, maturation and other subsequent viral events [ 37 , 39 ]. Furthermore, Vpr fused to the green fluorescence protein (GFP) has been recently used to tag HIV particles in order to follow intracellular virus behavior during the early steps of infection of target cells [ 40 , 41 ]. Vpr influences the fidelity of the reverse transcription process Following virus entry, the viral core is released into the cytoplasm of the target cell and the reverse transcription of the viral RNA takes place in the cytoplasm within a large nucleoprotein complex termed the reverse transcription complex (RTC) containing the two copies of viral RNA and the viral proteins: RT, IN, NCp7, Vpr and a few molecules of the matrix protein [ 42 - 46 ]. It is generally believed that the reverse transcription process is initiated in virus particles and is then completed, after virus entry, in the cystosol of the target cell. This process is likely concomitant of both virus uncoating and trafficking through the cytosol (for reviews, see [ 47 , 48 ]). Recent studies confirmed that Vpr co-localizes with viral nucleic acids and IN within purified HIV-1 RTCs [ 41 , 45 , 49 ], and remains associated with the viral DNA within 4 to 16 h after acute infection [ 43 ]. In addition to a potential role in the initiation step of the reverse transcription process [ 50 ], it has been shown that Vpr modulates the in vivo mutation rate of HIV-1 by influencing the accuracy of the reverse transcription. The HIV-1 RT is an error-prone RNA dependant DNA polymerase, and quantification of the in vivo rate of forward virus mutation per replication cycle revealed that the mutation rate was as much as fourfold higher in the absence of Vpr expression when measured in actively dividing cells using a genetically engineered system [ 51 , 52 ]. Furthermore, recent analysis in non-dividing cells shows that this phenotype is exacerbated in primary monocyte-derived macrophages (MDM) leading to a 18-fold increase of the HIV-1 mutation frequency [ 53 ]. This activity strikingly correlates with the interaction of Vpr with the nuclear form of uracil DNA glycosylase (UNG2) [ 54 ], an enzyme involved in the base excision repair pathway that specifically removes the RNA base uracil from DNA. Uracil can occur in DNA either by misincorporation of dUTP or by cytosine deamination. Initially identified from a yeast two-hybrid screening using Vpr as a bait, the interaction with UNG was confirmed both in vitro and ex vivo in Vpr-expressing cells. While the Trp residue in position 54 located in the exposed loop connecting the second and the third α-helix of HIV-1 Vpr has been shown critical to maintain the interaction with UNG, the Vpr-binding site was mapped within the C-terminal part of UNG2 and occurs through a TrpXXPhe motif. Currently, three distinct cellular partners of Vpr contain a WXXF motif including the TFIIB transcription factor, the adenosine-nucleotide translocator (ANT) and UNG2 [ 55 , 56 ]. The association of Vpr with UNG2 in virus-producing cells allows the incorporation of a catalytically active enzyme into HIV-1 particles where UNG2 may directly influence the reverse transcription accuracy [ 54 ], and this plays a specific role in the modulation of the virus mutation rate. The model supporting the direct contribution of incorporated UNG2 in the reverse transcription process was recently demonstrated by using an experimental system in which UNG2 was recruited into virions independently of Vpr. UNG2 was expressed as a chimeric protein fused to the C-terminal extremity of the VprW54R mutant, a Vpr variant that fails to recruit UNG2 into virions and to influence the virus mutation rate, even though it is incorporated as efficiently as the wild type (wt) Vpr protein. The VprW54R-UNG fusion is also efficiently packaged into HIV-1 virions and restores a mutation rate equivalent to that observed with the wt Vpr, both in actively dividing cells and in MDMs. In agreement with this phenotype on the virus mutation frequency, it was finally documented that the Vpr-mediated incorporation of UNG2 into virus particles contributes to the ability of HIV-1 to replicate in primary macrophages. When the VprW54R variant was introduced into an infectious HIV-1 molecular clone, virus replication in MDMs was both reduced and delayed whereas replication in PBMC was not altered by the lack of UNG2 incorporation into virus particles. Although it was proposed that the viral integrase was also able to mediate interaction with UNG2, Vpr seems the main viral determinant that allows for the incorporation of cellular UNG2 into virus particles. However, preliminary results obtained from in vitro binding assays suggest that both Vpr and IN associate with UNG to form a trimeric complex (ELR and SB, unpublished results), but further analyses are required to document the nature of the interactions between UNG2, Vpr, IN as well as RT both in virus-producing cells and then in target cells. HIV-1 and other lentiviruses are unusual among retroviruses in their ability to infect resting or terminally differentiated cells. While Vpr has been shown to facilitate the nuclear import of viral DNA in non-dividing cells, the virion incorporation of UNG2 via Vpr also contributes to the ability of HIV-1 to replicate in primary macrophages. This implies that UNG2 is a cellular factor that plays an important role in the early steps of the HIV-1 replication cycle (i. e. viral DNA synthesis). This observation is in good agreement with a recent report showing that the misincorporation of uracil into minus strand viral DNA affects the initiation of the plus strand DNA synthesis in vitro [ 57 ]. This observation suggests that UNG is likely recruited into HIV-1 particles to subsequently minimize the detrimental accumulation of uracil into the newly synthesized proviral DNA. While further work is needed to explain the precise mechanism for how UNG catalytic activity may specifically influence HIV-1 replication in macrophages, it is worth noting that nondividing cells express low levels of UNG and contain relatively high levels of dUTP [ 58 ]. Similarly, most non-primate lentiviruses, such as feline immunodeficiency virus (FIV), caprine-arthritis-encephalitis virus (CAEV) and equine infectious anemia (EIAV), have also developed an efficient strategy to reduce accumulation of uracil into viral DNA. These lentiviruses encode and package a dUTP pyropshophatase (dUTPase) into virus particles, an enzyme that hydrolyzed dUTP to dUMP, and thus maintains a low level of dUTP. Interestingly, replication of FIV, CAEV or EIAV that lack functional dUTPase activity is severely affected in nondividing host cells (e.g., primary macrophages). Taken together, these results indicate that uracil misincorporation in viral DNA strands during reverse transcription is deleterious for the ongoing steps of the virus life cycle. The presence of a viral dUTPase or a cellular UNG will prevent these detrimental effects for replication of non-primate and primate lentiviruses in macrophages, respectively. In addition, it is intriguing to note that two viral auxiliary proteins from HIV-1, Vpr and Vif, can both influence the fidelity of viral DNA synthesis. The Vif protein forms a complex with the cellular deaminase APOBEC-3G (CEM15) preventing its encapsidation into virions [ 59 - 63 ], while Vpr binds the DNA repair enzyme, UNG, to recruit it into the particles. It is tempting to speculate that the action of both viral proteins may influence the mutation rate during the course of HIV-1 infection, and their balance may play a key role during disease progression in infected individuals. Vpr and the nuclear import of the viral pre-integration complex Nondividing cells, such as resting T cells and terminally-differentiated macrophages, are important targets for viral replication during the initial stages of infection, since primary infection of these cell populations contributes to the establishment of virus reservoirs, crucial for subsequent virus spread to lymphoid organs and T-helper lymphocytes [ 64 ]. Infection of lymphoid histoculture using human tonsil or splenic tissue showed that Vpr greatly enhances HIV replication in macrophages but did not influence productive infection of proliferating or resting T cells [ 65 ]. After virus entry into the cell, the viral capsid is rapidly uncoated and the reverse transcription of the genomic HIV-1 RNA leading to the full length double-strand DNA is completed. This viral DNA associates with viral and host cell proteins into the so-called pre-integration complex (PIC). In contrast to oncoretroviruses which require nuclear envelope disintegration during mitosis to integrate their viral genome into host chromosomes, lentiviruses, such HIV and SIV, have evolved a strategy to import their own genome through the envelope of the interphasic nucleus via an active mechanism 4–6 h after infection (for review, see [ 66 ]). Vpr has been reported to enhance the transport of the viral DNA into the nucleus of nondividing cells [ 67 - 69 ], by promoting direct or indirect interactions with the cellular machinery regulating the nucleo-cytoplasmic shuttling [ 70 - 74 ]. PIC en route to the NE The exact composition of the PIC is still an area of debate but it contains the viral DNA at least associated with integrase, and many recent studies have confirmed that Vpr is also an integral component of this complex (for reviews, see [ 75 - 77 ]). Of course, the PIC likely contains cellular factors that participate in both intra-cytoplasmic routing and nuclear translocation of the viral DNA. While actin microfilaments seem to play a role in the early events of infection by acting as a scaffold for the appropriate localization and activation of the RTC [ 78 ], the PIC is tightly associated with microtubular structures in the cytoplasm. An elegant system using Vpr fused to GFP as a probe was developed to follow the movement of the PIC soon after virus entry in living cells [ 40 ]. It has been shown that the GFP-Vpr labeled-PIC progresses throughout the cytoplasm along cytoskeletal filaments and then accumulates in the perinuclear region close to centrosomes. More precisely, it was observed that the viral complex uses the cytoplasmic dynein motor to travel along the microtubule network to migrate towards the nucleus. It is not yet known whether Vpr plays an active role during this movement of the PIC along microtubules or whether it is only associated with the complex and then actively participates in the subsequent steps, including the anchoring of the PIC to the nuclear envelope (NE) and the nuclear translocation of the viral DNA. Vpr docks at the NE Indeed, Vpr displays evident karyophilic properties and localizes in the nucleus, but a significant fraction is anchored at the NE and can be visualized as a nuclear rim staining in fluorescence microscopy experiments [ 73 , 79 - 81 ]. The NE consists of two concentric inner and outer membranes studded with nuclear pore complexes (NPC) that form a conduit with a central aqueous channel which allows selective trafficking between the nucleus and cytoplasm and creates a permeability barrier to free diffusion of macromolecules or complexes. NPC corresponds to a 125-MDa structure consisting of 30 distinct nuclear pore proteins, named nucleoporins (Nups) [ 82 ]. A specific subset of Nups contain FG- or FxFG peptide repeats that constitute most of the filamentous structures emanating from both sides of the NPC and that provide docking sites for various transport factors [ 83 ]. Initial studies revealed that HIV-1 Vpr bound to the FG-rich region of several nucleoporins including the human p54 and p58 Nups, the rodent POM121, and the yeast NUP1P [ 71 , 73 , 74 ], but a direct interaction with the human CG1 nucleoporin was more recently reported [ 70 ]. This interaction is not mediated by the FG-repeat region of this Nup but rather via a region without consensus motif located in the N-terminus of the protein. Using an in vitro nuclear import assay, it has been demonstrated that the association with the N-terminal region of hCG1 is required for the docking of Vpr to the NE, whereas the FG-repeat region does not participate in this process [ 70 ]. The role of Vpr at the NE is not clear but two explanations can be proposed. First, this localization may account for the targeting of the PIC to the NPC before its translocation into the nuclear compartment. In this model, the virion-associated Vpr would be primarily involved, after virus entry and uncoating, in the initial docking step of the viral DNA to the NPC, while other karyophilic determinants of the PIC, such as IN, would then allow for the second step of nuclear translocation to proceed [ 81 , 84 - 86 ]. Alternatively, another explanation may come from the observation that Vpr was able to provoke herniations and transient ruptures of the NE [ 87 ]. The molecular mechanism supporting the local bursting induced by Vpr is not known but the interaction of Vpr with nucleoporins may cause initial misassembly of the NPC leading to alterations of the NE architecture. Consequently, these transient ruptures may provide an unconventional route for nuclear entry of the viral PIC [ 87 , 88 ]. Translocation of Vpr into the nucleus Despite the lack of any identifiable canonical nuclear localization signal (NLS), Vpr displays evident karyophilic properties and is rapidly targeted to the host cell nucleus after infection [ 89 ]. Even though the small size of Vpr does not strictly require an NLS-dependent process, experiments performed both in vitro or in transfected cells have shown that Vpr is able to actively promote nuclear import of a reporter protein, such as BSA, β-galastosidase or GFP [ 10 , 13 , 90 - 94 ]. Like proteins containing a basic-type NLS, it was initially proposed that Vpr uses an importin α-dependant pathway to access the nuclear compartment [ 72 , 73 ]. In addition, Vpr may enhance the inherently low affinity of the viral MA for importin α to allow nuclear import of MA [ 95 , 96 ], but conflicting data exists on the nuclear localization of this viral protein [ 81 , 85 ]. Finally, it was reported that Vpr nuclear import was mediated by an unidentified pathway, distinct from the classical NLS- and M9-dependant pathways [ 92 ]. Two independent nuclear targeting signals have been characterized within the HIV-1 Vpr sequence, one spanning the α-helical domains in the N-terminal part of the protein and the other within the arginine-rich C-terminal region [ 92 , 94 ]. These results are consistent with data showing that the structure of the α-helical domains of Vpr must be maintained both for its nuclear localization and for Vpr binding with nucleoporins [ 25 , 70 , 80 ]. In conclusion, the nucleophilic property of Vpr and its high affinity for the NPC, associated with its presence in the viral PIC, at least support a role during the docking step of the PIC at the NE, a prerequisite before the translocation of viral DNA into the nucleus. Even though there is no evidence that Vpr directly participates in the translocation process, it is worth noting that purified PICs also dock at the NE before nuclear translocation using a pathway also distinct from the NLS and M9 nuclear import pathways [ 49 ]. One can suggest that among the redundancy of nuclear localization signals characterized within the PIC, both in associated viral proteins (i.e. IN, MA, Vpr) and also in the viral DNA [ 97 ], Vpr primarily serves to dock the PIC at the NE, while IN and MA act in cooperation with the central DNA flap to target the viral DNA to the nucleus (for review, see [ 98 ]). Vpr, a nucleocytoplasmic protein In addition to its nonconventional NLS for targeting into the nucleus, Vpr is a dynamic mobile protein able to shuttle between the nucleus and cytoplasmic compartments [ 23 , 99 , 100 ]. Photobleaching experiments on living cells expressing a Vpr-GFP fusion confirmed that Vpr displays nucleocytoplasmic shuttling properties [ 70 ]. This shuttling activity has been related to the distal leucine-rich helix which could form a classical CRM1-dependant nuclear export signal (NES) [ 99 ]. The exact role of this NES in the function of Vpr is not known but since Vpr is rapidly imported into the nucleus after biosynthesis, the NES could redirect it into the cytoplasm for subsequent incorporation into virions through direct binding to the viral p55 Gag precursor during the late budding step of the virus life cycle [ 23 , 100 ]. Vpr and the cell cycle A further important biological activity of SIV and HIV Vpr proteins is related to their ability to induce an arrest in the G2 phase of the cell cycle of infected proliferating human and simian T cells [ 91 , 101 - 105 ]. Cell cycle arrest does not require de novo synthesis of Vpr, but is induced by Vpr molecules packaged into infecting virions [ 87 , 106 ]. This indicates that induction of the G2 cell cycle arrest might happen before the integration step of the viral DNA genome. It is noteworthy that the S. pombe fission yeast as well as S. cerevisiae overexpressing HIV-1 Vpr are also blocked in the G2 phase of the cell cycle [ 107 - 109 ], supporting the idea that the cellular pathway altered by Vpr is well conserved in all eukaryotic cells. Moreover, infection of caprine cells with a caprine arthritis encephalitis virus (CAEV) expressing the vpr gene from SIV similarly provoked a G2 arrest [ 110 ]. The biological significance of this arrest during the natural infection is not well understood, but the HIV-1 LTR seems to be more active in the G2 phase, implying that the G2 arrest may confer a favorable cellular environment for efficient transcription of HIV-1 [ 111 ]. In agreement, the Vpr-induced G2 arrest correlates with high level of viral replication in primary human T cells. The determinants of the G2 arrest activity are mainly located in the C-terminal unstructured basic region of HIV-1 Vpr and phosphorylation of the protein is required [ 112 , 113 ]. Regulators of the cell cycle, such as cyclin-dependant kinases (CDKs), control progression through the cell cycle by reversible phosphorylation [ 114 ]. The p34/cdc2 CDK associates with cyclin B1 in the G2 phase (for review, see [ 115 ]) to regulate the G2 to M transition. Accumulation of the cells expressing Vpr in the G2 phase has been correlated to the inactivation of the p34/cdc2-cyclinB kinase [ 102 , 103 ]. The activity of cdc2 is controlled by opposite effects of the Wee-1 and Myt1 kinases and the cdc25 phosphatase. Wee1 inhibits cdc2 activity through tyrosine phosphorylation, while dephosphorylation of cdc2 by the phosphatase cdc25 promotes cdc2-cyclinB activation that drives cells into mitosis. The activities of both cdc25 and Wee-1 are also regulated by phosphorylation/dephosphorylation. It was initially described that Vpr-expressing cells contained both hyperphosphorylated cdc2 and hypophosphorylated cdc25, their inactive status [ 101 - 103 ]. Consequently, these two regulators of the G2/M switch are blocked preventing any cell cycle progression. The molecular mechanism leading to this inhibition is not yet clear, but different cellular partners interacting with Vpr which could play a role in cell cycle regulation have been proposed as potential mediators of the Vpr-induced G2 arrest. hVIP/MOV34, a member of the eIF3 complex, was identified as a Vpr-partner in a yeast two-hybrid assay [ 116 ], and was associated with the cell cycle arrest activity of Vpr [ 117 ]. eIF3 is a large multimeric complex that regulates transcriptional events and is essential for both G1/S and G2/M progression. Intracellular localization studies revealed that expression of Vpr induces a relocalization of MOV34 that shifts from a cytoplasmic to a nuclear localization pattern [ 116 , 117 ]. Two other cellular partners of Vpr, UNG and HHR23A (i.e., the human homologue of the yeast rad23 protein), are implicated cellular DNA repair processes. Since a clear relationship exists between the DNA damage response pathway and the progression of the cell cycle, it was initially suggested that Vpr binding to these DNA repair proteins could account for the observed G2 arrest [ 118 - 120 ], but subsequent analyses indicated that there was no correlation between the association of Vpr with HHR23A and/or UNG and the block in G2 [ 121 , 122 ]. These analyses are in agreement with a previous report showing that the Vpr-mediated arrest is distinct from the cell cycle arrest in G2 related to DNA damage. However, it has also been reported that Vpr induces cell cycle arrest via a DNA damage-sensitive pathway [ 123 ]. The G2 DNA damage checkpoint is under the control of the phosphatidylinositol 3-kinase-like proteins, ATR and ATM [ 124 ], which lead to the inactivation of the cdc2-cyclinB complex. The ATR protein has been recently linked to the G2-arrest induced by Vpr [ 125 ]. Inhibition of ATR either by drugs, a dominant-negative form of ATR or by siRNA reverts the Vpr-induced cell cycle arrest while activation of ATR by Vpr results in Chk1 phosphorylation, the kinase regulating cdc25c activity. These authors suggested that the G2 arrest induced by Vpr parallels the ATR-DNA damage pathway, but additional work is needed to demonstrate that Vpr causes DNA damage or mimics a signal activating one of the DNA damage sensors. The protein phosphatase 2A (PP2A) has been shown to be directly associated with Vpr via its B55α subunit [ 126 ]. PP2A is a serine/threonine phosphatase involved in a broad range of cellular processes, including cell cycle progression. PP2A inactivates cdc2 indirectly both by the inactivation of the Wee1 kinase and by activation of cdc25 (for review, see [ 127 ]). Genetic studies performed in S. pombe suggest the involvement of PP2A and Wee1 in the Vpr-induced cell cycle arrest [ 128 ]. Intriguingly, expression of Vpr and B55α results in the nuclear localization of B55α subunit while it remains cytoplasmic in normal condition. Together, these studies emphasized the fact that Vpr might play a role in the subcellular redistribution of several regulatory protein complexes involved in the progression of the cell cycle. Indeed, the mitotic function of cdc2-cyclinB complex is triggered not only by the turn of phosphorylation/desphorylation of both subunits on specific residues, but also by spatio-temporal control of their intracellular distribution. For example, cyclinB is predominantly cytoplasmic throughout the G2 phase until it translocates rapidly into the nucleus 10 min before nuclear envelope breakdown [ 129 ]. As mentioned earlier, Vpr induces herniations and local bursting of the nuclear envelope leading to redistribution of key cell cycle regulators, including Wee1, cdc25, and cyclin B into the cytoplasm of the host cell [ 87 ]. It seems evident that alterations of the subcellular localization of segregated cell cycle regulators could explain the G2 arrest induced by Vpr; this may also explain the overall variety of cellular factors that have been involved in this process. Alternatively, nuclear herniations induced by Vpr could also affect chromatin structure leading to the activation of ATR. However, it not known if the Vpr-induced alteration of the NE architecture could cause DNA damage such as double-strand breaks, but disruption of the nuclear lamin structure is sufficient to block DNA replication, another abnormality recognized by the ATR protein (for reviews, see [ 130 , 131 ]). Vpr and apoptosis HIV infection causes a depletion of CD4 + T cells in AIDS patients, which results in a weakened immune system, impairing its ability to fight infections. The major mechanism for CD4 + T cell depletion is programmed cell death, or apoptosis, that can be induced by HIV through multiple pathways of both infected cells and non-infected "bystander" cells (for review, see [ 132 ]). Even though the exact contribution of Vpr as a pro-apoptotic factor responsible for the T cell depletion observed in the natural course of HIV infection is still unknown, it was repeatedly evidenced that Vpr has cytotoxic potential and is able to induce apoptosis in many in vitro systems. In addition, transgenic mice expressing Vpr under the control of the CD4 promoter show both CD4 and CD8 T cell depletion associated with thymic atrophy [ 133 ]. However, controversial results indicating that Vpr can also act as negative regulator of T cell apoptosis have been reported [ 134 , 135 ]. Initially proposed as a consequence of the prolonged cell cycle arrest [ 136 - 140 ], other investigations have then revealed that the Vpr-mediated G2 arrest was not a prerequisite for induction of apoptosis, suggesting that both functions are separated [ 79 , 87 , 141 , 142 ]. However, the recent observation that the activity of the cell cycle regulatory Wee-1 kinase is decreased in Vpr-induced apoptotic cells led to the hypothesis of a direct correlation between the G2 arrest and apoptotic properties of Vpr [ 143 ]. Hence, reduction of Wee-1 activity, probably related to its delocalization provoked by Vpr [ 87 ], results in an inappropriate activation of cdc2 leading to cell death with phenotypical aberrant mitotic features, a process known as mitotic catastrophe [ 144 , 145 ]. Using an established cell line expressing Vpr, it was observed that after the long G2 phase, cell rounded up with aberrant M-phase spindle with multiple poles resulting from abnormal centrosome duplication [ 138 , 146 ]. The cells stopped prematurely in pro-metaphase and died by subsequent apoptosis. However, works from the G. Kroemer's group have then well established that synthetic Vpr, as well as truncated polypeptides, are able to induce apoptosis by directly acting on mitochondria leading to the permeabilization of the mitochondrial membrane and subsequent dissipation of the mitochondrial transmembrane potential (ΔΨm) [ 56 ]. This direct effect of Vpr was related to its ability to interact physically with the adenine nucleotide translocator (ANT), a component of the permeability transition pore of mitochondria localized in the inner mitochondrial membrane. Since ANT is a transmembrane protein and presents a WxxF motif on the inner membrane face which is recognized by Vpr [ 56 , 147 ], this interaction implies that Vpr must first cross the outer mitochondria membrane to access ANT. The interaction between Vpr and ANT triggers permeabilization of the inner membrane followed by permeabilization of the outer mitochondrial membrane with consequent release of soluble intermembrane proteins, such as cytochrome c and apoptosis inducing factors, in the cytosol. Cytochrome c then associates with Apaf-1 in a complex with caspase-9 to create the apoptosome, allowing activation of effector caspases, such as caspase-3, and subsequently the final execution of the apoptotic process (for review, see [ 148 ]). While numerous reports have shown that Vpr mediated-apoptosis was associated with activation of caspase-9 and capase-3 [ 56 , 79 , 137 , 140 , 147 , 149 ], it is intriguing that Vpr was still able to induce cell death in embryonic stem cells lacking Apaf-1, caspase-9 and IAF [ 150 ]. These results suggest a model in which the direct action of Vpr on mitochondria may be sufficient to cause cell death in HIV-1 infected cells [ 149 ]. Although the causal role of Vpr in the induction of apoptosis is evident both in vitro and ex vivo , its real contribution with other viral determinants, such as gp120 envelope, Tat, Nef and the viral protease, in the physiopathology of AIDS needs to be further documented during the course of HIV infection [ 151 ]. However, it was recently revealed that long term non-progressor HIV-1 infected patients show a highest frequency of mutation at the position Arg77 of the Vpr protein than patients with progressive AIDS disease. Interestingly, this residue seems crucial for the capacity of the protein to induce apoptosis through permeabilization of the mitochondrial membrane [ 152 ]. Conversely, it was reported that mutation of the Leu64 residue enhanced the pro-apoptopic activity of Vpr [ 153 ], indicating that mutations affecting the C-terminal region of the protein may generate Vpr molecules with different pro-apoptotic potentials during the course of natural HIV-1 infection. In addition, soluble Vpr protein is found in the sera as well as in the cerebrospinal fluid of HIV-infected patients, and was proposed to play a role related to its pro-apoptotic activity in AIDS-associated dementia [ 154 , 155 ]. The involvement of Vpr in these neurological disorders has been suggested, since recombinant Vpr has neurocytopathic effects on both rat and human neuronal cells [ 156 - 158 ]. Neurons killed by extracellular Vpr display typical features of apoptosis evidenced by direct activation of the initiator caspase-8 that will lead to subsequent activation of effector caspases. These effects have been linked to the property of the first amphipathic α-helix of Vpr to form cation-selective ion channels in planar lipid bilayers, causing a depolarization of the plasma membrane [ 6 , 157 , 159 , 160 ]. These observations indicate that Vpr can trigger apoptotic processes by different alternative pathways depending of the target cells. Nuclear role(s) of Vpr The first reported function of Vpr was a modest transcriptional activity on the viral LTR promotor as well as on heterologous cellular promotors [ 161 , 162 ]. While the connection between cell cycle arrest and LTR-transactivation by Vpr is not well understood, it was concluded that activation of the Vpr-induced viral transcription is secondary to its G2/M arrest function [ 111 , 163 ]. An increase transcriptional activity is indeed observed from the viral LTR in arrested cells expressing Vpr [ 164 - 166 ]. The transactivation of HIV-1 induced by Vpr is mediated through cis -acting elements, including NF-κB, Sp1, C/EBP and the GRE enhancer sequences found in the LTR promotor [ 167 - 170 ]. Also related to this activity, Vpr regulates the expression of host cell genes such as NF-κB, NF-IL-6, p21 Waf1 and survivin [ 171 - 173 ]. Finally, Vpr seems also able to interact directly with the ubiquitous cellular transcription factor Sp1 [ 168 ], the glucocorticoid receptor [ 174 , 175 ], the p300 coactivator [ 163 , 176 ], and with the transcription factor TFIIB, a component of the basal transcriptional machinery [ 177 ]. This latter interaction is also mediated by a WxxF motif found within the TFIIB primary sequence [ 55 ]. Vpr displays high affinity for nucleic acids but no specific DNA sequence targeted by Vpr has been yet identified [ 19 , 29 ]. Interestingly, Vpr does not bind to the Sp1 factor or cis -acting elements alone but it associates with Sp1 in the context of the G/C box array [ 168 ], as well as in a ternary complex with p53 [ 178 ], indicating that Vpr might bind specific DNA sequence once associated with cellular partners to subsequently drive expression of both host cell and viral genes. Consistently, it has been reported that Vpr can directly bind to p300 via a LXXLL motif present in the C-terminal α-helix of the protein [ 179 ], suggesting that Vpr may act by recruiting the p300/CBP co-activators to the HIV-1 LTR promotor and thus enhance viral expression. Since p300 is a co-activator of NF-κB, Vpr can also mediate up-regulation of promotors containing NF-κB and NF-IL-6 enhancer sequences in primary T cells and macrophages. In addition, Vpr markedly potentiates glucocorticoid receptor (GR) action on its responsive promotors [ 174 , 175 ]. The Vpr-mediated LTR transcription was inhibited by the addition of the GR antagonist, RU486, in cultured macrophages [ 175 ]. That Vpr-mediated co-activation of the GR is distinct from the G2 arrest and required both LLEEL 26 and LQQLL 68 motifs contained within the first and third α-helical domains of HIV-1 Vpr [ 174 , 180 ]. Vpr may also function as an adaptor molecule for an efficient recruitment of transcriptional co-activators (GRE, p300/CBP...) to the HIV-1 LTR promotor and thus enhances viral replication. Additionally, it may be involved in the activation of host cell genes inducing cellular pathways in relation with the AIDS pathogenesis. Indeed, cDNA microarray analysis using isogenic HIV-1 either with or without vpr expression revealed that Vpr induces up and down regulation of various cell genes [ 181 ]. Conclusion By interfering with many distinct cellular pathways all along the virus life cycle, it is now evident that Vpr's contribution to the overall pathogenesis of HIV-1 infection in vivo is likely crucial. While major efforts have been made during the last years to define the molecular mechanisms and cellular targets of Vpr, additional work is needed for the complete understanding of its wide range of activities. An important issue now is to define the precise contribution of each activity to the viral replication and pathogenesis during the natural course of HIV infection. The involvement of Vpr in key processes of the early steps the viral life cycle (i.e., reverse transcription and nuclear import of the viral DNA) represents a good target for developing novel therapeutic strategies for AIDS therapy. In addition, this viral factor represents a valuable tool to elucidate many fundamental cellular processes. List of abbreviations HIV, human immunodeficiency virus; SIV, simian immunodeficiency virus; CypA, cyclophilin A; nup, nucleoporin; PIC, pre-integration complex; RTC, reverse transcription complex.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554975.xml
551602	Evolution of the relaxin-like peptide family	Background The relaxin-like peptide family belongs in the insulin superfamily and consists of 7 peptides of high structural but low sequence similarity; relaxin-1, 2 and 3, and the insulin-like (INSL) peptides, INSL3, INSL4, INSL5 and INSL6. The functions of relaxin-3, INSL4, INSL5, INSL6 remain uncharacterised. The evolution of this family has been contentious; high sequence variability is seen between closely related species, while distantly related species show high similarity; an invertebrate relaxin sequence has been reported, while a relaxin gene has not been found in the avian and ruminant lineages. Results Sequence similarity searches of genomic and EST data identified homologs of relaxin-like peptides in mammals, and non-mammalian vertebrates such as fish. Phylogenetic analysis was used to resolve the evolution of the family. Searches were unable to identify an invertebrate relaxin-like peptide. The published relaxin cDNA sequence in the tunicate, Ciona intestinalis was not present in the completed C. intestinalis genome. The newly discovered relaxin-3 is likely to be the ancestral relaxin. Multiple relaxin-3-like sequences are present in fugu fish ( Takifugu rubripes ) and zebrafish ( Danio rerio ), but these appear to be specific to the fish lineage. Possible relaxin-1 and INSL5 homologs were also identified in fish and frog species, placing their emergence prior to mammalia, earlier than previously believed. Furthermore, estimates of synonymous and nonsynonymous substitution rates ( d N /d S ) suggest that the emergence of relaxin-1, INSL4 and INSL6 during mammalia was driven by positive Darwinian selection, hence these peptides are likely to have novel and in the case of relaxin-1, which is still under positive selection in humans and the great apes, possibly still evolving functions. In contrast, relaxin-3 is constrained by strong purifying selection, demonstrating it must have a highly conserved function, supporting its hypothesized important neuropeptide role. Conclusions We present a phylogeny describing the evolutionary history of the relaxin-like peptide family and show that positive selection has driven the evolution of the most recent members of the family.	Background The relaxin-like peptide family includes: relaxin-1, relaxin-2, relaxin-3, and the insulin-like (INSL) peptides, INSL3, INSL4, INSL5 and INSL6. All share high structural similarity with insulin due to the presence of six cysteine residues, which confer two inter-chain and one intra-chain disulfide bonds. Thus, it was postulated that relaxin and insulin had derived from a common ancestral gene and were therefore grouped as the insulin superfamily, which later included insulin-like growth factors I and II (IGF-1 and -2) (reviewed in [ 1 ]). Despite less than 50% predicted sequence similarity across members of the insulin superfamily, primary structural determinants are retained, resulting in the similar tertiary structures of relaxin and insulin (reviewed in [ 1 ]). The structures of insulin, relaxin [ 1 ], relaxin-3 [ 2 ] and INSL3 [ 3 ] are formed by the cleavage of the pro-hormone peptide into three chains (A, B and C), removal of the C chain and the formation of three disulfide bridges between six invariant cysteine residues found on the A and B chains, to produce an active protein. Based on primary sequence similarity the native structures of human relaxin-1 (H1 relaxin), INSL4, 5 and 6 should be similar, but to date this has not been confirmed. Various studies have highlighted the importance of a [Arg-X-X-X-Arg-X-X-Ile] motif in the relaxin B chain for interaction with the relaxin receptor and biological activity, which is dependent upon the presence of this motif [ 4 ]. Coupled with the insulin-like cysteine bond pattern, the presence of this motif is used to distinguish relaxin sequences. Hence, three relaxin genes are present in humans: relaxin-1 is found only in humans and the great apes, its expression is limited to the decidua, placenta and prostate [ 5 ]; relaxin-2 is the major circulating form of relaxin in the human [ 6 ] and the functional equivalent to the relaxin-1 in all non-primates; while relaxin-3 was only recently discovered and shows brain specific expression [ 2 ]. Throughout this paper relaxin will be used to refer to relaxin-2 in humans and great apes, and the equivalent relaxin-1 in all other mammals (table 1 ). Table 1 The relaxin-2 gene in humans and great apes is the equivalent of the relaxin-1 gene in non-primate species and will be referred to as relaxin throughout this paper. Non-primate species only have RLN1 and RLN3 genes. Humans and Great Apes All other mammals e.g. mouse Peptide abbreviation Gene name Relaxin-1 Relaxin # RLX1 RLN1 Relaxin-2 # RLX2 RLN2 Relaxin-3 Relaxin-3 RLX3 RLN3 # indicates functional equivalence Relaxin has been well characterized in a reproductive context.. It is a product of the ovary and/or placenta in most species studied, and has various roles in pregnancy and parturition, which differ between species [ 7 ]. Recent advances have revealed relaxin to be a multifunctional hormone, with numerous non-reproductive roles (reviewed in [ 1 ]). Of the other relaxin-like peptides, INSL3, or relaxin-like factor (RLF), is closely related to relaxin and is critical for testis descent [ 8 , 9 ]. Early placenta insulin-like peptide (EPIL), placentin or insulin-like 4 ( INSL4 ) is primate specific and likely to have diverged from a common relaxin ancestor before the duplication that led to the two relaxin genes seen in humans and great apes [ 10 ]. Insulin-like 5 and 6 ( INSL5 and INSL6 ) were identified by searching EST databases with the cysteine motif conserved in all insulin-like peptides [ 11 , 12 ]. Both INSL5 and 6 peptides show higher sequence similarity to relaxin rather than insulin and are functionally uncharacterised. Until recently, receptors for all peptides of the relaxin family were unknown. Given the high degree of structural similarity between relaxin and insulin, it had been believed that the relaxin receptor would also be a tyrosine kinase receptor, similar to the insulin receptor. However, it was finally demonstrated in 2002, that relaxin activated two previously orphan leucine rich repeat containing heterotrimeric guanine nucleotide binding protein-coupled receptors (GPCR), LGR7 and LGR8 [ 13 ]. Further studies have shown that LGR7 is the relaxin receptor [ 13 ], although it also specifically interacts with relaxin-3 [ 14 ], and that LGR8 is the INSL3 receptor [ 15 , 16 ]. Even more recently, another two GPCRs which are activated by relaxin-3, the somatostatin- and angiotensin- like peptide receptor (SALPR or GPCR135) and GPCR142 have been identified [ 17 , 18 ]. More recently GPCR142 has been shown to be the receptor for INSL5, [ 19 ] while GPCR135 appears to be the specific receptor for relaxin-3. Surprisingly, while LGR7, LGR8, GPCR135 and GPCR142 are all Type I GPCRs, they are from different branches within this family and are only very distantly related [ 17 , 18 , 20 ]. No other relaxin-like peptides have been shown to interact with these receptors [ 13 , 14 , 16 - 18 ]. Relaxin and INSL3 have primarily been of interest as hormones of pregnancy and reproduction; therefore it was assumed that the largely uncharacterised INSL4, 5 and 6 would share similar functions. However the discovery of the brain specific relaxin-3, and the widespread expression of INSL5 and its receptor GPCR142 have resulted in a re-evaluation of these assumptions and raised the need for a new approach to investigating these peptides. Relaxin evolution has confounded researchers for decades. High sequence variability in relaxins across closely related species is a well-known feature of this peptide, however startling similarities have been observed between very distant species such as pigs and whales [ 21 ]. Other studies have reported the existence of an invertebrate relaxin, a hormone with "relaxin-like" properties has been described in protozoa ( T. pyriformis ) [ 22 ], ascidians ( H. momus ) [ 23 ] and tunicates ( C. intestinalis ) [ 23 ]. A cDNA and peptide sequence with almost 100% similarity to porcine relaxin was isolated from C. intestinalis [ 24 ]. Contrastingly, numerous efforts to identify and sequence a relaxin ortholog in bovine and other ruminants have been unsuccessful. A likely non-functional single copy relaxin-like gene has been found in the ovine [ 25 ]. Despite this controversy very few phylogenetic studies of relaxin, or the other relaxin-like peptides, have been reported. An in-depth analysis analysing only relaxin-1, -2 and INSL3 in several primates has been performed [ 26 ] and a more recent paper has discussed the evolution of the family without including a detailed phylogenetic analysis [ 27 ]. The increasing availability of genomic data has provided an opportunity to clarify the area of relaxin evolution using phylogenetic analysis. The relaxin-like peptide family phylogeny shows relaxin-3 is likely to be the ancestral relaxin, emerging prior to the divergence of fish, presumably with a function in the brain. Multiple relaxin-3 sequences are present in the fish and frogs, while the phylogeny also suggests that a relaxin with reproductive functions is present in these lineages. However, relaxin has been lost in the chicken and its genome instead contains two relaxin-3-like sequences. Evolutionary rate analysis shows positive Darwinian selection to be driving the most recent members of the family, INSL6, INSL4 and relaxin-1. Results Sequence similarity searches and multiple alignment Table 2 outlines the relaxin-like peptide sequences used in these phylogenetic analyses, from human, mouse, rat, dog, chimpanzee, pig, chicken, wallaby, R. esculenta , X. tropicalis , X. laevis , fugu fish, zebrafish, rhesus monkey and rainbow trout. New sequences identified during similarity searches have been highlighted. The source of each newly identified sequence is annotated as a footnote (i.e. genomic, ESTs). It should be noted that more mammalian relaxin and INSL3 sequences have been identified than were included in these analyses. Inclusion of these sequences did not improve the accuracy of the phylogeny, and as the aim of this study was to determine the evolutionary history of the entire family, they were omitted. All available non-mammalian relaxin-like sequences with accompanying nucleotide sequences were included. Table 2 GenBank accession numbers for relaxin-like sequences. Accession numbers for all sequences included in the phylogeny are listed. New sequences identified in this study are highlighted in bold and accession numbers shown in brackets underneath. Phylogenetic analyses showed TrRLX3f, DrRLX3c, XtRLX3a and XlRLX3 sequences (shown in italics) to be relaxin rather than relaxin-3 homologs and are therefore listed in the relaxin (RLX2) column. Species RLX1 RLX2 RLX3 INSL3 INSL4 INSL5 INSL6 Human P04808 P04090 Q8WXF3 P51460 Q14641 Q9Y5Q6 Q9Y581 Chimpanzee S42783 P51455 BK005156 b BK005155 b BK005152 b BK005153 b BK005154 b Mouse - CAA81611 Q8CHK2 O09107 - Q9WUG6 Q9QY05 Rat - J00780 Q8BFS3 AAD33663 - - Q9WV41 Dog - Q9TRM8 AAEX01023146 AAEX01022997 AAEX01053723 Wallaby AAM22209 R. esculenta CAC16108 Fugufish RLX3f BK005388 b RLX3a-c RLX3d,e Zebrafish RLX3c BK005255 b RLX3a BK005228 RLX3b BK005252 a,b RLX3d ( b,c ) Chicken RLX3a BK005146 a RLX3b BK005533 b Pig P01348 Q8HY17 P51461 X. tropicalis RLX3a BK005229 a RLX3b ( b,d ) RLX3c ( b,d ) X. laevis RLX3 BK005230 a O. mykiss RLX3 BK005147 a M. mulatta INSL4 BK005251 a a Sequences determined from multiple overlapping ESTs. b Sequences determined from genomic data. c Sequence from the Ensembl zebrafish genome assembly [68](GENSCAN00000013529 on supercontig NA8544). d Sequence from the Xenopus genome v2.0 [58]. Searches of the completed human, mouse and rat genomes did not identify any novel relaxin-like peptides, however, likely non-functional INSL5 genes were revealed in the rat and dog genomes. A sequence with high similarity to INSL5 was found (Genbank Accession No: NW_047717.1), with a frameshift mutation in the first exon of the gene, which introduces a stop codon, resulting in a protein truncated early in the B chain (data not shown). The recently completed dog genome also contains a sequence highly homologous to the human INSL5 peptide (Genbank Accession No: NW_ AAEX01024390.1). However this sequence does not encode an open reading frame, with two stop codons present in the C peptide sequence and one upstream of a homologous B chain sequence. Furthermore, there was no recognizable signal peptide or Methionine start codon upstream of this B chain sequence and it was missing the critical Cys-29 in the B chain. Orthologs of all relaxin-like peptides were identified in the completed chimpanzee genome (table 2 ). Two relaxin-like sequences were identified in the completed chicken genome (GgRLX3a, b in table 2 and figure 1 ), both with high similarity to human relaxin-3. Comparing the putative B and A domains of GgRLX3a and b with those of human relaxin-3 shows them to be 81% and 75% similar respectively (data not shown). Three sequences with a high similarity to human relaxin-3 (73%, 92% and 90% respectively; data not shown) were also discovered in the X. tropicalis genome (XtRLX3a-c, in table 2 and figure 1 ). Figure 1 Multiple sequence alignment of the relaxin-like peptide family. Amino acid sequences of the B and A domains only were aligned using ClustalW, then edited by hand using Seaview to remove gaps. This alignment was then used for all phylogenetic analyses. Newly identified sequences are highlighted in bold and italics. Invariant cysteine residues are indicated by asterisks ( * ) and the relaxin specific B-chain motif [RxxxRxxI/V] is shown. Sequences are clustered into subfamilies (A and B) based on primary sequence similarity and phylogenetic analysis. Hsa = Homo sapiens , Pt = Pan troglodytes , Mmul = Maca mulatta , Mm = Mus musculus , Rn = Rattus norvegicus , Cf = Canis familiaris , Ss= Sus scrofa , Re = Rana esculenta , Me = Macropus eugenii , Xl = Xenopus laevis , Xt = Xenopus tropicalis , Dr = Danio rerio , Tr = Takifugu rubripes , Gg = Gallus gallus , Om = Oncorhynchus mykiss . Sequences with high similarity to relaxin-3 have previously been reported in the fugu fish, TrRLX3a-e [ 27 ] and zebrafish, DrRLX3a [ 27 , 28 ] (see table 2 and figure 1 ). These searches identified a sixth relaxin-like sequence in the fugu fish, TrRLX3f, and another three in the zebrafish, DrRLX3b-d (table 2 and figure 1 ). Unlike the sequences previously identified in the fugu fish [ 27 ], TrRLX3f is more similar to human relaxin-2 (60%) than human relaxin-3 (54%) (Data not shown). Of the zebrafish sequences DrRLX3b and d both show 77% similarity to human relaxin-3 in their B and A domains (data not shown). DrRLX3c is only 60% similar to human relaxin-3 and 54% similar to human relaxin-2 (data not shown). Searches of partially completed genomes identified relaxin-like sequences in X. laevis (XlRLX3), Oncorhynchus mykiss (rainbow trout) (OmRLX3) and an INSL4 ortholog in the rhesus monkey (table 2 and figure 1 ). While OmRLX3 shows high similarity to human relaxin-3 (76%, data not shown), XlRLX3 is less similar (69%, data not shown). The presence of a relaxin gene in ruminants could not be determined due to the incomplete bovine genomic data currently available. No bovine ESTs with a similarity to relaxin were found and the presence of a bovine equivalent to the ovine likely non-functional genomic relaxin sequence could not be confirmed. However, a bovine EST was identified (BI682322) with high similarity to human relaxin-3 (79% identity to the translated EST product) starting from the end of the B chain (45F in the human pro-hormone sequence). Searches of invertebrate genomic and EST databases failed to identify a relaxin-like gene in any invertebrate or prokaryote. Although a C. intestinalis relaxin-like cDNA sequence has previously been reported [ 24 ], our searches failed to confirm this finding. The published relaxin sequence could not be found in the completed C. intestinalis genome. The B and A chains from all relaxin-like peptides identified in the following species: human, chimpanzee, rhesus monkey, pig, mouse, rat, wallaby, chicken, fugu fish, zebrafish, rainbow trout, R. esculenta , X. laevis and X. tropicalis were aligned using ClustalW and edited to remove all gaps, which are problematic in phylogenetic analysis (figure 1 ). Only the six cysteine residues responsible for conferring structure are conserved across all the relaxin-like peptides. Striking identity is seen amongst relaxin-3 sequences, especially in the B chain. Much lower similarity is seen amongst relaxin sequences, apart from the cysteine motif, only the relaxin-specific B chain motif is conserved. C peptide sequences show only negligible similarity between even closely related species making them impossible to align accurately. The C peptide is cleaved from the mature form of relaxin-1, 3 and INSL3, is believed to be cleaved from the mature form of all other relaxin-like peptides and was therefore excluded from all sequences. Phylogeny of the relaxin-like peptide family The alignment of B and A domains described above (figure 1 ) was used to construct phylogenetic trees with the maximum parsimony (MP), neighbour-joining (NJ) and maximum likelihood (ML) methods. MP and NJ methods produced conflicting trees, both with low bootstrap support for most of the major branches, and the ML tree failed to resolve many relationships (data not shown). Based on the relationships that could be determined with a degree of confidence, the sequences were divided into two clusters to be analysed separately (as shown in figure 1 ). Cluster A contained relaxin-3 and INSL5 sequences, while cluster B contained relaxin-1, relaxin-2, INSL3, INSL4 and INSL6 sequences. Several fish and frog sequences with lower sequence similarity to relaxin-3 (DrRLX3c, TrRLX3f, XlRLX3 and XtRLX3, figure 1 ) were grouped with cluster B based on preliminary phylogenetic analysis (data not shown). These sequences have, therefore, been listed as relaxin homologs rather than relaxin 3 (table 2 ). Also placed in cluster B are the sequences previously isolated from the tammar wallaby, MeRLX, [ 29 ] and edible frog, ReRLX [ 30 ]. Despite sequence similarity to relaxin-3, previous functional characterization and expression profiles of MeRLX and ReRLX indicates they are relaxin, rather than relaxin-3 homologs. None of the tree construction methods employed was able to completely resolve the phylogeny of either cluster. Bootstrap values in the MP and NJ generated trees were very low (below 50%, data not shown) and Tree-Puzzle failed to resolve the position of several sequences. This is primarily because of the short sequences used, as the C peptide can not be used to increase sequence length and thus improve the output from the tree generation methods, an inferred tree was produced instead. The Tree-Puzzle tree was resolved using topologies conserved between the MP and NJ trees and then reconciled against a species tree using GeneTree. The inferred gene trees were then edited to minimize the incongruence (the number of losses and inferred duplications) with the species tree. The inferred cluster A Additional file 1 and the cluster B trees Additional file 2 were combined to produce the phylogenetic tree of the complete relaxin-like peptide family (figure 2 ). Branch confidence levels are indicated on figure 2 ; branches without notation are inferred only. This gene tree was then reconciled with the species tree (figure 3 ). Figure 2 Evolutionary relationships among relaxin-like peptides. Topology shown is a consensus tree based on MP (maximum parsimony), ML (maximum likelihood) and NJ (neighbour-joining) analysis of the amino acid alignment shown in figure 1. Consensus tree was produced and edited using TreeView to correlate topology with known genomic information about the family. Human insulin used as an outgroup. Where possible, confidence values are shown at branches: * >50%, ** >75%, all other branches are inferred. Hsa = Homo sapiens , Pt = Pan troglodytes , Mmul = Maca mulatta , Mm = Mus musculus , Rn = Rattus norvegicus , Cf = Canis familiaris , Ss = Sus scrofa , Re = Rana esculenta , Me = Macropus eugenii , Xl = Xenopus laevis , Xt = Xenopus tropicalis , Dr = Danio rerio , Tr = Takifugu rubripes , Gg = Gallus gallus , Om = Oncorhynchus mykiss . Figure 3 Reconciled tree for the relaxin-like peptide family. The consensus tree of relaxin-like peptides (figure 2) from human, chimpanzee, mouse, dog, rat, pig, wallaby, chicken, fugu fish, zebrafish, rainbow trout, R. esculenta , X. laevis and X. tropicalis was reconciled using GeneTree with a species tree complied from a phylogeny of model organisms [65]. Squares indicate duplication events, red dotted lines indicate absent genes, either lost from those species (in grey), or not yet sequenced. While used to construct the gene tree as an outgroup, insulin has been removed from the reconciled tree. Hsa = Homo sapiens , Pt = Pan troglodytes , Mmul = Maca mulatta , Mm = Mus musculus , Rn = Rattus norvegicus , Cf = Canis familiaris , Ss = Sus scrofa , Re = Rana esculenta , Me = Macropus eugenii , Xl = Xenopus laevis , Xt = Xenopus tropicalis , Dr = Danio rerio , Tr = Takifugu rubripes , Gg = Gallus gallus , Om = Oncorhynchus mykiss . Analysis of the reconciled tree shows a major duplication event occurred early in the vertebrate lineage, giving rise to two subfamilies (clusters A and B respectively). Another duplication in subfamily A, prior to the emergence of fish, resulted in two lineages, which evolved into relaxin-3 and INSL5 in mammals. Interestingly, several non-mammalian relaxin-3-like sequences grouped with INSL5, implicating them as possible INSL5 homologs (GgRLX3b, OmRLX3, DrRLX3b, DrRLX3d, TrRLX3d and TrRLX3e). The reconciled tree also shows two additional fish-specific duplications in subfamily A. In the fugu fish genome a third duplication has occurred, resulting in three putative relaxin-3 (TrRLX3a, b, c) and two INSL5 homologs (TrRLX3d, e). In subfamily B there were four duplications, all were after the divergence of birds and reptiles and likely to have occurred during mammalian evolution. These events resulted in INSL3, INSL6, relaxin-1, relaxin-2 and INSL4. Synonymous ( d S ) and Nonsynonymous ( d N ) substitution rate estimates Results show the relaxin-like peptides are under varying selection pressures (table 3 ). Pairwise comparisons of human and chimpanzee orthologs provide the only way to compare all members of the family between two species. RLN1 , RLN2 and INSL6 have high d N /d S rate estimates, with results for RLN1 and INSL6 suggesting positive Darwinian selection. The extremely high estimate for INSL6 (99) is caused by having a d S of 0 (i.e. no synonymous substitutions), resulting in a division by 0 for the rate estimate, which is represented as 99 rather than infinity. All other human and chimpanzee sequences compared were identical and thus produced d N /d S estimates of 0. Table 3 Synonymous ( d S ) and nonsynonymous ( d N ) substitution rate estimates for all relaxin-like genes. Substitution rates were estimated using the Yang and Neilsen, 2000 method as implemented in yn00 in the PAML suite. Estimations were made using pairwise alignments of the nucleotide sequences of the B and A domains of human and chimpanzee or human and mouse genes. A d N / d S of 99 represents infinity and indicates that all substitutions detected are nonsynonymous, while na indicates that a d N / d S measurement is not possible as the sequences being compared are identical, or have only synonymous substitutions. human- chimpanzee human-mouse d N d S d N /d S d N d S d N /d S RLN1 0.055 0.039 1.4 _ a RLN2 0.017 0.023 0.7 0.56 1.06 0.5 RLN3 0 0 na 0.04 1.83 0.02 INSL3 0 0 na 0.17 1.31 0.1 INSL4 0 0.027 na _ a INSL5 0 0.023 na 0.16 1.24 0.1 INSL6 0.006 0 99 0.33 0.65 0.5 INS 0 0.13 na 0.033 1.44 0.02 a These genes are not present in the mouse. Human and mouse orthologs were used to estimate rates for the other members of the family. RLN2 and INSL6 show the highest estimates again, although are much lower than comparisons with chimpanzee sequences, the INSL6 estimate suggesting weak purifying selection instead of positive selection. The very low substitution rate observed for RLN3 (0.02) shows this peptide to be under strong purifying selection, at a similar rate to insulin ( INS ) (table 3 ). Rates vary among the other members of the family from ~0.1 for INSL3 and INSL5 to ~0.5 for RLN2 and INSL6 . As INSL4 is not present in mice and the human and chimpanzee sequences were identical, the INSL4 sequence from the rhesus monkey was used instead (data not shown). This comparison yielded a d N /d S estimate of 0.5, indicating weak purifying selection. Substitution rate estimates for the individual B and A domains were determined in a similar fashion, using human and chimpanzee comparisons for RLN1 , RLN2 and INSL6 ; human, rhesus monkey comparisons were used for INSL4 and human, mouse comparisons were used for RLN3 , INSL3 and INSL5 (figure 4 ). These comparisons show the B domains of relaxin-1 and INSL6 to be under positive selection (both estimates were 99). The B domains of INSL4 and relaxin-2 also have high substitution rates (1.0 and 0.7 respectively), but are not high enough to suggest positive rather than neutral or weak purifying selection. All A domains are under the effects of fairly strong purifying selection, except for that of relaxin-1, which is under only very weak selection pressures (0.8). Interestingly, while the B domains of relaxin-2, INSL6 and INSL4 all have very high d N /d S estimates, the A domains of these genes have very low estimates. This is in contrast with the other members of the family, relaxin-3, INSL3 and INSL5, which all have higher d N /d S estimates in the A domain than the B domain. Figure 4 Synonymous ( d S ) and nonsynonymous ( d N ) substitution rate estimates for individual B and A domains of each relaxin-like gene. Substitution rates ( d N / d S ) were estimated using the Yang and Neilsen, 2000 method as implemented in yn00 in the PAML suite. Human and chimpanzee comparisons were used for RLN1 and INSL6; human and rhesus monkey comparisons were used for INSL4 and human, mouse comparisons were used for RLN2, RLN3, INSL3 and INSL5. Positive selection tests To confirm the pairwise d N /d S (or ω) estimates, more sophisticated codon-based substitution models (reviewed in [ 31 ]) were used. As pairwise comparisons have already shown positive selection to be acting on RLN1 and INSL6 , INSL4 was analysed further. The phylogenetic tree of all sequences used is shown in figure 5 . Branch-specific likelihood analysis of the data, which assumes a constant ω ratio across all sites in a sequence, was used to test whether the INSL4 branch (branch A, figure 5 ) has a different ω ratio than all other branches. While the two-ratios model indicates a ω ratio of 1.1 for branch A (table 4 ), the LRT comparing this with the one-ratio model shows this to be statistically insignificant ( P = 0.6, d.f. = 1, table 5 ). Figure 5 Phylogeny of mammalian RLN2 and INSL4 genes. Tree generated using Tree-Puzzle using a gamma distribution, the Dayhoff model of substitution and 10 000 puzzling steps. Confidence values are shown as percentages on each branch. The INSL4 branch (labeled A on the tree) was tested for positive selection. Hsa = Homo sapiens , Mm = Mus musculus , Rn = Rattus norvegicus , Ss = Sus scrofa , Me = Macropus eugenii , Mmul = Maca mulatta , Cd = Camelus dromedaries , Gc = Galago crassicaudatus , Pt = Pan troglodytes , Ggo = Gorilla gorilla , Ph = Papio hamadryas , Fc = Felis catus , Cf = Canis familiaris , Oc = Oryctolagus cuniculus , Ma = Mesocricetus auratus , Ec = Equus caballus . Table 4 Parameter estimates for INSL4 under different branch models, site models and branch-site models. Models implemented in Codeml from the PAML suite. Parameters in boldface indicate positive selection. Sites potentially under positive selection are numbered using the human INSL4 sequence in figure 1 as the reference. Model ρ ℓ Parameter est. Positively selected sites 1 ratio (R0) 1 -1460.1 ω = 0.7 Branch specific 2 ratios (R2) 2 -1459.9 ω 0 = 0.7(background), ω 1 = 1.1 (branch A) Site specific Neutral (M1) 1 -1424.4 ρ 0 = 0.1, ρ 1 = 0.8 not allowed Selection (M2) 3 -1416.6 ρ 0 = 0.1, ρ 1 = 0.6, ρ 2 = 0.2, ω 2 = 2.9 15L ( P >0.99) 14H, 27H, 28R, 36V ( P >0.95) Discrete (M3) (k = 2) 3 -1421.4 ρ 0 = 0.2, ω 0 = 0.03, ρ 1 = 0.8, ω 1 = 1.0 37 sites a ( P >0.99) Discrete (M3) (k = 3) 5 -1413.8 ρ 0 = 0.1, ω 0 = 0.01, ρ 1 = 0.5, ω 1 = 0.6, ρ 2 = 0.3, ω 2 = 2.0 15L, 27H, 28R ( P >0.99) 14H, 25G, 26R, 30D, 36V ( P > 0.90) Beta (M7) 2 -1419.6 ρ 0 = 0.2,q = 0.2 not allowed Beta&ω (M8) 4 -1414.6 ρ 0 = 0.8,p = 0.3, q = 0.06, ρ 1 = 0.2, ω 1 = 2.6 15L, 28R ( P > 0.95) 14H, 27H, 36V ( P > 0.90) Branch-Site Model A 3 -1418.4 ρ 0 = 0.1, ρ 1 = 0.6, ρ 2 = 0.2, ω 2 = 3.0 In the foreground lineage: 13K, 37I ( P > 0.95) Model B 5 -1416.4 ρ 0 = 0.1, ω 0 = 0.01, ρ 1 = 0.6, ω 1 = 0.6, ρ 2 = 0.3, ω 2 = 3.2 In the foreground lineage: 13K, 37I ( P > 0.95) In the background lineages: no significant sites a 1A, 2A, 3E, 5R, 6G, 10R, 11F, 12G, 14H, 15L, 16L, 17S, 18Y, 20P, 25G, 26R, 27H, 28R, 29F, 30D, 31P, 32F, 35E, 36V, 37I, 39D, 40D, 41G, 42T, 43S, 44V, 45K, 47L. Note that these sites should be treated with caution as ω under this model is not significantly higher than 1. Table 5 Likelihood ratio test statistics (2δ) for the INSL4 data set. 2δ d.f. P -value LRT of ω at branch A 1 ratio (R0) vs. 2 ratios (R2) 0.3 1 0.6 LRTs of variable ω's among sites M0 vs. M3 (k = 3) 92.5 2 <0.0001 M1 vs. M2 15.7 2 0.0004 M7 vs. M8 10.1 2 0.006 LRT's of variable ω's along branch A M1 vs. Model A 12.0 2 0.002 M3 (k = 2) vs. Model B 10.0 2 0.007 Site-specific models, which allow the ω ratios to vary between sites in a sequence, were also applied to the data. The ω ratio was found to vary considerably among amino acid sites. The discrete model with K = 3 site classes was the best fit to the data with a log likelihood value of -1413.8 (table 4 ). This model suggests that 31% of sites are under positive selection (ω 2 = 2.00), while half (54%) are under weak purifying selection (ω 1 = 0.6) and the other 15% constrained under extreme purifying selection (ω 0 = 0.01) (listed in table 4 ). Eight amino acids are identified as under positive selection at the 90% cut off (14H, 15L, 25G, 26R, 27H, 28R, 30D, 36V, see table 4 ). All but two of these are within the A chain of INSL4 . The LRT of M3 with its null model (M0) shows these results to be significant ( P < 0.0001, d.f. = 2, table 5 ). Similar results are seen with Model M8. Lastly, branch-site models A and B were applied to the data. These models extend the site and branch specific models by allowing the ω ratios to vary among lineages and sites and were used to test for specific sites under positive selection on the INSL4 branch. Model A, which fits the data significantly better than its null model M0 ( P = 0.02, d.f. = 2, table 5 ) identifies 2 sites (13K and 37I) under positive selection in the INSL4 branch at the 95% cut off (table 4 ). Model B, which allows the ω ratio to vary both in the foreground lineage (the INSL4 branch) and in the background branches also fits the data significantly better than its null model, M3 with k = 2 ( P = 0.007, d.f. = 2, table 5 ), and identifies the same two positively selected sites as Model A (13K and 37I) in the INSL4 branch (table 4 ). Model B also confirms the results produced by the discrete model (M3 with k = 3 site classes), showing 27% of sites of under positive selection, 58% under weak purifying selection (ω 2 = 0.58) and 14% under very strong purifying selection (table 4 ). Discussion While relaxin evolution has been the centre of much controversy (relaxin is often cited as a gene that conflicts with the Darwinian theory of evolution [ 24 , 32 - 34 ]), this report is the first attempt to describe the evolutionary history of the whole relaxin-like peptide family from a phylogenetic perspective. Previous studies have concentrated on the primate relaxins and relaxin-like factors [ 26 ], or not included detailed phylogenetic analyses [ 27 ]. We have sought to expand upon these by incorporating sequences identified in all the available completed genomes with a subset of cloned relaxin-like sequences, particularly those from non-mammalian species. None of the phylogenetic tree construction programs used was able to completely resolve the evolution of the relaxin-like peptide family. This is likely due to variable divergence across the family and the short sequence length [ 35 ]. Incorporating results from the MP and NJ methods suggested positions for several branches that were unresolved after ML analysis. Minimizing incongruence between the gene and species trees by reducing the number of assumed duplications in the reconciled tree also provided a method to infer the evolutionary history of this family. Similarly to previously published results, searches of available genomic and EST data failed to identify any novel members of the relaxin-like peptide family [ 28 ]. Given the stringent and well-described insulin family signature that revolves around the invariant cysteine residues that confer the insulin-like structure seen across the superfamily, we find it improbable that any novel relaxin or insulin-like sequences will be identified. The presence of an invertebrate relaxin has been of speculation since 1983 when relaxin-like activity was first detected in the protozoa, T. pyriformis [ 22 ]. Similar activity was reported in H. momus [ 23 ] and in C. intestinalis , where a cDNA sequence almost identical to pig relaxin was found [ 24 ]. However, our searches of all completed invertebrate genomes (including C. intestinalis ) failed to identify any relaxin-like sequences, including the published sequence. Multiple insulin-like peptides have been found in several invertebrates, including: Bombyxi mori (silkworm) [ 36 ], D. melanogaster [ 37 ] and C. elegans [ 38 ]. As these sequences lack the relaxin-specific motif, and show no homology to other relaxin family peptides, they are not considered part of the relaxin subfamily and therefore have not been included in these analyses. Much of the controversy surrounding relaxin evolution concerns the identification of an invertebrate relaxin sequence (a cDNA sequence from Ciona intestinalis ) almost identical to pig relaxin (Georges and Schwabe, 1999). Completion of the C. intestinalis and other invertebrate genomes has allowed us to conclude that there is not a relaxin-like sequence in any invertebrate sequenced to date. If an invertebrate relaxin does exist, it does not contain the relaxin-specific motif characterized in vertebrates. A hallmark of relaxin sequences is their high variability, even amongst closely related species. Relaxin-like peptide sequences isolated from two whales are almost identical to porcine relaxin [ 21 ], however as these sequences were derived from amino acid sequencing, without nucleotide or and genomic sequence available, they have not been able to be included in these phylogenetic analyses. The presence of a functional relaxin in the ruminant lineage has yet to be confirmed [ 25 ]. More genomic data is required to confirm the presence of a non-functional relaxin gene sequence in the bovine, similar to that observed in the ovine [ 25 ]. Searches of the preliminary bovine genome assembly have failed to find a relaxin gene. Interestingly, a relaxin sequence has been identified in the camel [ 39 ] and relaxin expression found in the closely related llama and alpaca [ 40 ]. While classified as a ruminant, Camelidae have a unique reproductive anatomy and physiology [ 41 ]. A bovine EST (BI682322) with high similarity to exon 2 of human relaxin-3 was identified. Confirmation of the presence of relaxin and relaxin-3 orthologs in ruminants awaits further sequencing of the bovine and ovine genomes. The presence of an avian relaxin has also been of speculation. While relaxin-like activity has been reported in the chicken [ 42 ], an avian relaxin-like peptide or gene has not been identified. While two relaxin-3-like genes were identified in the nearly completed chicken genome, no avian relaxin gene was found. As no other relaxin-like genes were found, the reported relaxin activity may be due to one of the relaxin-3-like genes. The phylogeny of the relaxin-like peptide family indicates relaxin-3 is the ancestral relaxin, appearing prior to the divergence of teleosts. The finding of multiple relaxin-3-like genes in the fugu fish and zebrafish suggests multiple lineage-specific duplications of a single relaxin-3-like ancestor have occurred in fish [ 27 ]. However, the possibility the other mammal specific relaxin-like peptides emerged earlier before being lost in the teleost can not be excluded [ 27 ]. We find it more likely that these duplications, and the resulting multiple relaxin-3-like genes, are fish specific and due to genome wide duplications hypothesized to have occurred during fish evolution [ 43 ]. Phylogenetic analyses show multiple fish homologs of both the mammalian relaxin-3 and INSL5 genes, meaning that two relaxin-3-like genes existed prior to the genome duplication event proposed to have occurred in the teleost ancestor. The putative fish relaxin homolog was either, present in the teleost ancestor, duplicated and the second copy lost or emerged shortly after or, as a result of, the genome-wide duplication event. While termed relaxin-3-like based on sequence similarity, phylogenetic analysis indicates that several non-mammalian sequences (OmRLX3, DrRLX3b, DrRLX3d, TrRLX3d, TrRLX3e and GgRLX3b) could be INSL5 homologs. None of the sequences found in the complete X. tropicalis genome were placed in this group, while there are members present in the more ancient fish lineage and the younger avian lineage. It is possible that this gene has either been lost, or remains unidentified, in the X. tropicalis genome. A sequence with similarity only to the B chain of relaxin-3 was also found, but a corresponding A chain match was not, however, there is a gap in the genome assembly upstream which might contain the missing domain. Future assemblies of the Xenopus genome should resolve this issue. These results suggest that INSL5 could have emerged during teleost evolution, far earlier than previously believed. Unlike the mammal-specific relaxin-like genes, which are clustered together (on chromosome 9 in the human and chromosome 19 in the mouse), INSL5 is localized independently (chromosome 1 in the human and chromosome 4 in the mouse). These findings are of particular interest in the analysis of INSL5, which is still functionally uncharacterised. All the potential non-mammalian INSL5 homologs retain the relaxin-specific B chain [RxxxRxxI/V] motif, hence would be capable of interacting with the relaxin receptor, LGR7, and thus functionally classified as a relaxin. Recent studies have shown INSL5 is a high affinity ligand for GPCR142 but not GPCR135, LGR7 or LGR8 [ 19 ]. As the residues required for interaction with GPCR135 and GPCR142 are not known, it is unknown whether the non-mammalian INSL5 homologs would interact with GPCR142, GPCR135 and/or LGR7. Phylogenetic results from this study suggest the presence of a relaxin homolog in fish and frogs, although not in the chicken. Relaxin sequences have previously been isolated and peptide sequenced from either the ovaries or testes of the edible frog [ 30 ], little skate ( Raja erinacea ) [ 44 ], spiny dogfish ( Squalus acanthias ) [ 45 ], Atlantic stingray ( Dasyatis sabina ) [ 46 ] and the sand tiger shark ( Odontaspis taurus ) [ 47 ]. While having high similarity with relaxin-3, these sequences are not relaxin-3 homologs (as the B chain of the stingray sequence is lacking the relaxin-specific motif, it is not a functional relaxin [ 46 ] and has not been considered further). Based on the expression of all these genes in reproductive organs such as the testes and ovaries, and the failure to find the R. esculenta gene expressed in the brain using northern blot analysis [ 30 ], we believe these to be among the first relaxin peptides with a reproductive function. Based on the similarity with relaxin-3 observed in these sequences, the ancestral relaxin homolog, and its new reproductive function, is likely to have emerged prior to the divergence of teleosts. A complete picture of relaxin-like peptides present in non-mammalian genomes will be invaluable in understanding the evolution of relaxin from neuropeptide to reproductive hormone. The ancestral RLN3 gene is under very strong purifying selection, highlighting the importance of its highly conserved function, likely to be in the brain [ 2 ]. As high divergence is a hallmark of relaxin sequences, it is somewhat unsurprising that RLN2 is under only weak purifying selection. We suggest that this lack of selective pressure has contributed to the high sequence divergence seen between many relaxins (e.g. human and mouse) and the differences in relaxin's functions observed across mammals. Information about the selective constraints placed upon these peptides, can provide valuable insight into the nature of interactions with their receptors. Based on selection pressures we can conclude that the interactions between relaxin-3 and GPCR135, INSL5 and GPCR142 are very specific, while the binding of relaxin to LGR7 is much looser. In this context the cross-reactivity seen between LGR7 and INSL3 or H1 relaxin, which are both similar to relaxin in sequence but especially in structure, is understandable, as is the lack of binding between GPCR135 and GPCR142 with any other relaxin-like peptide. Unexpectedly, synonymous and nonsynonymous substitution rate estimates for RLN1 and INSL6 show these to be under positive selection. Positive selection is often difficult to observe using pairwise comparisons that average over the whole length of a sequence, making these results even more striking. While pairwise comparisons failed to confirm positive selection was acting on INSL4 , further statistical tests suggested that positive Darwinian selection acted on several sites in the INSL4 sequence after its emergence. Further analysis will be required to confirm these sites as important in the acquisition of a new receptor and a new function by INSL4, particularly in light of recent studies that question the reliability of ML methods to accurately detect positive selection acting on single sites [ 48 - 50 ]. We are encouraged that both branch-specific and site-specific ML models find positive selection to be acting on INSL4. When the B and A domains of each gene were analyzed separately, further differences in selection pressures became apparent. The interaction between relaxin and its receptor has been thought to be primarily mediated through the B chain of the peptide [ 4 ], so the finding that selection pressures are stronger on the A chain of relaxin-1, INSL4 and INSL6 was unexpected. We also find it noteworthy that INSL4, INSL6 and relaxin-1 are the most recent members of the family to emerge and all appear to be under the effects of positive Darwinian selection. INSL6 emerged during mammalian development, INSL4 and RLN1 during primate evolution, they remain functionally uncharacterized and INSL4 and INSL6 are without known receptors. The low selection pressure on the B domain and the strong constraints placed on the A domain of INSL4 and INSL6 suggests that, unlike the B chain mediated interaction of relaxin and INSL3 with their receptors, the interaction of these peptides with their receptors could be dependant on the A chain instead. The low d N / d S rate observed for INSL5 indicates this peptide to be evolutionary stable and of functional importance. In particular the constraints placed on both A and B chains of INSL5 suggest a well-defined receptor interaction system, while the total absence of these constraints on either chain within relaxin-1 suggests the opposite, that perhaps this peptide is still evolving its function. Conclusions We present here a phylogeny for the relaxin-like peptide family. Relaxin has long been used as an example of a gene that conflicts with the Darwinian theory of evolution [ 24 , 32 - 34 ]. However, we have shown that these can issues can be resolved when studied in the context of the rest of the relaxin-like peptide family, in particular the new, but likely ancestral relaxin, relaxin-3. We have demonstrated that positive selection has been a driving force in the recent expansion of the relaxin-like peptide family during mammalian evolution. While strong purifying selection has maintained the structural core of these peptides by constraining the insulin superfamily cysteine motif, outside these residues, positive selection has acted after at least three gene duplication events (which generated INSL6, INSL4 and relaxin-1) to allow these new genes to acquire a new receptor and novel functions. Given the known roles of relaxin and INSL3 in reproduction (and the likely similar roles of INSL4 and INSL6 given the specificity of their expression in reproductive tissues) these findings correlate with a general trend towards rapid evolution in several reproduction associated genes [ 51 - 54 ]. We anticipate that further analysis of the coevolution of the relaxin-like peptides with their receptors will contribute much towards our understanding of the pleiotropic actions of this family as well as mechanisms involved in the evolution of peptide hormone systems. Methods Sequences and sequence similarity searches Amino acid and nucleotide sequences of cloned relaxin-like peptide family members from the following species were obtained from GenBank [ 55 ]: human ( Homo sapiens ) H1 relaxin, H2 relaxin, H3 relaxin, INSL3, INSL4, INSL5, INSL6; mouse ( Mus musculus ) relaxin, relaxin-3, INSL3, INSL5, INSL6; rat ( Rattus norvegicus ) relaxin, relaxin-3, INSL3, INSL6; dog (Canis familiaris) relaxin; pig ( Sus scrofa ) relaxin, relaxin-3, INSL3; edible frog ( Rana esculenta ) relaxin and tammar wallaby ( Macropus eugenii ) relaxin (see table 2 for accession numbers). Five published relaxin-like sequences previously identified in the fugu fish (TrRLX3a-e) [ 27 ] and the zebrafish (DrRLX3a) [ 28 ] were also used. There are several partial relaxin-like peptide sequences available, however only sequences with corresponding nucleotide sequence data were utilized in this study. Sequence similarity searches using TBLASTN [ 56 ] were conducted using the B and A chain sequences of each family member to identify other mammalian, vertebrate and invertebrate relaxin-like peptides. The following databases were searched: human, mouse, rat, dog, chimpanzee ( Pan troglodytes ), fugu fish, zebrafish, fruit fly ( Drosophila melanogaster ), mosquito ( Anopheles gambiae ), Caenorhabditis elegans , all yeast, all plant and all bacterial genomes at NCBI [ 57 ], X. tropicalis [ 58 ] and C. intestinalis [ 59 ], Expressed Sequence Tags (EST), Genome Survey Sequences (GSS), and High-Throughput Genomic Sequences (HTGS) databases [ 55 ]. While the classical cysteine motif of the insulin superfamily was used to distinguish sequences as members of this family, relaxin homologs were distinguished by the additional presence of the specific relaxin motif [RXXXRXXI/V] in the B chain of the derived peptide sequence. Multiple sequence alignment and phylogenetic analysis Amino acid sequences were aligned using ClustalW [ 60 ] with default parameters. The alignments were edited to delete the C and signal peptide sequences, leaving only the B and A domains, which were further edited to minimize gaps and then concatenated. Human insulin was included as an outgroup. Phylogenetic trees were constructed using maximum parsimony (MP): implemented in PHYLIP [ 61 ] using ProtPars, Neighbour-joining (NJ): implemented in PHYLIP using ProtDist and Neighbour and maximum likelihood (ML): implemented in Tree-Puzzle [ 62 ]. Data analyzed in PHYLIP was bootstrapped 1000 times using SeqBoot and consensus trees derived using Consense. Tree-Puzzle was run with a two-rate model of heterogeneity, the Dayhoff model of substitution and 50 000 puzzling steps. Trees were edited using TreeView [ 63 ]. Reconciliation of gene and species trees Gene trees of relaxin-like peptides were reconciled with a species tree using GeneTree [ 64 ]. Reconciled trees are an attempt to resolve incongruence between gene and species trees by predicting gene duplications and losses [ 64 ]. The species tree was based on a phylogeny of model organisms [ 65 ]. The reconciled tree was edited to minimize incongruence, primarily by reducing inferred duplications. Estimation of synonymous and nonsynonymous substitution rates Pairwise nucleotide sequence alignments of human and chimpanzee, human and rhesus monkey ( Maca mulatta ) and human and mouse orthologs were constructed using ClustalW [ 60 ] and edited to limit alignments to the B and A domains only, which were then concatenated. Synonymous ( d S ) and nonsynonymous ( d N ) substitution rates were estimated using the methods of Yang and Nielsen [ 66 ] as implemented in yn00 in the PAML suite [ 67 ]. Testing for positive selection The following relaxin-1 and INSL4 nucleotide sequences were aligned using ClustalW: human H2 relaxin (X00948), INSL4 (L34838), chimpanzee( Pan troglodytes ) relaxin-2 (Z27245), INSL4 (BK005152); gorilla ( Gorilla gorilla ) relaxin-2 (Z27228, Z27237); rhesus monkey relaxin (A34936), INSL4 (BK005251); bush baby ( Galago crassicaudatus ) relaxin (AF317625); baboon ( Papio hamadryas ) relaxin (Z27246, Z27224); camel ( Camelus dromedarius ) relaxin (AF254739); cat ( Felis catus ) relaxin (AF233688); dog ( Canis familiaris ) relaxin (AF233687); guinea pig ( Cavia porcellus ) relaxin (S85964); rabbit ( Oryctolagus cuniculus ) relaxin (S45940); hamster ( Mesocricetus auratus ) relaxin (S79879) and horse ( Equus caballus ) relaxin (AB000201). The alignment was edited as described previously. A ML tree was constructed from this alignment using TreePuzzle [ 62 ] and the method of Yang and co-workers [ 31 ] was used to test for positive selection in the INSL4 branch. Using Codeml from the PAML suite, several models were fitted to the data. The branch specific models, One-ratio (R1) and Two-ratios (R2) were used to detect lineage-specific changes in selective pressure. The site specific models, Neutral (M1), Selection (M2), Discrete (M3) with 2 and 3 site classes, Beta (M7) and Beta&ω (M8), were also used to test for individual residues under positive selection. The branch-site models A and B were used to detect positive selection in a subset of sites in a specified branch. Likelihood ratio tests (LRT) were used to assess their goodness of fit, by comparing a model that does allow for dN/dS > 1 against a model that does not (i.e. a null model). Therefore, the branch specific LRT was R2 vs. R1. The site specific LRTs were M3, M2 and M8 against their respective null models, M0, M1 and M7. The branch-site models A and B were tested against M1 and M3 with k = 2 site classes respectively. Positively selected sites with a posterior probability of P (ω>1) >0.90 were listed. Authors' contributions TW performed all sequence and phylogenetic analysis and drafted the manuscript, TS participated in phylogenetic analysis, design and coordination of study, GWT participated in the design of the study, and RADB participated in phylogenetic analysis, conceived of the study and participated in its design and coordination. Supplementary Material Additional File 1 Phylogeny of cluster A- relaxin-3 and INSL5. Phylogeny of Cluster A constructed from a ClustalW alignment of the B and A domain amino acid sequences from relaxin 3 and INSL5 peptides. Consensus tree generated from MP (Protpars in PHYLIP), ML (TreePuzzle) and NJ (Neighbour in PHYLIP) methods and edited in Treeview to minimize species tree incongruence. Human insulin was used as an outgroup. Where possible, confidence values are shown at branches: * >50%, ** >75%, all other branches are inferred. Hsa = Homo sapiens , Pt = Pan troglodytes , Mm = Mus musculus , Rn = Rattus norvegicus , Cf = Canis familiaris , Ss = Sus scrofa , Xt = Xenopus tropicalis , Dr = Danio rerio , Tr = Takifugu rubripes , Gg = Gallus gallus , Om = Oncorhynchus mykiss . Click here for file Additional File 2 Phylogeny of cluster B- relaxin-1, 2, INSL3, INSL4 and INSL6. Consensus phylogeny of Cluster B constructed from a ClustalW alignment of the B and A domain amino acid sequences from relaxin 1, 2, INSL3, INSL4, INSL6 peptides. Consensus tree generated from MP (Protpars in PHYLIP), ML (TreePuzzle) and NJ (Neighbour in PHYLIP) methods and edited in Treeview to minimize species tree incongruence. Human insulin was used as an outgroup. Where possible, confidence values are shown at branches: * >50%, ** >75%, all other branches are inferred. Hsa = Homo sapiens , Pt = Pan troglodytes , Mmul = Maca mulatta , Mm = Mus musculus , Rn = Rattus norvegicus , Cf = Canis familiaris , Ss = Sus scrofa , Re = Rana esculenta , Me = Macropus eugenii , Xl = Xenopus laevis , Xt = Xenopus tropicalis , Dr = Danio rerio , Tr = Takifugu rubripes . Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551602.xml
544554	A Strategy for Developing an HIV Vaccine	The newly published strategic plan for developing an HIV vaccine is crucially important, say the PLoS Medicine editors, but it must be followed by clear milestones and a process for monitoring progress	In 1997, United States President Bill Clinton announced the challenge to develop an AIDS vaccine by 2007. Since 1997, the AIDS Vaccine Advocacy Coalition (AVAC) has published annual reports on the global status of the effort to meet Clinton's deadline. Last year's report, entitled “AIDS Vaccine Trials—Getting the Global House in Order,” officially ends the countdown. Saying that “we are on a long term mission,” AVAC concludes that there will not be a safe and efficient vaccine in 2007, and that we need to “focus on the long haul and set an agenda for sustained and sustainable action that stretches well beyond 2007.” It is not that there are no vaccine candidates in clinical trials, but there is little hope that any of the current candidates will turn out to be a cheap and safe vaccine that affords long-term protection. Among notable developments over the past 12 months, the AVAC report highlights the Global HIV/AIDS Vaccine Enterprise as an effort to improve coordination within the AIDS vaccine field. The Enterprise was announced in June 2003 and now shares its scientific strategic plan with everyone affected by the AIDS pandemic—that is, all of us—by publishing it in PLoS Medicine (DOI: 10.1371/journal.pmed.0020025 ). In its plan the Enterprise presents itself as a global endeavor and emphasizes the need for integration and capacity building around the world. It is not “a discrete organization with a pool of money” but a “coordinating group of individual funding agencies that will support specific areas of research using their own mechanisms, according to their own practices and policies, and following the Enterprise's principles.” These principles include collaboration, standardization, and coordination among international researchers and agencies. The plan focuses on specific scientific roadblocks that need to be overcome, but also looks ahead and mentions the need to build capacity for product manufacturing and clinical trials, and to address regulatory issues. These are noble goals, and the fact that they are stipulated jointly by many of the leaders in the field will generate excitement and expectations, even though much of what is said has been said before. The plan stresses collaboration and coordination; there are clear benefits from a concerted effort. But might a level of competition, rather than collaboration, be healthy, and, if so, what level of competition would work best? The Enterprise members seem to have wrestled with that question. The plan mentions an “appropriate balance between productive competition and effective collaboration,” and suggests that certain incentives could be provided by “the funders with greatest flexibility.” As long as it remains unclear where scientific breakthroughs will come from, diversity and flexibility should be encouraged and not stifled. David Ho, in his Perspective on the plan (DOI: 10.1371/journal.pmed.0020036 ), mentions the danger of “group think,” and the Enterprise must not fall into that trap. Notably absent from this initial plan is any mention of a timeline or milestones. The remit of the plan's authors was not to prescribe specific research but “to stimulate both researchers and funders to explore new, more collaborative, cooperative, and transparent approaches…in addition to continuing the productive, high-quality approaches already underway.” However, without a timeline, the plan fails to convey a sense of urgency. This is problematic, as any delays in developing a vaccine will increase the burden from HIV/AIDS in the parts of the world that can least afford it. To accelerate vaccine development, the plan urgently needs to be supplemented with a list of specific tasks, responsible individuals, necessary resources, and allocated time. The next document from the Enterprise must provide specifics on project management, although one problem with putting a time frame on HIV vaccine development is a fundamental one: we do not know whether it is actually possible to develop a safe and effective vaccine. (One assumes the Enterprise members agree, though there is no explicit acknowledgement of this uncertainty in the plan.) Moreover, provided it can be done, it is impossible to predict when the necessary scientific advances will happen. That said, without a list of specific projects, project leaders, and a time frame for achieving or at least evaluating specific goals, it will be impossible to define success and failure, review progress, and assure internal and external accountability. There is another reason why a best-guess timeline is essential: realistic expectations about an AIDS vaccine would stress the urgency of combating the AIDS pandemic over the next decade—and maybe longer—in the absence of an effective vaccine. The potential benefits of a vaccine cannot be overestimated, and its development has to be one top priority for the global scientific community. But its success cannot be taken for granted and will come too late for millions. Therefore, parallel efforts to prevent or reduce transmission and to treat infected individuals need to be accelerated now. The Enterprise's plan should be hailed as a crucially important outline for vaccine development, but the goodwill surrounding it won't last unless it is quickly followed up with a set of milestones, and a transparent process by which progress will be measured and course corrections implemented.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544554.xml
552323	Rapid self-assembly of DNA on a microfluidic chip	Background DNA self-assembly methods have played a major role in enabling methods for acquiring genetic information without having to resort to sequencing, a relatively slow and costly procedure. However, even self-assembly processes tend to be very slow when they rely upon diffusion on a large scale. Miniaturisation and integration therefore hold the promise of greatly increasing this speed of operation. Results We have developed a rapid method for implementing the self-assembly of DNA within a microfluidic system by electrically extracting the DNA from an environment containing an uncharged denaturant. By controlling the parameters of the electrophoretic extraction and subsequent analysis of the DNA we are able to control when the hybridisation occurs as well as the degree of hybridisation. By avoiding off-chip processing or long thermal treatments we are able to perform this hybridisation rapidly and can perform hybridisation, sizing, heteroduplex analysis and single-stranded conformation analysis within a matter of minutes. The rapidity of this analysis allows the sampling of transient effects that may improve the sensitivity of mutation detection. Conclusions We believe that this method will aid the integration of self-assembly methods upon microfluidic chips. The speed of this analysis also appears to provide information upon the dynamics of the self-assembly process.	Background There has been a rapid growth in the number of applications that are based upon DNA self-assembly, ranging from DNA microarrays (e.g. Affymetrix [ 1 ]) in the life sciences, through conformation-based mutation detection methods [ 2 , 3 ], to the ongoing development of DNA scaffolding methods of nanoassembly [ 4 ]. The control of the degree of DNA hybridisation requires elaborate and time consuming sample preparation (eg [ 5 ]) with methods that may require hours to achieve hybridisation [ 6 ], and on the order of an hour even within miniaturised systems [ 1 , 7 ]. However, a rapid method of controlling denaturation and renaturation within a microfluidic device would enable an inexpensive mutation detection method that could be performed within minutes. Microfluidic devices or 'microchips' are photolithographically-defined networks of microchannels in glass where the microchannels are similar in size to conventional capillaries. These microchips provide compelling advantages in terms of speed, reagent usage and integration over conventional capillary or gel-based methods. The potential of the microfluidic chip has led to the use of terms such as "micro-total analysis systems" and "lab-on-a-chip". These microchips have been demonstrated in conjunction with a range of applications that integrate the polymerase chain reaction (PCR) and capillary electrophoresis (CE) methods with some reaching nanolitre or smaller scale volumes. A powerful advantage of the microchip approach is that it can implement much the same molecular biology protocols and reagents as used with conventional equipment, thereby allowing a wealth of established expertise to be transferred to the microscale. Although the most effective method of mutation detection is sequencing, it is also by far the most expensive [ 8 ]. The microarray [ 8 ] technique, although powerful, is still handicapped by significant false positive rates and high cost [ 9 ]. Alternative methods based on DNA self-assembly are much faster than sequencing and these include single-strand conformation polymorphism (SSCP), denaturing high performance liquid chromatography (DHPLC) and heteroduplex analysis (HA). Although their cost has been shown to be far lower than sequencing, the achievable sensitivities (the percentage of mutations that are successfully detected) are only about 90 % [ 10 , 9 ]. Microfluidic chips may enable extremely high throughputs and high levels of integration. The achievement of this goal has been hindered by the lack of successful integrations of methods of mutation analysis based on single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA) – likely due to the difficulties in controlling the degree of hybridisation on chip without time consuming thermal processing. A great advantage would be provided by a method of enabling microchip-based control of a rapid DNA self-assembly process. The term wildtype is used to describe any given genetic sequence that does not contain mutations. Since individuals usually carry two copies of each gene, the genetic sequence of the two copies may be identical (homozygous) or may differ (heterozygous). DNA is normally double stranded, but under some conditions (e.g. high temperature), melts into single strands. Under other conditions, such as a lower temperature, these single strands will self-assemble into the double-stranded form again. The resulting double-stranded DNA is referred to as a homoduplex if the sequences are perfectly complementary, or a heteroduplex if the sequences are nearly complementary (i.e. a mutant sequence paired with a wildtype sequence). The misfit in a heteroduplex creates a "bulge" or "bubble" where the bases do not match and this affects the shape of the assembled molecule, typically lowering its velocity during electrophoretic movement, i.e. the heteroduplexes typically migrate more slowly than the homoduplexes. Any heterozygous sample will generate four different duplexes, two homoduplexes and two heteroduplexes, although the molecules often co-migrate so that fewer than 4 separate electropherogram peaks are resolved. In the HA method, electrophoretic conditions are chosen in order to enhance the velocity differences between the duplexes so that the process of duplex self-assembly can be used to determine the presence of a heterozygous state (hence indicating the presence of a mutation). In SSCP, isolated strands of ssDNA find near-complementary sequences on the same strand, with the result that the strand folds upon itself in a sequence dependent manner forming new conformations. This is a simplistic description since ssDNA without self-similar sequences, and homoduplex dsDNA, may also take complex forms. Techniques such as HA that aim to separate homoduplex fragments from heteroduplex fragments often use some combination of thermally and chemically denaturing conditions to cause the partial melting of the duplex, resulting in a shift in mobility or chromatography column retention time that increases with the degree of mismatch. Many medical diagnostics could be implemented on microchips if an effective implementation of a highly sensitive mutation analysis method could be integrated with PCR/CE. Considerable work has been done in developing SSCP [ 11 ] and HA [ 2 , 3 , 12 ]. An excellent review of such methods has been produced by Jin et al. [ 13 ]. The main drawback is the lower sensitivity of these methods. In macroscopic work Kozlowski and Krzyzosiak [ 5 ] and Kourkine et al . [ 14 ] greatly improved their sensitivities by combining SSCP and HA methods to develop capillary-based electrophoretic techniques with sensitivities of 90–94 % for SSCP and 75–81 % for HA. In a landmark analysis, Kourkine et al . achieved 100 % sensitivity by analysing denatured and non-denatured fragments in tandem. Despite being highly effective, the additional sample preparation required by these methods (i.e. separately preparing both single and double stranded DNA and maintaining this strandedness) complicates their implementation on microchips. In this work, we present an electrophoretic method in which DNA is denatured in a microchip (with formamide) and, depending upon the sequence of applied voltages, can be prepared with a widely varying degree of hybridisation (i.e. from almost entirely ssDNA to almost entirely dsDNA). Given the small volumes involved within the microchip, diffusion time plays a small role and the reassembly process can be fast, with dsDNA obtained within minutes. The rapidity of the manipulation possible on this system allows some investigation of the dynamics of the reassembly, indicating that there are well-defined intermediate states where both ssDNA and dsDNA exist in the reassembly process. We have applied our methods to the H63D and S65C mutations from the HFE gene associated with hereditary hemochromatosis (HH). The denaturation technique used enables a combined microchip-based method of HA and SSCP analysis. Results Heteroduplex Analysis In our electrophoretic analyses with a double-T chip (described below), the dsDNA arrives at the detection point before the ssDNA (after about 105 s of separation, versus 190 s of separation for the ssDNA). As shown in Fig. 1 , to demonstrate that our analysis conditions allow for the detection of mutations by means of HA, we analysed (undenatured) PCR products of a homozygous wildtype, a heterozygous and a homozygous H63D mutant. We found, as expected, that the heterozygous sample had two distinct peaks due to the transport of heteroduplexes as well as homoduplexes. However, the wildtype and homozygous mutant samples looked very similar, with the exception of a small peak following the main peak of the mutant sample. This small peak was only apparent with this sample and seems to indicate a PCR artefact. The size and shape of the bump remained consistent throughout the experiments and did not affect the peak intensities of either dsDNA or ssDNA. The bump is too small to add ambiguity when resolving the H63D mutation by HA and it should be noted that the emphasis of this work is on inducing the formation of dsDNA and ssDNA on-chip rather than upon improving mutation detection. Figure 1 Double-stranded DNA peak profiles prior to the addition of formamide (fluorescence in relative fluorescence units (RFU) vs. time). a) wildtype, b) homozygous H63D mutant, c) heterozygous H63D mutant. Simultaneous Analysis of ssDNA and dsDNA As expected, after the addition of formamide the electropherograms showed the presence of ssDNA peaks in addition to the dsDNA peaks (Fig. 2 ). The dsDNA profiles seen here are identical to those seen prior to the addition of formamide (Fig 1 ). The ssDNA peaks show differences in relative peak heights and in peak profile – and most notably the heterozygous sample shows a clefted peak. A comparison of the ssDNA profiles for the wildtype, homozygous mutant and heterozygous mutant would constitute a demonstration of SSCP analysis. Although the relative spacing of the ssDNA peaks differs between the wildtype and homozygous mutant, the most obvious difference is the clefted peak seen in the electropherogram of the heterozygous sample. This clefted peak was not present in the corresponding profiles of the homozygous samples (neither wildtype nor mutant). Under these conditions of electrophoresis, the mutational status of H63D is readily apparent. We have developed a combined HA and SSCP method and will report on it elsewhere (that report includes the detection of the common C282Y mutation). To our knowledge this is the first report of a method for performing combined on-chip HA and SSCP. Our emphasis here is on the ability to achieve rapid denaturation and renaturation processes on-chip. Figure 2 Electropherograms of H63D (ss and ds DNA) following addition of formamide (fluorescence in relative fluorescence units (RFU) vs. time). a) wildtype, b) homozygous H63D mutant, and c) heterozygous H63D mutant. Reassembly of dsDNA In order to confirm that we are reassembling DNA on chip rather than denaturing to varying degrees we investigated the on-chip production of heteroduplexes from two samples of homoduplexes (i.e. homozygous) samples. Fig. 3a ) shows the results of the analysis of a mix of the dsDNA from a homozygous H63D and its corresponding wildtype. This first analysis (done without the addition of formamide) showed a peak profile similar to that seen for the pure wildtype or homozygous mutant in Fig. 1 – i.e. no heteroduplexes are evident. We then added formamide to form a mixture in the sample well of homozygous mutant ssDNA with wildtype ssDNA. Once electrophoretically extracted from the sample well, the ssDNA reanneals to form heteroduplex mutants for analysis. As expected, the dsDNA profile of Fig. 3b ) is that of the heteroduplex profile seen in Fig. 1 (the signal to noise ratio of this electropherogram is low because the sample is still primarily ssDNA). This indicates that the DNA is extracted from the formamide-rich sample well as ssDNA and reassembles to dsDNA in the microchip channels. Figure 3 Double-stranded DNA peak profile of the mixture of H63D homozygous mutant with wildtype prior to and after the addition of formamide (fluorescence in relative fluorescence units (RFU) vs. time). a) prior to and b) after The reassembly of ssDNA can also be shown by denaturating the wildtype and homozygous mutant in separate wells. The two denatured samples were injected simultaneously and their ssDNA mixed in the injection channel of a Y-chip (described below). Subsequent separation and detection showed peak profiles (Fig. 4 ) similar to those obtained with the heterozygous mutant for both HA and SSCP. This suggests that the method of denaturation used here is a powerful tool for comparing test samples, either in the same or in separate sample wells. The testing of the wildtype, homozgyous and heterozygous mutants could be conducted by injecting samples from the desired wells without reloading the chip. This would greatly improve throughput. Figure 4 Separate injections of wildtype and homozygous H63D samples recombining on-chip (fluorescence in relative fluorescence units (RFU) vs. time) As will be described in the following section, by varying the electrophoretic parameters we can control the relative amount of dsDNA formed – a significantly larger amount could be obtained. Dynamics of DNA Reassembly After demonstrating that reassembly occurred within the microchannels after extraction from the formamide-rich sample well, it was of interest to investigate how the sequence and timing of the sample extraction and analysis might affect the degree of rehybridisation. In the work presented thus far we used a 60 s injection (although 20 s would probably have sufficed) as a means of drawing sample directly from the sample well to the intersection, from whence it could be analysed. After the addition of formamide and an analysis of the resulting sample (60 s injection and 180 s separation), a series of analyses were performed wherein each short injection (10 s) with a lower electric field was followed by a 180 s separation. These short injections sampled DNA that had remained in the microchannel since its extraction during the 60 s injection from the first analysis. The time required for the DNA to travel from the sample well to the intersection with the applied electric field during the 10 s short injections was calculated to be approximately 53 s. (The short injections are carried out at a lower field than the initial injection.) Thus, the two short injections of 10 s each were not enough to bring in fresh samples from the sample well. Fig. 5 indicates that after the initial 60 s injection the dsDNA concentration steadily increases as rehybridization occurs in the microchannel. Depending on extraction timing (e.g. short injections vs. longer), the relative intensities of the ssDNA and the dsDNA can be varied by a factor of approximately 10, ranging from primarily ssDNA to primarily dsDNA. Further optimisation is possible with changes in microchip geometry. (Shorter injection channels would allow for more ssDNA to be introduced). Figure 5 Electropherograms after successive short injections of H63D heterozygous mutant DNA that show the change of ssDNA to dsDNA in the channels after leaving the formamide-rich environment of the sample well. (fluorescence in relative fluorescence units (RFU) vs. time). a) H63D immediately after a 60 s injection. b) H63D after a subsequent 10 s injection c) H63D after a second subsequent 10 s injection Another interesting feature of Fig. 5 is that following the addition of formamide, the first peak of the ssDNA (marked *) seen in the first analysis after a 60 s (Fig. 5a ) injection is never present after a subsequent 10 s injection (Fig. 5b ) although it can be recovered by another 60 s injection (not shown). The strength of this peak is strongly dependent upon the sample tested (as discussed below). This interesting phenomenon was observed with wildtype, homozygous and heteroduplex samples corresponding to H63D and S65C (data not shown) and the transient peak was clefted for heterozygous S65C (Fig. 6 ) and not clefted for H63D (Fig. 5a ). It appears that this intermediate state may be used to investigate the dynamics of reassembly by a rapid microchip-based method. Figure 6 Electropherogram (ss and ds DNA) after initial 60 s injection of S65C heterozygous mutant DNA (fluorescence in relative fluorescence units (RFU) vs. time). Discussion The integration onto a microchip of an effective means of mutation detection is perhaps one of the most important technological barriers to the implementation of microchip-based medical diagnostics. The best means of attaining sufficiently high sensitivity is by integrating several existing methods of microchip-based mutation detection. The capillary-based analysis procedure developed by Kourkine et al . [ 14 ] is likely to be highly effective in conjunction with the microchip analysis of prepared samples, but since the procedure is based upon the thermal processing (95°C and snap cooling) of diluted PCR products, the integration of this processing onto the microchip may be problematic. The present method allows for such integrations, thereby enabling the mutation analysis throughputs predicted by Medintz et al. [ 15 ] – throughputs as much as 100 times higher that those presently attainable. Another issue is that of signal to noise ratios – rather than dilute our sample (possibly weakening its signal strength) we can analyse the sample essentially undiluted. Moreover, we can enhance the signal strength, as we choose, for either the ssDNA or the dsDNA. As demonstrated here, this method also allows on-chip comparisons of one type of DNA with another. A common problem encountered with HA methods is that they cannot distinguish homozygous mutant from homozygous wildtype – the present technique would allow an on-chip comparison of these samples to produce heteroduplexes that will then indicate the mutational status. The on chip denaturation is produced through the addition of formamide. The melting temperature for this sequence of DNA following the addition of formamide was found to be approximately room temperature, as determined by T m = 81.5 + 16.6(log M) + 0.41 (% G + C) - 0.72 (% formamide) (1) where T m is the melting temperature in degrees Celsius, M is the monovalent salt molarity, (% G + C) is the percent of the guanine and cytosine in the DNA strand of interest, and (% formamide) is the percentage of formamide added [ 16 ]. The melting of DNA was confirmed by forming heteroduplexes on-chip. The ability to quickly re-hybridise on chip allows for rapid investigation of self-assembly mechanisms. In addition, this re-hybridisation enables the formation of duplexes made from a sample and a set of DNA references – i.e. DNA self-assembly within a microchip could be used to form duplexes that, under electrophoretic analysis, would show the results of comparing the sample DNA with each type of DNA in the reference set. This could avoid the need for DNA sequencing. The rapidity of our method appears to provide additional information upon short-lived conformations. Although we have added a thermal re-annealing step as part of our PCR protocol, that step does not affect the results of analysis after adding formamide – i.e. by re-annealing on-chip the thermal reannealing is not needed. The thermal re-annealing stage was added to allow the direct comparison of heterozygous samples from the PCR with heterozygous samples after on-chip reassembly. After adding formamide, the electropherogram of the first separation analysis following any long injection shows a clearly defined transient peak. For H63D samples the transient peak is a single peak, whereas for S65C the transient peak is clefted. We have found that the transient peaks vary in size significantly depending upon the electrophoretic and PCR protocols used. Initially we had assumed that this transient peak indicated that the reassembly of the DNA was not 'random' but instead hybridised first in a high-melting point region, and only slowly thereafter. In this model, the presence of the split-peak would provide information upon the location of the mutation. This suggests that mutation S65C is within the higher melting point domain, while the H63D is not. However, as determined by the Meltmap program (generously provided by L. Lerman (MIT)), neither the H63D nor S65C mutations were within the high melting point region of the exon (data not shown). Several research groups have reported artefacts that arise from ssDNA-primer interactions [ 14 , 17 - 19 ]. Kourkine et al. [ 14 , 18 ] reported that primer-ssDNA complexes can give rise to extra peaks during SSCP. They performed tests with samples of PCR-amplified DNA with and without the removal of the PCR-primers after the amplification step and found that the presence of primers led to the appearance of extra peaks [ 18 ]. A reduction in primer concentration during PCR also proved to be effective in minimizing the appearance of these peaks. Kozlowski and Krzyzosiak [ 19 ] have reported similar effects and suggested that the primer-ssDNA complex may have a different mobility simply because of its changed mass, or perhaps due to a change in conformation induced by the binding. In the context of SSCP, they discussed two approaches for dealing with this effect 1) remove it through purification so as to obtain simpler profiles or 2) use the effect to advantage by achieving higher sensitivity in the detection of mutations. Hennessy et al. [ 17 ] performed similar tests and reported that variations in primer concentration are the likely source of irreproducible SSCP profiles. They too suggested that this effect could be used to increase the sensitivity of SSCP. We therefore suggest that the transient peak is due to the pairing of one product strand with one primer as a result of the renaturation process. The primer-ssDNA complex is primarily ssDNA with a small region of dsDNA at the end(s) of the strand. It is therefore expected to migrate with similar mobility as the ssDNA peaks. The disappearance of the transient peaks with the subsequent short injections may be a result of the complementary single strand binding and displacing the primer. However, the presence of the transient peaks may still provide useful information. The differences in the transient peaks (cleft versus no cleft) between S65C and H63D suggest that their shape may be dependent on the position of the mutation and that the position greatly affects the transport of the transient form of DNA. Thus, the phenomenon of the transient may be a general behaviour that could provide additional mutational information. In corroboration of past work by others [ 17 , 19 ], it therefore appears that the primer effects do provide mutational information. Moreover, this effect can be produced or avoided depending on whether the desire is to avoid the more complex profiles or to use them to achieve higher sensitivity. Conclusion We have developed a method of rapidly disassembling and re-assembling DNA within a microfluidic chip, allowing us control over the relative amount of ss and dsDNA and enabling the performance of rapid hybridisations under electrophoretic control. It has been reported that, when combined, HA and SSCP can provide sensitivities of 100% (e.g. [ 14 ]). In our work to date we have tested a large number of samples, predominantly of HFE, BRCA1 and BRCA2 sequences, and representing approximately several dozen different sequences. All samples containing a mutation have had their mutational status detected by at least one method. We expect then that the sensitivity of the combined methods will be close to 100%. We are now applying this method as part of a study of the application of DNA self-assembly based mutation detection methods (HA and SSCP) to the implementation of highly integrated microchips for performing medical diagnostics. The present work is also an early step towards directing and studying DNA self-assembly within microfluidic systems. The method applied here could be improved significantly by shortening the injection and separation channels and ultimately may even assist in providing the control needed to direct the assembly of DNA-based nanosystems within microfluidic channels. Methods Samples Volunteers who had given informed consent donated lymphocytes from which DNA was extracted and purified by using phenol-chloroform-isoamyl alcohol extractions [ 20 ] or the QIAmp DNA Blood kit (QIAGEN, Mississauga, ON). The purified DNA was solubilized in a Tris-EDTA buffer (TE, pH 8.0) and stored at 4°C. All genotypes were confirmed on an ABI Prism 377 Slab Gel Sequencer (Applied Biosystems, Streetsville, ON), using an ABI Prism BigDye Terminator v3.0 Ready Reaction Cycle Sequencing Kit with AmpliTaq DNA Polymerase (Applied Biosystems). The two mutations tested were H63D and S65C, from HFE Exon 2. PCR was performed on 25 μL reactions of both mutations. Thermal cycling was performed on all the samples as follows: 94 C for 2 min, 35 cycles of (94°C for 30 s, 55°C for 30 s, 72°C for 30 s), and finally 72°C for 10 min, 4°C thereafter. For H63D and S65C, the PCRs are performed with 5 μL of 30 ng/μL of genomic template DNA, 2 μL of 5 μmol/L each of HEX-HFE-2F primer and H63DR primer (Table 1 ), 2 μL each of 10 mmol/L dNTPs, 0.75 μL of 50 mmol/L of MgCl 2 , 2.5 μL of 10× PCR reaction buffer and 0.5 μL of Platinum Taq DNA Polymerase. All samples were re-annealed following PCR by first heating at 95°C for 3 min, followed by a subsequent ramping down of temperature by 1°C per minute until 65°C. The samples were then stored at -20°C. Table 1 Primers Used for PCR Amplicon – Primer Sequence 5' Label Final Concentration HFE Exon 2 – forward 5'-TCA GAG CAG GAC CTT GGT CTT TCC-3' HEX 0.4 μM HFE Exon 2 – reverse 5'-CAT ACC CTT GCT GTG GTT GTG ATT-3' N/a 0.4 μM Reagents PCR reagents (polymerases, buffers and primers) were obtained from Invitrogen (Burlington, ON). GeneScan™ polymer was used for microchip electrophoresis and obtained from PE Applied Biosystems (Foster City, CA). A polymer consisting of 5% GeneScan polymer and 10% glycerol (5GS10G), commonly used for SSCP, was made. Tris borate (Fisher Scientific, Fairland, NJ) with EDTA (Merck KGaA, Darmstadt, Germany) was used as the running buffer in concentrations of 1× and 0.1×. Glycerol (Sigma, Saint Louis, MO) is also added to each in 10% and 1% concentrations respectively (1 × TBE10G and 0.1 × TBE1G). De-ionised formamide (minimum 99.5%) was obtained from Sigma (F9037, Saint Louis, MO). The formamide was aliquotted and kept frozen until required. Microchip Electrophoresis The microchips were purchased from Micralyne (Edmonton, AB) and unless otherwise mentioned were a 4 port double T design (Fig. 7 ) consisting of 4 reservoirs (or wells) linked by two microchannels. One microchannel served as a separation channel approximately 80 mm in length and was nominally 50 μm wide and 20 μm deep. In order to demonstrate control of on-chip mixing we also used an 8-port Y-chip with 8 reservoirs, 2 of which are not connected by any channel and with a third reservoir connected by a 58 mm channel that was unused in this work (Fig. 8 ). Electrophoresis upon microchips was performed using the Microfluidic Tool Kit (μTK, Micralyne) as described previously [ 2 ], with a laser induced fluorescence (LIF) system that provides excitation at a wavelength of 532 nm and detection at 578 nm. The LIF signal was recorded by the μTK with sampling at 200 Hz and these data were recorded to a PC running a compiled LabVIEW interface (supplied by Micralyne). Figure 7 Glass microchip (Micralyne Inc.) with double-T intersection. Figure 8 Glass microchip (Micralyne Inc.) with Y-shaped intersecting channels. Microchip Loading and Electrophoresis The microchip was loaded with 5GS10G polymer without any pre-treatment. The sample well was loaded with 2.6 μL of 0.1 × TBE1G followed by 0.4 μL of DNA sample and mixed. The remaining wells were loaded with 3 μL of 1 × TBE10G. In the case of the Y-chip, 0.4 μL of wildtype DNA was added to 2.6 μL of 0.1 × TBE1G in the first sample well and mixed. The second sample well was filled with 2.6 μL of 0.1 × TBE1G and 0.4 μL of homozygous mutant. The operation of the μTK (injection and separation) was automated through the use of the LabVIEW interface. LIF detection took place 76 mm downstream from the intersection. We have found that the reproducibility of the peak arrival times is within 2 per cent from one run to the next. As such we have not needed to introduce size standards. Injection The sample DNA was brought from the sample well to the intersection and onto the sample waste well by applying 500 V/cm for 60 s. No initial injection was done with the Y-chip prior to denaturation. During this process the buffer and buffer waste well are left electrically disconnected. In doing so the intersection of the two (three) channels is filled with the sample DNA. This stage is referred to as an injection due to the injection of DNA into the separation channel in the sharply defined volume of the intersection of the channels. Separation Immediately following injection, the DNA caught within the intersection is separated by applying 714 V/cm for 180 s between the buffer and buffer waste wells. During this step, the sample and sample waste wells are left electrically disconnected. The effective separation distance was 76 mm from the intersection. Denaturation After the initial run on the 4-port chip, 1.5 μL of the sample mixture was removed and 1.5 μL of formamide was added and mixed. Following Howley et al. [ 16 ], this is sufficient to denature the DNA with a melting temperature of approximately 25.7°C. Since Fig. 4 clearly shows formation of heteroduplexes, we take this to indicate that the temperature was high enough to allow strands to interchange. Another run was then done with the same parameters as above. In the case of the Y-chip, denaturation of each sample was done immediately following the addition of the samples to the wells. A voltage of 400 V was applied between the sample and sample waste wells during a 60 s injection followed by a separation of 180 s. Subsequent electrophoretic runs followed with 10 s injections at 125 V/cm and 180 s of separation at 714 V/cm for both the 4-port and Y-chip. No additional mixing of the two samples for the Y-chip were required Authors' Contributions YZ performed the experimental work with some assistance from TF. DM performed additional protocol development. CB provided overall direction. All authors contributed to the writing of the manuscript and all made substantial contributions to the work.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552323.xml
549070	Integration with the human genome of peptide sequences obtained by high-throughput mass spectrometry	Peptides derived from protein tandem mass spectrometry data have been mapped to the human genome sequence forming an expandable resource for the proteomic data.	Background The recent definition of the complete nucleotide sequence of the human genome [ 1 , 2 ] has motivated the full annotation of the sequence. The true promise of the human genome project, to become the foundation for medical and biological research benefiting human health and quality of life [ 3 ], can only be realized if the coding sequences are conclusively identified, intron/exon structures are accurately described and the potential protein products from each gene in different tissues and cellular states are determined. Current methods for gene-prediction provide useful information but are still limited [ 4 ]. It is not presently possible to predict all features of the genome from its sequence alone. Therefore, the value of the human genome sequence can be enhanced through the collection of different types of experimental data and its integration and validation in a genomic context [ 5 ]. Current use of expressed sequence tags (EST) and full coding DNA (cDNA) sequences is extremely helpful in achieving complete genome annotation [ 6 - 9 ]. However, these data are not sufficient to unequivocally predict which proteins (and with what covalent structure) are expressed in a given tissue. The complete characterization of all proteins across disease states, tissues and stages of development can now be addressed through experimental protein identifications generated by proteomic methods. Experiments carried out over the past years have illustrated that peptides resulting from proteolytic digests of complex protein mixtures can be identified in a high-throughput mode using a combination of liquid chromatography (LC) and tandem mass spectrometry (MS/MS) (LC-MS/MS) [ 10 - 15 ]. Peptides are thus useful as the currency of MS/MS-based protein identification [ 16 ]. By combining a large number of experiments sampling different cell and tissue types, the observed peptides can be mapped onto the genome covering a significant part of its chromosomes. Results and discussion To begin annotating the human genome with protein-level information, we have built PeptideAtlas. The generally applicable procedure to annotate eukaryotic genomes with peptide sequences can be applied when datasets are acquired using different experimental protocols. In each case, sample proteins were first proteolytically cleaved into peptides using the enzyme trypsin. The resulting peptide mixture was then subjected to chromatographic separation by strong cation exchange and reverse-phase capillary chromatography. In addition, those experiments using the ICAT (isotope-coded affinity tag) reagent for quantification included an avidin affinity-purification step to select peptides containing biotinylated, stable-isotope-tagged cysteines [ 16 ]. The resulting peptide pools were then analyzed by electrospray ionization (ESI)-MS/MS. The database search program SEQUEST [ 17 ] was used to assign the resulting MS/MS spectra to a peptide sequence. The confidence of these peptide assignments was evaluated using PeptideProphet [ 18 ]. All of the experimental data products, including PeptideProphet probability scores, are loaded into SBEAMS - Proteomics, a proteomics analysis database built as a module under the Systems Biology Experiment Analysis Management System (SBEAMS) framework. All of the identifications above a certain probability threshold within a specific set of experiments are extracted from the main database tables into another set of tables containing the attributes of each distinct peptide. Resulting from this, 26,840 distinct peptide sequences were identified from 224,973 spectra with identifications of a high probability ( P ≥ 0.9) of being correct. Each peptide is given a unique and stable identifier with an eight-digit number in the form PAp00000001. We then attempted to map the 26,840 distinct peptides to the human genome sequence using the analysis pipeline shown schematically in Figure 1 . An example of visualizing the result in a genome browser is shown in Figure 2 (see Materials and methods for details). The mapping results are summarized in Table 1 . The result of this process is stored as a freely available public resource, the human PeptideAtlas database [ 19 ]. The current build of PeptideAtlas contains peptide sequences identified in 52 proteomic experiments in which proteins were extracted from a particular cell or tissue type, digested with trypsin and analyzed with a mass spectrometer. The 52 proteomic experiments comprised 14 published as well as 38 unpublished human datasets from various cell types such as T cells, B cells, lymphocytes, lymphoblasts, hepatocytes, intestinal cells, hepatoma cells and others. The 14 published datasets contain 47% of the distinct peptides in PeptideAtlas. A full listing of all the experiments and samples currently in PeptideAtlas can be found at the project website [ 19 ]. The raw data for all published datasets is also provided in a repository there. The cumulative number of distinct peptides as a function of the addition of identified spectra (with P ≥ 0.9) in the atlas is shown in Figure 3 . As most of the observable peptides in the proteome are matched with genes, the curve is expected to saturate and adding additional data will yield few new matched genes. However, the current behavior is still completely linear, with approximately 1 in 10 identified spectra contributing a previously uncataloged peptide. Each data point represents an added experiment; the experiments are presented approximately in chronological order of data collection. Among the 52 experiments, there is clearly great variability in the total number of identified spectra contributed as well as new distinct peptides contributed. A repeated, complex-sample experiment might yield many new spectra but few new distinct peptides, while a new sample of a type not previously analyzed might yield relatively few spectra but most of these might contribute a new distinct peptide. Applying our pipeline described in Materials and methods, 25,754 of the 26,840 distinct peptides in PeptideAtlas were mapped to 9,747 (28.6 %) of the 34,091 human Ensembl proteins (version 22.34d.1, 2004-06-02). These proteins represent unique proteins or splice forms from 6,423 genes (27%) of human genes in Ensembl. Some peptides have indistinguishable, perfect protein sequence matches to multiple proteins. These proteins are typically paralogs (protein families), protein isoforms or repeated protein domains in the human genome. We identified 3,718 proteins unambiguously by one or more 'discrete peptides' - peptides that map uniquely to a single protein - in the current build of PeptideAtlas. Those peptides are marked in the genome browser as 'discrete peptide'. 'Degenerate peptides' that map to several protein isoforms are also used to identify proteins. It would thus be more accurate to state that a product of a certain gene, rather than a certain protein, has been identified [ 20 - 22 ]. Moreover, the experimental data from those degenerate peptides generally do not allow differentiation between the sequence alternatives that exist in Ensembl. In fact, not all splicing variants that are in Swiss-Prot are also present in Ensembl and, therefore, it is impossible to ascertain the number of unambiguous identifications at the moment. This limitation underscores the requirement for mapping large-scale proteomic data to the human genome, such as presented in this report to aid in the generation of unambiguous sequence databases. A significant number of distinct peptides (1,086), assigned by SEQUEST/ProteinProphet from over 5,000 MS/MS spectra, could not be mapped to Ensembl database version 22.34d.1. These peptides were identified by SEQUEST searches against the IPI database [ 23 ] or ABCC non-redundant protein database (NCI) [ 24 ]. These peptides are of special interest as they often document interesting biological phenomena such as single-nucleotide polymorphisms (SNPs) and novel splice variants, demonstrating the need for annotating the human genome sequence with high-quality experimental data obtained from expressed proteins. The existence of these sequences also illustrates the flux in the genome annotation and sequence databases. For example, in Ensembl version 18.34.1, only 92% of genes from the previous build were transferred across to the new build. The missing 8% were predominantly inappropriate protein-coding genes coming from large-scale cDNA projects, which have a number of artifactual errors, or from chimeric cDNA clones from cancer cell lines. Experimentally observed, unmapped peptides are an ideal source of information for refining genome assembly and gene prediction. The absence of Ensembl matches does raise the question of whether these peptides are false positives or whether real proteins are missing in the Ensembl database. When these peptides were investigated in more detail it was found that nearly 100 were identified 10 or more times in several different experiments, and that many had protein sequence matches for Swiss-Prot entries. They are therefore likely to be true peptide attributions. For example, peptide PAp00000363 (AGKPVICATQMLESMIK) was identified 626 times at different charge states and with different mass modifications in 22 distinct experiments and mapped to KPY1_HUMAN, a pyruvate kinase M1 isozyme. Interestingly, the protein appears to have a likely SNP, which mutates the valine present in the Ensembl genome sequence to the isoleucine observed in PAp00000363. The 9,747 mapped proteins represent 28.6% of the predicted human proteome in Ensembl version 22.34d.1. The distribution of peptide matches to these proteins (Figure 4 ) revealed coverage of all chromosomes. Void areas were observed in the centromere region of chromosome 1 and the telomere regions of chromosomes 13, 14 and 15. These missing regions represent the unsequenced parts of human chromosomal heterochromatin structures and are therefore expected to be devoid of peptide matches. Very few peptides were observed mapped to chromosome Y. The development of PeptideAtlas and a method for mapping observed peptides to the genome allows us to determine the distribution of multiple peptide hits to specific proteins and the distribution of peptide sequences that are present in multiple proteins. Also, in some cases splice junctions and gene boundaries could be confirmed. Our method allows us also to identify peptides corresponding to abundant proteins such as actin, elongation factor and glyceraldehyde-3-phosphate dehydrogenase, which are commonly identified in high-throughput LC-MS/MS experiments. These proteins are products of housekeeping genes, which are expressed most of the time in almost every tissue [ 25 ], or are structural proteins which are also known to be abundant in cells. The identification of proteins that are specific to a given cell, tissue or disease state allows for the selection of marker proteins. The knowledge of a single marker, or a set of marker proteins, is crucial for the development of new strategies for rapid protein analysis and quantitative proteome profiling [ 16 , 26 ]. In PeptideAtlas we identify proteins to which two or more peptides map. In fact, for some proteins, 100 or more peptide matches were determined. These proteins were often unusually large in size and contained many exons. Examples of such proteins include the 1,462-amino-acid alpha-2-macroglobulin precursor (ENSP00000323929), which was matched by 161 peptides, or the 4,126-amino-acid DNA-dependent protein kinase catalytic subunit (ENSP00000313420) matched by 90 peptides (Figure 5 ), the 2,472-amino-acid spectrin alpha-chain protein (ENSP00000238302) with 102 peptides, and cytoplasmic 2 actin (ENSP00000331514) with 127 peptides. We also identified peptides whose amino-acid sequence is shared by members of protein families or shared domains among proteins in the genome. Peptides were matched to all identical sequences in all proteins. Multiple hits were possible and the resulting peptides were called degenerate peptides [ 22 ], in contrast to discrete peptides that matched one protein uniquely. For example, peptide PAp00001228 (CNGVLEGIR) matched to 26 proteins in the myosin family and peptide PAp00025728 (HCQLAIR) mapped to 23 proteins. Furthermore, our method was able to confirm intron/exon boundaries by identifying peptides that spanned these regions in a gene. We identified 4,800 intron/exon boundary-spanning peptides, corresponding to 2% of the splice junctions in the human Ensembl database, experimentally confirming specific intron/exon junctions. In most cases, these boundaries were already known to exist from cDNA information. However, using peptide information we were able to specifically confirm those boundaries on the level of expressed proteins. In one case (Figure 6 ) we observed a peptide confirming a skipped exon. This event was previously proposed to occur during expression of the A-type lamins in the lung adenocarcinoma cell line GLC-A1 [ 27 ]. The presence of some lamin A10 isoforms can easily be overlooked owing to their relatively low abundance. This new peptide information confirms the existence of this splice variant and shows that low-abundance proteins can be detected through the proteomics technologies described in this paper. The need for public proteomics data repositories is recognized [ 28 ] and we intend PeptideAtlas to become a growing database and public resource. We have structured the system in a way that allows scientists to submit their own MS data for incorporation into PeptideAtlas, thus increasing the number of experiments and identified peptides. Naturally, to be useful for the project, inclusion of third-party data is dependent upon data compatibility and consistent data quality. Consequently, only data with accurate statistical measures of confidence computed by, for example, PeptideProphet, or another published and tested statistical algorithm, will be included. Datasets for which such statistical analyses have been performed can be submitted for incorporation following the procedure detailed at the PeptideAtlas website. Alternatively, data contributors can submit raw MS/MS data directly. This information should preferably be formatted into mzXML [ 29 ] or mzData (HUPO Proteomics Standards Initiative) which are open file formats for the representation of MS data. Other traditionally used data formats are accepted as well. This data will then be searched by the PeptideAtlas curators using SEQUEST to correlate MS/MS spectra of peptides with amino-acid sequences using protein databases such as IPI, and the results will be further analyzed with PeptideProphet. An effort to add support for additional search engines is underway. This procedure will ensure the highest degree of consistency for the data in PeptideAtlas. In the future, the pipeline in general and the data submission process in particular, can be further improved and make compliant with the community accepted statistical data-validation standards and data file formats when such standards emerge [ 30 ]. Please see the submission section on the PeptideAtlas web-site for the most up-to-date submission methods and curator contact information. With an increasing number of included peptides, the utility of the resource will improve, as increasing numbers of genes, exons, transcripts and variant transcripts in many tissues and developmental stages will be verified on the protein level. All MS/MS spectra are stored in the SBEAMS - Proteomics database, from which PeptideAtlas is derived. While at present it is not possible easily to access the MS/MS spectra starting from the public PeptideAtlas interface, this possibility could be added in the future. All spectra for published experiments are available in the mzXML files in the repository. Access to raw spectra can be beneficial for many applications not related to the main purpose of PeptideAtlas. Furthermore, because peptide modifications (for example, phosphorylation) are stored, this information could be displayed as well. It is well understood and discussed in the literature [ 21 ] that all large-scale datasets obtained using high-throughput methods inherently contain a certain fraction of false-positive data. Thus, estimation of false-positive error rates is a very important but often challenging task. One significant advantage of the high-throughput pipeline implemented in this work is that computed peptide probabilities (here produced by PeptideProphet) allow estimation of the upper bound (most conservative estimate) of the false-positive identification error rates for any dataset submitted to PeptideAtlas. As the main purpose of PeptideAtlas is to map peptide identifications to the genome, the most relevant estimate of the false-positive error rates is the one at the level of distinct peptide assignments that have a defined mapping to Ensembl. Initial datasets of peptide assignments to MS/MS spectra, obtained by searching acquired MS/MS spectra using the database search program SEQUEST, were statistically validated using the computational tool PeptideProphet. For each peptide assignment to an MS/MS spectrum, PeptideProphet computes a probability of its being correct, based on its database search scores, difference between the measured and theoretical peptide mass, the number of termini consistent with the type of enzymatic cleavage used, the number of missed cleavage sites and other factors. Probabilities computed by PeptideProphet have been shown to be accurate in the entire probability range and, therefore, can be used to compute the false-positive identification error rate (fraction of all identifications passing the filter that are incorrect) resulting from filtering each dataset using any minimum computed peptide probability threshold [ 18 ]. The false-positive identification error rates for the combined dataset of peptide assignments (all 52 experiments) filtered using minimum probability thresholds 0.7, 0.9, 0.95 and 0.99 are shown in Table 2 . To assess the effect of using a particular probability threshold on the number of peptides in the atlas, we ran the PeptideAtlas pipeline using probability thresholds P ≥ 0.7, 0.9, 0.95 and 0.99. Decreasing the probability threshold increases the number of peptides, both correctly and incorrectly identified, and the corresponding proteins (Table 2 ). The most stringent threshold of P ≥ 0.99 produced 21,030 peptides with protein sequence matches (4,845 protein identifications), almost 8,400 fewer than the lowest threshold of 0.7 (2,252 fewer protein identifications). The P ≥ 0.9 threshold yielded 25,754 peptides with protein sequence matches at an estimated false-positive rate of less than 7%, and we selected this as an acceptable level for the default PeptideAtlas. The number of false-positive identifications could be reduced by selecting a higher threshold; however, a significant number of correct peptides and proteins would then also be eliminated. The additional peptides resulting from the low-probability threshold were valuable for adding additional peptide evidence in combination with higher-probability peptides corresponding to the same protein (peptides corresponding to proteins to which other peptides correspond are more likely to be correct than their probability value indicates [ 22 ]). We provide at our website the option for users to browse or download versions of the Atlas generated with the other P thresholds, which might be useful for some applications. To validate our approach for general use in eukaryote genomes, we have extended our methods to peptides obtained from Drosophila melanogaster LC-MS/MS experiments. We collected data obtained from cytoplasmic, nuclear and membrane fractions derived from a Drosophila S2 Schneider cell line. The resulting 4,406 different peptides with P > 0.9 were compared to the 18,289 proteins (Ensembl fly database version 18.3a.1, 2003-07-01) using the same pipeline as described for human. From the fly, 3,107 proteins could be validated, representing 1,876 (14%) of the fly's genes. These results show that our method could easily be adapted to other organisms, thus opening up the way for comparative proteome-level evaluations of eukaryotic organisms. Conclusions We have annotated the human genome with protein evidence for nearly 10,000 proteins. Although this number only represents a fraction of the genome and still contains some erroneous identifications, it is a first step towards the final goal: to fully annotate eukaryotic genomes via validation of expressed proteins. PeptideAtlas provides a method and a framework to accommodate proteome information generated by high-throughput proteomics technologies and is able to efficiently disseminate experimental data in the public domain. Its significance continues to grow as more data are submitted. Moreover, PeptideAtlas also allows one to address the important question of how big the human proteome is. Due to the technical limitations of current proteomics technologies, it is not possible yet to determine the complete proteome in one experiment. However, if the data from diverse experiments, using different cellular compartments and enrichment methods were combined, the determination of the complete proteome could eventually be achieved. PeptideAtlas offers the framework to answer this question accurately and to determine the size of the complete human proteome using pooled experimental data. Furthermore, PeptideAtlas provides a resource for the development of new avenues of research. The dataset will provide a rich source of data for computational scientists to develop and test new algorithms for proteomic analysis, gene discovery and splice-variant prediction. The methods described here, combined with the ever-increasing power of proteomics and bioinformatics technologies, will facilitate the determination or characterization of protein-coding genes, their features, and their processing and expression in relationship to the sequence of the human genome, thus contributing significantly to our understanding of genome structure. Materials and methods Pipeline The assembly of experimentally derived distinct peptides is mapped to the human genome in the following way. First, we use BLAST [ 31 ] to match the peptides to the Ensembl human protein database. The Ensembl database project [ 32 ] provides a bioinformatics framework to organize biology around the sequences of large genomes and, furthermore, extensive resources and visualization options as well as remote access to the underlying relational databases [ 33 ]. The human genome sequence (release 22.34d.1, 2004-06-02) contains 23,758 genes and 34,091 gene transcripts. Second, complete matches, spanning each peptide's complete length, were used to determine human chromosomal coordinates. The method for retrieving chromosomal coordinates within the human genome accounts for splice junctions; in cases where a peptide maps onto a splice junction, it is projected to both parts of the chromosome, generating multiple sets of coordinates. Third, the results are loaded into a relational database. This database schema (available at the project website [ 19 ]) is able to accommodate data for different PeptideAtlas builds, for different organisms or different reference protein sequence sets as starting material and is thus extremely versatile. Fourth, visualization of the results was achieved using the Distributed Annotation System (DAS) (Figure 2 ) in conjunction with the Ensembl database. DAS allows sequence annotations to be decentralized among multiple third-party annotators and integrated on an as-needed basis by the Ensembl genome browser [ 34 ]. Data collection LC-MS/MS analysis was performed on LCQ, Ion-trap (Thermo Finnigan LCQ) and Q-Tof (Micromass Waters) instruments. To estimate the false-positive error rate on the level of distinct peptide identifications, we first note that there is an almost 10-fold difference between the number of peptide assignments to MS/MS spectra and the number of resulting distinct peptide identifications. This can be explained by the fact that many peptides were sequenced multiple times, with some of the most abundant peptides sequenced more than 1,000 times (for example, peptides PAp00004784, PAp00003568, PAp00026910). While many correct peptide assignments to MS/MS spectra represent the same peptide sequence, the majority of incorrect peptide assignments are expected to be single identifications. As a result, the false-positive error rate on the level of distinct peptides is higher than that on the level of peptide assignments to MS/MS spectra. Second, it should also be taken into account that a considerable fraction of all distinct peptides did not match any Ensembl entry. This is due to the fact that MS/MS spectra were searched against larger databases, such as human IPI, which contained a number of protein sequences not present in Ensembl. The fraction of all distinct peptide identifications that did not map to any Ensembl entry can be estimated using information provided in Table 2 . Among peptides with probability of being correct of 0.99 or greater, only 2.6% of all distinct peptides did not map to any Ensembl sequence. The fraction of unmapped distinct peptides increases to 8.3% among peptides in the 0.95-0.99 probability range, 12.9% in the 0.9-0.95 range and 18.2% in the 0.7-0.9 range, reflecting the increase in the number of incorrect peptide identifications among peptides with lower probabilities. Thus, one can estimate that at least 18.2% of all incorrectly identified peptides did not map to any entry in Ensembl. The false-positive error rate among distinct peptides that mapped to Ensembl (peptides with protein sequence match) can then be estimated to be not higher than the maximum possible number of distinct incorrect peptide identifications that have protein sequence matches (computed by multiplying the total number of peptide assignments to MS/MS spectra by the corresponding false-positive error rate and applying an 18.2% correction to account for peptides with no mapping to Ensembl) divided by the total number of peptides with protein sequence matches. The corresponding estimates are 16%, 6%, 3% and 0.8% in the case of minimum-probability thresholds 0.7, 9, 0.95 and 0.99, respectively (Table 2 ). It should be noted that these are conservative (upper bound) estimates and the actual error rate may be significantly smaller. Population of the database The PeptideAtlas pipeline begins with the download of the Ensembl human protein database from [ 35 ]. Release 22.34d.1 (2004-06-02) was used here. PeptideAtlas peptides were then searched against the human proteins using BLAST with the following parameters adapted for searching small peptides [ 31 ]: -E 1 -W 2 -M PAM30 -G 9 -e 10 -K 50 -b 50 -F F. The BLAST results were then filtered for identical matches and mapped into chromosomal coordinates using Bio::EnsEMBL and Bioperl [ 36 ] Perl modules. The results are uploaded into the PeptideAtlas database and then the Ensembl genome browser. The PeptideAtlas database can handle different PeptideAtlas builds, different organisms and different versions of underlying genome data for maximum flexibility.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549070.xml
553996	Illness-related practices for the management of childhood malaria among the Bwatiye people of north-eastern Nigeria	Background A wide range of childhood illnesses are accompanied by fever,, including malaria. Child mortality due to malaria has been attributed to poor health service delivery system and ignorance. An assessment of a mother's ability to recognize malaria in children under-five was carried out among the Bwatiye, a poorly-served minority ethnic group in north-eastern Nigeria. Methods A three-stage research design involving interviews, participatory observation and laboratory tests was used to seek information from 186 Bwatiye mothers about their illness-related experiences with childhood fevers. Results Mothers classified malaria into male (fever that persists for longer than three days) and female (fever that goes away within three days) and had a system of determining when febrile illness would not be regarded as malaria. Most often, malaria would be ignored in the first 2 days before seeking active treatment. Self-medication was the preferred option. Treatment practices and sources of help were influenced by local beliefs, the parity of the mother and previous experience with child mortality. Conclusion The need to educate mothers to suspect malaria in every case of febrile illness and take appropriate action in order to expose the underlying "evil" will be more acceptable than an insistence on replacing local knowledge with biological epidemiology of malaria. The challenge facing health workers is to identify and exploit local beliefs about aetiology in effecting management procedures among culturally different peoples, who may not accept the concept of biological epidemiology.	Background Malaria is a major cause of death among children in many parts of the world, despite the availability of simple and effective treatments [ 1 ] and is one of the main causes of morbidity and mortality in Nigeria [ 2 , 3 ]. Although, treatment often begins early and at home, a mother's inability to correctly recognize malaria has contributed substantially to child morbidity and mortality due to malaria [ 4 ]. In Tanzania, as in western Nigeria, studies show that mothers were unable to recognize severe malaria despite perceiving the signs and symptoms of onset of childhood malaria as including high temperature and loss of appetite [ 5 ]. Malaria control depends on many factors, some of which have not been studied at the level of rural communities, and in different cultural settings. An understanding of the mother's ability to appropriately manage childhood malaria in the home is crucial. In order to provide a a community with the capacity for dealing with the management of malaria in the home, a study of mothers' illness-related experiences in the management of childhood malaria was carried out among the poorly-served Bwatiye ethnic group of north-eastern Nigeria, the results of which are presented in this article. Methods The study was carried out during the malaria season (May-September) of 1999 among the Bwatiye, a minority group living in six villages along the Benue valley in Adamawa State of north-eastern Nigeria. A multiple-stage approach consisting qualitative, quantitative and laboratory techniques was used. Local knowledge and help-seeking practices In order to orient the data collection process, informal conversations were held with women and lay people in markets and health facilities about malaria illness experience within communities, the method of recognition and help-seeking behaviour. On the basis of received information, key informants were identified and selected (medicine vendors, folk healers, health-workers) and formally interviewed about local knowledge and practices regarding malaria in children. The information was used for designing a cross-sectional household survey questionnaire. Households with children between 9 and 60 months in six Bwatiye villages were identified with the assistance of community health personnel. The study objectives and procedures were explained to household heads and mothers of children that were eligible participants in the study. It was emphasized that anyone was at liberty to decline participation and that those participating were also free to withdraw at any stage of the study. The pre-tested cross-sectional survey questionnaire was administered in the Hausa language to 186 mothers in their homes. Information on the mother's demographic, educational and parity status was collected. Mothers were also asked about symptoms used for recognizing and classifying malaria and to describe their help-seeking practices when a child has malaria. Previous history of child mortality was also recorded. The interviews were complemented with observations of fifteen mothers (with previous child mortality, first childbirth, and more than two children) identified during the interview process. The results were analysed and used for developing a group discussion guide. Six group discussions, each consisting of eight carefully selected mothers of different ages. family settings (polygamous or monogamous) and religions were held. Culturally appropriate explanations were further obtained from four key informants when the data had been analysed and a draft report written. Results Presumptive diagnosis Most mothers (88%) refer to malaria as zazzabi , meaning "hot body" with intermittent episodes of cold shivers ("when the child wants to stay in the sun") accompanied by headache as the main symptom of malaria in their children. Diarrhoea and teething were also important (44.6%) besides dehydration, sweating and vomiting (33.3%), loss of appetite (30.6%) and persistent crying (29.6%). Only very few mothers (6.5%) use faint spells, restlessness, moodiness and withdrawal as symptoms. Only 9.7% mentioned convulsion as an important symptom of malaria in children. Several symptoms were combined, although "child feels cold but has hot body" is an omnipresent clue to presuming that a child has zazzabi (Table 1 ). Table 1 Symptoms 1 used by Bwatiye mothers for the recognition of childhood malaria Symptom Age (years) Parity (childbirths) Previous child mortality Education Total <25 26–35 >35 1 st 2nd-3rd 4 th or more Yes No Literate Illiterate No in sample (%) 54(29) 59(32) 72(39) 42(23) 63(34) 81(44) 36(19) 150(81) 139(75) 47(25) 186(100) Hot body (>38 C ) 77.9 89.3 93.2 100 92.1 79.0 100 85.3 87.8 89.4 88.2 Diarrhoea/ enteric complaints 38.9 47.5 46.6 61.9 81.0 28.4 80.6 36.0 49.6 29.8 44.6 Vomiting 29.6 32.2 37.0 33.3 30.2 35.8 11.1 38.7 37.4 21.3 33.3 Loss of appetite 31.5 30.5 27.4 38.1 28.6 28.4 58.3 24.0 28.1 38.3 30.6 Persistent crying 38.9 32.2 20.5 47.6 15.9 30.9 19.4 32.0 20.9 55.3 29.6 Dehydration /sweating 37.0 33.9 32.9 35.7 31.2 35.8 66.7 26.7 38.1 23.4 34.4 Convulsions 9.3 10.2 9.6 9.5 9.5 9.9 5.6 10.7 8.6 12.8 9.7 Other 3 9.3 5.1 5.5 11.9 4.8 4.9 8.3 6.0 6.5 6.4 6.5 1 Multiple responses allowed 2 Vertical comparison only (based on number of responses) 3 Faint spells, insomnia, Restlessness, moodiness, withdrawal, colds While the number of childbirths and previous experience with child mortality were important factors in the recognition of malaria (Chisquare test, P > 0.05), age and education (Chisquare test, P < 0.05) were not. Association between age and number of childbirths on the one hand and age and child mortality on the other were significant (Chisquare test, P < 0.05). Young mothers had fewer children and fewer child mortality rates than the older mothers. Among first time mothers, 47.6% identified fever with persistent crying, an association unique to this group and used zazzabi as the major symptom (100%), while 79.0% of those with more than three child-births were content to presume malaria on noticing "hot-body with cold shivers" without any additional symptom. Mothers with more than three children (28.4%) and mothers with two or three children (81.0%) differed widely with respect to enteric symptoms. A history of child death in the family plays an important role in a mother's approach to recognition of malaria. Surprisingly, education played a minor role in the recognition of childhood malaria. For mothers who had experienced a child dying, emphasis was placed on diarrhoea. Classification Malaria was classified into two categories; the female and the male. The female form goes away after a couple of days without any particular treatment other than that provided by the household or in some cases the mother. When illness defies home remedies after several days, or deteriorates, it is considered to be the male form. However, should extra-community consultants become necessary, the illness is no longer regarded as zazzabi and "evil doers" are suspected to be using zazzabi to hide a more severe ailment. Convulsions or any form of complications including hallucinations were not regarded as symptoms of malaria but manifestations of other problems that the local healer ( boka ) rather than the health facility could handle. Help-seeking behaviour The sequence of action in progressive malaria infection among the Bwatiye is described in Figure 1 . Figure 1 Flowchart of the sequence of help-seeking habits. Self-medication with herbs and herbal potions was the preferred option (60.2%) for most mothers (Table 2 ). However, young mothers were less inclined to use self-medication with herbs (53.6%) than older women (65.7%). They either waited for their husbands to take action (42.9%) or consulted an older woman or a mother-in-law (25.0%). On the other hand, those who had previously experienced child death (69.4%) preferred self-medication. Most mothers went to the health-care centre as the third line of action. Only few women without history of child death (11.3%) took their children to the health-care centre for treatment. Mothers often give traditional treatments for childhood convulsions and wait until fits cease before the next action. First-time mothers tended to consult their husband before anyone else. Multigravid women would first go to friends with previous experience with malaria in children to collect leftover tablets before informing the child's father or other male relatives. Table 2 Help-seeking behaviour to malaria in children Young women (< = 30) Older women (>30) Previous mortality No history of child births Total (%) Sample (N = 186) 84 102 36 150 186 Behaviour* Self-medication (with herbs, herbal potions): 45(53.6) 67(65.7) 25(69.4) 87(58.0) 60.2 Self-medication with antipyretics from market/shops 9(10.7) 15(14.7) 2(5.6) 22(14.7) 12.9 Treatment with drugs from health facilities 12(14..3) 8(7.8) 4(11.1) 16(10.7) 10.7 Treatment with herbs and drugs from market/shops 14(16.7) 16(15.7) 5(13.9) 25(16.7) 16.1 If symptoms do not go away/get worse Visit the herbalist 6(7.1) 48(47.1) 23(63.9) 31(20.7) 29.0 Visit the health facility/staff 8(9.5) 15(14.7) 6(16.7) 17(11.3) 12.4 Wait for husband before any action 36(42.9) 9(8.8) 2(5.6) 43(28.7) 24.2 Consult mother-in-law 21(25.0) 19(18.6) 2(5.6) 38(25.3) 21.5 Consult a more experienced friend or neighbour 13(15.5) 11(10.8) 3(8.3) 21(14.0) 21.9 * Simultaneous treatment with herbs and modern medicine is very common Extra-communal effort at getting the child back to good health would then be made. Treatment which hitherto was confined to consulting immediate family members would now be extended to the wider community. Local herbalists and other neighbours then contribute to the management of the ailment on the basis of previous experience with similar or other illness. This is often done more as an attempt to show concern and solidarity rather than an understanding of the illness at hand. However if the child remains unwell, evil doers were regarded as blocking the effort to heal the child. The men would gather around the child to administer any medicine that may be brought by other community members. Women are excluded from the room at this stage. The "male form" of malaria is treated with both modern and traditional medicine of all sorts as precautions against spiritual dimensions to the illness. When convulsions set in, the child was immediately assumed to be severely ill and the treatment in any orthodox setting was most often ruled out. A furore of activities would begin to get the child out of the village to some other village, away from the reach of local enemies. The most common anti-malarial drug used by the few mothers, who bought over-the-counter drugs, was chloroquine that was often administered at inappropriate dosage. The 3-day course of chloroquine was never used in most cases mainly due to access and ignorance of the importance of full dose. Children with convulsions were taken first to the traditional healer and if it did not seem to be making any improvement the health-worker was consulted. Some health personnel thought convulsion was a different illness caused by cold and was best handled by a boka . Discussion Aetiological perspective and natural action In Bwatiye, the term zazzabi , a Hausa word, implies an ordinary illness that does not kill. If death eventually results, it could not have been malaria, but some other cause. As in other cultures, an attempt to differentiate between the severe and the mild forms of malaria has led to classification in order to determine at which stage a particular remedy would be required. While the Yoruba and Igbo are reported to have three classifications for malaria or fever episodes [ 6 , 7 ] the Bwatiye classified malaria into two as did the Kenyan tribe described by Munguti [ 8 ]. It has been demonstrated in many studies and for many illnesses, including river blindness and malaria [ 9 - 11 ], that many African communities are unable to perceive illness as a continuum of symptoms with one mild one giving way to a more severe one. Unlike the observations made by Linder [ 12 ] that the onset of fever in children often prompts mothers to seek immediate treatment, the natural primary action among the Bwatiye is to ignore the illness and hope it will go away. Persistent illness only elicits home remedy using available drugs or herbs within the household. Help is sought from the wider community (or from the local healer) only when the home remedy fails. Since there is a strong belief in the spiritual causes of illness, confidence and reliance in the local healer (the boka ) is strong, especially among the older mothers. The local belief is that convulsions are not curable in health facilities and a sick child taken to another village, out of reach of the local enemies, is likely to survive. Most child deaths occur at this stage. In some cases, the child is taken to the health facility, often too late to be helped. This experience, which seems common, has created a strain between the health personnel and mothers, one accusing the other of negligence. The delay action of first-time mothers is interpreted as stemming from fear of rebuke by their husband, while third-time mothers being more confident, would rather wait to be sure it was not an ordinary illness. Third-time mothers are often older, more experienced, but also more entrenched in the local beliefs about malaria aetiology than the young mothers. Although they have more responsibilities, they are economically more independent and are able to take vital decisions with respect to their children with little reference to the husbands, not the case with the younger, inexperienced mothers. Unlike the younger ones, older and multigravid mothers are assisted by their older children who serve as child-minders and help with domestic and economic tasks. When drugs are used, the habit of sharing health resources in times of illness signifying solidarity contributes to non-compliance with management regulations. A mother stops administering drug as soon as the child seems to have improved in health and the remaining drugs are kept for another episode. The act of sharing has been exploited for the development of community management of childhood malaria in Uganda. The Bwatiye and other similar societies could benefit from this management approach. The logical categorization of malaria as female (mild malaria) and male (severe), the interpretation of the class of ailment and its consequences, provide useful clues in the search for healing within the local milieu. Local consultants are useful allies in appropriately managing malaria or other febrile illnesses among the Bwatiye and perhaps other ethnic groups in north-eastern Nigeria. Implications for health education The Bwatiye have an all-pervading belief that enemies often use zazzabi to cover up their evil deeds until it is too late for the affected person to seek a potent cure. This belief may not be ignored when drafting health education material. During post-analysis discussions, the women saw the logic of promptly administering antimalarial drugs at the onset of any febrile illness as a means of eliminating malaria and exposing any underlying evil that may not be malaria. The belief that malaria is incapable of killing a child without enemy intervention is a challenge to health educationists in this and other parts of Africa, where such beliefs prevail. These findings raise a number of points about what is appropriate in health education provision and empowerment for community health-care delivery. Should the development of educational materials be based on promoting biomedical aetiological theory of disease and should we insist that others accept it first, or should we work within the traditional understanding of disease transmission, which includes concepts of evil and ill intent? The health authorities need to consider working within cultural epidemiological parameters if the programme objective is to reduce morbidity and mortality due to malaria. Insistence on acceptance of biomedical models of disease transmission belongs to another goal that may not be directly linked to reduction in morbidity and mortality. Conclusion Until mothers are empowered with the capability to recognize and treat malaria, its impact on child mortality will continue unabated. The need is to educate mothers to suspect malaria first in every case of febrile illness and take appropriate action to rid the child of malaria to expose the underlying "evil". This is a more acceptable form of education than an insistence on replacing local knowledge with biological epidemiology of malaria. The challenge before health personnel is to identify and exploit local beliefs about aetiology in effecting management procedures among culturally different peoples irrespective of their acceptance of biological models of disease transmission. Authors' contributions OBA conceived, designed, coordinated the implementation of the protocol, analysed and developed the manuscript while KKJ led the data collection team.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553996.xml
553983	Ethical challenges in surgery as narrated by practicing surgeons	Background The aim of this study was to explore the ethical challenges in surgery from the surgeons' point of view and their experience of being in ethically difficult situations. Methods Five male and five female surgeons at a university hospital in Norway were interviewed as part of a comprehensive investigation into the narratives of nurses and physicians about being in such situations. The transcribed interview texts were subjected to a phenomenological-hermeneutic interpretation. Results No differences in ethical reasoning between male and female surgeons were found. They reasoned in both action and relational ethical perspectives. Surgeons focused on their relationships with patients and colleagues and their moral self in descriptions of the ethical challenges in their work. Dialogue and personal involvement were important in their relationships with patients. The surgeons emphasized the importance of open dialogue, professional recognition, and an inclusive and accepting environment between colleagues. Conclusion The surgeons are personally challenged by the existential realities of human life in their relationships with patients. They realized that ethical challenges are an inherent part of performing surgery and of life itself, and say that they have to learn to "live with" these challenges in a way that is confirmed both socially and by their inner moral self. This means accepting their personal and professional limitations, being uncertain, being fallible, and being humble. Living with the ethical challenges of surgery seems to contribute to the surgeons' confidence and vulnerability in their professional identity.	Background It is important for surgeons to be and to act in a right and good way towards patients, relatives, and colleagues. Studies have shown, however, that physicians often are in doubt about the best and correct actions to take for the patients in specific situations [ 1 - 3 ]. This question is not only a medical one, but can be understood in both action ethics and relational ethics perspectives. An action ethics perspective concerns questions as to what ought to be done in ethically difficult situations and why. In this perspective, ethics often centres on difficult ethical dilemmas and decision-making. Ethical dilemmas occur when physicians have to choose between at least two alternative and equally difficult courses of actions. Because neither of the alternatives have positive outcomes, they have to choose between two evils [ 4 ]. Ethical dilemmas can also be understood as conflicts between different courses of action that result from general and mutually exclusive ethical principles in medicine [ 5 ]. A relational ethical perspective means reflecting on the challenges we encounter in our relationships with others and how to fulfil our social roles and obligations in a good way – as a human being, a surgeon, and a colleague. It tries to answer questions such as "How can I adequately meet the challenges that confront me in the relationships in which I am involved in this situation?" [ 4 ]. The qualities that make a person a good physician are not only individual traits but they are characteristics of the relationships. One way to describe a good physician or a surgeon is to count the number of characteristics or virtues which are portrayed. According to MacIntyre [ 6 ], another way is to speak about a good physician or a surgeon. Narrative ethics focus on what life demands from us in different situations and how we ought to respond to these challenges [ 4 , 6 , 7 ]. Action and relational ethical perspectives are not interchangeable as surgeons have a dual responsibility for their actions in specific situations as well as their way of being in their relationships [ 4 , 7 ]. Being a good surgeon presupposes both professional competencies based on scientific and clinical knowledge and skills, and being present and showing respect and compassion for patients [ 4 , 5 , 8 ]. Physicians become involved in the patients' problems both in a professional and moral sense [ 9 ]. Traditionally, it has been assumed that compassion can impair competence and that they cannot coexist [ 10 ]. Many authors argue that the respect and trust in the physician-patient relationship have eroded in recent years in spite of the physicians' increased therapeutic capabilities [ 11 , 12 ]. Shorter hospital stays and organizational changes in the hospitals are said to lead to surgeons spending more time in the operating theatre and less time talking to patients and establishing a trusting relationship [ 7 , 13 ]. An open and honest dialogue between physicians and patients can be difficult to achieve as medicine becomes more complex, fragmented, episodic, and impersonal, according to Jones [ 14 ]. Medical problems are always existential problems too because suffering, anxiety, life, death, and cure involve the core of human existence. Physicians are working with emotionally intense issues and have to accept the possibility of failures, continuing suffering, and death on the part of their patients [ 12 ]. Meeting patients who are emotionally distressed or tragically injured can make surgery emotionally challenging [ 15 ]. Patients may also elicit emotions of anger and frustration, fear, and despair in physicians [ 16 ]. Research suggests that the delivery of bad news can be particularly troubling for both patients and physicians because of the emotional component, and that physicians experience great discomfort in such situations [ 17 ]. Physicians have to function at an optimal level despite these challenges. Straume [ 18 ] regards the physicians' vulnerability as a result of their overwhelming responsibility and experience that patients' demands often exceed the physicians' ability. Although most clinicians are aware of the uncertainty and the limitations of medicine and their responsibility to try to reduce the likelihood of error, [ 19 , 20 ] the boundary between medical errors and accidents is not evident [ 21 - 23 ]. Physicians may have difficulty acknowledging personal errors because they can be experienced as personal defeats and thus confirm that physicians are vulnerable [ 20 , 24 ]. Surgeons' relationships with their colleagues have become more important as surgery has changed from relying heavily on the performance of individual surgeons to relying on a team of providers [ 8 , 25 ]. Several studies have found a lack of dialogue and support structures among physicians [ 26 , 27 ]. Physicians are said to have no tradition for open discussions about uncertainty and conflict areas in their practice, nor are they comfortable talking openly about their personal emotions and problems [ 12 , 16 , 24 ]. Adverse events are said to be generally managed by the conspiracy of silence [ 20 , 22 ]. The vulnerability of patients is emphasized in the literature of medical ethics. Less is written about the vulnerability of the physicians in their relationships with patients, relatives, and colleagues. Being involved may engender feelings of helplessness and vulnerability in the physician. MacLeod [ 28 ] argues that physicians have to accept their vulnerability and be able to express and share it in order to be able to live with the tensions. Little [ 29 ] suggests that understanding the peculiarities and intensity of the patient-surgeon relationship may help the surgeons understand the vulnerability of both patients and surgeons. Studies show that physicians seem to experience uncertainty and fallibility in different ways. Less experienced female physicians in paediatric care put on an air of certainty while the more experienced gained a kind of security by allowing themselves to feel uncertain [ 30 ]. Female physicians in geriatric care seemed to accept their own vulnerability and fallibility [ 31 ]. Experienced male physicians in pediatric care related personal security to their professional experience while the less experienced thought that advances in medical knowledge and ethical guidelines would make them more secure in their work [ 32 ]. Henriksen and Hansen [ 33 ] found that general practitioners seemed to strive for the ideal of a humble attitude towards problems; being too self-confident was regarded as a threat because of the increased risk of downfall. Few empirical studies have been found that explore the ethical challenges of surgery from the surgeons' point of view and their experience of being in ethically difficult situations. The present study is part of a comprehensive investigation of ethical reasoning among male and female physicians and nurses within surgical units. The results of this interview study with male and female surgeons will be presented in two articles. The present study describes the surgeons' experiences of being in ethically difficult situations from a relational ethics perspective. The other paper describes the ethical dilemmas as experienced by the surgeons from an action ethical perspective. The results from the interviews with the registered nurses (RNs) working in surgical units are in progress and will be addressed in a third paper. The aim of this study is to explore the meaning of being in ethically difficult situations in surgery as narrated by male and female surgeons. Methods Participants and setting Five male and five female surgeons working on surgical units at a university hospital in Norway participated in the study. All were experienced and had been working in health care from 9 to 31 years (median = 21.5), and in surgery between 5 to 21 years (median = 13). The surgeons worked full time and were on duty when the interviews were conducted. No individual characteristics will be disclosed in order to guarantee confidentiality. The surgeons gave their informed consent to participate in the study, which was also approved by the 5 th Regional Ethics Committee in Norway. Data collection Interviews The interviews were conducted by the first author and lasted from 35 to 75 minutes (median = 55). They were tape recorded and subsequently transcribed verbatim. The interviewees were asked to tell about one or more ethically difficult care situations that they had experienced in their work as surgeons. What constituted an ethically difficult situation was not defined, allowing the interviewees to determine what they considered ethically difficult themselves. The aim of the interviews was to obtain as many rich narratives as possible without interrupting the surgeons' narrative flow and reflection. If the surgeons did not spontaneously reflect on the events they talked about, their reflections were sought. Questions were asked when the interviewer wanted the interviewees to elaborate on their stories or had difficulty understanding the narration. These questions referred to the interviewees' thoughts, feelings, and actions [ 34 ]. Field notes were taken during the interview as aids to the interviewer's memory and in order to make it possible to understand the interview text in relation to its context, e.g. arrangements and interruptions. Nonverbal communications that seemed relevant were also noted, such as laughter and long pauses. The transcribed text was compared with the field notes and adjusted if necessary. Data analysis Interpretations The method of interpretation used was inspired by the French philosopher Paul Ricoeur's phenomenological hermeneutics [ 35 ], and developed at the University of Tromsø (Norway) and Umeå University (Sweden) and has previously been used by Lindseth et al., [ 36 ] Udén et al., [ 1 , 13 ] Søderberg et al., [ 37 ] and Sørlie et al. [ 30 , 32 , 38 ]. This method is useful to elucidate the narratives of people's experiences. The method of interpretation proceeds through three phases, which constitute a dialectical movement between the whole and the parts of the text and between understanding and explanation [ 35 ]. Each interview was regarded as a text. However, it was not what the texts said that was a subject matter to be investigated, but rather the focus was on the ethics expressed in them or the essential meaning of ethically good phenomena (or the essential meaning missing in ethically poor phenomena)[ 39 ]. First, a naïve reading was made of all the transcribed interviews as a whole to gain a first impression of the surgeons' experiences of being in ethically difficult situations during their clinical work. The repeated naïve reading was made as open-minded as possible, without any deliberate analysis of the text. The naïve reading shows the direction the structural analysis may take. Second, a structural analysis was performed in order to validate or refute the initial understanding obtained from the naïve reading and to explain what the text was saying. The interviews were divided into meaningful parts and patterns, i.e. one sentence, parts of a sentence, or a whole paragraph with a related meaning content. The meaning units were condensed and discussed among all the authors, and themes and subthemes were identified, and presented in 'Results'. Third, a comprehensive understanding was developed, taking into account the authors' pre-understanding, the naïve reading, and the structural analysis (results). The text was read as a whole and interpreted in relation to relevant theories of ethics and results from previous investigations into the meaning of being in ethically difficult care situations [ 39 ]. The comprehensive understanding is presented under the heading 'Discussion'. The analysis was conducted by all the authors and the interpretative agreement was considered satisfactory and to be the most useful understanding of the meaning of the surgeons' experiences of situations of ethical difficulty. The authors' interpretation was not shared or validated with the surgeons. In this study, the focus was on obtaining the meaning of the text, which cannot be validated by the interviewees. A kind of validation is accomplished by the structural analysis as the objective part of the interpretation process [ 39 ]. According to Ricoeur [ 35 ] a text has multiple but not infinite meanings. One particular phenomenological hermeneutic interpretation should therefore be seen as one of several possible interpretations and as arguments put into ongoing discourses, in this case, about ethical challenges in surgical care. Results Several readings of the interview texts revealed that the surgeons told about ethical challenges that confront them in their relationships with patients and colleagues and about their experiences of living with these challenges. They also reported their experiences of ethical dilemmas in surgical practice. No gender differences between the male and female surgeons were found in the analysis of the interviews. The results showed that each surgeon created many and detailed narratives. When the surgeons were asked to narrate their experiences, they did not differentiate between action and relational perspectives in their ethical reasoning. This is an analytical distinction made by the authors in order to structure the results. The authors therefore decided to separate the presentation of the results in two papers: one paper about the surgeons' experiences of being in ethically difficult situations from a relational ethics perspective and the other paper from an action ethical perspective according to the theory presented in the introduction [ 4 ]. This paper presents the ethical challenges of surgery as narrated by five male and five female surgeons. The themes and the subthemes from the structural analysis are shown in Table 1 and presented in the text below. Direct quotations from the interviewers are included to illuminate the results. Table 1 Themes and subthemes that emerged from the structural analysis of interviews with the surgeons. Themes Subthemes Dialogue with patients Openness and honesty Involvement Social confirmation Professional recognition Open dialogue Self confirmation Responsibility Uncertainty Fallibility Confidence Humility Dialogue with patients Openness and honesty The surgeons emphasized the importance of dialogue with patients and especially being open and honest about all aspects of their treatment and care. Talking to patients about difficult issues before or after the operation is experienced as an important part of the surgeons' responsibility. Openness and honesty is especially important when surgeons had to tell patients that they have cancer or a fatal disease, when something had gone wrong, or the operation did not turn out as successful as expected. Using frightening words such as "death" and "cancer" is also considered to be part of an open and honest dialogue. The surgeons also felt responsible for "not involving patients with bad news they were not ready to receive, and feeling one's way in what it's right to inform about". They said that if the patients are taken seriously and are talked to in a way that they understand, there is seldom difficulty reaching a mutual understanding about diagnostics and treatment. The patients are always told the truth about their disease, although it was experienced as an emotional burden for the surgeons to disclose bad news or present difficult decisions to the patients. "It's a burden to tell the patient that we will withdraw all active treatment. You feel a bit guilty; in a way you feel that it's your fault if the treatment does not succeed". They stressed that knowing the patient from previous meetings is important to them, as is having antennae for how the patients experience their life and the present situation. "If I do not manage to achieve what I am trying to do, I always tell [the patients] what the situation is. I never try to conceal anything. That will only torment them". Talking openly with patients is also important for the surgeons in situations when they are in doubt about the right thing to do. They experienced relief if the patient had an answer to their doubts. They said that patients who are seriously ill usually have thought about life and death issues and have a conception of their condition, and that the question of withholding or withdrawing treatment seldom comes as a surprise to them. Some patients say that they have lived a good life and do not want an operation. Others strongly want an operation even though their prognoses are poor. "We often reach an agreement about ending a treatment that either does not lead to a meaningful life afterwards or leads to a life that would be experienced as a heavy burden". Involvement The surgeons said that they become personally involved with their patients, focusing on patients as persons and their quality of life as much as on their medical treatment. Being involved and knowing the patients' background and what they really want in life makes difficult ethical decisions easier to handle. The surgeons explained that caring for the patient can be felt as a personal, emotional burden. "It's not easy when people you have established a relationship with die. The only way to run away from it is to relinquish your responsibility. But that means disassociating yourself from or rejecting the patient. So you have to care, to be involved and to be a human being". The surgeons are involved in many patients' lives and destinies and said that keeping a certain distance protects their feelings and is a way of caring for themselves. They said that keeping a certain distance is necessary in order to give the patient neutral advice and the most suitable medical treatment. The surgeons feel a responsibility to care in situations where they find it difficult, for instance when they dislike the patients' personality, behaviour, or values. They said they work hard to get involved and care for demanding and non-compliant patients. Social confirmation Professional recognition Ethical challenges are discussed in both formal and informal social arenas among the surgeons. All new patients are presented at the daily morning meeting between all the surgeons, including what had been done to them and why. Only questions and short objections to the patients' diagnostics and treatments are raised at these meetings or shortly after. The surgeons arrange separate meetings to discuss problematic cases. "We assemble the nurses, the anaesthesiologists, the surgeons, and even other specialists like the nephrologists when we have patients who reside a long time in the intensive care unit. In a way we create a meeting-place for the case and discuss whether we should withdraw a life-sustaining treatment or start additional treatment for a kidney failure for instance." The informal running dialogues during the day were equally important for the surgeons. "You have to make the decision yourself, but we always discuss the problem together before difficult decisions are made. The discussions do help and are experienced as mutual support". The surgeons expressed confidence in the consensus that usually is achieved in these discussions. "I know that I would be content to receive the treatment we agree upon myself". The surgeons emphasized the importance of having a caring relationship with their colleagues. They said that talking together and giving and receiving collegial support is necessary in order to live with the personal responsibility of being in ethically difficult situations and with the emotional burden of decision making. "People say that surgeons are a bit tough and do not talk about feelings, and that may be true. But in my experience we really care for each other. Perhaps we do not go all mushy about our feelings, but we understand when someone is in difficulty. I have experienced receiving good support in such situations. Colleagues contact you and say: "Ok, listen, a couple of years back the same thing happened to me", or: "You must not take this too hard, it could have happened to anybody". That helps". Personal and emotional support is informally and silently shared among trusted colleagues and great value is attached to it. After having presented a difficult decision about withholding treatment at a morning meeting, one of the surgeons commented: "That same afternoon, four or five of the most experienced surgeons came to me independently, and told me that they thought that what I had done was great. They said that most surgeons were not able to do what I had done. I remember it well because I think it was so well said". Open dialogue The surgeons emphasized the importance of "playing with an open hand" and that openness and honesty in the relationships with colleagues presupposes a trusting atmosphere that allows everybody to feel free to voice their opinion and be listened to. They feel that it is important that everyone who is involved in the treatment and care of the patient should have an opportunity to express their opinion and to be heard before any final decision is made. Openness and honesty are particularly important when medical errors occur or when something has gone wrong during an operation. Talking about medical errors or mistakes is considered an opportunity for learning and for improving surgical routines. "You have to have an including and accepting environment that allows you to say that you could have chosen a different solution. If there is no room for you saying something like that, then there will be a tendency to conceal it. We all make mistakes and we all make wrong deliberations and sometimes choose bad solutions. We have to live with that. Therefore it's important that we try to learn from the cases where something [erratic] happens". The surgeons focused on the necessity of dialogue and cooperation with their colleagues. "We are used to working close together in a team and it makes these difficult situations easier to handle". They found it satisfactory to work in a hospital because "there is always someone you can ask for advice when in doubt". They said they find it easier to talk to patients about difficult treatment options when the question has been discussed with experienced colleagues or senior surgeons first. This is especially important for less experienced physicians. The surgeons also make a contrast between the importance of cooperation with colleagues and the burden of being alone and being responsible decision makers. Self- confirmation Responsibility The surgeons said that they experience ethically difficult situations as an important part of their everyday activities that cannot be separated from the rest. "These situations are a part of our profession that are not necessarily experienced as difficult, but are sometimes unpleasant to be in". They said they have to experience these situations personally and be involved in order to understand them and to learn to live with them. Some ethically difficult situations are experienced as dramatic and tragic, especially when the surgeons feel the personal responsibility for saving the lives of trauma patients after major accidents. "If you are not able to cope with being in this situation and be responsible, and be the leader of the trauma team trying to save peoples' lives after a serious accident, if you cannot do that but go on wondering if you have done something wrong, then I think you will find something else to do". The surgeons told about situations where they are alone on duty and responsible for rapid deliberations and decisions in acute and emergency situations as especially challenging. "There are no other times in this job when you feel as lonely as you do in those situations". The decision whether to continue or withdraw treatment from traumatized young patients and children is experienced as especially challenging to make alone. "It's not easy [being alone]. It's the kind of decisions you often ponder about for several days afterwards, also at home. It's the sort of decisions you really try to closely think through, and it often troubles you even after you have made a decision". The surgeons said they have to learn to live with the unpredictable consequences of their decisions. "The practice of surgery is very specific and you feel more responsible in a way than in other areas of medicine. You have the feeling of being the direct cause when things go wrong, e.g. that you operated at the wrong time, you operated incorrectly or that you should have found the right diagnosis earlier". Uncertainty The surgeons spoke about "living with" the inherent uncertainty of surgery and emphasized that they can never be completely sure of the right thing to do in ethically difficult situations. They have to live and work with the uncertainty of the course of the disease, the patients' chances of survival, the risk of serious and fatal complications and the uncertainty of the patients' quality of life after extensive operations. The surgeons said they have to learn to live with the uncertainty of their deliberations and operations as there can be no right answer to ethical challenges, and no criteria to guide them when they make their decisions. Living with uncertainty is experienced as both frightening and satisfying. The surgeons said that not knowing what to do in an uncertain situation, finding a way, and succeeding in their attempt to restore health or save life is a satisfying aspect of their work. They also commented that they have to live with their doubts and fears of being too active and the risk of knowingly imposing severe complications and a poor quality of life on their patients. Fallibility The surgeons said that their aim is to make all patients better. Accepting the limitation of surgery and not being able to cure a particular patient or alleviate his or her suffering is not an easy task for surgeons. They commented that it feels difficult not to be able to or not to be allowed to help patients. Sometimes they feel guilty if the treatment does not succeed, but said that they have to accept the possibility of making mistakes as an essential part of the profession. "You have to face the reality of how things are. What we are dealing with are human beings in marginal situations where things can go wrong. Everybody who is dealing with these things makes errors of deliberation and judgment. It's a part of the game that you have to live with when the margins are so tight. It would not be better if you quit. Your dearly purchased experience will be of use to nobody then". Confidence The surgeons said that they are focused on healing and curing and that they try to operate on the patients in most situations. "I believe that surgeons feel that it's good to do something, curing, and saving lives. That is what we have learned to do. We are in a profession that can do many useful things and that is the gratifying part of practicing surgery, – that you can identify a problem and do something about it". The clinical experience of deliberating and choosing and finding workable solutions in clinical and ethical difficult situations makes the surgeons confident. "Most surgeons are action-oriented. If not, it's almost impossible to practice surgery, because you have to makes decisions all the time and be accountable for your decisions". Receiving social confirmation from patients, relatives, and colleagues when they succeed contributed to the surgeons' experience of confidence in their own decisions and actions. The surgeons also said that having the courage of your convictions and a set of personal ethical values is equally important in order to do a "proper and conscientious job". This means acting according to the patients' best interests and the standards of their profession. Their conscience helps them to decide which action is morally wrong, to make controversial decisions, and to voice personal, professional, and moral opinions to colleagues and patients. Humility The surgeons emphasized that although they deal with existential issues of life and death, they do not rule over these human conditions. "The essence is that we are not almighty. We do not save lives, we just postpone death. The only thing that is certain is that we are all going to die. You are just doing repairs and trying to help the patients to live a bit longer. " "Fortunately it's not we who rule over life and death. What we can do means less than you think and that is how it should be. I think it's important that we all acknowledge this and everyone else too". All the surgeons said they are against active euthanasia. They also commented that a humble attitude in their work helps them acknowledge their personal and professional limitations and to recognize what they cannot and ought not do. "There is a dimension in surgery of being right about having a good reason to operate and being considerate about the benefits of carrying out an operation". Discussion The surgeons in this study reasoned in a relational ethics perspective, focusing on dialogue, openness, and involvement in their relationships with patients and colleagues. While the surgeons in the interviews described their experience of being in ethically difficult situations, it seems at the same time that identity was central to their experience of ethics and their enactment of moral agency. The surgeons identified themselves by telling about their ethical experience, expressing the way they are living their identity. By narrating our lived experience, we give meaning to our experiences. We understand ourselves through the stories we tell and live, as well as those told about us and by interpreting them. Personal identity is said to be constructed through the stories we tell about our lives, stories which are in turn shaped by more general institutional, cultural, or national meta-narratives that live within the culture, of surgery, medicine, and society [ 40 - 42 ]. Thus, it seems that the surgeons' identity has a narrative structure and is narratively derived. Being open and honest about all aspects of patients' treatment and care is experienced as important for the surgeons. This is even true in situations when they have to disclose bad news and tell the patient that they are not able to or did not succeed in their efforts to restore health. Openness of speech is one of the spontaneous expressions of life, designated as "utterances of life" according to the writings of Løgstrup [ 43 ]. That openness and honesty are spontaneous expressions means that they are performed in an unconstrained manner and without ulterior motives. Human life means expressing oneself with the expectation of being met by others according to Løgstrup [ 43 ]. Openness and honesty are required in trusting relationships. When openness and honesty of speech are missing, closeness and dishonesty in the relationships result [ 43 ]. Sincerity in our relationships with others is a source of satisfaction and means being open and involved and allowing oneself to be moved and impressed by others according to Pahuus [ 44 ]. It is also probable that the surgeons' social confirmation increases when the patients meet surgeons who through dialogue turn out to be caring human beings. Although disclosing bad news to the patient is experienced as an emotional challenge by the surgeons, concealing the truth does not seem to be an option they consider. Concealing the truth from the patient is experienced as difficult. It seems that surgeons need to share their experiences by speaking openly to patients so they can live with their demanding professional life in a satisfactory way. Therefore, honesty and openness are important because the relationships with patients are important for the surgeons' lives. The surgeons said that it is important not to hide their uncertainty and doubts from themselves or from their patients. They accepted that uncertainty is an inherent part of their profession and realized that they often have to decide in spite of lack of scientific knowledge. They also deal with problems that may have no desirable solutions. Previous studies of surgeons [ 21 ], female physicians working in pediatrics, [ 30 ] and geriatrics [ 31 ] revealed that they are aware of and accept their own uncertainty and fallibility as inevitable in their professional lives. The surgeons said that they sometimes feel guilty when they have to disclose the bad news to patients that the treatment did not succeed. We feel guilty when we fail to do what is required of us in a situation according to Løgstrup [ 43 ], either by not answering the ethical demand of the other or by betraying something valuable in our own life – values and beliefs we hold to be good and right. The surgeons are often in situations in which they have to use their clinical competence and practical wisdom to choose between conflicting values and obligations and being responsible for their decisions and actions. They can never know for certain in a given situation if their actions are absolutely right or wrong. They have to take risks, knowing that they may not succeed in their efforts to give patients a chance of survival. Feeling guilty is our fundamental ethical condition as human beings according to Pahuus [ 44 ] because we have only limited possibilities at hand in particular situations and in life as a whole. Choosing one solution in a situation means excluding a multiple of other desirable possibilities. Thus, feeling guilty is a side effect of trying to give meaning and direction to our lives. Reflecting on our feelings of guilt means having a dialogue with and negotiating with our inner self, thereby evaluating what we hold to be important in our life with and for others. Feeling and being guilty is a heavy burden because human existence also contains that which is final, irreparable and cannot be changed [ 44 ]. Feeling guilty can also be understood in relation to what Ricoeur [ 40 ] has called an ethics of memory – that people never will or can and must not forget the bad or good things in history. The surgeons cannot and must not forget situations in which they failed to do the right or good thing. The surgeons said that they experience a relief when the patients give answers to their own uncertainty and doubts about the right or good thing to do. It seems as if the surgeons are relieved when patients want to decide for themselves and take responsibility for their lives, appreciating that they do not hand over this responsibility to the surgeons. The surgeons in this study do not fit into the traditional and stereotypical picture of the paternalistic and authoritarian physicians [ 45 ]. On the contrary, the surgeons in this study seem to value being in a dialogue with patients and acknowledging their autonomy. Having to take responsibility for patients' lives is experienced as an ethical challenge and a personal burden in situations when the patients are unable to express their autonomy. The surgeons emphasized the importance of being personally involved with the patients by focusing on patients as persons and their quality of life as much as their medical treatment. However, they experienced the challenge of finding the balance between involvement and distance in the relationships, between caring for the patients and themselves, and in giving professional and neutral information. Distance based on knowledge, skills, and experience with previous cases and situations is necessary in order to help the patients in a professional manner [ 43 , 46 ]. Personal involvement in a situation is necessary in order to know which fundamental values are threatened in both the patients' and the surgeons' lives. The aim of a dialogue is to abolish the distance between patients and surgeons in order to establish a space for mutual understanding. Involvement does not only mean acting according to the patients' expectations because surgeons are also responsible for their professional conduct as well as their moral integrity. Professional distance means taking the other's perspective in the situation, reflecting, deliberating, and using one's experience of similar cases. The surgeons also learned from their experiences through discussions and dialogue with experienced and trusted colleagues. They may throw new light on situations and increase the surgeons' abilities to deliberate and decide in their patients' best interests [ 46 ]. Finding the 'right' balance between involvement and distance in a situation is an expression of what Aristotle [ 47 ] calls practical wisdom (phronesis) or the practical knowledge of virtuous persons. A virtuous person is able to find the middle path between two extremes. The right and good thing to do in a particular situation involves "to the right person, to the right extent, at the right time, with the right motive, and in the right way" according to Aristotle [ 47 ]. In their narratives, the surgeons focused on cooperation and the relationships with their colleagues. They experienced being listened to and together they seem to create and recreate a collegial environment where they can express their doubts and admit their errors of deliberation and decisions. The surgeons said that they experience taking difficult decisions alone as a burden. Working in a hospital is experienced as satisfying because they appreciate working together. They seemed to accept that they are mutually dependent on each other and appreciated the support they receive in dialogue with colleagues. Interdependence and exchange is a fundamental way of human lives according to the ethics of Løgstrup [ 43 ]. He calls this "display of life" which means that human beings are fundamentally dependent on one another and that we always are giving and receiving something in our relationships with others. This also means that each individual is challenged to take care of that part of the other's life that is in one's own power. To be seen and listened to by colleagues, be taken seriously, and comforted when something goes wrong is an answer to fundamental human needs. Openness about personal doubts and failures also make the surgeons vulnerable. The interdependency among surgeons in this study is not experienced as a burden, but as mutual support. The surgeons' acceptance and appreciation of their collegial dependency may be an answer to why surgeons are able to stand in ethical difficult situations and live with them in a satisfying way. The dialogue with colleagues reduces the surgeons' doubts, although their doubts cannot and do not disappear. The surgeons receive social confirmation for being good surgeons through dialogues with patients, relatives, and colleagues. Their experience of being of use to society as a whole also confirms their identity as good surgeons. To be confirmed is to be seen, listened to, and accepted by other persons. A person's identity is dependent on other persons in good or bad relationships according to Ricoeur [ 40 ], as well as an inner voice from the person's own consciousness. The surgeons in this study emphasized the importance of dialogue with patients and colleagues. All identities are socially bestowed, socially sustained, and socially transformed and people make meanings of their actions by participating in communicative contexts. We all need recognition and social confirmation from others in order to construct our identities [ 41 , 48 ]. Thus the self-image of a surgeon can be maintained only in social contexts in which others are willing to recognize him or her in this identity. There is a close relationship between identity management and values in a society since human beings usually want to be something they find valuable. What constitutes morality is created, recreated and communicated in social relations [ 41 , 49 ]. Narratives are a display of moral identity according to Jordens and Little [ 42 ], where the speakers present themselves as experienced, knowledgeable, and ethical. Institutional settings provide the narrative auspices under which identities are articulated. Telling stories is necessary for moral agency as we must be able to account for our actions and the actions of relevant others. We also tell stories as a sort of self-examination by self-exposure. We often find out what we think and who we are by listening to what we say. The modern self is described as multi-voiced and dialogical and moral action is determined through a process of inner dialogue. To open up to the different voices within and enter into a dialogue, one can either reach consensus with oneself or experience conflict and dissatisfaction with oneself [ 41 , 49 ]. The surgeons' identities are constructed through negotiation with themselves, their patients, and colleagues within a context of social organization. The view of ourselves is shaped by the values of the larger society and what collectively is tacitly deemed to be right and good actions. Conversely, one's conception of how ethical the collective group is arises through one's personal view of what constitutes right and good actions [ 21 , 49 ]. Surgeons attempt to reconcile their relational experiences and their actions with their identity as moral agents as they live and practice relationally with others within the institutional values of their workplace. The modernization and secularization of western societies emphasize the ideology of individual autonomy and freedom of choice. At the same time, the way in which social structures control and restrict individual performance and freedom becomes less visible. It is not in every person's or profession's power to choose, in other words, act as his/her conscience directs [ 41 ]. The surgeons in this study seem to have a wide range of freedom to act in a way that is confirmed both from their patients, colleagues, society and by their inner selves. According to the surgeons, their choices of actions and ways of being in relationships with patients and colleagues have to be accepted and confirmed by their inner moral selves. They speak about "having the courage of your own conviction", "taking a stand", "having a personal set of ethical values" as important in order to "live with" the ethical challenges of surgery. The surgeons in this study revealed their identities through their narratives about being in ethically difficult situations. Being a surgeon means having the necessary specialized knowledge, skills, and experience to perform complicated, extensive, and technically advanced operations in principally uncertain and unpredictable situations. This identity is socially sustained through dialogues with patients, relatives, and colleagues. To have an identity as a surgeon is also to understand that one is vulnerable and exposed in these relationships and to know what is at stake in one's own life as well as in the lives of the patients and colleagues. In this study uncertainty, fallibility, and humility seem to be equally important in the surgeons' identities as are their responsibility and confidence. The social importance and benefit of their work to patients and society is confirmed both socially and by their moral selves. This seems to outweigh their personal uncertainty and vulnerability. Thus, surgeons seem to appreciate their work and have learned to live with, and to a certain extent even appreciate, its most difficult dimensions. The meaning of being a surgeon is to be able to live with the tensions of contradictions, ambivalence, dilemmas, and paradoxes in their practice and to find workable solutions in these situations. Competing interests The author(s) declare that they have no competing interests. Authors' contributions KT participated in the design of the study, carried out the interviews, participated in the analysis and completed the manuscript. AN read the interviews, participated in the analysis and helped to draft the manuscript. VS participated in the design of the study, read the interviews, participated in the analysis and helped to draft the manuscript. All authors have read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553983.xml
553982	Docetaxel plus cisplatin is effective for patients with metastatic breast cancer resistant to previous anthracycline treatment: a phase II clinical trial	Background Patients with metastatic breast cancer (MBC) are frequently exposed to high cumulative doses of anthracyclines and are at risk of resistance and cardiotoxicity. This phase II trial evaluated the efficacy and toxicity of docetaxel plus cisplatin, as salvage chemotherapy in patients with MBC resistant to prior anthracyclines. Methods Patients with MBC that had progressed after at least one prior chemotherapy regimen containing anthracyclines received docetaxel 75 mg/m 2 followed by cisplatin 60 mg/m 2 every 3 weeks for a maximum of 6 cycles or until disease progression. Results Between Jan 2000 and May 2002, 24 patients with tumors primary resistant and 15 with secondary resistant disease were accrued. All 39 patients were evaluable for safety and 36 for efficacy. The objective response rate was 31% (95% CI, 16–45%) with 3 complete responses. The median time to disease progression was 7 months, and the median overall survival was 23 months (median follow-up of 41 months). Neutropenia was the most frequently observed severe hematologic toxicity (39% of patients), whereas asthenia and nausea were the most common non-hematologic toxicities. No treatment-related death was observed. Conclusion In conclusion, we found docetaxel plus cisplatin to be an active and safe chemotherapy regimen for patients with MBC resistant to anthracyclines.	Background In the management of breast cancer, anthracycline-based chemotherapy regimens remain standard adjuvant or first-line palliative treatment. Furthermore, some patients cannot be treated with anthracyclines due to impaired cardiac function. It is thus important to identify active, well-tolerated, not anthracycline cross-resistant, salvage regimens [ 1 ]. Taxanes (docetaxel and paclitaxel) are currently the most extensively studied new chemotherapeutic agents for metastatic breast cancer (MBC). Single-agent docetaxel has demonstrated significant survival advantages over other recognized regimens in 2 large randomized trials in patients with anthracycline-pretreated MBC [ 2 , 3 ]. Phase II data suggest that docetaxel is the most active agent yet available for the treatment of MBC [ 4 ]. Docetaxel also has some activity in paclitaxel-resistant MBC [ 5 ]. Cisplatin monotherapy has shown response rate of 9% in salvage settings, and 50% as first-line therapy [ 6 ]. Because docetaxel and cisplatin are both active and have different mechanisms of action, this combination may provide additive effect against MBC. Although docetaxel and cisplatin were not actively synergistic in preclinical studies [ 7 , 8 ], this combination chemotherapy has been widely used for treatment of a variety of tumor types [ 9 - 13 ]. Prior phase I studies showed the feasibility of this combination and its activity on different tumors [ 14 ]. Considering the single-agent activity of both drugs, their different mechanism of action and distinct toxicity profile, we designed this phase II study, in which the combination was evaluated as a salvage therapy in patients with anthracycline-resistant MBC. Methods Eligibility Eligible patients had measurable or assessable histologically confirmed breast cancer that had progressed after at least one prior chemotherapy regimen containing anthracyclines. All patients had Eastern Cooperative Oncology Group performance status of 2 or lower and adequate bone marrow, hepatic, and renal function, defined as white blood cells ≥4000/μL, absolute neutrophil count ≥1500/μL, platelets ≥100,000/μL, total bilirubin ≤2.0 mg/dL, transaminases ≤3 times the upper normal limit, and serum creatinine ≤1.5 mg/dL. Concurrent radiation or hormonal therapy was not allowed; however, patients with clinically stable metastases of the brain or other sites who had completed radiation therapy were permitted. Patients were eligible regardless of the nature of prior therapy, including high-dose therapy with stem cell support and prior exposure to paclitaxel. Any prior antitumor treatment had been completed at least 1 month before entering this study. Patients were excluded if they had severe comorbid illness, symptomatic peripheral neuropathy of any origin, or a history of anaphylaxis of any type. All patients were provided a thorough explanation of the study, and they signed informed consent prior to enrollment into the study. Definition of anthracycline-resistance Patients were classified as primary and secondary anthracycline-resistant. Primary anthracycline-resistance was defined as relapse during or within 12 months after anthracycline-based adjuvant therapy, or disease progression with no intervening response during anthracycline-containing chemotherapy for MBC. Secondary resistance was defined as relapse more than 12 months after receiving adjuvant anthracycline chemotherapy, or disease progression at some time after a documented clinical response to anthracycline-based chemotherapy for MBC. Pretreatment and follow-up evaluation Pretreatment evaluation included a complete medical history and physical examination, a computed tomographic (CT) scan of the chest and abdomen, magnetic resonance imaging (MRI) of the brain, and a bone scintigraphy to assess the extent of disease. Follow-up consisted of physical examination, monitoring of toxic effects, a complete blood count, liver function tests, chest radiography, and CT scan as clinically indicated. Tumor response and toxicity were recorded in accordance with the World Health Organization (WHO) criteria. Responses were assessed every 2 cycles of chemotherapy, and patients were evaluated before each new treatment cycle for toxicities. Treatment plan The treatment consisted of docetaxel (Taxotere ® , Aventis, Bridgewater, NJ) 75 mg/m 2 given by a 1-hour intravenous infusion immediately followed by cisplatin 60 mg/m 2 in a 2-hour infusion. Cycles were repeated every 3 weeks if the patient's blood count had returned to normal and non-hematologic toxicities had resolved. Dosage of subsequent cycles was adjusted according to the observed toxicities that developed during the preceding cycle. The treatment continued for a maximum of 6 cycles or until disease progression. All patients received standard supportive care regimen consisting of adequate hydration, dexamethasone, and antiemetic therapy according to the guidelines of the American Society of Clinical Oncology [ 15 ]. Statistics The primary end points of the trial were the efficacy of the therapy, which were measured as objective response rate and time to progression, with a secondary end point of toxicity. All patients who received at least 2 cycles of treatment or who progressed after the first cycle were considered assessable for response. The study was conducted using a single-stage, phase II trial design [ 16 ]. Based on the results from previous phase II trials reporting response rates of about 30% in patients with anthracycline-resistant MBC, we selected a 50% target response probability and a 30% null response probability, with type I and type II error set at 10%. The number of patients required was 39. Statistical calculations were performed using SPSS software, version 11.5 (SPSS, Inc, Chicago, IL). Comparisons were performed using chi-square test and survival curves were generated using Kaplan-Meier method. Results were considered significant at the P = .05. Results Patients characteristics Between Jan 2000 and May 2002, 24 patients with tumors primary resistant and 15 with secondary resistant disease were accrued. Among 39 patients assessable for safety, 3 were not evaluable for response due to early discontinuation of treatment. Baseline characteristics of the eligible patients are listed in Table 1 . Median age was 51 years (range, 41–64) and median performance status was 1. Thirty-three patients (85%) had visceral metastases as the dominant site of disease. All patients had received prior anthracycline-based chemotherapy with a median number of prior chemotherapeutic regimens of 2 (range, 1–3). Three patients (8%) had exposed to paclitaxel in their previous chemotherapy regimen. Efficacy According to the "intent-to-treat" analysis, the objective response rate was 31% (95% confidence interval [CI], 16–45%). There were 3 complete responses, 12 partial responses and 15 patients had stable disease. The three patients who obtained a complete response had non-visceral metastatic pattern (neck lymph nodes and/or soft tissue lesions). Most patients who achieved a response did so at the end of the third cycle (median time to response, 2.2 months; 95% CI, 2.1–2.4 months). At a median follow-up of 40.5 months (95% CI, 36.4–44.6 months), the median time to progression was 6.6 months (95% CI, 3.9–9.4 months) and the median overall survival was 22.9 months (95% CI, 17.2–28.5 months). The time to progression in responders were 16.7 months (P < 0.0001). The Kaplan-Meier method was used to estimate overall survival and time to progression, as shown in Figure 1 . At the time of present analyses, 21 patients (54%) died. We observed no significant differences in overall response rate or survival between primary and secondary anthracycline-resistant groups. The overall survival and time to progression were higher in patients with objective response (P < 0.0001 and P = 0.05, respectively). Toxicity Patients received a median of 6 cycles (range, 1–6) of docetaxel plus cisplatin. Relative dose intensities were 84% (95% CI, 73–96%) and 86% (95% CI, 75–97%) for docetaxel and cisplatin, respectively. Twenty-four patients (62%) completed 6 cycles of chemotherapy as planned; Nine patients had progression of disease while receiving treatment and 6 patients discontinued treatment because of toxicity. Treatment delay was occurred in 54 cycles (29%). Toxicity was evaluable in all 39 patients and 189 chemotherapy cycles. Even if the most frequent hematological toxicity was neutropenia (grade 3/4 in 22% of treatment cycles), only 12 episodes of non-fatal febrile neutropenia were observed. The most commonly encountered non-hematologic toxicities were asthenia and nausea. Other toxicities are reported per patient and per cycle in Table 2 . Six patients had severe toxicity precluding further treatment. No treatment-related deaths were observed. Discussion Single-agent docetaxel has been considered a standard treatment for patients with anthracycline-pretreated MBC for several years. Docetaxel monotherapy often demonstrated response rates of 50% or more in this setting [ 17 , 18 ]. Ahn et al . also reported, in a similar patient population, 36% response rate and 69 weeks median survival with docetaxel 75 mg/m 2 every 3 weeks[ 19 ]. Platinum-based combinations also have significant activity in previously treated as well as previously untreated patients. In several small clinical trials testing the cisplatin-based regimens, response rates of 50%-83% were reported [ 20 ]. The North Central Cancer Treatment Group (NCCTG) evaluated carboplatin combined with paclitaxel as first-line chemotherapy for MBC [ 21 ]. The overall response rate was 62% and the median time to progression was 7.3 months. Clearly, platinum-based combinations are very active against MBC, but the relative degree of toxicity compared with other agents limited their use in routine clinical practice. Recently, synergism between platinum and trastuzumab, a novel monoclonal antibody directed against the protein product of the HER2/neu oncogene, awaked interest in the use of cisplatin for breast cancer [ 6 , 22 ]. There are relatively few completed clinical studies involving platinum-based combination chemotherapy for anthracycline-resistant MBC. Spielmann et al . evaluated docetaxel 75 mg/m 2 plus cisplatin 80 mg/m 2 every 3 weeks in 38 patients with anthracycline-resistant MBC, giving an objective response rate of 36% [ 23 ]. Japanese investigators performed a phase II study of docetaxel 60 mg/m 2 and cisplatin 80 mg/m 2 in patients with anthracycline-pretreated MBC [ 24 ]. They reported an overall response rate of 64%. In a phase II study of docetaxel and carboplatin, overall response rate of 61% was achieved in patients with chemotherapy-pretreated MBC [ 25 ]. Gelmon et al . combined biweekly paclitaxel with cisplatin and achieved a response rate of 85% with few septic events [ 26 ]. We classified "anthracycline-resistant" into primary and secondary. Ando et al . suggested that the status of anthracycline-resistance is important for the prediction of response to second-line treatment with docetaxel [ 27 ]. However, we could not observe significant difference in the efficacy of docetaxel plus cisplatin between primary and secondary resistant groups. The different outcomes might be due to the current situation that there have been no standard criteria defining anthracycline-resistance and often intermingled with "anthracycline-pretreated" or "anthracycline-refractory." This study demonstrated that a combination of docetaxel plus cisplatin in a 3-week cycle was an effective and well-tolerated regimen for patients with anthracycline-resistant MBC. In this study, the use of docetaxel plus cisplatin resulted in an overall response rate of 31% and the median time to progression was 7 months. Significant improvements in the actuarial survival rate and time to progression were observed in the group of patients who achieved objective responses. The main toxicities, gastrointestinal, hematological and asthenia, were manageable with dose adjustment and supportive care. Dose intensities of more than 80% were delivered and 61% of patients completed 6 cycles of chemotherapy, which is considered acceptable and expected. We used the planned dose of cisplatin 60 mg/m 2 every 3 weeks, rather than 75 mg/m 2 cited in a phase I study [ 14 ], because we believed that tolerability of treatment is indispensable in the salvage setting in the management of solid tumors. With the increasing use of anthracycline-based chemotherapy as adjuvant treatment, as well as first-line chemotherapy against MBC, patients are frequently exposed to high cumulative doses of anthracyclines and are therefore at risk of resistance and cardiotoxicity [ 28 ]. This combination of docetaxel with cisplatin may be particularly useful in patients previously treated with anthracyclines (but naïve to either docetaxel or cisplatin). In addition, for patients who have already had cardiac failure and have not received chemotherapy with taxanes as an adjuvant or as first-line treatment, use of docetaxel plus cisplatin is considered a better option. More recently, reports that trastuzumab has a powerful synergistic interaction with docetaxel and with cisplatin [ 29 ], have prompted evaluation of the combination of trastuzumab with docetaxel and/or platinum in the treatment of MBC [ 30 , 31 ]. Conclusion In summary, the combination of docetaxel plus cisplatin is active and safe in patients with anthracycline-resistant MBC. The activity observed in anthracycline-resistant and heavily-pretreated patients suggests relative non-cross-resistance with other drug combinations. Therefore, we hope that this study could result in a prospective trial to determine whether this activity translates into actual improvement in survival and quality of life in patients with anthracycline-resistant MBC. Competing interests The author(s) declare that they have no competing interests. Authors' contributions SHP collected the data, performed the statistical analysis and drafted the manuscript. EKC, SB, JHL, YDL followed the patients. DBS designed the study, followed the patients and helped with the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553982.xml
549065	Identification of the prokaryotic ligand-gated ion channels and their implications for the mechanisms and origins of animal Cys-loop ion channels	Acetylcholine receptor type ligand-gated ion channels are well known in animals. Homologs are identified in prokaryotes that may act as chemotactic receptors.	Background The flux of ions across excitable cellular membranes is a signaling mechanism that is extensively utilized by organisms from all the three major superkingdoms of life. This directional flow of ions across cellular membranes is mediated by a wide range of ion channels that may be gated by a variety of signals, such as voltage, mechanical forces or chemical first messengers [ 1 ]. Ion-dependent signaling is particularly critical for the functions of the animal nervous system, where propagation of signals along neuronal processes and the transmission of signals from neurons or receptor cells to their targets is mediated by the action of ion channels. The neuronal ligand- or neurotransmitter-gated ion channels (LGICs) combine the functionalities of a receptor and ion channel in a single protein, and mediate fast synaptic signaling [ 1 ]. The neurotransmitter released by the presynaptic cell, within a few microseconds binds to the extracellular ligand-binding module of the ion channel and causes the channel to open. This results in a selective flow of ions down their electrochemical gradients through the water-filled pore of the channel, and the excitation or inhibition of the train of action potentials in the postsynaptic cells. Furthermore, within a few milliseconds the neurotransmitter dissociates from the receptor and thereby terminates the synaptic signal. Thus, the LGICs act as molecular switches to provide a specific impulse of ion flux in response to a neuronal signal [ 1 ]. One of the most prominent superfamilies of the animal LGICs has as its prototype the acetylcholine-gated channels and includes the receptors for a variety of neurotransmitters in both vertebrates and invertebrates ([ 2 ], also see [ 3 ]). The known endogenous ligands bound by these receptors are acetylcholine, GABA, serotonin, glycine, histidine, glutamate and cationic zinc [ 4 - 8 ]. The receptors are also the targets of plant toxins such as nicotine and strychnine, conotoxins of snails, lophotoxins of corals, and many of the neurotoxins of elapid snakes [ 4 - 6 ]. This superfamily is commonly referred to as the Cys-loop superfamily (named after a conserved cystine bridge seen in the animal representatives of this superfamily) or the acetylcholine-receptor-type LGIC superfamily (ART-LGIC). All the known members of this superfamily possess stereotypic domain architectures, with an all-β amino-terminal ligand-binding domain (LBD) and a carboxy-terminal transmembrane domain comprised of four membrane-spanning helices (4-TM). The members of this superfamily exhibit a pentameric quaternary structure, with the second transmembrane helix from each monomer (helix M2) contributing to the wall of a transmembrane pore through which the ion passes. The animal ART-LGICs may exist as heteropentamers, containing up to four distinct paralogous monomers. The ligand is bound at the dimer interface of two adjacent LBDs, and residues from both subunits form a box-like cavity to accommodate the ligand [ 9 , 10 ]. In the case of most animal neurotransmitter receptors in their open state, only two (or occasionally three) of the five subunit junctions in the pentameric receptor are occupied by the ligand [ 4 - 6 ]. The ART-LGICs characterized to date show ion selectivity. The excitatory channels, such as the acetylcholine and serotonin receptors, the mammalian Zn receptors and some invertebrate GABA receptors, allow the flow of cations, whereas the inhibitory receptors, such as those for glycine and GABA, invertebrate glutamate and histamine receptors, and some invertebrate serotonin receptors (such as Caenorhabditis elegans MOD-1), allow the flow of anions. Cation or anion selectivity of the channel is principally governed by the charge distribution in the linker between the transmembrane helices M1 and M2 [ 11 , 12 ]. Several recent studies based on the X-ray structure of the recombinant homopentamer of the soluble acetylcholine-binding domain (ACHB) from the snail Lymnaea stagnalis [ 9 ] and the electron microscopic structure of the transmembrane domain [ 13 ] have thrown light on the possible mechanisms of ligand interaction and channel gating of the ART-LGICs. The current model for the mechanism of these channels posits that the binding of the ligand causes a preferential rotation of one of the β sheets of the LBD. The resultant conformational change is believed to be transmitted via interactions with the loop between helices M2-M3 to the hydrophobic constriction in the middle of the M2 helices that line the channel walls [ 13 ]. This causes a relaxation of the middle of the girdle and allows the flow of the ions. Despite intense studies, there remain several unresolved issues with respect to the mechanism by which the binding of the ligand to a segregated site transmits the conformational change to the rest of the LBDs to trigger the rotation. Furthermore, the extent of the applicability of the conclusions drawn from the acetylcholine receptor model for other members of the superfamily remains somewhat unclear. Thus far, the ART-LGIC superfamily is known only from multicellular animals (metazoans). Phylogenetic analysis suggests that the common ancestor of the bilateral animals already possessed multiple members belonging to two major families of the superfamily that correspond, respectively, to the excitatory cationic channels, including the acetylcholine and serotonin receptors, and the inhibitory anionic channels, including the GABA, glycine and invertebrate histamine and glutamate receptors [ 2 , 14 , 15 ]. This restricted phyletic pattern is in contrast with what has been previously observed for the voltage-gated potassium channels of the Shaker-type superfamily and the voltage-gated sodium channels. In both these cases, several representatives are known from both non-animal eukaryotes, as well as numerous prokaryotes, suggesting that they were employed in signaling in other contexts well before the origin of the animal nervous system [ 16 - 19 ]. This prompted us to investigate if distant representatives of the Cys-loop/ART-LGIC superfamily could be detected in organisms outside the animal lineage. We also sought to use these distant relatives in comparative sequence-structure and genomics studies to understand the most general functional and mechanistic features that typify this superfamily. We report here the identification of several prokaryotic members of the ART-LGIC superfamily and discuss the general implications of these proteins for the mechanisms and origin of the Cys-loop receptors of the animal nervous system. Results and discussion Identification of prokaryotic versions of the ART-LGIC superfamily To investigate the origins of the animal ART-LGIC superfamily, we tried to obtain a complete picture of their phyletic spread in all organisms with currently available genomic sequence information. All bona fide animal members of this superfamily (with the exception of snail ACHB) contain a globular, extracellular, amino-terminal LBD and a carboxy-terminal 4-TM domain. The membrane-spanning helices, being compositionally biased, tend to frequently recover false positives in iterative sequence profile searches. Accordingly, we only used the globular extracellular domains of the known ART-LGIC receptors, which are typically around 200-220 amino acids in length, for our iterative sequence profile searches with the PSI-BLAST program. Iterative searches from a number of starting queries, such as the human acetylcholine receptor α7 chain (gi: 2144875; region 24-230), C. elegans MOD-1 receptor (gi: 25154135; region 32-238) or the human GABA receptor α4 chain (gi: 1346079; 46-256) recovered a consistent set of receptors from diverse animals with significant expect (e)-values prior to convergence (run with inclusion threshold of 0.01). Interestingly, in addition to the animal sequences these searches also recovered sequences from different bacteria. For example a search initiated with the above-mentioned acetylcholine receptors as the seed recovered Gloeobacter violaceus , Crocosphaera watsonii (both cyanobacteria) in iteration 3 (e-values = 10 -5 -10 -7 ) and Rhodopseudomonas palustris (α-proteobacteria) in iteration 6 (e-value = 10 -4 ). However, no significant hits belonging to any of the other eukaryotic lineages, such as the fungi, Dictyostelium , plants, alveolates or Giardia were detectable. To further investigate the occurrence of ART-LGIC homologs in bacteria, we constructed a PSI-BLAST profile of the LBDs recovered in the above searches and used it to systematically search all the bacterial genomes, which are available as whole-genome shotgun reads or as completely assembled chromosomes. As a result of these searches we recovered statistically significant hits to the ART-LGIC LBDs from several other phylogenetically diverse bacteria including Cytophaga hutchinsonii , α-proteobacteria like Bradyrhizobium japonicum and Magnetospirillum magnetotaticum , γ-proteobacteria, like Erwinia chrysanthemi , Microbulbifer degradans and Methylococcus capsulatus , several cyanobacteria and a single archaeal genus Methanosarcina . All these bacterial hits corresponded to the full length of the animal LBDs, which were used as seeds to build the sequence profiles. When signal peptides and the transmembrane helices were predicted for the bacterial proteins, all of them showed a general structure similar to the animal receptors; that is, an amino-terminal signal peptide and a LBD followed by a carboxy-terminal 4-TM domain. However, some of the bacterial proteins showed additional domains between the amino-terminal signal peptide and the ART-LGIC superfamily ligand-binding and channel domains (see below for further discussion). Reciprocal searches with either just the region corresponding to the LBD or the whole unit comprising both the LBD and the following 4-TM domain of the bacterial proteins recovered the animal Cys-loop proteins with significant e-values (0.001-10 -17 in iterations 1-3). For example, a search with the sequence of Chut0841 (gi: 23135736) from C. hutchinsonii recovered the animal receptors with e-values in the range 10 -4 -10 -6 in the second iteration. The secondary structure was predicted for the region corresponding to the LBD in the bacterial proteins using the programs PHD [ 20 ] and JPRED2, using the combined information from the multiple alignment, a PSI-BLAST position-specific score matrix and a hidden Markov model derived from the alignment [ 21 ]. The predicted secondary structure of the bacterial proteins precisely corresponded to the secondary structure of the conserved core of the animal LBDs typified by the ACHB (PDB:1UV6), with an amino-terminal helix followed by nine β strands, which form a β sandwich [ 9 ]. Taken together, the above observations suggested that the bacterial proteins were bona fide homologs of the animal neurotransmitter receptors of the ART-LGIC/Cys-loop superfamily. Mechanistic and functional implications of the comparative sequence-structure analysis of the bacterial and animal ART-LGIC receptors To obtain information regarding the potential functional and structural similarities and differences of the predicted bacterial ART-LGIC and the animal receptors we prepared a multiple alignment of the bacterial sequences with the representatives of all the major classes of animal Cys-loop proteins (Figure 1 ) using the T_Coffee program [ 22 ]. The alignment was further refined on the basis of secondary structure predictions and comparisons with the available structure of the stand-alone animal LBD, ACHB. The multiple alignment shows that the majority of the highly conserved positions in the LBD are in the conserved strands, and when mapped onto the structure of ACHB, they correspond to the positions stabilizing the hydrophobic core of the β-sandwich (Figure 1 , see also Additional data file 1). This observation strongly suggests that the bacterial versions would adopt a tertiary structure similar to the animal LBDs. The bacterial LBDs differ notably from the animal LBDs, however, in lacking the characteristic cysteine residues which form the disulfide bridge in practically all known animal receptor subunits (Figure 1 ). However, in place of the second cysteine the bacterial sequences possess a highly conserved hydrophobic position that is likely to be buried in the hydrophobic core of the sandwich and, thereby, similarly stabilize the region corresponding to the Cys-loop of the animal sequences (Figure 1 ). This absence of the cysteines in the bacterial versions of these family is reminiscent of what was previously observed in the bacterial homologs of several animal extracellular protein domains, such as the SCP1/PR1 domain, the immunoglobulin domains and the MAC-perforin domains [ 23 - 25 ]. Eukaryotic cells typically possess an extensive secretory compartment, with a strongly oxidizing environment, in the form of the endoplasmic reticulum, through which a protein passes before secretion [ 26 ]. In contrast, in bacteria most disulfide bond formation occurs after extrusion to the periplasmic compartment [ 27 ]. The presence of this extensive secretory compartment in eukaryotic cells might have allowed a greater role for stabilization through disulfide bonds, and thereby favored the emergence of interacting cysteines in eukaryotic versions of domains as opposed to the bacterial counterparts. Over and above the general conservation of hydrophobic residues in the 4-TM domain, there are some potential functionally relevant conserved positions shared by the bacterial and metazoan proteins. One of these is the helix-bending position in helix M1 (usually occupied by a proline (P), glycine (G) or serine (S), and corresponding to P221 in Torpedo californica ACHR α-chain), which is predicted to be critical for the flexibility of the structure to conformational change [ 13 , 28 ]. Another position of interest is in the middle of helix M2, and is occupied by a small residue (corresponding to S252 in T. californica ACHR α-chain) that initiates a bend in the helix resulting in the hydrophobic constriction or girdle that forms the channel gate [ 13 ]. Glycine 275 of T. californica ACHR α-chain, in the loop between helices M2 and M3, has been implicated as one the residues that may be critical for the rotational freedom of the ACHR M2 helix during the gating process [ 13 ]. The strong conservation of a small residue at this position in both the bacterial and animal members of this family suggests that it is likely to support this function throughout the superfamily. Less obvious is the role of a polar residue just before the start of helix M4 that is highly conserved across both bacterial and animal members of this superfamily. From its position in the structure, it is possible that interactions of residue with solvent water might play a role in stabilizing one of the conformational states. One of the major determinants of ion selectivity is the sequence just amino-terminal to the helix M2 on the cytoplasmic side. The cation channels usually have a sequence motif of the form glutamate (E)-[arginine (R)/lysine (K)] with the glutamate playing a role in cation selection. The anion channels usually have a motif of the form alanine (A)-[RK] with the basic residue participating in anion selection [ 11 , 12 , 29 ]. A glutamate corresponding to that of the cation channels is seen in about eight of the bacterial sequences and a basic residue similar to the anion channels is seen in six of the bacterial sequences, suggesting that both selectivities are likely to be encountered in the bacterial sequences (Figure 1 ). In addition, like the animal sequences, the bacterial sequences contain poorly conserved polar or charged residues at the carboxyl terminus of the M2 helix, which might play a role in fine-tuning their selectivity [ 4 , 6 , 11 , 13 ]. The long hydrophilic linker between helices M3 and M4 is highly variable in length and sequence in the animal proteins. It has been implicated in cytoplasmic interactions with functional partners such as the P2X family of ATP receptors [ 30 ] and the cytoskeletal receptor-clustering protein gephyrin [ 31 ]. In contrast to the animal members of the superfamily, all bacterial versions possess an abbreviated cytoplasmic M3-M4 loop and are unlikely to have functional interactions that are seen in the animal versions. The ligand-binding box in ACHB has been termed the aromatic box as it is bounded by multiple aromatic residues (Figures 1 , 2 ). In several metazoan receptors the positively charged group on the ligands has been suggested to form cation-π interaction with the π-orbitals of different aromatic residues in the binding box [ 32 - 34 ]. An examination of the ACHB structure [ 9 ] revealed that the side chains of eight residues almost completely envelop the ligand, and are the principal constituents of the ligand-binding box (Figure 2 ). Of these, the dyad of two consecutive cysteines, which are amino-terminal to the final strand of the LBD is observed only in a subset of the animal cation channels, and does not represent a conserved interaction position. Of the remaining six positions, two are from one of the subunits while the remaining four are from the other subunit (Table 1 ). The average number of aromatic residues in these positions in the bacterial proteins is 2.1, whereas in the animal sequences it is 2.6. Every sequence in our representative set, animal or bacterial, with the exception of the human Zn receptor [ 8 ], contains at least a single aromatic residue in one of these positions. This suggests that aromatic residues are critical for ligand interaction throughout this superfamily, though the exact position in the ligand-binding box that is occupied by an aromatic residue does not seem to be preserved. However, the smaller number of aromatic residues in the ligand-binding box of bacteria may indicate some differences in the type of ligand and the nature of the interactions. Furthermore, an interesting difference is noted in the aromaticity of the positions corresponding to leucine (L) 112 (subunit D) and tryptophan (W) 143 (subunit C) of the ACHB structure between the bacterial and animal sequences (Figure 2 ). The ratio of aromatic residues at these positions is anti-correlated, and this anti-correlation is strongly preserved in the individual sequences. This suggests that these two positions might represent mutually exclusive, but functionally equivalent, surfaces for ligand interaction. The presence of at least one aromatic residue in most of the predicted ligand-binding pockets could imply that cation-π interactions with the bound ligand are widespread in the entire superfamily. However, other explanations are also possible. For example, one or more aromatic residues could have a possible structural role in constraining the pocket to favor a particular ligand or ligand orientation. Alternatively, they could provide the requisite hydrophobic environment in the pocket or interact with the ligand through aromatic stacking. In addition to the residues discussed above, there are several other conserved residues in the LBD that may have a role in transmission of conformational changes. Among the most highly conserved features is the aPaD signature (where 'a' is any aromatic residue, and D is aspartate) in the middle of the region corresponding to the Cys-loop (Figure 1 ) and these residues are essential for wild-type receptor function [ 5 ]. They lie far away from the ligand-binding region and close to another nearly universally conserved basic residue at the end of the terminal strand of the LBD (Figure 1 ). This basic residue is known to be mutated in the glycine receptor α1 subunit in the human genetic disease sporadic hyperekplexia [ 35 ]. The aspartate from the aPaD motif and the basic residue could potentially form a salt bridge to stabilize the 'outer sheet' of the β sandwich and thereby regulate the preferential movement of the sheets after ligand binding. This proposal is consistent with recent studies that implicate some of these charged residues, especially the aspartate in the Cys-loop, in coupling ligand binding to further conformational changes leading to channel gating [ 36 , 37 ]. The other highly conserved positions are a tryptophan at the end of strand 2 (W58 in ACHB) and an aromatic or hydrophobic position (W82 in ACHB) that are in hydrophobic interaction with each other (Figures 1 , 2 ). These residues are at the center of a set of fairly conserved positions (including D61, P84, D108, G109 and isoleucine (I) 150 in ACHB) in both bacterial and eukaryotic proteins that form a chain on either side from the ligand-binding box to the surface of the 'inner sheet' at the top of the LBD [ 9 , 28 ]. It is likely that these residues form a conduit for the propagation of the conformational change from the ligand-binding box to the inner sheet (Figure 2 ). The conservation of certain key features in both the LBD and the 4-TM domains of the bacterial and eukaryotic receptors suggests that despite their extensive sequence divergence they are likely to share general functional and mechanistic properties. In the pentamer these residues appear to form a continuous ring passing through the top surface of the LBD, and undergo conformational changes in relation the presence of a bound ligand (Figure 2 ) [ 9 ]. Functional significance of domain architectures and gene neighborhoods of bacterial ART-LGICs We analyzed the domain architectures and gene neighborhoods of the predicted bacterial ART-LGICs to glean further insights regarding their biological functions. Unlike the animal ART-LGICs, the bacterial receptors show a greater diversity in their domain architectures, while preserving the core module which comprises the extracellular LBD and 4-TM channel-forming domain (Figure 3 ). The representatives from cyanobacteria, Rhodopseudomonas and one of the three versions from C. hutchinsonii show a simple architecture identical to the animal forms. Some versions, like those from the α-proteobacteria, M. magnetotacticum and B. japonicum , show a further amino-terminal fusion to a domain of the periplasmic binding protein type I (PBP-I) superfamily (Figure 3 ). The archetypal domains of the PBP-I superfamily are the bacterial proteins such as the lysine/arginine/ornithine-binding protein, that bind amino acids and other small molecules in the extracellular or periplasmic space and facilitate their uptake by ABC-family transporters [ 38 ]. Interestingly, PBP-I domains also form the LBDs of two distinct superfamilies of animal neuronal receptors. The NMDA-type receptors, which comprise a class of ligand-gated channels distinct from the ART-LGIC/Cys-loop superfamily, contain an amino-terminal PBP-I domain and a carboxy-terminal domain belonging to the second major superfamily of bacterial periplasmic binding proteins (the PBP-II superfamily, for example, HisJ) [ 39 , 40 ]. The channel-forming transmembrane domain in these proteins is inserted into the carboxy-terminal PBP-II domain. The metabotropic glutamate receptor and vertebrate taste receptors, which are G-protein-coupled receptors, contain a PBP-I domain amino-terminal to their 7-TM domains [ 39 , 40 ]. A third architectural theme in the bacterial ART-LGICs is a fusion of two additional amino-terminal domains to the core receptor module, namely the MCP-N (methyl-accepting chemotaxis protein-N domain) and Cache domains [ 41 ]. This version is seen in a number of phylogenetically distant prokaryotes, such as the archaeon Methanosarcina and the bacteria Cytophaga and Microbulbifer (Figure 3 ). The MCP-N and Cache domains are prevalent prokaryotic sensor domains that bind a variety of extracellular or periplasmic ligands and regulate signal transduction via a variety of carboxy-terminal signaling domains. In an interesting parallel to the PBP-I/II domains, the MCP-N and Cache domains are found in a regulatory subunit (α2-δ) of the animal voltage-gated calcium channels, and appear to comprise the binding site for the drug GABApentin and possibly an as-yet unknown endogenous ligand [ 41 ]. Thus, these architectures suggest that many of the predicted bacterial receptors might possess multiple ligand-interaction domains and that an interplay of allosteric effects could regulate their function. Remarkably, the additional domains found with the bacterial ART-LGIC proteins are also encountered in animal neuronal receptors, suggesting that all these domains belong to a common network of ancient sensory modules that have been utilized in diverse contexts [ 42 ]. Contextual information in the form of conserved gene neighborhoods or predicted operons in prokaryotes often provides hints to identify gene products that functionally or physically interact or belong to the same pathways or signaling cascades [ 43 , 44 ]. Accordingly, we examined the gene neighborhoods of all the predicted bacterial ART-LGICs to identify conserved neighborhoods or persistent patterns of genomic clustering of genes with similar functions. In some bacteria, the gene for the ART-LGIC was found in a conserved gene neighborhood along with a gene for a stand-alone version of the PBP-II superfamily (Figure 3 ). This is analogous to the above-noted fusion of the PBP-I domain to the ART-LGIC in other bacteria, and suggests that these stand-alone PBP-II domains probably functionally cooperate with the receptors. In one bacterium, namely Rhodopseudomonas , there is a similar predicted operon, but instead of a gene for a PBP-II superfamily protein, there is one for a stand-alone Cache domain. This situation parallels the fusion with the Cache domain in some of the receptor versions and these two independent proteins may similarly cooperate functionally. Taken together, these observations suggest that bacterial ART-LGICs may function as chemotaxis receptors. As most bacterial genomes in which they are present contain only a single member of the ART-LGIC superfamily, it is likely that, in contrast to many of the well studied metazoan receptors, they function as homopentamers. The PBP-I, PBP-II MCP-N and Cache domains that are either fused or operonic with many of the predicted bacterial receptors may help in a preliminary concentration or sensing of amino acids or other small-molecule ligands. These ligands may then bind to the channel's LBD domain and activate an ionic flux across the cell membrane that in turn regulates the motility of the bacterium in response to the ligand. This proposal is analogous to the recently reported activity of a voltage-gated Na + channel in the bacterium Bacillus pseudofirmus in chemotaxis, motility and the regulation of the Na + -cycle [ 16 ]. Interestingly, in at least one bacterium, Microbulbifer degradans , the ART-LGIC with a predicted cation selectivity is in a predicted operon with a Na + /H + symporter, suggesting possible interactions with the Na + cycle. Phyletic patterns and phylogenetic relationships of the bacterial and eukaryotic ART-LGICs Comparative genomics of ART-LGICs suggests that they show a highly non-uniform phyletic patterns. Among the eukaryotes they are only seen in animals, and could not be detected in the currently available genomic sequences of other crown-group eukaryotes such as plants, fungi, Dictyostelium , Entamoeba , apicomplexans or earlier-branching eukaryotic taxa such as Giardia and Trichomonas . Among the prokaryotes, too, they show a highly sporadic distribution: very distantly related taxa may possess similar receptors (for example, Cytophaga and the archaeon Methanosarcina , Figure 3 ), whereas closely related taxa may differ from each other in possessing or lacking a predicted ART-LGIC. These phyletic patterns are similar to those observed for several signaling receptors in prokaryotes and are suggestive of a high degree of mobility through lateral transfer, and frequent gene loss [ 45 ]. We constructed phylogenetic trees of the ART-LGICs by using an alignment that spanned the entire length of the LBD and the 4-TM channel domain, and included all bacterial members and representatives of all the major animal receptor groups. The trees constructed using several different methods - maximum likelihood, Bayesian inference, minimum evolution and neighbor-joining - produced congruent tree topologies (Figure 3 ). As expected, the tree showed a strongly supported monophyletic animal branch that in turn split up into the two major families corresponding to the great split between the classical acetylcholine-serotonin type (usually cationic) receptors and their relatives and the classical glycine-GABA type (usually anionic) receptors and their relatives [ 2 , 7 , 14 , 15 ]. All the animal sequences are much closer to each other to the exclusion of all other prokaryotic sequences (Figure 3 ). They possess several unique sequence and structure features, including the characteristic cysteines of the Cys-loop and the extra large variable region between the transmembrane helices M3 and M4. Its absence in the bacterial forms suggests that they are 'simpler' versions, which are closer to the primitive state. The mean intra-group distance of the metazoan versions, measured using the JTT substitution matrix on an alignment of 368 positions, is 1.7. This value is much lower than the intra-group distance of 3.01 that is observed for the bacterial forms (the overall mean distance being 2.8). The prokaryotic proteins also show greater diversity of architectures in comparison to the stereotypic architecture of all the animal members of this superfamily. These observations suggest that the diversification of the prokaryotic forms preceded the emergence of the eukaryotic forms and thus that the root of the tree is more likely to lie in the bacterial lineage than within the metazoan lineage. Certain bacterial versions (those from Crocosphaera , Gloeobacter , Erwinia and Rhodopseudomonas ) are markedly more similar in sequence to the eukaryotic forms (Figures 1 , 3 ). Specifically, these similarities include the extension of strand 2 of the LBD, before the universally conserved W, and the hWxP motif (where h is a hydrophobic residue and x any residue) amino-terminal to strand 4 of the LBD. Constrained trees, where the animal branch was artificially grouped with the more distantly related bacterial sequences, were significantly worse (using the Kishino-Hasegawa and Bayesian posterior probability tests; data not shown) than the trees in which they were grouped with their closer bacterial homologs. This observation, taken together with the greater likelihood of the root being amongst the prokaryotes, suggests that the above features shared by some of the bacterial sequences and the animal versions are synapomorphies or derived characters. Taken together, the phyletic patterns and the specific relationship of the animal sequences to a subset of the bacterial forms suggests that the common ancestor of the animal ART-LGICs probably arose via an early lateral gene transfer from a bacterium to the ancestral lineage leading to the modern metazoans. Following this transfer, the ancestral eukaryotic version acquired the characteristic cysteines of the Cys-loop and duplicated and diverged to give rise to the two major metazoan Cys-loop families. By the time of the common ancestor of the bilateral animals the two major families appear to have diversified into about nine distinct lineages (Figure 3 ). The biased sampling of eukaryotic genome sequences and the high frequencies of gene loss in the eukaryotes could imply that the transfer of the ART-LGICs from bacteria to the eukaryotes might have occurred well before the emergence of the animal lineage, and has been lost repeatedly in the other eukaryotes. While this possibility cannot be ruled out until more representative eukaryotic sequences become available, it is likely that there was a single precursor for all the animal sequences, which was acquired at some point from a bacterial source, and the massive radiation of the Cys-loop receptors occurred only after the animals branched off from the rest of the crown group. In principle it is possible that the bacterial sequences emerged through a secondary transfer from the animals. However, the potentially greater antiquity of the prokaryotic lineages possessing these proteins, combined with their greater diversity, makes this direction less likely given the current data. In addition, as discussed below, the case of the ART-LGIC receptors seems to fit the general pattern, which is observed for many other eukaryotic signaling proteins that appear to have a bacterial provenance. It is of interest to note that several other animal neurotransmitter receptors show connections to bacterial signaling systems. In addition to the MCP-N and Cache domains shared by the metazoan voltage-gated Ca 2+ channels, and the PBP-I domains of various G-protein-linked and NMDA-type receptors, there are similar parallels in the receptors for the gaseous neurotransmitter nitric oxide (NO). The NO receptors of animals share two domains, namely the HNOB and HNOBA, which are involved in heme-dependent NO sensing with several bacterial signaling proteins [ 46 ]. Likewise, a recent analysis of the enzymes in the biosynthetic pathways of all common metazoan neurotransmitters suggested that many of them may have been laterally transferred from bacteria to eukaryotes at different points in eukaryotic evolution [ 47 ]. Some of these include some potentially late transfers, analogous to previous observations for the NO receptors and the present report on ART-LGICs. Furthermore, parallel instances of connections to bacterial sensory proteins have been noted in the case of plant receptors for cytokinin, ethylene and light (phytochromes), and certain small-molecule receptors of the cellular slime mold Dictyostelium (see [ 48 ] and references therein). Thus, the ART-LGICs appear to belong to a larger sensory network that probably first emerged in the bacterial signaling systems and was subsequently recruited by the eukaryotes in contexts unique to their own functional milieus. Conclusions We report here the identification of several prokaryotic homologs of the animal acetylcholine receptor-type ligand gated ion channels (Cys-loop receptors). The pattern of the residues conserved in both the metazoan and bacterial receptors suggests that a common mechanism of channel-gating is likely to operate throughout this superfamily. Furthermore, the ligand-binding box appears to preserve at least one aromatic residue, although its exact position may not necessarily be conserved. The conservation pattern also suggests that a chain of positions leading out on either side from the ligand-binding box may mediate the transmission of the conformational change through the 'top' of the LBD, which may then transmit through the rest of the structure. The charge interactions between the acidic residue in the middle of the Cys-loop region and a basic residue the extreme carboxyl terminus of the LBD, just before the transmembrane domain also appear be universal features that might be involved in the process of channel gating. On the basis of the domain architectures and operon organizations, we predict that the bacterial ART-LGICs are likely to function as chemoreceptors for low-molecular-weight solutes in the environment. Phyletic and phylogenetic analyses suggest that the ancestor of the animal lineage probably acquired a single progenitor from a bacterial source, and it subsequently radiated to give rise to all the Cys-loop receptor subunits of the extant metazoans. Materials and methods The nonredundant (NR) database of protein sequences (National Center for Biotechnology Information (NCBI)) was searched using the BLASTP program [ 49 ]. Unfinished microbial and eukaryotic genomes were searched using the TBLASTN program with protein queries [ 49 ]. Iterative database searches were conducted using the PSI-BLAST program with either a single sequence or an alignment used as the query, with a position-specific score matrix inclusion expectation (E) value threshold of 0.01 (unless specified otherwise); the searches were iterated until convergence [ 49 ]. For all searches with compositionally biased proteins, the statistical correction for this bias was used. Multiple alignments were constructed using the T_Coffee [ 22 ] or PCMA programs [ 22 ], followed by manual correction based on the PSI-BLAST results and structural information. All large-scale sequence-analysis procedures were carried out using the SEALS package [ 50 ]. Transmembrane regions were predicted in individual proteins using TMPRED [ 51 ], TMHMM2.0 [ 52 ] and TopPred II [ 53 ] with default parameters. For TopPred, the organism parameter was set to 'prokaryote' or 'eukaryote' depending on the source of the protein. Signal peptides were predicted using the SIGNALP program [ 54 ]. Protein structure manipulations were performed using the Swiss-PDB viewer program [ 55 ]. Protein secondary structure was predicted using a multiple alignment as the input for PHD [ 20 ] or JPRED2 [ 21 ]. Similarity-based clustering of proteins was carried out using BLASTCLUST [ 56 ]. Phylogenetic analysis was carried out using the maximum-likelihood, neighbor-joining, Bayesian inference and minimum evolution (least squares) methods. The MrBayes program was used for the Bayesian inference of phylogeny [ 57 ]. The alignment for phylogenetic analysis was prepared by visually deleting all those columns that contained non-conserved residues from five or fewer sequences. Regions with substantial gaps, which are replaced by numbers in Figure 1 , were also entirely deleted from the alignment. The resulting alignment with 49 sequences and 368 columns was used for all subsequent phylogenetic analysis. Maximum-likelihood distance matrices were constructed with the TreePuzzle 5 program [ 58 ] using 1,000 replicates generated from the input alignment and used as the input for construction of neighbor-joining trees with the Weighbor program [ 59 ]. Weighbor uses a weighted neighbor-joining tree construction procedure that has been shown to correct effectively for long-branch effects. The minimal evolution trees were constructed using the FITCH program [ 60 ] of the Phylip package on 1,000 bootstrap replicates prepared from the input sequence. For maximum-likelihood analysis two different procedures were used. In the first, a minimum evolution tree obtained using FITCH was provided as a input for the Protml program [ 61 , 62 ], which then produced a maximum-likelihood tree using local rearrangements. The statistical significance of the internal nodes of this maximum-likelihood tree was assessed using the relative estimate of logarithmic likelihood bootstrap (Protml RELL-BP) [ 61 , 62 ], with 10,000 replicates. In the second procedure an initial full maximum likelihood tree was constructed using the Proml program of the Phylip package [ 60 ]. A gamma distribution with one invariant and four variable sites with different rates was used for constructing this tree, which was then used as the guide tree to generate further full maximum-likelihood trees using the PhyML program with 100 bootstrap replicates generated from the input alignment [ 63 ]. The consensus of these 100 trees was derived using the Consense program of the Phylip package to obtain the bootstrapped full maximum-likelihood tree. Gene neighborhoods were determined by searching the NCBI PTT tables with a custom-written script. These tables can be accessed from the genomes division of the Entrez retrieval system. Additional data files The following additional data are available with the online version of this paper. Additional data file 1 contains the conservation pattern of the ART-LGIC superfamily plotted onto the three-dimensional structure of the ACHB protein. Additional data file 2 contains the alignment of the proteins in Figure 1 in Word format. Supplementary Material Additional data file 1 The conservation pattern of the ART-LGIC superfamily plotted onto the three-dimensional structure of the ACHB protein Click here for additional data file Additional data file 2 The alignment of the proteins in Figure 1 Click here for additional data file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549065.xml
532398	Cro-Magnons Conquered Europe, but Left Neanderthals Alone	null	After miners unearthed a skull and bones in a Neander Valley cave in Germany in 1856—three years before the publication of On the Origin of Species —the remains were initially described as either those of a “brutish” race or of someone disfigured by disease. As Darwinian evolution caught on, so did the realization that these fossils were evidence of an earlier human species. Scientists have been debating Neanderthal's place in human evolution ever since. An ongoing question concerns the possibility that Neanderthals and early humans mated, since they likely crossed paths during thousands of years of European cohabitation. In a new study, Mathias Currat and Laurent Excoffier present a simulation model based on what we know about the population density and distribution of Neanderthals and Cro-Magnons. Their results complement recent genetic and morphological evidence indicating that early human and Neanderthal interbreeding was unlikely. The notion that modern Europeans directly descended from Neanderthals has mostly yielded to two competing models: One postulates that modern humans arose in Africa about 130,000 years ago and completely replaced coexisting archaic forms with no interbreeding, while the other proposes a gradual transition with interbreeding. Though mounting genetic evidence (based on mitochondrial DNA extracted from fossils) suggests Neanderthals and early humans did not breed, the evidence has been inconclusive. It's possible, for example, that any Neanderthal gene “leakage” could have been lost through genetic drift if the mating populations were small. And because so few fossils are available to analyze, previous studies could rule out only Neanderthal contributions over 25%. Currat and Excoffier's model refines various parameters—such as geographic boundaries, local population variations, range expansion, and competition for resources—based on archeological and demographic data for both populations. Evidence suggests modern humans replaced Neanderthals over 12,500 years, for example, which constrains the speed at which modern humans could expand. The authors started with a scenario based on a set of “plausible” parameter values—their basic scenario—and then varied the local interbreeding rate and, for example, the population size and location of Cro-Magnons, to produce eight alternate scenarios describing how Cro-Magnon colonization of Europe might have proceeded. They estimated the likely proportion of Neanderthal gene contributions to the modern gene pool using “coalescent simulations,” which generate the genealogies and diversity of genes in local populations based on simulations of their population densities and migration histories. The simulations show that if Neanderthals bred with Cro-Magnons without constraints over thousands of years, Neanderthal contributions to the modern gene pool “would be immense.” Surprisingly, the simulations also show that even a very small mixing should lead to high levels of Neanderthal DNA in modern humans. Reconstruction of Neanderthal woman What could account for this counterintuitive result? Given a low population density with small local breeding populations, any introduction of Neanderthal genes would decrease the frequency of Cro-Magnon genes of that population; if these Neanderthal integrations take place as the Cro-Magnon population is expanding, newly acquired Neanderthal genes would also be amplified Since no Neanderthal mitochondrial DNA has been found in modern-day Europeans, the authors modeled the maximum number of interbreeding events that would support this observation. The estimated maximum number of events, it turns out, falls between 34 and 120—extremely low values, Currat and Excoffier conclude, “given the fact that the two populations must have coexisted for more than 12,000 years.” While the authors acknowledge their simulations suggest rather than reflect reality, their model does incorporate real historical data such as Cro-Magnon expansion over time and local population growth. At a value of only 0.1%, their new estimate of the rate of interbreeding is about 400 times lower than previous estimates and provides strong support that Neanderthals and Cro-Magnon didn't interbreed and may even have been different species.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532398.xml
521496	Caring for family members with chronic physical illness: A critical review of caregiver literature	This article reviews 19 studies (1987–2004) on quality of life for family caregivers helping those with chronic physical illness. Here we explore the concepts of and instruments used to measure caregivers' quality of life. We were particularly interested in understanding stress-related variables and documenting factors influencing quality of life based on family stress theory. Findings show that various positive and negative terms equated with quality of life were used to measure them. Results indicate that stress-related variables as possible predictors influencing caregivers' quality of life include: patient and caregiver characteristics, stressors, stress appraisal, stress coping methods, and social support. Our recommendations touch upon applying theory for intervention, developing measurement, making operable the concepts for measuring, and the need for longitudinal and comprehensive study.	Background Recent reforms in U.S. health care systems mean that individuals with long-term, complex health problems are being cared for at home by family members [ 1 ]. Specifically, changes in medical practice resulting in shorter impatient hospital stays and the search for outpatient substitutes such as home-based care have brought cost savings to both hospitals and consumers. A study reported that home-based care reduced the cost per patient treated by 44% overall compared with hospital-based treatment [ 2 ]. Despite such cost-effectiveness, this trend means that an increased financial, physical, and emotional responsibility falls upon family members who care for a person with chronic physical illness [ 3 ]. Now, more than 25 million Americans serve as family caregivers for that population. Their work, if it were part of the market economy, would have an economic value of nearly $257 billion in 2000, which is equal to 20 % of the total for all health-care expenditures [ 4 ]. For example, family caregivers are more frequently called upon to use daunting and complex equipment at home. They also deal with extensive coordination of care, including symptom management, disability, mobility, and dressings. In the face of these increasing challenges and responsibilities, caregivers often feel tired, isolated, and overwhelmed, because they lack support, training, information and a sympathetic ear. Furthermore, some family caregivers who are employed report missing work, taking personal days, and quitting or retiring early to provide care [ 5 ]. Thus, chronic illness affects not only the lives of those suffering from disease but also those of family members who care for them. Attending to the impacts of chronic illness on family members is important because the physical and emotional health of family caregivers has the potential to influence the health, welfare and successful rehabilitation of persons with such chronic illness [ 6 ]. Existing studies document how caring for chronically ill family members or significant others at home influences multiple aspects of caregivers' lives. These effects are physical, psychological and social and may include worsened physical health, impaired social and family life, and increased stress, anxiety and depression ([ 7 - 9 ]). Placing these conditions experienced by caregivers in the context of family stress theory and quality of life advances our understanding of caregivers' experiences by examining how multiple aspects of caregivers' lives – their quality of life – may be partly influenced by other existing environmental stressors, stress appraisal, coping methods and social support. The importance of family stress theory in studying normative family transitions and adaptation to major life changes and illness is based on the central role that family strengths and capabilities play in understanding and explaining psychological and behavioral outcomes [ 10 ]. In family stress theory, the family is "viewed as encountering hardships and changes as an inevitable part of family life over the life cycle" [ 11 ]. Given that caring for a seriously ill family member arguably is (or is quickly becoming) an inevitable part of family life, current research has begun examining family stressors, stress appraisal, coping methods, and social support as they influence QOL outcomes, or as they attenuate the effects of other patient or caregiver characteristics or health-related variables on caregiver QOL ([ 12 - 15 ]). Quality of life is a construct that encompasses health and functioning, socioeconomic status, psychological, emotional and spiritual aspects, and family [ 16 ]. Ferrell [ 17 ] confirmed the influence of such multidimensional aspects on the QOL for breast cancer survivors. Wyatt & Friedman [ 18 ] also identified concerns related to QOL in those with chronic physical illness, and suggested that considering multidimensional aspects of QOL for them is essential. Although some scholars have different points of view regarding the dimensions of QOL, most researchers generally agree that QOL is multidimensional, subjective, and relating to a state of physical, psychological, social, spiritual well-being [ 19 ]. However, QOL for caregivers includes more aspects such as burden and family functioning [ 20 ]. Further, some researchers use QOL together with life satisfaction, adaptation, health, and distress ([ 12 , 13 , 21 - 23 ]). Such comprehensive consideration has led to an awareness of QOL as a broader and more appropriate concept for determining how caregiving affects family members [ 1 ]. Therefore, it is meaningful to address QOL comprehensively in this paper. This work will also consider 1) the lack of consensus on concepts like stressors, stress appraisal, coping methods, and social support, and 2) inconsistent results regarding factors influencing caregivers' QOL. To address the lack of review papers on this topic as well as inconsistent results among empirical studies, we undertook a thorough review of the literature. We were particularly interested in research that reported psychosocial or QOL outcomes and accounted for family stress variables as correlates or explanatory variables. The purpose of this paper is twofold: to review caregiver QOL impacts (loosely-defined) organized around family stress theory, and to discuss the implications of findings for future research (i.e., measurement, hypothesis testing, refinement of concepts and constructs). Methods Only published peer reviewed research articles were included in this review. Several methods were applied in searching the literature. First, articles were limited to those published in English between January 1, 1987, and January 31, 2004. Second, participants were caregivers of patients over 21 years of age with chronic physical illnesses. Third, a study was excluded if patients were receiving hospice care. Fourth, a computer search was conducted in February 2004 to review the databases of MedLine and PubMed by using the following key words: caregivers, caregiving, chronic illness, quality of life, adjustment, life satisfaction, burden, distress, and family stress theory. The terms "health," "stress," and "coping" were also used for study retrieval. Databases also were searched for review articles published during the same time period. Additional sources for empirical reports included reference lists from published studies. Of the more than 220 articles identified, there were only 16 empirical studies based on family stress theory that focused on the caregivers of patients with chronic physical illness and measured the QOL of caregivers. Two additional articles were identified from searching review articles. Finally, one more article was found in the references of these 18 articles. Thus, a total of 19 articles are included in this review (Appendix 1 [See additional file ]). Results Concepts and Instruments Used to Measure Caregiver Quality of Life Reviewed studies used diverse outcomes and models to indicate caregivers' QOL. Psychosocial outcomes falling with varied dimensions of QOL and measured included adaptation, mental health, life satisfaction, stress, emotional distress, health, caregiver burden, and depression. Given the variety of terms equated with QOL, both positive and negative terms were used to measure the QOL. Three studies ([ 12 , 13 , 21 ]) used positive terms such as 'adaptation' and 'life satisfaction.' Seven studies ([ 15 , 22 - 27 ]) measured QOL using negative terms such as 'caregiving burden,' 'depression,' 'stress outcomes,' and 'emotional distress.' Other research used neutral terms including 'health,' 'mental health outcomes,' and 'quality of life.' Of 19 reviewed articles, only one used a single measure of QOL per se : the Caregiver Quality of Life Index containing items related to physical, emotional, social, and financial wellbeing [ 28 ]. Eight papers reported use of a single instrument based either on a modified patient questionnaire or on other concepts (burden, general health, and mood status) indicating various QOL dimensions. Ten reported studies administered a composite measurement combining some concepts. Table 1 shows diverse outcomes and instruments used to measure caregivers' QOL. Table 1 Concepts and measurements of quality of life Outcome (QOL) Single Measurement Composite Measure Quality of life 1. Caregiver quality of life index – cancer [28] 1.1) Stress, 2) Anxiety, 3) Depression, 4) Health [7] 2. Multidimensional quality of life scale – cancer [33] 2.1) Physical health, 2) Emotional health, 3. Quality of life index [32] 3) Use of psychotropic drugs, 4) Caregivers' social life, 5) Financial status [30] 1.1) Depression, 2) Life satisfaction, 3) Health [12] Adaptational outcome (Adaptation) 2.1) Caregiver's level of life satisfaction, 2) Depression 3) Subjective symptoms of stress [13] Mental health outcome 1.1) Depression, 2) Quality of life [29] Life satisfaction 1. 8-item, open-ended questionnaire [21] Stress outcome (stress response, distress) 1. Brief symptom inventory [34] 1.1) Yielding of role, 2) Physical health, 3) Anxiety [24] 2.1) Anxiety, 2) Depression, 3) Stress [25] 3.1) Burden, 2) Depression, 3) Anxiety [26] Emotional (Psychological) distress 1. Profile of mood states – short form [22] 2. General health questionnaire [23] Health 1. the Medical Outcome Study 36-item 1.1) Caregiver mental health, Short Form Health Survey [31] 2) Caregiver physical health [14] Caregiving Burden 1. Zarit burden scale 27 Depression 1.1) depression, 2) strain [15] Factors Influencing Caregiver Quality of Life Studies of variables influencing caregivers' QOL are summarized in Table 2 . Variables examined are categorized here into patient characteristics, caregiver characteristics, stressors, stress appraisal, stress coping methods, and social support. Table 2 Factors influencing caregiver quality of life Predictors Significant Mediating variables Quality of life Significant Not found Patient characteristics Performance status [7], [24], [26], [28], [34] Age [7] [15] Gender [15] Depression [7], [29] The kind of illness [30] Pain / symptoms (severity of illness) [29, 32] Caregiver characteristics Age [7], [24] [15] Gender [7], [15] [24] Physical disability [7] Income [29], [32] Initial quality of life [29] Educational level [22], [28] Health problem [26] Depression [29] Anger [26] Anxiety [26] STRESSOR Primary Objective context [13] [14] Caregiving demands [13], [22] Patient impairment [12], [15] Duration of care [29], [25] Intensity of care [29], [25] ADL Dependency [7], [24], [26] [29] Stress types [25] Caregiver overload [24] Recurrence [15] Problem behavior [24] Secondary Subjective context [13] Caregiving demands [13] Role change [21] [29] Responsibility [21] Caregiver experiences [29] Life style interference [22] Stress Appraisal Appraisal [12], [14] [12], [14], [23], [31] Perceived control [13] Differences in the perception (pt & caregiver) [33] Stress Coping Methods Coping responses [12], [13], [25], [14], [15] [12], [14], [26], [31] [24], [32] Social support Perceived adequacy of social support [14], [25], [34] [14], [21], [26], [31] Social life and social network [12], [15], [12], [7] Family life (Quality of relationship and Marital adjustment) [7], [29], [32] Loneliness [7] Resources [26] Formal support [24] Patient characteristics Nine articles examined the association between patient characteristics (including performance status, age, gender, depression, type of illness, pain, and symptoms) and caregiver QOL. Generally, there were significant correlations between caregiver QOL and the patient's physical and emotional characteristics as related to the illness. Seven studies ([ 7 , 29 , 24 , 26 , 28 , 31 ]) found that the patient's performance status, type of illness, and depression were related to the caregiver's QOL. However, two ([ 29 , 32 ]) showed that pain and physical symptoms were not related to the caregiver QOL. Two other articles investigated the relationship between patient age and caregiver QOL, but there were no consistent results. Schumacher, Dodd, & Paul [ 15 ] reported a relationship between patient gender and caregiver QOL, with caregivers of male patients reporting higher levels of strain. Caregiver characteristics Eight of 19 articles examined the relationship between caregiver characteristics and QOL. A caregiver's age, gender, physical disability, income, initial QOL, educational level, health problem, depression, anger, and anxiety were addressed as characteristics. Three studies examined the association between a caregiver's age and QOL, with two ([ 7 , 24 ]) reporting that older age of the caregiver is associated with increased stress. Two of three articles investigating the relationship between a caregiver's gender and QOL reported that females were more likely to be depressed ([ 7 , 15 ]). In addition, the caregiver's physical disability, income, initial quality of life, educational level, health problem, depression, anger, and anxiety were consistently found significant in reducing their QOL. Stressors Eleven articles investigated the relationship between stressors and caregiver QOL, but they did not show consistent results. To varying degrees, caregiving demands, patient impairment, the duration and intensity of care, ADL (activities of daily living) dependency, stressor types, caregiver overload, how much assistance is given the caregiver, recurrence of illness and problem behavior in the patient were identified as primary stressors. Secondary stressors were caregiving demands, role change, responsibility, caregiver experience, and life-style interference. Seven studies ([ 7 , 12 , 13 , 15 , 24 - 26 ]) found that primary stressors were related to reductions in caregiver QOL. However, five articles ([ 13 , 14 , 22 , 29 , 25 ]) did not find any association between primary stressors and QOL. Two ([ 13 , 21 ]) found a significant relationship between secondary stressors and lower QOL. In contrast, another [ 29 ] found that secondary stressors were unrelated. Vedhara, Shanks, Anderson, & Lightman [ 25 ] investigated the relative importance of stressor types on stress outcomes. Their study demonstrated that stressor types (e.g. daily hassles, caregiving-specific stressors, and life events) determined the stress outcomes, with the proportion of variance accounted for by the stressor indices (which ranged from 20% to 53%). Winslow [ 24 ] found that caregiver overload was positively related to caregiver anxiety. The hypothesized direct effect of a care receiver's problem behavior on the caregiver's yielding up his/her role was also supported by the findings. That is, higher levels of a care receiver's problem behavior as the primary stressor were more likely to lead to patient institutionalization. Examining the stress process in family caregivers of persons receiving chemotherapy, Schumacher et al. [ 15 ] corroborated that modest but significant negative relationships were found between caregiver strain and patient functional status as well as disease recurrence. Haley, Levine, Brown, & Bartolucci [ 12 ] also found a significant positive correlation between patient impairment on the IADL (instrumental activities of daily living) and caregiver depression scores. Wallhagen [ 13 ] found that the subjective context and subjective demands of caregiving as well as the objective context were associated with the caregiver's adaptation, including level of life satisfaction, depression, and subjective symptoms of stress. Aspects of the caregiving situation assessed by both the objective and subjective context indices included caregiver competence, social resources, the physical environment, and socioeconomic status or perceived financial adequacy. The objective and subjective demands of caregiving included caregiving responsibilities, instrumental activities of daily living, and personal demands. Examining determinants of caregiver outcomes through a longitudinal study, Nijboer, Trienmstar, Tempelaar, Sanderman, & Van den [ 29 ] considered various stressors as both mediating and predicting variables. Some stressors, including duration of care, intensity of care, ADL dependency of patients, and role change, were not related to caregiver QOL. On the other hand, caregiver experience (i.e. disrupted schedule, financial problems, lack of family support, loss of physical strength, and self-esteem) was a significant mediating variable affecting the relationship between caregiver, patient, care characteristics and a caregiver's mental health. They found that all caregiver experiences were related in the expected direction to the caregiver's level of depression. With regard to a caregiver's QOL, only the loss of physical strength and its impact on the caregiver's self-esteem appeared to be related significantly, also in the expected direction. Negative caregiver experiences were associated with low income, living with the patient exclusively, distressed relationship, high level of patient dependency, and high involvement in caregiving tasks. Cameron, Franche, Cheung, & Stewart [ 22 ] specifically examined the mediation of lifestyle interference as the secondary stressor between the amount of care provided and emotional distress. The results supported that lifestyle interference mediated the relationship between caregiving assistance and overall mood disturbance. However, they did not find the relationship between caregiver QOL and caregiving demands to be the primary stressor. Coping methods Eight reports examined the relationship between coping methods and caregiver QOL. Five studies ([ 12 - 15 , 25 ]) found that stress coping methods operated as a mediating variable affecting the relationship between predictor variables and caregiver QOL. Four studies ([ 9 , 12 , 14 , 26 ]) concluded that coping methods significantly predicted the caregiver's QOL. Two studies ([ 12 , 14 ]) showed that stress coping methods operated as both predictor and mediating variables, where two others ([ 32 , 24 ]) found no such significant relationship. Haley et al. [ 12 ] found that coping responses were significant mediators and predictors of all three outcome variables: depression, life satisfaction, and self-related health. They showed that when caregivers used logical analysis and problem-solving coping strategies, they enjoyed a higher QOL. Information seeking was related to a better health outcome, and affective regulation was related to better outcomes in health and life satisfaction. Emotional discharge was actually related to higher levels of caregiver depression. Goode, Haley, Roth, & Ford [ 14 ] found that initially higher proportions of approach versus avoidance coping predicted better health over time. This suggests that relatively greater use of approach coping may help optimize caregivers' health over time. They also examined mediated effects on physical and mental health outcomes. Changes in approach coping percentage were directly related to changes in depression for self-care stressors and memory and behavior problems, indicating that as relative levels of approach coping increase, depression decreases. Of the four coping strategies, Wallhagen [ 13 ] discovered that, only wishful thinking mediated the perceived control and outcome variables. That is, wishful thinking coping behavior had a negative relationship with all adaptation variables. The higher levels of perceived control also reported using wishful thinking behavior. Schumacher et al. [ 24 ] studied how coping mediated the relationship between strain and depression. A modest but significant negative relationship was found between caregiver strain and coping efficacy. Predictably, caregiver depression was also significantly related to coping efficacy. That is, caregivers who experienced less coping efficacy were more depressed. Stress appraisal Six articles examined the association between stress appraisal and caregiver QOL. Two ([ 12 , 14 ]) demonstrated that stress appraisal operated as both a predictor and mediating variable. Four articles ([ 33 , 13 , 23 , 9 ]) investigated how stress appraisal was associated with caregiver QOL as a predictor variable only. Thus, most research investigating the relation between stress appraisal and caregiver QOL showed their significant relationship. Haley et al. [ 12 ] found that measures of caregiver appraisal were consistently related to caregiver outcome and operated as mediator and predictor variables. When caregivers appraised patients' behavioral problems and disability as highly stressful and appraised themselves as lacking in self-efficacy, they experienced higher levels of caregiver depression. Goode et al. [ 14 ] found that changes in one domain of caregiving stress, memory and behavior problems produced changes in stressfulness appraisals as a mediating process. Changes in stressfulness appraisals were then positively associated with changes in depression and health symptoms. These results may suggest that the appraised stressfulness of memory and behavior problems mediates the relation between these problems and caregiver health outcomes. In contrast, Wallhagen [ 13 ] did not validate the hypothesis that perceived control mediates the objective and subjective aspects of caregiving and caregiver adaptation. However, he found that perceived control had a direct relationship with life satisfaction and depression. Thus, perceived control was associated with a higher level of life satisfaction and lower levels of depression and subjective symptoms of stress. In addition to its direct relationships with the outcome variables, perceived control also had an indirect relationship with both life satisfaction and depression through its direct connection to wishful thinking. Miaskowski, et al. [ 33 ] investigated whether differences in patients' and family caregivers' perceptions of pain experience influence patient and caregiver outcomes. In terms of the QOL measures, significant differences were found for psychological well-being, interpersonal well-being, nutrition and the total QOL score, with reporting lower scores for patients whose pain intensity scores were non-congruent with their family caregivers. These data suggest that non-congruence in the patient's and the caregiver's perceptions of the patient's pain may result in a poorer QOL score for the patient. Social support A majority of the articles addressed social support as a mediator and predictor variable. Six articles ([ 12 , 14 , 15 , 24 , 25 , 34 ]) reported that social support mediated both predictor and outcome variables. Seven ([ 7 , 12 , 14 , 21 , 29 , 26 , 32 ]) showed a direct relationship between social support and caregiver QOL. Specific items of social support addressed here were perceived adequacy of social support, social life, social network, family life (quality of relationship and marital adjustment), loneliness, resources, and formal support. Examining predictors of adaptational outcome among dementia caregivers, Haley et al. [ 12 ] found that social support and activity were significant predictor variables of caregiver outcome. They discovered that higher levels of social network size, activity, and satisfaction with network were related to better outcomes, particularly life satisfaction and health. Social support and activity also mediated the stressors and caregiver outcomes, including depression, life satisfaction, and health. In the study conducted by Schumacher et al. [ 15 ], caregiver depression was significantly related to perceived adequacy of social support. Predictably, caregivers with less social support were more depressed. Social support was found to mediate the relationship between functional status and depression. Goode et al. [ 14 ] found that initial levels of social support also protected physical health changes over time. Those caregivers who reported higher initial levels of social support resources actually showed improved health over time. Initial satisfaction with level of social support provided the same beneficial effect in preventing physical health problems or promoting health improvements. Vedhara et al. [ 25 ] assessed the predictive stability of psychosocial mediators over a six month period. The results of the regression analyses revealed stable predictive relationships between the mediator factors and the stress response indices. Anxiety was predicted by seeking social support at six months. Ergh, Rapport, Coleman, & Hanks [ 34 ] also examined predictors of caregiver distress among 60 caregivers of patients with a traumatic brain injury. In this study, social support showed a direct relationship to family functioning. As well, social support powerfully moderated the caregiver's psychological distress. That is, in the absence of adequate social support, caregiver distress increased with performance status in care recipients. On the other hand, in studying how formal supports affect stress outcomes in family caregivers of Alzheimer's patients, Winslow [ 24 ] considered formal supports as mediators. He found that no formal support mediated primary stressors and caregiver characteristics in the directions hypothesized. Others ([ 21 , 26 , 7 , 29 , 32 ]) showed that social support was a predictor variable affecting caregiver QOL. Discussion Nineteen studies have been reviewed to understand stress-related variables and to examine how each factor influences a caregiver's QOL. Factors were identified from the literature based on family stress theory and included patient characteristics, caregiver characteristics, stressors, stress appraisal, coping methods, and social support. In this section, we will discuss method as well as family stress theory based on the literature reviewed. Theoretical and Clinical Considerations Family stress theory provides a way of viewing the family's efforts over time to adapt to multiple stressors through using family resources and perceptual factors as a coping process aimed at achieving family balance [ 35 ]. A family situation addresses multiple changes and demands simultaneously, not single stressors. Secondary stressors, such as role change, responsibility, and caregiving demands, emerge from the primary stressors and these strains often may be difficult to resolve. They become instead a source of chronic strain. Chronic strain causes a build-up of unresolved stressors and contributes to undesirable characteristics in the family environment [ 36 ]. Although most of the reviewed papers addressed stressors as factors influencing caregiver QOL, the focus was on primary rather than secondary stressors. As a result, researchers may be overlooking stressors without considering chronic illness as a source of chronic strain changing the family system. Therefore, it is imperative to consider secondary as well as primary factors for understanding stress on caregivers of patients with chronic physical illnesses. Resources for the family are the psychological, social, interpersonal, and material characteristics of individual family members, of the family unit, and of the community. That meets family demands and needs. When families have insufficient resources, their needs and demands are not adequately met. As a result, this contributes to increased conflict in the family environment [ 35 ]. Although this theory emphasizes resources, the focus on resources is remarkably broad [ 26 ]. For instance, some works ([ 13 , 14 , 23 , 26 ]) have addressed social support and coping as components of resources. A lack of clarity is likely to bring inconsistent research results and thus may cause problems of generalization. Instead we need to clearly prescribe each concept. Family stress theory suggests that stress may be perceived or experienced both as a crisis and a challenge to be overcome [ 10 ]. The family may perceive the stressor as having caused a crisis. Or the family may accept it and see it as a challenge. Perception of the stressor as a challenge suggests that, over time, families engaged in a constructive effort to manage the stressor will redefine their total situation [ 35 ]. This explanation regarding stress perception may artificially dichotomize the stressor as a crisis or a challenge. However, perception is so subjective that each family member may differently interpret an event occurring within a family system. The subjective characteristics of perception may impede measuring the cognitive patterns of each person. Even if caregivers perceive stressors as a challenge and have volition to cope with stressors, we must consider the possibility that recovery from crisis may not occur among caregivers of patients with chronic illness. In other words, many caregivers may have financial burdens caused by continuous treatment and tests. Additionally, although personal characteristics change, environmental characteristics of patients and caregivers, such as social prejudice against the chronically ill, may not change. As shown by inconsistent results from previous research findings, the assumption that caregivers recover from the crisis and achieve some level of adaptation needs to be re-considered. Family stress theory also emphasizes the need for intervention following clear assessment. In the Resiliency Model [ 37 ], practitioner interventions were directed at restoring the balance between family stressors and resources. This can be the first intervention to assess whether family behavior is adaptive or maladaptive. Robinson [38] maintains that in collaboration with the family it is necessary to develop a plan, for managing stress. The plan includes the following points: (a) commitment of all family members to work on the problems; (b) inclusion of all past successful coping strategies; (c) brainstorming of all possible strategies; (d) use of strategies that are flexible, reality-oriented, and open to expression of emotions; and (e) discussion of possible outcomes of all strategies. Figley & McCubbin [ 36 ] suggested that a family crisis should be an opportunity for family interventionists to promote family well-being. Social workers can not only make use of the community and its programs and services in support of families under stress, but also more importantly use the situation to improve the family's problem-solving skills, coping repertoire, and overall interpersonal relationships. Besides those interventions mentioned above, the reviewed papers suggested other interventions for caregivers and health care professionals on the basis of family stress theory. These interventions emphasized the need for health professionals' involvement in the ongoing care of cancer patients and their families to monitor increasing demands [ 29 ], the importance of educating family members in effective ways to communicate [ 33 ], and the need for long-term counseling and early involvement of caregivers [ 6 ]. Based on strategies mentioned above, social workers can use individual therapy, family therapy, education, and problem-solving programs as interventions for caregivers of patients with chronic illness. When social workers so apply family stress theory, they must maximize resources because resources can positively influence coping, perception, and adaptation. After finding resources, it may be helpful to educate caregivers, focusing on their perceptions, problems-solving, and coping skills. In family stress theory, resources, perception, and coping mutually interact, and thus may influence QOL among family members. As a result, synergistic interventions that integrate all factors may be more effective. Finally, family interventions must consider longer-term effects or the sustained effect, because coping methods and perception may change over time, depending on circumstances. Methodological Considerations Some studies of caregivers and family stress theory address various outcome variables, including QOL, adaptation, life satisfaction, emotional distress, and caregiving burden ([ 12 , 13 , 28 ]). As well, most research uses concepts such as stress, perception, coping, and social support as predictor or mediating variables. These concepts are so subjective that they are hard to define and measure using concrete methods. For example, Schumacher et al [ 15 ]'s study measured both perceived efficacy of coping strategies and perceived adequacy of social support with only one-item indicators. Thus, at first, operationalization of concepts is required to reduce the gap between theory and research and clear the way for measure. As well, specific instruments to measure the complex and multidimensional phenomena need to be developed. For a number of reasons, we need longitudinal studies in research applying family stress theory to caregivers of the chronically ill. Otherwise we cannot understand longer-term or sustained effects. First, although a patient with chronic illness may have completed treatment, recurrence of the illness is possible. Second, coping methods and appraisal may change depending on the circumstances, because family stress theory is dynamic and influenced by both internal and external environments. A longitudinal study may see the attrition of participants over time. Due to the nature of chronic illness, many patients may have died or experienced recurrence at the follow-up evaluation. Therefore, caregivers of these patients need to be followed through the end to assess predictors and consequences of patient death and recurrence of illness. Family stress theory is very complex and comprehensive. As a result, it is difficult to conduct research using the full model. Examples of this complexity include interactions among various factors, interactions and transactions over time, individual dynamics within the family system, and the balance between family demands and the family. Considering these multidimensional aspects, it would take much hard work to test the full model of family stress theory. However, we need to try. Otherwise, we may never understand the complex relation between individual, family, and environment. Supplementary Material Additional File 1 Appendix 1: Review of empirical findings on quality of life of family caregivers. Author, topic, significant predictor variables, mediating variables, outcome variables, measurement, and intervention of 19 reviewed articles are shown in detail. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521496.xml
549064	Genome-scale approaches for discovering novel nonconventional splicing substrates of the Ire1 nuclease	Three different genome-scale screens indicate that the HAC1 mRNA is the only substrate for the Ire1 nuclease in yeast.	Background The unfolded protein response (UPR) regulates the protein-folding and secretory capacity of eukaryotic cells by monitoring conditions within the endoplasmic reticulum (ER) and regulating a downstream gene-expression program (reviewed in [ 1 - 3 ]). In yeast, about 5% of the genome is under the transcriptional control of the UPR [ 4 , 5 ]. Induction of this vast set of genes is thought to lead to a restructuring of the secretory pathway to allow an increased protein flux to the cell surface and enable the cell to tolerate protein-folding stress. Hence, the UPR adjusts the secretory capacity of cells by feedback regulation. The identification and characterization of the UPR signaling components revealed a unique mechanism of signal transduction whose salient features are conserved among all eukaryotes. The UPR is initiated when the amino-terminal portion of the serine/threonine ER-transmembrane kinase Ire1 detects unfolded proteins within the ER lumen [ 6 , 7 ]. Accumulation of unfolded proteins sequesters chaperones and thereby allows Ire1 molecules to oligomerize in the plane of the ER membrane. Oligomerization, in turn, results in trans -autophosphorylation of the cytosolic kinase domain, providing the means by which the signal is transmitted across the ER membrane. Activated Ire1 acts as a site-specific endoribonuclease, cleaving the mRNA encoding the transcription activator Hac1 at two discrete positions, and removing a 252-nucleotide nonclassical intron [ 8 , 9 ]. A second enzyme, tRNA ligase (Rlg1), then joins the severed exons to produce a spliced version of HAC1 mRNA, termed HAC1 i mRNA ( i for UPR-induced) [ 10 ]. The HAC1 mRNA splicing reaction mediated by Ire1 and Rlg1 is spliceosome-independent, and utilizes chemistry that closely resembles pre-tRNA splicing [ 9 , 11 ]. As in pre-tRNA processing, 5' and 3' splice-site cleavage of HAC1 mRNA occur independently. In contrast, spliceosome-mediated splicing is a series of two transesterification reactions that have to be strictly ordered: 3' splice-site cleavage cannot occur before 5' splice-site cleavage. Intriguingly, this Ire1-mediated splicing reaction happens on polyribosome-bound HAC1 mRNA in the cytosol [ 12 , 13 ]. In the unspliced HAC1 mRNA, the intron forms a long-range base-pairing interaction with the 5' untranslated region (UTR) that is responsible for preventing its translation [ 12 , 13 ]. Splicing abolishes translational inhibition, allowing production of the Hac1 transcription factor and induction of UPR target genes. Hence, removal of the intron provides a key regulatory step in the signaling pathway. The amino-acid sequence of the nuclease region of Ire1 reveals significant similarities to that of RNase L, a mammalian endoribonuclease that is activated by interferon during viral infection [ 14 , 15 ]. RNase L functions to eliminate infected cells by nonspecific degradation of cellular RNA. Mutagenesis analysis maps the endoribonuclease activity of both proteins to homologous carboxy-terminal domains [ 16 , 17 ]. The nuclease domains in both proteins are preceded by kinase domains, and only the oligomerized forms of each protein appear to be active nucleases. Despite these similarities, the endoribonuclease activity of Ire1 is specific for the HAC1 mRNA, whereas the nuclease activity of RNase L shows no sequence specificity. Previously, to demonstrate the sequence specificity of Ire1, we devised an in vitro assay which utilized a purified, recombinant Ire1 protein [ 9 ]. This protein, here referred to as Ire1, is composed of a linker region (bridging between the membrane anchor of full-length Ire1 and the cytosolic kinase domain), the kinase domain itself, and the RNase domain. Ire1 cleaves HAC1 mRNA faithfully at both 5' and 3' exon-intron junctions and thus recapitulates the substrate specificity of full-length Ire1 in vivo . Other RNAs, including actin mRNA and poly(U) RNA, which has been demonstrated to be an RNase L substrate [ 18 ], are not cleaved by Ire1* [ 9 , 16 ]. In metazoans, multiple parallel pathways originate from the ER and contribute to the UPR. Two bona fide Ire1 orthologs IRE1α and IRE1β [ 19 , 20 ] are present in higher eukaryotes. In addition, a Hac1 ortholog (XBP1) is activated by a similar nonconventional splicing step that changes the protein's carboxy-terminal sequence by introducing a frameshift [ 21 - 23 ]. Another transmembrane kinase, PERK, shares structural similarity to Ire1 in its ER-luminal unfolded-protein-sensing domain and is also activated upon accumulation of unfolded proteins [ 24 , 25 ]. Activated PERK phosphorylates the translation initiation factor eIF2α, thereby inactivating it. This branch of the pathway leads to a global repression of translation, which is thought to lessen the secretory burden on the ER. In cells with reduced eIF2α activity, mRNAs containing small open reading frames (ORFs) in their 5' UTRs become preferentially translated. One such mRNA codes for the transcription factor ATF4, which collaborates with XBP1 to induce UPR target genes [ 26 ]. A third branch of the metazoan UPR activates the transcription factor ATF6, which is initially synthesized as an ER transmembrane protein [ 27 , 28 ]. Accumulation of unfolded proteins allows ATF6 to leave the ER and move to the Golgi compartment, where it encounters proteases that release its cytosolic portion as a soluble protein [ 27 , 29 - 32 ], which participates along with XBP1 and ATF4 in executing the transcriptional program of the UPR [ 33 ]. To date, HAC1 mRNA is the only known RNA substrate for yeast Ire1, prompting the question of whether other substrate RNAs exist or whether Ire1 and HAC1 mRNA function as a matched enzyme-substrate pair that interact exclusively with each other. Here we describe three independent genome-scale methods that address this question. Each approach successfully identifies HAC1 mRNA as a substrate of Ire1; none of the approaches identifies any other mRNA as a qualified candidate. Two of these approaches represent novel applications of cDNA array technology and could be adapted to the study of other signal transduction pathways regulated by RNA processing. Results and discussion We first devised a molecular screen to identify mRNAs cleaved by the Ire1 nuclease (Figure 1 ). In brief, we isolated a total poly(A) + RNA fraction from cells and subjected it to in vitro cleavage reactions in the presence or absence of Ire1. Fragments that lost their poly(A) + tails due to cleavage by Ire1 were re-isolated, reverse transcribed, fluorescently labeled and hybridized to genomic cDNA microarrays to identify Ire1 substrates. For these experiments, we expressed and purified a recombinant Ire1* tagged with glutathione- S -transferase (GST). The GST moiety was removed during the purification by protease cleavage. Upon incubation of Ire1* with in vitro transcribed HAC1 mRNA, we observed efficient and accurate cleavage at both splice junctions as previously described [ 9 , 11 ]. We optimized reaction conditions such that HAC1 mRNA contained in a total poly(A) + RNA fraction from yeast would be efficiently cleaved, even in the presence of significant excess of other mRNAs. We incubated total poly(A) + RNA with Ire1* and then fractionated the products using oligo(dT) cellulose (Figure 2 ). Input RNA and the bound and unbound fractions of the cleavage reaction were analyzed by Northern blotting using a HAC1 -specific probe covering the 5' exon (Figure 2 , lanes 1-3). Note that HAC1 mRNA present in the input fraction (Figure 2 , lane 1) was efficiently converted to a faster migrating species corresponding to the released 5' exon and intron, which, lacking poly(A) tails, were recovered exclusively in the unbound fraction after oligo(dT) chromatography (Figure 2 , lane 3). Conversely, analyzing the same RNA fractions with a HAC1 probe directed to the 3' exon, the resulting 3' exon (still containing its poly(A) tail) was recovered exclusively in the oligo(dT)-bound fraction (Figure 2 , lane 5). No significant levels of uncleaved HAC1 mRNA were detectable after Ire1* cleavage (Figure 2 , lanes 2, 3, 5 and 6). Using these reaction conditions, we isolated Ire1-cleaved RNA fragments from total poly(A) + RNA, which were recovered in the unbound fraction after oligo(dT) chromatography. We prepared a mock-treated reference sample using reaction conditions which were identical except that Ire1 was omitted. We reverse-transcribed both samples and fluorescently labeled the resulting cDNA with Cy3 (Ire1-treated; green) or Cy5 (mock-treated; red), respectively. The two probes were then simultaneously hybridized to yeast microarrays. We expected that contaminating poly(A) - RNA or uncleaved poly(A) + RNA would be equally represented in the oligo(dT) unbound fractions of both Ire1-treated and mock-treated reactions, thus leading to equal representation of both fluorescent probes in the mixture. Indeed, scatter plots generated by quantitating the fluorescence levels of both Cy3 and Cy5 (Figure 3a ) showed that most spots appeared on a tight diagonal, having hybridized to roughly equal amounts of both the Cy3 and Cy5 probes. We expected those mRNAs that are specifically cleaved by Ire1* to be enriched in the oligo(dT) unbound fraction of the Ire1-treated reaction compared to the unbound fraction of the mock-treated reaction, resulting in a higher Cy3/Cy5 fluorescence ratio in the microarray hybridization. Depletion of particular mRNAs by Ire1 digestion thus provides an enzymological tool to fractionate substrate from nonsubstrate mRNAs. We therefore plotted a histogram of the log 2 Cy3/Cy5 ratio for all mRNAs (Figure 3b , and see Additional data file 1). The histogram approximates a tight normal distribution (mean = 0.06; σ = 0.41) with only one significant outlier: At a log 2 Cy3/C5 ratio of 2.3, HAC1 mRNA falls 5.6σ from the mean of the distribution, clearly identifying this mRNA as an Ire1* substrate. Indeed, under these conditions, HAC1 mRNA is the only substrate for cleavage by Ire1* represented in the poly(A) + RNA fraction. As the Ire1 cleavage sites on other mRNAs could, in principle, be located within the 5' or 3' UTRs, we also hybridized the same probe to genomic microarrays containing yeast intergenic regions in addition to the ORFs. Our results were similar to those observed using ORF-only microarrays, with HAC1 being the only significant outlier (data not shown). To exclude the possibility that other potential RNA substrates might be hiding in the scatter of the data, our second approach combined the microarray analysis with a bioinformatics approach to search the genome for potential Ire1 cleavage sites. We had previously defined a consensus stem-loop motif by comparing the two splice sites of HAC1 mRNA and mutational analysis (Figure 4a ) [ 11 ]. Both cleavages occur between the third and fourth nucleotide of a predicted seven-nucleotide loop bounded by a stem; the first nucleotide of both loops is a C; the third and sixth nucleotide of both loops is a G; and the first nucleotide of the 3' leg of the stem is in both cases a G. The primary and secondary sequence information in this consensus is illustrated in Figure 4b . In vitro , Ire1* can cleave short RNA substrates containing only the 5' or 3' stem-loop sequences [ 11 ]. Mutational studies at nucleotide resolution demonstrate that each of the shared primary and secondary sequence elements is essential for efficient cleavage. We computationally searched the yeast genome for the presence of sequences fitting the consensus stem-loop motif. In this analysis, we searched ORFs, as well as 1,000 nucleotides upstream and downstream to include potential stem-loop motifs in 5' and 3' UTRs. The search yielded a total of 52 hits (Additional data file 2). In the list of genes identified, only a single gene, HAC1 , contains two predicted stem-loop structures, consistent with the notion that HAC1 mRNA is the only Ire1-dependent splicing substrate in yeast cells. For the remaining 52 predicted stem-loop structures, the possibility remained open that Ire1 would cleave some substrates at only a single site, perhaps to downregulate particular mRNAs. However, in the light of the data obtained from the microarray analysis described in Figure 3 we consider this possibility unlikely. We would have expected to see the 5' fragments resulting from such cleavage events in the oligo(dT) unbound fraction of Ire1-treated poly(A) - mRNA, resulting in high outlying Cy3/Cy5 ratios; however, none of the genes predicted to bear single HAC1 splice consensus sequences fell more than 3σ from the mean of the distribution (that is, none fell outside a 99% confidence interval). Because previous in vitro analyses have shown that stem-loop structures matching the consensus are sufficient for cleavage, we consider it most likely that the predicted stem-loop structures are either not included in the transcripts or do not fold as predicted in the context of the full mRNA sequences. In a third genome-scale approach, we exploited the phenotype of a mutant in tRNA ligase that is defective in UPR induction [ 10 ]. Characterization of this mutant, rlg1-100 , previously showed that Ire1-dependent cleavage of HAC1 mRNA occurs normally in the absence of ligation, but that the cleavage products are rapidly degraded. We assume that any other mRNA following this pathway should suffer the same fate, and that we could identify substrates of Ire1 by looking for mRNAs whose steady-state levels drop upon UPR induction in the rlg1-100 mutant, but not in a wild-type cell. The Ire1-dependent selective disappearance of HAC1 mRNA in rlg1-100 cells thus provides us with a tool to assess the spectrum of mRNAs that utilize the Ire1-mediated splicing pathway. We treated rlg1-100 cells with either tunicamycin (to induce the UPR by inhibition of N -linked glycosylation in the ER) or with no drug as a control, isolated total mRNA, reverse transcribed and fluorescently labeled the cDNA with Cy5 (for the tunicamycin-treated sample) and Cy3 (for the untreated sample) before simultaneous hybridization to genomic microarrays. mRNAs which are depleted during tunicamycin treatment should have Cy3/Cy5 ratios greater than 1 (log 2 Cy3/Cy5 ratios greater than 0). As shown by the data in Figure 5 , the steady-state levels of most mRNAs remain unchanged upon tunicamycin treatment (Figure 5a , and Additional data file 3). As before, a histogram of the log 2 Cy3/Cy5 ratios followed a tight quasi-normal distribution (mean = 0.11, σ = 0.27) with a single outlier. HAC1 , at a log 2 Cy3/5 ratio of 1.5, HAC1 falls 5σ from the mean, and is thus successfully identified as a splicing target of the UPR. Once again, no other mRNA clearly satisfied the criteria for identification as an Ire1 substrate. The depletion of HAC1 mRNA upon tunicamycin treatment was specific to rlg1-100 cells. A similar analysis of wild-type cells showed the expected induction of UPR target genes and no depletion of HAC1 mRNA relative to an untreated control (data not shown). Thus, HAC1 mRNA again stands out as the unique substrate for Ire1-dependent cleavage in yeast. Conclusions The experiments presented in this paper represent three genome-scale approaches for identifying mRNA substrates of the Ire1-dependent mRNA splicing pathway in yeast. Each approach successfully and selectively identifies the HAC1 mRNA as a target of Ire1 nuclease. In vitro cleavage of mRNAs by Ire1, followed by microarray detection of the fractionated mRNAs, identifies HAC1 mRNA as significantly enriched in the population of mRNAs specifically cleaved by Ire1. A computational search for the experimentally determined Ire1 consensus cleavage sites identifies HAC1 as the unique gene containing two such sequences. Finally, in vivo induction of splicing in a cell containing wild-type tRNA ligase or the mutant rlg1-100 , and subsequent microarray detection of 'genetically fractionated' mRNAs, identifies HAC1 as selectively depleted in the absence of the wild-type ligase. No other mRNA met any of these criteria for identification as an Ire1 substrate. We consider it reasonable to conclude that among the set of robustly expressed genes, HAC1 mRNA is the lone substrate of Ire1. In principle, there are many reasons why each of the approaches presented here could have missed identifying an Ire1 substrate other than HAC1 mRNA. For example, a poorly expressed substrate would exhibit a low signal-to-noise ratio in the microarray readout of the in vitro cleavage assay, or a substrate cleaved close to the 5' end would still hybridize to cDNA arrays with an efficiency comparable to that of the uncleaved mRNA, and hence could have escaped detection. In our computational screen, we might have missed cleavage sites that are significantly divergent from the experimentally defined consensus. Finally, because tRNA ligase would not take part in a cleavage-only reaction on a single-site substrate, we would also not expect our third method to lead to a relative reduction of the abundance of such mRNAs in rlg1-100 mutant cells as compared to the wild-type (though we would have expected such substrates to be identified in the in vitro Ire1* cleavage experiment). Thus rigorously, we cannot conclude that HAC1 mRNA is the only Ire1 substrate. However, because the potential caveats are noncongruent and because HAC1 stands out unambiguously in each of the three methods applied, we consider it highly unlikely that additional substrates exist. Previous studies in metazoans have suggested that other RNAs are degraded when Ire1 is activated [ 17 , 34 ]. When overexpressed, wild-type IRE1α mRNA accumulated at levels that were greatly reduced compared to those in which an RNase-dead mutant was overexpressed, raising the possibility that Ire1 might downregulate its own mRNA by degradation. Similarly, overexpression of IRE1β appeared to cause fragmentation of 28S rRNA. In both cases, however, no Ire1 cleavage consensus stem-loop structures were found. It therefore remains questionable at this time whether these degradation events are directly due to cleavage by Ire1 itself or to indirect secondary effects. According to the current evidence, therefore, Ire1 has a single identified target in yeast and thus functions solely for the purpose of post-transcriptional regulation of HAC1 . In contrast, other signaling pathways in the cell are commonly branched, using one enzyme to activate multiple different substrates. The utilization of common principles of protein modification, such as phosphorylation or ubiquitination, readily allows cross-talk among different signaling pathways that use common components. One phosphatase, for example, can dephosphorylate substrates that have been phosphorylated by several different kinases. In contrast, communication between the ER and nucleus through Ire1 appears to be a very private affair. Ire1-dependent splicing offers unique mechanistic advantages, such as the ability to quickly complete the synthesis of partially made Hac1 transcription factor by releasing the translational arrest of stalled polyribosomes. Given that the ER is a topologically distinct compartment of the cell with metabolic considerations quite distinct from those of the cytosol, a communication route that is wired differently and therefore more insulated from other informational 'chatter' in the cell may be particularly beneficial in the secure passage of information from the ER to the nucleus. The linear connectivity (activation of Ire1 → production of Hac1) of the two key players of the yeast UPR is supported by genetic analysis. Both Δire1 and Δhac1 mutant cells display indistinguishable growth phenotypes and highly correlated gene-expression profiles [ 4 ]. In higher eukaryotes, signaling through the UPR is more complex, with multiple ER-proximal components each activating distinct downstream targets. Metazoan Ire1 exists as two isoforms (Ire1α and Ire1β) [ 20 , 35 ]; in addition, the transmembrane kinase PERK is activated and the membrane-tethered transcription factors ATF6α and ATF6β are released when unfolded proteins accumulate in the ER [ 27 , 36 ]. The diversity of these signaling components varies widely between different tissues, and the relative contribution of each of the parallel pathways to the induction of the downstream transcriptional program is currently the subject of intense study. Given this increased complexity, it is possible that there are other metazoan Ire1 splicing substrates in addition to the HAC1 ortholog XBP1, and the methods developed here may prove useful in searching for such putative, additional Ire1 substrates in metazoans. Materials and methods RNA isolation RNA samples were prepared as described previously [ 6 ]. Briefly, yeast cells were grown to mid-log phase in selective media. rlg1-100 cells were treated with 1 μg/ml tunicamycin (Calbiochem) for 40 min in the experiments described in Figure 5 . Total RNA was isolated in SDS-high salt buffer by hot phenol extraction. For the microarray experiments described in Figures 3 and 5 , poly(A) + RNA was isolated from total RNA using the PolyA + Tract mRNA isolation system (Promega) according to the manufacturer's instructions. Poly(A) + RNA used for in vitro Ire1* nuclease reactions was prepared by two rounds of purification using the PolyA + Tract system. Northern blot analysis Total cellular RNA and RNA recovered from in vitro nuclease reactions was analyzed by Northern blot hybridization as described previously [ 8 ] by separation on 1.5% agarose gels containing 6.7% formaldehyde. Hybridization probes were generated by random labeling of PCR fragments using [ 32 P]α-dCTP according to the manufacturer's instructions (Amersham). All probes prepared from DNA fragments were generated by PCR amplification of yeast genomic DNA. In vitro Ire1* nuclease reaction In vitro nuclease reactions were performed by incubating yeast poly(A) + RNA with recombinant Ire1* (expressed and purified from Escherichia coli ) as described previously [ 9 , 11 ]. Following phenol/chloroform extraction, the reaction mixture was fractionated on oligo(dT) magnetic beads using the PolyA + Tract mRNA Isolation System (Promega) according to the manufacturer's instructions. Control nuclease reactions were performed and processed identically, except that Ire1* was omitted from the reaction mixture. Microarray hybridization of amino-allyl coupled cDNA probe General protocols for microarray hybridization and for preparation of probe were as described previously [ 37 ]. A sample of oligo(dT) unbound RNA was reverse transcribed with StrataScript (Stratagene) using random primer A (5' GGTTCCCAGTCACGATCNNNNNNNNN 3', where N is any nucleotide) followed by addition of Sequenase to synthesize the second strand. The resulting double-stranded cDNA above was amplified in the polymerase chain reaction (PCR) (20 cycles) using primer B (5' GGTTCCCAGTCACGATC 3') complimentary to the specific sequence portion of primer A. This extra PCR amplification step was added because of the low abundance of Ire1*-cleaved RNA fragments. After ethanol precipitation, one fourth of these PCR reactions was used to re-PCR in the presence of fluorescently labeled nucleotides (either Cy3 or Cy5 dTTP). The resulting Cy3- and Cy5-labeled probes were combined. Unreacted fluorescent dye was quenched. Probes were cleaned up in a QIA-quick PCR purification spin column and hybridized at 50°C for 24 h to glass slide microarrays containing the entire yeast genome. Synthesis of the cDNA for poly(A) + RNA (2 μg) used for the experiment in Figure 5 was carried out by reverse transcription in the presence of aminoallyl-dUTP at 42°C for 2 h. cDNA prepared from untreated cells was coupled with Cy3, and cDNA prepared from tunicamycin-treated cells was labeled with Cy5. Data analysis Hybridized microarrays were scanned with a GenePix 4000A microarray scanner (Axon Instruments). GenePix Pro was used to analyze and to display the data. All data points with absolute fluorescence intensity less than 250 in either channel were discarded. Cy3 and Cy5 fluorescence intensities were normalized against one another to adjust for differences in labeling efficiency and scanner gain. Quantitative analysis was performed in Microsoft Excel. Stem-loop consensus search Pattern searching was performed using the public-domain software scan_for_matches [ 38 ]. Sequence file input was a database of all yeast ORFs ± 1,000 nucleotides obtained from the Saccharomyces Genome Database [ 39 ]. Pattern file input was r1 = {au,ua,gc,cg,ug,gu} RNA base-pairing rules were used; allowed base-pairs are the fields of argument r1. p1 = 4..4 YCNGNNGNG ~p1 These parameters instructed the program to use RNA base-pairing rules to find a sequence of four nucleotides (one arm of stem), followed by a sequence matching YCNGNNGNG (where Y is pyrimidine; the seven-nucleotide loop flanked the conserved closing nucleotide pair), followed by the reverse complement of the four nucleotides at the beginning (the second arm of the stem). The output of the computational screen is listed in Additional data file 2. Additional data files The following additional data are available with the online version of this paper. Additional data file 1 contains supplementary Table 1 which lists all the data displayed in Figure 3 ; Additional data file 2 contains supplementary Table 2, which lists the complete output of the computational screen; Additional data file 3 contains supplementary Table 3 which lists all data displayed in Figure 5 . Additional data file 4 is a Word file containing the captions and keys to the tables. Supplementary Material Additional data file 1 Supplementary Table 1 which lists all the data displayed in Figure 3 Click here for additional data file Additional data file 2 Supplementary Table 2, which lists the complete output of the computational screen Click here for additional data file Additional data file 3 Supplementary Table 3 which lists all data displayed in Figure 5 Click here for additional data file Additional data file 4 The captions and keys to the supplementary tables Click here for additional data file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549064.xml
551603	Comparative linkage analysis and visualization of high-density oligonucleotide SNP array data	Background The identification of disease-associated genes using single nucleotide polymorphisms (SNPs) has been increasingly reported. In particular, the Affymetrix Mapping 10 K SNP microarray platform uses one PCR primer to amplify the DNA samples and determine the genotype of more than 10,000 SNPs in the human genome. This provides the opportunity for large scale, rapid and cost-effective genotyping assays for linkage analysis. However, the analysis of such datasets is nontrivial because of the large number of markers, and visualizing the linkage scores in the context of genome maps remains less automated using the current linkage analysis software packages. For example, the haplotyping results are commonly represented in the text format. Results Here we report the development of a novel software tool called CompareLinkage for automated formatting of the Affymetrix Mapping 10 K genotype data into the "Linkage" format and the subsequent analysis with multi-point linkage software programs such as Merlin and Allegro. The new software has the ability to visualize the results for all these programs in dChip in the context of genome annotations and cytoband information. In addition we implemented a variant of the Lander-Green algorithm in the dChipLinkage module of dChip software (V1.3) to perform parametric linkage analysis and haplotyping of SNP array data. These functions are integrated with the existing modules of dChip to visualize SNP genotype data together with LOD score curves. We have analyzed three families with recessive and dominant diseases using the new software programs and the comparison results are presented and discussed. Conclusions The CompareLinkage and dChipLinkage software packages are freely available. They provide the visualization tools for high-density oligonucleotide SNP array data, as well as the automated functions for formatting SNP array data for the linkage analysis programs Merlin and Allegro and calling these programs for linkage analysis. The results can be visualized in dChip in the context of genes and cytobands. In addition, a variant of the Lander-Green algorithm is provided that allows parametric linkage analysis and haplotyping.	Background The oligonucleotide Mapping 10 K arrays [ 1 ] have been used for linkage analysis [ 2 - 4 ] and their advantages in genome coverage and information content compared to microsatellite-based assays has been demonstrated. The array contains 11,550 SNPs with an average heterozygosity rate of 0.32 and an average marker distance of 0.31 cM. However, the commonly used multi-point linkage analysis software packages such as GeneHunter [ 5 , 6 ] and Merlin [ 7 ] are command-line programs and it is not straightforward to find genes in the regions of high linkage scores. In addition, the haplotyping results are represented commonly in a text format without any gene context. Here we report the development of a new software tool called CompareLinkage that can be used for automated conversion of Mapping 10 K genotype data into the "Linkage" format for linkage analysis in Merlin, GeneHunter and Allegro [ 8 ]. In addition the program can convert the pedigree information and SNP marker information into the "Linkage" format. After performing the linkage analysis using one or more of these programs, the CompareLinkage software can export the linkage score information into the dChip software [ 9 - 11 ] to visualize the results within a chromosome window. In addition, we implemented a variant of the Lander-Green [ 5 , 12 ] algorithm into the dChipLinkage module to analyze pedigrees with up to 18 bits (bits = 2n-f ; with n = number of non-founders and f = number of founders) using the parametric linkage analysis method. We are currently testing and validating the implementation of the algorithm which will be described in detail elsewhere. The linkage score curves, genotypes and haplotypes are graphically displayed in a dChip chromosome window which has the genes, cytoband and SNP marker information included. Together the CompareLinkage and dChip software programs provide for the first time a graphical user interface (GUI) and an automated procedure for comparative linkage analysis utilizing three commonly used linkage software programs. Implementation The CompareLinkage software for comparative linkage analysis using Merlin and Allegro To analyze large pedigrees rapidly and to compare the linkage analysis results of different software packages, we developed a software tool called CompareLinkage to automate the following processes: (1) Converting of Affymetrix Mapping 10 K genotype data, pedigree files and marker information into the "Linkage" format [ 13 ], and detecting and fixing incompatibilities in pedigree genotypes. The input genotype text file for CompareLinkage can be a single text file containing genotypes for each sample or a combined text file as exported by the Affymetrix GDAS 3.0 software. (2) Automatically calling the software packages Merlin and Allegro for linkage analysis and converting the analysis results (LOD or non-parametric linkage (NPL) scores) into the input files for dChip to visualize the results in the context of genes and cytobands. (3) The SNP genotype data in the "Linkage" format can be converted into the dChip input files (genotype, pedigree and marker information files) to perform parametric linkage analysis by dChipLinkage. All steps are discussed in detail at the CompareLinkage software manual provided on the software website. All these functionalities are useful for cross-validation of linkage results and to identify concordance and discordances between different linkage analysis programs as well as between parametric and non-parametric linkage results. A graphical user interface (GUI) for Windows was also implemented in Java. In this GUI users are allowed to set their own working directory and the location of the Perl interpreter through the "Setting" menu. CompareLinkage's functions of converting file formats and getting dChip input files are incorporated through the "Convert" and "GetCurve" menu (Figure 1 ). Since computing is usually time-consuming, the code of calling the Perl program is executed in separate thread to provide better interaction. The output of the Perl program can be viewed in the message window (Figure 2 ). Figure 1 The CompareLinkage GUI dialog for choosing pedigree, genome information and genotype call files. Figure 2 The intermediate output of the CompareLinkage GUI. The dChipLinkage software module The Affymetrix Mapping 10 K array CEL files and genotype TXT files can be imported into dChip and visualized along cytobands and genes as previously reported [ 9 , 11 ]. The information of the SNPs such as their genetic and physical distance and allele frequencies from three ethnic groups (Asian, African American and Caucasian) is obtained from the Affymetrix website [ 14 ] and converted into the genome information files for dChip. The information of the reference genes and cytobands is obtained from the UCSC genome bioinformatics database [ 15 ] for the matching human genome assembly (hg12 or hg15) of the SNP information, and is organized into the refGene and cytoband file provided with dChip. We implemented a variant of the Lander-Green [ 5 , 6 , 16 ] algorithm in the dChipLinkage module of dChip to perform multipoint parametric linkage analysis and compute a LOD score at each SNP position. Disease allele frequencies, penetrance information and phenocopy information for dominant and recessive disease models can be selected by the user through a dialog (Figure 3 ). The Mendelian genotype errors inconsistent with parental genotypes are detected and set to missing genotypes. To handle other genotyping errors or wrongly mapped SNP markers, we assume a conservative genotyping error rate of 0.01 [ 1 ] (user adjustable) and regard observed genotypes as phenotypes in the likelihood computation [ 17 ]. As a result, the computation of the probability of the observed genotype data at one marker given an inheritance vector v involves the summation over all the possible real genotypes (or equivalently the founder allele configurations): Figure 3 The dChipLinkage dialog for specifying linkage analysis parameters. where F i represents the i th of all the possible founder allele configurations and is independent of v . P (real genotypes i \| v , F i ) is 1 since an inheritance vector and founder allele configuration uniquely determines the real genotypes, and P (observed genotypes \| real genotypes i ) involves comparing the real genotype and observed genotype for all the individuals and multiplying the probability by the error rate of 0.01 (default value) for each disagreement and 0.99 for each agreement. We also use the matrix-vector multiplication algorithm and bit reduction due to founder phase symmetry described in [ 16 ], and the founder allele factoring technique reported in [ 6 , 17 ] to speed up the computation of single-locus and accumulative likelihood vectors as well as the likelihood vector of disease phenotypes. We use the forward-backward computation in the Lander-Green algorithm to obtain the marginal probability distribution of inheritance vector at each SNP marker position given the data of all the markers on a chromosome. In addition the most likely inheritance vector at each marker given the genotype data of all the markers on this chromosome is calculated [ 6 ]. Conditioned on the most likely inheritance vector at a marker and the observed genotype data, we can find the most likely founder allele configurations. When there are competing inheritance vectors with the same largest marginal probabilities at a marker, we select the one with fewer crossover events from the last marker since the distance between adjacent markers are small (average 300 kb) and it is therefore less likely to have multiple crossover events between two markers in a pedigree [ 7 ]. Together these procedures give the haplotyping results of the pedigree data. dChipLinkage visualizes the haplotyping result in either the haplotype view or the ordered genotype view. Results The comparative linkage analysis using Merlin, GeneHunter, Allegro and dChipLinkage CompareLinkage can format Affymetrix Mapping 10 K SNP genotype output files and genotype files into the "Linkage" format and convert genome information and pedigree files into the formats suitable for Merlin (Version 0.10.2), GeneHunter (Version 2.1) and Allegro (Version 1.2). CompareLinkage removes all non-informative markers and calls the PedCheck software [ 18 ] to detect genotype incompatibilities using the pedigree information. A Mendelian genotype inconsistency at a SNP is handled by setting the genotype of this SNP in all the individuals to missing. For the analysis in GeneHunter, overlapping segments of large chromosomes are prepared, with each segment containing 150 or fewer markers with 75 markers in common between adjacent segments. Linkage scores are computed as the mean of two scores for the same marker from the two overlapping fragments. We ran genome-wide linkage analysis using all the three software packages and dChipLinkage for the 10 K SNP genotype data of three families: 5026.10 (Figure 4 ; autosomal recessive non-syndromic deafness disease, 13 bits, Asian), CR (Figure 5 ; recessive, 17 bits, Asian) and ER (Figure 6 ; dominant, 17 bits, Caucasian). For the parametric analysis, we use a disease frequency of 0.001, a penetrance value of 0.99 and a phenocopy of 0.01 for all the families and all the software packages. For GeneHunter and Allegro we ran both nonparametric and parametric analysis. For Merlin, the NPL_all statistic is computed. The allele frequencies are calculated based on the actual genotype data in each family. The LOD score or NPL score are computed at the position of the SNP makers. After running the analysis for all chromosomes, the two chromosomes with the largest LOD scores were selected from each pedigree and compared below. Figure 4 The pedigree structure of family 5026.10. The PED 4.2 software is used to draw the pedigrees. Figure 5 The pedigree structure of family CR Figure 6 The pedigree structure of family ER. Figures 7 , 8 , 9 , 10 , 11 , 12 show the comparative LOD score and nonparametric score plots in dChip for these chromosomes analyzed with GeneHunter, Merlin, Allegro and dChipLinkage. The vertical red line in the figures indicates the significance threshold and is set to 3 for parametric analysis (LOD scores) and to 3.7 for non-parametric analysis (NPL score) based on statistical significance recommended by Lander and Kruglyak [ 19 ]. The linkage scores largely agree with each other in the regions with significant LOD/NPL scores. GeneHunter, Merlin and Allegro detect the peaks in the chromosome 1 and 3 of the consanguineous family 5026.10 but compute lower LOD scores than dChipLinkage (indicated by arrows in Figure 7 and 8 ). For another consanguineous family CR with a recessive disease, all software packages detect similar peak regions in the two chromosomes denoted as A and B (Figure 9 and 10 ). For the family ER with a dominant disease, dChipLinkage computes similar overall patterns but reports a possible sporadic and non-significant peak (LOD < 1.6) in each chromosome (indicated by arrows in Figure 11 and 12 ). Figure 7 The comparative linkage results of the chromosome 1 of the family 5026.10 using CompareLinkage and dChipLinkage. The genotype calls are displayed on the left in yellow (AB), red (AA) and blue (BB), with SNPs on rows and samples on columns. The sample names and the disease status (1 = Unaffected and 2 = Affected) are displayed on the top. The linkage scores of different software are displayed on the right in the shaded box. The lower and upper limits of the shaded box (such as [-10, 6]) are in the brackets on the bottom of the curve. The red vertical line indicates the threshold of 3.0 for LOD scores and 3.7 for NPL scores. This line is user adjustable. Figure 8 The comparative linkage results of the chromosome 3 from the family 5026.10. The figure format is the same as Figure 7. Figure 9 The comparative linkage results of the chromosome A of the family CR. Figure 10 The comparative linkage results of the chromosome B of the family CR. Figure 11 The comparative linkage results of the chromosome A of the family ER. Figure 12 The comparative linkage results of the chromosome B of the family ER. Linkage analysis and visualization using dChipLinkage To do parametric linkage analysis in dChipLinkage, a pedigree file is needed (Figure 13C ). The file is similar to the standard pedigree file format but has an additional "Array" column matching each individual in the pedigree file to the corresponding genotype information in the genotype file through array names (the header line in Figure 13A ). The data importing and analysis steps are: Figure 13 The genotype file (A), genome information file (B) and pedigree file (C) used by dChipLinkage for analysis. 1. Open dChip. 2. Select the Analysis menu and the Get External Data function to read in the genotype file in the text format (Figure 13A ). 3. Select the genome information file downloaded from the dChip website (Figure 13B ). This file is provided in three versions, each containing the SNP information like TSC SNP ID and genetic map locations but having different allele frequencies for each of the three ethnic groups (Asian, Caucasian and African Americans). 4. Select the Analysis menu and the Chromosome function to display the genotype calls, genes and cytobands along the chromosome 5. After the program has displayed the genotype data, select the Chromosome menu and the Linkage function to start the dChipLinkage module (Figure 3 ). Specify the pedigree file (Figure 13C ) and other linkage parameters. Depending on whether the dChip " Chromosome View " displays one or all chromosomes, the linkage analysis will be performed for one or all chromosomes accordingly. For the analysis of the 5026.10 family, the recessive disease model is assumed, and a penetrance of 0.99, phenocopy of 0.01, disease allele frequency of 0.001 and a SNP marker error rate of 0.01 are used. The SNP allele frequencies in the genome information file are used and truncated to values between 0.001 and 0.999. This family has 13 bits and it takes about 20 minutes for the whole genome linkage analysis. Using dChipLinkage to analyze the 5026.10 family, we were able to identify a region on the chromosome 1 (Cytogenetic region: 1p36.32 – 1p36.22) with LOD scores of greater than 2.3 (Figure 7 , indicated by arrow). The most interesting gene in this region is ESPIN, which has previously been shown to be involved in deafness in mice [ 20 ] and two frameshift mutations in the gene have just recently been associated with deafness in two consanguineous families [ 21 ]. Sequence analysis of the locus revealed that the parents (the individual 1 and 2 in Figure 4 ) and the unaffected child (the individual 6) are heterozygous for the insertion mutation and the affected children are homozygous (data not shown). In addition a novel locus with a maximum LOD score of 2.77 was identified on the chromosome 3 (Figure 8 , indicated by arrow). The peak region on the chromosome 3 is about 2 Mb wide (Figure 14C ). Using the GeneHunter software, we compute a maximum expected LOD score of 2.78 for this family under the specified parameters. Therefore we extract the most linkage information based on the dense SNP markers in this region. Figure 14 shows the LOD score curve together with genotype calls, inferred haplotypes and ordered genotypes based on haplotyping. In Figure 15 the results are presented in the context of cytobands and genes. The Chromosome/Export SNP data function can also export the text information of the SNPs, genes and cytobands in the region with linkage scores exceeding the threshold. Figure 14 (A) In the genotype view, the red, blue, yellow and white colors represent genotype call AA, BB, AB and No Call. (B) The inferred haplotypes indicating ancestor origins are displayed in correspondence to the genotype view. The different colors represent distinct founder chromosomes. For each individual (column), the father allele haplotype is displayed on the left and mother allele haplotype on the right. (C) In the ordered genotype view, the red and blue colors represent the A and B genotype of father allele (left) and mother allele (right) in each individual (column). The LOD score curve is displayed in the shaded box on the right. The left boundary and right boundary of the box represent value of -2 and 3, and the red vertical line represents 2. Figure 15 (A) The peak LOD score region is enlarged and displayed proportionally to real chromosomal distance in the context of genes and cytobands. LOD score peaks are shown at the q-arm of chromosome 3 (114.18 -117.00 Mb, maximal LOD = 2.77). The shaded curve region has the same range as Figure 14. (B) A enlarged view of the peak region with more details of the individual SNPs and genes. The transcription starting site of the genes are used to display their positions. After the linkage computation is finished, the inferred haplotype information can be visualized. In the haplotype view (Figure 14 and 16 ), one can view the inference on how the founder chromosomes are crossed over and inherited by the descendants. The different colors represent distinct founder chromosomes, and for each individual, the father allele haplotype is displayed on the left and mother on the right. Since a pedigree contains no phase information of the founders [ 6 ], in the linkage computation we can assume that one child of each founder always inherits the whole grandfather-descent chromosome. This assumption does not affect the LOD score computation but reduces the number of bits in the Lander-Green algorithm by the number of founders and consequently reduces the analysis time. This is the reason that in Figure 14B the individual 1 has both father and mother haplotypes in pure color and individual 2 has only the father haplotype in pure color. By inspection of the observed genotype and the inferred haplotypes (Figure 16 ), one can see that only in the peak LOD score region all the affected children (individual 3, 4, 5 and 7) are homozygous and that the unaffected child (individual 6) is heterozygous. All the affected individuals share two copies of the identical chromosome segment (the pink color between the two arrows) presumably containing the disease locus. By two very close crossover events respectively in individual 6 (indicated by the black arrow) and individual 7 (indicated by the white arrow), the LOD score implicates the possible disease gene in a 2 Mb region and one can easily search the physical map for candidate disease genes in this region in the dChip chromosome view (Figure 15 ). Figure 16 The genotypes (A) and inferred haplotypes (B) from family 5026.10 on the peak score region of chromosome 3 are shown (for more details see the legend in Figure 14). In the peak LOD score region all the affected children (3, 4, 5 and 7) inherited the same ancestral allele in the consanguineous family and the unaffected child (6) inherited two different ancestral alleles. Discussion and conclusions We have developed the CompareLinkage software for easy comparison and analysis of genotype datasets with common multi-point linkage analysis software programs. It provides functions such as automated data formatting and the calling of linkage analysis software programs to facilitate comparative linkage analysis. The results can be visualized in a chromosome window in the context of genes, cytobands and SNPs in dChip's user friendly graphical interface. The linkage scores of other linkage software packages can be saved into the dChip score file format through CompareLinkage and viewed in the dChip chromosome viewer. This provides the interface to view other computed statistics such as linkage disequilibrium scores along the chromosomes. We have also implemented a variant of the Lander-Green algorithm as the dChipLinkage module for parametric linkage analysis of small pedigrees. It can analyze all chromosomes for families with up to 18 bits within one hour on a PC with one gigabyte memory. This is useful for recessive and consanguineous families whose bits are often small. The comparison analysis of three Mapping 10 K array data sets show similar results in regions with significant LOD scores across all the four software packages. The regions with concordant LOD/NPL scores should provide more confidence in the candidate disease loci. However, there are clear differences in isolated regions. This emphasizes the challenge of a comparative analysis using different linkage algorithm implementations. We hypothesize that the differences between the software programs in peak locations are attributable to: 1. The specific algorithm implementation in each program. 2. The difference between parametric – and non-parametric analysis. 3. The existence of undetected genotype errors in the data sets which could falsely deflate LOD scores [ 17 , 22 ]. dChipLinkage uses an error model to automatically handle genotype errors and avoid sporadic LOD score peaks due to undetected non-Mendelian errors, and results in a smoother LOD curve as seen in Figure 7 , 8 , 9 , 10 , 11 , 12 . However, this error handling algorithm involves more iterations and increases the computation time. There are further techniques to reduce the memory and time requirement of the Lander-Green algorithm [ 7 , 8 , 23 , 24 ] In light of the discordance between the results from common linkage software packages and from dChipLinkage, we will validate dChipLinkage implementation using additional datasets and the CompareLinkage software. In summary, the CompareLinkage and dChipLinkage software automate the comparative linkage analysis and visualization using multiple software packages. With these tools users will be able to increase their confidence in candidate regions and can use the visualization tools to explore the disease associated genome regions. Availability and requirements Project name: The CompareLinkage software and the dChipLinkage software module Project home page: Operating system(s): Windows (dChipLinkge); Windows (CompareLinkage and its graphical interface), Unix (CompareLinkage command line version) Programming language: Visual C++ 6.0 (dChipLinkge); Perl and Java (CompareLinkage software) Other requirements: None License: None. Any restrictions to use by non-academics: No restrictions Authors' contributions CR, CL and WHW conceived of the study, and participated in its design and coordination. NM and RJHS generated the 5026.10 family data, and MP generated the CR and ER family data. IL implemented the CompareLinkage software and performed the comparative analysis using multiple linkage analysis software packages. JC implemented its graphical user interface (GUI). CL implemented the dChipLinkage module. KH participated in the design and analysis of the study. IL, CR and CL drafted the manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551603.xml
522809	Reconstruction of ancestral protein sequences and its applications	Background Modern-day proteins were selected during long evolutionary history as descendants of ancient life forms. In silico reconstruction of such ancestral protein sequences facilitates our understanding of evolutionary processes, protein classification and biological function. Additionally, reconstructed ancestral protein sequences could serve to fill in sequence space thus aiding remote homology inference. Results We developed ANCESCON, a package for distance-based phylogenetic inference and reconstruction of ancestral protein sequences that takes into account the observed variation of evolutionary rates between positions that more precisely describes the evolution of protein families. To improve the accuracy of evolutionary distance estimation and ancestral sequence reconstruction, two approaches are proposed to estimate position-specific evolutionary rates. Comparisons show that at large evolutionary distances our method gives more accurate ancestral sequence reconstruction than PAML, PHYLIP and PAUP*. We apply the reconstructed ancestral sequences to homology inference and functional site prediction. We show that the usage of hypothetical ancestors together with the present day sequences improves profile-based sequence similarity searches; and that ancestral sequence reconstruction methods can be used to predict positions with functional specificity. Conclusions As a computational tool to reconstruct ancestral protein sequences from a given multiple sequence alignment, ANCESCON shows high accuracy in tests and helps detection of remote homologs and prediction of functional sites. ANCESCON is freely available for non-commercial use. Pre-compiled versions for several platforms can be downloaded from .	Background Present-day protein sequences can be used to reconstruct ancestral sequences based on a model of sequence evolution. Such knowledge about ancestral sequences is helpful for understanding the evolutionary processes as well as the functional aspects of a protein family. Existing methods of ancestral sequence reconstruction can be divided into two main categories: Maximum Parsimony (MP) methods [ 1 , 2 ] and Maximum Likelihood (ML) methods [ 3 - 5 ]. MP methods do not take into account biased substitution patterns between amino acids or different tree branch lengths, and cannot distinguish those equally parsimonious reconstructions [ 3 ]. ML methods do not have these limitations and generally give more reliable results than the MP methods [ 6 ]. Yang et al. [ 3 ] first developed a ML method for ancestral sequence reconstruction. Yang [ 7 ] also made a distinction between "joint" reconstruction and "marginal" reconstruction. Joint reconstruction methods intend to find the most likely set of amino acids for all internal nodes at a site, which yields the maximum joint likelihood of the tree [ 5 ]. Marginal reconstruction compares the probabilities of different amino acids at an internal node at a site and selects the amino acid that yields the maximum likelihood for the tree at that site. Marginal reconstruction can also compute probabilities of all other amino acids for that node [ 4 ]. Koshi and Goldstein [ 4 ] developed a fast dynamic programming algorithm for marginal reconstruction in the framework of Bayesian statistics, while Pupko et al. [ 5 ] proposed a fast algorithm for joint reconstruction. The computational complexities for both algorithms scale linearly with the number of sequences. Both marginal and joint reconstruction algorithms are implemented in our program. All reconstruction methods require a phylogenetic tree inferred from a given alignment. The quality of the tree is crucial for the reliability of reconstruction. A number of methods exist for phylogenetic inference, such as maximum likelihood [ 8 ], distance-based [ 9 ] and parsimony [ 1 ]. Distance-based methods have the advantage of being simple and are able to handle a large set of sequences. They require evolutionary distances estimated for all the sequence pairs. The most common method to infer phylogeny from distances is based on the neighbor-joining algorithm [ 9 ]. Bruno et al. [ 10 ] introduced a distance-based phylogeny reconstruction method called "Weighbor", i.e. "weighted neighbor joining", which takes into account the fact that errors in distance estimates are larger for longer distances. Giving similar results, Weighbor is much faster than ML phylogeny reconstruction. It is also better than other methods such as BIONJ [ 11 ] and parsimony [ 1 ], in aspects of "long branches attract" and "long branch distracts" problems [ 10 ]. Weighbor is used in our program for phylogenetic inference. Overwhelming evidence exists for substitution rate variation across sites [ 12 - 15 ]. For a protein family, rate heterogeneity reflects the selective pressure imposed by folding, stability and function. Gamma distribution is widely used to model the rate variation among sites [ 13 , 16 , 17 ] because of its simplicity. Nielsen [ 18 ] suggested a method for site-by-site estimation of rate factors by a Maximum Likelihood approach. Rate variation among sites has not been taken into account in the early work of ML reconstruction of ancestral sequences [ 4 , 5 ]. Recently, Pupko et al. [ 19 ] introduced rate variation into joint reconstruction by a branch-and-bound algorithm, assuming a gamma distribution of rates among sites. In our package, two methods are proposed to estimate a rate factor for each site. The first one is based on our observation that the substitution rate at a site is correlated with the conservation of the site. The more conserved the site is in a multiple sequence alignment, the smaller its substitution rate is. This empirical method, the result of which we call Alignment-Based rate factors or α AB , relies only on a multiple sequence alignment and a general model of amino acid exchange. The other one is a maximum likelihood method ( α ML ), which requires a tree. In our implementation, we incorporate α AB or α ML in the joint and marginal reconstruction algorithms [ 4 , 5 ]. α AB is also used in the Maximum Likelihood estimation of evolutionary distances [ 20 ] for tree inference. We implement a method of evolutionary simulation that introduces site-specific rate variations in a natural way by imposing structural and functional constraints [ 21 ]. We show by simulations that the reconstruction methods can give reasonable results and that the problem of evolutionary distance underestimation [ 22 ] is alleviated by considering rate variation across sites. Background (or equilibrium) amino acid frequencies ( π ) are usually estimated from the target set of sequences or from large databases of protein families. Background amino acid frequencies estimated from a small dataset tend to have bias, while amino acid frequencies from large databases may not be suitable for the specific protein family under analysis. Here, we propose a ML method to optimize the amino acid frequency vector π . The optimized π vector can give significant improvement over the likelihood of a alignment. Information obtained from ancestral sequence reconstruction is used for two applications: homology detection and prediction of functional sites. For homology detection, ancestral sequences represent an enlargement of the sequence space around native sequences. We demonstrate that adding reconstructed ancestral sequences to a native alignment improves the detection of homologs in database searches. A number of methods have been developed to predict functional sites from amino acid sequences [ 23 , 24 ]. One simple way to infer functional sites is by positional conservation of a multiple sequence alignment [ 25 ]. Lichtarge et al. [ 26 ] proposed a method called evolutionary trace to predict functional sites by analyzing the conservation of sequence subgroups. Functional divergence during the evolutionary process can be reflected in the variation of amino acid usage across different functional subgroups. We propose a new approach that uses information from ancestral sequence reconstruction to identify sites that are well conserved within individual sub-trees but exhibit variability among different sub-trees. By several examples, we show that these sites frequently contribute to the functional specificity of a protein family. Results and discussion We developed a package (ANCESCON) to reconstruct ancestral protein sequences considering rate variation among sites. Rate factors can be estimated either by an empirical method or by a maximum likelihood method. Consideration of rate variation among sites not only improves evolutionary distance estimation, but also gives more accurate ancestral sequence reconstruction. Ancestral sequences are used to improve profile-based sequence similarity searches. We also propose a new approach to predict positions with functional specificity based on the reconstruction of ancestral sequences. Observed α , Alignment Based Rate Factor α ( α AB ) and Rate Factor α estimated by Maximum Likelihood ( α ML ) Evolutionary simulations based on a Z -score model introduce rate variation across sites in a natural way by incorporating structural and functional constraints specific for a protein family [ 21 ]. The simulation procedure is a Monte Carlo simulation of the amino acid substitution process. The fixation of substitutions is dictated by a simple scoring function, which is derived from the template structure and an alignment of its homologs. The number of substitutions occurring at each site can be recorded during the simulation process and the observed α at a site equals the number of recorded substitutions at that site divided by the average substitution number for all sites. To reduce sampling variance, an average observed α vector is calculated from 100 simulations. For the alignment consisting of all the leaf node sequences generated by the simulation process, an α AB vector was calculated according to equation (11) (for details see Methods). An average α AB vector was derived from 100 simulations. Correlation coefficient between the average α AB vector and the average observed α vector was high (data not shown). However, we found that for large observed α values, the corresponding α AB values were smaller. A constant β was introduced to correct this underestimation in equation (11). Here, α i is Alignment-Based rate factor at site i . K is the number of sites in a given alignment. C i is the value assigned to site i (for details see Methods). We optimized the β value by fitting the average α AB vector and average observed α vector to y = x line. Alignments for three different protein families (trypsin, carboxypeptidase and pdz domain) gave a good empirical estimation for β of about 1.3. The relation between this corrected average α AB vector and average observed α vector is shown in Figure 1a for a typical example, the pdz domain (correlation coefficient 0.973). Figure 1 a) Correlation between average α AB and average observed α . b) Correlation between average α ML and average observed α . α AB is Alignment-Based rate factor solely depending on the given alignment. α ML is rate factor estimated by maximum likelihood method, which requires an alignment and evolutionary tree inferred from the alignment. The protein family used here is the PDZ domain. We also estimated an α ML vector for each alignment generated from the simulation (for details see Methods). The average α ML vector shows good correlation with the average observed α vector (Figure 1b ) (correlation coefficient 0.945). α AB or α ML can be incorporated in likelihood calculation in marginal or joint reconstruction. Table 1 shows that improvement of logarithm likelihood of the alignment is significant when α AB or α ML is used. Table 1 Difference of logarithm likelihood and CPU time when using different α vectors α = 1.0 α AB α ML Δ l P* Δ l P* Logarithm Likelihood -5324.56 -5087.72 236.84 <0.0001 -4987.27 337.29 <0.0001 CPU Time (s) + 213 213 359 The alignment tested here is a subset of SH2 family. It includes 44 sequences and each sequence contains 83 amino acids (including gaps). * The likelihood ratio test (LRT) [58] is used to test whether α AB and α ML are significantly different from α = 1.0. The difference in number of free parameters between α AB , α ML and α = 1.0 model is 82. + CPU times were computed on a Dell PowerEdge 8450 server (CPU 700MHz, RAM 8G). Rate variation across sites can be modeled by assuming that the rate factors follow a certain type of statistical distribution. Gamma distribution [ 13 , 27 ] and its discrete approximations [ 28 ] are frequently used for DNA or protein sequences. Rate variation for a protein family reflects different selective pressure at different sites to maintain structure and function. Fewer substitutions are expected to occur in more conserved sites. This hypothesis has prompted us to estimate rate factors ( α AB ) based on sequence conservation in an empirical way. The α AB is compared and calibrated using the observed α as standards. Our method of estimating α ML is similar to the one proposed by Nielson [ 18 ]. One problem with site-by-site rate factor estimation is the small sample size at each site, especially with a small alignment. We have used α AB to eliminate outliers with very large α ML estimates (for details see Methods). Site-specific rate factors improve distance estimation Evolutionary distances tend to be underestimated when rate homogeneity among sites is assumed [ 22 ]. This was tested using the simulation with structural and functional constraints. For the arbitrarily selected tree shown in Figure 2 , we obtained leaf node sequences in the simulation and estimated an evolutionary distance for each sequence pair by Maximum Likelihood, either incorporating α AB or setting α equal to 1.0 (equation (16)). Evolutionary distances were severely underestimated (average underestimation: 0.894) without considering rate variation among sites (Figure 3a ). Introducing α AB in the maximum likelihood method gave more accurate distance estimation (Figure 3b ), although the distances were still underestimated, especially for small distances (average underestimation: 0.286). We believe that more accurate distances will give more accurate phylogeny reconstruction using "Weighbor" [ 10 ]. Since a tree is required to estimate α ML , α ML is not incorporated in estimating evolutionary distance. Figure 2 The tree used to test ancestral sequence reconstruction. This is an arbitrarily selected evolutionary tree. Evolutionary distances are shown to scale. Figure 3 Comparison of pairwise distances between the rebuilt tree and original tree. a) distance estimation assuming no rate variation among sites; b) distance estimation with α AB . The rebuilt tree is inferred from the alignment that is generated by evolutionary simulation performed on the original tree. The original tree is arbitrarily selected. Optimization of equilibrium frequencies A continuous minimization method by simulated annealing was used to optimize the equilibrium frequency vector π , with the objective function being the logarithm likelihood of the alignment. Our π vector optimization program was tested on four alignments, which were taken from the SH2 and SH3 superfamilies in Pfam database (version 7.3) [ 29 ]. Two alignments from the SH2 superfamily have 44 and 87 sequences respectively and both alignment lengths are 83 amino acids (including gaps). The other two alignments from SH3 superfamily have 39 and 94 sequences respectively and both alignment lengths are 57 amino acids (including gaps). For each alignment, we ran optimization 3 times starting from different random initial points. The optimized π vectors did not converge to exactly the same point, but they had a high correlation with each other (always > 0.95) and the difference of logarithm likelihood function values was small (less than 0.1%). The logarithm likelihood of the alignment, using optimized π vector, increased slightly, but significantly (Table 2 ), compared with the logarithm likelihood using the π vector calculated from the alignment. Table 2 Difference of logarithm likelihood and CPU time with and without optimization of π vector α AB & Calculated π α AB & Optimized π Δ l P* Logarithm Likelihood -5087.72 -5055.97 31.75 <0.0001 CPU Time (s) + 213 14902 The alignment tested here is the same alignment used in Table 1. Calculated π means frequency vector calculated from the alignment. * The likelihood ratio test (LRT) [58] is used to test whether optimized π is significantly different from calculated π . The difference in number of free parameters between these two models is 19. + CPU times were computed on a Dell PowerEdge 8450 server (CPU 700MHz, RAM 8G). Optimization of the π vector is time consuming. The running time for reconstruction with or without optimizing π vector is 14,902 seconds and 213 seconds for SH2 alignment (44 sequences), respectively, on a Dell PowerEdge 8450 server (CPU 700MHz, RAM 8G) (Table 2 ). In our program, the default π vector is calculated from the alignment while the user has the option to optimize the π vector for ancestral sequence reconstruction. Testing reconstruction Two different methods for simulations of the evolutionary process were used, as described in Methods, to test the reliability of the reconstruction results. In the first simulation method, starting from a randomly generated root sequence, we simulated the evolutionary process to obtain leaf node sequences based on a tree and a rate matrix. This process was repeated 100 times for a given root sequence R to produce 100 alignments consisting of all leaf node sequences. For each of the 100 alignments, we used the marginal reconstruction method to obtain an amino acid probability vector for each site at the root. To reduce sampling variance, the amino acid probability vector was averaged over the 100 simulation trials. At each site, the amino acid with the highest average probability was chosen as our result of the "reconstructed amino acid" at that site. All "reconstructed amino acids" formed the reconstructed sequences R' . There is no difference between R and R' , that is, the accuracy of reconstruction is 100% for the tree shown in Figure 2 . For each individual simulation and its reconstruction, we checked the amino acid with the highest probability in the reconstructed probability vector of the root. If it is indeed the "reconstructed amino acid", the prediction for that simulation is correct according to the average reconstructed results. The fraction of individual predictions that are correct according to the average reconstructed results is almost always higher than the average probability of the "reconstructed amino acid", suggesting that the average probability of the "reconstructed amino acid" gives a lower estimation of the reconstruction reliability (Figure 4a ). Figure 4 a) Correlation between the average probability of "the reconstructed amino acid" and the fraction of correct predictions. b) Correlation between the fraction of correct predictions and average α AB at each site. The protein family used here is the PDZ domain. Red filled points are sites with incorrect reconstruction. For the second simulation method, we introduced rate heterogeneity across sites with structural and functional constraints [ 21 ]. For the same tree, the accuracy of reconstruction was about 90%. Sites with larger substitution rates are expected to have less reliable reconstructions. Figure 4b shows the relationship between the average α AB and the fraction of individual predictions that are correct according to the "reconstructed amino acid". Sites with incorrect "reconstructed amino acids" all have large α AB values. These values reflect the difficulty of reconstructing sites with large numbers of substitutions. The probabilities of the "reconstructed amino acids" are all small for sites with incorrect reconstructions (less than 0.15), suggesting that the information content of the reconstruction is low. The second simulation method was also used to test ANCESCON along with the reconstruction programs from PAML [ 30 ], PHYLIP [ 31 ] and PAUP* [ 32 ]. All tree topologies used in reconstruction tests were inferred from real alignments. All original root sequences were taken from PDB database [ 33 ]. We had three different types of alignment testing sets. The first testing set used the same tree topology but different root sequences to generate 100 alignments (for details see Methods). The second testing set used the same root sequence but different tree topologies. The third testing set randomly selected a root sequence and a tree topology to generate 100 alignments. After 100 alignments were generated, we reconstructed the root sequence for each alignment and found the consensus root sequence for the 100 reconstructed root sequences. Finally, the consensus root sequence was compared with the original root sequence to calculate the reconstruction accuracy, i.e. the fraction of correctly reconstructed sites for the root sequence. In addition, for the third test, the paired t-test was used to calculate the one-tail probability between ANCESCON and other three methods. In order to make different tree topologies comparable, those trees were scaled to make the average distance from root to all leaf nodes ( d a ) the same for all trees and equal to the tree of pii1 (a signal transduction protein) ( d a = 4.23). If d a was too small (e.g. 0.5), the reconstruction accuracy was always close to 1 for all reconstruction methods used. The value d a = 4.23 was large enough to generate diverse sequences to differentiate 4 different ancestral sequence reconstruction methods. For ANCESCON we had 3 different parameter settings, which included site-specific rate factors estimated by maximum likelihood method ( α ML ), Alignment-Based rate factors ( α AB ) and no rate factors (equal rates among sites). Different parameters were also used for the reconstruction programs from PAML and PHYLIP to find their best reconstructions. For PAML, reconstruction was tested with parameter α (rate factor) estimated from alignment and without α . For PHYLIP, 4 different parameter settings were tried, which were combinations of with/without α estimated from alignment by PAML and with/without branch length dwelling in input tree topology. For PAUP, default settings were used. Table 3 shows a comparison of the reconstruction accuracy for these 4 methods. The reconstruction accuracy of ANCESCON with α ML is higher than the other three methods in almost every test. Also the reconstruction accuracy of ANCESCON with α AB and without α is comparable with PAML and PHYLIP methods and is much better than PAUP. For the first testing set, the best average accuracy for ANCESCON is about 0.5, while the best average reconstruction accuracies for PAML, PHYLIP and PAUP* are 0.45, 0.39 and 0.32 respectively. Testing set 2 and 3 produce similar results. Using the paired t-test in the third testing set, we show that ANCESCON method with α ML gives significantly better reconstruction than the other 3 methods. Because the site-specific α ML is very close to the true mutation rate at a site (Figure 1b ), using the site-specific α ML can improve our ability to reconstruct the amino acids for ancestral sequences correctly. These reconstruction tests suggest that ANSCESCON may be a better tool to reconstruct ancestral sequences compared to PAML, PHYLIP and PAUP* if the given alignment contains more diverse sequences. Table 3 Ancestral sequence reconstruction accuracy by different programs Root Seq. Tree Leaf Node Num. Methods ANCESCON PAML PHYLIP $ PAUP* α ML α AB - α + α - α +L + α -L + α +L - α -L - α 1em2 pii1 25 0.45 0.32 0.35 0.41 0.37 0.29 0.27 0.21 0.29 0.26 1g9o pii1 25 0.56 0.46 0.47 0.53 0.53 0.51 0.54 0.40 0.51 0.47 1rgg pii1 25 0.60 0.42 0.47 0.60 0.62 0.47 0.58 0.32 0.56 0.47 1sgt pii1 25 0.38 0.34 0.33 0.33 0.32 0.32 0.33 0.27 0.33 0.32 1zm2 pii1 25 0.33 0.29 0.3 0.28 0.25 0.21 0.25 0.21 0.27 0.16 2a8v pii1 25 0.62 0.45 0.42 0.56 0.55 0.44 0.46 0.28 0.50 0.36 2ctb pii1 25 0.53 0.40 0.39 0.41 0.38 0.24 0.24 0.21 0.29 0.22 Average accuracy 0.496 0.383 0.390 0.446 0.431 0.354 0.381 0.271 0.393 0.323 2ctb gef 27 0.54 0.37 0.38 0.35 0.35 0.29 0.17 0.24 0.22 0.22 2ctb LacI 54 0.66 0.64 0.57 0.44 0.37 0.49 0.35 0.42 0.33 0.34 2ctb pdz 39 0.54 0.41 0.42 0.44 0.39 0.22 0.34 0.18 0.32 0.22 2ctb ph 30 0.79 0.74 0.75 0.53 0.55 0.45 0.25 0.43 0.37 0.32 2ctb pii1 25 0.53 0.40 0.39 0.41 0.38 0.24 0.24 0.21 0.29 0.22 2ctb ptb 29 0.58 0.39 0.43 0.39 0.38 0.29 0.23 0.26 0.24 0.23 2ctb sh2 34 0.61 0.42 0.40 0.43 0.40 0.30 0.22 0.20 0.27 0.22 2ctb sh3 43 0.83 0.82 0.80 0.62 0.55 0.69 0.45 0.66 0.46 0.54 2ctb GST 140 0.76 0.73 0.73 @ @ # # 0.47 0.38 0.33 Average accuracy & 0.635 0.524 0.518 0.451 0.421 0.371 0.281 0.325 0.313 0.289 1em2 pdz 39 0.45 0.35 0.36 0.44 0.44 0.29 0.43 0.23 0.4 0.24 1g9o pii1 25 0.56 0.46 0.47 0.53 0.53 0.51 0.54 0.40 0.51 0.47 1rgg sh2 34 0.64 0.48 0.46 0.61 0.61 0.56 0.59 0.34 0.6 0.41 1sgt gef 27 0.49 0.39 0.40 0.48 0.44 0.42 0.44 0.36 0.45 0.41 1zm2 ptb 29 0.66 0.47 0.48 0.57 0.57 0.53 0.51 0.32 0.52 0.41 2a8v ph 30 0.81 0.78 0.81 0.71 0.74 0.60 0.61 0.50 0.65 0.50 2ctb LacI 54 0.66 0.64 0.57 0.44 0.37 0.49 0.35 0.42 0.33 0.34 Average accuracy 0.610 0.510 0.507 0.540 0.529 0.486 0.496 0.367 0.494 0.397 Probability Δ 0.0026 0.0023 0.0248 0.0328 0.0007 0.0168 0.0001 0.0143 0.0005 All root sequences are taken from PDB database and the names listed in the table are PDB IDs. Tree topologies for gef (guanine nucleotide exchange factor), LacI (PurR/LacI family of bacterial transcription factors), pdz, ph, pii1 (a signal transduction protein), ptb, sh2, sh3 and GST (glutathione S-transferase) are inferred from multiple sequence alignments chosen from Pfam database (version 7.3). All tree topologies are generated from real alignments and the distances are rescaled in order to make the trees comparable. The value in this table represents the accuracy of reconstruction, i.e. the fraction of correctly reconstructed sites for the root sequence. The best reconstruction accuracy in each test is shown in bold. α ML means that the site-specific rate factors were estimated by maximum likelihood method. α AB means that the site-specific rate factors were estimated by our empirical equation based on the given alignment (for details see Methods). - α means that the rate factors were not considered in reconstruction. + α means that the rate factors were considered in reconstruction. +L means that branch lengths of the input tree were used in reconstruction, while -L means that branch lengths were estimated by the reconstruction program itself. @: tree topology for GST had 140 leaf nodes that were too many for PAML to run through. $ : rate factors estimated by PAML were used by PHYLIP in ancestral sequence reconstruction. #: tree topology for GST had 140 leaf nodes, which were too many for PAML to estimate rate factors for GST. & :GST is excluded in calculation of the average. Δ : paired t-test method [40] was used to estimate the one-tail probability between ANCESCON and the other three reconstruction methods. Ancestral sequences used in homology detection Thirty-eight OB (Oligonucleotide/oligosaccharide binding)-fold [ 34 ] proteins and ten other alignments (adenylyl kinase, gef, globin, pdz, ph, ptb, ras, sh2, sh3 and subtilase) from the Pfam database (version 7.3) [ 29 ] were chosen to perform homology detection tests. Given an alignment with N sequences, we had four different methods, "BEST", "SECOND BEST", "SHUFFLE" and "RANDOM", to generate another N -1 sequences (for details see Methods). For each combined alignment (2 N -1 sequences), PSI-BLAST [ 35 ] searches were performed starting from each sequence and seeded with the combined alignment (-B option in the program BLASTPGP, e-value cutoff 0.01), and all found hits were pooled together. The benchmark experiment was PSI-BLAST seeded with the native alignment ( N sequences). For each type of the four combined alignments, we checked hits not found by the native alignments. New hits were verified to be true positives or false positives by running PSI-BLAST or HMMER [ 36 ], followed by manual inspections. Using the 48 native alignments, a total of 13973 hits were found by the benchmark. Compared to the benchmark, the "BEST" method detected 120 new homologs and the other three methods found 69, 74 and 9 new homologs, respectively (Figure 5 ). Among those new homologs, "BEST", "SECOND BEST", "SHUFFLE" and "RANDOM" methods had 3, 2, 6 and 3 false positives, respectively (Figure 5 ). Also, "BEST", "SECOND BEST", "SHUFFLE" and "RANDOM" methods missed 61, 1070, 60 and 7811 homologs as compared to the benchmark. Figure 5 Comparison of "BEST", "SECOND BEST", "SHUFFLE" and "RANDOM" methods in the number of new homologs detected when compared with the benchmark experiment. The methods are defined in "Methods" section. The blue portion of the bar shows the number of true positives. The red portion of the bar shows the number of the false positives. Adding non-native sequences to the native alignment results in a change of sequence profile for PSI-BLAST searches. Random sequences can dilute the position-specific amino acid exchange characteristics of native alignments. This effect should not improve the profile. Indeed, few new homologs are found by the "RANDOM" method. However, sequences generated by shuffling each position of the native alignment have the same conservation properties as the native alignment, and the "SHUFFLE" method detects a total of 74 new homologs. Two effects may account for this finding. First, addition of shuffled sequences to the native alignment can slightly change the estimates of pseudocount frequencies of amino acids and thus the position specific scoring matrix [ 35 ]. Second, the new version of PSI-BLAST program uses composition-based statistics with e-value estimation related to the composition of the query sequence [ 37 ]. Each shuffled sequence has its own amino acid composition that is different from the native sequences. This difference can affect the e-values of hits. The "BEST" method detects the most number of new homologs, suggesting that the reconstructed ancestral sequences resemble the native sequences. Ancestral sequences may therefore be more similar to some remote homologs than to the native sequences. The "SECOND BEST" method detects less new homologs than the "BEST" method but more than the "RANDOM" method, suggesting that the second most probable amino acids in reconstruction can still reflect some properties of native sequences. Table 4 shows homology detection results of OB-fold structures using reconstructed ancestral sequences. Table 4 Homology detection results of OB-fold structures using reconstructed ancestral sequences SCOP Superfamily/family PDB structure New homologs NCBI annotation Nucleic acid-binding proteins/ Anticodon-binding domain 1b7yB, 39–151 N/A - 1b8aA, 1–102 N/A - 1bbuA, 64–151 13431467 DNA polymerase II small subunit 15598836 DNA polymerase III, alpha chain 1c0aA, 1–106 11261591 DNA polymerase III, alpha chain 11499379 conserved hypothetical protein 1169392 DNA polymerase III alpha subunit 118794 DNA polymerase III alpha subunit 13620707 putative DNA polymerase III, alpha chain 14194684 DNA polymerase III alpha subunit 14194702 DNA polymerase III alpha subunit 14195653 DNA polymerase III alpha subunit 14195659 DNA polymerase III alpha subunit 15594924 DNA polymerase III, subunit alpha 15598836 DNA polymerase III, alpha chain 15601899 DnaE 15642243 DNA polymerase III, alpha subunit 15669005 M. jannaschii predicted coding region MJ0818 15679404 DNA polymerase delta small subunit 3914611 ATP-dependent DNA helicase recG 1cuk, 1–64 N/A - 1e1oA, 64–148 11261591 DNA polymerase III, alpha chain XF0204 14194684 DNA polymerase III alpha subunit 1fguA, 181–298 15219507 hypothetical protein 15230563 putative protein 15790309 Vng1255c from Halobacterium sp. 6166145 DNA polymerase III alpha subunit 8778702 T1N15.20 1fl0A 10957481 hypothetical protein 1g51A, 1–104 14520587 hypothetical protein 14591565 hypothetical protein 15595886 hypothetical protein 3914638 ATP-dependent DNA helicase recG 1otcB, 36–126 N/A - 1quqA, 62–152 15387767 probable replication protein a 28 Kd subunit 1qvcA, 1–114 N/A - Nucleic acid-binding proteins/Cold shock DNA-binding domain like 1a62, 48–125 N/A - 1ah9 N/A - 1bkb, 75–139 15790688 translation initiation factor eIF-5A; Eif5a 1c9oA 6014735 Cold shock protein CspSt 1csp N/A - 1d7qA N/A - 1mjc N/A - 1rl2 N/A - 1sro 15671445 N utilization substance protein A 15794781 N utilisation substance protein A 15803711 transcription pausing; L factor 2eifA, 73–132 N/A - Nucleic acid-binding proteins/DNA ligase, mRNA capping enzyme, domain2 1a0i, 241–349 N/A - 1dgsA, 315–400 N/A - 1ckmA, 238–302 N/A - 1fviA, 190–293 N/A - Nucleic acid-binding proteins/Phage ssDNA-binding proteins 1gpc N/A - 1gvp N/A - 1pfs N/A - Nucleic acid-binding proteins/RNA polymerase subunit RBP8 1a1d N/A - Staphylococcal nuclease/Staphylococcal nuclease 1eyd 13422779 aldose 1-epimerase * Bacterial enterotoxins/Bacterial AB5 toxins, B units 1c4qA N/A - 1prtF N/A - Bacterial enterotoxins/Superantigen toxins 1an8, 19–94 N/A - TIMP-like/Tissue inhibitor of metalloproteases 1ueaB, 14–106 N/A - Inorganic pyrophosphatase/ Inorganic pyrophosphatase 2prd N/A - MOP-like/BiMOP, duplicated molybdate-binding domain 1b9mA, 127–262 10639288 probable ATP-binding protein 10955070 AgtA 1175513 Putative ferric transport ATP-binding protein afuC 15598450 probable ATP-binding component of ABC transporter 3978166 ATPase FbpC 4895001 glucose ABC transporter ATPase * Histidine kinase CheA, C-terminal domain/ Histidine kinase CheA, C-terminal domain 1b3qA, 540–671 N/A - * Putative false positives as assessed by manual inspection. Prediction of functional sites Ten well-studied protein families (adenylyl kinase, gef, globin, pdz, ph, ptb, ras, sh2, sh3 and subtilase) from the Pfam database (version 7.3) [ 29 ] were selected to test the prediction of functional sites. To define functional sites, we considered residues falling within 5Å of any ligand to be functionally important (i.e. AP5 for adenylyl kinase). As a simple quantification of prediction accuracy, we counted the number of predictions that lie within 5Å from the ligands and consider these sites to be true positives. Our method intends to identify those sites with high similarity within individual sub-trees and high variation among sub-trees. These sites are likely to contribute to functional specificity. Based on a tree partition and the reconstructions at the cutting nodes (details see Methods), we have developed a measure called specificity score (equation (27)). We expect that both highly variable sites and highly conserved sites tend to score low in our method. Ten top-ranking sites were selected as our predicted functional sites for each family. For comparison, we also implemented a simple conservation (SC) method [ 25 ], the evolutionary trace (ET) method [ 26 ] and the conservation difference (CD) method [ 21 ] on the 10 protein families. The results are shown in Table 5 . Here, the results from these three methods tend to include invariant or highly conserved sites while the result from our method scores those sites low. Still, the number of true positives of our method is comparable to other methods for several families. For some protein families, such as gef, pdz and subtilase, our method predicts no fewer functional residues than the other three methods. Table 5 Comparison of the true hits among the top 10 predicted sites for ANCESCON, evolutionary trace (ET), simple conservation (SC), and conservation difference (CD) methods Protein Family PDB ID # Ligand/ substrate Number of sites * * ANCESCON ET SC CD adkinase 1aky AP5 188 42 20 18 3 9.5 9.1 8 gef 1bkd H-Ras 245 47 4 0 3 3 3 2 globin 1a6g HEM 147 21 1 1 2 5.5 6 6 pdz 1be9 + 81 15 2 1 6 4 4 2 ph 1mai I3P 109 11 2 0 2 2 3 2 ptb 1shc PTR 157 27 2 1 6 5 5 9 ras 821p GTN 185 29 10 9 2 5.6 8.7 5 sh2 1a09 ACE 83 17 2 1 3 5 4 4 sh3 1nlo ACE 57 9 1 1 2 5 4 0 subtilase 1av7 SBL 278 22 8 4 5 4.6 3.8 4 #: Representative protein structure : Number of sites within 5Å to ligand or substrates : Number of invariant sites, which may contain gaps *: Number of invariant sites within 5 Å to ligand or substrates +: C-terminal peptide of protein CRIPT Figure 6 shows the mapping of our predictions on the structure for the PDZ domain family. In green color is the ligand and in red color are the functional residues predicted by our method. Six of the predicted residues are within 5Å to the peptide ligand. Nine of the predicted residues are around the ligand binding area. Only one is distant from the ligand (Figure 6 ). Figure 6 Mapping top 10 predictions by ANCESCON to PDZ domain (PDB ID: 1be9) [50]. The color code scheme: ligand is shown in green and the predicted functional residues are shown in red. Another example is the family of adenylyl kinases. Our method identified 3 residues within 5 Å to the ligand while the other 3 methods identified more such residues, most of which are in highly conserved positions such as the catalytic residues. Highly conserved residues, however, are not selected by our method since our measure is designed to emphasize on sites contributing to specificity. Figure 7 shows the N-terminal part of the alignment of adenylyl kinases, with our predictions highlighted in red and orange colors. The evolutionary tree for the alignment is shown in Figure 8 . The first cutting layer (for details see Methods) results in two well-separated sub-trees. Functional annotations suggest that they contain enzymes with different substrate preferences: adenylyl kinases and uridylate kinases, respectively. Three residues (27, 54 and 89) from our predictions (red colored in Figure 9 ) contribute to substrate-binding specificity, as have been noted before in the structural studies of the UMP kinases [ 38 ]. Figure 9 highlights our predicted functional residues on the adenylyl kinase protein structure. Most of our predictions fall within the specificity pocket. Figure 7 A partial alignment of the N-terminal part of adenylyl kinases. Sites colored in red are our predictions that are within 5Å from the ligand. Sites colored in orange are our predictions more than 5Å apart from the ligand. Figure 8 The evolutionary tree for the adenylyl kinase family generated by "Weighbor". The first cutting layer is shown. Evolutionary distances are shown to scale. Figure 9 Mapping top 10 predictions by ANCESCON to adenylyl kinase domain (PDB ID: 1aky) [47]. The color code scheme: ligand is shown in green and the predicted functional residues are shown in red. Conclusions We developed a package (ANCESCON) to reconstruct ancestral protein sequences that takes into account the variation of substitution rates among sites. Two methods were proposed to estimate site-specific evolutionary rates ( α ), namely Alignment-Based rate factor ( α AB ) and rate factor α estimated by maximum likelihood ( α ML ). Consideration of rate variation among sites can alleviate the underestimation of evolutionary distances. Accuracy of ancestral sequence reconstruction by our method is higher than that of PAML, PHYLIP and PAUP when the given alignment contains more diverse sequences. We show that reconstructed ancestral sequences help to improve detection of distant homologs and prediction of functional sites with specificity. Methods Transition probability and likelihood calculations For all models discussed in this paper, we assume all sites in an alignment evolve independently and according to a homogeneous, stationary and time reversible Markov process. The probability of an amino acid i to be replaced by amino acid j after a time interval t is P ij ( t ). The transition probability matrix of 20 amino acids is written as P ( t ), which can be calculated as P ( t ) = exp( Q t ) (2) Here, Q is the rate matrix. The non-diagonal elements q ij are the instantaneous rates of change from amino acid i to amino acid j and diagonal elements q ii are such that each matrix row sums up to 0. Q can be calculated by: Q = S * diag ( π ) (3) S is the matrix of amino acid exchangeability parameters [ 39 ]. π i is the equilibrium frequency for amino acid i . Time reversibility implies that S is a symmetric matrix. In our program, the S matrix is taken from Whelan and Goldman [ 39 ] and the default π vector is estimated from the given alignment. Q can be decomposed into eigenvalues ( λ i ) and eigenvectors ( u i ). U = ( u 1 , ..., u 20 ) (5) P ij ( t ) can be calculated using the following equation, The likelihood function [ 8 ] for an evolutionary tree T shown in Figure 10 is: Figure 10 An evolutionary tree topology. Nodes C, D, E and F represent given protein sequences, while nodes A and B represent ancestral protein sequences, i.e. unknown sequences. d YZ represents the evolutionary distance between nodes Y and Z. Here, is the equilibrium frequency of the amino acid at the node A. is the transition probability from the amino acid at node A to the amino acid at node B after an evolutionary distance d AB . Considering that each site i has a rate factor α i [ 13 , 18 ], we have: t in equation (6) can be expressed as: t = α · d (9) d is the evolutionary distance and α is rate factor. The following restriction on the vector α holds: Here, K is the number of sites. Alignment-Based Rate Factor α ( α AB ) and Rate factor α estimated by Maximum Likelihood ( α ML ) Our program supports two methods to estimate a rate factor for each site: Alignment-Based rate factor α ( α AB ) and Maximum Likelihood-estimated rate factor α ( α ML ). The estimation of α AB is empirical and based on the observation that the substitution rate at a site is correlated with the conservation of the site, which, in turn, is correlated with the average transition probability among the amino acids at that site. Conserved sites are dominated by highly similar amino acids and thus have high average transition probabilities among the amino acids. The algorithm to calculate α AB is as follows: 1. Set t equal to 1.0 and use equation (6) to calculate a transition probability matrix P for 20 amino acids. Equation, , is used to compute a symmetric matrix P' . 2. Calculate the average transition probability for each site and take the reciprocal: , where is the number of non-gapped amino acid pairs in site i and the denominator is the sum over the transition probabilities between all amino acid pairs ( j , k ) at a site i . 3. For invariant sites, C i is set to 0 to make it consistent with the Maximum Likelihood estimation. 4. Equation (11) is used to calculate α AB , so that equation (10) holds. If an evolutionary tree is assumed for the alignment, we can estimate the α ML factors by maximizing the likelihood (equation (8)) for each site: If some sites are highly variable, the α ML at those sites can be very large, as has been previously noticed [ 18 ]. We consider these rate factors to be outliers. For these sites, we have observed that likelihood changes very little over a wide range of the α values. An empirical method is used to reduce the values of α ML outliers, guided by the α AB values. a Z -score of the ratio of α ML to α AB is calculated for each site except invariant sites: Here, is the ratio of to for site i ; is the number of sites excluding the invariant sites. If Z i is greater than 3, it is reduced to 3 by decreasing the value of . We repeat this procedure until no Z i for any site i is greater than 3. After removing the outliers, we scale the values so that equation (10) holds. Amino acid frequency vector π optimization Two methods are implemented to estimate the equilibrium frequency vector π , one derived directly from the given alignment (Alignment-Based π or π AB ) and the other estimated by Maximum Likelihood ( π ML ). The likelihood for the entire alignment is a function of π with 19 variables. A continuous minimization method by simulated annealing [ 40 ] is used to optimize π , with the objective function being the logarithm likelihood of the alignment. The simulated annealing is computationally intensive and is the major reason for the long CPU time given in Table 2 . Distance matrix calculation and tree inference A Maximum Likelihood approach is used to estimate the evolutionary distances among sequences, either considering rate variation across sites or not. The logarithm likelihood for replacing one protein sequence (A) with another protein sequence (B) after an evolutionary distance d can be written as: Here, is the equilibrium frequency for the amino acid at site j in sequence A. is the transition probability from amino acid at site j in sequence A to amino acid at site j in sequence B after an evolutionary distance α j · d . α j is 1 if all sites are assumed to evolve at the same rate; otherwise the α AB at site j is used for α j . An estimate of the evolutionary distance between two sequences is obtained by maximizing the likelihood function of equation (15): Equation (16) can be solved by the bisection root-finding method [ 40 ]. After the distance matrix is calculated, the "Weighbor" method, i.e. weighted neighbor joining, is used to infer an evolutionary tree [ 10 ]. Ancestral sequence reconstruction Two methods are implemented to reconstruct ancestral sequences. One is a marginal reconstruction method [ 4 ], and the other is a joint reconstruction method [ 5 ]. Below are their brief descriptions. The marginal reconstruction method [ 4 ] We calculate P ( A r \|{ A l } T ), which is the conditional probability of amino acid A r at the root, given leaf node amino acid set { A l } and a tree T . Since time reversibility is assumed, any internal node can serve as a root. Using Bayes' theorem, we have: Here, P ( A r ) is used here instead of P ( A r \| T ) because the frequency of the root amino acid A r , i.e. π r , does not depend on tree T . P ({ A l }\| A r T ) is the conditional probability of the known amino acids at the leaf nodes, given T and A r . P ({ A l }\| T ) does not depend on A r , so it is calculated as a normalization constant for P ( A r \|{ A l }, T ) terms over all 20 possible values of A r to make the sum equal to 1. For Figure 10 , P ({ A l }\| A r T ) can be expanded as: Here, is the transition probability from amino acid at node A to amino acid at node B after an evolutionary distance d AB . Equation (18) can be calculated using a recursive method suggested by Felsenstein [ 8 ]. If rate factors are used in the reconstruction of the root sequence, we have: Here, α i could be either α AB or α ML at site i . P ( A C , A D , A E , A F \| A A , T ) i is the conditional probability P ( A C , A D , A E , A F \| A A , T ) at site i . The joint reconstruction method [ 5 ] The objective of a joint reconstruction method is to find the combination of amino acids for an internal node set { A i } that maximize the conditional probability of this amino acid combination, given the leaf node amino acid set { A l } and a tree T , P ({ A i }\|{ A l }, T ). Using the Bayes' theorem, we have: Because P ({ A l }\| T ) is the same for all amino acid combination at internal node set { A i } this problem becomes finding the maximum of P ({ A l }\|{ A i }, T ) * P ({ A i }). The details of a fast algorithm to solve equation (20) can be found in Pupko et al. [ 5 ]. We also incorporated site-specific rate factors in this algorithm, in a similar way as equation (19) Gaps Due to difficulties with the probabilistic models of gaps, a simplified empirical approach is used to alleviate the problem. We assume that gaps are "supersede" letters. Gaps are considered for each site independently. If a leaf node has a gap instead of an amino acid at a site, this node will be deleted from the tree for this site. After dealing with leaves, we check all internal nodes for children. If an internal node has no children or only one child due to the leaf removal because of gaps, it will be removed from the tree and a gap will be assumed as its reconstructed state. Simulations of evolutionary process Two methods of simulating amino acid substitution process were used to test the reliability of reconstruction, rate factors and evolutionary distance estimation. The first simulation method was based on a homogeneous time reversible Markov model. The parameters from Whelan and Goldman [ 39 ] were chosen for our model, including the equilibrium frequency vector π and the S matrix. Given the length of a branch from a parent node to one of its child nodes and the amino acid for the parent node, we simulated the substitution process to generate an amino acid for the child node based on the transition probabilities that were calculated using equation (6). For the arbitrarily selected tree shown in Figure 2 , we first generated a random sequence of 100 amino acids as the root sequence based on the amino acid frequencies from Whelan and Goldman [ 39 ]. We then simulated the random substitution process to obtain all leaf node sequences. This simulation was repeated 100 times. The resulting 100 alignments were used to test the reliability of the reconstruction result. In this simulation, each site evolved independently according to the same tree topology and branch lengths, thus there was no rate heterogeneity across sites. The second simulation method, based on a Z -score model, introduced rate variation across sites by using structural and functional information for a specific protein family [ 21 ]. We selected three protein families for the Z -score simulations under structural and functional constraints: pdz domain (Protein DataBank (PDB) ID: 1g9o) [ 41 ], trypsin (PDB ID: 1sgt) [ 42 ] and carboxypeptidase A (PDB ID: 2ctb) [ 43 ]. Given a rooted tree, the native sequence with known structure was used as the root sequence. Simulations were made along the tree to generate sequences at any internal node or leaf node. If the evolutionary distance from a parent node to a child node was d , the child sequence was obtained after l * d accepted substitutions starting from the parent sequence, where l is protein sequence length. Simulations of the substitution process were repeated 100 times. For each site, the number of accepted substitutions was recorded and averaged over 100 simulations. Rate factors (observed α ), representing site mutability, were calculated from these average substitution numbers, such that the average of rate factors is 1 (equation (10)). 100 simulated alignments were used to test the rate factor estimators ( α AB and α ML ), distance calculation methods and ancestral sequence reconstruction. Homology detection Testing dataset 38 OB (Oligonucleotide/oligosaccharide binding)-fold [ 34 ] proteins with known structures were selected for homology detection test. OB-fold has a 5-stranded β -barrel structure. In the SCOP (Structure Classification of Proteins) database (version 1.55) [ 44 ], there are 7 OB-fold superfamilies. The superfamily of nucleic acid binding proteins is the most populated. Diversity of many OB-fold homologs extends beyond detection by automatic PSI-BLAST searches. Multiple sequence alignments of native sequences were obtained from PSI-BLAST searches starting from the 38 OB-fold sequences with known structures. We also selected 10 alignments (adenylyl kinase, gef, globin, pdz, ph, ptb, ras, sh2, sh3 and subtilase) from the Pfam database (version 7.3) [ 29 ] for homology detection test. Four different methods For each alignment with N sequences, ancestral sequences for the N -1 internal nodes were reconstructed. The idea is to test whether adding more sequences to a native alignment can help homology detection. Four types of combined alignments were generated, adding different sets of N -1 sequences to the native alignment. In the first case, the added sequence at each internal node consisted of amino acids with the largest probability at each position. In the second case, the added sequences were made up of amino acids with the second largest probability. In the third case, we shuffled the native alignment at each position while keeping the gap pattern as in the native alignment. After shuffling, we added N -1 sequences resulted from the shuffling to the native alignment. In the fourth case, N -1 random sequences were generated with the overall amino acid frequencies of the native alignment. These four methods are named "BEST", "SECOND BEST", "SHUFFLE" and "RANDOM", respectively. Prediction of functional sites Our objective is to find sites that are well conserved within each sub-tree, but show high variability between different sub-trees. These sites are likely to contribute to functional specificity [ 26 , 45 , 46 ]. Sequence datasets Multiple sequence alignments of ten protein families were chosen from the Pfam database (version 7.3) [ 29 ]. These families are: adenylyl kinase (adkinase) (representing structure PDB ID: 1aky; its ligand or substrate: AP5) [ 47 ], guanine nucleotide exchange factor (gef) (1bkd; H-Ras) [ 48 ], globin (1a6g; HEM) [ 49 ], pdz domain (1be9; C-terminal peptide of protein CRIPT) [ 50 ], ph domain (1mai; I3P) [ 51 ], ptb domain (1shc; PTR) [ 52 ], ras (821p; GTN) [ 53 ], sh2 domain (1a09; ACE) [ 54 ], sh3 domain (1nlo; ACE) [ 55 ] and subtilase (1av7; SBL) [ 56 ]. Most of these alignments contain many sequences. We pruned and clustered the sequences in each alignment according to the length and diversity. Representative sequences were kept and used for tree inference and ancestral sequence reconstruction. This procedure was done in three steps: 1) removing fragments, 2) single-linkage clustering and 3) complete-linkage clustering, as described below. 1. For each family, there is a template sequence with known structure. The sequences, which cover less than 75% of the non-gapped positions in the template sequence with amino acids, were considered to be fragments and discarded. 2. A sequence identity matrix was calculated for the remaining sequences. A single linkage clustering was done to form sequence groups at sequence identity threshold 0.8. For each group, we chose the longest sequence as a representative, discarding other members. This step reduced redundancy in the dataset. 3. An average sequence identity was calculated for the remaining sequences. We used this average identity as a threshold for complete linkage clustering to form new sequence groups. Four groups with the largest sequence numbers were chosen to form our new alignment. Any group with the same number of sequences as the fourth group was also included in the new alignment. The purpose of this step is to keep the major sequence subgroups of a family while leaving out highly divergent sequences that might be deleterious for tree inference. Rooting The "Weighbor" method gives an unrooted tree. For our purpose of predicting functional sites, we need to find a point on the tree that serves as the root. We used a least-squares modification of the midpoint rooting procedure to define the root [ 57 ]. Tree partitioning The tree was partitioned into sub-trees at several levels and compared the amino acid usages within each sub-tree and among the sub-trees. For this partitioning, we "cut" the tree into a fixed number of equal-distanced layers, using the midpoint as the root (Figure 11 ). Several criteria were tried for selecting the distance between adjacent layers. Empirically we found that a simple partition of the tree into 5 layers usually gave the best results. If the average distance from the root to all leaf nodes is d r , then the distance between adjacent layers is d r /5 (Figure 11 ). Each place of a "cut" between the layers corresponds to a certain ancestral sequence. We term the location of a "cut" as a "cutting" node. The marginal reconstruction method was used to reconstruct amino acid probability vectors for all the cutting nodes (Figure 11 ). The reconstructed probability vector of a cutting node reflects the amino acid usages of the sub-tree under it. Figure 11 An example showing the different cutting layers in a rooted tree. d r is the average distance from the root to all leaf nodes. Nodes i and j are neighboring cutting nodes. Calculating specificity score for each site We use { L K } to represent the set of cutting nodes for layer L K , K = 0,1,5. { L 0 } is the root and L 1 is the closest layer to the root, etc. A dissimilarity score between any neighboring cutting node pair is calculated. The definition of a neighboring cutting node pair ( i , j ) (Figure 11 ) is: 1. i ∈ { L K } 2. j ∈ { L K +1 } 3. Node i is an ancestor of node j (all points on the path from j to root node are ancestors of node j ), so that the distance between i and j is exactly d r /5 . Each cutting node has only one ancestral cutting node neighbor. The dissimilarity score for cutting node j and its ancestral cutting node neighbor i, i.e. anc ( j ), at site m is defined as: and are the reconstructed probabilities of amino acid A at cutting node j and its ancestral cutting node neighbor i ( anc ( j )), respectively. Let , K = 1,...,5 (22) Here, is the average dissimilarity score for layer K . N K is the number of cutting nodes in layer K . The specificity score is defined as: reflects the difference of amino acid compositions among the major sub-trees defined by the first layer. to reflect the average difference of amino acid compositions within each sub-tree. If the amino acids are highly conserved within each sub-tree but show variability among the sub-trees, to are small and is large, leading to a large value of S m . We set S m to 0 for invariant sites. We sort the sites by their specificity scores and choose the 10 top scoring sites as our predicted functional sites. Those predicted functional sites that lie within 5 Å from the ligand(s) are considered to be true positives. Comparison with other methods We compared our method with three other methods for prediction of functional sites. The first method (Simple Conservation or SC) is based on sequence conservation. Highly conserved sites are considered to be functional. For each family, we sorted the sites by positional conservation [ 25 ] and chose the 10 top-ranking sites as the predictions. There might be ties for sites. For example, if there were 5 sites tied at the tenth conservation value and only one of them was within 5Å from the ligand(s), then its contribution to the total number of "correct predictions" was 1/5. The second method is the evolutionary trace (ET) method [ 26 ], which partitions a sequence identity dendrogram into sub-trees at varying sequence identity thresholds. Sites that are invariant within each individual sub-tree are picked as functional sites. A higher identity threshold gives rise to more sub-trees and, since conserved sites are more frequent in the sub-trees with smaller sizes, lead to more predicted sites. ET analysis was performed from a low identity threshold to higher thresholds until the number of predicted sites was 10 or just above 10 (in the cases of ties). Ties were resolved similarly to the simple conservation method. The third method (conservation difference or CD) is based on the conservation differences between a native alignment and an alignment derived from the Z -score sequence design [ 21 ]. The basic idea was to differentiate sites conserved due to structural stability and sites conserved due to function. Since the pairwise potential in the Z -score design tends to weaken the conservation caused by function, functionally conserved sites tend to have a large conservation difference between the native alignment and the alignment of designed sequences. We chose 10 top ranking sites sorted by conservation difference as predictions by CD. Authors' contributions NVG conceived and initiated the study. All authors took part in developing methods and designing experiments. WC wrote the source code and JP analyzed the data. All authors read and approved the final manuscripts.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522809.xml
543575	Respiratory symptoms in relation to residential coal burning and environmental tobacco smoke among early adolescents in Wuhan, China: a cross-sectional study	Background Cigarette smoking and coal burning are the primary sources of indoor air pollution in Chinese households. However, effects of these exposures on Chinese children's respiratory health are not well characterized. Methods Seventh grade students (N = 5051) from 22 randomly selected schools in the greater metropolitan area of Wuhan, China, completed an in-class self-administered questionnaire on their respiratory health and home environment. Results Coal burning for cooking and/or heating increased odds of wheezing with colds [odds ratio (OR) = 1.57, 95% confidence interval (CI): 1.07–2.29] and without colds (OR = 1.44, 95% CI: 1.05–1.97). For smoking in the home, the strongest associations were seen for cough (OR = 1.74, 95% CI: 1.17–2.60) and phlegm production (OR = 2.25, 95% CI: 1.36–3.72) without colds among children who lived with two or more smokers. Conclusions Chinese children living with smokers or in coal-burning homes are at increased risk for respiratory impairment. While economic development in China may decrease coal burning by providing cleaner fuels for household energy use, the increasing prevalence of cigarette smoking is a growing public health concern due to its effects on children. Adverse effects of tobacco smoke exposure were seen despite the low rates of maternal smoking (3.6%) in this population.	Background Residential coal burning and cigarette smoking are the most common sources of indoor air pollution in Chinese households [ 1 - 3 ]. Although use of coal stoves and smoking have been linked to respiratory morbidity among adult populations in mainland China [ 1 , 4 - 6 ], little is known about how these exposures affect Chinese children's respiratory health. Children's developing lungs are especially vulnerable to indoor air pollution because children spend much of their time indoors at home [ 7 , 8 ]. Coal has been widely used for cooking and heating in China [ 1 , 9 ]. Domestic coal stoves and boilers produce high indoor concentrations of sulfur dioxide (SO 2 ), carbon monoxide (CO), particulate matter, and other pollutants [ 2 , 3 , 10 , 11 ]. Some studies suggest that indoor concentrations of emitted pollutants may exceed international guidelines and national ambient air pollution standards in China [ 1 , 2 ]. An increasing proportion of Chinese children are exposed to tobacco smoke because smoking prevalence in China has increased rapidly, especially among men, during the last decades [ 1 , 12 ]. Over two thirds of the Chinese population is currently exposed to environmental tobacco smoke (ETS) in the home [ 12 ]. Literature from Western populations has generally shown that maternal smoking is more strongly associated with children's respiratory symptoms than paternal smoking [ 13 , 14 ]. China provides a unique opportunity to examine effects of parental smoking because prevalence of smoking remains very low among women. We examined the associations between respiratory symptoms and exposure to residential coal burning and environmental tobacco smoke in a cross-sectional study of seventh grade students in the greater metropolitan area of Wuhan, China. Methods Participants and study procedure In the spring of 1999, 5231 seventh grade students at 22 public schools in and around Wuhan, China were invited to complete a self-administered questionnaire on respiratory symptoms and home environment. Two schools were randomly selected from each of the 11 administrative units governed by the city of Wuhan, the capital of Hubei Province. The resulting schools included 14 schools from urban (53.0% of the students), three schools from suburban (25.1%), and five schools from rural areas (21.9%). Of the 5231 students, 5051 (97 %) provided parental consent and completed the questionnaire in class with study staff in attendance. Before completing the questions students viewed a video demonstrating wheezing [ 15 ]. The study protocol was approved by the Institutional Review Boards at the Wuhan Public Health and Anti-Epidemic Station and the National Institute of Environmental Health Sciences. Questionnaire data The questionnaire included questions on respiratory health and potential risk factors such as exposure to cooking and heating fuels, smokers in the home, and personal smoking. We incorporated respiratory health items from a standardized questionnaire (ATS-DLD-78-C) translated into Chinese for a previous study in Wuhan and other Chinese cities [ 16 ]. Our primary outcome measures were responses to questions regarding respiratory symptoms in the past 12 months. We asked whether children had cough and/or phlegm production almost everyday during the past 12 months, with and without colds. Additionally, we asked whether children had wheezing over the past 12 months, with and without colds. We assessed exposure to residential coal burning by questions defining the types of fuels used for cooking and heating in the child's home. We combined information on cooking and heating with coal into a single variable with the following categories: no coal stove, coal stove used only for heating, coal stove used only for cooking, and coal stove used for both cooking and heating. To assess environmental tobacco smoke exposure, we asked the child to list all household members and indicate whether each person currently smoked. We created three exposure categories: no smokers in the home, one smoker (1) in the home, and two or more smokers (2 +) in the home. Personal smoking was not considered in the analysis because of very low prevalence (0.6%). Statistical analysis We calculated prevalence odds ratios (95% confidence intervals) for each of the six outcome measures (cough, phlegm production and wheezing in the past 12 months, each with colds and without colds) by unconditional logistic regression (Proc Genmod in SAS System for Windows, Version 8.01). Although the odds ratio is the most common measure of association in cross-sectional studies [ 17 ], divergence between odds ratios and risk ratios increases as the outcome becomes more common [ 18 , 19 ]. However, we present odds ratios as our effect measures to estimate the associations between respiratory symptoms and residential exposures (coal burning and environmental tobacco smoke); using the log-binomial model (Proc Genmod in SAS System for Windows, Version 8.01) [ 20 ] to estimate prevalence proportion ratios for the outcomes did not alter any of our conclusions. We excluded 521 subjects with missing data on any of the outcome or exposure variables leaving 4,530 subjects for the analysis. The following covariates were considered as potential confounders or modifying factors within the logistic models: child's gender, presence of animals in the household, presence of pests (cockroaches, ants, rodents), crowding in the household, presence of older siblings, parental asthma, physical activity, living area (school district), and time spent indoors and outdoors. To account for variation due to the type of neighborhood the children lived in, we included school district (22 districts) in the models using CLASS and REPEATED statements within Proc Genmod in SAS. The models reported here are adjusted for coal use, smokers in the home, school district, and child's sex because inclusion of the other variables did not appreciably change the associations. Results Characteristics of the study population are presented in Table 1 . The mean age of the seventh grade students was 13.6 years (SD = 0.7 years). The majority of the students (94.2%) were life-long residents of the Wuhan area. Although 7.1% of the students reported wheezing without colds, doctor-diagnosed asthma was relatively uncommon in this population (3.2%). Coal was used for cooking and/or heating in almost half of the homes. Few children smoked (0.6%), but 73.2% of the students lived with household members who smoked. The prevalence of ETS exposure was similar across the study area (74.5% in urban areas, 70.2% in suburban areas, and 73.5% in rural areas). Fathers (69.1%) were much more likely to smoke than mothers (3.6%). Table 1 Characteristics of the study population of 4530 students at 22 schools in greater Wuhan, China Characteristic % Subjects Age (mean, SD) in years 13.6, 0.7 Gender Male Female 52.5 47.5 Respiratory symptoms Wheezing with colds 19.4 Wheezing without colds 7.1 Bringing up phlegm with colds 16.7 Bringing up phlegm without colds 5.7 Coughing with colds 24.7 Coughing without colds 4.5 Exposures Smokers in child's household No smokers 1 smoker 2+ smokers Father smokes Mother smokes Personal smoking by students 26.8 62.3 10.9 69.1 3.6 0.6 Exposure to coal burning No coal use Coal used only for heating Coal used only for cooking Coal used for cooking and heating 54.2 8.8 25.9 11.1 After adjusting for gender, ETS, and living area, residential coal burning was primarily associated with wheezing in the past 12 months (Table 2 ). For those who used coal only for cooking or only for heating, wheezing was more strongly associated with cooking. However, the association between coal use and recent wheezing tended to strengthen when coal was used for both cooking and heating (OR = 1.78, 95% CI: 1.08–2.91 for wheezing with colds; OR = 1.57, 95% CI: 0.94–2.64 for wheezing without colds). Table 2 Respiratory symptoms in relation to residential coal burning Cough with colds Cough without colds No Yes No Yes Exposure N N OR* (95% CI) N N OR* (95% CI) Total 3413 1117 4327 203 Coal use No 1833 622 1.00 2347 108 1.00 Yes 1580 495 0.92 (0.76,1.11) 1980 95 1.03 (0.80,1.33) Heating 300 99 0.96 (0.76,1.22) 381 18 1.02 (0.67,1.55) Cooking 926 249 0.79 (0.67,0.94) 1120 55 1.04 (0.74,1.46) Both 354 147 1.22 (0.93,1.59) 479 22 0.99 (0.66,1.49) Phlegm with colds Phlegm without colds No Yes No Yes N N OR* (95% CI) N N OR* (95% CI) Total 3772 758 4274 256 Coal use No 2051 404 1.00 2315 140 1.00 Yes 1721 354 1.04 (0.90,1.20) 1959 116 0.96 (0.75,1.23) Heating 331 68 1.04 (0.86,1.27) 376 23 1.02 (0.62,1.66) Cooking 994 181 0.92 (0.76,1.10) 1114 61 0.86 (0.64,1.16) Both 396 105 1.34 (1.05,1.73) 469 32 1.12 (0.81,1.54) Wheeze with colds Wheeze without colds No Yes No Yes N N OR* (95% CI) N N OR* (95% CI) Total 3652 878 4210 320 Coal use No 2058 397 1.00 2309 146 1.00 Yes 1594 481 1.57 (1.07,2.29) 1901 174 1.44 (1.05,1.97) Heating 329 70 1.10 (0.76,1.57) 368 31 1.35 (0.86,2.15) Cooking 892 283 1.66 (1.01,2.73) 1077 98 1.42 (1.05,1.92) Both 373 128 1.78 (1.08,2.91) 456 45 1.57 (0.94,2.64) * Odds ratios (OR) adjusted for gender, ETS, and school district. Dichotomous and multilevel odds ratios are computed in separate models. After adjusting for gender, coal use, and living area, living with smokers (Table 3 ) was significantly associated with chronic cough and phlegm production in the past 12 months. The strongest associations were seen for cough (OR = 1.74, 95% CI: 1.17–2.60) and phlegm production (OR = 2.25, 95% CI: 1.36–3.72) without colds among children who lived with two or more smokers. Living with smokers was not appreciably associated with wheezing. Table 3 Respiratory symptoms in relation to living with smokers Cough with colds Cough without colds No Yes No Yes Exposure N N OR* (95% CI) N N OR* (95% CI) Total 3413 1117 4327 203 Smokers in the home No 954 259 1.00 1165 48 1.00 Yes 2459 858 1.29 (1.05,1.58) 3162 155 1.19 (0.86,1.65) 1 smoker 2105 717 1.26 (1.02,1.55) 2700 122 1.10 (0.77,1.57) 2+ smokers 354 141 1.47 (1.11,1.95) 462 33 1.74 (1.17,2.60) Phlegm with colds Phlegm without colds No Yes No Yes N N OR* (95% CI) N N OR* (95% CI) Total 3772 758 4274 256 Smokers in the home No 1036 177 1.00 1164 49 1.00 Yes 2736 581 1.24 (1.08,1.43) 3110 207 1.60 (1.11,2.29) 1 smoker 2327 495 1.25 (1.09,1.43) 2657 165 1.49 (1.04,2.14) 2+ smokers 409 86 1.23 (0.92,1.64) 453 42 2.25 (1.36,3.72) Wheeze with colds Wheeze without colds No Yes No Yes N N OR* (95% CI) N N OR* (95% CI) Total 3652 878 4210 320 Smokers in the home No 993 220 1.00 1125 88 1.00 Yes 2659 658 1.11 (0.93,1.31) 3085 232 0.96 (0.74,1.25) 1 smoker 2265 557 1.10 (0.93,1.30) 2619 203 0.99 (0.75,1.30) 2+ smokers 394 101 1.13 (0.85,1.49) 466 29 0.78 (0.45,1.37) * Odds ratios (OR) adjusted for gender, coal use, and school district. Dichotomous and multilevel odds ratios are computed in separate models. Discussion Domestic coal use and exposure to ETS in the home were both associated with adverse respiratory effects in this population of Chinese adolescents. Coal burning was associated with increased wheezing, whereas living with smokers was associated with increased cough and phlegm production. Coal burning produces high concentrations of particulate matter, SO 2 , and other pollutants [ 2 , 3 , 11 ]. Exposure to these pollutants may impair clearance mechanisms, and lead to airway inflammation [ 21 , 22 ]. Decreased pulmonary function has been associated with exposure to particulate matter and SO 2 in several air pollution studies during the past decades [ 21 ]. Although residential coal burning has been linked to decreased pulmonary function and asthma among children [ 23 - 25 ], conflicting data exist. In two European studies, domestic coal burning has been associated with lower risk for childhood asthma and allergic diseases [ 26 , 27 ]. The findings in these two studies, however, may reflect some early life or other lifestyle factors related to coal use in Europe. In our study, residential coal burning was predominantly associated with wheezing. Coal cooking was a stronger risk factor for wheezing than was coal heating. This may be explained by relatively low heating use in the Wuhan area, whereas cooking is a year around activity. The greater association with coal use for both cooking and heating may suggest an exposure-dependent relationship. Although wheezing is often closely related to asthma, coal use was not positively associated with asthma diagnosis (data not shown) in this population. The majority of the diagnosed asthmatics (76.4%) lived in urban areas, where prevalence of coal use was lower than in non-urban areas. The diagnostic ascertainment of asthma most likely was greater in the urban than in the rural areas. The harmful effects of ETS in children, primarily from living with smokers, have been widely studied [ 14 , 28 - 31 ]. In general, evidence that ETS causes cough, phlegm, and wheezing has not been as strong for school-aged children as it has been for infants and preschool children [ 28 ]. There are few data among Chinese populations where smoking behavior differs from Western populations. In utero exposure, via maternal smoking, that is believed to contribute to adverse effects of ETS in children [ 32 , 33 ] is uncommon in China. Thus, it is of interest that in this group of middle school children, where maternal and personal smoking were low, exposure to ETS in the home was clearly associated with chronic cough and phlegm production, with and without colds. Our results indicated an exposure dependent response to ETS; having two or more smokers in the household increased the odds of cough and phlegm production compared to having only one smoker in the household. We did not find strong evidence suggesting modifying effects by gender, although the effect of ETS on persistent cough without colds was more pronounced among boys than girls (data not shown). Exposure levels may be influenced by time-activity patterns that can differ by gender. Boys may be more likely to spend time in close proximity with their smoking fathers or male relatives than girls. Mechanisms responsible for the respiratory effects of ETS have been proposed in the literature [ 28 ]. In addition to decreased mucociliary clearance and goblet cell hypertrophy/hypersecretion, local and central nervous system components are thought be involved in cough and phlegm production [ 28 , 34 ]. Although exposure to ETS may affect childhood lung growth and result in lower pulmonary function [ 14 , 35 ], wheezing was not appreciably related to the presence of smokers in our study. Genetic susceptibility may influence the effects of ETS on bronchial obstruction. For example, parental atopy was found to modify the effects of ETS on bronchial obstruction and asthma considerably in a Norwegian birth cohort study [ 36 ]. However, we were unable to examine potential interactions between family history and ETS in relation to atopic illness in our population because, consistent with previously published data on Chinese children [ 37 , 38 ], the prevalence of asthma (3.2%) and hay fever (1.8%) was very low. In general, our findings agree with available data on Chinese children's respiratory health [ 16 , 23 ]. However, residential exposures in the current study were more selectively associated with the respiratory symptoms than in previous studies. This may reflect differences in the study settings. In the previous studies [ 16 , 23 ], for example, most of the children were younger in age than in the current study. Prevalence of symptoms and factors associated with childhood respiratory symptoms may differ between different age groups [ 39 ]. It is also possible that using students rather than parents as a source of information on child's symptoms may contribute to the observed differences. Exposure to indoor air pollutants is not only influenced by the source strength and other emission characteristics, but also by air exchange rates. A recent study showed that ventilation could modify effects between respiratory health outcomes and indoor air pollutants [ 40 ]. In that study, the modifying effects were found most relevant when air exchange rates were low. Residences in Wuhan, however, were not energy-efficiently built [ 16 ]. Air conditioning was uncommon, and most of the homes, both in urban and non-urban areas, relied on natural ventilation. In this study, we were unable to evaluate the effects of ventilation rates, because we did not collect detailed information on ventilation practices. We thought that children would not be able to give this information accurately. The composition of pollutants produced by residential coal burning and smoking can be highly variable, but both exposures contribute substantially to inhalable and respirable particulate matter in indoor environments [ 2 , 3 , 41 ]. Existing data suggests that coal burning and smoking may have synergistic effects on respiratory symptoms [ 5 ]. In our data, we did not find consistent evidence of interaction between coal burning and ETS exposure. Our outcome and exposure measures were determined by questionnaire alone, which is one of the major limitations of the study. Nonetheless, large epidemiological studies of respiratory health often rely on reports on recent symptom history because self-reported measures are cost efficient, practical and their repeatability is good [ 42 , 43 ]. Generally, respiratory symptoms have been reported consistently across populations [ 43 ]. To improve the quality of our self-reported outcomes we included audiovisual presentation of wheezing symptoms [ 15 ]. Because the temporal relationship between outcome(s) and exposure(s) can be difficult to determine in cross-sectional studies we focused on respiratory symptoms in the past 12 months to minimize recall bias. We did not use parents as source of information on child's symptoms. Some studies suggest that Chinese parents may deny or underreport child's symptoms or illnesses [ 44 , 45 ]. In addition, parents living in non-urban areas around Wuhan have lower educational level than parents living in urban areas [ 16 ], and their literacy level may be lower than their children attending middle school. Therefore, adolescents' reports on their own symptoms and health status may be more accurate than their parents'. Because children were answering in school about exposures in their home, we were not able to acquire very detailed information on exposure characteristics. Given that questionnaires have limited ability to quantify exposures, the possibility of exposure misclassification cannot be excluded. However, serious differential misclassification either of the exposures or outcomes is unlikely because health hazards of indoor air pollutants were not widely known among Chinese school children at the time when the survey was conducted [ 46 ]. Although urban air pollution has long been a major environmental concern in China, we do not believe that outdoor air pollution alone could explain the observed associations. Exposures to indoor air pollutants are likely to dominate the total exposure burden [ 47 ], especially among children, who spend much of their time inside the home [ 8 ]. In Chinese homes with coal stoves and smokers, not only levels of particulate matter, but also levels of many other air pollutants, including concentrations of SO 2 , often exceed the levels outdoors [ 2 , 3 ]. In Wuhan, where coal stoves are not usually vented via flue, concentrations of respirable particulate matter (291 μg/m 3 ) and SO 2 (173 μg/m 3 ) can reach high levels indoors [ 2 ]. Concentrations of these pollutants have been found to be lower in ambient air. For example, a study investigating long-term air pollution in Wuhan estimated that the annual means for PM 2.5 , PM 10 , and SO 2 in urban areas were 73 μg/m 3 , 129 μg/m 3 , and 73 μg/m 3 , respectively [ 48 ]. Because indoor air quality is influenced by infiltration of outdoor air, we cannot fully exclude possible confounding effects of ambient air pollution [ 1 ]. However, the effects of living area, as measured by school districts, were taken into account in our models, providing some control for differing air pollution levels in the study area. The major strength of this study is that the public school system ensured a large and representative sample of rural, suburban, and urban populations in the Wuhan area. Our study is one of the few studies that have examined effects of major indoor pollutants in relation to children's respiratory health in mainland China [ 16 , 23 - 25 ]. Conclusions Coal burning and living with smokers contributed to persistent respiratory symptoms in this cohort of Chinese adolescents. Adverse effects of tobacco smoke in the home were seen despite the very low prevalence of maternal smoking. Even if exposure to residential coal burning declines in response to economic changes in China, the increasing prevalence in smoking augur an increase in children's exposure to environmental tobacco smoke. Because many men initiate smoking during adulthood, and the rate of quitting and desire to quit smoking are low [ 49 ], future prospects for children's health are worrisome. The rise in cigarette smoking in China is a growing public health concern, not only in the adult population but because its effects on children. Although rates of childhood asthma have remained low in China, common indoor air pollutants, coal and tobacco smoke, impair children's respiratory health. List of abbreviations CI = confidence interval CO = carbon monoxide ETS = environmental tobacco smoke OR = odds ratio PM 10 = particulate matter with an aerodynamic diameter less or equal to 10 μm PM 2.5 = particulate matter with an aerodynamic diameter less or equal to 2.5 μm SO 2 = sulfur dioxide Competing interests The authors declare that they have no competing interests. Authors' contributions Contributors: PMS analyzed the data and wrote the manuscript with input from all investigators. YL, JX and CL are key investigators for the data collection. ELA assisted with data collection, and GK assisted with data analysis. CAJ was involved in design of the study. SJL is the principal investigator and guarantor of the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543575.xml
497042	Caveolin-2 associates with intracellular chlamydial inclusions independently of caveolin-1	Background Lipid raft domains form in plasma membranes of eukaryotic cells by the tight packing of glycosphingolipids and cholesterol. Caveolae are invaginated structures that form in lipid raft domains when the protein caveolin-1 is expressed. The Chlamydiaceae are obligate intracellular bacterial pathogens that replicate entirely within inclusions that develop from the phagocytic vacuoles in which they enter. We recently found that host cell caveolin-1 is associated with the intracellular vacuoles and inclusions of some chlamydial strains and species, and that entry of those strains depends on intact lipid raft domains. Caveolin-2 is another member of the caveolin family of proteins that is present in caveolae, but of unknown function. Methods We utilized a caveolin-1 negative/caveolin-2 positive FRT cell line and laser confocal immunofluorescence techniques to visualize the colocalization of caveolin-2 with the chlamydial inclusions. Results We show here that in infected HeLa cells, caveolin-2, as well as caveolin-1, colocalizes with inclusions of C. pneumoniae (Cp), C. caviae (GPIC), and C. trachomatis serovars E, F and K. In addition, caveolin-2 also associates with C. trachomatis serovars A, B and C, although caveolin-1 did not colocalize with these organisms. Moreover, caveolin-2 appears to be specifically, or indirectly, associated with the pathogens at the inclusion membranes. Using caveolin-1 deficient FRT cells, we show that although caveolin-2 normally is not transported out of the Golgi in the absence of caveolin-1, it nevertheless colocalizes with chlamydial inclusions in these cells. However, our results also show that caveolin-2 did not colocalize with UV-irradiated Chlamydia in FRT cells, suggesting that in these caveolin-1 negative cells, pathogen viability and very likely pathogen gene expression are necessary for the acquisition of caveolin-2 from the Golgi. Conclusion Caveolin-2 associates with the chlamydial inclusion independently of caveolin-1. The function of caveolin-2, either in the uninfected cell or in the chlamydial developmental cycle, remains to be elucidated. Nevertheless, this second caveolin protein can now be added to the small number of host proteins that are associated with the inclusions of this obligate intracellular pathogen.	Background The Chlamydiaceae are gram-negative obligate intracellular bacterial pathogens that replicate entirely within membrane bound inclusions that develop from the phagocytic vacuoles in which they enter. By mechanisms not understood in detail, inclusions avoid fusing with endosomes or lysosomes that might destroy the pathogen [ 1 - 3 ]. Instead, the inclusions provide an intracellular niche in which the pathogen can survive and complete its developmental cycle. The initial pathogen-containing endocytic vacuoles co-associate with microtubules, and dynein, and traffic to a perinuclear region, ultimately residing at the microtubule organizing center and in close proximity to the Golgi [ 4 , 5 ]. Here, they intercept exocytic vesicles of the biosynthetic pathway that are derived from the Golgi and continue their developmental cycle [ 6 - 9 ]. There are two genera in the family of Chlamydiaceae [ 10 ]. The genus Chlamydia is comprised of the C. trachomatis species, which contains human pathogens, as well as the mouse and swine strains, C. muridarum and C. suis respectively, while the Chlamydophila genus is more diverse, consisting of six species. Of these, C. pneumoniae , and C. psittaci are relevant in human diseases. C. trachomatis is primarily a pathogen of humans, containing the trachoma biovariant (biovar), and the lymphogranuloma biovar (LGV). The mouse pneumonitis agent (MoPn), frequently used in vaccine work [ 11 , 12 ], is now classified separately as C. muridarumn [ 10 ]. Some serological variants (serovars) of the trachoma biovar primarily infect the urogenital tract (e.g., E, K, and F) and are the major cause of significant venereal disease in the United States [ 13 ]. Other, trachoma serovars primarily infect the conjunctiva (e.g., A, B, and C), causing trachoma, the major form of preventable blindness, worldwide [ 14 ]. The LGV biovar also causes chronic venereal disease, which differs clinically from the sexually transmitted disease caused by strains of the trachoma biovar. C. pneumoniae , another human pathogen, is now implicated in atherosclerosis and cardiac disease, as well as respiratory infection [ 15 , 16 ]. C. psittaci primarily infects birds, but can also cause disease in humans. The mechanism by which chlamydiae enter host cells is not well defined, largely because of conflicting reports that these organisms can enter by both clathrin-dependent, as well as clathrin-independent pathways. Recently, we examined the entry mechanisms of representative members of each of the above Chlamydia species and biovars. Our concern was whether their entry might be mediated by the cholesterol and sphingolipid enriched micro-domains in host cell plasma membranes known as lipid rafts [ 17 ]. Application of biochemical criteria and confocal microscopy demonstrated that five of the ten strains we examined indeed enter by a mechanism that is dependent on intact lipid rafts [ 18 , 19 ]. These strains included C. trachomatis serovars E, F, and K, C. pneumoniae (strain AR39), and C. caviae (Guinea Pig Inclusion Conjunctivitis strain; GPIC). In contrast , C. trachomatis serovars A, B, C, biovar LGV (L2 strain), and MoPn do not enter via lipid rafts. Caveolae are now considered to be a specialized kind of lipid raft. These 50–100 nm flask-shaped invaginations form in lipid rafts when the caveolae marker protein, caveolin-1, is expressed [reviewed in [ 20 ]]. In contrast, lipid rafts are ubiquitous entities that form regardless of caveolin-1 expression. The inclusion membranes of all of the chlamydial strains that entered via lipid rafts, as indicated by biochemical criteria, also appeared to acquire caveolin-1 at entry, and accumulated more of the protein during development in caveolin-1 expressing HeLa cells [ 18 , 19 ]. In contrast to the strains that entered by a raft-dependent process, none of the other strains examined acquired caveolin-1 at entry. The fact that entry of some chlamydial strains is dependent on lipid rafts in general, rather than on the more specialized caveolae, was demonstrated using Fischer Rat Thyroid (FRT) cells, which do not express caveolin-1 or form caveolae. All ten strains that we examined were able to enter the caveolin-1 negative FRT cells, but not if those cells were treated with pharmacological agents that specifically disrupt lipid rafts [ 19 ]. The Golgi is a likely source of most caveolin-1 on the inclusions of those strains entering via lipid rafts. This is suggested by the fact that caveolin-1 immunostaining of those inclusions actually increases as they enlarge and mature, and is consistent with reports that caveolin-1 first becomes associated with raft domains in the Golgi, and subsequently is transported to the plasma membrane on vesicles derived from the Golgi [ 21 - 23 ]. Moreover, chlamydial inclusions are known to intercept vesicles of the biosynthetic pathway that originate from the Golgi, from which they acquire sphingolipids and perhaps other nutrients as well[ 6 , 7 ]. However, since inclusions of all strains appear to acquire sphingolipids from the Golgi, whereas only a subset of strains acquires caveolin-1, there may be more than one sorting process and class of vesicles with which Chlamydia can interact. The caveolin gene family also contains caveolin-2 [ 24 ] which has 38% sequence identity with caveolin-1 and 58% similarity. Caveolin-1 and -2 are co-expressed in most cells, in which they form hetero-oligomeric complexes. These complexes form in the endoplasmic reticulum (ER), becoming associated with lipid rafts in the Golgi, from which they are transported to the plasma membrane. All of the putative functions of caveolae are thus far associated with caveolin-1, and the function of caveolin-2 is essentially unknown. Importantly, unlike caveolin-1, caveolin-2 can not induce caveolae to form in lipid rafts. An experimental finding that is particularly relevant to the current study is that caveolin-2 can not exit the Golgi in caveolin-1 negative FRT cells. However, it can exit when recombinant caveolin-1 is expressed [ 25 , 26 ]. Thus, formation of hetero-oligomeric complexes of caveolin-1 and -2 facilitates transport of caveolin-2 out of the Golgi and its sorting to the plasma membrane. In the present study, we ask whether caveolin-2, like caveolin-1, might be associated with inclusions formed by particular chlamydial strains. Unexpectedly, we found that in HeLa cells, caveolin-2 was associated with inclusions of strains that did not enter via lipid rafts and that did not acquire caveolin-1 either at entry or at later times. These strains included C. trachomatis serovars A, B, and C. In the same cells, inclusions of C. trachomatis serovars E, F, and K, C. pneumoniae (A39), and C. caviae (GPIC) were marked by both of the caveolin proteins. Each of these eight strains acquired caveolin-2 even in caveolin-1 negative FRT cells. In addition, the ability of each of these strains to acquire caveolin-2 in the FRT cells was dependent on pathogen gene expression. Thus, chlamydiae appear to activate or induce an additional pathway for the transport of caveolin-2 that is independent of caveolin-1. Inclusions of two strains, LGV (L2) and MoPn, were not marked by either caveolin-1 or -2. Methods Chlamydial strains C. pneumoniae AR39 (Cpn), C. caviae, Guinea Pig Inclusion Conjunctivitis (GPIC strain), C. trachomatis serovars A/Har-13, Har-36B, C/TW-3, E/VW-KX, F, K/VR887, mouse pneumonitis agent, (MoPn) and Lymphogranuloma venereum, (LGV 434) were grown in HeLa 229 cells without centrifuge assistance. Infectious elementary bodies (EBs) were purified by renografin (Squibb diagnostics, New Bronswick, NJ) density gradient centrifugation; alternatively lysates from infected cells were used to infect monolayers. Cell lines used HeLa 229 cells were obtained from the American Type Culture collection. Fischer rat thyroid (FRT) cells were a gift from Dr. Michael P. Lisanti (Albert Einstein College of Medicine, Bronx, NY). Cells were grown in minimum essential medium with insulin (IMEMZO, Irvine Scientific, Santa Ana, CA) with 5% fetal bovine serum (FBS, Atlanta Biologicals, Norcross, GA.). Immunostaining of Chlamydia infected cells HeLa 229 or FRT cells were grown to confluence on 12 mm coverslips in 24 well plates (Becton Dickinson Labware, Franklin Lakes, NJ). The cells were infected using the various chlamydial strains listed above at an MOI of 3.0 in complete cycloheximide overlay media (Bio-Whittaker, Walkersville, Md.) containing 10% FBS, 1X L-glutamine (CCOM) for 48 h at 37°C and 5% CO2. Coverslips with the cell monolayers were harvested, rinsed with phosphate buffered saline (PBS), fixed with 70% cold methanol, stored and subsequently immunostained following protocols similar to that described in detail previously [ 18 , 19 ]. Briefly, infected cells were immunostained with a guinea pig anti-chlamydia polyclonal antibody (Biomedia, Foster City, CA) and the monoclonal mouse anti-caveolin-2 antibody (Transduction Laboratories, Lexington, KY) for 1 h at 37°C. Following 4 washes with PBS, the bound antibodies were detected using a 1:50 dilution of TRITC-conjugated goat anti-guinea pig and FITC-conjugated goat anti-mouse secondary antibodies (Jackson Immuno Research, West Grove, PA). Following incubation for 1 h at RT and 4 rinses with (PBS), coverslips were mounted onto slides using Fluoromount-G (Southern Biotechnology Associates Inc., Birmingham, AL). Slides were examined at 600X using a Bio-Rad MRC-600 Laser Confocal Microscope system. Images were captured and as relevant, merged using the Confocal Assistant™ version 4.02 Image Processing Software. UV-treatment of EBs and infection of HeLa and FRT cells C. trachomatis serovars A, B, C, E, F, LGV, C. muridarum (MoPn), C. caviae (GPIC), and C. pneumoniae EB concentrates were exposed to short wavelength ultraviolet light for 1 h in a Biosafety hood. A portion of the EBs used was removed before UV treatment to serve as a control for initial viability. The EBs were then diluted in CCOM to an MOI of 4.0 per cell and used to infect HeLa 229 and FRT cells grown to confluence on 12 mm coverslips in 24 well plates. Plates were incubated for 2 h at 37°C and 5% CO2. After the 2 h incubation, the inoculum was removed and replaced with fresh CCOM; cells were then re-incubated for 34 h. At the end of the incubation period, the coverslips with cells were rinsed with sterile PBS and fixed with methanol for 10 minutes as detailed above. Following fixation, the cells were immunostained with a polyclonal guinea pig anti-chlamydia and mouse monoclonal anti-caveolin-2 antibodies as described above. FITC-conjugated goat anti-mouse and TRITC-conjugated goat anti-guinea pig secondary antibodies were used respectively to visualize the caveolin-2 and EBs within the cells. Slides were examined at 600X with a Laser Confocal Microscope system and images processed as described above. Western blot of purified EBs and anti-Caveolin-2 C. trachomatis serovar K and C. pneumoniae AR39 purified EBs were electrophoresed on a 4–12% NuPAGE gel (Invitrogen Life technologies, Carlsbad, CA) using MES running buffer. The separated proteins were transferred to a PVDF membrane at 25 V for 1 h using a 1X Novex ® Tris Glycine buffer with 20% methanol in the XCell II™ Blot Module and a XCell Surelock ™ Mini-Cell apparatus (Invitrogen). The blot was then rinsed with 0.1% BSA/PBS, blocked with 5% non-fat dry milk for 3 h, then rinsed and cut into two. One half was stained with a 1:200 dilution of mouse monoclonal anti-caveolin-2 antibody (Transduction Labs, Lexington, KY), while the second half was stained with a 1:100 dilution of a rabbit anti- Chlamydia for 2 h with shaking. The blot was then rinsed and a 1:1000 dilution of AP-conjugated goat anti-mouse secondary antibody added and incubated for 1 h at RT. Following several washes, a BCIP/NBT alkaline phosphatase substrate was added and the color allowed to develop. Results Caveolin-2 on chlamydial inclusions in caveolin-1 positive HeLa cells We asked whether caveolin-2 might be associated with the inclusions of each of the ten chlamydial strains and species that we had previously assessed for raft-mediated entry and for inclusion-associated caveolin-1 [ 19 ]. We began by looking at infections in caveolin-1 positive HeLa cells. Inclusions of C. trachomatis serovars E, F, and K, C. pneumoniae (A39), and C. caviae (GPIC), previously were seen to enter HeLa cells via lipid rafts, and their inclusions were marked by caveolin-1 in those cells (ibid). Immunostaining with caveolin-2 specific monoclonal antibodies demonstrated that in HeLa cells, inclusions of each of these strains also were marked by caveolin-2 (Figure 1 , Table 1 ). As in the case of caveolin-1 [ 19 ], optical sections through the Z-axis clearly indicated the caveolin-2 as packets associated with the vacuolar membrane (Figure 2 ). The disappearance of the caveolin-2 staining as the depth of the optical sections increased, with no change in the anti- Chlamydia staining, confirms that the caveolin-2 proteins are associated with the vacuolar membranes of these inclusions and are not associated with the individual EBs and RBs inside of the chlamydial vacuole. This profile is similar to that seen when Z-sections were taken of the caveolin-1 staining. Importantly, immunostaining for caveolin-2 appears nonrandom on the inclusion membrane (Figure 2 , Z section #1). Indeed, the series of merged optical sections shows that the caveolin-2 staining appears in apposition to those pathogen cells that are at the inclusion membrane (Figure 2 , see Discussion). In contrast to the above strains, C. trachomatis serovars A, B, and C, do not enter via lipid rafts, and their inclusions do not acquire caveolin-1 in HeLa cells [ 19 ]. Nevertheless, in HeLa cells inclusions of each of these strains also acquired caveolin-2 (Figure 1 , Table 1 ). In our earlier study, entry of strains LGV (L2) and MoPn was seen to be independent of lipid rafts, and their inclusions were not marked by caveolin-1. In the present study, their inclusions did not display caveolin-2 and therefore they did not accumulate this host protein during their development (Figure 1 , Table 1 ). Caveolin-2 is present on chlamydial inclusions in caveolin-1 negative FRT cells Others have demonstrated that caveolin-2 is not transported out from the Golgi in caveolin-1 negative FRT cells, but that expression of a transfected caveolin-1 gene in those cells restores caveolin-2 transport [ 25 , 26 ]. These experimental findings imply that caveolin-2 transport is dependent on its association with caveolin-1. Thus, it was somewhat surprising that during development in HeLa cells, inclusions of C. trachomatis serovars A, B, and C acquired caveolin-2 without acquiring caveolin-1. To confirm that chlamydial inclusions indeed could acquire caveolin-2 independently of caveolin-1, we further examined infections using caveolin-1 negative FRT cells. Normally these cells do not express caveolin-1 or form caveolae, but do so when transfected with caveolin-1 cDNA [ 27 - 29 ]. Inclusions of C. trachomatis serovars A, B, C, E, F, and K, as well as C. pneumoniae , and C. caviae each were marked by caveolin-2 in the caveolin-1 negative FRT cells (Figure 3 ). This clearly confirms that trafficking of caveolin-2 to chlamydial inclusions does occur independently of caveolin-1. Note that earlier we demonstrated that among those we tested, MoPn and GPIC are the only strains that develop large inclusions in FRT cells, and confirmed that these cells indeed do not express caveolin-1 [ 19 ]. To confirm the specificity of the caveolin-2 antibody for the caveolin-2 protein in these cells, and to rule out the possibility of cross-reactivity of this antibody with the chlamydial EBs, we performed western blot on renografin gradient purified EBs from C. pneumoniae and C. trachomatis and stained the blot with anti-caveolin-2 antibodies (figure 4A ). As seen in the representative blot, the caveolin-2 antibody does not cross-react with purified EBs. This is also true for anti-caveolin-1 antibodies (data not shown). The presence of chlamydial EBs on the blot was assessed by loading the same amount of EBs onto the second half of the gel in figure 4A and staining with a rabbit anti- Chlamydia serum (figure 4B ). Note that there is adequate material on this stained blot, confirming that if the caveolin antibody was cross-reactive with EB material one would be able to see such a reaction. Acquisition of caveolin-2 by inclusions requires pathogen gene expression The apparent inability of caveolin-2 to traffic from the Golgi independently of caveolin-1 in uninfected cells [ 25 , 26 ] implies that in a manner yet unknown, the chlamydial inclusions might produce a factor that activates an additional sorting and transport process for caveolin-2. To test this hypothesis, we asked whether UV-inactivated EBs, the infectious form of the pathogen, might sequester caveolin-2 on their vacuoles in caveolin-1 negative FRT cells. Pathogen-containing vacuoles of each of the strains indeed were unable to acquire caveolin-2 when the EBs had been inactivated by UV irradiation prior to infection. This result therefore supports the premise that vacuole/inclusion acquisition of caveolin-2 is dependent on pathogen gene expression (Figure 5 ). In caveolin-1 positive HeLa cells, a minimal amount of caveolin-2 protein was observed to co-localize to vacuoles with UV-treated EBs (Figure 6 , see Discussion). Discussion Experimental results presented here demonstrate that for a number of chlamydial strains and species, the intracellular inclusions indeed acquire caveolin-2 during the developmental cycle. These strains include C. trachomatis serovars A, B, C, E, F, and K, C. pneumoniae (A39), and GPIC (Table 1 ). C. trachomatis serovars E, F, and K, and the C. pneumoniae and C. caviae species of Chlamydophila previously were shown to enter host cells via lipid rafts. Their vacuoles/inclusions were demonstrated to acquire host cell caveolin-1 at entry and to accumulate it during later stages of infection [ 18 , 19 ]. Consequently, in host cells that express both caveolin proteins, inclusions of these strains display caveolin-2, as well as caveolin-1. However, in this same cell type, inclusions of serovars A, B, and C, are marked only by caveolin-2, a result implying that for chlamydial inclusions this acquisition can occur independently of caveolin-1. This implication is somewhat surprising since there is experimental evidence demonstrating that caveolin-2 usually can not exit from the Golgi and traffic to the plasma membrane in the absence of caveolin-1[ 25 , 26 ]. However, if recombinant caveolin-1 DNA is expressed by caveolin-1 negative cells, then caveolin-2 can transit from the Golgi and be delivered to the plasma membrane, as a hetero-oligomeric complex with caveolin-1[ 25 , 26 , 30 , 31 ]. To confirm that transport of caveolin-2 to chlamydial inclusions indeed might occur independently of caveolin-1, we asked whether chlamydial inclusions might similarly acquire caveolin-2 in FRT cells, a cell line that does not express caveolin-1 [ 27 - 29 ]. In our earlier report [ 19 ], we confirmed that FRT cells in fact do not express caveolin-1. Also, we demonstrated that whereas for some chlamydial strains, entry into those cells was dependent on intact lipid rafts (Table 1 ), in no case was entry dependent on caveolin-1 and caveolae. We show here that in the caveolin-1 negative FRT cells, inclusions of each of the above chlamydial strains in fact acquired caveolin-2. Moreover, since FRT cells do not contain caveolin-2 at the plasma membrane, pathogen-associated caveolin-2 could not have been acquired at entry. Rather, it had to derive from an intracellular source. Chlamydial inclusions thus acquired caveolin-2 despite the fact that in the absence of caveolin-1 this protein usually does not traffic from the Golgi. We therefore suggested that the pathogen might influence the host cell and actually induce a transport and sorting pathway that normally may not actively function in the uninfected cell. Consistent with this suggestion is the fact that for all strains and species tested in FRT cells, if the pathogens were inactivated by prior EB treatment with UV irradiation, none of the EB containing vesicles acquired caveolin-2. In contrast, in HeLa cells, there were low levels of caveolin-2 associated with vesicles containing UV-treated EBs. The source of these low levels of caveolin-2 on vesicles/inclusions in HeLa cells is not yet clear. Perhaps the EBs are acquiring minimal amounts of caveolin-2 from the membranes of these cells upon entry. This suggestion is supported by the fact that the vesicles of these UV irradiated EBs contain both caveolin-1 and -2 (data not shown); suggesting that the proteins might be co-localizing in hetero-oligomeric complexes. It is well known that Chlamydia -containing vacuoles and later, the developing inclusions, are able to intercept and fuse with exocytic vesicles of the biosynthetic pathway that originate from the Golgi [ 6 , 7 ]. These Golgi-derived vesicles provide the inclusions with sphingolipids and perhaps other key metabolites [ 7 , 32 ]. In addition, the ability of at least some chlamydial strains to intercept these Golgi-derived vesicles is dependent on pathogen viability and therefore likely on gene expression. Considering these points, and the fact that caveolin-1/2 hetero-oligomers traffic to the plasma membrane from the Golgi [ 25 , 26 , 30 , 31 ], it is reasonable to suggest that the previously described pathogen induced interception of Golgi-derived vesicles might account for the acquisition of caveolin-2 by inclusions, as seen in the current study. However, although studies by others indicate that all chlamydial strains appear to intercept vesicles of the biosynthetic pathway, not all strains acquire caveolin-2 and this fact provides a strong counter argument. Thus for example, in the current study LGV (L2) inclusions did not acquire caveolin-2 (Table 1 ), although they have been reported to intercept Golgi-derived vesicles of the biosynthetic pathway [ 7 ]. Furthermore, in the current study, acquisition of caveolin-2 by inclusions of some strains was independent of caveolin-1, while among other strains, inclusions acquired both caveolin proteins. Thus there may be more than one pathway by which inclusions acquire caveolin-2. Importantly, acquisition of caveolin-2 by inclusions of any particular strain did not correlate with entry by a raft-mediated pathway. In contrast, acquisition of caveolin-1 at early as well as late stages in pathogen development did correlate with a route of entry involving lipid raft microdomains. Together, these facts are consistent with the conclusion that acquisition of caveoloin-1 and -2 can represent independent and distinct processes. As expected, similar incubations with anti-caveolin-3 demonstrated this component was not present in any of these cells whether infected with Chlamydiae or not. To date, no chlamydial protein has been identified that might be secreted into the host cell cytosol to influence host intracellular trafficking. However, several chlamydial proteins, termed Incs, have been identified in the inclusion membrane [ 33 - 35 ]. Several of these Inc proteins have cytoplasmic domains, making them potential mediators for interactions with the host that might influence trafficking [ 36 ]. One of these proteins, IncG, interacts with host protein, 14-3-3β [ 37 ]. This particular interaction can not underlie the trafficking of caveolin-2 to the inclusions since GPIC ( C. caviae ) and C. pneumoniae (AR39) do not express IncG [ 37 ], although, as shown here, they do acquire caveolin-2. Furthermore, the reverse also is true. Thus, as demonstrated above, mouse pneumonitis strain, MoPn and LGV (L2) do not acquire caveolin-2, but they do express IncG [ 37 ]. There are, however, other Inc proteins and they may underlie the phenomena we have presented here. Despite their presence on chlamydial inclusions, it remains unclear what role, if any, caveolin proteins may play in the developmental cycle of Chlamydia . As noted above, all ten of the chlamydial strains and species we examined were able to enter caveolin-1 negative FRT cells [ 19 ]. Since these caveolin-1 negative cells do not express caveolin-2 at the plasma membrane [ 25 , 26 ], it would appear that for any of these strains neither of the caveolin proteins is necessary for entry per se. Although following internalization all ten strains and species remained viable in the FRT cells [ 19 ], the only strains able to generate large inclusions in the FRT cells were GPIC which acquires both caveolin-1 and -2 in caveolin-positive cells, and MoPn which acquires neither caveolin protein. Despite these latter two instances, our findings remain consistent with the possibility that caveolin-1 may yet play a post entry role in the development of at least some strains. The fact that successful MoPn development does occur in inclusions that acquire neither caveolin-1 nor -2 [ 19 ] (and current study) could mean that neither caveolin-1 nor -2 has a necessary role in the development of MoPn. The ability of GPIC to develop mature inclusions in FRT cells where it acquires only caveolin-2, despite acquiring both caveolin proteins in caveolin-1 positive cells, may imply that for GPIC caveolin-2 rather than caveolin-1 is a key for development. Likewise, inclusions of C. trachomatis serovars A, B, and C, whether developing in HeLa or FRT cells, display only caveolin-2, suggesting a potentially broader chlamydial requirement for the caveolin-2 protein during development. The series of merged optical sections demonstrating the association of caveolin-2 with the inclusions (Figure 2 ) implies that caveolin-2 is in apposition to pathogen cells that are located at the inclusion membrane. Nevertheless, the caveolin-2 is not in direct contact with the pathogens. Caveolin is not a transmembrane protein, a fact that can be deduced from the experimental finding that cell-surface biotinylation does not label caveolin proteins [ 38 ]. Thus, these membrane proteins are not accessible from the extracellular milieu, and this originally extracellular membrane surface, is topologically equivalent to the lumen of the chlamydial inclusion. Hence, whereas our findings imply a specific association between the pathogens and caveolin-2, this association appears to be indirect. The paucity of information concerning caveolin-2 does not enable us to suggest an identity for the linking molecule. Moreover, how this indirect association between caveolin-2 and the pathogen might relate to the protein's function in the chlamydial developmental cycle, or in its acquisition by the pathogen, is also not yet clear. Conclusions Among the Chlamydiae , depending on the serovar, or the species, one or both of the caveolin proteins may play an important role in the developmental cycles. Further studies of the interaction of the Chlamydiae with these enigmatic host cell proteins may well help to clarify caveolin functions in the host cell. Likewise, such studies may indicate factors influencing caveolin expression and trafficking from the Golgi, and elucidate their significance to the developmental cycle of these pathogens. The function of caveolin-2, either in the uninfected cell or in the chlamydial developmental cycle, remains to be elucidated. Nevertheless, this second caveolin protein can now be added to the small number of host proteins that are reported as associated with the inclusions of this obligate intracellular pathogen. Competing interests None Declared. Authors' contributions ES and LN conceived of the study and participated in its design and coordination. LN was instrumental in the design of the study and prepared the initial draft of the manuscript. WW carried out all the experimental work, took photographs, did statistical analyses where necessary, and participated in the design and coordination of the project. All authors contributed to the preparation of the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497042.xml
552336	Fructose, insulin resistance, and metabolic dyslipidemia	Obesity and type 2 diabetes are occurring at epidemic rates in the United States and many parts of the world. The "obesity epidemic" appears to have emerged largely from changes in our diet and reduced physical activity. An important but not well-appreciated dietary change has been the substantial increase in the amount of dietary fructose consumption from high intake of sucrose and high fructose corn syrup, a common sweetener used in the food industry. A high flux of fructose to the liver, the main organ capable of metabolizing this simple carbohydrate, perturbs glucose metabolism and glucose uptake pathways, and leads to a significantly enhanced rate of de novo lipogenesis and triglyceride (TG) synthesis, driven by the high flux of glycerol and acyl portions of TG molecules from fructose catabolism. These metabolic disturbances appear to underlie the induction of insulin resistance commonly observed with high fructose feeding in both humans and animal models. Fructose-induced insulin resistant states are commonly characterized by a profound metabolic dyslipidemia, which appears to result from hepatic and intestinal overproduction of atherogenic lipoprotein particles. Thus, emerging evidence from recent epidemiological and biochemical studies clearly suggests that the high dietary intake of fructose has rapidly become an important causative factor in the development of the metabolic syndrome. There is an urgent need for increased public awareness of the risks associated with high fructose consumption and greater efforts should be made to curb the supplementation of packaged foods with high fructose additives. The present review will discuss the trends in fructose consumption, the metabolic consequences of increased fructose intake, and the molecular mechanisms leading to fructose-induced lipogenesis, insulin resistance and metabolic dyslipidemia.	Emerging epidemic of the Metabolic Syndrome The new millennium has witnessed the emergence of a modern epidemic, the metabolic syndrome, with frightful consequences to the health of humans worldwide. The metabolic syndrome, also referred to as "Diabesity" [ 1 ] describes the increasing incidence of diabetes in combination with obesity as a result of changes in human behaviour, available nutrition, and the adoption of more sedentary lifestyles. Obesity and type 2 diabetes are occurring at epidemic rates in the United States [ 2 - 4 ] and developing countries including China [ 5 ] and India [ 6 ]. From 1935 to 1996, the prevalence of diagnosed type 2 diabetes climbed nearly 765% [ 7 ]. The global figures are predicted to rise 46% from 150 million cases in 2000 to 221 million in 2010 [ 8 ]. This epidemic of type 2 diabetes is complicated by the fact that it is a multi-factorial disease, frequently associated with a cluster of pathologies including obesity, hypertriglyceridemia, impaired glucose tolerance, and insulin resistance, collectively referred to as the metabolic syndrome (formerly known as syndrome X and insulin resistance syndrome). Although there is no universally accepted definition of the metabolic syndrome, most would agree that the syndrome includes a cluster of common pathologies: obesity, insulin resistance, dyslipidemia, and hypertension. It is present in 25–50% of the United States population [ 9 ]. There has been a heightened awareness of the metabolic syndrome and a subsequent increase in clinical attention directed towards prevention, due to its strong association with premature morbidity and mortality [ 8 , 10 ]. In particular, these risk factors predispose the individual to greater risk for developing cardiovascular disease and Type 2 diabetes. Recently, the National Cholesterol Education Panel (NCEP) has officially described and identified a number of these risk factors for cardiovascular diseases [ 11 ]. These include: 1) abdominal obesity, 2) elevated TG levels, 3) low high density lipoprotein (HDL)-cholesterol levels, 4) increased blood pressure, and 5) impaired fasting glucose [ 12 ]. There is also now consensus that insulin resistance and obesity are actually part of one common pathologic mechanism of the metabolic syndrome [ 13 , 14 ]. Evidence shows that the metabolic syndrome process begins early in life and persistence from childhood to adolescent/adult life produces type 2 diabetes and cardiovascular disease [ 15 , 16 ]. The symptoms of metabolic syndrome are not necessarily manifestations of age, but develop over a predisposed background established at a young age [ 17 , 18 ]. This is a dangerous predisposition, with trends in modern diet and habit likely influencing health and behaviour in increasingly younger populations. The main driving forces for the increased prevalence of insulin resistance are modern Westernized diets and patterns of eating associated with the dramatic rises in obesity. Insulin resistance is often linked to the macronutrient content in the diet. In the past, diets high in saturated fats have been shown to induce weight gain, insulin resistance, and hyperlipidemia in humans and animals [ 19 - 22 ]. Recent research suggests that a high intake of refined carbohydrates may also increase the risk of insulin resistance [ 23 - 26 ]. In addition, diets specifically high in fructose have been shown to contribute to a metabolic disturbance in animal models resulting in weight gain, hyperlipidemia [ 27 ], and hypertension [ 28 ]. Nutritional factors influencing the development of the Metabolic Syndrome Nutrition represents a lifestyle element that can be controlled, and that can directly influence health; therefore preventative nutrition and weight control should become a main focus of consumers and prepared-food providers [ 29 ]. The Westernization of diets, with an increase in availability of high calorie foods certainly contributes to the epidemic of metabolic syndrome. In the past, physicians and scientists have made an association between dietary energy from fat and body fat. A large market has developed for the popularity and promotion of low fat diets. Interestingly, however, the decline in dietary fat consumption has not corresponded to a decrease in obesity – in fact, the opposite trend has emerged [ 30 ]. Certainly, diets high in saturated fats have been shown to induce weight gain, insulin resistance, and hyperlipidemia in humans and animals [ 19 - 22 , 31 ], but the emphasis on fat reductions has had no significant benefits relative to the obesity epidemic. More importantly, the focus on dietary fat is more likely a distraction to more significant causes of metabolic syndrome [ 30 ]. If fat is not the culprit in metabolic disorders, then what is? Increasing evidence now suggests that the rise in consumption of carbohydrates, particularly refined sugars high in fructose, appears to be at least one very important contributing factor. Carbohydrates and the link to the Metabolic Syndrome The general increases in consumption of calories, and specifically of refined carbohydrates and fructose, is clear and correlates positively with an alarming increases in metabolic syndrome. Can these seemingly harmless nutrients actually be directly associated with metabolic syndrome? Recent studies appear to support this link. In a 2004 study, Gross et al examined nutrient consumption in the United States between 1909 and 1997, and discovered there was a significant correlation in the prevalence of diabetes with fat, carbohydrate, corn syrup, and total energy intakes. Most striking was the fact that when total energy intake was accounted for, corn syrup was positively associated with type 2 diabetes, while protein and fat were not [ 32 ]. High fructose corn syrups (HFCS) are quite commonly found in soft drinks and juice beverages, and are incorporated into many convenient pre-packaged foods, such as breakfast cereals and baked goods. Fructose consumption has thus largely increased over the past few decades most likely as a result of this increased use of HFCS, which contains between 55–90% fructose. The use of HFCS has increased an alarming 1000% between 1970 and 1990 [ 33 ]. In 1970, individual consumption of fructose was only 0.5 lb/year. However, in 1997, this figure rose to an alarming 62.4 lb/year [ 34 ]. The type of common, general use sweeteners represent as large an impact as the dramatic increase in the use of these caloric sweeteners. Between 1909 and 1997, sweetener use increased by 86%; and specifically, corn syrup sweeteners now represent over 20% of total daily carbohydrate intake, at an increase of 2100% [ 32 ]. These documented trends have inspired a number of consumption studies and recommendations towards HFCS intake. In 1992, the USDA recommended that only 40 g of extra sugars should be added to a standard 2000 calorie a day diet [ 35 ]. The amount of HFCS found in only one 12-oz soft drink equals this total proportion of daily intake. HFCS consumption trends are further exacerbated by the fact that soft drink and fruit juice consumption itself has increased dramatically, adding even more extraneous calories and fructose to the diet. From 1965 to 1996, a food consumption study involving 11 to 18 year olds revealed that total energy and fat intakes were decreasing. There were significant decreases in milk consumption but large increases in the consumption of soft drinks and non-citrus juices [ 36 ]. Increasingly, children seem to be choosing mass-produced, 'tasty' artificial juices and sodas over healthier alternatives. In a recent letter to the editor, Jacobson [ 37 ] illustrates some important factors that contribute to increased consumption of soft drinks, and the link to obesity; a) Society is constantly bombarded by huge million-dollar advertising campaigns for soft drinks, offered extra-extra-large serving sizes with free refills, and surrounded by ubiquitous access to soft drink vending machines even in schools, and b) children's standard drinks to accompany meals, and especially fast food, have become soft drinks. The increased use of HFCS in soft drinks and food products are thus exacerbated by increased exposure, and consumption of these products. HFCS are the main caloric sweeteners utilized in soft drinks in the United States, with fructose representing over 40% of sweeteners added to prepared foods and beverages [ 33 ]. In a study of females aged 12 to 19 years milk intake decreased by 36%, whereas sodas and fruit drink consumption increased to nearly double from the 1970s to the mid 1990s. From 1994 to 1996, it was found that even though intake of soda, juices, tea, and alcoholic beverages remained constant, the steady decrease of milk intake continued [ 38 ]. This becomes a major problem, because while these high-calorie beverages are being consumed, calories from the rest of the diet are not subsequently reduced. The reality is that people do not eliminate or reduce their food portions because they drank a can of soda that day. Data indicate that energy from beverages generally does not displace or decrease energy from other foods consumed, leading to energy imbalances [ 39 ]. The main diet issues involve a general lack of education and/or understanding of the implications with recent consumption patterns. Despite education programs to prevent obesity and diabetes worldwide, there has been little focus on the reduction of fructose and HFCS in beverages. Fructose metabolism Fructose is readily absorbed and rapidly metabolized by human liver. For thousands of years humans consumed fructose amounting to 16–20 grams per day, largely from fresh fruits. Westernization of diets has resulted in significant increases in added fructose, leading to typical daily consumptions amounting to 85–100 grams of fructose per day. The exposure of the liver to such large quantities of fructose leads to rapid stimulation of lipogenesis and TG accumulation, which in turn contributes to reduced insulin sensitivity and hepatic insulin resistance/glucose intolerance. These negative effects of fructose are the reason that fructose metabolism has gained recent research attention. Interestingly, small catalytic quantities of fructose can have positive effects, and actually decrease the glycemic response to glucose loads, and improve glucose tolerance. These effects are also observed without any changes in insulin responses and non-esterified fatty acid (NEFA) and TG levels [ 40 , 41 ]. In 1976, sugar substitutes such as fructose had been found to offer the 'advantage' of a 'better' utilization in conditions of limited insulin production. Fructose had a smaller influence on serum insulin concentrations than glucose, and no influence on plasma glucose levels. At that time, this evidence was considered to support fructose as a positive treatment for diabetic control [ 40 ]. In 1986 HFCS were even proposed as a low-cost substitute for fructose in diabetic management. Based on these early observations, nutritive sweeteners were considered safe by the Food and Drug Administration, although, it has now been found that intakes above 25% of total energy consumed will cause hypertriglyceridemia and gastrointestinal symptoms [ 42 ]. Even with the early positive results, researchers noticed accompanying "unfavorable" influences of these so-called diabetic sugars on obesity and weight gain. Certain metabolic differences exist between glucose and fructose, and the results that were once thought favorable, proved exacerbating to insulin resistance and obesity. In a study comparing normal and diabetic patients, glycemic effects of HFCS were compared to glucose. The negative results of HFCS on immunoreactive insulin, glycemic effect, and immunoreactive C-peptide did not support its use as a substitute for glucose in diabetic patients [ 43 ]. Unfortunately, one out of every four children in the United States consumes above the recommended 25% of total energy intake from sweeteners [ 42 ] and the harmful effects of fructose have been extensively studied in healthy, non-diabetic patients. Studies involving commonly consumed fruit juices showed that natural fructose carbohydrates can alter lipid and protein oxidation biomarkers in the blood, and mediate oxidative stress responses in vivo [ 44 ]. A comparative study by Raben et al . examined overweight men and women who consumed fructose-containing sucrose, as opposed to artificial sweeteners as supplements to their diet. Weight, fat mass, and blood pressure were found to be lower in the artificial sweetener-consuming group compared to the sucrose-consuming group, and the sucrose group did not decrease intake of other nutrients to compensate for their increased calorie consumption from the sucrose. Subjects consuming the sweetener did not exhibit increases in energy intake, weight, and blood pressure that seen in the sucrose-consuming subjects [ 45 ]. Research in the metabolism of fructose has left more questions about the difference between short-term positive effects, and the negative effects of chronic, long-term use of fructose sugars [ 46 ]. The long-term negative effects can include changes in digestion, absorption, plasma hormone levels, appetite, and hepatic metabolism, leading to development of insulin resistance, diabetes, obesity, and inevitably cardiovascular disease. When the metabolic pathways and characteristics of fructose are examined more closely, many of the questions about its positive and negative effects can be answered. Fructose is a potent regulator of glycogen synthesis and liver glucose uptake. Therefore any catalytic improvements are due to hepatic glucokinase and glucose uptake facilitation. However, as mentioned, the beneficial effects do not continue with chronic fructose utilization [ 47 ]. Because of its lipogenic properties, excess fructose in the diet can cause glucose and fructose malabsorption, and greater elevations in TG and cholesterol compared to other carbohydrates [ 48 ]. There are key differences in the metabolic pathways that glucose and fructose follow. Upon gastric absorption both fructose and glucose are delivered via the portal vein to the liver. It is believed that the ability of the liver to metabolize high doses of fructose is responsible for the disruption in energy stores and fuel metabolism observed [ 49 - 52 ]. In the liver, fructose is metabolized into glyceraldehyde and dihydroxyacetone phosphate. These particular fructose end products can then readily converge with the glycolytic pathway. Of key importance is the ability of fructose to by-pass the main regulatory step of glycolysis, the conversion of glucose-6-phosphate to fructose 1,6-bisphosphate, controlled by phosphofructokinase. Thus, while glucose metabolism is negatively regulated by phosphofructokinase, fructose can continuously enter the glycolytic pathway. Therefore, fructose can uncontrollably produce glucose, glycogen, lactate, and pyruvate, providing both the glycerol and acyl portions of acyl-glycerol molecules. These particular substrates, and the resultant excess energy flux due to unregulated fructose metabolism, will promote the over-production of TG (reviewed in [ 53 ]). The glycemic index (GI) has been commonly used to differentiate and compare various nutrients, as well as to describe how different foods produce different plasma glucose levels after ingestion. The GI can range from 100 for glucose and baked potato compared to approximately 20 for fructose and whole barley [ 54 ]. Foods with varying GIs have different time courses associated with satiety. High GI carbohydrates have been reported to reduce appetite in the short term, whereas low GI carbohydrates possess a more delayed effect on energy intake controls [ 55 ]. Fructose appears to have differing effects on appetite compared to glucose, contributing to its negative properties. Anderson et al . determined the association between food intake and blood glucose, comparing glucose and a fructose mixture. Glucose was administered as a high GI preload, which resulted in lower mealtime energy intakes compared to the low GI preload of the glucose-fructose mixture. An inverse relationship was seen between GI (and blood glucose concentrations), and appetite with consequent increased food intakes seen with fructose [ 56 ]. In 2002, Vozzo et al . studied the comparative effects of glucose and fructose on blood glucose, insulin, and acute food intake. When subjects drank equienergetic preloads of glucose or fructose before an ad libidum buffet lunch, glucose concentrations were lower in the fructose group compared to glucose, and insulin concentrations were 50% higher in the fructose group in type 2 diabetics than in non-diabetics. The authors concluded that fructose may be a suitable replacement for glucose in diabetic patients – although it was found that satiating efficiencies of fructose certainly offered no advantages [ 57 ]. This study differs from others with regards to insulin secretion, but the trend is clear between GI, glucose concentrations, and appetite. An explanation for the variation in glucose and fructose glycemic responses appears to be dependent on rates of hydrolysis and absorption of glucose, and gastric emptying [ 58 ]. The variations observed in GI and appetite control of glucose and fructose can also be explained by differences in stimulation of insulin and leptin, important players in the long-term regulation of energy homeostasis. Fructose will generally produce smaller insulin excursions upon consumption because it does not stimulate the secretion of insulin from pancreatic beta cells, whereas glucose does. Insulin-regulated leptin will also have a reduced concentration and a decreased net effect on reducing appetite. Limited effects on appetite suppression, combined with the fact that fructose is favoured by the liver to be metabolized into lipid, will subsequently lead to weight gain, hyperinsulinemia, and the associated insulin resistance [ 59 ]. Glucose and fructose comparison studies continued examining new hormonal targets. In 2004, Teff et al . showed that subjects served meals with either 30% glucose beverages, or 30% fructose beverages, had differing hormonal and metabolic responses. Glycemic excursions and insulin responses were reduced by 66% and 65%, respectively, in the fructose-consuming subjects. There was a concomitant reduction in circulating leptin both in the short and long-term as well as a 30% reduction in ghrelin (an orexigenic gastroenteric hormone) in the fructose group compared to the glucose group. A prolonged elevation of TG was also seen in the high fructose subjects [ 60 ]. Both fat and fructose consumption usually results in low leptin concentrations which, in turn, leads to overeating in populations consuming energy from these particular macronutrients. An adipocyte hormone, adiponectin, also plays an important role in lipid homeostasis and insulin action [ 61 ]. The insulin sensitizer agonist, peroxisome proliferator-activated receptor-gamma, stimulates adiponectin production and adiponectin is in fact thought to be part of this agonist's mechanism lowering circulating fatty acids and increasing fat oxidation. The net effect is to decrease liver TG and increase insulin sensitivity [ 62 ]. Chronic fructose consumption reduces adiponectin responses, contributing to insulin resistance [ 63 ]. Animal studies have illustrated various differences between glucose and fructose metabolism. In 2002, Miller et al . injected fructose into the cerebroventricles of rats, and observed enhanced food intake, whereas similar concentrations of injected glucose suppressed appetite-agonist stimulated food intake [ 64 ]. Feeding rats either 32% glucose, fructose, or sucrose solutions, resulted in increased weight gain, and energy consumption compared to chow fed controls. Rats given the fructose and sucrose solutions also had a decreased ability to tolerate a glucose load, and fructose animals had greater serum TG levels over all other conditions ([ 65 ]. This is likely because the hepatic metabolism of fructose favours de novo lipogenesis. In combination with alterations in insulin signaling and leptin regulation, weight gain and unregulated energy intake can occur [ 33 ]. In 1986, Levine et al . found that fructose, administered in the form of the disaccharide sucrose, promotes obesity more than glucose because fructose does not stimulate thermogenesis [ 58 ]. These hormonal and physiological changes illustrate the important connections between energy intake, appetite control, weight gain, and insulin resistance. Fructose and insulin resistance Increasingly, questions have been raised as to whether dietary carbohydrate and fructose intake are directly related to the development of type 2 diabetes. As insulin resistance is often associated with circulating C-peptide concentrations, a cross-sectional study was performed to assess dietary fructose and carbohydrate, and glycemic loads related to C-peptide concentrations. It was found that the highest quintile of fructose intake had 13.9% higher C-peptide concentrations than the lowest quintile. Of note, subjects with high intakes of cereal fiber had 15.6% lower C-peptide concentrations, indicating that these types of nutrients may have opposing roles in the development of insulin resistance [ 66 ]. A definite relationship has also been found between metabolic syndrome and hyperhomocysteinemia, which is associated with cardiovascular and cerebrovascular diseases. Rats fed a fructose-enriched diet had a 72% higher homocysteine levels after 5 weeks compared to chow-fed controls [ 67 ]. Elevated homocysteine levels are an important risk factor for vascular disease. Homocysteine was found to be higher in patients with stenotic vessels and coronary artery disease scores, and was in fact highest in diabetic patients [ 68 ]. This is consistent with the increased TG, very low density lipoprotein (VLDL) secretion, and atherosclerosis associated with chronic fructose feeding. Although fructose does not appear to acutely increase insulin levels, chronic exposure seems to indirectly cause hyperinsulinemia and obesity through other mechanisms. One proposed mechanism involves GLUT5, a fructose transporter that is found to have significantly higher expression levels in young Zucker obese rats compared to lean controls. As the rats age and become diabetic, GLUT5 abundance and activity is compromised, causing an even more marked insulin resistance over lean rats, implying a possible role of GLUT5 receptors in the pathology of metabolic syndrome associated with fructose feeding and insulin resistance [ 69 ]. In rats fed 66% fructose for 2 weeks, insulin receptor mRNA, and subsequent insulin receptor numbers in skeletal muscle and liver were significantly lower compared to rats fed a standard chow diet. Also, blood pressure and plasma TG increased in the fructose-fed rats, even though there was no change in plasma insulin, glucose, or body weight [ 70 ]. Evidence shows these early steps in insulin signaling are important for insulin's metabolic effects. In a different study, it was found that after 28 days of fructose feeding there were no changes in insulin receptor concentration, but, insulin stimulated autophosphorylation, a mechanism necessary for insulin action, was reduced to 72% in the liver. Insulin receptor substrate (IRS) protein levels were similar, but there were significant decreases in insulin induced IRS (1/2) phosphorylation in both the liver and muscle of the fructose fed rats [ 71 ]. These changes are important, because it has been shown that the products of these insulin independent metabolic pathways lead to polyol formation and advanced glycation end products, which can contribute to the numerous complications and premature atherosclerosis seen in diabetic patients [ 58 ]. It is also known that such inflammations can lead to the pathogenesis of diabetes, and there is strong evidence suggesting that increased free fatty acids (FFA) in diabetic subjects and fructose fed models play a role in the inflammatory state of insulin resistance. If FFA are not removed from tissues, as occurs in fructose fed insulin resistant models, there is an increased energy and FFA flux that leads to the increased secretion of TG. Insulin resistance has also been correlated with intracellular TG stores, which are involved in lipotoxicity and beta cell failure leading to diabetes [ 72 ]. Another theory explaining how chronic fructose overnutrition can lead to type 2 diabetes is the hexosamine hypothesis, where hexosamine flux is thought to regulate glucose and satiety-sensing pathways. With overexpression of glutamine:fructose-6-phosphate amidotransferase, the key regulatory enzyme in hexosamine synthesis, the liver produces excess fatty acids, skeletal muscle becomes insulin resistant, and hyperinsulinemia results. This pathway of excess hexosamine flux leads to long-term storage of energy, and eventually obesity and type 2 diabetes [ 73 ]. Fructose: a highly lipogenic nutrient There is considerable evidence supporting the ability of high fructose diets to upregulate the lipogenesis pathway, leading to increased TG production [ 74 ]. Insulin and glucose are known to directly regulate lipid synthesis and secretion. Insulin controls hepatic sterol regulatory element binding protein (SREBP) expression, which is a key transcription factor responsible for regulating fatty acid and cholesterol biosynthesis. SREBP binds to sterol responsive elements (SRE) found on multiple genes, and can activate a cascade of enzymes involved in cholesterol biosynthetic pathways, such as HMG-CoA reductase [ 75 ] and fatty acid synthase (FAS) [ 76 ]. Miyazaki et al . reported an induction of the hepatic SREBP-1 isoform and lipogenic gene expression including FAS, acetyl-CoA carboxylase (ACC), and stearoyl-CoA desaturase (SCD) in mice following 7 days on a 60% fructose diet [ 77 ]. It is known that SREBPs are regulated by intracellular sterol concentrations. However, more recently, it has been established that hormones such as insulin and platelet derived growth factor play a role in regulating these transcription factors. Expression of SREBP is enhanced by insulin in all three major insulin target tissues, liver, fat, and skeletal muscle [ 78 - 81 ]. Similarly, levels of SREBP are enhanced in the presence of hyperinsulinemia [ 82 , 83 ]. There is evidence that the insulin-mediated stimulation of SREBP occurs through the MAP kinase pathway [ 84 ], with ERK1/2 being shown to activate the SREBP-1a isoform by phosphorylating serine 117) [ 85 ]. Despite the fact that SREBP-1 is directly stimulated via insulin signaling, the depletion of insulin and insulin signaling through streptozotocin (STZ) treatment paradoxically induces SREBP-1c expression upon glucose, fructose, or sucrose feeding. It would have been expected that SREBP-1c would be downregulated concomitantly along with the reduced insulin availability, but this is not the case. Glucose feeding causes a short-term peak induction, whereas fructose caused a gradual extended increase in SREBP-1c activity, providing evidence that lipogenesis can be independent of insulin signaling, given carbohydrate, and particularly fructose, availability [ 86 ]. Emerging evidence suggests that a protein phosphatase, known as PTP-1B, may link high carbohydrate feeding, insulin resistance, and lipogenesis. Recently, PTP-1B has been linked to lipogenesis and SREBP regulation. Shimizu et al . found that overexpression of protein tyrosine phosphatase 1B (PTP-1B), which is associated with dysfunctional insulin signaling, leads to increased mRNA and promoter activity of SREBP-1c, and subsequent increases in the expression of FAS. PTP-1B may therefore regulate the lipogenesis and hypertriglyceridemia associated with insulin resistance syndrome [ 87 ]. In insulin resistant fructose fed rats, it has been reported that the increase of hepatic SREBP-1 mRNA [ 88 ] occurs in correlation with an increased PTP-1B expression [ 87 ]. The authors established a role for PTP-1B in enhancing SREBP-1 gene expression through upregulation of Sp1 transcriptional activity, via an increase in protein phosphatase 2A activity [ 87 ]. FAS, an important downstream component of lipid synthesis, was extensively studied in rat livers. Dietary carbohydrates increased the transcriptional rate of FAS in comparison to proteins. Specifically, fructose feeding increased FAS mRNA concentrations, and somewhat increased transcriptional rate. This suggests that fructose may increase the stability of FAS mRNA, while carbohydrates stimulate FAS through increased transcriptional rate [ 89 ]. Other studies using animal models of insulin resistance, for example, the Wistar fatty rats, showed the effects of dietary carbohydrates on TG production. Feeding rats fructose stimulated FAS, and created a 56% increase in TG secretion rate, and an 86% increase in plasma TG. Feeding glucose, however, did not have this effect on TG production, nor did it affect induction of FAS. This is likely because glucose stimulates both TG production, and TG removal, maintaining homeostasis. Fructose stimulates TG production, but impairs removal, creating the known dyslipidemic profile [ 90 ]. The human liver possesses a large capacity to metabolize fructose to lipids because of its ability to shunt metabolism toward serum TG production. TG stores supply an energy 'sink', providing an almost unlimited TG production capacity. Conversely, glucose as opposed to fructose would decrease serum TG [ 91 ]. As discussed earlier, the effects of fructose in promoting TG synthesis are independent of insulinemia. Hirsch argued that carbohydrate overload results in elevated TG because the large amounts of sugar that need to be absorbed so rapidly from the intestine lead to the involvement of other metabolic pathways, such as the hexose monophosphate shunt, that that favour the synthesis of FFA [ 92 ]. Again, the liver takes up dietary fructose rapidly where it can be converted to glycerol-3-phosphate. This substrate favours esterification of unbound FFA to form the TG [ 93 ]. It has also been found that increases of 1,2-sn-diacylglycerol and elevated expression of a PKC isoenzyme are associated with the enhanced synthesis of TG observed with high fructose diet models [ 94 ]. In these scenarios, where there is excess hepatic fatty acid uptake, synthesis and secretion, 'input' of fats in the liver exceed 'outputs', and hepatic steatosis occurs [ 95 ]. The mechanisms of steatosis and liver enlargement due to fructose intake are not well understood, but it is believed to be related to microsomal enzyme induction, increased storage of lipids, peroxisome proliferation, and hyperfunction due to excessive hepatic 'workloads' [ 96 ]. All of these factors contribute to fructose being a highly lipogenic nutrient, and to the resultant hepatic steatosis. Mechanisms of fructose induced lipoprotein overproduction There is growing evidence that the insulin resistant state developed upon fructose feeding is also associated with stimulated hepatic VLDL secretion. Several animal models have been employed to examine the mechanisms of this induction of VLDL, and the subsequent increases in plasma TG observed. Mechanistic studies based on carbohydrate versus lipid metabolism have recently become important because carbohydrate induced hypertriglyceridemia shares a metabolic basis with high fat diet induced endogenous hypertriglycerolemia. The similarly induced dyslipidemias would therefore have the same or similar atherogenic risks [ 97 ]. Carbohydrate induced hypertriglycerolemia results from a combination of both TG overproduction, and inadequate TG clearance [ 97 , 98 ]. These disease processes and the hepatic steatosis caused by stimulated lipogenesis have been illustrated by fructose fed animal models showing how aberrant leptin signaling, hyperinsulinemia, and dyslipidemia are related to TG induction [ 95 ]. Animals maintained on a chronic high fructose diet develop elevated NEFA and hyperinsulinemia at the expense of glycemic control [ 99 ]. This is not surprising, as fructose-induced metabolic dyslipidemia is usually accompanied by whole body insulin resistance [ 100 ] and reduced hepatic insulin sensitivity [ 101 ]. In the fructose fed hamster model, animals showed decreased glucose disappearance rates, increased plasma NEFA and increased plasma and liver TG [ 27 ]. Figure 1 (adapted from ref. 100) shows clear in vivo evidence of fructose-induced insulin resistance as assessed by euglycemic hyperinsulinemic clamp studies. Taghibiglou et al . further characterized the fructose fed hamster model demonstrating the development of a metabolic dyslipidemic state characterized by high plasma levels of VLDL-TG and apolipoprotein B (apoB) due to hepatic lipoprotein overproduction [ 100 ]. Serum TGs are elevated via both an increased secretion, and decreased clearance of VLDL [ 102 ]. Also, high rates of lipolysis in visceral adipose depots can increase availability of NEFAs and promote hepatic TG synthesis. The TG is then packaged with apoB, and secreted as VLDL particles [ 93 ]. Evidence has shown that there is a complex interplay of cellular enzymes regulating lipid synthesis and uptake, as well as export and oxidation. Observations of the actions of insulin affecting lipid secretion as well as inhibition of TG has brought research interests towards the effects of chronic insulin stimulation on VLDL secretion and transport. Excess VLDL secretion has been shown to deliver increased fatty acids and TG to muscle and other tissues, further inducing insulin resistance [ 103 ]. Induced cellular changes include alterations in hepatic pyruvate dehydrogenase, changes in insulin signaling phosphorylation, and increases of inflammatory cytokines [ 104 , 105 ]. It is evident that the metabolic effects of fructose occur through rapid utilization in the liver due to the bypassing of the regulatory phosphofructokinase step in glycolysis. This in turn causes activation of pyruvate dehydrogenase, and subsequent modifications favoring esterification of fatty acids, again leading to increased VLDL secretion [ 53 ]. Increases in VLDL secretion can then lead to chain reactions in other lipoproteins and lipids, such as low density lipoprotein (LDL). Resultant LDL cholesterol levels induced by high fructose intake are illustrated by comparison of a diet including 20% fructose, contrasted to a starch diet of less than 3% fructose. The 20% fructose diet initiated a cycle of increased fasting serum total and LDL cholesterol of 9% and 11%, respectively, over the starch feeding [ 106 ]. Increased evidence was shown in transgenic apo AI-CIII-AIV mice, fed a fructose solution for 9 months, where differential expressions of the apo AI and apo AIV genes were found. This indicated general perturbations in response to dietary intakes, causing long-term adverse effects in this hyperlipidemia mouse model [ 107 ]. The male Wistar fatty rat model of obese type 2 diabetes has also shown hyperglycemia. Remarkably, the female Wistar rats only develop this hyperglycemia when given sucrose, containing the responsible element of fructose, which causes increases in gluconeogenic enzymes and decreases in glucokinase. A hypertriglyceridemic effect is seen, presumably due to hepatic overproductions of VLDL and induction of lipogenic enzymes via dietary fructose [ 108 ]. Figure 1 Fructose-induced insulin resistance: evidence from euglycemic hyperinsulinemic clamp studies. Mean glucose levels ( A ) were slightly but significantly higher in fructose-fed vs. control animals during the last 30 mins of the clamp period (p < 0.01). Mean insulin levels ( B ) were slightly but not significantly higher in the fructose-fed vs. control hamsters during the clamp period. The glucose infusion rate (Ginf) ( C ) during the clamp period was significantly lower in fructose-fed vs. control animals (p < 0.01). The calculated insulin sensitivity index (S I – see methods) ( D ) was also significantly lower in the fructose-fed vs. control hamsters (p = 0.03). Fructose-fed (n = 9), control hamsters (n = 10). ( adapted from Taghibiglou et al. [100] ). Another contributing factor to VLDL overproduction includes fructose effects on lipid peroxidation. High fructose diets can have a hypertriglyceridemic and pro-oxidant effect, and fructose fed rats have shown less protection from lipid peroxidation. Replacing the fructose in these diets with a more natural source of high fructose, honey, reduces this susceptibility and lowers plasma nitrite and nitrate levels [ 109 ]. In 2004, Kelley et al . hypothesized that pro-oxidant stress response pathways may mediate hepatic increases in VLDL secretion and delayed clearance upon fructose feeding. Hypertriglyceridemic fructose fed rats were treated with lipoxygenase inhibitors, which reversed the inflammatory protein activity response, and the lipid dysregulation observed [ 102 ]. Recent findings have also shown that the hyperlipidemic and pro-oxidant effect induced by a high fructose diet can be decreased by oligofructose consumption. Oligofructose administered to fructose fed rats did not alter insulin concentrations, and lowered plasma leptin by 50% compared to control groups. Oligofructose prevented TG changes induced by fructose feeding, and decreased hepatic TG accumulation. The peroxidation effect of fructose was also decreased by oligofructose, and had beneficial protective effects [ 110 ]. Oxidative stress has often been implicated in the pathology of insulin resistance induced by fructose feeding, and lipid peroxides, diene conjugates, and reactive substances are undeniably elevated in fructose fed animals, especially accompanying a deficient antioxidant system. Administration of alpha-lipoic acid (LA) has been shown to prevent these changes, and improve insulin sensitivity [ 111 ]. LA treatment also prevents several deleterious effects of fructose feeding: the increases in cholesterol, TG, activity of lipogenic enzymes, and VLDL secretion, the reductions in lipoprotein lipase and HDL cholesterol and may even normalize a dyslipidemic cholesterol distribution of plasma lipoproteins [ 112 ]. Taken together, this evidence shows a clear role of peroxidative stress pathways involved in VLDL oversecretion. Observations made in our own laboratory have also shown aberrant lipogenesis activity. In primary hepatocytes isolated from fructose fed hamsters, there were significant increases in LXRα, SREBP-1, FAS and SCD, which indicate increased activity of the lipogenic pathways (unpublished observations). Fructose has also been implicated in reducing PPARα levels in rat hepatocytes. PPARα is a ligand activated nuclear hormone receptor that is responsible for inducing mitochondrial and peroxisomal β-oxidation. Nagai et al . found that following 8 weeks of a high fructose diet, rats showed decreased PPARα mRNA and protein levels [ 88 ]. In addition, primary rat hepatocytes treated with fructose also showed decreased PPARα expression, suggesting that fructose or its metabolites can directly regulate lipid oxidation. We have also recently detected decreased mRNA levels of PPARα in both liver and intestine of the fructose fed hamster (unpublished observations). Hence, decreased PPARα expression can result in reduced oxidation, leading to cellular lipid accumulation. For example, PPARα null mice have extensive hepatic steatosis because of diminished β-oxidation capacity, such as seen in the insulin resistant state [ 113 ]. Other mechanisms have been illustrated by Taghibiglou et al ., who found evidence for enhanced lipoprotein assembly, reduced intracellular apoB degradation, and increased microsomal triglyceride transfer protein (MTP) mass, mRNA and activity in the fructose fed hamster [ 100 ]. These metabolic changes also coincided with a decrease in ER-60, a cysteine protease that may play a role in apoB degradation, and an increase in synthesis and secretion of apoB [ 101 ]. It appears that a complex relationship exists in the fructose fed animal model that links insulin resistance and dyslipidemia through NEFA flux, SREBP-1 expression, de novo lipogenesis and MTP expression. Amplified MTP activity and expression would be expected to stimulate the assembly and secretion of apoB-lipoproteins, as an association has been demonstrated between MTP levels and VLDL production [ 114 ]. As insulin is a negative regulator of MTP gene expression [ 115 ], the upregulation of MTP that has been observed in insulin resistance states is predictable. MTP is also negatively regulated by SREBP through sterol response element (SRE) regions located within -124 and -116 of the 5' MTP gene promoter [ 116 ]. However, in fructose fed animals [ 87 ] as well as other models of insulin resistance [ 117 ] where increased levels of MTP and SREBP have been established, the regulatory effects of SREBP may play a minor role in regulating MTP expression. Rather increased hepatic NEFA and increased TG stores might stimulate MTP expression [ 118 ]. Recent observations in our laboratory show that oleic acid can stimulate the MTP promoter and the stimulation occurs independently of SRE activity (unpublished observations). Thus, in insulin resistance states, increased MTP may occur through another mechanism that may block SREBP-mediated inhibition of the promoter. These phenomena help explain the increased assembly and secretion of apoB in fructose fed models. In addition, increased levels of small dense LDL particles have been observed in insulin resistant states [ 119 ]. Early studies by Verschoor et al . showed that fructose diets altered the structure and function of VLDL particles causing and increase in the TG: protein ratio, and an increased total cholesterol and phospholipid content [ 120 ]. LDL particle size has been found to be inversely related to TG concentration [ 121 ] and therefore the higher TG results in a smaller, denser, more atherogenic LDL particle, which contributes to the morbidity of the metabolic disorders associated with insulin resistance. Several theories are proposed for the overproduction of VLDL: more TG per VLDL particle, increases in particle number, changes in the production rates of VLDL TG or apoB, decreased TG clearance, increased lipoprotein lipase activity, and increased de novo lipogenesis. It is likely a combination of some or all of these factors that contribute to the elevated TG seen in a fructose rich carbohydrate fed model of metabolic disorder. High fructose, which stimulates VLDL secretion, may initiate the cycle that results in metabolic syndrome long before type 2 diabetes and obesity develop [ 103 ]. More recently, our studies have identified an interesting link between the development of insulin resistance and deregulation of intestinal lipoprotein metabolism [ 122 ]. Chronic fructose feeding stimulated intestinal secretion of apolipoprotein B48-containing lipoprotein particles accompanied by enhanced intestinal lipid synthesis in the form of free cholesterol, cholesterol ester, and triglyceride, as well as increases in both MTP mass and activity. These results suggest that in insulin resistant or diabetic animals, there may be a mechanism causing enhanced intestinal secretion of lipoproteins in the fasting state. Fructose feeding may enhance this basal level of lipoprotein secretion through increased de novo lipogenesis and increased MTP availability. Comparison of plasma lipoproteins from fructose-fed animals showed a significant shift toward secretion of larger, less dense, chylomicrons in the insulin resistant animals [ 123 ]. Concluding remarks The alarming increase in fructose consumption may be an important contributor to the epidemic of obesity and insulin resistant diabetes in both pediatric and adult populations. For thousands of years, the human diet contained a relatively small amount of naturally occurring fructose from fruits and other complex foods. Adaptation of humans to a high glucose/low fructose diet has meant that hepatic carbohydrate metabolism is designed to actively metabolize glucose with a limited capacity for metabolizing a small daily intake of fructose. The increasing application of high fructose sweeteners over the past few decades has resulted in a considerable rise in the dietary intake of fructose. A high flux of fructose to the liver, the main organ capable of metabolizing this simple carbohydrate, disturbs normal hepatic carbohydrate metabolism leading to two major consequences (Figure 2 ): perturbations in glucose metabolism and glucose uptake pathways, and a significantly enhanced rate of de novo lipogenesis and TG synthesis, driven by the high flux of glycerol and acyl portions of TG molecules coming from fructose catabolism. These metabolic disturbances appear to underlie the induction of insulin resistance commonly observed with high fructose feeding in both humans and animal models. Fructose induced insulin resistant states are commonly characterized by a profound metabolic dyslipidemia, which appears to result from hepatic and intestinal overproduction of atherogenic lipoprotein particles. Taking into consideration that a typical western diet not only contains high levels of fructose but is also rich in both fat and cholesterol, synergistic interactions among these nutrients can readily occur leading to a greater degree of insulin resistance and dyslipidemia. In conclusion, emerging evidence from recent epidemiological and biochemical studies clearly suggests that the high dietary intake of fructose has rapidly become an important causative factor in the development of the metabolic syndrome. There is an urgent need for increased public awareness of the risks associated with high fructose consumption and greater efforts should be made to curb the supplementation of packaged foods with high fructose additives. Figure 2 Hepatic fructose metabolism: A highly lipogenic pathway. Fructose is readily absorbed from the diet and rapidly metabolized principally in the liver. Fructose can provide carbon atoms for both the glycerol and the acyl portions of triglyceride. Fructose is thus a highly efficient inducer of de novo lipogenesis. High concentrations of fructose can serve as a relatively unregulated source of acetyl CoA. In contrast to glucose, dietary fructose does NOT stimulate insulin or leptin (which are both important regulators of energy intake and body adiposity). Stimulated triglyceride synthesis is likely to lead to hepatic accumulation of triglyceride, which has been shown to reduce hepatic insulin sensitivity, as well as increased formation of VLDL particles due to higher substrate availability, increased apoB stability, and higher MTP, the critical factor in VLDL assembly. List of Abbreviations ACC: acetyl-CoA carboxylase apoB: apolipoprotein B ERK: extracellular signal related kinase FAS: fatty acid synthase GSK-3: glycogen synthase kinase-3 HDL: high density lipoprotein HFCS: high fructose corn syrup IR: insulin receptor IRS: insulin receptor substrate LA: alpha-lipoic acid LDL: low density lipoprotein LXR: liver X receptor MAPK: mitogen activated protein kinase MTP: microsomal triglyceride transfer protein NEFA: non-esterified fatty acids PA-1: plasminogen activator inhibitor-1 PI3-kinase: phosphatidylinositol 3 kinase PKB: protein kinase B PKC: protein kinase C PPAR: peroxisome proliferator activated receptor PTP-1B: protein tyrosine phosphatase-1B SCD: stearoyl-CoA desaturase SREBP: sterol regulatory element binding protein TG: triglyceride VLDL: very low density lipoprotein	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552336.xml
524511	Characterization of an endogenous retrovirus class in elephants and their relatives	Background Endogenous retrovirus-like elements (ERV-Ls, primed with tRNA leucine) are a diverse group of reiterated sequences related to foamy viruses and widely distributed among mammals. As shown in previous investigations, in many primates and rodents this class of elements has remained transpositionally active, as reflected by increased copy number and high sequence diversity within and among taxa. Results Here we examine whether proviral-like sequences may be suitable molecular probes for investigating the phylogeny of groups known to have high element diversity. As a test we characterized ERV-Ls occurring in a sample of extant members of superorder Uranotheria (Asian and African elephants, manatees, and hyraxes). The ERV-L complement in this group is even more diverse than previously suspected, and there is sequence evidence for active expansion, particularly in elephantids. Many of the elements characterized have protein coding potential suggestive of activity. Conclusions In general, the evidence supports the hypothesis that the complement had a single origin within basal Uranotheria.	Background ERV-Ls are retroviral elements (retroelements) lacking the envelope gene ( env ) and exhibiting homology to the class of human endogenous retroviruses designated as HERV-L by [ 1 ]. Similar retroelements have been identified in several eutherian groups (see below), but their incidence in metatherians and monotremes is not known at present. They presumably arose from successful germ-line infection by foamy -like viruses, but when or how many times this might have occurred during the course of eutherian evolution is unknown. From the perspective of evolutionary biology it is of great interest that some classes of ERVs are known to retain original functions, including the capacity to produce infectious viral particles [ 2 ]. Others have gained novel regulatory functions in the mammalian genome [ 3 ]. Formation of the human placenta may depend on expression of a HERV-W element env gene [ 4 ]. Human immunodeficiency virus (HIV) shares specific functionally homologous sequences with endogenous retroviruses, suggesting the possibility that recombination with ERVs could change the properties of exogenous retroviruses [ 5 ]. Thus, ERVs may serve as a variable pool from which exogenous viruses may diversify. Exogenous retroviruses may have originated from ERVs and ERV-Ls in particular may represent an intermediate between retrotransposons and exogenous viruses [ 6 ]. Comparison of ERV-L polymerase ( pol ) gene sequences from 22 mammalian species revealed ERV-Ls that have expanded in copy number and remained active over long periods of time [ 1 ]. Phylogenetic analysis of these sequences demonstrated that primates and rodent ERV-L sequences are both diverse and, with few exceptions, monophyletic, whereas carnivore and ungulate ERV-L sequences were polyphyletic. The phylogenetic picture reflects the particularly robust expansion of the primate and rodent ERV-L complement. Importantly, the primates and rodents were the only groups that included ERV-L sequences with protein coding potential and therefore potential transpositional activity. These points suggests that, if the history of active expansion of retroelements within a group can be deciphered, it might be possible to use this information in the same way that parasite data are conventionally used [ 7 ], to perform tests of host phylogenetic relationships that are at least logically independent of other data sources. In this connection, the superorder Uranotheria is of particular interest. Uranotheria [ 8 ] is the most recent nomen for a constellation of relationships that has, in fact, been supported by the majority of ungulate specialists throughout the past century. Simpson [ 9 ] grouped proboscideans, hyraxes, embrithopods and sirenians under the group-name Paenungulata, but was not certain of its monophyly. Most other authorities have supported this clade, albeit with some variation in content, in the years since Simpson's [ 9 ] publication (e.g., [ 10 - 14 ]). McKenna and Bell [ 8 ] divided Uranotheria into three major groups, Hyracoidea, Embrithopoda, and Tethytheria. The last is further subdivided into Sirenia and Behemota; behemotans consist of Proboscidea and Desmostylia. Only Hyracoidea, Sirenia, and Proboscidea possess living members. Morphologically, there is considerable evidence that supports the association of Proboscidea and Sirenia as sister taxa to the exclusion of Hyracoidea [ 10 ], and little that appears to contradict it. Fischer and Tassy [ 15 ] take the position that alleged hyracoid morphological resemblances to tethytheres are either convergences or misconstrued, on the argument that hyraxes are in fact perissodactyls or closely allied to them. This aspect of the Fischer-Tassy hypothesis is not supported by molecular data [ 14 , 16 ]. On the other hand, it must also be admitted that sequence data have not provided especially strong support for Tethytheria (and, by extension, the monophyly of Uranotheria) [ 17 ]. In the most recent exercise in this arena, Asher et al [ 18 ] were able to recover Tethytheria under certain conditions when fossil and morphological data were combined with sequence information, but not when sequence data were used alone. To investigate whether ERV-L and other retroelements may be useful in resolving phylogenetic questions involving uranotheres at multiple taxonomic levels, we utilized an ERV-L polymerase gene ( pol ) fragment using degenerate primers tested in other mammalian orders. Extending our previous work [ 19 ], we found that ERV-L sequence diversity was high in all members of this group and that phylogenetic analysis of our data to a limited extent supported Uranotheria as a distinct clade when sequences that lack coding potential are used. By contrast, sequences that are potentially active form separate monophyletic groups, indicating a more recent origin. Thus, it appears that ancient ERVs reflect the phylogeny of their host like classic genes and more recently active ERVs will tend to be more similar to one another as opposed to their host. Results Among-clone comparisons A ~330 bp PCR product was amplified for African elephant, Asian elephant, manatee, and rock hyrax. The products were cloned and 10 clones sequenced for each product. Of the 40 sequences thus developed, only one Asian elephant sequence had no homology to ERV-Ls and was removed from analysis (not shown). No identical sequences were shared among taxonomic groups. All nine Asian elephant and all 10 manatee clones were unique. However, one Asian elephant clone, designated Max3 (accession number AY394573), was a recombination product of clone Max2 (accession number AY394572) and clone Max6 (accession number AY394576). Whether this represents a PCR artifact or is a genomic recombination event is not known. However, it is not expected that recombinational PCR would be observed in modern undamaged DNA [ 20 ]. Among the African elephant runs, four clones differed at 0–1 positions. As PCR errors probably account for these minor differences we assume only 6 unique ERV-Ls were discovered for this individual. Similarly, the hyrax sample yielded 3 groups comprised of 2 identical sequences, while two other sequences differed at 5 positions. Thus, 5 unique ERV-Ls were also obtained for Procavia . Recovered sequences were compared to a mouse element with full coding potential in the gag and pol genes (MuERV-L, GenBank no. Y12713). Twelve clones were in frame with no stop codons. However, only 6 of the total 12 were unique (Figure 1 ). This is surprising, as 87 sequences from 22 mammalian species previously revealed only 7 sequences with coding potential [ 1 ]. Among the 39 sequences determined here, 6 unique sequences had coding potential among only 4 species. The observed sequence diversity and frequency of observed coding potential is consistent with active ERV-L expansion in these four species and consistent with results with a smaller internal fragment from the same groups (plus extinct woolly mammoth in the proboscidean sample) [ 19 ]. Figure 1 Endogenous retrovirus type L (ERV-L) phylogeny. Non-uranothere sequence designations taken from [1]. Uranothere designations are, Max ( Elephas maximus ), Lox ( Loxodonta africana ), Mana ( Manatus trichechus ), and Hyrax ( Procavia capensis ). Neighbor-joining tree of all uranothere sequences including representative ERV-L elements from other mammalian orders. "*" designates sequences with coding potential. Phylogenetic analysis A heuristic search of the entire uranothere ERV-L data set yielded 12 equally parsimonious trees, a strict consensus of which (Figure 2 ) showed poor recovery of accepted clades within Uranotheria and low bootstrap support at each node. While neighbor-joining analysis produces a tree with uranothere sequences as a monophyletic group, branch lengths in some cases were very short and bootstrap support under any method used was not statistically significant (Figure 1 ). On the assumption that sequences with potential RT activity may have been under different evolutionary constraints and may differ in their phylogenetic resolution, the sequences with no stop codons in the retrieved pol gene were analyzed separately from those with stop codons (Figure 3 , 4 and 5 ). Those with potential RT activity grouped as distinct monophyletic groups, possibly reflecting their more recent activity and thus showing closer affinity to one another as opposed to other related ERV-L sequences (Figure 3 ). Those with stop codons showed a different picture with modest support for Uranotheria as one might expect for single or low copy sequences that have been transmitted vertically over time (Figure 4 ). Likelihood analysis of the data produced similar groupings, though with weak support, suggesting the associations found are not an artifact of the phylogenetic analysis methodology (Figure 5 ). However, ME analysis did not produce statistically significant resolution for any ERV group examined including non-uranothere ERVs (not shown). Poor resolution within Uranotheria with all analyses could be due to several factors, although the likeliest is different ages of individual element copies. This is not unlike the situation with various sequences recovered from primates and rodents: some sequences reconstruct accepted ordinal groupings, while others do not. For example, in the study by Bénit et al. [ 1 ], one New World monkey sequence (As2) grouped with two dog sequences and not with other primates. In other mammals, sequences (e.g., those retrieved from cow and horse) were dispersed in no evident pattern. Figure 2 Maximum parsimony bootstrap consensus tree of all uranothere ERV-L sequences. Bootstrap values over 50% are shown. The scale bar indicates the number of steps. Figure 3 Bootstrap consensus tree of Uranothere ERV-L sequences with coding potential only. Figure 4 Bootstrap consensus tree of Uranothere ERV-L sequences without coding potential. Figure 5 Quartet Puzzle maximum likelihood tree of sequences without coding potential. Puzzle support for each node is indicated. Discussion The most important finding resulting from this study is that elements that have undergone expansion–i.e. have remained transpositionally active–are the ones that are most likely to group monophyletically and those that have not tend to be consistent in their higher-level taxonomic distribution (in this case, at the superordinal level). Thus, there are different elements within the same family demonstrating different evolutionary trajectories. It is reasonable to suppose that for some groups, such as rodents, primates, and uranotheres, continued expansion of active ERV-L elements was tolerated by the host. By contrast, in other ungulate and carnivore lineages active elements were not inherited or were silenced early during their evolution and ERV-L expansion did not occur. Older elements tend to evolve as typical orthologous sequences. An advantage of the great diversity of elements is that with a single PCR, cloning, and determination of multiple clone sequences one retrieves multiple independent sequences with which to do phylogenetic analysis. In these regards, the uranothere evidence is consistent with results previously reported for primates and rodents in which older elements and elements that have undergone bursts of transposition were found coexisting. In each of these groups there is now good sequence evidence for retroelements that have retained coding capacity, which is of some interest since HERV-L pol (for example) is known to be expressed in specific tissue types [ 21 ]. This indicates that, in addition to potential transpositional activity, retroelements other than syncytin may have acquired biological functions important for their hosts. Conclusions In terms of our general results, elephants and the manatee were found to contain the most diverse sequences, while hyrax showed comparatively less diversity. This finding is consistent with results from our previous study involving extinct elephantids [ 19 ]. Elements that have been investigated in other ungulates (bovids and suids) do not yield monophyletic groupings and are represented by low copy numbers [ 1 ], suggesting that expansion did not occur (or has not recently occurred) in these taxa. Although critical studies would have to be undertaken to demonstrate the matter conclusively, it appears that amplification and diversification of ERV-L elements were independent events in primates, rodents and uranotheres. Although this study shows that the value of ERV-Ls for the narrow purpose of phylogenetic reconstruction is limited at higher taxonomic levels, ERV-L is only one class among many different groups of ERVs in mammalian genomes. Some regions of ERVs have been used successfully to reconstruct phylogenies at lower taxonomic levels [ 22 ]. Additional transposable elements could serve as phylogenetic markers in a manner similar to ERV-L in the present study, while providing multiple independent sequences to test ordinal level phylogenies. Methods Samples African elephant ( Loxodonta africana ) DNA was supplied by N. Georgiadis of the M'Pala Research Centre, Kenya. Asian elephant ( Elephas maximus ) blood was provided by J. Hektor of the Tierpark Hellabrunn, Munich. Manatee ( Trichechus manatus ) blood was provided by D. Murphy of the Lowry State Park, Florida. Hyrax ( Procavia capensis ) muscle was provided by G. Amato of the Bronx Zoo, New York. DNA extractions One ml of blood or approximately one gram of tissue was incubated in 1–2 ml 10 mM Tris-/Cl (pH 7.5), 10 mM EDTA (pH 8.0), 50 mM NaCl, 2% SDS, and 0.6 mg/ml Proteinase K overnight at 37 C, extracted with phenol and chloroform, and subsequently concentrated with 50 ul Millipore Ultrafee MC 30,000 NMWL columns or precipitated in 2.5 volumes ethanol and 1% NH4 Oac. PCR, cloning, and sequencing PCR primers for the ERV-L pol gene are described in [ 1 ]. Three μl of extract was added to 50 μl PCR containing standard buffer supplied by Boehringer Mannheim and 30 PCR cycles performed. PCR products were cloned using the pGEM-T cloning system (Promega). After heat shock into bacteria, ampicillin and blue/white selection, colonies were picked with a sterile pipette tip and added to 30 μl PCR reactions where M13 forward and reverse primers were used to amplify inserts for 25 cycles using the same buffer system described for ERV-L amplifications and as described in [ 19 ]. Five μl of the colony PCR products were visualized on ethidium-stained gels. Insert positive PCR reactions were purified with QIAquick columns and sequenced with T7 and SP6 primers using an ABI 377 sequencer. Phylogenetic analysis Alignment Representatives of each clade determined by Bénit et al. [ 1 ] were included in an alignment with the elephant, manatee, and hyrax sequences determined. HERV-L, X89211; MERV-L, Y12713; NWM (AS2), AJ233633; Lemur CM8, AJ233645; horse1, AJ233650; horse24, AJ233654; horse26, AJ233655; horse27, AJ233656; pig1, AJ233661; cow1, AJ233662; cow2, AJ233663; dog1, AJ233665; rabbit4, AJ233627. Alignments were performed using ClustalX [ 23 ] and adjusted where necessary. Elephant, manatee and hyrax sequences have been deposited in GenBank (accession numbers AY394571-AY394609) Phylogenetic methods Maximum parsimony and neighbor joining analysis was performed using PAUP 4.0b [ 24 ]. Heuristic searching including all uranothere ERV-L sequences yielded 12 maximum parsimony trees (MPTs). 100 bootstrap replicates were performed to test MPT robustness, the strict consensus of which is shown in Figure 2 . A bootstrap consensus tree using only sequences with no stop codons yielded 2 trees. The strict consensus of these trees following 1000 bootstrap replicates is shown in Figure 3 . A consensus tree of 1000 bootstrap replicates, this time excluding sequences with coding potential, is shown in Figure 4 . Maximum likelihood was performed using quartet puzzle in PAUP 4.0b after determining the evolution model as HKY +G using Modeltest 3.5 [ 25 ]. 10,000 puzzling steps were employed to determine the tree topology. Minimum evolution trees were generated using the program Mega2 [ 26 ]. Authors' contributions ADG oversaw the molecular genetic studies, coordinated the study, participated in the sequence alignment and phylogenetic studies and co-drafted the manuscript. CCE participated in the phylogenetic analysis. RDEM participated in the design and coordination of the study and co-drafted the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524511.xml
522821	Evaluating the cytotoxicity of innate immune effector cells using the GrB ELISPOT assay	Background This study assessed the Granzyme B (GrB) ELISPOT as a viable alternative to the 51 Cr-release assay for measuring cytotoxic activity of innate immune effector cells. We strategically selected the GrB ELISPOT assay because GrB is a hallmark effector molecule of cell-mediated destruction of target cells. Methods We optimized the GrB ELISPOT assay using the human-derived TALL-104 cytotoxic cell line as effectors against K562 target cells. Titration studies were performed to assess whether the ELISPOT assay could accurately enumerate the number of GrB-secreting effector cells. TALL-104 were treated with various secretion inhibitors and utilized in the GrB ELISPOT to determine if GrB measured in the ELISPOT was due to degranulation of effector cells. Additionally, CD107a expression on effector cells after effector-target interaction was utilized to further confirm the mechanism of GrB release by TALL-104 and lymphokine-activated killer (LAK) cells. Direct comparisons between the GrB ELISPOT, the IFN-γ ELISPOT and the standard 51 Cr-release assays were made using human LAK cells. Results Titration studies demonstrated a strong correlation between the number of TALL-104 and LAK effector cells and the number of GrB spots per well. GrB secretion was detectable within 10 min of effector-target contact with optimal secretion observed at 3–4 h; in contrast, optimal IFN-γ secretion was not observed until 24 h. The protein secretion inhibitor, brefeldin A, did not inhibit the release of GrB but did abrogate IFN-γ production by TALL-104 cells. GrB secretion was abrogated by BAPTA-AM (1,2-bis-(2-aminophenoxy)ethane-N,N,N', N'-tetraacetic acid tetra(acetoxymethyl) ester), which sequesters intracellular Ca 2+ , thereby preventing degranulation. The number of effector cells expressing the degranulation associated glycoprotein CD107a increased after interaction with target cells and correlated with the stimulated release of GrB measured in the ELISPOT assay. Conclusions Because of its high sensitivity and ability to estimate cytotoxic effector cell frequency, the GrB ELISPOT assay is a viable alternative to the 51 Cr-release assay to measure MHC non-restricted cytotoxic activity of innate immune cells. Compared to the IFN-γ ELISPOT assay, the GrB ELISPOT may be a more direct measure of cytotoxic cell activity. Because GrB is one of the primary effector molecules in natural killer (NK) cell-mediated killing, detection and enumeration of GrB secreting effector cells can provide valuable insight with regards to innate immunological responses.	Background Cytotoxic T-lymphocytes (CTL) and natural killer (NK) cells play an important role in host defense against intracellular pathogens and tumor cells. CTL recognize target cells through processed antigenic peptides presented via MHC. In contrast, NK cells mediate lysis of numerous cellular targets without classical MHC restriction. NK cells appear to use a variety of different, non-rearranging receptors to initiate cytoxicity and cytokine production [ 1 ]. Although CTL and NK differ in the receptors they use to recognize target cells, they both utilize the granule exocytosis and the Fas ligand (FasL)-mediated pathways to eliminate altered-self targets [ 2 - 6 ]. The granule-mediated pathway is the dominant pathway in CTL and NK [ 5 ]. CTL and NK cell granules contain a number of proteins, including perforin and granzymes, with GrB being the most abundant granzyme present [ 7 , 8 ]. Upon recognition and conjunction of the effector cell with the target, preformed granules containing GrB polarize in cytolytic lymphocytes at the point of contact and are secreted into the intercellular space formed between the effector and target cell [ 9 - 14 ]. The secretion of GrB occurs quite rapidly, is Ca 2+ -dependent, and mediates the lethal hit that kills virus-infected and tumor cells [ 2 , 7 , 8 , 10 , 15 - 19 ]. Cell-mediated cytotoxicity has conventionally been measured using the standard 51 Cr-release assay [ 20 ]. Recently, the use of the IFN-γ ELISPOT assay as a surrogate measure for CTL and NK responses has gained increased application. However, the IFN-γ ELISPOT assay may not be an accurate measure of cytotoxic lymphocytes as non-cytotoxic cells can secrete IFN-γ. Since GrB is exclusively present in the granules of CTL and NK cells and is a key mediator of the granule exocytosis-mediated cytolytic pathway [ 21 - 23 ], it is an excellent candidate marker for immunological monitoring of innate immunity by the ELISPOT method. The ELISPOT method has been successfully applied to measure GrB secretion by GrB-transfected CHO cells and for assessing antigen specific T-cell cytotoxic activity [ 24 , 25 ]. In this study, we utilized human NK-like and lymphokine-activated killer (LAK) effector cells to assess whether the GrB ELISPOT assay could accurately measure the MHC non-restricted cytolysis that occurs upon recognition of appropriate target cell ligands by activating receptors on these effector cells. Additionally, we evaluated whether the ELISPOT assay measured GrB release due to degranulation of stimulated effector cells. The GrB ELISPOT assay was strategically compared to the IFN-γ ELISPOT and the standard 51 Cr-release assays to determine if the GrB ELISPOT assay is a viable or better alternative to measure innate immunity. Methods Target cell lines K562 cells (Human myelogenous leukemia cell line, ATCC, Manassas, VA) were cultured at 37°C, 5% CO 2 in tissue culture medium (TCM) consisting of RPMI 1640 (BioWhittaker, Walkersville, MD) supplemented with 10% fetal bovine serum (FBS; Hyclone, Logan, UT), 1 mM non-essential amino acids, 2 mM glutamine, 100 U/ml Penicillin, 100 μg/ml Streptomycin, 20 mM HEPES and 1 mM sodium pyruvate (Gibco BRL Life Technologies, Grand Island, NY). TALL-104 effector cells The TALL-104 cell line is a T-cell derived line that is highly cytotoxic for NK-sensitive targets and when grown in the presence of IL-2, can be stimulated to secrete IFN-γ and GrB [ 26 , 27 ]. The cell line was cultured in Iscove's modified Dulbecco's medium (BioWhittaker) supplemented with 20% FBS (Hyclone), 4 mM glutamine, 100 U/ml Penicillin and 100 μg/ml Streptomycin (Gibco BRL Life Technologies) at 8% to 10% CO 2 . Recombinant human IL-2 (100 U/ml; Hoffmann-LaRoche, Nutley, NJ) was added every 2–3 days to ensure optimal growth and maintenance of cytotoxic activity. Generation of human lymphokine-activated killer (LAK) cells Peripheral blood mononuclear cells (PBMC) were isolated from venous blood of normal human volunteers by buoyant density centrifugation over Ficoll-Paque (Pharmacia, Piscataway, NJ). Aliquots of effector cells were cryopreserved in the vapor phase of liquid N 2 for future use in functional testing. The PBMC were thawed and resuspended at 2 × 10 6 cells/ml in 20 ml of RPMI 1640 (BioWhittaker) containing 10% human AB serum (Mediatech, Herndon, VA), 1 mM non-essential amino acids, 2 mM glutamine, 1 mM pyruvate, 20 mM HEPES, 100 U/ml Penicillin and 100 μg/ml Streptomycin (Gibco BRL Life Technologies). Cell suspensions were stimulated with 100 U/ml of IL-2 (Hoffmann-LaRoche) on day 0 and cultured for 5 to 6 days at 37°C, 5% CO 2 . LAK cultures consisted of 30.0 ± 2.1 % NK cells (CD3 - ,CD16/CD56 + ) and 26.9 ± 4.9 CD8 + T-cells as determined by flow cytometric analysis. Secretion Inhibitors TALL-104 cells were treated with inhibitors of cellular secretion prior to use in the GrB and IFN-γ ELISPOT assays. Inhibitors used included 1,2-bis-(2-aminophenoxy)ethane-N,N,N', N'-tetraacetic acid tetra(acetoxymethyl) ester (BAPTA-AM, Molecular Probes, Eugene, OR), a cell permeant chelator of intracellular Ca 2+ , and γ,4-dihydroxy-2-[6-hydroxy-1-hep-tenyl]-4-cyclopentanecrotonic acid λ-lactone (brefeldin A; Sigma, St. Louis, MO) which blocks protein secretion. TALL-104 cells were resuspended at 2.5 × 10 5 cells/ml and 2.5 × 10 4 cells/ml in PBS without divalent cations for BAPTA-AM and brefeldin A treatment, respectively. Cells were pretreated with the indicated concentrations of BAPTA-AM for 45 min or with 5 μg/ml brefeldin A for 1 h at 37°C, washed twice, assessed for viability by trypan blue exclusion, and resuspended at 2.5 × 10 4 cells/ml in assay media. 51 Cr-release assay Cytotoxicity of TALL-104 and LAK cells was assessed using the standard 51 Cr-release assay. Briefly, one million target cells were labeled at 37°C for 1 h with 100 μCi Na 2 51 CrO 4 (New England Nuclear, Boston, MA). Target cells were washed and resuspended in TCM at 5 × 10 4 cells/ml. Five thousand target cells per well (100 μl) were added to a 96 well plate (Costar, Cambridge, MA) following the appropriate number of effector cells (100 μl/well). The defined effector:target (E:T) ratios were plated in triplicate. Cytotoxicity assays were performed at 37°C for 4 h. Percent specific lysis was calculated using the following equation: (ER - SR)/(MR - SR) × 100, where ER = experimental release, SR = spontaneous release and MR = maximum release. Granzyme B ELISPOT assay Granzyme B secretion was measured using the GrB ELISPOT assay as previously described [ 25 ]. Briefly, MultiScreen-IP plates (PVDF membrane, Millipore, Bedford, MA) were coated overnight at 4°C with 100 μl/well of anti-human GrB antibody (7.5 μg/ml in PBS, clone GB-10, PeliCluster, Cell Sciences, Norwood MA). Effector cells (100 μl/well) were added to triplicate wells at specified concentrations followed by 5 × 10 4 target cells per well (100 μl). After the specified effector-target cell incubation, the plates were washed and 100 μl/well of biotinylated anti-human GrB detecting antibody (0.25 μg/ml in PBS/1% BSA/0.05% Tween 20, clone GB-11, PeliCluster, Cell Sciences) was added. Plates were incubated for 3 h and 50 μl of Streptavidin-Alkaline Phosphatase (1:1500 in PBS/1% BSA, Gibco BRL Life Technologies) was added for 1 h. Spots were visualized with 100 μl/well of BCIP-NBT phosphatase substrate (KPL, Gaithersburg, MD) and subjected to automated evaluation using the ImmunoSpot Imaging Analyzer system (Cellular Technology Ltd, Cleveland, OH). IFN-γ ELISPOT assay For assessing IFN-γ secretion, MultiScreen-IP plates (PVDF membranes, Millipore) were coated overnight at room temperature with 50 μl/well of anti-human IFN-γ antibody (20 μg/ml in PBS, Biosource, International, Camarillo CA) as previously described [ 25 ]. After effector and target cells were incubated at 37°C, the plates were washed and 50 μl/well of biotinylated anti-human IFN-γ antibody (1.3 μg/ml in PBS/1% BSA/0.05% Tween 20, BD PharMingen, San Jose, CA) was added. Plates were incubated with biotinylated antibody, washed and 50 μl of Streptavidin-Alkaline Phosphatase (1:1500 in PBS/1% BSA, Gibco BRL Life Technologies) was added. Spots were visualized and enumerated as described above. CD107a mobilization assay Degranulation of TALL-104 and LAK cells in response to target cell recognition was assessed by monitoring surface antigen expression of CD107a (lysosomal-associated membrane protein-1), a surface antigen transiently present on the cell surface after release of cytolytic granules. Expression of CD107a has been used as a marker to measure degranulation by flow cytometry [ 28 , 29 ]. To distinguish target cells from effector cells, K562 cells were first labeled with PKH67 dye (Sigma) according to the manufacturer's instructions and washed extensively prior to use in the assay. Effector cells (2 × 10 5 ) were resuspended in 400 μl of phenol red-free TCM in polystyrene tubes (Falcon, Franklin Lakes, NJ). PKH67-labeled K562 target cells (1 × 10 5 cells in 50 μl) were then added to appropriate tubes, the tubes were spun for 30 sec at 500 rpm and incubated for the specified periods in a 37°C water bath. Reactions were quenched with cold PBS and then 20 μl of CD107a-PE-Cy5 (PharMingen) was added. Mouse IgG 1 -biotin (Beckman-Coulter, Miami, FL) and streptavidin-PE-Cy5 (Caltag, Burlingame, CA) were utilized for controls. Tubes were incubated with antibody for 20 min at room temperature, washed with buffer and analyzed using a FACScan instrument (Becton-Dickinson Immunocytometry Systems, San Jose, CA) equipped with a 15 mW argon-ion laser. The percent of effector cells expressing CD107a was determined by gating the PKH67 negative (effector cells) population. Within the PKH67 negative gate, CD107a expression was determined from histograms based on forward and side scatter analysis. Effector cells run in the absence of target cells were evaluated for baseline expression of CD107a. The percentage of effector cells induced by target cells to express CD107a was calculated by subtracting background (spontaneous) expression from experimental samples. Statistical analysis Statistical analysis was performed using Student's t test and Pearson correlation coefficient (R 2 ). Results GrB release by stimulated TALL-104 cells GrB ELISPOT assays were performed using a constant number of K562 targets and decreasing numbers of TALL-104 effector cells per well. As the number of TALL-104 cells decreased, the resulting number of GrB spots decreased linearly. Limited GrB spots were detected in wells containing TALL-104 alone demonstrating that the ELISPOT assay measured primarily stimulated release of GrB (Fig. 1 ). K562 alone did not secrete GrB (data not shown). The correlation between the number of spots per well and the number of TALL-104 was highly significant with a Pearson correlation coefficient of R 2 = 0.98. Therefore, the GrB ELISPOT assay is capable of measuring the frequency of effector cells stimulated to secrete GrB. Figure 1 Granzyme B release by stimulated TALL-104 cells measured in the ELISPOT assay. TALL-104 cells (1 × 10 3 , 2 × 10 3 or 5 × 10 3 cells per well) were incubated with K562 targets (5 × 10 4 cells per well) for 4 h at 37°C in the Granzyme B ELISPOT assay (effector:target cell ratios are in brackets). Data are presented as average number of spots per well ± SD and are representative of 3 experiments with similar results. A significant correlation between the number of spots per well and effector cell number was observed (R 2 = 0.98). Correlation of GrB secretion and cytotoxic activity In tandem with the GrB ELISPOT assay, TALL-104 cells were also utilized as effectors against K562 targets in the standard 51 Cr-release assay to assess lytic activity. Significant specific lysis was observed at effector:target ratios from 10:1 to 1:1 (Table 1 ). In contrast to the ELISPOT assay, that can measure the frequency of cytotoxic cells at effector:target ratios as low as 0.02:1, specific lysis was not significant in the 51 Cr-release assay at cell ratios below 1:1 (Table 1 ). However, when optimal ratios were used in each individual assay, similar trends between the number of GrB spots per well in the ELISPOT and specific lysis in the 51 Cr-release assay were observed. These data demonstrate that the release of GrB is indicative of cytolytic activity. Table 1 Comparison of the GrB ELISPOT and 51 Cr-release assays for measuring TALL-104 cell activity. Effector:Target Ratio 51 Cr-Release Assay a (% Cytotoxicity ± SE) GrB ELISPOT Assay b (Spots/well ± SE) 10:1 65 ± 4 TNTC c 5:1 57 ± 10 TNTC 2.5:1 47 ± 11 TNTC 1:1 20 ± 5 TNTC 0.5:1 9 ± 2 TNTC 0.2:1 2 ± 1 316 ± 50 0.1:1 2 ± 1 144 ± 13 0.05:1 1 ± 1 80 ± 6 0.02:1 0 ± 1 33 ± 4 a The 51 Cr-release assay was performed with the effector:target cell ratio as specified for 4 h at 37°C. Data presented as average cytotoxicity ± SE (n = 3). b The granzyme B ELISPOT assay was performed with K562 cells at a constant number of 5 × 10 4 cells per well and effector:target cell ratio as specified for 4 h at 37°C. Data presented as average spots/well ± SE (n = 3). c TNTC= Too numerous to count accurately. Secretion kinetics of GrB and IFN-γ To measure the secretion kinetics of GrB compared to IFN-γ, TALL-104 cells were incubated with K562 target cells for the indicated times in the ELISPOT assays. Secretion of GrB was compared to IFN-γ because IFN-γ is currently the standard utilized to measure the frequency of cytotoxic cells via the ELISPOT method. GrB secretion was observed within 10 min of effector-target cell interaction with optimal secretion around 3–4 h (Fig. 2 ). This observation is in contrast to maximum IFN-γ production, which was observed at 24 h of effector-target interaction (data not shown). These data are consistent with rapid effector cell degranulation (within minutes) upon contact with target cells [ 18 , 19 ]. Figure 2 Kinetics of Granzyme B and IFN-γ secretion by TALL-104 cells in the ELISPOT assays. TALL-104 cells (2.5 × 10 3 cells per well) were incubated with K562 cells (5 × 10 4 cells per well) in the Granzyme B and IFN-γ ELISPOT assays for 0.2, 0.5, 1, or 4 h at 37°C. Data are presented as average spots per well ± SD. Results are corrected for background and are representative of 3 experiments with similar results. Mechanism of GrB secretion Effects of secretion inhibitors on the release of GrB and IFN-γ in the ELISPOT assays as well as expression of CD107a on effector cells were evaluated to determine the possible mechanism of GrB release measured in the ELISPOT assay. TALL-104 cells were treated with inhibitors of cellular secretion and assessed for their ability to release GrB and IFN-γ in the ELISPOT assays. BAPTA-AM chelates intracellular Ca 2+ and was utilized to block degranulation. Brefeldin A, which blocks protein secretion, was utilized to prevent secretion of newly synthesized proteins. Loading TALL-104 cells with BAPTA-AM resulted in a dose-dependent inhibition of GrB secretion (Fig. 3 ). Inhibition of GrB was not attributed to alterations in cell viability as TALL-104 cells loaded with as high as 100 μM of BAPTA-AM remained viable as assessed by trypan blue exclusion (data not shown). Brefeldin A treatment did not alter the secretion of GrB when TALL-104 cells were stimulated with K562 targets but did significantly ( p < 0.05) inhibit IFN-γ secretion (Fig. 4 ). Additionally, increased expression of the degranulation marker, CD107a, on TALL-104 cells was observed when TALL-104 cells were incubated with K562 target cells. Compared to TALL-104 cells alone, the addition of K562 targets induced 13.75% of TALL-104 cells to express CD107a (Fig. 5 ). Taken together, these data are consistent with synthesis of IFN-γ de novo while GrB is released from preformed granules upon effector-target interaction. Figure 3 Effect of BAPTA-AM on Granzyme B secretion by TALL-104 cells. TALL-104 cells were pretreated with the indicated concentrations of BAPTA-AM for 45 min at 37°C. TALL-104 cells (2.5 × 10 3 cells per well) were incubated with K562 (5.0 × 10 4 cells per well) for 4 h in the Granzyme B ELISPOT assay. Data are presented as spots per well ± SE (n = 3). Results are corrected for background. Figure 4 Effect of Brefeldin A on Granzyme B and IFN-γ secretion in the ELISPOT assays. TALL-104 cells were preincubated with brefeldin A (5 μg/ml; 1 h at 37°C) prior to use in the ELISPOT assays. TALL-104 cells (2.5 × 10 3 cells per well) were incubated with K562 (5 × 10 4 cells per well) for 4 h in the Granzyme B and 24 h in the IFN-γ ELISPOT assays at 37°C. Data are presented as percent inhibition ± SE (n = 3). Results are corrected for background. Figure 5 Stimulated TALL-104 cells express the degranulation marker CD107a. TALL-104 (2 × 10 5 cells) were incubated with K562 (1 × 10 5 cells) and surface expression of CD107a was determined by flow cytometric analysis. K562 were prelabeled with PKH67 dye to differentiate effector cells from targets. The histograms represent the PE-Cy5 CD107a positive cells determined by forward versus side scatter of the gated effector cells. TALL-104 cells alone are represented by the open histogram and TALL-104 cells incubated with K562 for 1 h are represented by the shaded histogram. The data is representative of 3 experiments with similar results. Application of the GrB ELISPOT assay to assess innate immunity To address the potential clinical relevance of the GrB ELISPOT assay, LAK cells derived from PBMC, rather than the TALL-104 cell line, were utilized as effector cells. As shown in Figure 6 , the number of GrB spots per well correlated with the number of LAK cells added, results comparable to those obtained with TALL-104 cells. Only wells containing both LAK and K562 target cells contained a substantial number of spots (Fig. 6 ). Figure 6 Granzyme B release by stimulated LAK cells measured in the ELISPOT assay. Human peripheral blood mononuclear cells (PBMC) were cultured with 100 U/ml of IL-2 for 5–6 days to induce LAK cells. LAK cells (1 × 10 3 , 3 × 10 3 or 1 × 10 4 cells per well) were incubated with K562 (5 × 10 4 cells per well) for 4 h at 37°C. Effector:target cell ratios are in brackets. Data are presented as average number of spots per well ± SD and are representative of 5 experiments with similar results. A significant correlation between the number of spots per well and effector cell number was observed (R 2 = 0.94). The secretion kinetics of GrB and IFN-γ by LAK cells were similar to that observed for TALL-104 cells. GrB secretion was seen within 10 min of effector-target cell interaction with optimal secretion around 3 h and maximal IFN-γ production was observed at 24 h of effector-target interaction (Fig. 7 ). Similarly, stimulated secretion of GrB by LAK cells was associated with increased surface expression of the degranulation marker, CD107a. The correlation between the number of GrB spots and the number of cells expressing CD107a per 100,000 effector cells was significant (R 2 = 0.89, Fig. 8 ). These data demonstrate that the GrB ELISPOT assay accurately measures the rapid release of GrB by stimulated LAK cells and can enumerate GrB-secreting effector cells. Figure 7 Kinetics of Granzyme B and IFN-γ secretion by LAK cells in the ELISPOT assays. Human peripheral blood mononuclear cells (PBMC) were cultured with 100 U/ml of IL-2 for 5–6 days to induce LAK cells. LAK cells (5 × 10 3 cells per well) were incubated with K562 cells (5 × 10 4 cells per well) in the Granzyme B and IFN-γ ELISPOT assays for 0.2, 0.5, 1, 4 or 24 h at 37°C. Data are presented as average spots per well ± SE (n = 6). Results are corrected for background. Figure 8 Correlation between CD107a expression measured in the flow cytometric assay and Granzyme B release measured in the ELISPOT assay. Human peripheral blood mononuclear cells (PBMC) were cultured with 100 U/ml of IL-2 for 5–6 days to induce LAK cells. LAK cells (5 × 10 3 cells per well) were incubated with K562 cells (5 × 10 4 cells per well) in the Granzyme B ELISPOT and 2 × 10 5 LAK cells were incubated with 1 × 10 5 K562 in the CD107a mobilization assay for 0.5, 1 and 3 h. The ELISPOT data are presented as spots per 1.0 × 10 5 effectors ± SE (n = 9) and the CD107a data as the number of positive cells per 1.0 × 10 5 effectors ± SE (n = 9). All results are corrected for background. A significant correlation between the number of GrB spots and CD107a positive effector cells was observed (R 2 = 0.89). LAK cells were utilized in the GrB ELISPOT, IFN-γ ELISPOT and 51 Cr-release assays to evaluate the different assays for measuring innate immunity (Table 2 ). Both ELISPOT assays were significantly more sensitive than the 51 Cr-release assay. At effector:target ratios of 20:1-5:1, significant specific lysis can be measured by the 51 Cr-release assay but the spots per well in the ELISPOT assays were too numerous to count accurately. Significant IFN-γ and GrB secretion was evident at effector:target ratios as low as 0.02:1 (1000 effectors/well) therefore, the optimal ratios for the ELISPOT assays are below the level of sensitivity of the 51 Cr-release assay (Table 2 ). However, when the optimal number of LAK cells are used in each individual assay, the amount of GrB and IFN-γ secreting cells in the ELISPOT assays and cytotoxicity in the 51 Cr-release assay have shown cross-correlation with R 2 values greater than 0.86 for all combinations. When comparing the two ELISPOT assays, the GrB ELISPOT is more rapid and may be a more direct measure of cytotoxic activity than the IFN-γ ELISPOT. Therefore, the GrB ELISPOT assay can elucidate the frequency of GrB secreting cells in primary LAK cultures and can be applied to measure innate immunity in clinically relevant samples. Table 2 Comparison of the GrB ELISPOT, IFN-γ ELISPOT and 51 Cr-release assays for measuring Lymphokine Activated Killer (LAK) cell activity. Effector:Target Ratio 51 Cr-Release Assay a (% Cytotoxicity ± SE) IFN-γ ELISPOT Assay b (Spots/well ± SE) GrB ELISPOT Assay c (Spots/well ± SE) 20:1 42 ± 10 TNTC d TNTC 10:1 30 ± 10 TNTC TNTC 5:1 22 ± 6 TNTC TNTC 2.5:1 10 ± 3 TNTC TNTC 1:1 5 ± 1 TNTC TNTC 0.4:1 4 ± 1 227 ± 16 282 ± 18 0.2:1 2 ± 1 182 ± 8 193 ± 12 0.1:1 2 ± 2 121 ± 14 117 ± 12 0.04:1 1 ± 0 61 ± 13 54 ± 11 0.02:1 1 ± 0 29 ± 8 37± 7 a The 51 Cr-release assay was performed with the effector:target cell ratio as specified for 4 h at 37°C. Data presented as average cytotoxicity ± SE (n = 3). b The IFN-γ ELISPOT assay was performed with K562 cells at a constant number of 5 × 10 4 cells per well and effector:target cell ratio as specified for 24 h at 37°C. Data presented as average spots/well ± SE (n = 3). c The granzyme B ELISPOT assay was performed with K562 cells at a constant number of 5 × 10 4 cells per well and effector:target cell ratio as specified for 4 h at 37°C. Data presented as average spots/well ± SE (n = 3). d TNTC= Too numerous to count accurately. Discussion NK cells are a key part of innate immunity due to their ability to secrete cytokines and mediate cytolytic activity and act as an important first line of defense against virally infected cells and tumor cells [ 30 ]. Although cytokine secretion by NK cells plays a role in regulating the adaptive immune response, cell-mediated cytotoxicity is the major effector function of NK cells. NK cells can mediate cytotoxicity by two main pathways, FasL-mediated and granule-mediated. The granule-mediated pathway, however, is dominant in NK cells with the release of GrB as one of the key factors in the lethal interaction that kills virus-infected and tumor cells [ 2 , 5 , 7 , 8 , 10 , 15 - 19 ]. Therefore, evaluating the secretion of this molecule provides an indirect measure of cell-mediated cytotoxicity mediated via the release of granules. The GrB ELISPOT assay has been previously applied to enumerate the frequency of antigen-specific T cells and this release of GrB by T cells was indicative of cytolytic ability [ 24 , 25 , 31 ]. Moreover, the GrB ELISPOT assay requires significantly less effector cells than the standard 51 Cr-release assay [ 24 , 25 ]. In this study, we demonstrated that the GrB ELISPOT assay can be utilized to determine the frequency and potential cytolytic ability of innate immune cells. When TALL-104 or LAK cells were used as effectors, our results were in excellent agreement with results obtained when GrB-transfected CHO cells, T-cell lines or CTL stimulated in vitro , were employed as effector cells [ 24 , 25 ]. TALL-104 or LAK cells secreted significant amounts of GrB only when stimulated with an NK-sensitive target cell line indicating that the GrB ELISPOT assay primarily detected stimulated, and not spontaneous (constitutive), release of GrB. Although assessing GrB release is an indirect measure of target cell lysis, a similar trend between the number of GrB spots per well in the ELISPOT assay and specific lysis in the 51 Cr-release was observed. Therefore, the GrB ELISPOT assay is a viable alternative to the 51 Cr-release assay for measuring MHC non-restricted cytolytic activity. Preformed granules that contain mature GrB are constitutively expressed in NK cells, thus NK cells are always armed with functional GrB. Preformed GrB is rapidly released upon recognition and conjunction of the effector cell and this process is Ca 2+ -dependent [ 18 , 19 , 32 ]. In contrast, IFN-γ is produced de novo upon activation and is secreted within hours [ 33 ]. The differences observed in the pattern of GrB and IFN-γ secretion suggest that the GrB measured in the ELISPOT assay is due to degranulation of pre-formed GrB whereas IFN-γ secretion results from protein synthesis de novo . To confirm that GrB measured in the ELISPOT assay was released via degranulation, effector cells were treated with inhibitors of cellular secretion: brefeldin A, BAPTA-AM and EGTA. Brefeldin A reversibly disrupts protein translocation from the endoplasmic reticulum to the Golgi apparatus, blocking the production and subsequent secretion of newly synthesized proteins [ 34 ]. BAPTA-AM and EGTA are both Ca 2+ -chelating agents and therefore can block Ca 2+ -dependent degranulation. Cell permeant BAPTA-AM contains acetoxymethyl ester groups that are removed by cytosolic nonspecific esterases, trapping the chelator within the cell. In a series of experiments, treatment of effector cells with brefeldin A significantly decreased IFN-γ, but not GrB secretion, whereas, BAPTA-AM abrogated GrB secretion but its effect on IFN-γ secretion was not apparent (data not shown). EGTA, which sequesters extracellular Ca 2+ , could also abrogate GrB secretion measured in the ELISPOT assay. However, significantly higher concentrations of EGTA (4–40 mM) than BAPTA-AM were needed to block GrB secretion (data not shown). These data are in good agreement with recently published findings [ 35 ] which demonstrate that higher concentrations of EGTA are needed to block degranulation compared to BAPTA-AM. Although adherence of cytolytic cells to their targets is a prerequisite for granule release, this interaction is dependent on Mg 2+ , but not Ca 2+ [ 19 ]. Thus, the decrease in GrB secretion after BAPTA-AM and EGTA treatment cannot be attributed to inhibition of adhesion between the effector and target cells, but results from inhibition of degranulation. Recently, a flow cytometric assay based on CD107a (lysosomal-associated membrane protein-1) mobilization was developed to measure degranulation of cytolytic cells [ 28 , 29 ]. CD107a is a vesicle membrane protein of cytolytic granules that is transiently expressed on the surface of effector cells during degranulation. Correlations between direct lytic ability and surface expression of CD107a on effector cells has been shown, indicating that CD107a expression is a reliable measure of cytolytic capacity [ 28 , 29 ]. In our study, CD107a expression correlated with GrB secretion (R 2 = 0.89). Both the release of GrB and the expression of CD107a was observed as early as 30 min after LAK cells were stimulated with relevant target cells, and increased with extended effector-target cell interaction. These data confirm that the GrB ELISPOT assay is an excellent measure of cytotoxic capacity mediated by effector cell degranulation. When the ELISPOT assays were directly compared to the 51 Cr-release assay, they demonstrated higher sensitivity with both TALL-104 and LAK cells as effector cells, data consistent with previous studies [ 24 , 25 , 36 , 37 ]. The ELISPOT assays enumerate antigen specific lymphocyte frequency by measuring secretion of specific immune proteins engaged in the specific pathway utilized to mediate lysis of target cells, whereas the 51 Cr-release assay measures cytotoxic ability regardless of the mechanisms of the killing. Therefore, unlike the 51 Cr-release assay, the ELISPOT assays are both qualitative and quantitative. It is important to emphasize that 51 Cr-release and the GrB ELISPOT assay measure different aspects of cell-mediated killing – target cell death and effector cell function, respectively. A limitation of the GrB ELISPOT assay is that it measures degranulation, not direct target cell lysis. As such, degranulation may not always equate to cell death if target cells contain serpin proteinase inhibitor 9 (PI9), a protein that inhibits the proteolytic activity of GrB, [ 38 ] or if effector cells are perforin deficient. The GrB ELISPOT assay also does not account for cytotoxicity mediated by FasL pathway. Therefore, when appropriate, the two assays should be used in concert. However, the high sensitivity and specificity of the ELISPOT assay are beneficial for monitoring clinical trials where frequently there are limited numbers of patients' cells available or target cells cannot be effectively labeled. Target cells that resist or do not tolerate labeling, such as some primary tumor cells, can still be utilized in the GrB ELISPOT assay. Additionally, compared to the 51 Cr-release assay, the GrB ELISPOT provides the precusory frequency of cells with the potential to kill targets. Although a number of flow cytometric assays have been developed to assess target cell cytotoxicity as well as identify the phenotype of the effector cells mediating the immune response, these assays are not as sensitive as the ELISPOT assay. The use of the IFN-γ ELISPOT assay as a surrogate measure for CTL and NK responses has recently gained increased application as an alternative to the 51 Cr-release assay [ 36 , 39 - 43 ]. However, the IFN-γ ELISPOT assay may not be an accurate measure of cytotoxic lymphocytes since 1) non-cytotoxic cells can secrete IFN-γ and 2) CTL with lytic activity do not always secrete IFN-γ [ 44 - 48 ]. The release of GrB is a more specific measure of cytotoxic lymphocytes than IFN-γ because the expression of GrB is restricted to CTL and NK cells [ 8 , 49 , 50 ]. Therefore, the GrB ELISPOT assay may be a more direct measure of innate immunity compared to the IFN-γ ELISPOT. The high sensitivity and specificity of the ELISPOT assays are beneficial for monitoring clinical trials where frequently there are limited numbers of patients' cells available. As such, simultaneous use of the IFN-γ and GrB ELISPOT assays may provide important immunological insight into patient responses that may then be directly assessed against clinical outcome. Conclusion Our data demonstrate that the GrB ELISPOT assay is a viable alternative to the standard 51 Cr-release assay to measure granule-mediated cytotoxicity and could be easily adapted to reliably and accurately measure MHC non-restricted cytotoxicity indicative of NK and LAK activity. The GrB ELISPOT assay measured GrB release due to degranulation of stimulated effector cells. Additionally, GrB is a more specific candidate marker than IFN-γ to measure the cytotoxic capacity of innate immune effector cells such as NK cells. Competing interests None declared. Author's contributions KSW and AM designed the study, analyzed the data and drafted the manuscript. KSW carried and prepared the cell lines and performed the majority of the immunogical assays. MWB performed and analyzed the flow cytometric CD107a assay data. SS participated in the design of the study and preparation of the manuscript. TS, DK and MB participated in the design of the study and served as expert advisors. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522821.xml
554784	Reporting quality of randomized trials in the diet and exercise literature for weight loss	Background To adequately assess individual studies and synthesize quantitative research on weight loss studies, transparent reporting of data is required. The authors examined the reporting quality of randomized trials in the weight loss literature, focusing exclusively on subject characteristics as they relate to enrollment, allocation, and follow-up. Methods An extensive literature review, which included a computerized search of the MEDLINE database, manual searches of bibliographic references, and cross-referencing of 92 review articles was conducted. A checklist, based on CONSORT recommendations, was used to collect information on whether or not authors reported age, gender, co-morbid disease, medication use, race/ethnicity, and postmenopausal status. Also tracked was whether or not initial and final sample size was reported and stratified by gender. Results Of 604 possible articles, 231 articles met eligibility criteria. Important subject characteristics were not reported as the following breakdown indicates: age (11%), gender (4%), race/ethnicity (86%), co-morbid disease states (34%), and medication use (92%). Additionally, 21% of articles failed to report initial sample size by gender while 69% neglected to report final sample size by gender. Conclusion Inadequate reporting can create difficulties with interpretation and can lead to biased results receiving false credibility. The quality of reporting for weight loss studies needs considerable improvement.	Background Current statistics indicate that approximately 64.5% of US adults can be considered overweight while 30.5% can be classified as obese [ 1 ]. Obesity is also linked to a variety of chronic diseases, and is associated with approximately 300,000 deaths each year and annual economic costs of over $117 billion [ 2 ]. According to the latest Behavioral Risk Factor Surveillance Survey (BRFSS) data, obesity and overweight continue to be major pubic health concerns, and reports indicate that no state had met the Healthy People 2010 objective of reducing obesity to 15% [ 3 ]. As a result, there is need for a strong evidence base for prevention and treatment strategies. Systematic reviews can offer the most reliable sources of information on which to develop guidelines and base treatment policy. However, such reviews need to contain a high proportion of all relevant evidence, which relates both to the need to find all trials and the need to analyze data on all participants [ 4 ]. To improve the quality of reporting for randomized controlled trials (RCTs) in the overweight and obesity literature, as well as any intervention trial, investigators are encouraged to publish reports in clear and unambiguous language, accounting for all events that occurred during the conduct of the investigation, and providing an accurate and thorough description of subjects who were selected, excluded, withdrawn or did not complete the study. Failing to report essential variables, such as subject characteristics, does not allow readers to adequately judge the validity or applicability of the study results, and may influence the interpretation of findings [ 5 ]. In response to the need for quality reporting, a panel of clinical investigators, epidemiologists, biostatisticians, and journal editors published a statement called the Consolidation of the Standards of Reporting Trials (CONSORT), which is designed to improve the standard of written reports of RCTs [ 6 ]. The CONSORT statement includes a checklist of 21 items and a flow diagram that can be used by authors to mark the page of the manuscript in which each of the items is addressed. In addition, the flow chart provides a detailed description of the progression of subjects through the intervention trial, from the number of potentially eligible participants for inclusion in the study to the number of intervention subjects in each group who completed the trial [ 6 ]. Using a similar format, we adapted the CONSORT form to examine the frequency of explicit reporting of subject characteristics as they relate to enrollment, allocation, and follow-up for weight loss studies published in the diet and exercise literature. Methods For the present study, we used information collected from a meta-analysis of the diet and exercise literature. The primary purpose of the meta-analysis was to statistically integrate and analyze published research studies on the effects of diet restriction only, diet restriction and exercise, or exercise only on weight loss, body composition, fat distribution, metabolism, and aerobic fitness. We examined the frequency of explicit reporting of subject characteristics (i.e., age, gender, co-morbid disease states, medication use (in addition to intervention drug), race/ethnicity, and postmenopausal status). We also examined whether or not initial and final sample size was reported and if so, we recorded if the final sample size was reported by gender. Search strategy. With the assistance of a clinical medical librarian, we developed search strategies, key words, and check tags to begin our initial search for relevant articles to include in the meta-analysis. We identified studies from a computerized search of the MEDLINE database (US National Library of Medicine) from 1966 – 2003, manual searches of bibliographic references of relevant published articles, and extensive cross-referencing and manual searches of 92 review articles. Overall study inclusion criteria Subject types. We required that intervention trials must study overweight or obese (BMI ≥ 25 kg/m 2 ) adult participants (≥ 18 years old). Ambulatory patients who were kept on wards exclusively for study purposes were also included. However, studies including pregnant women, patients with serious medical conditions, military personnel, and trained or professional athletes were excluded. Intervention types. For inclusion, trials were required to meet each of the following criteria: 1) active intervention trials with one group assigned to a weight-loss program involving energy restriction, exercise (aerobic or weight training) or both energy restriction and exercise; 2) outcome measures included weight loss; however, weight loss was not required to be a primary outcome; 3) the minimum duration of the intervention for weight loss must be greater than one day; and, 4) the article must be published in the English language. Case studies, crossover trials (due to possible carryover effects of weight reduction), and exploratory studies were excluded. We did not require studies to have been conducted in the United States. Pharmacotherapy, hormonal therapy, or surgical treatment studies, which did not include a separate group receiving a weight loss intervention (i.e., diet and/or exercise) were not included in the review process. Initial screening process. We independently screened all records resulting from our MEDLINE search strategies and manual searches by examining the titles and abstracts. We rejected articles if we could determine from the title and/or abstract that the study did not meet our selection criteria. If we could not determine whether or not to reject the article or if there was disagreement among reviewers, we retrieved the full text of the article for further evaluation. Quality control measures To control for abstraction bias, we trained five graduate research assistants to extract information from articles, using sample articles. Reliability checks were completed on the first several sample articles before graduate students were allowed to code independently. On a monthly basis thereafter, inter- and intra-rater reliability checks were completed to control for drift. All research assistants were required to attend weekly meetings with the principal investigator and co-investigators to discuss coding issues and to resolve coding disagreements. To control for potential selection biases after our initial screening procedure, we selected or rejected studies based on an examination of the research design and methodology. Acceptance or rejection for inclusion in the meta-analysis was not based on the study's outcome. We required coders to list specific reasons for all excluded studies. In addition, we tried to enhance data reporting for studies with missing covariate data by contacting authors by email or letter. Data entry quality control measures included the design of screens to match the coding sheet, automatic range limits placed on items entered, and queries to check for outliers. Values greater than two standard deviations from the mean were checked against the article. In addition, a random sampling of 58 articles (~25%) was pulled to check for accuracy. Standardized coding sheet for data abstraction. A separate checklist was devised based on CONSORT recommendations and quality indicators suggested by other authors [ 6 - 10 ]. Specific subject information was extracted from the meta-analysis database to examine if investigators reported age, race/ethnicity, and gender of their study subjects. Coders also recorded whether or not investigators reported the subjects' health status and presence of co-morbidities (i.e., diabetes, cardiac problems, cancer, or hypertension). Whether or not investigators reported medication use by subjects, other than the intervention drug, was also recorded. In addition, we tracked reporting of the postmenopausal status of subjects in investigations which included women 45 years or older. When analyzing the reporting quality for initial and final sample size separated by gender, only studies that included a group(s) consisting of both men and women were selected. Statistical analysis The frequency of reporting in the journal articles for the selected variables was obtained by performing a proc frequency using SAS (Version 8.2, Cary, NC). Results are presented as totals. Results Of the 604 randomized clinical trial articles that were reviewed, 231 (38%) met eligibility criteria for the present study. Figure 1 displays the frequency of articles included in the analysis by year of publication. Of these, 123 (53%) articles were randomized studies without a control group and 108 (47%) were randomized controlled trials. Seventy-six articles (33%) were dietary intervention only, 60 (26%) were exercise intervention only, and the remaining 95 articles (41%) included both exercise and dietary intervention for weight loss. Figure 1 Frequency of randomized trials published by year. Contribution of journals For this analysis, the American Journal of Clinical Nutrition contributed the most articles that met the eligibility criteria (47), followed by the International Journal of Obesity (32), Medicine & Science in Sports & Exercise (24), Journal of the American Medical Association (11), Metabolism (11), New England Journal of Medicine (7), Archives of Internal Medicine (6), and Journal of the American Dietetics Association (6). In addition, 52 journals had 5 or fewer articles in this analysis. Subject characteristics Of the 231 articles that met eligibility criteria for the present study, subjects' age was not reported in 25 (11%) articles. Investigators of ten (4%) different studies failed to report the gender of their study participants. Race and/or ethnicity of study subjects were not reported in 199 (86%) studies. In addition, 78 (34%) studies failed to report the health status (i.e., presence or absence of co-morbidities) of their study subjects at baseline. Further, medication use was not reported in 213 (92%) of the articles (Figure 2 ). For postmenopausal status, 89 articles were included in the overall analysis with only 7 (8%) studies that did not report the postmenopausal status of their female subjects. Figure 2 Proportion of selected subject characteristics not reported in randomized trials. Values reported are means. Co-morbid disease status included the reporting of diabetes, cardiac diseases, cancer, hypertension, or any other metabolic disease. Sample size and attrition Initial sample size was not reported in 14 (6%) studies while final sample size was not reported in 133 (58%) of the 231 eligible articles. When reporting initial and final sample size by gender, 68 articles were included in the analysis. Of the eligible articles, 14 (21%) did not report initial sample size by gender while 47 (69%) failed to report final sample size by gender (Figure 3 ). Figure 3 Proportion of articles that did not report initial and final sample size by group or gender in randomized trials. Values reported are means. Discussion The present study evaluated the quality of reporting of RTs and RCTs in the diet and exercise literature for weight loss, focusing exclusively on subject characteristics, including age, gender, race, health status, medication use, postmenopausal status, and attrition. Transparent reporting of subject data is important in the scientific literature so that readers can efficiently evaluate outcomes in RTs and RCTs. Inadequate reporting creates numerous difficulties with interpretation and can lead to biased conclusions. For example, the effects of diet and/or exercise can vary based upon an individual's age. Schwartz et al. 1991[ 11 ] compared the effects of 6 months of endurance exercise (4 days/wk, 45 minutes/day, 85% of heart rate reserve) in older (Mean = 67.5 yrs, SD = 5.8 yrs) versus younger (Mean = 28.2 yrs, SD = 2.4 yrs) on body weight and composition. They found that older men had a 2.5 kg reduction in body weight, a 2.3% decrease in percent body fat, and a 2.4 kg decline in fat mass compared to no changes in the younger individuals. The findings suggest that subjects' age is an important determinant in the response to exercise. However, despite the difference in response to exercise based on the age of subjects, we found that 11% of authors failed to report their subjects' ages within their published report. Reporting gender of the study participants is also important as men and women respond differently to diet and/or exercise treatments. For example, Donnelly et al. [ 12 ] reported that following 16 months of verified, supervised aerobic exercise at 45 minutes per day, 5 days per week resulted in a decrease in body weight of 6% for the men compared to no change in the women. It also has been shown that individuals from different races and/or ethnic background may respond differently to diet and/or exercise. Jakicic et al. [ 13 ] reported that resting energy expenditure was lower in African-American women (Mean = 7279 kJ/d, SD = 825 kJ/d) compared to Caucasian women (Mean = 7807 kJ/d, SD = 854 kJ/d) even after correcting for body weight and lean body mass. The authors concluded that such differences might partially explain the smaller weight losses typically seen in African-American women when compared to Caucasian women enrolled in a weight loss program. In a recent meta-analysis on walking and resting blood pressure in adults, the investigators reported that only 13% of their included studies reported information on race [ 5 ]. In the current investigation 86% of authors failed to report the race/ethnicity of the subjects, which highlights the need for caution when extrapolating findings from one population to another. In the present study, co-morbid health conditions of study subjects were only marginally reported (66%) in the literature. It is well known that certain disease states, such as diabetes and cancer, affect metabolism [ 14 - 17 ], which can influence the effect of diet and exercise on body composition. Closely related to the need for reporting health conditions of study participants is the issue of medication use. Davis et al. [ 18 ] determined the effects of taking an antihypertensive medication (i.e., atenolol or chlorthalidone) compared to a placebo group combined with weight loss. They found that at 6 months those taking chlorthalidone lost (6.9 kg) the most weight compared to either the placebo (4.4 kg) or atenolol (3.0 kg) groups. They speculated that the group taking chlorthalidone might have reduced appetite or increased fat mobilization due to the volume depletion and increases in serum and urinary catecholamines. However, despite the profound effects that medication may have on body weight and composition 92% of authors failed to report medication use in our study. Without an adequate description of subject characteristics, readers are unable to judge the comparability of study groups. Menopause is associated with decreases in lean mass and increases in fat-mass [ 19 - 22 ]. Therefore, post-menopausal status and the associated decline in estrogens, androgens, and other hormones may significantly influence the effects of dietary/exercise interventions on body composition. It appears from the present study the majority of investigators (92%) of dietary/exercise studies report post-menopausal status of their subjects. Attrition may threaten the internal and external validity of the scientific literature as well as the efficacy of a specific dietary/exercise intervention [ 23 , 24 ]. For example, failure to report the number of dropouts and completers prevents the reader from calculating attrition rates for different experimental conditions, which can result in an overestimate of treatment effectiveness. In the present study, we found that end of study sample size was not reported in the majority (58%) of studies. Without knowledge of the number of subjects who were lost to follow-up, readers are unable to judge the effectiveness of a clinical treatment or ascertain whether or not a research finding has practical significance. Conclusion In the present critical appraisal of the methodological quality of weight loss studies, we found major shortcomings in the reporting of subject characteristics as they relate to enrollment, allocation, and follow-up in trials that evaluated diet and exercise interventions. Many studies did not report variables that may explain some of the variance in outcomes. These findings are consistent with those of similar studies, which indicate inadequate reporting of subject characteristics [ 5 , 25 ] and reveal that poor adherence to published standards of reporting is common [ 26 ]. Clearly, additional attention should be paid to ensure compliance with reporting standards for diet and exercise intervention studies. List of abbreviations Consolidation of the Standards of Reporting Trials (CONSORT), randomized controlled trials (RCTs), randomized trials (RTs), kilojoules per day (kJ/d), standard deviation (SD), kilograms (kg), years (yrs) Competing interests The author(s) declare that they have no competing interests. Authors' contributions CAG made substantial contributions to the conception and design, acquisition of data, analysis and interpretation of data, and drafting and revising the manuscript for important intellectual content. EPK, JDL, BWB and GH made substantial contributions to the submitted manuscript by assisting with the acquisition of data, analysis and interpretation of data, drafting of the manuscript, statistical analyses, and critical revisions. JED made substantial contributions to the manuscript by helping with the conception and design, critical revisions of the manuscript, and obtaining funding. All authors read and approved the final manuscript. Funding source Supported by National Institutes of Health NIH DK56303, Joseph E. Donnelly, Ed.D., principal investigator Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554784.xml
544568	Small area mapping of prostate cancer incidence in New York State (USA) using fully Bayesian hierarchical modelling	Background As part of a long-term initiative to improve cancer surveillance in New York State, small area maps of relative risk, expressed as standardized incidence ratios (SIRs), were produced for the most common cancers. This includes prostate cancer, the focus of this paper, since it is the most common non-dermatologic malignancy diagnosed among men and the second leading cause of cancer deaths for men in the United States. ZIP codes were chosen as mapping units for several reasons, including the need to balance between protecting personal privacy and public demand for fine geographic resolution. Since the population size varies greatly among such small mapping units, hierarchical Bayes spatial modelling was applied in this paper to produce a map of smoothed SIRs. It is further demonstrated how other characteristics of the large sample from the stationary posterior distribution of SIRs can be mapped to investigate various aspects of the statewide spatial pattern of prostate cancer incidence. Results Thematic mapping of the median and 95 percentile range of SIRs provided, respectively, a map of spatially smoothed values and the uncertainty associated with these smoothed values. Maps were also produced to identify ZIP codes expressing a 95% probability, in the Bayesian paradigm, of being less than or greater than the null value of 1. Conclusion The model behaved as expected since areas that were statistically elevated coincided with areas identified by the spatial scan statistic, plus the relative uncertainty increased as a ZIP code's population decreased, with an exaggerated effect for low population ZIP codes on the edge of the state border. The overall smoothed pattern, along with identified high and low areas, may reflect difference across the state with respect to socio-demographics and risk factors; however, this is confounded by potential differences in screening and diagnostic follow-up. Nevertheless, the Bayes modelling approach is shown to provide not only smoothed results, but also considerable other information from a large empirical distribution of outcomes associated with each mapping unit.	Background Geographic surveillance of chronic disease is central to understanding spatial or spatial-temporal patterns that may help to identify discrepancies in disease burden among different regions or communities. As part of ongoing efforts in New York State to understand spatial patterns of cancer and to help implement cancer prevention and control programs, small area maps of cancer relative risk, expressed as standardized incidence ratios (SIRs), have been produced and shared with the public [ 1 ] for the most common anatomical cancer sites. Prostate cancer, the focus of this paper, was included because it is the most common non-dermatologic malignancy diagnosed among men and the second leading cause of cancer deaths for men in the United States (US) [ 2 ]. Although mortality from this disease in the US has statistically significantly decreased at a rate of 2.6% per year from 1990 to 2000 [ 3 ], unexplained geographic discrepancies in mortality rates do exist [ 4 ]. Furthermore, several treatment options appear to be associated with excellent long-term disease-specific survival for otherwise healthy men with localized disease [ 5 ]. Results for prostate cancer (all stages combined) are reproduced in Figure 1 , where ZIP code-level standardized incidence ratios (SIRs) are presented along with results from analyzing these data with the spatial scan statistic [ 6 ]. The circles in Figure 1 represent statistically elevated regions based on Poisson likelihood ratios comparing rates inside the circle to those outside the circle. Details of how the scan statistic results were reduced to the circles presented in Figure 1 are found in Boscoe et al [ 7 ]. Figure 1 All stage prostate cancer Incidence by ZIP Code in New York State, 1994–1998 . ZIP code-level ratios of observed incidence to age- and race-adjusted expected incidence, along with significant spatial scan statistic circles that are non-overlapping within specified ranges of standardized incidence ratios, based on reference [1]. Select cities and regions overlaid for reference. It is well recognized that the stability of population-based statistics like the SIRs in Figure 1 can vary greatly among small geographic areas due to varying population size. Different methods of smoothing have been developed to address this issue, where all are based on the phenomenon that observations close together in space are more likely to share similar properties than those that are far apart [ 8 ]. While this positive spatial autocorrelation may be problematic for statistical methods that require independent observations, it can also be embraced to help smooth noisy maps by borrowing strength from neighbors for those mapping units with small populations. Of the different approaches to spatial smoothing, only a few appear to have gained acceptance in spatial epidemiology. Non-parametric approaches include spatial filtering [ 9 , 10 ] and the head-banging algorithm [ 11 ], both of which are basically variations on a moving window kernel-type smoother. The parametric approach of generalized linear modelling [ 12 ] treats the observed response, y , as a random variable that has arisen from a probability distribution with expectation θ . This expectation is modeled, via an appropriate link g (·), as a linear function g ( θ ) = α + x' β + ε , for a common value α , explanatory covariates x' β and a random effect ε that captures unexplained variation. If the random effect is associated with exchangeable spatial heterogeneity, estimates are smoothed towards a global mean, whereas if the random effect is associated with local spatial autocorrelation, estimates are smoothed towards a local neighborhood mean, which is typically more meaningful in geographic epidemiology. There are different approaches to modelling local spatial dependence, and section 6.3 of Cressie [ 13 ] presents several arguments in favor of the conditional autoregressive (CAR) model originally conceived by Besag [ 14 ]. Estimation of model parameters can proceed by maximum likelihood [ 13 ]; however, the hierarchical nature of generalized linear models lends itself well to Bayesian analysis whereby linear terms in the model are assigned prior distributions that, in turn, have "hyperprior" parameters. Earlier applications employed empirical Bayes methods [ 15 ], where hyperparameters are estimated directly from the data. This approach is limited because it assigns a point estimate to the hyperparameter without allowing for variability that may be associated with it, and this variability can be large [ 16 , 17 ]. Fully Bayesian modelling assigns hyperprior distributions to these hyperparameters, so that every parameter of the hierarchical model is allowed to vary over a prior distribution and no single point estimate is used to represent an unknown parameter value. Furthermore, the fully Bayesian approach allows the convolution model that incorporates both a heterogeneous and spatially structured random effect [ 18 ], thus allowing the most flexibility in model development. Several reviews consistently support the fully Bayesian approach over empirical Bayes modelling [ 16 , 17 , 19 ]. Since there are no closed form analytical solutions for parameter estimates of a fully Bayesian model, nor likelihood profiles to maximize, Markov Chain Monte Carlo (MCMC) methods are used to generate large samples from the posterior distributions of all stochastic nodes of the hierarchical model, given the likelihood describing the original data distribution and all appropriate prior and hyperprior distributions of the likelihood parameters [ 16 ]. Estimation and inference in the fully Bayesian paradigm are based upon these large sample approximations of the posterior distributions. In what follows, the fully Bayesian approach is applied to simulating large samples from the posterior distribution of prostate cancer relative risk in each of 1412 ZIP codes in New York State. Various aspects of these distributions are then mapped to reveal information on the geographic patterns of prostate cancer. Results The model defined by equations 1–5 was applied to simulate a sample of 1000 independent observations from the stationary posterior distribution of standardized incidence ratios for each ZIP code. Summary statistics and graphical analysis of these empirical distributions indicated that they arose from generally symmetric posterior distributions. Since the sample mean, median and mode were very similar for each ZIP code, the median was chosen to represent central tendency and is mapped in Figure 2 . This "smoothed" map of SIRs provides a picture of spatial pattern inherent in the raw data mapped in Figure 1 . Uncertainty associated with these estimates of relative risk is mapped in Figure 3 as the 95 percentile range (97.5 th – 2.5 th percentile) of the 1000 values sampled from the posterior distribution of SIRs for each ZIP code. Figure 2 Bayesian smoothed prostate cancer incidence . Median of the posterior distribution of ZIP code-level standardized incidence ratios. Thematic categories based on natural breaks method, with slight adjustment. Figure 3 Uncertainty of Bayesian Smoothed prostate cancer incidence . ZIP code-level 95 th percentile range of posterior distribution of standardized incidence ratios. Thematic categories based on Natural Breaks Method. The posterior distributions can also be used to identify ZIP codes where a specified mass of the distribution of relative risk is greater or less than the null value. For example, Figure 4 shows ZIP codes where 95% of the simulated SIRs exceed the value of one. In the Bayesian paradigm, those ZIP codes highlighted in Figure 4 have a 95% probability of higher than expected risk. Likewise, Figure 5 highlights ZIP codes expressing a 95% probability of lower than expected risk. Figure 4 ZIP codes with a 95% probability of relative risk exceeding 1 . The lower 5 th percentile of the posterior distribution of standardized incidence ratios exceeds or equals 1. Figure 5 ZIP codes with a 95% probability of relative risk being less than 1 . The upper 95th percentile of the posterior distribution of standardized incidence ratios is less than or equal to 1. Discussion Methodology The Poisson model applied in this paper is a particular application of hierarchical Bayes spatial generalized linear models for the exponential family of likelihoods [ 20 ]. Particular models are specified by the likelihood that is assumed to give rise to the observations, the structure of the prior, and the hyperprior distributions of variance components, which are typically vague to allow learning from the data. An aspect of these models that will influence outcomes is the neighborhood weights, as in Equation (4); however, defining these weights remains an open area of research. Most applications to date use the first order binary weighting scheme where w ij = 1 if a mapping unit j shares a common border with unit i , and w ij = 0 otherwise. This weighting scheme actually has its roots in image analysis, for which this type of modelling was developed [ 14 ], and it makes sense when the spatial units are equal size and shape pixels and the response variable has a constant variance for each pixel. However, this is not the case when smoothing disease maps where the mapping units are of irregular size and shape, and stability of the response variable estimates varies with changing population size. This problem is especially relevant for mapping units like ZIP codes. A proper approach may be to define weights as a decay function of geographic distance between population-weighted centroids of the mapping units. This function may be obtained by fitting a model correlogram to residuals that are obtained from a model that does not include a random effect to account for spatial autocorrelation. Cressie and Chan [ 21 ] used the empirical variogram of the response variable to determine the range of spatial autocorrelation. For neighborhood distances within this range, weights were defined as a function of Euclidean distance. Meanwhile, Griffith [ 22 ] provides some "rules of thumb" for defining geographic weights, but they are very general. Ferrandiz, et al [ 23 ] used a weight of n i n j / d ij for neighboring mapping units separated by geographic distance d ij and of population sizes n i and n j . These authors applied such a weight to prostate cancer mortality mapping; however, this gravity-type weighting may be better suited for infectious disease, not chronic disease. If a decay function is fit from the data, the varying stability of disease rates among the mapping units presents a challenge. Since the Bayesian model is designed to adjust for varying stability, perhaps a hierarchical model like the one applied in this paper can be extended so that the weights in Equation (4) are defined as a decay function whose unknown parameters can be assigned "hyperprior" distributions. Application to prostate cancer mapping Maps like in Figure 1 present a compromise between the need to protect personal privacy and public demand for fine geographic resolution. Such small mapping units are necessary for discerning among communities that can vary drastically across a region with respect to possible risk factors and both population density and demographics. However, this comes with the cost of unstable risk estimates for many mapping units that have small populations. Smoothing is therefore applied to help visualize spatial pattern that is inherent in the data of Figure 1 . It is demonstrated how hierarchical Bayes spatial modelling has the appealing feature of providing a whole distribution of possible outcomes that can be used for not only smoothing, but also to explore other aspects of spatial pattern. Viewing Figures 1 through 3 indicate that the Bayesian model is behaving as expected since the smoothed estimates are increasingly dependent on the prior model as uncertainty increases due to decreasing population, whereas for ZIP codes with large populations, like in New York City, the smoothed estimates are similar to the raw SIRs. We also see that many edge ZIP codes in less populated areas tend to have greater uncertainty relative to their non-edge neighbors because there are fewer neighbors to borrow strength from. For the heterogeneous Poisson model applied in this paper, Lawson et al [ 24 ] suggest treating edge mapping units as a guard and not as part of the actual study area. However, presenting the width of Bayesian posterior distributions provides a way to retain the edge units while also showing the relative uncertainty associated with their smoothed values. The smoothed pattern in Figure 2 is highlighted for areas of high and low incidence in Figures 4 and 5 , respectively. Geographic patterns seen in these maps are potentially influenced by many factors, including differences between regions of the state in terms of racial, ethnic and socio-demographic composition. Yet it is well recognized that interpreting any possible relationships with risk factors is confounded by differences in screening and diagnostic practices across the state. Prostate-specific antigen (PSA) testing remains high among US males over 40 [ 25 ] and there is evidence of a steady increase in testing rates in New York State during the years corresponding to the data analyzed in this paper [ 26 ]. Along these lines, we note that a relatively large proportion of the four most populated counties of New York City reveal a high probability of less than expected incidence (see Figure 5 inset). This may be partially explained by the large immigrant population in these four counties, as indicated by much higher proportions of people who are foreign-born and/or do not speak English at home [ 27 ], which may translate to lower screening rates. Although the patterns seen in Figures 2 , 4 and 5 may be partially explained by geographic variations in PSA testing and diagnostic follow up, such variation is not actually known, therefore we cannot adjust for this potential confounder. In the neighboring state of Connecticut, geographic variation of invasive prostate cancer incidence was large and revealed some consistency before and after the introduction of PSA testing, while the pattern was completely different and variation was much smaller during the years of PSA introduction [ 28 ]. These authors suggest that such a space-time pattern reflects the impact of introducing PSA testing, although this cannot be confirmed in the absence of data on geographic differences in PSA use. Some areas appear elevated that are in popular vacation spots, such as the eastern forks of Long Island and the "north country" of New York State including the Thousand Islands area along the Saint Lawrence River (near Watertown in Figure 1 ) and the Adirondack region. This may possibly be due to a seasonal residence effect whereby vacation areas tend to have artificially inflated chronic disease rates [ 29 ]. This occurs when seasonal residents provide a health care provider with a local address of a vacation home, while their primary residence is where they are counted by the decennial US census. Consequently, if their residence at time of diagnoses is in the vacation area, this record inflates the SIR numerator for that area, while they are counted in the denominator for the area of their primary residence as captured by the census. This effect is enhanced since the population spending extended periods in vacation areas tends to be over age 55, which is the age cohort at highest risk of chronic disease. Boscoe and Mclaughlin [ 29 ] have presented evidence of increased overall cancer rates in areas with seasonally resident populations in New York State, especially the Thousand Islands area. This uncertainty is reflected in Figure 3 where the width of the posterior distribution of SIRs for these areas is relatively large. While the smoothed results in Figure 2 present an advantage over mapping raw data, a limitation of smoothing is that the pattern we decipher is subject to confounding by spatially varying population sizes [ 30 ]. In other words, smoothed maps like in Figure 2 reveal high and low relative risk in areas with larger populations, while areas with small populations tend to be smoothed towards the null value. While this means that areas with small populations that actually have abnormally high or low disease rates may be obscured, it is still well recognized that many extreme values associated with small populations may simply reflect random noise. Other methods like the spatial scan statistic can be used in conjunction with Bayes smoothing to strengthen overall spatial analysis. Statistically significant scan statistic circles like those in Figure 1 can vary in size, potentially encompassing many ZIP codes, so are not restricted to only pre-defined neighborhoods like the conditional autoregressive model used by the Bayes smoothing algorithm. In this regard, the spatial scan statistic is similar to smoothing by spatial filtering with variable-radius circles [ 10 ]. General regions of statistically elevated relative risk may be identified by the scan statistic, with supporting evidence from Bayesian posterior distributions to help identify the mapping units that contribute strongly to a scan statistic circle. Indeed, each spatial scan statistic circle reported in Figure 1 contains at least one elevated ZIP code identified in Figure 4 by the Bayesian model. There is ample flexibility for exploratory analysis by varying the display parameters for results from these two methods. For example, the results in Figure 4 can be either more generalized or further focused by identifying ZIP codes where, say, 90% or 99%, respectively, of the posterior probability mass exceeds the null value of one. At the same time, we can display the set of non-overlapping scan statistic circles that correspond to lower levels of relative risk than are shown in Figure 1 , thus capturing more geographic area. In fact, when this is done for scan statistic circles corresponding to 15–49% relative risk (not shown), all of the elevated ZIP codes in Figure 4 are spatially associated with significant scan statistic circles. There is extensive literature on hierarchical Bayes spatial modelling for disease mapping; however, most papers are theoretical in nature and use illustrative examples, often with the same data sets. One exception was recently published by Short et al [ 31 ], who applied Bayes modelling to produce maps of cancer control variables. Specifically, they smoothed maps of different outcomes (mortality, incidence, staging and screening) for each of breast, colorectal and lung cancer in Minnesota (USA) counties. Cancer control maps were created for each cancer site by obtaining a weighted sum of each smoothed outcome, and an overall cancer control map was obtained by a weighted sum of the individual cancer control map values. These results can help guide resource allocation for state cancer prevention and control efforts. While there are open areas for improvement in the methodology of hierarchical Bayes spatial modelling, it is a valuable tool for geo-spatial assessment of disease patterns that can help identify differences among communities. This may in turn indicate patterns of health care access, screening and diagnostic follow up and possibly indicate etiologic clues about causal relationships. Methods Data Observed and expected values of prostate cancer incidence used to calculate SIRs for 1412 New York State ZIP codes [ 1 ] were obtained for the years 1994 to 1998 from the New York State Cancer Registry (NYSCR). The expected values used in the SIR denominator are based on indirect standardization using the age-by-race distribution in each ZIP code and the statewide age- and race-specific incidence rates as a reference. Age and race distributions correspond to the year 1997, as estimated by the Claritas Corporation™ based on prior census values. ZIP code boundaries were delineated by the GDT Corporation™ in 1999. ZIP Code delivery areas are prone to change over time [ 32 ], particularly in rapidly growing parts of the country. According to the NYSCR [personal communication], a review of all of the issues of the Postal Bulletin, where these changes are documented, from 1990 to the present revealed that New York has had stable ZIP Code delivery areas. Approximately 50 small, rural post offices were closed, 3 new post offices were added, and none were realigned. ZIP codes were combined in instances where service delivery area changed between 1990 and 1999 or for confidentiality reasons where necessary [ 33 ]. Modelling Letting the geographic domain (New York State) be subdivided into i = 1, ..., n distinct mapping units ( n = 1412 ZIP codes for our application), the number of cases within each unit, Y i , conditional on location i , is defined as a Poisson random variable with expectation E i θ i , where E i equals the age- and race-adjusted expected number of prostate cancer cases, and θ i equals the area-specific relative risk. Given an observed response y i , note that the maximum likelihood estimate of relative risk is = y i / E i , the standardized incidence ratio (SIR). The relative risk parameter θ i is assigned a log-normal prior distribution, log( θ i ) ~ N( μ i , ), where the expectation and variance are defined by a linear function of a common value (intercept), α , and two independent random effects, a heterogeneous component, u i , that does not depend on geographic location (exchangeable) and an autocorrelated component, v i , that reflects local spatial structure by incorporating the influence of neighboring geographic units. Prior distributions are then assigned to these linear terms and consequent hyperprior distributions are assigned to the variance terms, thus creating a 4-level hierarchical model as follows. Level 1, define the likelihood: Y i ~ Poisson( E i θ i ) (1) Level 2, link to a linear function: log( θ i ) = α + u i + v i (2) Level 3, assign prior distributions: α ~ N(0,0.0001), noting that 0.0001 is the precision, thus defining a vague prior, for a neighborhood of geographic units δ i with respect to unit i and w ij is a weight defining the relationship between geographic unit i and its neighbor j . The weight is defined simply as w ij = 1 if ZIP codes i and j are adjacent (share a common border) and w ij = 0 otherwise. Level 4, assign hyperprior distributions to precision terms: τ u = 1 / ~ Gamma(a,b) and τ v = 1 / ~ Gamma(c,d) (5) for shape parameters a and c, and inverse scale parameters b and d. This is the convolution model originally proposed by Besag, York and Mollie [ 18 ], where the random effect associated with spatial autocorrelation, v i , is defined according to the conditional auto-regressive model (CAR) [ 14 ]. Note that the distribution of v i is conditional on geographic location, whereby its expectation equals a local neighborhood average. The Bayesian model puts increasing emphasis on this term as the underlying population at location i decreases. Although covariates can be incorporated into the log-linear expression at the second level of the model, our interest is with estimating and mapping the relative risk, θ i = exp( α + u i + v i ). Choosing Gamma Hyperpriors While it is established that a vague prior is acceptable for the linear term α in Equation (2) (i.e. Ghosh et al [ 20 ]), the model should be evaluated for sensitivity to choice of the Gamma hyperprior distributions of the precision terms, as in Equation (5). Two very different hyperprior specifications that appear in the literature for this convolution model were experimented with. Hyperparameters were specified for one model as Γ(1,1), which yields a probability of 99% that the precision lays between 0.01 and 4.6, and for the other model as Γ(0.5,0.0005), which yields a probability of 99% that the precision lies between 0.16 and 6635, with most of the probability concentrated towards 0. Note that these parameter choices also satisfy sufficient conditions for ensuring a proper joint posterior distribution of all the stochastic nodes [ 20 ]. For the statewide collection of New York ZIP code log(SIR)s to be smoothed in this paper, the sample precision equals 6.25 (variance = 0.16) and the precision of first order neighborhood means equals 20 (variance = 0.05). Therefore, it may be desired to retain the model with hyperprior specification of Γ(0.5,0.0005) to at least capture the sample-based precision estimates, while also defining a vague hyperprior that allows more learning from the data. Final smoothed results from each model are compared in Figure 6 where we see, in agreement with Bernardinelli et al [ 34 ], that it essentially makes no difference which hyperprior is used. Therefore, the Γ(0.5,0.0005) hyperprior specification was chosen. It is not the intention of this paper to perform a rigorous sensitivity analysis with respect to hyperprior specification; however, the two models assessed in Figure 6 represent very different distributions and therefore indicate that the fully Bayes hierarchical model is quite robust with respect to hyperprior specification when smoothed relative risks are the objective. Figure 6 Comparison of smoothed standardized incidence ratios for two different specifications of the hyperprior distributions . Each point represents a ZIP code with two medians of the posterior distribution of standardized incidence ratios obtained from different models. Running the Gibbs Sampler WINBUGS 1.4 [ 35 ] was used for running three independent Markov Chains. Initial values of all stochastic nodes of the model were chosen to provide dispersed initial values without being excessively overdispersed. For the common intercept, α , and heterogeneous random effect, u i , zero (0) was used to initiate one chain, plus/minus four standard deviations of the statewide log(SIR) were used to initiate the other respective chains. Zero is the statewide average log(SIR) that provides a point estimate for α , plus it is the expected value of u i . For the random effect associated with local spatial clustering, v i , the initial values were based on the average plus/minus four standard deviations of the log of first order neighborhood average SIRs. For the precision terms τ u and τ v , the inverse of the sample variances of the log(SIR) and the log of the first order neighborhood average SIRs were used respectively for one chain, plus lower and higher values were chosen for the other two chains in order to be well dispersed from the middle value, but not wildly so with respect to what is reasonably expected based on the observed spatial variability. Convergence of relative risk for the three independent chains was confirmed by graphing their traces and observing random mixing of all chains, which revealed white noise variation around a common value, with no trend. This was supported by observing Brooks-Gelman-Rubin diagnostics that clearly satisfied convergence criteria [ 36 ]. After a burn-in of 10,000 iterations, which was far more than actually necessary, the following 1000 iterations were sampled from each of the three chains by choosing every third iteration to help avoid possible autocorrelation within a chain. This large sample approximation of the stationary posterior distribution for each ZIP code relative risk was then summarized in WINBUGS and brought into a Geographic Information System [ 37 ] for mapping. Model Selection Variations of the model defined above were compared by evaluating the mean deviance of 1000 iterations chosen from the three independent Markov Chains after burn-in. This was done by obtaining the mean of -2(log likelihood) for each iteration, as provided by the deviance node in WINBUGS. The mean deviance was then calculated as D ( y , μ ) = 2 [ l ( y , y ) - l ( y , μ )], where l ( y , y ) is the mean maximum achievable log likelihood, obtained for a saturated model where a parameter is assigned to each datum, and l ( y , μ ) is the mean log likelihood obtained for the model in question. This takes the conventional assessment of deviance for generalized linear models [ 12 ] and applies it to the many outcomes of Monte Carlo simulation, as per Spiegelhalter et al [ 38 ]. Incorporating a random effect associated with local spatial structure (CAR term) provides much stronger prior information than the exchangeable random effect alone (Table 1 ), which assumes purely heterogeneous variation across the state. This agrees with findings by Spiegelhalter et al [ 38 ], who developed the Deviance Information Criterion as a penalized version of deviance and applied it to Scottish lip cancer data. The convolution model was therefore chosen, which incorporates both random effects. Table 1 Deviance analysis. See text for explanation. Model Mean (-2 LL) Mean Deviance Saturated 8082.0 Convolution 8194.0 112.0 CAR only 8191.0 109.0 Exchangeable only 10800.0 2718.0	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544568.xml
544597	Murine leukemia virus (MLV) replication monitored with fluorescent proteins	Background Cancer gene therapy will benefit from vectors that are able to replicate in tumor tissue and cause a bystander effect. Replication-competent murine leukemia virus (MLV) has been described to have potential as cancer therapeutics, however, MLV infection does not cause a cytopathic effect in the infected cell and viral replication can only be studied by immunostaining or measurement of reverse transcriptase activity. Results We inserted the coding sequences for green fluorescent protein (GFP) into the proline-rich region (PRR) of the ecotropic envelope protein (Env) and were able to fluorescently label MLV. This allowed us to directly monitor viral replication and attachment to target cells by flow cytometry. We used this method to study viral replication of recombinant MLVs and split viral genomes, which were generated by replacement of the MLV env gene with the red fluorescent protein (RFP) and separately cloning GFP-Env into a retroviral vector. Co-transfection of both plasmids into target cells resulted in the generation of semi-replicative vectors, and the two color labeling allowed to determine the distribution of the individual genomes in the target cells and was indicative for the occurrence of recombination events. Conclusions Fluorescently labeled MLVs are excellent tools for the study of factors that influence viral replication and can be used to optimize MLV-based replication-competent viruses or vectors for gene therapy.	Background Efficient and long-lasting gene delivery is the major challenge in the development of vectors for gene therapy. Replication-competent retroviruses (RCRs) encoding suicide genes linked via an internal ribosome entry site (IRES) offer a significant advantage over replication-deficient vectors in cancer gene therapy, since they are able to spread efficiently in vivo [ 1 - 4 ]. Uncontrolled virus spread is, however, associated with serious risk of adverse events due to viral-integration mutagenesis. Therefore, for a therapeutic application, RCRs have to be equipped with additional safety features, e.g. transcription controllable by exogenous agents or viral entry restricted to the diseased cells. The selective delivery of a therapeutic gene by targeting retroviral entry would immensely reduce unfavorable side effects and ease the clinical application of gene therapy. The ecotropic MLV envelope protein does not recognizes receptors on human cells. An obvious challenge has been to extend the host range of vectors carrying the ecotropic envelope glycoprotein to a predetermined human cell type. This change in host range requires the inclusion of a novel attachment site and the induction of fusion via a novel receptor interaction. It has been shown before that it is possible to modify ecotropic Env and change its binding specificity, however, the efficient triggering of the membrane fusion or the escape from endosomes of viral particles targeted to e.g. epidermal growth factor (EGF)-receptor is still missing [ 5 , 6 ]. The further development of such targeted vectors requires the understanding of the mechanisms that are involved in adsorption and internalization of retroviruses. Investigating murine leukemia virus (MLV) replication is technically inconvenient because MLV infection does not cause a cytopathic effect in the infected cell. Viral replication can only be studied by immunostaining, measurement of reverse transcriptase activity or syncytia formation. We have developed a tool to simplify these analyses. We generated an MLV tagged with a fluorescent envelope protein, which allows viral replication and Env attachment to target cells to be followed by flow cytometry. This method will be useful for optimizing RCRs or retroviral vectors for gene therapy. Results Construction of GFP-tagged MLVs and their replication We previously constructed a modified ecotropic murine leukemia virus (Mo-MLV) bearing the green fluorescent protein (GFP) from Aequoria victoria in its envelope. A replication competent ecotropic MLV variant was generated (GFP-EMO) that had the 53 aas of the epidermal growth factor (EGF) fused to the N-terminus of Env and the GFP sequences inserted into the proline-rich region (PRR) [ 7 ]. We deleted the EGF sequences by replacing a Pfl MI fragment of GFP-EMO with wt sequences. This resulted in a replication-competent virus expressing the chimeric GFP-Env protein (GFP-MOV) (Fig. 1A ). NIH3T3 cells were transfected with 10 μg plasmid DNA encoding GFP-MOV or GFP-EMO using the calcium-phosphate procedure and were cultured for 13 days. Viral replication was monitored as GFP-positive cells by flow cytometry. As indicated in Figure 1B , both viruses replicate with similar kinetics. Untransfected NIH3T3 cells did not show green fluorescence. Figure 1 Generation and replication of the GFP-Env-tagged viruses. (A) Schematic representation of the GFP-Env-tagged viruses. EGF, epidermal growth factor; PRR, proline rich region; GFP, green fluorescent protein; L, signal peptide.(B) Viral replication kinetic in transfected NIH3T3 cells monitored by the percentage of GFP-positive cells.(C) PCR analysis of genomic DNA from FLY-Jet cells transfected with GFP-EMO. The N-terminal sequences of the EGF-Env gene were analyzed by PCR using the primers MLV-5'-Env and BS-5. GFP-EMO plasmid DNA was used as a positive control and gave rise to a 900 bp fragment. Predominantly faster migrating fragments were amplified from genomic DNA (gDNA) of GFP-EMO transfected FLY-Jet cells 32 days after transfection. Sequestering of EGF-Env-containing viral particles has been described before [ 8 , 9 ]. Viral particles containing EGF-Env were rapidly trafficked to endosomes and became degraded. This effect was dominant over the normal entry pathway, because mouse cells expressing the ecotropic receptor and the EGF-receptor showed a severely decreased infectivity of EGF-Env containing vectors [ 8 ]. We were interested, if replication competent GFP-EMO might be useful to select viral variants able to escape the degradation in the endosomes. Transfection of GFP-EMO into cells expressing only the EGF-receptor (A431, COS-7) did not result in viral replication (data not shown). Therefore, GFP-EMO and GFP-MOV were transfected into FLY-Jet cells [ 10 ], which express the human EGF-receptor and the receptor for ecotropic MLV. Viral replication of GFP-EMO could be observed in FLY-Jet cells, although strongly delayed, after 10 days only 7.4 % of the cells were GFP-positive. After 38 days, all cells were GFP-positive and the N-terminus of the Env gene was analyzed by PCR amplification of genomic DNA isolated from infected cells. Predominantly a band migrating faster than the GFP-EMO fragment was amplified (Figure 1C ), which was verified by sequence analysis to contain wt Env sequences. The less abundant, slower migrating fragments still contained the EGF sequences in Env. This confirms the sequestering of EGF-Env containing retroviral particles via the EGF-receptor. The selection of viruses able to escape the endosomal degradation was not possible and shows that degradation of viral particles in the endosomes favors the selection of wt Env-containing MLV, which escapes the sequestering by EGF-receptor. Cell binding of GFP-tagged MLV Viral entry is initiated by the binding of the envelope protein (Env) to the retrovirus receptor at the target cell surface. To test whether labeling of Env with GFP allows viral attachment to be monitored, we incubated supernatants of NIH3T3 cells producing GFP-EMO or GFP-MOV with cells that either express mCAT, the receptor for ecotropic MLV [ 11 ] (NIH3T3), do not express it (293T, A431) or do express the human EGF receptor (A431). As illustrated in Figure 2A , NIH3T3 cells incubated with cell culture supernatants showed a shift to green fluorescence, indicating specific binding of GFP-tagged Env to mCAT. The shift to green fluorescence could not be increased by larger amounts of viral supernatants or longer incubation times (data not shown), which shows that already after 5 min. all receptors are occupied by Env. For GFP-MOV supernatants a shift in fluorescence was only observed with mCAT-expressing cells, while GFP-EMO supernatants also produced a shift with A431 cells. This indicates additional specific binding to the EGF receptor. The shift was more pronounced on A431 cells than COS-7 cells, correlating with the amount of EGF receptor expressed by the target cells (data not shown). Figure 2 Binding of GFP-Env to cells. (A) Supernatants of GFP-EMO- or GFP-MOV-infected NIH3T3 cells were incubated with the indicated target cells and analyzed by flow cytometry. Binding of GFP-Env was detected by a shift to green fluorescence (FL-1).(B) Supernatants from GFP-MOV-infected NIH3T3 cells were incubated with the indicated target cells and analyzed by flow cytometry. Soluble receptor binding domains of the ecotropic or the amphotropic MLV Env (E-sRBD, A-sRBD) were added prior to the virus, as supernatants from 293T cells transfected with the expression constructs. After 5 mins., supernatants of GFP-MOV-infected NIH3T3 cells were added for an additional 5 mins. Binding of GFP-Env was detected by a shift to green fluorescence (FL-1). NIH3T3 i -MLV: chronically MLV-infected NIH3T3 cells. The specificity of cell staining by supernatants containing GFP-MOV was further examined using chronically Mo-MLV-infected NIH3T3 cells (NIH3T3 i -MLV). These cells have only negligible numbers of mCAT molecules on the cell surface, because Env expression leads to their retention within the cell (receptor interference). As expected, NIH3T3 i -MLV cells produced no shift when incubated with GFP-MOV supernatants (Fig. 2B ). Furthermore, binding of GFP-MOV supernatants could be inhibited by preincubation of NIH3T3 target cells with a soluble Env fragment containing the receptor binding domain (sRBD) derived from the ecotropic Env [ 12 ], but not with the equivalent sRBD derived from the amphotropic Env [ 12 ], which binds to a different receptor (Fig. 2B ). This shows that GFP-tagging can be used to investigate Env-binding properties by flow cytometry. Replication of semi-replicative retroviral vectors The size of a retroviral genome is limited to roughly 11 kb. The capacity for the insertion of a therapeutic gene for gene therapy is, however, increased by the use of semi-replicative retroviral vectors (SRRVs), where the gag / pol and env genes are split between two viral genomes. We constructed split viral genomes and used fluorescent proteins to monitor the replication of the resulting SRRVs. A packagable MLV Gag/Pol expression vector, GAG/POL-RFP, was generated by deleting of the env gene and replacing it with the red fluorescent protein (RFP) (Fig. 3 ). RFP is encoded by the spliced mRNA and its expression can be monitored by red fluorescence (Fig. 4C ). The GFP-Env protein was cloned into the retroviral vector pczCFG5 IEGZ (Lindemann, unpublished) (Fig. 3 ). This vector has additional GFP sequences linked via an IRES element, but GFP expression derived from IRES-GFP in transduced cells is barely detectable. GFP expressing cells always showed staining of the endoplasmatic reticulum (ER)/Golgi and plasma membrane but not of the nucleus. This is the expected pattern for Env, indicating that the green fluorescence detected derived from GFP-Env (Fig. 4B ). Co-transfection of equal amounts of both plasmids into NIH3T3 cells resulted in the spread of both genomes, which was detecteable by the appearance of green and red fluorescence (Fig. 4A , green, red and double positive). Separation of the viral genomes strongly delayed viral growth and we did not observe 100% double-positive cells in any of the transfections. Since the expression of Env in the target cell leads to receptor down-regulation (receptor interference), Env-expressing cells should no longer be transducible. This could explain the selected appearance of GFP-positive cells, but their rapid increase starting day 12 also points towards the generation of full-length MLV genomes containing GFP-Env. We therefore, analyzed the integrity of the viral genomes by PCR. Both split genomes were co-transfected in different ratios into NIH3T3 cells and genomic DNA was isolated at the time points indicated in Figure 5 . Primers derived from the pol and the env regions (p1, p2; Fig. 3 ) were used to study the generation of full-length MLV from the split genomes. As indicated in Figure 5A , lane 3, a 600 bp fragment can be amplified from full-length MLV DNA using these primers. The split genomes do not give rise to a DNA fragment, because the primer binding sites are on separate genomes (Fig. 5A , lane 2). After 13 days of culture, the appearance of a full-length MLV recombinant could be observed when the vector genomes were co-transfected in a ratio of 1:1 (gag/pol:env) (Fig. 5A , lane 5) and after 32 days, wt MLV could be detected in all samples (Fig. 5A , lanes 9, 10 and 11). This illustrates that full-length MLV was generated from the split viral genomes after prolonged passage. Figure 3 Schematic representation of fluorescently labeled semi-replicative retroviral vectors. The env open reading frame was replaced with the gene for red fluorescent protein (RFP) in the gag/pol-expressing construct, GAG/POL-RFP, and GFP-tagged Env was expressed from a packagable vector (GFP-Env). Positions of primers used to analyze the appearance of replication-competent viruses and the stability of the inserted GFP sequences by polymerase chain reactions (PCR) are indicated as p1 to p4. SA: splice acceptor site; SD: splice donor site. Figure 4 eplication of semi-replicative retroviral vectors. (A) Replication of semi-replicative retroviral vectors in transfected NIH3T3 cells, monitored by detection of green, red or double fluorescent cells by flow cytometry.(B) NIH3T3 cells expressing GFP-Env. The green fluorescence of the GFP-Env fusion protein can be detected in regions surrounding the nucleus (ER/golgi) and in the plasma membrane.(C) NIH3T3 cells expressing GAG/POL-RFP. RFP expression can be detected all over the cell, since RFP is not fused to a viral protein and is able to freely diffuse. Figure 5 PCR analysis of genomic DNA from NIH3T3 cells transfected with semi-replicative retroviral vectors. (A) The generation of full-length MLV genomes was analyzed by PCR using the primers p1 and p2 (see Fig. 4). Full-length MLV generates an 800 bp PCR fragment, semi-replicative retroviral vectors should not give rise to a DNA fragment because the primers do not bind to the same genome. DNA was transfected in different molar ratios as indicated. The first number indicates the molar ratio of the gag/pol plasmid and the second the Env encoding plasmid.(B) The stability of the GFP sequences inserted into the Env gene was analyzed by PCR using the primers p3 and p4 (see Fig. 4). The gfp-env sequence gives rise to a 1.5 kb fragment and wt env to an 800 bp fragment. Untransfected NIH3T3 cells were cultured in parallel and analyzed identically. The data are given as negative at days 13 and 32. NTC, no template control. In addition, we examined the stability of the GFP-tagged Env in the split genome approach. As shown in Figure 5B , PCR analysis with primers flanking the GFP sequences in Env (p3, p4; Fig. 3 ) clearly demonstrated that GFP-Env is stable and the GFP sequences were not deleted from the viral genome after 32 days of culture (Fig. 5B , lanes 5, 6 and 7). Discussion Our data demonstrate that labeling the MLV Env with a fluorescent protein is an easy method of monitoring MLV replication and the attachment of Env to target cells. This is especially useful for the development of novel cancer gene therapies that use replication-competent MLV encoding a cytotoxic gene [ 3 ]. Labeling Env with GFP in the PRR leaves the 3' untranslated region at the Env boundary available for the insertion of IRES-linked therapeutic genes [ 1 ]. These recombinant viruses could be monitored by GFP expression and would allow the study of replication kinetics in vitro and in vivo . The biodistribution of replication-competent viruses in animal models and their safety for cancer treatment could, thereby, be assessed. A further improvement of replication-competent viruses would be tumor cell-specific entry. The inclusion of tumor-specific ligands into Env is one option to potentially expand the ecotropic host range of MLV to human tumor cells [ 6 , 5 ]. Ecotropic MLV containing GFP-tagged Env can be used to analyze the receptor-dependent binding of the viral Env proteins to target cells. Labeling Env in the PRR leaves the N-terminus or the receptor binding site [ 13 ] available for further insertions of ligands to target tumor cell specific receptors. The use of GFP-tagged Env to determine receptor binding is very simple and in addition GFP-tagged Envs are helpful for the identification of recombinant viruses from retroviral library screens. GFP-Env fusions will therefore be very useful for the development of targeted vectors and as a screening system for retroviral-receptor antagonists. However, selecting EGF-Env containing MLV on cells that express both receptors (EGF- and ecotropic receptor) did not permit the isolation of a virus with an EGF-receptor specific tropism. EGF sequences were deleted from the viral genome in this setting. EGF sequences in Env, however, did not alter the replication kinetics in mouse fibroblasts (Fig. 1 ), which further indicates that targeting retroviruses to membrane spanning receptor tyrosine kinases inactivates retroviral particles. In our experiments using semi-replicative retroviral vectors, we found that a rapid increase in GFP-positive cells correlated with the appearance of recombinations and the formation of full-length MLV genomes. This indicates that semi-replicative vectors have to be improved to avoid intergenomic recombination before they can be considered to be used for gene therapy. The recombinants did contain the GFP-Env gene, providing further proof that insertion of GFP into the proline-rich region of Env did not interfere with viral fitness. Conclusions Fluorescently labeled MLVs are excellent tools for the study of factors that influence viral replication and can be used to optimize MLV-based vectors or viruses for gene therapy. This method is not limited to ecotropic Env, but can be extended to amphotropic MLV, since it has been shown recently that the amphotropic MLV Env can also be tagged with GFP [ 14 ]. Methods Cell lines NIH3T3, A431, 293T and COS-7 cells were grown in Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal calf serum, 4 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in 10% CO 2 . Plasmids The construction of GFP-EMO has been described previously [ 7 ]. GFP-MOV was generated by replacing a Pfl MI fragment of pGFP-EMO with wt MLV sequences using standard cloning procedures [ 15 ]. GAG/POL-RFP was generated starting with the genomic MLV clone, pKAΔenv-egfp, which contains a 30 nucleotide-linker with an Sfi I-site introduced at position 5893 (all positions according to GenBank Accession No. J02255) and an additional Sfi I-site at position 5389 removed by mutation. The start codon of MLV env (position 5777) was deleted to allow translation to start at the inserted GFP sequence [ 16 ]. We replaced GFP with RFP, which was introduced as a Sfi I-Cla I fragment. GFP and RFP sequences were derived from vectors purchased from Clontech (BD Biosciences Clontech, Heidelberg, Germany) Transfections Plasmids encoding the MLV genomes or soluble receptor binding fragments (sRBDs) [ 12 ] were transfected using the calcium phosphate procedure [ 15 ]. For the sRBDs, supernatant was collected two days after transfection, filtered through a 0.45 μm pore filter (Millipore, Eschborn, Germany) and 1 ml was used per binding assay. Cell binding assay Supernatants of tissue culture cells were collected, filtered through a 0.45 μm pore filter (Millipore, Eschborn, Germany) and added to target cells. After 5 min. at room temperature, the cells were spun down, redispersed in PBS and immediately monitored by fluorescence-activated cell sorting (FACScan, Becton Dickinson, Heidelberg) using the Cellquest software. Fluorescence-activated cell sorter (FACS) analysis Green fluorescence protein (GFP) expression was monitored by a shift to green fluorescence (FL-1) and red fluorescent protein (RFP) by a shift to red (FL-2). FACS analysis was performed with FACScan (Becton Dickinson, Heidelberg) using the Cellquest software. Polymerase chain reaction (PCR) Genomic DNA was isolated after proteinase K digestion and phenol/chloroform extraction. PCR was performed using the manufacturers protocol (Qiagen, Hilden, Germany). N-terminal EGF-Env sequences were analyzed using the primers BS-5: 5'-TCT GAG TCG GAT CCC AAA TGT AAG and MLV-5'-Env: 5'-TAA CCC GCG AGG CCC CCT AAT CC, which amplified a 899 bp fragment from GFP-EMO and a 726 bp fragment from wt MLV. The generation of full-length genomes was analyzed using the primers p1: 5'-GAA TAG AAC CAT CAA GGA GAC and p2: 5'-CTC GAG AAG CTT AGT ACT GA, which amplify a 600 bp fragment from full-length MLV. No fragment should be amplified from the semi-replicative vectors, because the primers bind to genes on separate constructs. The stability of the GFP-Env fusion gene was analyzed using the primers p3: 5'-GTC AGT AAG CTT CTC GA and p4: 5'-GGT TTT GTC AGG ACT GGT GAG, which amplify a 1.5 kb fragment from gfp-env and an 800 bp fragment form wt env. Competing interest The author(s) declare that they have no competing interests. Authors' Contributions Katja Sliva and Alexandra Bittner performed the experiments. Katja Sliva, Otto Erlwein and Barbara Schnierle participated in the design of experiments, oversight of the conduction of the experiments, and in the interpretation of the results.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544597.xml
544583	Treatment of pregnancy-related pelvic girdle and/or low back pain after delivery design of a randomized clinical trial within a comprehensive prognostic cohort study [ISRCTN08477490]	Background Pregnancy-related pelvic girdle and/or low back pain is a controversial syndrome because insight in etiology and prognosis is lacking. The controversy relates to factors eliciting pain and some prognostic factors such as the interpretation of pain at the symphysis. Recent research about treatment strategies also reflects those various opinions, in fact suggesting there is professional uncertainty about the optimal approach. Currently, physiotherapists often prescribe a pain-contingent treatment regime of relative rest and avoiding several day-to-day activities. Additionally, treatment more often includes an exercise program to guide rectification of the muscle imbalance and alignment of the pelvic girdle. Effectiveness of those interventions is not proven and the majority of the studies are methodologically flawed. Investigators draw particular attention to biomedical factors but there is growing evidence that important prognostic issues such as biopsychosocial factors appear to be even more important as point of action in a treatment program. Methods/design This pragmatic randomized controlled trial is designed to evaluate the effectiveness of a tailor-made treatment program with respect to biopsychosocial factors in primary care. The effect of the experimental intervention and usual care are evaluated as they are applied in primary health care. The trial is embedded in a cohort study that is designed as a longitudinal, prospective study, which studies prevalence, etiology, severity and prognosis during pregnancy until one year after delivery. The present paper focuses on choices regarding recruitment procedures, in-/exclusion criteria and the development of a well-timed intervention. Discussion This section briefly discusses the actions taken to minimize bias in the design, the proper time-window for the experimental intervention and the contrast between the experimental intervention and usual care.	Background Since 1962[ 1 ], diagnosis, prognosis and treatment of pregnancy-related pelvic girdle and/or low back pain have inflicted debate and have led to considerable differences of opinions. Many articles appeared mainly in International journals and some etiological mechanisms were hypothesized. However, the subject remains controversial, mainly because insight in etiology and prognosis is lacking. Moreover, diagnostic investigation into the exact definition and classification of pregnancy-related pelvic girdle and/or low back pain shows various opinions between leading experts on this topic. The controversy relates to factors eliciting pain[ 2 ] and prognostic factors such as the interpretation of pain at the symphysis [ 3 , 4 ], the question whether pelvic girdle pain is a syndrome separate from low back pain [ 4 , 5 ] and the importance of questions about limitations in activities [ 6 ]. Also recent research about treatment strategies reflects those various opinions [ 7 ], in fact suggesting there is professional uncertainty about the optimal approach. Investigators draw particular attention to biomedical factors but there is growing evidence that important prognostic issues such as biopsychosocial factors appear to be even more important as basis in a treatment program[ 8 , 9 ]. Although the group of musculoskeletal disorders holds many different biomedical labels, the process of developing chronic disability has shown surprising similarities with regard to biopsychosocial factors [ 10 ]. For the moment, pregnancy-related pelvic girdle and/or low back pain is a subjective experience comprising pain and limitations in activities for which classification criteria are insufficient in guiding to a treatment approach (Bastiaenen et al. personal communication). Results of various therapeutic interventions have been published but excepting one recent study[ 11 ], their effectiveness remain unproven. Furthermore, the majority of the studies are methodologically flawed [ 7 ]. Currently, physicians and physiotherapists usually prescribe a pain contingent treatment regimen of relative (bed) rest and avoiding several day-to-day activities such as using the stairs, bending, twisting, lifting and cycling. Additionally, the usual treatment approach of a physiotherapist more often includes an exercise program to guide rectification of the muscle imbalance and alignment of the pelvic girdle [ 12 ]. Therapists rely on knowledge of pain duration and intensity during goal-setting for treatment, for a great deal. Why publish a study protocol There are several reasons for publishing a study protocol before obtaining research data. The main reason is to reflect on the study design independently of the results. Considerations and choices concerning methodology and treatment can be described more detailed. The present paper focuses on choices about recruitment procedures, in-/exclusion criteria and the development of a well-timed experimental intervention. We also present details about the enrollment of women with pregnancy-related pelvic girdle and/or low back pain in the controlled trial. Methods/design Study design and research question The trial is embedded in a cohort study that is designed as a longitudinal, prospective study, which studies the prevalence, etiology, severity and prognosis of pregnancy-related pelvic girdle and/or low back pain until one year after delivery (Figure 1 ). The present study is designed as a pragmatic trial aimed to compare the effects of interventions carried out in primary health care. Figure 1 Design of the study We performed a randomized controlled trial in primary care to determine whether a tailor-made program with respect to biopsychosocial factors (intervention group) benefits women with pregnancy-related pelvic girdle and/or low back pain more in terms of effectiveness and costs than usual care would on a traditional pain contingent basis (control group). Recruitment and informed consent The medical ethics committee of the Maastricht University Hospital approved the intervention and cohort study. The study is performed in the Southeast of the Netherlands. Midwives and gynecologists recruited the women during early pregnancy (10–14 weeks). Participation of midwives and gynecologists in the recruitment of eligible pregnant women is of major importance for the success of the cohort and intervention study [ 13 ]. We have paid a lot of attention to difficulties in recruitment such as busy consultation hours and not feasible recruitment procedures. Therefore, we designed a recruitment protocol that is as simple as possible, not restrictive, and demanding a minimum of time from midwives and gynecologists. Standardized written information about the cohort and intervention study is available for every potentially eligible woman and to be handed out by the midwife or gynecologist. Several steps are taken to encourage participation of the midwives and gynecologists. We distributed newsletters about the developments in the cohort and intervention study every three months and visited the practices and meetings of midwives on a regular basis. Any questions regarding trial questions received prompt feedback. The flexibility of the trial procedure is also guaranteed by assessing potential candidates for the trial at home. Women are included in the cohort if they are at least 18 years old, pregnant and well versed in Dutch language. Women are given written information explaining the aims and contents of the cohort and intervention study before they decide to participate. Concerning the intervention study they are told that to current knowledge the two investigating treatment options are considered to be equally effective. The moment of inclusion for the intervention study lies at about three weeks after delivery. An individual woman enters the intervention study after signing informed consent for both the cohort and intervention study during early pregnancy and meets the in- and exclusion criteria of the intervention study three weeks after delivery. Women are included when having pain in the pelvic girdle and/or low back with an onset during pregnancy or just after delivery (cohort data), are restricted in their normal daily activities because of pelvic girdle and/or low back pain and if there is a delay in recuperation (not yet in the condition to participate satisfactory in housekeeping and care of children because of the complaints under investigation). The severity of symptoms must be varying with physical activities and time during the day. Women diagnosed with a relevant specific pathology (such as nerve root pathology, rheumatoid disorders, carcinoma, obstetric complications) that affects pain and activities of daily life are excluded. Exclusion also occurs in case of family related or psychosocial problems or when a disablement procedure is not yet finished. Final important aspects for in-/exclusion are the willingness of a woman to participate in the study or having a clear treatment preference[ 13 ]. We only included women who did not indicate such a preference and who were willing to take the 50% risk of receiving a referral to a participating physiotherapist (and treatment option) or the freedom of choosing a therapist by themselves (usual care). Including only women who are naïve (who never received treatment for their complaints during this pregnancy or earlier pregnancies) will result in an unacceptable reduction in the number of eligible patients. However, we excluded all the women who already received treatment after their current delivery. A basic principle for selection of eligible women in this study is that inclusion criteria must have a meaningful influence in goal setting for treatment. We therefore focused on criteria such as a delay in recuperation and restrictions in normal daily activities caused by pregnancy-related pelvic girdle and/or low back pain. However, other studies in this field formulated inclusion/exclusion criteria based on certain diagnostic classification strategies. Although rationales of these strategies greatly differ, they all attach great importance to the outcomes of particular (albeit different) diagnostic tests. In the absence of a clear definition and reference standard to diagnose pregnancy-related pelvic girdle and/or low back pain, the outcomes of these procedures not only led to different selections of women having complaints, the prognostic and diagnostic importance of these subgroups also remain unclear (Bastiaenen et al. personal communication). Exclusion of differential diagnoses is a major point of concern. For that reason we included a history taking and a physical examination protocol that focuses on differential diagnoses at first and then on the formulated inclusion criteria. The various specific physical examination tests to diagnose pregnancy-related pelvic girdle pain are left aside. For a better understanding of the complaints and tailoring treatment, application of these tests has no supplemental value (Bastiaenen et al. personal communication). An experienced research-physiotherapist visited women at home, about three weeks after delivery. This visit is called for on the basis of a short self-administered questionnaire and/or initiated by midwives. A positive answer from a participating woman from the cohort and/or her midwife on the question: "Do you or does this woman need treatment?" took a central position in these questionnaires. In advance of a possible home visit, a short history taking by telephone took place about two weeks after delivery. History taking focuses on exclusion criteria such as: willing to participate in this part of the study, a diagnosis with relevant specific pathology, limitations in daily life caused by pregnancy-related pelvic girdle and/or low back pain and a delay in recuperation. During a home visit, a standardized history is taken and physical examination to exclude specific pathology is performed. Self-administered questionnaires are used to question the women about pain, limitations in activities, restrictions in participation, pain-related fear, pain catastrophizing, positive and negative affectivity, depression, expectancy of treatment result and quality of life. The questionnaires contain clear instructions for completion with no help or support from others. If a woman meets the selection criteria, she is informed about the aim and method of the intervention study and if she is willing to participate, the informed consent procedure is completed. The research-physiotherapist collecting the baseline data is trained in performing the measurements in a standardized way and is unaware of the women's treatment assignments. Randomization and blinding Randomization takes place after collecting the baseline data. In this study we used a block randomization (size of four). An independent research assistant (unaware of the baseline data) carried out the randomization procedure according to a random computer-generated list. When a woman is allocated to the intervention group, the participating physiotherapist in the environment of the woman is contacted and we ensured that treatment could start as soon as possible (within one week). Treatment is covered for all participants in the intervention group on a research-physician's referral. Women, allocated to the usual care group, are free to choose usual treatment by a (not participating) physiotherapist. Information about a possible guidance by a general practitioner and feasible treatments received after randomization is collected by means of questionnaires in the follow-up period. Women are blinded to a certain extent to the allocated treatment because they are kept naïve of the exact content of both treatment options. Participating physiotherapists are not blinded to the treatment option but not involved in the baseline and effect measurements. Researchers dealing with the baseline and outcome data are unaware of the treatment assignments. History and physical examination During a home visit a standardized history is taken [ 8 ] and physical examination is performed. History taking focuses on on-going pain, its location, intensity and modalities, variation of symptoms with physical activities, radiation into the legs, back pain versus leg pain, neurological signs, deformity, obstetric complications, a case history of low back and pelvic girdle pain prior to this pregnancy and other differential diagnoses. The format of the answers is presented as a dichotomous "yes or no". Demographic characteristics and data about education, work, income, use of alcohol, smoking, medication, the onset of pain and functional status during pregnancy have already been gathered as part of the cohort study at 14 and 30 weeks gestation period and two weeks after delivery. After history taking a short standardized clinical examination program is performed, which includes tests of nerve root radiation (exclusion)[ 8 ]. The research-physiotherapist fills out the Pain Behavior Scale, a standardized observation scale for quantifying pain behavior [ 14 , 15 ], after clinical examination. Interventions Usual care Prior to the trial, detailed information is gathered about the contents of traditional treatment options. Part of the information is collected by means of group discussions with experienced physiotherapists and occupational therapists and interviews on an individual basis with affected women out of the cohort. An independent rehabilitation specialist, specialized in pain treatment chaired the meetings with the therapists. Some subjects for discussion were: differences in clinical spectrum seen by the therapists, contents of treatment programs during pregnancy and after delivery, common knowledge by the therapists about etiology, prognosis and prevalence of the syndrome, the optimal time-window for treatment in the course of complaints and the therapist-patient relationship. Items that provided important topics of conversation between the therapists were: the moment of taking up and finishing off treatment, the contents of education and advice given to the patient and the (lack of) compliance. The most striking characteristics of a traditional treatment were the character of the therapist-patient relation and the way of goal setting, focusing on disease management [ 16 ]. There was an explicit professional input and an accent on biomedical factors. A pain contingent regimen of avoiding and limiting several day-to-day activities was important. Compliance and adherence based on these goals played an important part. Therapists were often highly concerned about their patient's pain themselves. However, interviews with affected women made clear that most of the women were irritated about this regimen in an increasing degree after starting the treatment sessions. The regimen was too strict and on a number of points not geared to the wishes and concerns of the women. These aspects caused a lack of compliance and an unremitting hesitation about a good prognosis and in particular about reassuming certain day-to-day activities after delivery. Therapists did not realize the nature of this problem although they did mention problems with compliance. Some women were not able to get a grip on their condition and left management of their pain and activities of daily life to the therapist. A larger part of the women was more or less uncertain about picking up their full range of activities again after delivery. Their beliefs and concerns about origin and prognosis of their complaints clearly bore the stamp of the introduced biomedical label. The relatively favorable prognosis after delivery was largely unknown to the women as well as to the physiotherapists. Experimental therapy Women, allocated to the intervention group, are referred to a participating physiotherapist in their own neighborhood. These physiotherapists received an educational course about the treatment protocol prior and during the study. All physiotherapists were already experienced and specialized in treating women with pregnancy-related pelvic girdle pain prior to the study. The contents of the experimental therapy are based on the latest literature, results of interviews with affected women (participating in the cohort study) and group discussions with experienced physical and occupational therapists. A search procedure in literature resulted in various therapeutic interventions. However, effectiveness of those interventions remain unproven. An important common goal of these treatments is restoration of optimal biomechanics, although this is not based on established theoretical principles [ 7 ]. The search did not provide enough possibilities to design a treatment protocol. However, as mentioned above, results of interviews and group conversations showed interesting contradictions. During development of the experimental intervention we focused on the following contradictions: patient-therapist relationship, education, and hesitation or avoiding of activities. Theoretical concepts of self-management [ 16 , 17 ] and fear-avoidance [ 18 ] were integrated in the treatment protocol. A treatment program that demands a much more active involvement of a participating woman was designed. Interventions with a self-management approach are considered to be able to build a bridge between patients' needs and caregivers' services to meet those needs. Self-management refers to the individual's ability to manage the symptoms, treatment, physical and psychosocial consequences and life style changes inherent to living with a chronic condition [ 17 ]. Self-management approaches are either group-based or individualized. We performed an individualized approach of 7–9 sessions of 30 minutes once a week. Standardized information is presented through a treatment protocol for the therapists and booklets for the patients [ 16 , 19 ]. Topics included back and pelvis anatomy, "red flags" indicating a serious medical condition, factors contributing to fluctuations in pain, appropriate pacing of exercises [ 12 ] and activity, handling pain flare-ups, cognitive restructuring, some graded exposure techniques [ 18 , 20 , 21 ], communication and social persuasion. Therapists had to employ problem-solving techniques that engaged women in identifying day-to-day problems or limitations related to pelvic girdle and/or low back pain, setting personal goals, brainstorming options for achieving these goals and developing personal action plans. In subsequent sessions, women reviewed their action plans and their progress towards goals and engaged in problem-solving skills around difficulties that arose in trying to implement their plans. Information about two opposing behavioral responses of pain-related fear (avoidance and confrontation) is given, and a hierarchy of individual fear-eliciting movements and activities is made. Therapists encouraged women in making action plans for specific activities that were avoided. Complaint-related problem solving is a key skill. The role of the therapist is to encourage women to identify possible causes of a problem, find a number of potential solutions, select one, then try it and finally evaluate the results and possibly adjust the solution. The second important key skill is action planning or goal setting. Often a plan must have been generally unacceptable (such as "go skiing") for a therapist in the usual care. Nevertheless, the protocol of the experimental intervention embraced the point of view that a woman is her own best judge of what is possible. Another major point of action planning is that a woman could not only receive but also give feedback on her own accomplishments. Endorsement by the therapist is very important for a woman to accept her new role. This way of collaborating with a therapist on short-term action planning enabled women to master new skills and to make changes that are realistic and feasible for them. Therapists also have a role in assisting women in understanding their symptoms. Knowledge of the course of the complaints during pregnancy and after delivery including pain flare-ups in the year after delivery, factors contributing to fluctuations in pain, evidence-based knowledge about etiology and the concept about pain-related fear are essential. Symptoms are explained as having many but not alarming causes, which offers the possibility to choose different actions by the concerning woman. Finally, therapists have a task in practicing social persuasion. A woman is more likely to change her behavior and have confidence in doing so if she perceives those around her, including the therapist to be supportive. A relationship in which the physiotherapist and the woman make health care decisions together is the basic assumption of the intervention. Generally, a time contingent policy is followed in which women set the pace by means of action plans. The expertise of the physiotherapists of the condition in general and of the women about their own specific condition and lives are equally important [ 22 ]. Outcome measurements Outcome measures (Table 1 ) chosen to explore the success of any intervention need to match the desired aims of that intervention. It is a process in which a standardized attempt is made to observe an often complex clinical picture. Primary domain for improvement of the treatment under investigation is limitations in activities. Other important domains are the severity of the main complaints, the woman's global feeling of recovery, pain and participation. Table 1 Timing of measures Baseline (about 3 weeks after delivery) 12 weeks after randomization 6 months (after delivery) 1 year (after delivery) History taking X Physical Examination: X PBS X GPE X X X MC X X X X MPQ(VAS) X X X X RDQ X X X X QBPDS X X X X TSK X X X X PCS X BDI X NEM X PEM X Expectancy treatment result: X SF-36 X X X X EuroQol X X X X IPA X X X X Cost-diary X X X Satisfaction treatment : X X Recurrence X X Co-interventions X Compliance X X Subsequent pregnancy : X X PBS = Pain Behavior Scale GPE = Global erceived Effect MC = Main Complaint MPQ = McGill Pain Questionnaire RDQ = Roland Disability Questionnaire QBPDS = Quebec Back Pain Questionnaire TSK = Tampa Scale For Kinesiophobia PCS = Pain Catastrophizing Scale BDI = Beck Depression Inventory NEM = Negative Emotionality Scale PEM = Positive Emotionality Scale SF-36 = Short-Form-36 IPA = Impact on Participation and Autonomy Limitations in activities are measured with the Dutch translation of the Roland Disability Questionnaire (RDQ) [ 23 ] and the Quebec Back Pain Disability Scale (QBPDS) [ 24 , 25 ]. We added the phrase "because of my back and/or pelvic pain" in both questionnaires. Subjective measurements like global feeling of recovery (global perceived effect, GPE) and severity of the main complaints (MC) reflecting a patient-specific approach are also selected. Global Perceived Effect (GPE) is measured by self-assessment on a 7-point scale (1 = completely recovered, 7 = worse than ever). The main complaints (MC) are selected by the woman in a standardized approach by selecting three activities, which are an essential and frequently performed part of her everyday life. However, the performance is difficult or impossible because of low back and/or pelvic girdle complaints at the moment of baseline measurement. Severity of a main complaint is rated on a visual analog scale (VAS). [ 26 , 27 ]. Pain is measured with two VAS-scales of the McGill Pain Questionnaire (MPQ-DLV) [ 28 , 29 ] to record the intensity of pain the last week and day. The impact on participation and autonomy (IPA) is used to measure person-perceived restriction in participation and autonomy [ 30 , 31 ]. The used subscales are autonomy in self-care, mobility and leisure, social relationships and family role. Other important prognostic factors that can influence treatment results are fear of movement, pain catastrophizing, depression, negative and positive affect, expectancy of treatment result and pain behavior. Fear of movement is measured by the Dutch translation of the Tampa Scale for Kinesiophobia(TSK)[ 32 , 33 ]. We used the TSK and the both subscales "fear avoidance" and "harm"[ 34 , 35 ] Pain catastrophizing is measured by the Pain Catastrophizing Scale (PCS)[ 36 , 37 ]. The Beck Depression Inventory (BDI) [ 38 ] measures depressive symptoms [ 39 ]. Analyses of the BDI in this study did not include items concerning weight loss, sleeping disturbance and work inhibition [ 40 ] To measure the experience of negative affect we used the 14-item Negative Emotionality Scale (NEM) [ 41 ]. To measure positive affect we used the 11-item Positive Emotionality Scale (PEM) [ 41 ]. Both are subscales of the Multidimensional Personality Questionnaire. Health status is evaluated by the Short-Form 36 (SF-36)[ 42 , 43 ] and the EuroQol [ 44 ]. We used the subscale "general health". A cost-diary [ 45 ] is used to obtain data on physical activities, health care utilization, and days of sick leave. Women are instructed to record costs on a weekly basis until one year after delivery. Expectancy of therapy result [ 46 ] is measured by means of a 100 mm visual analog scale (VAS). The woman is asked to what extent she believes that a treatment is beneficial to her. The Pain Behavior Scale (PBS)[ 14 , 15 ] is an observation scale tapping 8 pain behaviors that the physiotherapist completes after physical examination. These are verbal complaints, vocal complaints, facial grimaces, standing posture, mobility, body language, use of visible supportive equipment and stationary movement. Follow-up Women are asked to complete follow-up questionnaires at 12 weeks after randomization, 6 months after delivery and one year after delivery. Women who did not return their follow-up questionnaires were contacted by mail or phone and were asked to continue participation. Compliance, other interventions and confounding The follow-up questionnaires ask all women how many treatment sessions they have followed in the previous period of time. Furthermore, information on contents, satisfaction and the aspects of the (experimental) treatment which benefited them most, is gathered. Co-interventions, medication, aids, additional medical consumption, recurrence of complaints, return to gainful employment and a possible subsequent pregnancy are also registered. Physiotherapists who treat the participants of the intervention group also answered questions about the number and contents of the treatment sessions after conducting the last meeting. Statistical analyses Statistical analyses are carried out according to the "intention-to-treat" approach. The baseline status of the study groups is compared with respect to the distribution of all independent prognostic variables and the baseline values of the outcome variables. For the outcome measures recorded at baseline and at follow-up, we computed the difference between the baseline and the follow-up score for each woman. Differences between groups and 95%CI are calculated for each outcome measure according to the intention to treat approach. Primary analysis is done by means of analysis of an independent t-test (for continuous outcome variables) and chi-square test (for categorical outcome variables). In order to adjust for possible baseline differences a multiple linear regression analysis for continuous outcome measures is performed with the change scores as dependent variable, treatment option as independent variable and base line scores of the prognostic variables as co-variables. Missing data at the baseline-measurement are substituted by the "mean of series" imputation method. Longitudinal missing data are substituted with the "last value carried forward method". In all comparisons between the two treatment options a two-tailed p-value of 0.05 is considered to indicate statistical significance. Prognostic status at baseline for women with and without missing values for the outcome variables is compared for both groups. Analyses are done by using SPSS statistical software, version 12.0 (SPSS, Inc., Chicago, Illinois). Short term and long term effect analyses are performed separately. Economic analyses A cost-effectiveness analysis compares the costs and health effects of the experimental intervention to assess whether it is beneficial from an economic perspective. The costs of the intervention are calculated separately for the intervention group. For the whole study group all relevant health care costs, production loss and patient and family costs are measured by means of a cost-diary [ 45 ] and follow-up questionnaires collected 6 months and one year after delivery. Both direct health care costs (such as physician visits, the number of treatment sessions and medication), direct non-health care costs (such as transport to therapist) and indirect costs associated to the complaints (like sick leave, professional as well as voluntary aid and extra baby sitter) are registered until one year after delivery. Quality of life is measured using the EuroQol. [ 44 ]. For the validation of the healthcare costs, patient and family costs, an update of the Dutch manual for costing in economic evaluations is used. The primary outcome measure for the cost-effectiveness analysis is the difference in limitations in activities (RDQ)[ 23 ]. Details about enrollment in the study During the study, 397 of the 7526 women (5%) signed only for the cohort study (n = 7526) and were therefore beforehand excluded for taking part in the intervention study. Throughout pregnancy, 73% of all women in the cohort reported pain in the lumbar/pelvic region leveling off to 35.9% three weeks after delivery (Figure 2 ). The "three weeks after delivery" prevalence rate of "wanted to be referred for treatment" was 4.8% at that moment and remained remarkably stable in the year after delivery. Figure 2 Prevalence of pelvic girdle and/or low back pain during pregnancy and after delivery Since November 2000 (Figure 3 ), 682 women reported that they need treatment during pregnancy (9% of the total cohort). 384 times midwives indicated that a woman need treatment at the time of 10 days after delivery (5% of the total cohort). On 197 occasions, both the woman and her midwife responded positive. The outcomes resulted in 869 possible eligible participants (11.5% of the total cohort). However, these data resulted in only 147 home visits, 99 visits indicated by a midwife (67 times in combination with the woman concerned) and 115 indicated by the woman (Figure 3 ). On basis of history taking by telephone, 722 women were excluded from participation. Ten women did not give informed consent for the intervention study, 3 women moved outside the area intervention was provided, 13 women were excluded because of specific pathology, 49 women did not want to be randomized (clear treatment preference deviating from the study protocol) and 12 women did not feel like participation on second thought. The majority, 635 women, were excluded because of a quick recovery. Figure 3 Enrollment in the intervention study After the home visits, 21 women were excluded. Indicated by a midwife; one because of family reasons, one because of specific pathology and two women because of quick recovery. Indicated by themselves, 17 women were excluded. Two women because of family and social problems, one woman because of specific pathology, one because of a clear treatment preference and 13 women because of quick recovery. Eventually, 126 women were included in the intervention study. 93 times indicated by midwives, 56 times indicated both by the woman and her midwife and 89 times by themselves. Finally, only 24.2% of the women indicated by a midwife were included and 13% indicated by themselves. Indicated by both the woman and her midwife, 28.4% was included. Discussion This study is designed to evaluate the effectiveness of a tailor-made program with respect to biopsychosocial factors. A pragmatic design provides the opportunity to evaluate the value of the experimental intervention without depriving participating patients of the best current treatment option. Including only women who would take the 50% risk of depriving any treatment at all for their complaints during the first 12 weeks after delivery was not a realistic option. The effects of the experimental intervention and usual care are evaluated as they are applied in primary health care. It is not feasible to blind a woman to the applied treatment option, which increases the risk of information bias. We have tried to minimize this type of bias by assessing treatment preference before randomization and excluding women with a clear treatment preference. Details about the enrollment of the trial underscored this necessity (n = 50 excluded because of a clear treatment preference). The research-physiotherapist dealing with the baseline measurement was therefore unaware of treatment allocation. Details about the enrollment out of the cohort into the trial also show that the start of the experimental intervention is well timed. Most women have complaints during pregnancy. However, a considerable drop in the number of women having persistent complaints in the first weeks after delivery is observed. Then again, numbers of women having one or more episodes of pain complaints remained stable in the year after delivery. We have seen similar trends of proportions of women with a request for treatment for their complaints during pregnancy (9%), just after delivery (4.8%) and in the year following delivery. The aim of the experimental intervention is to increase the level of activities. Therefore the primary outcome measure is limitations in activities. The contrast between both interventions is an important issue in this study. Major features that underscore the contrast are the character of the patient-therapist relationship, pain-contingent versus time-contingent treatment, compliance to a regime of avoiding and limiting activities versus action planning and personal goal setting by the women themselves. Among therapists, the approach of the experimental intervention is not widespread at all. The participating therapists are explicitly asked to not give any information about the contents of the experimental treatment to therapists who do not participate in the experimental intervention. It is necessary to interest physiotherapists in the trial for an efficient performing of the experimental treatment option, which can be achieved by a relevant research question and in practice applicable results. Competing interests The author(s) declare that they have no competing interests Authors' contributions CHGB: first author, is involved in the design, data collection, statistical analyses, and the development of the experimental intervention. RAdB: participated in the design, coordination, statistical analyses, have made substantial contributions to the development of the experimental intervention, is involved in revising the article for important intellectual content. PMJCW: participated in the design, has made substantial contribution to the development of the experimental intervention and is involved in revising the article for important intellectual content JWSV: has made substantial contributions to the development of the experimental intervention and is involved in revising the article for important intellectual content JMB: participated in the design and statistical analyses, is involved in revising the article for important intellectual content ABAK: participated in the design and collecting of the data AH: participated in the design, recruitment and collecting of the data PvdB and GGME; participated in the design, have made substantial contributions to conception, design and experimental intervention and revising the article critically for important intellectual content. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544583.xml
546419	Diet and physical activity behavior among users of prescription weight loss medications	Background There is limited population-based data on diet and physical activity behaviors and weight loss among users of prescription weight loss medications. Most findings are from clinical settings or from research that includes organized behavioral programs. Methods We analyzed data from the 1998 Behavioral Risk Factor Surveillance System, an annual telephone survey conducted in all fifty states, the District of Columbia and Puerto Rico. The sample consisted of 135,435 noninstitutionalized adults aged 18 years old and older. We determined the prevalence and odds of prescription weight loss medication use, odds of 10% weight loss, and among current weight loss medication users, the prevalence and odds for diet and physical activity behaviors. Results 10.2% of obese women and 3.1% of obese men reported using prescription weight loss medications in the past 2 years. Of users, 28.2% had lost at least 10% of their pretreatment body weight. The odds of losing at least this much weight were higher among women, those who usually consumed ≥ 5 fruits and vegetables daily and those who met physical activity recommendations. Among current prescription weight loss medication users, 26.7% reported both eating fewer calories and meeting recommended leisure-time physical activity levels (<40% of any group met both). Of those meeting both recommendations, almost half (47.2%) had lost 10% of their pretreatment body weight. Of current users, 9% reported using the medications for weight maintenance. Conclusions Only 26.7% of prescription weight loss medication users reported following recommended diet and physical activity behaviors. Further research is needed to assess whether behavioral changes are associated with greater weight loss and maintenance among prescription weight loss medication users.	Background Lifestyle modifications, including behavior therapy, diet, and physical activity, are the cornerstone of weight management [ 1 ]. Current weight management guidelines state that prescription weight loss medication only may be used as part of a comprehensive weight loss program including diet and physical activity modifications [ 1 ]. These modifications include increasing physical activity to 30 to 45 minutes on most days of the week and reducing daily caloric intake by 500 to 1000 calories. Clinical research also suggests that increasing consumption of whole grains, fruits, and vegetables in place of calorie-dense foods may increase satiety and decrease overall caloric intake aiding weight management [ 2 ]. The initial goal of weight loss therapy is to reduce body weight by approximately 10% over a 6-month period [ 1 ]. Even small losses in weight, such as 5% to 10%, have been found to reduce blood pressure, cholesterol and triglycerides levels, and reduce blood glucose levels among overweight and obese persons without diabetes [ 3 , 4 ]. Weight loss attributable to prescription weight loss drugs is modest, about 3% to 8% of body weight, and is based on a small number of patients assigned to drug treatment in clinical trials [ 5 - 8 ]. Lifestyle modification in combination with pharmacotherapy has been shown to improve weight loss more effectively than pharmacotherapy alone [ 9 ]. However, in the general population, data suggests that only one-fifth of individuals who were trying to lose weight reported using the recommended combination of eating fewer calories and meeting weekly physical activity recommendations [ 10 ]. Using population-based data we estimated that 4.6 million American adults used prescription weight loss medications between 1996 and 1998 [ 11 ]. During this period, obesity medications included phentermine, fenfluramine, dexfenfluramine, and sibutramine. Because of the paucity of population-based data on behaviors among persons using prescription weight loss medications, the purpose of our study was to further examine demographic and behavioral characteristics related to the reported use of prescription weight loss medications collected in the late 1990s as part of a national dataset. These data were collected the same year as issuance of guidelines for assessment and treatment of obesity and therefore provide baseline estimates. Specifically, we examined 1) prescription weight loss medication use by demographic and behavioral characteristics, 2) 10% weight loss among prescription weight loss medication users by demographic and behavioral characteristics, and 3) diet and physical activity behaviors among those who used prescription weight loss medications at the time of the survey. Methods We examined data from the 1998 Behavioral Risk Factor Surveillance System (BRFSS). The BRFSS is a telephone survey of health practices that is conducted each year by all state health departments. Each state, the District of Columbia, and Puerto Rico select an independent probability sample of noninstitutionalized residents aged 18 years and older. In 1998, 149,806 persons responded to the BRFSS; several investigators have published detailed descriptions of survey methods and quality control indices [ 12 ]. The average 1998 state cooperation rate (completed interviews divided by completed, refused, and terminated interviews) was 73.4% (range, 45.4%–95.4%). To determine prescription weight loss medication use, respondents were asked, "In the past two years, have you taken any weight loss pills prescribed by a doctor? Do not include water pills or thyroid medications." Responses were coded as: 1) "Yes, I am currently taking them", 2) "Yes, I have taken them but I am not currently taking them", or 3) "No, I have not taken them". Respondents with a positive response were asked about their prepill weight, "How much did you weigh just before you started taking prescription weight loss pills for the first time?" At the end of the interview, respondents were asked to report their current height and weight without shoes. We used current weight and pretreatment weight to calculate current and pretreatment body mass index (BMI), respectively. BMI was calculated as weight (either current or pretreatment) in kilograms divided by current height in meters squared. BMI was categorized into 3 groups: normal weight (<25.0), overweight (25.0–29.9), and obese (≥ 30.0) [ 1 ]. Percent weight loss was defined as the difference between pretreatment body weight and current body weight [pretreatment body weight minus current body weight divided by pretreatment body weight times 100]. A 10% weight loss was used in the analyses as this is considered a reasonable individual initial weight loss goal by current clinical guidelines [ 1 ]. We used a 6-item question screener to assess the usual consumption of five or more fruits and vegetables per day [ 13 ]. Respondent were asked, 1) "How often do you drink fruit juices such as orange, grapefruit, or tomato? 2) Not counting juice, how often do you eat fruit? 3) How often do you eat green salad? 4) How often do you eat potatoes not including French fries, fried potatoes, or potato chips? 5) How often do you eat carrots?", And 6) "Not counting carrots, potatoes, or salad, how many servings of vegetables do you usually eat?" We used the following questions to determine weight control strategy, "Are you now trying to lose weight?" Those who responded "no" were then asked, "Are you now trying to maintain your current weight, that is to keep from gaining weight?" Additional questions on dietary practices and physical activity were asked of the subset of respondents (95,100) who responded "yes" to trying to lose or maintain their body weight. These respondents were asked "Are you eating either fewer calories or less fat to lose/maintain weight?" Their responses were either: 1) "Yes, fewer calories", 2) "Yes, less fat", 3) "Yes, fewer calories and less fat", or 4) "No". They were also asked "Are you using physical activity or exercise to lose/maintain weight? Their responses were either "yes" or "no". All individuals were also asked "In the past 12 months, has a doctor, nurse, or other health professional given you advice about your weight?" Responses were "Yes, lose weight", "Yes, gain weight", "Yes, maintain current weight", "No". In an earlier section individuals were asked, "About how long has it been since you last visited a doctor for a routine checkup? The latter question was used to distinguish those who had not received advice due to not seeing a doctor for a routine checkup in the past 12 months. For respondents who were currently using prescription weight loss medication and were trying to lose weight or maintain their current weight, we attempted to determine if they were meeting the minimal lifestyle recommendations. Respondents who stated they were eating fewer calories – either as fewer calories or as both fewer calories and less fat – were classified as meeting minimal dietary recommendations. Respondents were asked about the two physical activities or exercises they engage in most often and about the frequency, duration, and distance (as appropriate) of each activity [ 14 ]. Responses were then classified as one of 56 selected activities. Moderate activity was defined as any of the 56 selected activities, and vigorous activity was defined as aerobic physical activity classified as vigorous-intensity based on estimated metabolic expenditure (MET). To classify an activity as vigorous, it must be aerobic with an assigned MET value that was at least 60% of a person's maximal cardiorespiratory capacity. To have achieved recommended levels of physical activity, a person must have reported engaging in moderate-intensity physical activity ≥ 5 times per week for ≥ 30 minutes each time, vigorous-intensity physical activity ≥ 3 times per week for ≥ 20 minutes each time, or both during the preceding month. Persons reporting some activity during the preceding month but not enough to be classified as moderate or vigorous were classified as insufficient. Persons classified as inactive reported no physical activity outside of their occupation during the preceding month. In addition to these activity categories, we calculated minutes per week of leisure-time (non-occupational) physical activity levels. We also requested demographic and socioeconomic information such as age in years, sex, race/ethnicity, marital status, educational level, and household income. Of the 149,806 respondents, we excluded those who did not report information on prescription weight loss medication use (n = 1561), had a missing current weight (n = 5229) or height (n = 1369), were pregnant (n = 1780), were missing current weight goal (trying to lose weight or trying to maintain weight) (n = 749), or were missing information on sociodemographic variables, fruit and vegetable intake, or physical activity (n = 3595). We excluded an additional 88 respondents that had a questionable reported weight, height or BMI that was outside the sex-specific reference values from the Third National Health and Nutrition Examination Survey, 1989–1994 [ 15 ]. The final study sample included a total of 135,435 respondents. We used SAS and SUDAAN for the statistical analysis to account for the complex sampling design [ 16 , 17 ]. Student's t tests were used to test between-group differences for means and chi-square tests were used to test between-group differences for proportions. We set statistical significance at P < 0.05 for all comparisons. Multivariate logistic regression analyses were carried out to examine independent predictors of our outcome measures including pharmacotherapy use, 10% weight loss, and meeting diet and physical activity recommendations. Models of pharmacotherapy use were stratified by sex and adjusted for age, race/ethnicity, BMI, education level, household income, marital status, and geographic region. Models of 10% weight loss adjusted for sex, age, race/ethnicity, BMI, education level, household income, marital status, geographic region, smoking status, usual fruit and vegetable consumption, and past month leisure-time physical activity. Models of meeting both diet and physical activity recommendations adjusted for sex, age, race/ethnicity, BMI, education level, household income, marital status, geographic region, and smoking status. Results The results of our multivariate analyses show that the reported use of prescription weight loss medications in the past 2 years was higher among women (4.0%) than men (0.9%). For women, the odds of use was 19% lower among those aged 35–54 years and 75% lower among those 55 years old than among those aged <35 years. The odds of prescription weight loss medication use was 34% lower among non-Hispanic black women than among non-Hispanic white women; and 53% and 39% lower among those women residing in the Northeast and Midwest regions of the United States, respectively, than among women residing in the West region (Table 1 ). We also found the odds of prescription weight loss medication use was 62% higher among women with some college education than among those with less than a high school education. For men, the odds of use was 62% higher in Hispanics than among non-Hispanic whites. Table 1 Prevalence of Prescription Weight Loss Medication Use in the Previous 2 Years by Characteristics. Women Men (N = 78127) (N = 57308) n % (SE) OR* (95% CI) n % (SE) OR* (95% CI) Characteristic Age group (yrs) ≥ 55 26787 1.61 (0.14) 0.25 (0.20–0.32) 16484 0.87 (0.13) 0.92 (0.58–1.47) 35–54 31256 5.54 (0.22) 0.81 (0.71–0.93) 24320 1.06 (0.09) 1.03 (0.73–1.45) 18–34 20084 4.80 (0.23) 1.00 16504 0.64 (0.10) 1.00 Race/ethnicity Other 2613 2.80 (0.72) 0.69 (0.40–1.13) 2245 0.26 (0.10) 0.41 (0.20–0.86) Hispanic 6248 5.31 (0.48) 1.12 (0.91–1.38) 4447 1.26 (0.27) 1.62 (1.02–2.57) Non-Hispanic black 7086 3.90 (0.35) 0.66 (0.54–0.80) 4027 0.58 (0.15) 0.69 (0.40–1.19) Non-Hispanic white 62180 3.93 (0.12) 1.00 46589 0.88 (0.07) 1.00 Body mass index (kg/m2) ≥ 30 14296 10.2 (0.42) 2.27 (1.98–2.60) 10566 3.11 (0.26) 6.24 (4.43–8.79) 25.0–29.9 22005 4.88 (0.23) 1.00 26036 0.51 (0.08) 1.00 <25.0 41826 1.45 (0.10) 0.24 (0.20–0.28) 20706 0.20 (0.04) 0.41 (0.24–0.69) Education level College graduate 19203 3.50 (0.23) 1.13 (0.86–1.48) 17525 0.92 (0.11) 1.00 (0.61–1.64) Some college 22179 5.24 (0.25) 1.62 (1.28–2.06) 15151 0.93 (0.13) 1.04 (0.63–1.69) High school 26614 3.83 (0.18) 1.26 (1.00–1.59) 17840 0.75 (0.10) 0.85 (0.52–1.39) <High school 10026 3.00 (0.30) 1.00 6734 0.92 (0.19) 1.00 Household income Don't know/refused 11947 2.14 (0.21) 0.88 (0.69–1.13) 6402 0.55 (0.18) 0.90 (0.42–1.95) ≥ $75,000 6859 5.13 (0.41) 2.06 (1.60–2.65) 7523 1.41 (0.19) 2.24 (1.38–3.62) $50–74,999 9044 5.24 (0.38) 1.68 (1.34–2.10) 8690 1.03 (0.15) 1.51 (0.91–2.52) $35–49,999 12108 5.19 (0.35) 1.54 (1.24–1.92) 10975 0.88 (0.16) 1.28 (0.75–2.19) $20–34,999 19766 3.91 (0.22) 1.16 (0.97–1.39) 14929 0.68 (0.10) 0.99 (0.61–1.62) <$20,000 18403 3.52 (0.24) 1.00 8789 0.72 (0.16) 1.00 Marital status† Married 39608 4.28 (0.16) 0.90 (0.79–1.04) 33318 1.00 (0.08) 1.02 (0.74–1.42) Not married 38428 3.72 (0.1) 1.00 23923 0.65 (0.09) 1.00 Geographic region South 17887 5.11 (0.34) 0.90 (0.77–1.07) 14385 0.86 (0.17) 1.3 (0.83–2.04) Northeast 26747 4.70 (0.18) 0.47 (0.37–0.60) 18007 1.11 (0.11) 0.85 (0.50–1.44) Midwest 13508 2.32 (0.22) 0.61 (0.51–0.73) 10165 0.72 (0.12) 0.71 (0.44–1.15) West 18524 3.48 (0.21) 1.00 13839 0.62 (0.08) 1.00 Models are adjusted for age, race/ethnicity, body mass index, education level, household income, marital status, and geographic region. † Marital status: respondents were classified as "not married if they were divorced, widowed, separated, never been married, member of an unmarried couple. Women who were obese had more than two times the odds of prescription weight loss medication use compared to women who were overweight and men who were obese had more than six times the odds of use compared to men who were overweight. In addition, women and men who had an annual household income of at least $75,000 had more than two times the odds of prescription weight loss medication use compared to those respondents with an annual household income <$20,000. At the time of the survey, the average percentage of body weight that was reported lost by prescription weight loss medication users was 8.1% for women and 7.0% for men. About one fifth of users gained weight (women: 19.8%, men: 16.5%), one tenth lost no weight (women: 12.9%, men: 13.4%), one-tenth lost between 1% and 5% (women: 12.7%, men 21.0%), and one fifth lost between 5% and 9% of body weight (women: 19.6%, men: 24.1%), data not shown. Additionally among users, one-third of women (35.0%) and one quarter of men (25.0%) reported a 10% weight loss, Table 2 . The odds of having a 10% weight loss was 37% less among men than women and 43% less among those 55 years old and older than among those 18–34 years. Those who were obese at pretreatment had more than twice the odds of having a 10% weight loss compared to those who were overweight. The odds of having a 10% weight loss was 39% higher among those who were current smokers compared to non-smokers. Those who met minimal recommendations for leisure-time physical activity had more than two twice the odds compared to those who were inactive. The odds of a 10% weight loss was 64% higher among those who usually consumed 5 or more fruits and vegetables daily than among those who usually consumed less than 3 fruits and vegetables a day. Inclusion of weekly minutes of physical activity instead of the meeting physical activity recommendations variable into the 10% weight loss final model resulted in significant odds ratios for individuals with 150–<320 minutes odds ratio (OR) 1.50 (95% confidence interval (C.I.) 1.09–2.17), 320–<420 minutes OR 2.10 (95% C.I. 1.33–3.31), and OR 1.61 (95% C.I. 0.98–2.64) compared to those individuals reporting less than 150 minutes. Table 2 Prevalence of 10% Weight Loss Among Past 2 Year Users of Prescription Weight Loss Medication. Characteristic n % (SE) OR 95% CI Sex Men 497 25.0 (2.8) 0.63 (0.44–0.92) Women 3012 35.0 (1.4) 1.00 Age group (yrs) ≥ 55 506 23.1 (2.9) 0.57 (0.38–0.85) 35–54 1927 34.0 (1.8) 0.95 (0.73–1.24) 18–34 1076 36.7 (2.3) 1.00 Race/ethnicity Hispanic/other 350 30.4 (7.4) 0.77 (0.53–1.14) Non-Hispanic black 258 35.2 (4.6) 0.76 (0.50–1.15) Non-Hispanic white 2404 35.6 (1.6) 1.00 Pretreatment body mass index (kg/m2) ≥ 30 1799 45.5 (2.0) 2.24 (1.17–2.94) 25.0–29.9 269 27.2 (2.3) 1.00 <25.0 292 7.1 (1.7) 0.22 (0.13–0.38) Education level College graduate 675 33.1 (2.9) 1.00 (0.62–1.61) Some college/tech 1106 37.2 (2.5) 1.16 (0.74–1.82) High school 993 35.9 (2.4) 1.35 (0.89–2.06) <High school 237 27.5 (4.4) 1.00 Household income Don't know/refused 231 32.8 (5.1) 0.80 (0.46–1.37) ≥ $75,000 343 33.0 (3.9) 0.66 (0.41–1.06) $50–74,999 467 33.4 (3.4) 0.65 (0.41–1.01) $35–49,999 625 32.2 (3.1) 0.66 (0.44–1.00) $20–34,999 800 36.5 (2.9) 0.82 (0.57–1.19) <$20,000 564 39.9 (3.6) 1.00 Marital status Married 1696 33.4 (1.8) 0.92 (0.70–1.21) Not married 1314 37.4 (2.4) 1.00 Geographic region South 1184 32.9 (1.8) 1.01 (0.74–1.39) Northeast 243 30.9 (4.8) 0.91 (0.70–1.21) Midwest 653 40.9 (3.2) 1.13 (0.79–1.64) West 884 34.9 (3.3) 1.00 Smoking status Current smoker 695 39.8 (3.2) 1.39 (1.04–1.86) Former smoker 709 30.1 (2.8) 0.87 (0.65–1.17) Never smoker 1608 34.7 (1.9) 1.00 Usual fruit and vegetable consumption ≥ 5 day 693 29.1 (3.1) 1.64 (1.19–2.27) 3–4 day 1205 35.4 (2.3) 1.16 (0.90–1.51) <3 day 1114 30.2 (4.4) 1.00 Past month leisure-time physical activity† Meeting recommendations 1331 42.9 (2.2) 2.19 (1.60–2.98) Insufficient 834 30.3 (2.5) 1.16 (0.83–1.60) Inactive 847 26.6 (2.7) 1.00 Models are adjusted for sex, age, race/ethnicity, body mass index, education level, household income, marital status, geographic region, smoking status, usual fruit and vegetable consumption, and leisure-time physical activity in the past month. † To have achieved recommended levels of physical activity, a person must have reported engaging in moderate-intensity physical activity ≥ 5 times per week for ≥ 30 minutes each time, vigorous-intensity physical activity ≥ 3 times per week for ≥ 20 minutes each time, or both during the preceding month [14]. We also assessed professional advice, diet, and physical activity among current prescription weight loss medication users by current weight intention (Table 3 ). Although most of the respondents who reported current medication use were trying to lose weight, 9% reported they were trying to maintain weight. Examination of past year health care professional advice about weight among those who had seen a doctor for a routine checkup showed that almost three quarters of those trying to lose weight had been advised to lose weight. However, almost one-quarter of current prescription weight loss medication users who were trying to lose weight were given no advice about their weight. Over half of users who were trying to maintain their weight were given no advice about their weight. Assessment of usual fruit and vegetable consumption showed that about one quarter of those trying to lose weight (28.3%) and one quarter of those trying to maintain their weight (27.6%) reported consuming 5 fruits and vegetables a day. Almost two thirds of current prescription weight loss medication users who reported they were trying to lose weight reported reducing caloric consumption (62.0%), 70.9% reported increasing physical activity, and 47.3% met physical activity recommendations. However, these behaviors were reported less often in current prescription weight loss medication users who reported they were trying to maintain their weight, 39.6%, 49.6%, and 29.9%, respectively. A similar proportion of weekly minute categories were observed among those trying to lose weight and those trying to maintain weight, with 8.0% and 9.7% respectively, reporting at least 420 minutes per week. Table 3 Prevalence of Professional Advice, Diet, and Physical Activity Behaviors Among Current Users of Prescription Weight Loss Medication. Current Weight Management Intention Lose Weight Maintain Weight Total (n = 563) (n = 59) (n = 622) % (SE) % (SE) % (SE) Health professional past year weight advice Lose weight 73.9 (3.6) 35.6 (10.7) 71.0 (3.4) Maintain weight 1.9 (0.6) 7.2 (5.2) 2.5 (0.7) Gain weight 0.1 (0.1) 0.0 (0.0) 0.3 (0.2) No advice 24.1 (3.5) 57.2 (10.6) 26.3 (3.3) Diet Fewer calories for weight control 62.0 (3.2) 39.6 (7.5) 60.2 (3.0) ≥ 5 fruits and vegetables daily 28.3 (3.4) 27.6 (7.7) 28.2 (3.1) Physical activity Increased physical activity for weight control 70.9 (3.4) 49.6 (9.3) 69.1 (3.2) Past month Inactive 25.1 (3.5) 47.9 (9.5) 27.0 (3.3) Insufficient 27.6 (3.3) 22.2 (7.1) 27.2 (3.1) Meeting recommendations** 47.3 (3.8) 29.9 (7.7) 45.8 (3.5) Past month (weekly minutes) <150 minutes 53.0 (4.1) 58.4 (11.1) 52.1 (3.9 150–<320 minutes 27.9 (3.8) 20.2 (8.0) 27.5 (3.5) 320–<420 minutes 11.1 (2.8) 11.7 (5.6) 11.1 (2.6) ≥ 420 minutes 8.0 (1.9) 9.7 (7.1) 9.4 (1.9) Among those who reported that had visited a doctor for a routine checkup in the past 12 months (n = 514). †To have achieved recommended levels of physical activity, a person must have reported engaging in moderate-intensity physical activity ≥ 5 times per week for ≥ 30 minutes each time, vigorous-intensity physical activity ≥ 3 times per week for ≥ 20 minutes each time, or both during the preceding month [14]. Only one-quarter (26.7%) of current prescription weight loss medication users met both lifestyle recommendations – eating fewer calories and attaining recommended physical activity levels. Our results indicate that among current users, the odds of meeting both recommendations was 62% lower among men than among women and 56% lower among other race/ethnic groups than among non-Hispanic whites (Table 4 ). We also found that the odds of meeting both recommendations was more than three time higher among those with a household income of at least $75,000 (OR 3.23) compared to those with a household income <$20,000 and those who usually consumed 5 or more fruits and vegetables a day had twice the odds compared to < 3 fruits and vegetables a day. In a sensitivity analysis, we found that adding to this model past year health professional advice about weight among those who had a routine checkup (n = 559) did not appreciably change the results and that advice to lose weight was not an independent predictor of meeting both recommendations (OR = 1.37, 95% C.I. 0.73–2.56). Table 4 Meeting diet (fewer calories) and Physical Activity Recommendations* Among Current Users of Prescription Weight Loss Medication. Characteristic n Total (SE) OR† 95% CI Sex Men 98 14.8 (4.4) 0.38 (0.17–0.84) Women 524 29.7 (3.3) 1.00 Age group (yr) ≥ 55 91 21.6 (6.9) 0.92 (0.34–2.46) 35–54 338 29.5 (4.1) 1.10 (0.58–2.11) 18–34 193 25.7 (4.1) 1.00 Race/ethnicity Other 154 16.9 (4.8) 0.44 (0.20–0.96) Non-Hispanic white 468 32.1 (3.4) 1.00 Pretreatment body mass index (kg/m2) ≥ 30.0 403 25.7 (3.1) 1.02 (0.54–1.91) <30.0 187 28.8 (6.0) 1.00 Education level College graduate 147 29.8 (6.4) 1.78 (0.76–4.15) Some college/tech 201 34.7 (5.1) 1.97 (1.00–3.90) <High school 274 19.7 (4.0) 1.00 Household income ≥ $75,000 91 39.0 (7.4) 3.23 (1.18–8.83) $50–74,999 80 29.0 (2.9) 2.49 (0.85–7.25) $35–49,999 140 32.4 (5.9) 2.45 (0.82–7.35) $20–34,999 157 27.1 (4.7) 2.14 (0.89–5.14) <$20,000‡ 154 14.9 (3.9) 1.00 Marital status Married 370 28.0 (4.9) 0.87 (0.47–1.59) Not married 252 24.6 (4.0) 1.00 Geographic region South 241 27.6 (4.0) 0.78 (0.34–1.78) Midwest/Northeast 189 26.1 (4.3) 1.03 (0.45–2.32) West 159 26.5 (6.3) 1.00 Usual fruit and vegetable consumption ≥ 5 day 158 34.7 (6.4) 2.30 (1.05–5.03) 3–4 day 227 26.3 (4.1) 1.30 (0.67–2.53) <3 day 237 21.3 (3.8) 1.00 Smoking Current smoker 149 31.5 (6.6) 1.78 (0.94–3.36) Former smoker 146 26.4 (5.5) 1.10 (0.54–2.26) Never smoker 327 24.3 (3.3) 1.00 *To have achieved recommended levels of physical activity, a person must have reported engaging in moderate-intensity physical activity ≥ 5 times per week for ≥ 30 minutes each time, vigorous-intensity physical activity ≥ 3 times per week for ≥ 20 minutes each time, or both during the preceding month [14]. † Models are adjusted for sex, age, race/ethnicity, body mass index, education level, household income, marital status, geographic region, smoking status, and usual fruit and vegetable consumption. ‡ Includes "Don't know" and "Refused" annual household income responses. In a secondary analyses we assessed prevalence of a 10% weight loss among current prescription weight loss medication users who were trying to lose weight (not maintain) by whether or not recommendations were met for eating fewer calories and physical activity (n = 237). 39.6% (SE 5.3) of individuals who reported both eating fewer calories and who met physical activity recommendations had lost at least 10% of their pretreatment weight; 36.5% (SE 6.4) who reported eating fewer calories but who did not meet the physical activity recommendations had lost at least 10% of their pretreatment weight; 54.3% (SE 9.8) who reported not eating fewer calories but met physical activity recommendations had lost at least 10% of their pretreatment weight; and 25.5% (SE 5.9) who did not meet either lifestyle recommendation had lost at least 10% of their pretreatment weight, (chi-square p = 0.03). Discussion Although in 1998 clinical guidelines recommended that prescription medication for obesity be used in conjunction with lifestyle modifications [ 1 ], we found that during that same year only one quarter of current prescription weight loss medication users reported minimal diet (eating fewer calories) and physical activity behaviors recommended for weight loss. Our analyses indicated that meeting these recommendations was low across all sociodemographic groups (<40%) and, men were less likely to meet recommendations compared to women. Although patterns of prescription medication use differed by sex, for both men and women the odds of use was higher among those who were obese compared to those who were overweight although at different magnitudes. Our data show than more overweight women used prescription weight loss medications (4.88%) than did obese men (3.11%). It is possible that pharmacotherapy prescribing patterns differ by sex since more women than men attempt weight loss [ 10 ] or that for men, a higher BMI is needed before personal action regarding weight is taken. Eating fruits and vegetables and taking part in recommended levels of physical activity through leisure-time activities were more frequent behaviors among prescription weight loss medications users who had lost at least 10% of their body weight. Although we had no direct measurements of caloric intake, fruits and vegetables in their natural state are low in energy density. Thus, adding or substituting fruits and vegetables for energy-dense snacks and sweets can impact satiety and weight gain [ 2 , 18 , 19 ]. Our analyses also indicate that those respondents who met the minimal recommendations for leisure-time physical activity had twice the odds of having lost 10% of their body weight compared to those who were sedentary. We observed higher odds of 10% weight loss for those with at least 150 weekly minutes, but the magnitude did not appear to have a dose effect. We believe additional research is needed to determine whether physical activity modifications require greater duration, such as exercising 60 minutes or more daily to prevent weight regain as has been suggested by the Institute of Medicine [ 20 ], to improve the efficacy of currently approved obesity pharmacotherapy for weight management. About 40% of current prescription weight loss medications users who both reduced calories and met leisure-time physical activity recommendations had lost 10% of their body weight. Our data also show that respondents who consumed usually less than 5 fruits and vegetables per day were less likely to meet both diet and physical activity recommendations than did those who consumed 5 or more fruits and vegetables a day. Because fewer than half of physicians counsel obese patients about weight control [ 21 , 22 ] we believe the opportunity exists for physicians and other health care professionals to provide counseling that emphasizes appropriate weight control practices that include regular physical activity and a balanced low-calorie diet which includes increasing consumption of fruits and vegetables. The current study had several limitations. Although the participants self-reported pre-medication body weight and current weight, we did not ask for the date prescription weight loss medication use began and it could not be established whether weight loss was solely attributable to pharmacotherapy usage. We were also unable to determine if respondents had lost a substantial amount of weight but then regained the weight (weight history). Self-reported weights and weight history may also be a concern because individuals, especially those who are overweight, may underreport their weight [ 23 ]. We lacked dietary assessment measures to quantify the number of calories consumed. Thus, we could not determine whether individuals had actually reduced their caloric consumption to levels that would lead to substantial weight loss. Research suggests that some obese subjects underestimate their caloric intake and overestimate their physical activity [ 24 ]. In addition, we lacked data on whether respondents were counseled specifically about physical activity and/or nutrition for weight management or had undergone behavioral therapy. Although we found that about one fifth of participants had lost between 5% and 9% of their pretreatment body weight and one third of participants had lost at least 10% of their pretreatment body weight, the pill dosage and length of medication use were not collected. Thus, some users may have used pills for a short time, such as less than 6 months, and would not be expected to have lost 5% to 10% of their body weight. In current practice, physicians treating obesity can prescribe phentermine, sibutramine, and orlistat (available April 1999). Analysis of national data on patient visits found that phentermine was the most common antiobesity medication in 2001 and in early 2002 (based on January through March figures) [ 25 ], suggesting that our data, although collected in 1998, has relevance to current weight management. Most nonsurgical obesity treatments lead to weight loss for the first 6 months followed by regain [ 26 ]. Thus, some health care professionals suggest that prescription weight loss medication be prescribed either to enhance weight loss during the active weight-loss phase or to prevent later regain [ 26 , 27 ]. In our study, 9% of current medication users reported that they were currently trying to maintain their weight. We observed that the behaviors of these individuals differed from those who said they were currently trying to lose weight, e.g. fewer reported eating less calories, fewer met physical-activity recommendations, and fewer were advised by a health care professional to either lose or maintain weight. Additional research is needed on use of pharmacotherapy for weight maintenance. Conclusions We found that one in ten obese women and one in 33 obese men reported using a prescription weight loss medication in the past two years. Among medication users, one-third of women and one quarter of men reported a 10% weight loss. Only one quarter of current medication users reported the minimal diet and physical activity behaviors recommended for weight loss. Ongoing population-based surveillance is needed to assess the prevalence of prescription weight loss medication use and whether pharmacotherapy users are making appropriate lifestyle changes in order to lose or maintain weight. List of abbreviations BRFSS: Behavioral Risk Factor Surveillance System BMI: body mass index OR: odds ratio 95% C.I.: 95% confidence interval Competing interests The author(s) declare that they have no competing interests. The questions that were asked of all authors: Have you received reimbursements, fees, funding, or salary from an organization that may in any way gain or lose financially from the publication of this paper in the past five years? No. Have you held any stocks or shares in an organization that may in any way gain or lose financially from the publication of this paper? No. Do you have any other financial competing interests? No. Are there any non-financial competing interests you would like to declare in relation to this paper? No. Author contributions HB performed the statistical analysis and drafted the manuscript. LKK conceived the study questions, provided statistical input, and edited the manuscript. MKS conceived the study questions, provided statistical input, and edited the manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546419.xml
547908	Clinical and cost effectiveness of mechanical support for severe ankle sprains: design of a randomised controlled trial in the emergency department [ISRCTN 37807450]	Background The optimal management for severe sprains (Grades II and III) of the lateral ligament complex of the ankle is unclear. The aims of this randomised controlled trial are to estimate (1) the clinical effectiveness of three methods of providing mechanical support to the ankle (below knee cast, Aircast ® brace and Bledsoe ® boot) in comparison to Tubigrip ® , and (2) to compare the cost of each strategy, including subsequent health care costs. Methods/design Six hundred and fifty people with a diagnosis of severe sprain are being identified through emergency departments. The study has been designed to complement routine practice in the emergency setting. Outcomes are recovery of mobility (primary outcome) and usual activity, residual symptoms and need for further medical, rehabilitation or surgical treatment. Parallel economic and qualitative studies are being conducted to aid interpretation of the results and to evaluate the cost-effectiveness of the interventions. Discussion This paper highlights the design, methods and operational aspects of a clinical trial of acute injury management in the emergency department.	Background Sprains of the lateral ligaments of the ankle joint account for between 3 and 5% of all emergency department (ED) attendances in the UK [ 1 ], with approximately 5600 injuries each day [ 2 ]. The injury is painful and incapacitating, and, for all but the most minor injuries, weight bearing is difficult to tolerate. Activities of daily living can be significantly compromised in the first few weeks, and although acute symptoms resolve, persistent symptoms are reported to occur in between 30 to 50% of people [ 3 - 6 ]. Persistent symptoms include recurrent sprains, instability, swelling, unsightly appearance and pain. Crichton has developed a classification of severity of ankle injuries [ 7 ]. • GRADE I – the ligament is stretched but not torn and the anterior talofibular ligament is usually involved. The anterior draw test is negative. • GRADE II – the ligaments are partially torn; laxity may be present and there is moderate swelling. • GRADE III – complete rupture of the ligament resulting in joint instability. The anterior draw test is positive. The focus of this trial is Grade II and III sprains (referred to as severe sprains from this point). Severe sprains have poorer prognosis, and necessitate more intensive management than Grade I sprains (minor sprains). Minor sprains are considered to be self-limiting and require minimal treatment. Recent systematic reviews have highlighted the lack of high quality evidence to support clinicians in managing severe sprains [ 2 , 7 ]. There are few reliable studies describing long-term outcome. Most trials required x-ray evidence of talar tilt or an arthrogram for inclusion and are, therefore, not generalisable to clinical practice in the ED. Current practice ranges from no intervention, physiotherapy, and different types of mechanical support to surgical repair of the ligaments. A recent UK survey of 83 EDs demonstrated the most popular treatments were ice, elevation, Tubigrip ® support stocking and exercise, each of which was reported as being used in most cases by over 70% of respondents [ 8 ]. In addition, over half of responding departments reported that crutches, early weight bearing, and non-steroidal anti-inflammatory drugs were used in most cases. Follow up was used only in selected cases. Practice has been gradually changing from immobilisation of the injured joint to early mobilisation. To aid mobilisation the use of mechanical supports has been suggested. These supports vary in the amount of ankle movement they allow, but all encourage ankle flexion/extension and aim to minimise inversion/eversion that theoretically reduces the risk of further ligament injury. Movement is also proposed as a way of retraining the ankle proprioception which may reduce recurrent injury rate [ 9 ]. Other hypothesised benefits of mechanical supports are early restoration of functional mobility, rapid return to usual activities, reduction of pain because the joint is stabilised and protected, and improved quality of life. However, there is also the possibility of discomfort and inconvenience owing to restriction of joint range, and delayed healing, skin and circulatory problems. The present study has two aims 1. To estimate the clinical effectiveness of three different methods of mechanical support (below knee cast, Aircast ® ankle support and Bledsoe ® boot) in comparison to Tubigrip ® in a. The recovery of mobility (primary outcome) b. The recovery of usual occupation c. Avoidance of residual symptoms including recurrent instability, lasting limitation of physical activity, and need for further medical, rehabilitation or surgical treatment. 2. To measure the cost of each strategy, including treatment and subsequent health care costs. The NHS National Co-ordinating Centre for Health Technology Assessment has guided the selection of treatments. Tubigrip ® has been chosen as the reference treatment because it is the cheapest and one of the most commonly used [ 10 ]. The Bledsoe ® boot is lightweight and incorporates a novel design to promote ease of walking. However it is considerably more expensive, and it's clinical and cost effectiveness is yet to be proven. The below knee cast is Scotch ® Cast is commonly used for casting in the NHS [ 9 ]. There is a range of lightweight mechanical supports available, and we have selected the Aircast ® Brace. Methods This pragmatic randomised controlled trial is being run in 6 trusts (covering eight hospitals) across the UK. The trusts are: Birmingham Heartlands and Solihull NHS Trust, North Bristol NHS Trust, Oxford Radcliffe Hospitals NHS Trust, South Warwickshire General Hospitals NHS Trust, University Hospitals Coventry and Warwickshire NHS Trust and the Worcestershire Acute Hospitals NHS Trust. Figure 1 provides an overview of the trial method and patient journey. Multi-centre Research Ethics Committee approval has been awarded by the Northern and Yorkshire Regional Office and local ethical and research governance approvals have been obtained. Informed consent to participate in the trial and allowing researchers to access hospital records is obtained from all participants. Figure 1 Study design Study population The target population is people attending the ED with a severe sprain of the lateral ligament complex of the ankle. Inclusion criteria People who attend with sprain of the ankle who are unable to weight bear, aged 16 years and older, and give informed consent. Weight bearing is used as a proxy for severe sprains as clinical grading is not possible in the acute phase [ 7 ]. Flake fractures (<=2 mm) of the lateral malleolus are included as they are normally treated as soft tissue injuries [ 11 ]. Exclusion criteria Age less than 16 years old, ankle or other fracture sustained in addition to the sprain (such as to the wrist, head etc). An x-ray is used to exclude fracture where this is clinically indicated and as guided by the Ottawa guidelines [ 12 ]. People who can weight bear and those with fractures are ineligible for the trial, and are managed in accordance with normal local practices. Age is used as an exclusion criterion because of the complications involved in the management of epiphyseal injuries (growth plate injuries). Growth plate injuries would not normally be managed using the treatments being tested [ 13 ]. Patients are also excluded if they have a contra-indication to any of the four arms of the trial. This is most likely to occur if a patient has a DVT, high risk of DVT or other circulatory disturbance. Other contra-indications include poor skin viability preventing splinting or casting. The decision to exclude on these criteria is made by the attending clinician. The process of identifying participants A standard approach to identifying potential participants has been instituted across all participating centres. The approach was designed in consultation with the departments to dovetail with current procedures, and eradicate duplication of clinical and research data collection. The attending clinician (nurse practitioner, physiotherapist or doctor) assesses weight-bearing status. These data are recorded on a form that includes details of the remainder of the clinical examination, including if indicated, x-ray results. At this stage, people able to weight bear are excluded. The attending clinician gives a brief explanation of the trial, an information pack, and arranges an appointment at a follow-up clinic two to three days later. Delay is an accepted practice in the application of mechanical supports, allowing for the resolution of acute swelling. It also filters out patients in whom the diagnosis of severe sprain at initial contact was incorrect, and gives participants sufficient time to consider whether they wish to participate in the trial. A physiotherapist who staffs the follow-up clinic is responsible for checking eligibility and providing a further opportunity for patients to ask questions. If appropriate, the physiotherapist recruits the patients, contacts the randomisation centre, and arranges application of the treatments. Reasons for declining to participate in the trial are recorded. Assessing the generalisability of the results Data on all ankle sprains attending the participating EDs during the recruitment period are collected using a standardised proforma. A copy of this proforma, which is entered into the medical record, is anonymised and passed to the research team. It provides descriptive data on the injury pattern, occupation, age, and sex of patients, including those who are ineligible or decline trial participation at various time points. Completed proformas are audited against ED attendance records to check whether attending clinicians have referred appropriate patients to the follow-up trial clinics. This allows clinicians who are unfamiliar with the trial to be identified and approached individually to be briefed on trial procedures. Treatment allocation Randomisation is stratified by centre and provided via a phone-in service, which utilises a computer generated random allocation. Allocation concealment, which shields people who enter patients into a trial from knowing future allocations, is thereby ensured. The randomisation service is independently administered and quality controlled by the Birmingham Cancer Trials Unit. Treatment protocol The mechanical supports are fitted to each individual by a trained health professional (physiotherapist, nurse or plaster technician) in the follow-up clinic, to ensure comfort and correct fit. Participants are provided with standardised written and verbal instructions including when to remove the support, encouragement of normal walking with limits of tolerance, simple exercise advice, what to do in the event of experiencing difficulties with the support, and washing instructions. The protocols for duration of support, weight-bearing status, and activity have been determined by the manufacturers recommendations and the results of a national survey of practice completed in the planning phase of the trial [ 8 ]. Treatments are applied within three days of the injury, and within an hour of randomisation. Other treatments All other treatments are standardised and include the provision of crutches, advice to elevate and pain relieving medications if needed. Withdrawal of the latter treatments would be inappropriate as they constitute normal and accepted care. Physiotherapy is not provided as part of the trial treatment protocol, and is not part of routine care provided in the participating centres [ 9 ]. If the participant receives these types of treatment in addition to the trial protocol, this is recorded as an outcome. Baseline and outcome measures Clinical status is measured at baseline, 4 weeks, 12 weeks and 9 months. The baseline assessment also includes date of birth, sex, body mass index, ethnicity, an assessment of pre-injury abilities including usual levels of mobility, engagement in sporting activity, usual occupation and employment (including hours worked and type of work undertaken) and completion of the baseline version of the outcome questionnaires. The selection of outcome measures is based on a systematic review completed during the preparatory stage [ 14 ], and are shown in Table 1 . There is a paucity of information on the psychometric properties of self-completed questionnaires for ankle conditions. The Foot and Ankle Outcome Score (FAOS) is a questionnaire that ascertains functional limitations (including mobility) and the severity of other symptoms after ligament sprains, and has evidence of validity and reliability [ 10 ]. Previous versions have been validated against objective tests of ankle function [ 15 ]. The Functional Limitations Profile (FLP) is the British version [ 16 ] of the Sickness Impact Profile [ 17 ]. We are using the FLP occupation and mobility sub-scales to provide more detailed information on the impact of the injuries and treatments including adaptations that occur after the injury. The FLP has not been used in studies of ankle injuries previously. Health related quality of life will be measured using the SF-12 version 1 and EQ-5D [ 18 ]. Return to normal occupation and leisure activities will be recorded as the date that people return to work and normal activities. Patients are asked to make a record of significant dates on a calendar to aid recall. The effectiveness of these calendars is being tested in a separate study. Table 1 Outcome measures Outcome Outcome measure Functional limitations and severity of symptoms after ligament sprains Foot and Ankle Outcome Score (FAOS) Impact of injuries and treatments including adaptations that occur after injury Functional Limitations Profile (FLP) occupation and mobility sub scales Health related quality of life Short Form 12 version 1 (SF-12 v1) Health related quality of life EuroQol (EQ-5D) Health economics Resource Use Questionnaire Return to normal occupation and leisure activities Date of return to work and normal activities recorded on a trial calendar Timing of follow-up The mechanical supports are likely to have maximal impact and benefit during the first three months of recovery, and this time period is defined as the primary time point. We anticipate that some participants will still be wearing the mechanical support at 4 weeks. The natural time course of recovery of ankle sprains is for functional limitations to stabilise between 3 and 9 months and it is expected that the difference between treatments will narrow in the longer term as the majority of people will recover [ 3 , 5 , 6 ]. Participants will be followed to 9 months to ensure that there are no longer-term complications from injury or their treatments. Follow-up and masking Follow-up data are collected by postal questionnaire. We have implemented an intensive approach to follow-up. Participants are mailed a follow-up questionnaire, and if a response is not received within one week, they are telephoned to ensure they have received the questionnaire. A second questionnaire is mailed if necessary. If there is still no response, the participant is contacted one week later, and a core set of outcomes are collected over the telephone. All follow-up data are collected and analysed by individuals who are independent of the recruitment, randomisation, baseline assessment and are masked from the treatment provision. This level of masking will be maintained until final analysis of the data has been completed. The only exception to this rule will be if the data monitoring committee require unmasked data, and in this circumstance, only the independent Chairperson will be aware of assignments. Quality assurance Trial procedures are audited at regular intervals to ensure compliance with the protocol. Maintaining trial profile and competence in trial procedures is challenging in the ED, because of a high turn-over of staff and a diversity in the staff groups dealing with injuries in different hospitals. To combat these difficulties, training is provided on cycles that coincide with staff rotation, as well as various other time points (e.g. in-service training). Data analyses The analysis will be conducted as intention to treat. An analysis of all people who completed the trial will be undertaken, and in addition, a sensitivity analysis will be undertaken to assess the range of potential biases that could result from loss to follow up or withdrawal. Numerical and graphical summaries of all the data will be compiled, including a detailed description of missing data at the clinic visit, questionnaire and individual level. Logistic and log-linear multinomial regression models will be used to provide estimates of the recovery rates and the prevalence of residual symptoms, with confidence intervals. As there are multiple centres and repeat assessment, random effect (or hierarchical) models will be used to investigate the components of variation. The common current and cheapest therapy is Tubigrip ® , so each of the three other treatments will first be compared with it. Any treatments found to be more effective than Tubigrip ® will be compared with each other. Regression modelling will allow an assessment of factors that might indicate the appropriate choice of treatment. As these analyses are pre-specified, issues of multiple comparisons are minimal. The sensitivity of the above analyses to missing data at various levels will be assessed and quantified using modern statistical methods for incomplete multivariate data [ 19 ]. Economic analysis Severe sprains may have a range of direct cost consequences across primary and secondary health care, they may also have cost consequences for patients themselves in terms of their personal expenditure and return to work. The costing study will seek to adopt a broader societal perspective, including patient costs, when estimating differences in the cost of resources used in the four arms of the trial. The economic analysis will compare resource use (costs) with any measurable changes in health outcomes (benefits). The health outcomes of the four technologies will be measured in terms of changes in specific disease parameters and validated measures of health-related quality of life (HRQL). Any uncertainties in the cost and outcomes data will be incorporated into a sensitivity analysis. The cost of each mechanical support protocol will be determined through observation and will include staff time, overheads, equipment and transport; plus follow-up visits to hospital, GP surgeries, physiotherapists or others. The consequences for patients in terms of GP and hospital visits, including travel expenses and time off work, and expenditure on aids or private practitioner input are obtained from structured resource use questions added to patient follow-up questionnaires. Further treatments recorded include pain relieving medications, anti-inflammatory and other topical agents, bandages, supports or footwear. Patients are also asked to distinguish whether these are NHS-funded or private treatment paid for by the individual or private provider. Patient self-reported information on service use has been shown to be accurate in terms of intensity of use of different services [ 20 ]. Hospital notes and records will also be audited for information on service use. The cost of primary and hospital services will be estimated from a variety of sources, including the finance departments of the trial hospitals and services concerned and national sources [ 21 , 22 ]. Differences in resource use and outcomes are ascertained in follow-up questionnaires at 12 weeks and 9 months. The appropriate technique of economic evaluation will depend on the results of the study [ 23 ]. The simplest eventuality would be where the least expensive intervention is found to be better on at least one outcome measure and no worse on any other i.e. dominant. Another is where two interventions have the same outcomes in which case cost-minimisation analysis will be used to compare the two. However, where an intervention is clearly better in terms of outcome but is also more costly, a different approach is required. One accepted method is to compare the different interventions in terms a single outcome measure identified as clinically important for the condition being treated. The primary outcome meets this requirement. Therefore, the costs per unit improvement in mobility will be used to provide an estimate of overall cost-effectiveness; average and incremental (relative to reference treatment) cost-effectiveness ratios will be estimate for the different treatments used. However, this approach does not allow comparison with other types of intervention/condition combinations and also does not consider the value (utility) of differences in improvement in health status. In the present study, the EQ-5D will be used to generate such utility scores that can be compared to costs. A cost-utility analysis will estimate outcomes in terms of differences in quality-adjusted life years (QALYs), representing the period of life subsequent to a health care intervention weighted or adjusted for the quality of life experienced by the patient during that period, and compare these with cost for the four interventions [ 23 ]. The cost-utility analysis will present the incremental cost of the extra benefit gained both in summary form in terms of incremental cost per QALY, and also using a 'disaggregated' approach where the extra costs are presented alongside the HRQOL dimensions such as pain. Sample size Although there was a paucity of published data available to inform the sample size estimate at the outset of the trial, the preliminary phase of the study has enabled us to make a more accurate estimation of the sample size required. The presented estimate is based on a standard sample size calculation for a two-sample t-test with equal variances and a significance level of 0.05, using the variance estimated from an ANOVA of the 4-week data (n = 100). The minimal clinically important difference is set at 10 percent, which represents a small to moderate effect size. A target of 600–650 participants will allow us to detect clinically important outcomes at 4 and 12 weeks. We are powered >90% to detect differences of 10 percent in the primary outcomes, and have sufficient power to detect differences in a range of secondary outcomes at 80% power. To account for the possibility of loss to follow-up, the estimate includes an inflation of 20%. However, the trial should be able to report whether any of the treatments have a sustained negative effect on outcome at 9 months. We are taking a pragmatic approach to estimating the sample size, and the Data Monitoring Committee are reviewing assumptions underlying the calculation at 6-monthly intervals. Qualitative sub-study A purposive sample of randomised patients will be interviewed to obtain qualitative data on the patients' experiences of the various treatments in a semi-structured interview. Patients will also be asked for their views on their willingness to pay a deposit for the boots/splints to encourage their return. Up to 20 patients will be interviewed using a semi-structured format. The data will be analysed using thematic content analysis. Conclusions We have described the protocol and conduct of a large scale UK randomised controlled trial of mechanical supports for the management of acute severe ankle sprains. There has been a paucity of good quality research conducted to date, which may be partly explained by the challenges implicit to studying acute events within EDs [ 24 , 25 ]. These include a very short window of opportunity in which patients satisfy the criteria of acute injury and reliance on attending clinical staff to identify and approach for inclusion potential participants who are often in a state of discomfort. The trial is currently in the middle of the recruitment phase and is running well. Trial procedures have been well received by both patients and clinicians. Much effort has gone into maintaining the profile of the trial and disseminating trial procedures across all disciplines of ED staff (e.g. doctors, nurses, reception staff and plaster technicians). This has been paramount to the success of the trial. By referring potential patients to the follow-up trial clinics ED personnel, with no direct involvement in the trial, play a vital role in the trial process. The intense follow-up has produced good completion rates. Using a pragmatic approach to sample size estimation was useful, in particular, re-estimation after an adequate run in period, and is recommended where little prior data exists. Competing interests The author(s) declare that they have no competing interests. Authors' contributions SEL wrote the original protocol, secured funding, act as co-principal investigator and lead in the writing of this manuscript. MWC wrote the original protocol, secured funding and act as co-principal investigator. JRD and SW contributed to the development of the protocol and securing funding. JLH and AS contributed to the development of the protocol and securing funding and wrote the statistical and economic aspects of the protocol respectively. RAN is responsible for the day to day co-ordination of the trial, has contributed to refining the protocol and outcome measure package. JLM revised the protocol and the sample size calculations. MC refined the economics section of the protocol and is providing economic input. EJW is responsible for the data management and trial support. All authors have contributed to and approved the final manuscript. CAST research team Lamb SE, Cooke MW, Nakash RA, Withers EJ, Hutton JL, Marsh JL, Szczepura A, Wilson S, Clark M, Dale JR, MacNamara A, Kendall J, Skinner D, Sakr M, Kelly C, O'Byrne G, Morrell R, Dunn M, Bateman S, Kempson S, Hooker F, Parker N, Noakes E, Vaux N, Doughty G, Nichols V, Ritchie C, Sabine E. Trial Steering Committee Professor Bill Gillespie, Professor Sallie Lamb, Dr Matthew Cooke, Professor Jeremy Dale, Professor Jane Hutton, Dr Jen Marsh, Professor Ala Szczepura, Dr Sue Wilson, Rachel Nakash, Vicki Staples. Data Monitoring Committee Dr Janet Dunn, Professor Damian Griffin, Pat Overton-Brown. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547908.xml
524505	Leptin receptor in the chicken ovary: potential involvement in ovarian dysfunction of ad libitum-fed broiler breeder hens	In hens, the ovarian follicles committed to ovulation are arranged in an ordered follicular hierarchy. In standard broiler breeders hens genetically selected for high growth rate the reproductive function is clearly dysfunctional. Feed restriction is needed during reproductive development to limit the formation of excessive numbers of ovarian yellow follicles arranged in multiple hierarchies. To determine whether leptin is involved in the nutritional and reproductive interactions controlling follicular hierarchy in hens, blood leptin levels and ovarian expression of the leptin receptor mRNA were determined during follicle maturation in three chicken lines; a slow growing broiler "Label" genotype without reproductive dysfunction, a fast growing "Standard" genotype fed ad libitum or restricted and a fast growing "Experimental" line with intermediate reproductive performance levels. Whereas expression of the leptin receptor mRNA did not change in the theca, it clearly decreased with follicular differentiation in the granulosa of slow growing hens. In fast growing standard hens fed ad libitum and presenting significant reproductive dysfunction, the decrease was disrupted and dramatic up-regulation of granulosa cell expression of the leptin receptor was observed. On the other hand, feed restriction decreased the overall level of expression of the leptin receptor mRNA and restored the decrease with follicular growth. The level of expression of the leptin receptor probably modulates the action of leptin on follicular differentiation. Since blood leptin and other metabolic factors were not affected by the genotype or by nutritional state, the factors involved in the regulation of leptin receptor gene expression remain to be determined. This study demonstrates the involvement of leptin in the nutritional control of reproduction in birds. Leptin action on the ovary probably controls follicular hierarchy through the regulation of steroidogenesis.	Background The ovary of the mature hen contains a hierarchy of yellow yolky follicles and several thousand smaller follicles from which the large yolky follicles are recruited. The yellow follicles are arranged in a size hierarchy and are committed to ovulation. In each follicle, the granulosa cells are surrounded by theca tissue and are separated from it by a basement membrane. Each compartment of the largest yellow follicles (theca and granulosa cells) can be anatomically separated to follow the individual functions of these two ovarian compartments in follicle growth and differentiation. At an early stage of follicular development the small ovarian follicles produce estrogens and androgens. As follicles begin to sequester yolk their production of estrogens from theca cells decreases to become very low at ovulation. As follicles are recruited into the yolky follicular hierarchy, estrogen and androgen production by theca cells diminishes and production of progesterone by granulosa cells increases. The largest F1 yellow follicle then attains the highest progesterone production at the time of ovulation. In most domestic animals, reproductive function is considerably affected by nutrition [ 1 , 2 ]. Various hormones, including growth hormone, insulin-like growth factors (IGFs) and insulin, have been proposed as potential mediators affecting reproductive function [ 3 ]. However, the interactions between the reproductive endocrine axis and the metabolic axis have not been clearly determined. Leptin represents also a good candidate for such reproductive-metabolic interactions. Leptin, the protein hormone synthesized and secreted mainly by adipose tissue, has primarily been shown to regulate food intake and energy expenditure (for review see [ 4 - 6 ]). Recent studies have demonstrated that leptin may also be involved in the regulation of reproductive mechanisms in human and rat ovaries [ 7 - 10 ]. Exogenous leptin can rescue reproductive function in ob/ob leptin-deficient mice that are not only obese but also infertile. This leptin action is independent from weight loss since feed restriction in ob/ob female mice fails to restore fertility [ 11 ]. Leptin can also advance the onset of puberty or at least reverse the delay caused by feed restriction in rodents [ 12 , 13 ]. In chickens, leptin attenuates the negative effects of fasting on ovarian function. Injections of leptin during fasting delays cessation of egg laying, attenuates regression of yellow hierarchical follicles, alters ovarian steroidogenesis and limits apoptosis [ 14 ]. Leptin exerts its effect by binding to a receptor which belongs to the cytokine receptor super-family [ 15 ]. The chicken leptin receptor has been cloned and sequenced [ 16 , 17 ]. Its expression at the level of the ovary [ 14 , 17 ] suggests that leptin might act directly on the ovary to regulate chicken reproductive function. Standard broiler breeders have been submitted to high selection pressure on growth and feed efficiency. Male traits have been favoured resulting in poorer reproductive performances of the hens. As a result of such selection, broiler breeder hens are subject to metabolic disorders and reproductive dysfunction. Overfeeding during reproductive development is associated with the formation of excessive numbers of ovarian yellow follicles which can be arranged in multiple hierarchies, with increased production of unsettable eggs [ 18 - 21 ]. Severe feed restriction during rearing reduces the production of yellow follicles, the incidence of double ovulation and considerably improves the laying rate [ 19 ]. Up-regulation of yellow follicles has been related to excessive recruitment and rapid growth rate of follicles to maturity, especially under ad libitum feeding [ 19 ]. However, the mechanisms that regulate these processes are still not fully explained. This study investigated the potential involvement of leptin and its receptor in ovarian abnormalities observed in broiler breeder hens fed ad libitum. We aimed to determine the evolution of leptin receptor gene expression in both granulosa and theca cells from the four largest yellow follicles of 32 week-old hens during the laying period. The effects of genotype and diet on plasma leptin levels and ovarian expression of the leptin receptor gene were also quantified. For this purpose standard broiler breeder hens fed ad libitum or feed-restricted were compared to a French "Label" genotype and a dwarf "Experimental" line. Compared to fast growing standard broiler breeders, the French "Label" is a dwarf, slow-growing broiler genotype and the "Experimental" line is a dwarf genotype with a growth potential of progeny chicks close to that of the standard broiler chicks. The "Experimental" line is specifically selected for reproductive traits and viability at partial expenses of growth performances. The "Label" line tolerates ad libitum feeding of breeders and does not have reproductive problems under ad libitum feeding whereas the «Experimental» line can be fed ad libitum on a low energy diet during the growing period but presents more reproductive problems than the Label genotype but less than the standard genotype fed ad libitum. Methods Animals Three lines of broiler breeder hens supplied by Hubbard primary breeder (Chateaubourg, France) were used in this experiment. The S line is a standard fast growing broiler, the French Label (L) line is a slow growing broiler breeder strain used for the quality market and the experimental (E) line is a broiler breeder strain bearing the "dw" dwarf gene. This strain has a decreased need for rationing. S and L hens were given the same regime in accordance with Hubbard nutritional recommendations (2724 kcal/kg) in fine meal form. During the growing period (0–20 weeks), half of the S hens were feed restricted (SR) on the same diet in order to match a reference body weight curve provided par Hubbard primary breeder, the other half (SA) and all the L hens were fed ad libitum. The feed intake was equivalent to 37% of the SA group up to point of lay [ 22 ]. A special diet was designed for the E hens, consisting of a series of finely ground meal diets with a lower energy content (2550 kcal/kg). The interest of the E group is mainly practical, it represents an actual alternative for severely restricted standard broiler breeder hens [ 22 ]. Transition between the grower and breeder feed occurred at 20 weeks of age for the ad libitum-fed hens (L, SA and E). Transition occurred at the beginning of laying for the restricted hens that were then allowed ad libitum access to food. At 32 weeks of age, blood samples were collected from 12 hens of each experimental group (SA, SR, E and L) and six hens were sacrificed by an overdose of pentobarbital (Sanofi-Santé Animale, Libourne, France). The ovaries and liver were immediately removed. Granula and theca compartments from the first (F1), second (F2), third (F3) and fourth (F4) largest ovarian yellow follicles were dissected as previously described [ 23 ]. Since SA birds presented a greater average number of yellow follicles per ovary (9.36, 8,00, 7.42, and 6.33 yellow follicles/ovary for the SA, SR, E and L hens respectively) and a higher proportion of pairs of yellow ovarian follicles undergoing simultaneous development [ 22 ], follicles were assigned to the same follicular rank when the difference of weight between two follicles was of 0.4 g or less. In that case, one follicle per pair was collected and dissected. Tissues were immediately snap frozen in liquid nitrogen and stored at -80°C until used for total RNA extraction. This experiment was carried out with due regard to the legislation governing ethical treatment of animals, and investigators were certificated by the French government to carry out animal experiments. RNA extraction and leptin receptor RT-PCR Total RNA was extracted from liver, granulosa and theca cells using RNA InstaPure (Eurogentec, Angers, France) according to the manufacturer's recommendations. After DNAse treatment using Ambion's DNA-free kit (Clinisciences, Montrouge, France), 2 μg of total RNA were reverse-transcribed (RT) in a final volume of 20 μl using RNAse H - MMLV reverse transcriptase (Superscript II, Invitrogen, Cergy Pontoise, France) and random hexamer primers (Promega, Charbonnières, France). cDNA was then diluted to 1:8. For normal PCR amplification, five microliters of the RT reaction were amplified for 35 cycles in a 50 μl reaction volume containing 2.5 units of Taq DNA Polymerase (Amersham Biosciences, Orsay, France), 2.5 mM MgCl 2 , 0.2 mM dNTPs (Promega, Charbonnieres, France) and 0.2 μM of each forward and reverse primer. Leptin receptor forward (5'-GTC CAC GAG ATT CAT CCC AG-3') and reverse (5'-CCT GAG ATG CAG AGA TGC TC-3') primers were chosen according to the previously determined sequence of the chicken leptin receptor cDNA [ 17 ]. This pair of primers amplifies a 271 bp cDNA fragment located in the coding sequence of the extra-cellular domain. The amplification conditions were as follows: denaturation at 94°C for 30 sec, annealing at 60°C for 30 sec and primer extension at 72°C for 60 sec. After final extension at 72°C for 10 min, PCR products were resolved on 1.5% agarose gel containing ethidium bromide. Real-time RT-PCR Real-time RT-PCR was performed as previously described [ 24 ]. Briefly, forward leptin receptor primer 5'-GCATCTCTGCATCTCAGGAAAGA-3' and reverse leptin receptor primer 5'-GCAGGCTACAAACTAACAAATCCA-3'(nucleotides 362 to 448 of the chicken leptin receptor cDNA sequence) [ 16 , 17 ] were designed to be intron-spanning to avoid co-amplification of genomic DNA using Primer Express Software (Applied Biosystems, Courtaboeuf, France). A 20 μl master mix containing 12.5 μl SYBR Green PCR Master Mix, 1 μl forward primer (300 nM), 1 μl reverse primer (300 nM) and 5.5 μl water was prepared to perform real-time PCR (Applied Biosystems, Courtaboeuf, France). Five microliters of cDNA dilution was added to the PCR Master Mix to a final volume of 25 μl. The following PCR protocol was used on the ABI Prism 7000 apparatus (Applied Biosystems, Courtaboeuf, France): initial denaturation (10 min at 95°C), followed by a two-step amplification program (15 sec at 95°C, followed by 1 min at 60°C) repeated 40 times. Quantification was performed using ABI integrated software as previously described [ 24 ]. 18S ribosomal RNA was chosen as the reference gene. The level of 18S RNA was determined using the Pre-developed TaqMan Ribosomal RNA control kit (Applied Biosystems, Courtaboeuf, France) according to the manufacturer's recommendations. The results were expressed as the leptin receptor mRNA/18S RNA ratio. Each PCR run included a no template control and replicates of control and unknown samples. Runs were performed in triplicate. Plasma lipid, glucose and hormone analyses Total cholesterol, phospholipid, and triglyceride plasma concentrations were calculated with "Cholesterol RTU", "Phospholipides Enzymatique PAP 150", and "Triglycérides Enzymatique PAP 150" kits (bioMérieux, Charbonnières les Bains, France) according to the manufacturer's recommendations. Plasma glucose levels were measured by the glucose oxidase method using an automated analyzer. Plasma insulin levels were determined by a radioimmunoassay with a guinea pig anti-porcine insulin antibody using chicken insulin as the standard [ 25 ]. Plasma concentrations of leptin were determined by a multi-species leptin RIA kit (LINCO Research Inc, CliniSciences, Montrouge, France) according to the recommendations of the manufacturer. Statistical analysis The results of plasma lipid, glucose and hormone analyses as well as leptin receptor mRNA expression in the liver were analyzed by one-way ANOVA and means were compared by Student Newman Keuls multiple comparison test. The effects of the groups of birds (E, L SA and SR), follicular rank and possible interaction on the logarithm of leptin receptor mRNA levels were tested by two-way ANOVA using the General Linear Model (GLM) procedure of SAS (SAS Institute, 1999. SAS User' Guide, Version 8 ed. SAS Institute Inc., Cary, NC). An additional effect of the subject was introduced into the model in order to take into account the fact that measurements of leptin receptor mRNA expression for the different follicular ranks were performed on the same animal. Pairwise comparisons of means for each significant effect of the ANOVA were performed by Scheffe test with the least means squares statement of the GLM procedure. The level of significance was set at P < 0.05 . Results Leptin receptor mRNA expression in granulosa and theca cells We demonstrated the expression of leptin receptor mRNA in granulosa and theca cells from the three genotypes fed ad libitum or restricted for the S line. The expression of leptin receptor mRNA for both ovarian cells was detected in each hierarchical yellow follicle studied (F1 to F4) (Figure 1 ). Figure 1 Amplification by RT-PCR of leptin receptor mRNA (271 bp) in granulosa and theca cells of the four largest yellow follicles (F1, F2, F3 and F4) inbroiler breeder hens from a dwarf Experimental line (E), a French Label line (L) and a Standard line fed ad libitum (SA) or restricted (SR). Evolution of leptin receptor mRNA expression with follicular development The evolution of expression of leptin receptor mRNA during follicle development was investigated in both granulosa and theca cells from F4 to F1 yellow follicles using real-time RT-PCR. Since leptin receptor mRNA levels did not follow a normal distribution (skewness of 5.48 and Kurtosis of 32.88), they were log transformed. The resulting distribution was closer to the normality with skewness of 0.87 and kurtosis of 0.69. Variance analysis was performed on transformed data. As shown in Table 1 and Figure 2 , expression of leptin receptor mRNA in granulosa cells from the L genotype clearly decreased as the follicle developed. Significant statistical differences were measured between hierarchical yellow follicles at each stage. In the E line, expression of the leptin receptor mRNA decreased between F4 and F1 yellow follicles. However the high variability of the expression of leptin receptor mRNA in F4 follicles prevented the decrease from reaching statistical significance. Compared to the L line, the level of expression of the leptin receptor in the granulosa cells was lower in the E group, especially for the F4 and F3 yellow follicles but statistical significance was reached only for the F4 follicles (Table 1 ). In the S line fed ad libitum, expression of the leptin receptor in the granulosa was dramatically up-regulated. This up-regulation was clearly evident in F4 F3 and F1 follicles. Wide variability was also observed in F4 and F3 follicles. Feed restriction of the standard hens (SR) induced a general decrease in the expression of leptin receptor mRNA. The overall level of expression of the leptin receptor mRNA measured in the SR line was similar to that observed in the L and E lines. Compared to the SA birds, feed restriction has restored the decrease in the expression of leptin receptor mRNA with follicle development. Table 1 Leptin receptor mRNA levels normalized to the level of 18S rRNA (expressed as arbitrary units) in the granulosa of ovarian F1, F2, F3 and F4 yellow follicles. E L SA SR ANOVA F1 18.63 ± 5.26 a 16.34 ± 2.26 a 219.01 ± 72.23 b 24.73 ± 4.20 a P < 0.0001 F2 34.60 ± 6.18 a 32.78 ± 4.73 a 11.57 ± 4.31 a 26.89 ± 3.89 a NS F3 29.78 ± 3.43 a 92.80 ± 18.25 a 1062.90 ± 580.82 b 71.31 ± 20.92 a P < 0.001 F4 91 ± 48.85 a 290 ± 57.93 b 844 ± 420.56 b 135.22 ± 32.26 a P < 0.01 Data are expressed as mean ± SEM (n = 6). Data with different letters are significantly different between groups. Figure 2 Relative levels of leptin receptor mRNA assessed by real time RT-PCR in the granulosa cells of the four largest yellow follicles (F1, F2, F3 and F4) in broiler breeder hens from a dwarf Experimental line (E), a French Label line (L) and a Standard line fed ad libitum (SA) or restricted (SR). The results were corrected by the corresponding levels of 18S rRNA. Data are expressed as mean ± SEM, n = 6. Bars with different letters are significantly different within groups (P < 0.05). Expression of leptin receptor mRNA in the theca cells remained stable during yellow follicle development, whatever the group of birds considered. Statistical analysis did not reveal any difference in leptin receptor mRNA expression between the 4 groups of birds (Figure 3 ). Figure 3 Relative levels of leptin receptor mRNA assessed by real time RT-PCR in the theca cells of the four largest yellow follicles (F1, F2, F3 and F4) in broiler breeder hens from a dwarf Experimental line (E), a French Label line (L) and a Standard line fed ad libitum (SA) or restricted (SR). The results were corrected by the corresponding levels of 18S rRNA. Data are expressed as mean ± SEM, n = 6. Expression of leptin receptor mRNA in the liver In the liver, the expression of leptin receptor mRNA was up-regulated in S birds fed ad libitum. Feed restriction of S birds restored the level of expression of leptin receptor mRNA similar to that measured in the E and L birds (Figure 4 ). Figure 4 Relative levels of leptin receptor mRNA assessed by real time RT-PCR in the livers of broiler breeder hens from a dwarf Experimental line (E), a French Label line (L) and a Standard line fed ad libitum (SA) or restricted (SR). The results were corrected by the corresponding levels of 18S rRNA. Data are expressed as mean ± SEM, n = 6. Bars with different letters are significantly different (P < 0.05). Plasma lipid and glucose and hormone concentrations At 32 weeks of age plasma leptin and insulin concentrations were found to be similar in the three genotypes. Food restriction of the standard hens did not alter plasma leptin, glucose or insulin levels (Table 2 ). Triglyceride, cholesterol and phospholipid levels were also measured. Cholesterol and phospholipid levels were not affected by genotype or diet. On the other hand, triglyceride levels seemed to be affected by the genotype. Lower triglycerides levels were measured in Standard birds (SA, SR). However, statistical significance was reached only for the restricted SR birds. Table 2 Lipid and hormone plasma levels. E L SA SR Leptin (ng/ml) 1.68 ± 0.17 2.08 ± 0.23 2.06 ± 0.19 1.97 ± 0.21 Insulin (pmol/ml) 118.6 ± 17.9 101.1 ± 12.0 129.7 ± 29.7 128.3 ± 16.3 Glucose (mmol/l) 10.99 ± 0.39 10.60 ± 0.22 11.10 ± 0.22 10.71 ± 0.33 Triglycerides (mmol/l) 31.36 ± 1.89 33.48 ± 1.63 25.01 ± 3.41 19.15 ± 3.34* Phospholipids (mmol/l) 12.10 ± 0.29 12.75 ± 0.47 10.53 ± 1.14 11.04 ± 1.24 Cholesterol (mmol/l) 9.16 ± 0.39 9.48 ± 0.40 8.40 ± 0.63 9.07 ± 0.77 Data are expressed as mean ± SEM (n = 12). * indicates statistical difference ( P < 0.05) Discusssion Several studies conducted on theca and granulosa cells have shown that leptin may have direct negative effects on ovarian steroidogenesis in various mammalian species. Leptin inhibits insulin-induced progesterone and 17β-estradiol production by isolated bovine granulosa cells [ 26 ] and impairs the hormonally-stimulated in vitro release of 17β-estradiol by rat granulosa cells [ 27 ]. In granulosa cells from fertile women, leptin inhibits FSH and IGF-I stimulated estradiol production [ 28 , 29 ] Since leptin has a more potent inhibitory action of insulin-induced aromatase activity of granulosa cells from small than large follicles, it has been proposed that the numbers of leptin receptors in granulosa cells might decrease as follicles develop in order to make mature Graafian follicles less sensitive to the negative action of leptin [ 26 , 30 ]. As shown in this study and in previous reports [ 14 , 17 ], the leptin receptor was expressed in the hen ovary in both granulosa and theca cells, suggesting a direct action of leptin at the level of the ovary. It seemed that leptin might affect ovarian steroidogenesis in laying hens during fasting [ 14 ] but the involvement of leptin on steroidogenesis during normal follicle development remained to be determined. In this study we demonstrated that the direct action of leptin on the ovary might be modified during follicle development since the level of expression of its receptor clearly decreased during maturation of yellow follicles. This decrease was particularly evident in slow growing broiler breeder hens from the "Label" genotype and from the feed-restricted standard line. Given that fast growing chickens (ad libitum-fed standard and Experimental broiler breeder hens) have the highest reproductive problems, genetic or nutritional control of the growth rate might regulate ovarian leptin receptor gene expression and improve reproductive function. Such evolution of receptor expression in the follicular hierarchy has previously been shown for the FSH receptor. FSH-stimulated steroidogenesis declined during follicle maturation and was associated with a decrease in FSH receptor numbers [ 31 ]. Conversely, the expression of mRNA encoding the IGF-I receptor and the related efficacy of binding of IGF-I to granulosa cells increased as the follicle matured [ 32 , 33 ]. Since leptin receptor gene expression was modified during follicle development, leptin might also be involved in regulation of the follicular hierarchy and onset of preovulatory steroidogenesis, as has been proposed for gonadotrophins and growth factors including FSH and IGF-I. Unlike mammals, progesterone in chickens is synthesized and secreted mainly by granulosa cells whereas theca cells generate estradiol [ 34 ]. Progesterone produced by granulosa cells from mature follicles provides the positive feedback necessary to stimulate a preovulatory surge of LH [ 35 ]. IGF-I has been involved in the regulation of ovarian steroidogenesis in both mammals and birds. IGF-I stimulates progesterone production from avian granulosa cells [ 36 ] whereas it up-regulates estradiol from mammalian granulosa cells [ 27 - 29 ]. Since leptin is considered to be an inhibitor of insulin and IGF-I action on steroidogenesis in mammals, leptin might have similar negative action in birds. Thus, the decrease in its receptor in the granulosa suggests that the inhibiting action of leptin would decrease during follicle development and consequently favours the stimulatory effect of gonadotrophins and IGF-I on follicular maturation. This hypothesis is also consistent with the weaker steroidogenic response of granulosa cell culture of ad-libitum fed standard broiler breeder hens when stimulated by IGF-I compared to granulosa cell culture from feed-restricted birds [ 37 ]. Moreover, Onagbesan et al (2004) demonstrated in an experiment similar to that performed in the present study and using the same genotypes that plasma progesterone levels were clearly affected in SA birds. They demonstrated that plasma progesterone levels remained relatively stable between 25 and 37 weeks of age in the E, L and SA birds with a significant lower level in the SA birds (2.2 ± 0.62 ng/ml for SA birds compared to 3.9 ± 0.36 and 4.2 ± 0.54 for L and E birds respectively). In restricted standard birds, plasma progesterone levels dramatically increased and reached values (3.8 ± 0.26 ng/ml) closed to that measured in the E and L lines (Onagbesan et al., 2004, data from progesterone levels were personal communication from Dr Onagbesan, Catholic University of Leuven, Belgium). The erratic pattern of oviposition in standard broiler breeder hens fed ad libitum has been previously demonstrated to be related to abnormal maturation of steroidogenesis, particularly in the two largest yellow follicles [ 35 ]. Since F2 and F1 yellow follicles presented similar endocrine profiles, the preovulatory surge of LH probably triggers ovulation of the two largest follicles [ 21 , 36 ]. In this study we have shown that ad libitum feeding of broiler breeder hens dramatically up-regulated expression of the leptin receptor in the granulosa cells of yellow follicles and changed the evolution of expression of this receptor with follicle development. These results suggest a strong action of leptin on the ovaries of ad libitum fed birds. Feed restriction reduced the level of expression of the leptin receptor and on the whole restored the evolution of expression of the receptor with follicle maturation. Since ad libitum feeding affects the hierarchical endocrine order of the follicles, as a potential inhibitor of hormonally induced avian steroidogenesis leptin represented a good candidate to explain the affects of follicular hierarchy. Up-regulation of the expression of the leptin receptor gene was also demonstrated in the liver. This up-regulation may be related to the control of lipogenesis. The liver plays a key role in lipid metabolism and lipogenesis in avian species [ 38 , 39 ] and the standard broiler breeder hens were the fattest birds of this experiment. Since expression of its receptor was dramatically up-regulated in SA hens, leptin probably played an important role in the increased number of large yellow follicles and abnormal follicle hierarchy. However the factors involved in regulation of the expression of the leptin receptor within the hen ovary remains to be determined. Among the plasma hormones and lipids analyzed in this study only triglycerides were found to be different between strains, with a lower level in the restricted standard broiler breeder hens that were also the leanest birds. Down regulation of the expression of the leptin receptor by homologous and heterologous signals have previously been demonstrated in both mammals [ 40 , 41 ] and chickens [ 24 ]. Leptin and insulin are able to down-regulate expression of the chicken leptin receptor in vitro. In the present study, plasma leptin and insulin levels were similar for each genotype and were not altered by feed restriction in the standard genotype. We have previously demonstrated that during the first 5 weeks of age, plasma leptin levels remained relatively stable in both broiler and layer chicken despite increased body weight [ 42 ]. However the absence of leptin levels differences may be related to the fact plasma leptin levels were measured at 32 weeks of age. SR birds were relaxed at the start of lay, switched to breeding feeding and allowed ad libitum access as the other groups of birds. We therefore suggested that leptin and insulin are probably not involved in the regulation of ovarian leptin receptor gene expression in ad libitum or feed-restricted standard broiler breeder hens. Evidence of the regulation of expression of the leptin receptor gene in the granulosa related to follicle maturation and nutritional state strongly suggest that leptin played an important local and sequential role in the dysfunction of the follicular hierarchy observed in standard broiler breeder hens fed ad libitum. This study suggests that the level of expression of the leptin receptor regulates the action of its ligand at the level of the ovary. This provides an interesting perspective to understanding the physiological role of leptin in the ovary.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524505.xml
539232	Optimal cDNA microarray design using expressed sequence tags for organisms with limited genomic information	Background Expression microarrays are increasingly used to characterize environmental responses and host-parasite interactions for many different organisms. Probe selection for cDNA microarrays using expressed sequence tags (ESTs) is challenging due to high sequence redundancy and potential cross-hybridization between paralogous genes. In organisms with limited genomic information, like marine organisms, this challenge is even greater due to annotation uncertainty. No general tool is available for cDNA microarray probe selection for these organisms. Therefore, the goal of the design procedure described here is to select a subset of ESTs that will minimize sequence redundancy and characterize potential cross-hybridization while providing functionally representative probes. Results Sequence similarity between ESTs, quantified by the E-value of pair-wise alignment, was used as a surrogate for expected hybridization between corresponding sequences. Using this value as a measure of dissimilarity, sequence redundancy reduction was performed by hierarchical cluster analyses. The choice of how many microarray probes to retain was made based on an index developed for this research: a sequence diversity index (SDI) within a sequence diversity plot (SDP). This index tracked the decreasing within-cluster sequence diversity as the number of clusters increased. For a given stage in the agglomeration procedure, the EST having the highest similarity to all the other sequences within each cluster, the centroid EST, was selected as a microarray probe. A small dataset of ESTs from Atlantic white shrimp ( Litopenaeus setiferus ) was used to test this algorithm so that the detailed results could be examined. The functional representative level of the selected probes was quantified using Gene Ontology (GO) annotations. Conclusions For organisms with limited genomic information, combining hierarchical clustering methods to analyze ESTs can yield an optimal cDNA microarray design. If biomarker discovery is the goal of the microarray experiments, the average linkage method is more effective, while single linkage is more suitable if identification of physiological mechanisms is more of interest. This general design procedure is not limited to designing single-species cDNA microarrays for marine organisms, and it can equally be applied to multiple-species microarrays of any organisms with limited genomic information.	Background Expression microarrays are powerful tools for human disease diagnosis, prognosis and treatment [ 1 ] offering unparalleled insight into the function of the entire genome and the dynamic interactions among genes. The ability of microarrays to identify gene expression signatures, specific subsets of genes that respond to particular stimuli, make them valuable tools for characterizing organisms' response to environmental conditions and host-parasite interactions. This method relies on organisms as sentinel markers of environmental changes. Since aquaculture marine species are easy to keep in a captive environment, they can be used as convenient sentinels by profiling their physiological responses. An efficient and economic method to quantify their physiological responses is to collect the expressed sequence tags (ESTs) with the purpose of constructing cDNA microarrays, which can be used to screen their transcriptomes. Therefore, several pilot studies have been initiated in economically important marine species to generate genomically enabled tools for the purpose of elucidating the role of biological and environmental factors in ultimately determining the difference between survival, morbidity and mortality [ 2 - 4 ]. The growing need for a marine functional genomics approach using microarrays bespeaks a general-purpose cDNA microarray probe selection tool to identify which ESTs to spot on the microarray from large collections of ESTs with unknown functions and variable redundancies. The two most widely used expression microarray systems are oligonucleotide and cDNA microarrays. Oligonucleotide microarrays are generated by chemically synthesizing short oligo probes (20–70 bp) onto the slides [ 5 ]. In contrast, cDNA microarrays are created by spotting long strands of amplified cDNA sequences (e.g., the expressed sequence tags) [ 6 ]. In this paper, the sequences spotted on the arrays are referred to as "probes." Although many algorithms have been developed for selection of oligonucleotide [ 7 - 11 ] or gene-specific probes [ 12 , 13 ], only one application was found by the authors for cDNA microarray probe selection [ 14 ]. However, this algorithm was designed specifically for organisms with extensive genomic data, not for the organisms with limited genomic information. In the absence of cDNA microarray probe selection algorithms, EST selection for spotting on microarrays has been approached using various informal methods. These methods included spotting ESTs without sequencing information, spotting only sequenced ESTs with annotations, or forcing the selection on gene-oriented clusters [ 15 ]. The choice of method typically reflects cost/benefit ratios and the stage of development of the EST collection. A comprehensive review of microarray probe selection can be found in Tomiuk and Hofmann [ 16 ]. Gene or transcript oriented clusters are generally formed by gene indexing projects, such as TIGR [ 17 , 18 ], Stack [ 19 ], or Unigene [ 20 ]. Gene indexing projects involve three general steps. First, the quality control step filters out contaminating sequences such as vector or bacterial sequences. Second, ESTs are partitioned into smaller clusters, often using the hierarchical single-linkage method with an arbitrarily chosen cut-off threshold [ 21 , 22 ]. Finally, although not all projects include a assemblage step, sequences are often assembled into contigs using existing software, such as CAP3 [ 23 ] or PHRAP [ 24 ]. In this study, we propose a probe selection procedure for cDNA microarray that tracks both sequence redundancies and functional representativeness of the selected probes in an integrated sequence diversity plot (SDP). SDP includes a sequence diversity index (SDI) to measure the sequence similarities within EST clusters quantitatively. The issue of how many probes are sufficiently representative for all collected ESTs is approached in a manner similar to the choice of dimensions to retain in principle component analysis (PCA). This approach reflects the fact that there is no definitive right answer to the question [ 25 ]; the number of "clusters" of ESTs may vary as the stringency of microarray hybridization condition changes. All collected ESTs are automatically annotated using Gene Ontology [ 26 ] terms, and then a unique probe GO index (UPGI), a functional index, was devised to access functionally how representative the selected probes are. This integrated and flexible method using SDP allows users to decide which clustering method and stringency to use when designing a cDNA microarray for organisms with limited genomic information based on their logistical constraint and experimental purposes. A small data set of ESTs was used to test this algorithm so that the detailed results of this algorithm could be examined. Results A small data set of 1047 ESTs from Atlantic white shrimp ( Litopenaeus setiferus ) from the Marine Genomics website [ 27 ] was analyzed. After pre-processing, 971 sequences longer than 100 bp were further used in the analysis (details see methods; Figure 1 ). The ESTs were progressively grouped using different hierarchical linkage methods from 1 to n (n = 971) clusters (details see methods). The sequence diversity plot (SDP) summarizes sequence properties within clusters and the functional representativeness of the selected probes using three indexes: the sequence diversity index [SDI; Eq. (1)], the contiguity index [CI; Eq. (2)], and the unique probe GO index [UPGI; Eq. (3)] (Figure 2 ). Sequence diversity index (SDI) measures within-cluster sequence dissimilarity This index is the ratio of within-cluster sequence dissimilarities to the total sequence dissimilarity when m clusters are formed ( m = 1,2,... n ): where d mi is the distance (dissimilarity), the E-value from blast result (details see methods), between the i th pair of sequences for a total k m pairs of within-cluster comparisons when m clusters are formed. D is defined as , the average distance of the total N pair-wise distances among all n sequences (where in this data set). Contiguity index (CI) measures the sequence contiguity within clusters The within-cluster sequence contiguity is evaluated using CAP3 [ 23 ], commonly used sequence assembly software (see methods). The number of putative unique genes, denoted as PG m , is the sum of the number of assembled contigs and singlets (single sequences, which cannot be assembled with any other sequences) when m clusters ( m = 1,2...n) are formed. The contiguity index (CI) at a given number of clusters ( m ) is defined as the inverse of the average number of putative genes per cluster, which equals the number of clusters per gene: This index reflects how contiguous the sequence members are within a cluster. Maximum value of CI is 1 when all the members are contiguous (one cluster per gene). Unique probe Gene Ontology (GO) index (UPGI) measures functionally how representative the selected probes are The unique probe GO index (UPGI) when m clusters of ESTs are formed is defined as the number of unique GO terms associated with all m probes (m = 1, 2...n) divided by the number of the GO terms associated with all n sequences (n = 971). where ProbeGO mj is the number of unique GO terms associated with the probe representing the j th cluster when m clusters are formed and sequenceGO i is the number of GO terms of the i th sequence. This index measures functionally how representative the selected probes are among all functionally unique sequences in the entire EST collection. Three UPGIs are calculated for three GO domains, respectively: molecular function (UPGI-MF), biological process (UPGI-BP), and cellular components (UPGI-CC) (Figure 3 ; see more about Gene Ontology in methods). Sequence diversity plot (SDP) used as an aid to decide how many probes to spot on microarray The dissimilarities among sequences within a cluster, measured by SDI, decrease as total number of the clusters increases; sequences within a cluster share higher similarity as the number of clusters formed increases (Figure 2 ). From the collection of 971 Litopenaeus setiferus ESTs, the first break point of SDI using single linkage method was 442 clusters (Figure 2 ). An elbow (bend) in SDI, analogous to an elbow of scree plot of the principle component analysis (PCA), indicates that the remaining within-cluster diversity is very low after this number of clusters formed [ 25 ]. The selected probes presented 93% unique molecular functions, 94% unique biological processes, and 96% unique cellular components when 442 clusters were formed using single linkage method (Figure 3 ). Other amalgamation algorithms produced clusterings with different properties. The average and complete linkage methods reduced the sequence dimensionality more efficiently than that by using the single linkage method (Figure 2 ). For the complete linkage method, the break point was observed at 289 clusters, at which, the selected probes represented only 50% of unique molecular functions while the selected probes represented 71% of unique molecular functions using single linkage, and 56% using average linkage (Figure 2 ). The probes selected using single linkage were functionally more unique in all three domains (molecular functions, biological processes, and cellular components) than the ones selected using average or complete linkage methods (Figure 3 ). Exceptions to this rule were found when very small (<60 clusters) or large (>442 clusters) numbers of probes were selected. The functional representativeness of the probes at very high or low ends (<60 or >442 clusters) was comparable using any of the three linkage methods. When 442 probes were selected, 93 – 95% unique biological process, ~92% within-cluster biological process, and ~96% unique cellular component was represented by the selected probes (Figure 3 ). Although fewer annotated EST clusters (number of clusters containing at least one annotated sequences) were formed using single linkage method compared to those selected using the other two linkage methods given a fixed number of cluster within the middle range (~60–442 clusters), more functionally unique probes were selected among the formed clusters by single linkage method (Figure 4 ). Contig assemblage using CAP3 yields a similar result as that of cluster analysis using the single linkage method (Figure 2 ). A total of 461 putative genes was generated using sequence assembly software CAP3 without partitioning the sequences into subgroups (by cluster analysis). These putative unique genes included 356 singlets (single ESTs) and 110 assembled contigs. This result followed closely the result of cluster analysis with single linkage method, which indicated 442 clusters. The EST members in each putative gene were in general agreement with the result of single-linkage cluster analysis with some exceptions. For example, sequence 59 (Penaeidin 2), sequence 10 (Penaeidin 3a), and sequences 177 (Penaeidin 3c) were not assembled into any contigs using CAP3, but they were clustered together when 422 clusters were formed using single linkage method. These sequences share high similarities and high percent identities (E-values < 10 -37 ; Table 1 ), and they are likely to hybridize with each other. Probes selected using clustering methods reflect the hybridization potential compared to the assembly approach. Some sequences, on the other hand, were not clustered into a group although they could be assembled into one putative contig. For instance, sequences 79 and 158 were not clustered in a group because the overlapping segment is marginally short (61 bp/64 bp identical) and this segment is composed of low-complexity sequences (31 pairs of GA repeats, which were masked when using BLAST). The different characteristics of three linkage methods could be further illustrated by local sequence percent identity and the lengths of high scoring pair segments (HSP) (Figure 5 ). Sequences within a cluster formed using single linkage method do not always have to overlap with each other as long as the distances between some of the "linking sequences" are short (the similarities are high). That is, the fragmented ESTs could be "linked" by fragmented (or incomplete sequenced) ESTs and the average within-cluster percent identity is not necessary high when using the single linkage method (Figure 5 ). The sequences within same clusters using the average linkage methods, as expected, have the highest average percent identity (before all three methods converge around 545 clusters). Sequence contiguity assessed by CAP3 (Eq. (2)) has shown similar results observed using the probe functional index, UPGI (Eq. (3); Figure 2 ). Clusters formed using the single linkage method contained slightly more contiguous EST members while the other two linkage methods generated fewer contiguous sequences in the mid range (Figure 2 ). Similarly, when the number of clusters was either very low or high, the results were comparable. ESTs were annotated based on Gene Ontology (GO) terms (details see methods). Three types of functionally unassigned sequences were generated through the GO annotation process: the first type was the sequences having no similar sequences found in the GO database. The majority of ESTs (63%) belonged to this category (607 out of 971 ESTs; Figure 6 ). The second type was similar sequences found in the GO database with the function of those sequences annotated as "unknown." The last type of "unknown" was similar sequences found in the GO database, but only certain domains of GO annotation were complete. For example, it could only have molecular functional annotation associated with the sequence but biological process and cellular components are unknown. The last two types of sequences were combined into one "unknown category" in that particular functional domain (Figure 6 ). Twenty five percent of sequences was annotated in molecular function while 12% was unknown; 27% was annotated in biological process while 10% was unknown; and 27% was annotated in cellular components with 11% unknown. Among the annotated sequences, 36%, 49%, and 22% of annotated sequences were associated with unique GO terms in each of the three domains (molecular function, biological process and cellular component), respectively (Figure 6 ). Both functional and sequence indexes for the three clustering methods converge around the threshold of 442 clusters. When the user-defined number of probes is fewer than this threshold value (442 clusters), the functional uniqueness of the selected probes using single linkage method is superior than that of the other two methods while average linkage is the most effective method for dimension reduction (Figure 2 ). Discussion cDNA microarray is one of the most common microarray platforms, but it is also known to have cross-hybridization potentials. The hybridization potentials between sequences may also vary as the experimental condition changes. This changing nature and the potential of cross-hybridization could be depicted by the index developed in this study, the sequence diversity index (SDI). The magnitude of SDI decreases as the number of clusters increases; sequences are more similar within clusters as the number of clusters increases. SDI is analogous to the F-statistics. That is, SDI is the "within" variation divided by the "total" variation while the F-statistics is "within" variation divided by "between" variation. Two ancillary indexes (a functional index (UPGI) and a sequence contiguity index (CI)) were designed to evaluate the functional representativeness of the selected probes and identify the numbers of putative genes each probe potentially would cross-hybridize. These indexes aid the probe selection processes by bringing in the functional annotations of ESTs as the main goal of the microarray experiments is generally to interpret the biological significances and interactions of genes of interest. A common goal of microarray experiments is to identify co-regulated genes. This is based on the assumption that if two genes are co-expressed, they are likely to be co-regulated through the same mechanism [ 28 ]. It has been shown experimentally, at least in yeast, that combining expression data and sequence functional annotation information results in a better predictive model than using microarray expression data alone [ 29 ]. The integrated procedure in our study including both probe sequence and functional annotation allows a user-defined flexibility based on the purpose of experiments and the limitation or experimental conditions, such as different hybridization stringencies, budget limitations for numbers of probes to spot on the array, or physical size constraint of the array. Different clustering processes mimic different scenarios of cross hybridization between sequences. Sequences from the same transcript will hybridize with each other, and this is reflected in the clusters formed using the single linkage method. In contrast, some of the sequences in the clusters formed by the complete or average linkage methods could be paralogs or alternative splicing variants of the same gene. It might be argued that if a sequence, for example Penaeidin 2, was chosen as a probe from the cluster of sequences containing different subtypes to spot on the microarray, this sequence will likely hybridize to the sequences in the same cluster, for example, Penaeidin 3a and Penaeidin 3c. The contiguity index and probe functional index developed in our study will identify the cross hybridization potential for the users. Potential cross-hybridization has become a more apparent problem for the transcriptomics community. A tool was developed to identify potential cross-hybridized probes lately [ 30 ], however, this tool is designed for species with rich genomic information. Our method provides an integrated approach for cDNA microarray design for any organisms, especially for projects with very limited genomic information. Cross-hybridization potential between long cDNA seqeuences is harder to model than that between short (oligonucleotide) sequences. Although several studies have shown that local sequence percent identity seem to be a reasonable predictor for cross hybridization for cDNA microarray experiments [ 31 - 33 ], the cross-reactivity varies in a wide range (0.6 – 57% signal) even when percent identity is high and within a similar range (80–85% identity) for sequences in different gene families [ 33 ]. Currently, there is (are) no good predictor(s) to model the cross-hybridization on cDNA microarray. The similarity measurement between sequences in our study ( d mi ) could be easily replaced in the future by any good cross-hybridization predicting parameter(s) developed for long cDNA sequence hybridization. The design procedure we described here will work in the exact same fashion. In a similar manner, although the traditional hierarchical clustering algorithm with three linkage methods was used in our study, any bottom-up clustering algorithm (e.g., K-nearest means clustering) or top-down approach (e.g., principle component analysis, single value decomposition) could be easily performed, and the corresponding SDIs, UPGIs and CIs will be generated in the same way and summarized in the SDP. The performances of these different bottom-up or top-down algorithms (to group or partition the sequences) could be compared using the SDP. In brief, other distance matrix and clustering algorithms other than what we used in this study could be easily applied using our algorithm, and their performances could be evaluated quantitatively using the suite of indexes in SDP. Annotation of functionally unknown sequences is not a trivial task itself. Gene Ontology has become a standard ontology to annotate unknown sequences. Sequence similarity search against the GO database using BLAST was used in this study. The completeness of the GO database and sensitivity/selectivity of the BLAST procedure would dictate the annotation capability. Several different approaches could potentially improve the annotation in the future. One example found in this study was Penaeidin family, a unique family of antimicrobial peptides with both proline and cysteine-rich domains that were first identified and characterized as peptides in the hemolymph of the Pacific white shrimp, Litopenaeus vannamei [ 34 ]. No homologous proteins are found in GO database. Future research could emphasize how to integrate other sources of knowledge (database) to enrich the functional annotation process, especially as very limited knowledge is available for marine organisms in the public domain. Different approaches to annotate the EST sequences could also be adapted. For instance, position-dependent method (such as using HMMER [ 35 ] to search against Pfam database [ 36 , 37 ]) could be used to search the existing database. This may increase the chance to annotate sequences with lower sequence homologies with the sequences stored in the database. The current functional representativeness of the selected probes was quantified using the unique GO terms associated with the probes among all sequences. The quantification of how representative the probes are could be modified in the future to include the hierarchical nature of the GO terms. The integration of this cDNA probe selection procedure with the database through Marine Genomics web-based interface [ 27 ] is currently in progress, and the marine genomics community will be directly benefited, and it will be equally applicable to any organisms with limited genomic information. Conclusions The sequence diversity index (SDI) was developed in this study to select probes using ESTs for designing cDNA microarrays. Two ancillary mathematic indexes (sequence contiguity index [CI] and unique probe GO index [UPGI]) were used to identify potential cross-hybridization between different transcripts (or paralogs) and to quantify biologically how representative the probes were. These three indexes were summarized in a sequence diversity plot (SDP) and were used to assist cDNA microarray probe selections for organisms without any genomic information. This method allows the user-defined number of probes to be selected for the cDNA microarray experiments. Different clustering methods balance the representativeness of the probe functional annotations and minimization of the sequence redundancies. Accordingly, different linkage methods can be used to decide between microarray designs for biomarker discovery or for functional genomics. It is clear that sequence assembly into contigs is not necessary for microarray probe selection although it is informative to identify the relationship among sequence members within clusters based on the CI. The microarray design procedure described here could also be used for multi-species or cross-species microarray design in a scenario where the sequences with high similarity from different species cross hybridize to each other [ 32 ], but not necessary be assembled into contigs. This method is not limited to the ESTs collected from single or multiple marine organisms. Furthermore, this method can be applied to any organisms without the complete sequenced genomes. Methods Sequence availability and pre-processing Twenty six thousand and six hundred fifty-six (26,656) Expressed Sequence Tags (ESTs) from 14 marine species were generated and stored in a postgresSQL database through a user-friendly interface at the Marine Genomics website [ 27 ]. All the sequences are freely available to the public. One thousand and forty seven ESTs from Atlantic white shrimp ( L. setiferus ) were used in this study. Pre-processing included customized low quality filtering, poly-A tail, vector, adaptor screening, trimming, and low-complexity masking by DUST [ 38 ]. After pre-processing, 971 sequences longer than 100 bp were further analyzed (Figure 1 ). Sequence similarity comparison All against all pair-wise BLASTN [ 39 ] was performed between these 971 ESTs. In the BLASTN result, with sufficiently large sequence lengths q and n , the statistics of HSP (high-scoring segment pairs) scores are characterized by two parameters, K and lambda . The E-value, the expected number of HSPs with score of at least S , given by the formula E = Kqne -λS , was used as the distance measurement ( d mi in Eq (1) in results) between ESTs for cluster analysis to determine sequence redundancies. d mi is the distance between the i th pair of sequences for a total k m pairs of within-cluster comparisons when m clusters are formed. Sequence redundancy reduction by cluster analyses Hierarchical cluster analyses with three common linkage methods (single linkage, average linkage, complete linkage) were performed to reduce the redundancies among sequences. Two sequence indexes were used to quantify sequence diversity and contiguity within clusters Two sequence indexes, the sequence diversity index [SDI; Eq. (1) in results] and the contiguity index [CI; Eq. (2) in results], were used throughout the sequence redundancy reduction. SDI was used to aid the number of probes to select. The within-cluster sequence contiguity (CI) is evaluated using CAP3 assembly software with default parameters [ 23 ]. Unweighted average within-cluster percent identity of the HSP segments and HSP length from BLAST results were quantified throughout the process of clusterings. Probe selection To maximize the hybridization probability between the selected probe and the sequences within the cluster, the sequence has the highest similarity to all the other sequences within the cluster is selected. That is, the centroid EST, the sequence has the minimum average distance to all the other sequences within each cluster was spotted on the array. Sequence functional annotation using Gene Ontology (GO) terms Functions of all 971 sequences were annotated using the functional categorizations of similar sequences stored in Gene Ontology (GO) database [ 40 ]. GO terms are commonly used for functional categorization in three domains (biological process, molecular function, and cellular component) for gene products (proteins) or nucleotide sequences. The GO terms and associated protein sequences were downloaded from the GO website [ 41 ] in the format of mySQL database [ 42 ]. The ESTs were annotated by the top BLASTX hit after blasting them against the proteins with GO terms associated in the database. The sequences with the E-value threshold set at 10 -6 for GO annotation are considered as similar, and they potentially share the same molecular functions, cellular components, or biological processes. The GO terms found associated with the EST sequences, if any, were recorded separately for each of the three domains. If there were multiple GO terms in any single domain (e.g., molecular function), the inverse of the number of GO terms in that domain is used for functional quantification (i.e., the traditional pie-chart summary of the functional categories of ESTs). For example, there are three molecular functional annotations (GO:0005515, GO:0004866, GO:0004867) associated with the sequence 1046, then each of them is considered 1/3 in the GO quantification for this particular sequence. Therefore, the quantification for each GO domain will sum up to the original analyzed sequence numbers at the end when we quantify the percentage of each category (n = 971). A functional index to quantify how representative the selected probes are The unique probe GO index [UPGI; Eq. (3) in results] was used to quantify functionally how representative the selected probes were within EST clusters. Number of probes to retain using the sequence diversity plot (SDP) Two sequence indexes (CI and SDI) and one functional index (UPGI) mentioned above were included in the sequence diversity plot (SDP) (Figure 1 ). Sequence similarity was measured by SDI (Eq. (1) in results), and within-cluster sequence contiguity was measured by CI (Eq. (2) in results). The unique probe GO index (Eq. (3) in results) was used to quantify functional annotation levels represented by the selected probes. This integrated information will allow user-defined flexibility of probe selection involving both sequence similarity and functional annotation. List of abbreviations SDP: sequence diversity plot, SDI: sequence diversity index, CI: sequence contiguity index, and UPGI: unique probe GO index	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539232.xml
514607	Epidemiology of the human immunodeficiency virus in Saudi Arabia; 18-year surveillance results and prevention from an Islamic perspective	Background data on HIV epidemiology and preventive measures in Islamic countries is limited. This study describes the results of 18-year of HIV surveillance in Saudi Arabia (SA) and the preventive measures implemented from an Islamic perspective. Methods surveillance for HIV has been underway in SA since 1984. Indications for HIV testing include clinical suspicion, screening of contacts of HIV-infected patients, and routine screening of blood and organ donors, prisoners, intravenous drug users, patients with other sexually transmitted infections, and expatriates pre-employment. This is a case series descriptive study of all confirmed HIV infections diagnosed in SA from 1984 through 2001. Results a total of 6046 HIV infections were diagnosed, of which 1285 (21.3%) cases were Saudi citizens. Over the 18-year surveillance period the number of HIV infections diagnosed annually among Saudi citizens gradually increased and, over the period 1997–2001, it reached to 84 to 142 cases per year. The number of cases per 100,000 population varied widely between regions with a maximum of 74 cases and a minimum of 2 cases. The infection was most common in the age group 20–40 years (74.6%) and predominantly affected men (71.6%). The modes of transmission among Saudi citizens and expatriates, respectively, were as follows: heterosexual contact, 487 (37.9%) and 1352 (28.4%) cases; blood transfusion, 322 (25.0%) and 186 (3.9%) cases; perinatal transmission, 83 (6.5%) and 19 (0.4%) cases; homosexual contact, 32 (2.5%) and 38 (0.8%) cases; intravenous drug use, 17 (1.3%) and 33 (0.7%) cases; bisexual contact, 10 (0.8%) and 14 (0.3%) cases; unknown, 334 (26.0%) and 3119 (65.5%) cases. The number of HIV infections transmitted by blood or blood products transfusion declined to zero by year 2001 and all such infections occurred due to transfusions administered before 1986. At HIV diagnosis, 4502/6046 (74.5%) patients had no symptoms, 787 (13.0%) patients had non-AIDS defining manifestations, and 757 (12.5%) patients had AIDS. A total of 514/1285 (40%) Saudi patients died by year 2001. Conclusions the number of HIV cases in SA is limited with heterosexual contact being the main mode of transmission. From an Islamic perspective, preventive strategies include prevention of non-marital sex and intravenous drug use with encouragement of "safe sex" through legal marriage.	Background The human immunodeficiency virus (HIV) continues to be a major health problem worldwide despite enormous efforts to control its spread. As of December 2002, the estimated number of people living with HIV is 42 millions [ 1 ]. In year 2002 alone, the HIV epidemic claimed more than 3 million lives and an estimated 5 million people acquired HIV [ 1 ]. Available data point to increasing HIV infection rates in the Middle East and North Africa, with an estimated 83000 people having acquired the virus, and 37000 having died, in 2002. This brings to 550000 the estimated number of people living with HIV in this region [ 1 ]. Sexual contact remains the main mode of transmission of the disease worldwide followed by intravenous drug use and perinatal transmission [ 1 ]. Transfusion of blood and blood products is no longer a significant risk factor for acquiring HIV infection since the introduction of routine pre-transfusion HIV screening in 1985–1986 in most countries. Saudi Arabia occupies most of the Arabian Peninsula with an area of about 2,240,000 sq km (figure 1 ). It comprises 13 administrative provinces, namely, Makkah province (which includes the holy city of Makkah, Jeddah and Tayef), Madinah province (which includes the holy city of Madinah), Riyadh province (which includes the capital city Riyadh), the Eastern province (which includes Dammam, Ahsa, and Hafr Albaten), Asir province (which includes Abha and Bisha), Joaf province (which includes Joaf and Qerayyat), Hudud Shamaliyah (North borders) province (which includes Arar), and Baha, Jizan, Najran, Hail, Qassim, and Tabook provinces. The capital and largest city is Riyadh. The population in Saudi Arabia is 24,293,844, including 5,576,076 non-nationals as estimated in July 2003. Approximately, 42.3% of the population is below 15 years of age, 54.8%, between 15 and 64 years, and 2.9%, above 64 years of age. Ninety nine percent of the population is Muslim and the country is governed according to the Islamic law. Information on HIV epidemiology in Saudi Arabia and other Islamic countries is limited. Islam prohibits non-marital sex, homosexuality, and intravenous drug use. Therefore, the prevalence and incidence of HIV and other sexually transmitted infections are expected to be low in Islamic communities. This study describes the results of HIV surveillance activities that have been underway in Saudi Arabia from 1984 through 2001 and the preventive measures implemented by the government. Methods Surveillance for HIV infections has been underway since 1984 when the first case of HIV was diagnosed in Saudi Arabia. HIV cases are detected by HIV testing for various indications including clinical suspicion, screening of contacts of HIV-infected patients, routine screening of blood and organ donors, and testing of all prisoners, intravenous drug users, patients with other sexually transmitted infections, and expatriates pre-employment. In addition to being HIV tested in their homeland as a compulsory pre-requisite for employment in Saudi Arabia, expatriates are routinely retested for HIV upon arrival to Saudi Arabia before they are allowed to work and then regularly every 2 years to have their legal residence permits renewed. Only HIV-negative expatriates are allowed to work in Saudi Arabia. Routine screening of blood, blood products, and organ donors for HIV has been a standard procedure in Saudi Arabia since 1986. Enzyme-linked immunosorbent assays (ELISA) are used for routine HIV1/HIV2 testing. Positive ELISA results are confirmed by Western blot test performed in selected referral laboratories. The expanded World Health Organization (WHO) case definition was used to define AIDS [ 2 ]. All HIV infections diagnosed in governmental or private health care facilities are notified to the Ministry of Health (MOH) using unique identifying codes. This policy is strongly enforced by law. In addition to notifying the MOH, the Ministry of Interior is also informed about expatriates infected with HIV. HIV-infected expatriates are treated for any acute or life threatening complications and they are then returned back to their homeland. Saudi patients are referred to tertiary care governmental HIV-specialized centers where highly active antiretroviral therapy (HAART) medications and essential laboratory tests, such as HIV viral load and CD4/CD8 counts, are available. Epidemiological data are collected from patients by the attending physicians on standardized data collection forms. The likely mode of transmission is determined by the attending physician after interviewing the patient and taking complete history regarding high risk behaviors. The mode of transmission is considered to be unknown if the patient fails to admit to any high risk behavior. Collected information is subsequently submitted to the Department of Preventive Medicine in the Central MOH office in Riyadh where all surveillance data are compiled. Annual reports are issued but they are only utilized internally by the concerned officials in the MOH and the Ministry of Interior and they are not made available for the public. Results From 1984 through 2001, a total of 6046 cases with HIV infection were diagnosed; 1285 (21.3%) cases among Saudi citizens and 4761 (78.7%) cases among expatriates. Table 1 shows the demographic data, risk factors, and mortality of the infected patients. The likely mode of HIV transmission among 340 Saudi infected female patients was as follows: blood transfusion, 86 (25.3%) patients; marital sex, 74 (21.8%) patients; maternal transmission to female babies, 27 (7.9%) patients; non-marital sex, 8 (2.4%) patients; intravenous drug abuse, 3 (0.9%) patients; and in the remaining 142 (41.7%) patients the mode of transmission was unknown. Among 1376 non-Saudi HIV infected female patients, the likely mode of transmission was as follows: non-marital sex, 209 (15.2%) patients; blood transfusion, 32 (2.3%) patients; marital sex, 4 (0.3%) patients; intravenous drug abuse, 3 (0.2%) patients; maternal transmission to female babies, 3 (0.2%) patients; and in the remaining 1125 (81.8%) patients the mode of transmission was unknown. A total of 157/1285 (12.2%) Saudi patients had AIDS at the time of HIV diagnosis whereas the rest had either no symptoms (884 patients, or 68.8%) or non-AIDS defining manifestations (244 patients, or 19.0%) such as generalized lymphadenopathy, oral or vaginal thrush, oral hairy leukoplakia, recurrent herpes simplex, herpes zoster, molluscum contagiosum, condyloma acuminata, thrombocytopenia, or aphthus ulcers. Similarly, 600/4761 (12.6%) non-Saudi patients had AIDS at the time of HIV diagnosis, 3618 (76.0%) patients had no symptoms, and 543 (11.4%) had non-AIDS defining manifestations. Among 4761 HIV-infected expatriates, 3771 (79.2%) patients were from African countries, 624 (13.1%) patients, from Asian countries, 352 (7.4%) patients, from Middle East countries, and 14 (0.3%) patients, from Western countries. Ninety two percent of patients were from 10 countries, namely, Ethiopia (2214 patients or 46.5%), Nigeria (343 patients or 7.2%), Chad (329 patients or 6.9%), Yemen (309 patients or 6.5%), Sudan (267 patients or 5.6%), Eritrea (248 patients or 5.2%), India (219 patients or 4.6%), Somalia (176 patients or 3.7%), Pakistan (152 patients or 3.2%), and Bangladesh (133 patients or 2.8%), and the remaining 371 (7.8%) patients were from 41 other countries. Table 2 shows the number of HIV cases per 100,000 population by region. HIV cases have been reported from virtually all regions of Saudi Arabia but the highest prevalence was observed in Jeddah and Makkah in the western province (figure 1 ). Table 3 shows the indications for HIV testing in the Saudi patients. Approximately, one third of cases were identified because of clinical suspicion and the rest of cases were asymptomatic subjects identified following HIV screening for the indications mentioned in table 3 . Figure 2 shows the annually reported HIV infections among Saudi patients over the surveillance period 1984 through 2001. There has been a gradual increase in the number of cases reported annually but this seems to have plateaued in the period 1997 through 2001. Figure 3 shows the annually reported number of HIV infections transmitted by transfusion of blood or blood products by year of HIV diagnosis from 1984 through 2001. The number of these infections has declined to zero by year 2001. All such infections were due to transfusions administered to patients before 1986. Discussion The prevalence and the annually reported HIV infections in Saudi Arabia were limited. More than three quarters of HIV cases were expatriates. Expatriates are routinely tested for HIV upon arrival to Saudi Arabia and then regularly every 2 years to have their legal residence permits renewed. Therefore, one possible explanation for the higher prevalence of HIV infections among expatriates is that HIV testing rates might be higher among expatriates compared to Saudi citizens. The number of HIV infections diagnosed annually among Saudi citizens gradually increased to reach a plateau of 84 to 142 cases per year in the last 5 years of the study period. The infection was more prevalent in Jeddah than it was in other cities in Saudi Arabia. A possible explanation for this increased prevalence in Jeddah is that this city is the main air and sea port for expatriates who come to work in the western province of Saudi Arabia and for those who come to visit the holy places in Makkah and Madinah. Additionally, the population in Jeddah is multi-cultural and inhomogeneous from the religion point of view. HIV infection in Saudi Arabia, as the case worldwide, was more common in the age group 20–40 years and predominantly affected men. All modes of transmission were registered but the main mode was heterosexual contact. Even though blood transfusion caused one quarter of HIV cases among Saudi patients, all such infections occurred due to transfusion of blood or blood products before 1986. This risk factor was virtually eliminated since the introduction of routine HIV screening of donated blood and blood products in 1986. The vast majority of Saudi HIV-infected women with identifiable risk factors acquired the infection through blood transfusion or marital sex with their infected husbands. The large proportion of patients with "unknown" risk factors was likely due to deliberate concealment of the actual risk factors, likely to be illegal sex, by the infected patients as non-marital sex and homosexuality are prohibited in Islam, the religion of all Saudi citizens and the vast majority of expatriates living in Saudi Arabia. The international efforts to control the HIV pandemic have failed to control it on a global scale despite their partial success in many of the developed countries and some of the developing countries. Limitation of resources in underdeveloped countries makes health education and other preventive strategies difficult to implement. Additionally, the limited access to antiretroviral medications in the underdeveloped countries has compounded the problem and perhaps contributed to some extent to the spread of HIV from untreated patients to new victims. Substance abuse is an increasing problem in Saudi Arabia as it is in the rest of the world [ 3 ]. Substances abused include injectable drugs such as heroin and cocaine and non injectable drugs such as cannabis and amphetamine-type stimulants. The estimated annual prevalence of heroin and amphetamine abuse in Saudi Arabia in 2000 as percentage of the population aged 15 and above was 0.01% and 0.002%, respectively [ 3 ]. The number of drug abusers annually admitted to detoxification centers in Riyadh, Jeddah, Dammam, and Qassim from 1996 through 2001 ranged from 4740 to 6650 patients with an average annual increment of 5.1% (unpublished data). Several studies were conducted in Saudi Arabia to describe the socio-demographic characteristics, pattern of substance abuse, and prevalence of blood-borne infections among drug abusers. For instance, 799 drug abusers from a voluntary detoxification unit in Jeddah were studied [ 4 ]. Sixty eight percent of subjects were under 35 years of age and 64% initiated drugs before age 25. Eighty seven percent used heroin or alcohol and 14% were dependent on more than one drug. Among heroin users, 91% injected the drug. The prevalence of hepatitis C virus infection among these patients was 69% [ 4 ]. In another study of 349 drug abusers in Jeddah, 281 (80.5%) subjects were intravenous drug users. The prevalence of HBsAg, anti-HBs, and anti-HBc was 12.6%, 49.0%, and 53.6%, respectively, suggesting that sharing of needles was a common practice [ 5 ]. In a more recent study in Jeddah including 1321 drug abusers, 1038 (78.6%) subjects were intravenous drug users and the prevalence of HBsAg and anti hepatitis D virus was 6.1% and 0.8%, respectively [ 6 ]. The prevalence of confirmed HIV infection among 2628 drug abusers in Jeddah, of whom 80% were intravenous drug users, was found to be only 0.15% [ 7 ]. Among 116 drug users in the Eastern province, 83% of subjects were below 32 years of age, 52.6% were unemployed, and the majority were of intermediate education [ 8 ]. Eighty-four percent of the patients abused heroin either alone or in combination with other drugs, 31% used alcohol, 26% used cannabis, and 10% used stimulants. The use of other drugs was rare [ 8 ]. Non-marital sex is expected to remain the main risk factor for acquiring HIV infection in Saudi Arabia as it is worldwide for several reasons. The ever-decreasing religious values, the ever-increasing ease of international transportation and communication, and the increasing poverty and unemployment are the main driving forces for non-marital sex, not only in Saudi Arabia but all over the world. Intravenous drug use is also expected to play a more important role for HIV transmission in Saudi Arabia in the years to come as the population of intravenous drug users increases. Preventive strategies should thus be directed towards these risk factors. Some of the preventive strategies that are advocated and used in other non-Islamic countries such as "Safe Sex" and "Needle Exchange Programs" contradict the Islamic rules and values and as such can not be used as valid and acceptable strategies to prevent the spread of this infection in Islamic countries. The concept of "Safe Sex" basically promotes the use of condoms for non-marital sexual relations, considered in Islamic countries a way of promoting non-marital sex which is absolutely prohibited in Islam. Needle Exchange Programs, likewise, is viewed by Islam as a way of encouraging people engaged in intravenous drug use to continue this prohibited practice. Strategies to prevent HIV infection in Islamic countries have to abide by the Islamic rules and values. Proposed preventive strategies that can be used in Islamic countries include strengthening of both Islamic and health education, encouraging people to follow and implement the Islamic rules and values that prohibit adultery, homosexuality, and intravenous drug use, and to practice safe sex only through legal marriage. Encouragement of legal marriage is a multidisciplinary task that should involve both governmental and non-governmental charitable organizations and the population at large to cut the cost of marriage and support programs that help the youth to get married. There are several such programs in Saudi Arabia. For example the "Charitable project to help the youth to get married" in Jeddah has helped 2500 young men to get married at a cost of 10 million Saudi Riyals (2.7 million US dollars) collected from donations (personal communication). Other Islamic aspects that are important in preventing non-marital and irresponsible sexual relationships include the fact that men are allowed to be married to up to four women and the fact that there is no age limit for marriage, thus permitting adolescents to get married. Further, Islam obliges women to cover themselves with the so called "Hijab" or veil and to be segregated from men in educational institutes and other mass-gathering. places. Islam also fights poverty, a driving force for commercial sex and prostitution, through a well established system of obligatory charity, known as "Zakat", and voluntary charity, known as "Sadaqa", taken from the rich people and given to the poor and needy. Additionally, Islam obliges the rulers to eliminate all means and factors that are conducive for indulging in non-marital sex and intravenous drug use such as sex trade, prostitution, and drug smuggling and to implement the Islamic penalties on those involved in such illegal acts. It should be noted, however, that HIV preventive strategies in some Islamic countries do not necessarily abide by the Islamic doctrine and that knowledge, attitude, and practice of Muslims in various Islamic societies do not necessarily conform to Islamic norms. Unfortunately, there is little published data on the effectiveness of the Islamic preventive strategies in reducing HIV transmission in Muslim communities. According to the United Nations and the World Health Organization data on HIV prevalence in different countries, however, the prevalence of HIV infection in Islamic countries is strikingly low compared to other countries [ 9 , 10 ]. A recent study showed that among 38 sub-Saharan African countries, the percentage of Muslims within countries negatively predicted HIV prevalence [ 11 ]. A survey of published journal articles containing data on HIV prevalence and religious affiliation showed that six of seven such studies indicated a negative relationship between HIV prevalence and being Muslim [ 11 ]. Conclusions Even though the number of HIV cases in Saudi Arabia is still limited, there is a potential for a rapid spread of this virus particularly among the youth. Strategies to prevent the spread of this virus in Saudi Arabia as well as in other Islamic countries should conform to the Islamic rules and values to be successful. Competing interests None declared Authors' contributions TAM designed the study, analyzed the data, and wrote the manuscript. YYM, MHJ, and NSH supervised the data collection and revised the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514607.xml
549514	Nosocomial Plasmodium falciparum infections confirmed by molecular typing in Medellín, Colombia	Three cases of nosocomial malaria are reported from patients of the Internal Medicine Ward of a tertiary University teaching hospital in Medellin, Colombia. Epidemiological research, based on entomological captures, medical records review and interviews of nursery staff about patient care practices potentially involving contact with blood, were carried out. Molecular characterization of Plasmodium falciparum was based on the amplification of MSP1, MSP2 and GLURP genes. This method enabled confirmation of the same P. falciparum genotype in all three patients as well as in a fourth one (index case). The presence of nosocomial malaria was confirmed and it was concluded that the most likely source of transmission was through multi-dose preparations of heparin applied to heparin locks.	Background The city of Medellin is located in the north-western region of Colombia and, although malaria is endemic in the country, absence of anophelines within the city makes vector transmission of the infection impossible. However, this city, located within the Andes Mountains (6°13'N, 75°36'W) at 1,588 m above sea level (a.s.l.), is surrounded by highly endemic regions and imported cases are often referred to the University hospital for treatment. Here, three cases of nosocomial malaria, which occurred in hospitalized patients in the Internal Medicine Ward of Hospital A, are described. These patients were being treated for diverse conditions and they shared the hospital facilities during different periods. One of these patients was in the Emergency Ward at the same time as a Plasmodium falciparum -infected patient. In order to confirm an association between all the cases, P. falciparum genomic DNA was amplified using specific primers for polymorphic streches of Merozoite Surface Antigen-1 (MSP-1), Merozoite Surface Antigen-2 (MSP-2) and Glutamate Rich Protein (GLURP), and the profile observed among the different isolates was compared. Case presentations Patient 1. Index case A 23-year-old male woodcutter, resident in Vigia del Fuerte (6.35N, 76.53W), was referred to the hospital on 11 April 2001, diagnosed with yellow fever. Vigia del Fuerte is a highly endemic malaria region located in the Atrato River basin with a mean Annual Parasite Index (API) of 72.6 during the past five years. The patient presented 15 days with fever, headache, chills, jaundice, arthralgia, melena, dark urine and bilious vomiting. He was given a heparin lock for administration of I.V. treatment. The patient deteriorated throughout the night and died before thick smear results were available. Peripheral blood thin and thick smears confirmed the presence of P. falciparum asexual forms (>100,000 per μl). A post-mortem confirmed as cause of death complicated malaria. Dengue and yellow fever IgM antibodies were absent. The patient remained in the emergency ward for 18 hours prior to his death. Patient 2 A 19 year-old male, unemployed, resident in Angelopolis (Antioquia, 6.06°N 75.42°W), a non-malarious area, located 1,900 m a.s.l., attended the emergency service of the hospital on 10 th April 2001, and remained there for 24 hours before being transferred to the Internal Medicine Ward with diagnosis of systemic lupus erythematosus (SLE). In addition to symptoms and signs of SLE, he presented fever on 23 April, and received treatment with fenitoin, dipirone, ampicillin, ranitidine, ceftriazone, methilprednisolone, oxacillin, amikacin and enoxaparine (via an heparin lock). As part of the treatment for his condition, he was administered chloroquine 150 mg daily from 24 th April 2001. He was discharged on 19 th May 2001. Then, on 6 th June 2001, he was hospitalized with a diagnosis of a febrile syndrome. P. falciparum malaria was confirmed on 20 th June 2001 after observation of the parasite in a blood cell count. A thick smear revealed the presence of 14,440 asexual forms per μl in peripheral blood. The patient was administered quinine and sulphadoxine-pyrimethamine with good treatment response. Patient 3 A 46 year-old male, unemployed, resident of Santa Barbara (Antioquia, 5.57°N 75.91°W), a non-malarious area located 1,800 m a.s.l., was hospitalized in the Internal Medicine Ward on 8 th May with diagnosis of status asthmaticus. He was administered aminophyllin, hydrocortisone, ranitidine, enoxaparine (via a heparin lock) and was discharged on 15 th May 2001. He returned to the hospital on 5 th June 2001 complaining of fever, chills and adynamia and was admitted with diagnosis of toxic hepatitis due to the presence of dark urine, jaundice and hepatomegaly. On 6 th June 2001, diagnosis of complicated malaria was confirmed by observation of the parasite in a blood cell count. Later, the presence of 87,500 trophozoites of P. falciparum per μl was confirmed by a thick smear. He received treatment with quinine, sulphadoxine-pyrimethamine and was discharged on June 23 rd 2001 Patient 4 A 41 year-old male, flower-grower, resident of El Carmen del Viboral (Antioquia, 6.09°N 75.34°W), a non-malarious area located 2,150 m a.s.l.., was admitted directly to the Internal Medicine Ward of the hospital on 7 th May 2001 with superior vena cava syndrome (100% obstruction) and was applied a heparin lock for administration of I.V treatment. The patient was diagnosed with a mediastinal mass, compatible with lymphoma, and received treatment with dexametasone (I.V.) and radiotherapy. On June 9 th 2001, he evidenced fever and on 9 th June 2001, he was diagnosed as complicated malaria by a cell blood count. A thick smear confirmed the presence of 166,440 asexual forms of P. falciparum per μl. He received treatment with quinine, sulphadoxine-pyrimethamine with good treatment response. However, on 1 st July 2001, he died as a consequence of multiple non-malaria related complications. Entomological captures Search for Anopheles was carried out by expert personnel from the Secretariat of Health, both within the ward and in the surrounding gardens, around the time of malaria diagnosis of patients 2 and 4. These confirmed the absence of the vector in the hospital area. Molecular analysis P. falciparum genomic DNA was extracted from whole blood collected onto filter paper or from paraffin embebbed brain tissue (for the index case). Paraffin was removed from post-morten material using xylene, followed by 2 washes with 100% ethanol. DNA was isolated by proteinase K digestion, followed by four rounds of phenol-chloroform extraction. Purified DNA was stored at -20°C until amplification. The Region 2 of MSP-1 and the central region of MSP-2 were amplified by a nested Polymerase Chain Reaction (PCR); the region RII of GLURP was amplified by a semi-nested PCR [ 1 ]. Products obtained after the first PCR were amplified using specific primers for region 2 of MSP-1 corresponding to MAD20, K1 and RO33 allelic families, and FC27 and IC-1 for the central region of MSP-2. Briefly, PCR was carried out in a total volume of 20 μL, containing 10 mm Tris-HCl (pH 9,0 at 25°C), 50 mm KCl and 0,1% Triton X-100, 125 mM dNTPs, 0,4 units Taq DNA polymerase (Promega, Madison, WI), 1,6 mM MgCl2 and 125 nM of each primer. Initial denaturation was 5 min at 95°C, 1 min at 94°C, 2 min at 58°C annealing (all first PCR reactions and second reaction for GLURP) or at 61°C (for all the second reactions of MSP-1 and 2). This was followed by extension for 2 minutes at 72°C. This first reaction underwent 25 amplification cycles and the second 30 cycles. Positive (strains HB3, K1 and RO33) and negative controls (healthy individuals), were included. Products were electrophoresed on an agarose (MetaPhor) gel (2,5% for MSP-1 and 2% for MSP-2 and GLURP), stained with ethidium bromide and visualized under ultraviolet light. Size analysis of the amplified fragments revealed identical pattern distribution for all the examined markers assayed, confirming the presence of P. falciparum infection by matching strains in all 4 patients (Fig 1 ). Figure 1 Dates, chronology of exposure, place of contact and time of malaria diagnosis of the index case and the 3 patients with nosocomial malaria. Each bar represents the location of each patient during the stay at the hospital. Conclusions The presence of P. falciparum malaria infection in three patients without history of malaria or travel to malaria-endemic regions is described. The first of these patients (Patient 2) shared the Emergency Ward with a fatal case of falciparum malaria. The incubation period for this nosocomial infection was 12 days, however, since he was administered high doses of chloroquine, fever receded for about two weeks. Resistance of P. falciparum to chloroquine is highly prevalent in the Antioquia region [ 2 ], this explains the lack of efficacy of the antimalarial in eradicating parasites in this patient. Patient 2 shared the Internal Medicine facilities with patients 3 and 4 at least for one week during May. Patient 3 evidenced an incubation period of 22 days, while in Patient 4 this was 26 days. Since patients 3 and 4 had no contact with the index case, further analysis of the incubation periods allowed us to conclude that transmission occurred from index case to patient 2 and the latter was the source of infection for patients 3 and 4. This means that at least 2 separate episodes of contamination were involved in transmission to the different patients (see figure 2 ). Since entomological captures were negative for vectors and Medellin is not an area of vector transmission for malaria, parenteral contamination might be the most likely mechanism of transmission. None of the patients was administered blood before diagnosis of malaria, but all of them had at least one heparinized lock during their stay either in the Emergency or in the Internal Medicine Ward. Figure 2 Agarose gel electrophoresis showing the size of the amplified products of the genes MSP-1 (panel A), MSP-2 (panel B) and GLURP (panel C). MW represents the molecular weight marker,(-) a negative control and (+) a positive control. The numbers are equivalent to the case number. Panel A corresponds to allelic family MAD-20 of MSP-1 showing amplification products of ~150 bp in the cases, panel B corresponds to allelic family IC50 showing amplification products of ~450 bp in the cases, panel C corresponds to GLURP showing amplification products of ~700 bp in the cases. Transmission of malaria via heparin locks has been reported previously by other authors, the source of contamination was a multidose heparin container [ 3 ]. Although the practice of multiple dose preparations of heparin into large volume syringes to be distributed among the several patients of the ward is forbidden at this tertiary institution, it has been difficult to eradicate due to the high volume of patients and the limited economical resources of the hospital. Other researchers have reported on the use of molecular typing of P. falciparum to confirm nosocomial transmission of malaria [ 4 ]. Genotyping of all 4 cases confirmed the epidemiological suspicion of nosocomial transmission. Considering the previous report on the limited genetic polymorphism of P. falciparum observed in the Antioquia region [ 5 ], the presence of an identical genotype of P. falciparum among unrelated parasites may be possible, but unlikely, since examination of these 3 different genes has been routinely used to discriminate unrelated parasites [ 6 ]. Moreover, the molecular findings together with the epidemiological characteristics contribute to confirm nosocomial transmission of the infection. These observations are of major relevance to sanitary and health authorities, since they confirm the importance of biosafety during minor procedures, such as application of heparin to peripheral locks. In addition, it highlights the need to rule-out malaria infections in all febrile patients sharing hospital facilities with malaria patients as well as in those not responding as expected to antimicrobial therapy. Authors' Contributions GL was involved in the review of clinical records, staff interviews and preparation of the manuscript. OJ coordinated the epidemiological and entomological researches. FL coordinated the microscopy diagnosis and following of the patients. AM participated on the epidemiological research. RE carried out the molecular analysis of the samples. BS, the Grupo Malaria leader, participated in the treatment and following of the patients. MA attended the patients, coordinated the clinical and molecular aspects of the research and prepared the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549514.xml
553969	Automated generation of heuristics for biological sequence comparison	Background Exhaustive methods of sequence alignment are accurate but slow, whereas heuristic approaches run quickly, but their complexity makes them more difficult to implement. We introduce bounded sparse dynamic programming (BSDP) to allow rapid approximation to exhaustive alignment. This is used within a framework whereby the alignment algorithms are described in terms of their underlying model, to allow automated development of efficient heuristic implementations which may be applied to a general set of sequence comparison problems. Results The speed and accuracy of this approach compares favourably with existing methods. Examples of its use in the context of genome annotation are given. Conclusions This system allows rapid implementation of heuristics approximating to many complex alignment models, and has been incorporated into the freely available sequence alignment program, exonerate.	Background George Box [ 1 ] once remarked that, "All models are wrong, but some are useful." This statement bears much relevance to biological sequence alignment, where there is no guarantee that the alignment model will accurately represent the evolutionary (or other) processes which separated the two sequences. All that is certain is that exhaustive dynamic programming (DP) algorithms (such as Smith-Waterman [ 2 ]) will yield the optimally scoring path in terms of the given model. Heuristics for sequence comparison (such as BLAST [ 3 ]) generate alignments which are valid paths through the underlying alignment model, but are not guaranteed to be optimal. Alignments generated by these heuristics can be calculated much more rapidly, and often closely match alignments which would be generated by exhaustive methods. Furthermore, many problems in the context of genome analysis consist simply of alignment of gene products (cDNA or protein) back onto the gene from which they were produced, and consequently do not require very sensitive alignment. Hence the aim here is not to attempt to develop models which are correct, but to facilitate development of models which are more useful in the context of large scale analyses. Transformation between a finite state automata describing an alignment model and the recurrence relations used in DP is a powerful technique [ 4 , 5 ] as it allows modification of the alignment algorithm by direct manipulation of the alignment model. The Dynamite compiler [ 6 ] allows automated implementation of alignment algorithms from a description of the alignment model. This allows development of complex models which can be used to generate accurate alignments. However, calculation of alignments using these models always requires quadratic time, which is prohibitively slow for many large scale applications. This method can be applied to any alignment problem which can be represented by a regular grammar (or the equivalent finite state machine). This includes the simple three state model required for affine gaps in the Smith-Waterman-Gotoh algorithm [ 2 , 7 ], but also more complex models such as that used by EST_GENOME [ 8 ], where splice site prediction is integrated into the DP, allowing alignments to include introns. Other alignment models which may be expressed by a regular grammar include those allowing non-equivalenced regions such as the ABC model [ 9 ] for improved modelling of divergent loops regions in protein alignments, and DNA Block Aligner [ 10 ] which finds co-linear conserved blocks in the alignment of genomic sequences. This framework also allows models such as that used by PairWise [ 11 ] where the sequences are translated during alignment allowing for frameshifts, and GeneWise [ 12 ] which integrates translated alignments with modelling of introns for alignment of proteins against genomic DNA. The availability of vast amounts of sequence data has generated a need for faster alignment algorithms [ 13 ]. Many of these approaches use fast algorithms to identify closely matching words which seed un-gapped alignments that are subsequently joined to form the final gapped alignment. For example, BLAST [ 3 ], FASTA [ 14 ], and sim4 [ 15 ] all operate by first finding matching words before building alignments. Such word finding can be done by a multitude of techniques, such as finite state machines used in BLAST [ 3 ], table-lookup used in SSAHA [ 16 ], or by suffix arrays as used in QUASAR [ 17 ]. Novel methods for word-based seeding of these alignments are not presented here, but rather a general system of joining these seeds to produce gapped alignments. The alignment program FASTA [ 14 ], generates alignments from sets of seeds by performing DP confined to a diagonal band surrounding the initial matches [ 18 ]. In contrast, when building alignments, Gapped BLAST [ 19 ] permits gapped extension which allows the DP to be applied to an arbitrary high-scoring region surrounding the HSP seeds. Such heuristics allow very fast sequence alignment, but it is difficult to apply to more complex models. Furthermore, features such as introns cannot easily be incorporated into the resulting alignments without necessitating a large amount of DP to ensure that both very short exons and large introns can be included in alignments. The aim here is to combine the strengths of Gapped BLAST and Dynamite, in a system which allows automated generation of heuristics in terms of the underlying model. This allows the development of heuristics for complex models such as those used in GeneWise. Firstly, we describe bounded sparse dynamic programming (BSDP), a novel heuristic for sequence alignment, then we describe the system for automated implementation of the complex alignment algorithms required. We present proof of principle results, but this paper is primarily focussed on developing a framework for the general case. Implementation The basic strategy of seeding alignments used here is the same as for BLAST, in that alignments seeds are generated, and then extended to form High-scoring Segment Pairs (HSPs), which are then joined together to form the alignments. The alignments are seeded using an Aho and Corasick [ 20 ] type finite state machine (FSM) built using the word neighbourhood of the query sequence. This generates the seeds which are extended to form the HSPs. For large scale analyses, the FSM is multiplexed using word neighbourhoods from multiple sequences. This allows analysis of multiple queries during a single pass of a genomic database, in a manner similar to that used in MPBLAST [ 21 ]. However the methods for seeding HSPs are independent from those used for building the alignments, and this paper only deals with algorithms involved in the generation of gapped alignments from sets of HSPs, and not in the calculation of the HSPs themselves. The following strategies are employed to enable alignments to be built efficiently from sets of HSPs: • To connect the underlying alignment model to the heuristics, a portal describes a set of states in the model which correspond to a set of HSPs, a span refers to a looping state for large alignment features such as introns, and a SAR (Sub-Alignment Region) describes a rectangular region on an HSP to which DP is applied. • To avoid DP in every SAR, upper bounds are generated for the best alignment score for each SAR, and BSDP (Bounded Sparse Dynamic Programming) exploits these bounds to yield alignments in an efficient manner. • To perform various types of DP in these SARs, the required models are generated automati cally, including C code to produce an efficient viterbi implementation for each model. Bounded sparse dynamic programming Dynamic programming (for any alignment model which can be represented by a regular grammar) requires quadratic time, and hence is the most computationally expensive part of building an alignment. For pairs of sequences more than a few kilobases long, DP becomes prohibitively slow. The approach used here is similar to sparse dynamic programming [ 22 ], and the fragment chaining approaches used in the program sim2 [ 23 ], in that DP is applied to rectangular regions surrounding alignment seeds. However, there are two major differences in our approach. Firstly, DP is only applied to two small discrete regions on each HSP, as it is assumed that most of the HSP itself should appear in the alignment. These sub-alignments improve the quality of the overall alignments, and they allow complex alignment models, and large gaps such as introns to be integrated into a sparse DP framework. Secondly, as it would take too long to apply DP to every sub-alignment region (SAR), upper bounds are calculated for the DP scores for each of these SARs. This allows the sub-alignment DP to be avoided in cases where it joins HSPs which cannot feature in the final alignment, so that alignments can still be generated very rapidly even when large numbers of HSPs and SARs are involved. Before the BSDP can be performed, a single point on each HSP is selected which will feature in any alignment generated using that HSP. This point corresponds to a pair of equivalenced symbols which must feature in any alignment to include that HSP. A point is chosen where half the HSP score is generated by equivalenced symbols on either side of it, as this is likely to be in the highest quality portion of the HSP. As shown in the example in Figure 1 , this strategy is particularly beneficial where one end of the HSP has a much lower quality than the other. The five types of region used for sub-alignments are classified in Figure 2 . Each of these require a slightly different alignment algorithm. The alignment path must meet corners of the SARs that contain an HSP, so that the sub-alignments can be integrated with the HSPs to produce the final alignment. This approach has been primarily designed for local models, but BSDP may also be used for global and semi-global models, in which case constraints are added such that both the terminal regions (cases A and E in Figure 2 ) and the resulting sub-alignments must contain the relevant sequence ends. In the case of C and D, the two HSPs and their SARs are separated by a span, allowing large gaps or introns in the alignment. In these cases DP must be applied to the SAR before the span, and the end state scores must be integrated in an intermediate matrix before being provided as start state scores for the DP in the SAR after the span. Regions for the sub-alignments are selected within the area between the centre points of the two HSPs to be joined, or in the case of terminal HSPs, between the HSP and the ends of the sequences. In addition, the positions of the SARs must be constrained to limit their size, and so that the HSPs correctly intersect with the corners of the SAR. In the case of overlapping HSPs, where there is a choice of positions for placement of the SAR, the position is chosen such that the highest scoring parts of the HSPs are outside the SARs. Once the SARs have been selected, an upper bound is placed on the score for each sub-alignment. The calculation of these bounds is described in a later section describing automation of this method. The BSDP approach can then utilise these upper bounds to avoid application of DP in some SARs, as demonstrated by the example in Figure 3 . In the case of a real alignment, a much greater number of HSPs will be involved, so the amount of DP avoided will be larger. The BSDP is mediated through a set of priority queues, one of which is associated with each HSP, and will contain an entry for each partial alignment that ends at the HSP. The key for these priority queues is the highest score for any partial alignment ending in that HSP. Additionally, there is one global priority queue containing the highest scores from each of the other priority queues. The upper bound scores are confirmed by DP in the SARs in the current highest scoring putative alignment path. The highest scoring path will change as the scores are updated during this process. DP is applied to SARs in this way until the highest scoring path does not contain any bound scores, and then the alignment may be extracted. Upper bounds dictate that there can be no better alignment using these HSPs. This algorithm is similar to A* search, (but differs in that many different points may be the start or end of the search), and retains the admissibility property of A* search, such that the result of the BSDP computation is guaranteed to be the same as if DP had been performed on every candidate SAR prior to calculation of the alignment. This is because no alignment can be extracted until all the alternative sub-alignments (which have upper bounds that indicate they could contribute to a higher-score) have been eliminated. Suboptimal alignments The BSDP alignment process can be iterated to generate sub-optimal alignments similar to those generated by the Waterman-Eggert algorithm [ 24 ], with only minimal recalculation of the partial alignments in the SARs. Each HSP may only appear in a single alignment, but further constraints are required to prevent overlapping alignments arising from overlapping SARs. The likelihood of this is occurring is greatly increased during translated alignments when HSPs in different reading frames may overlap each other. After the first alignment has been reported, the scores for any SARs which have already been confirmed by DP, but which are not yet included in a reported alignment, are then considered as an upper bound. SARs are disallowed before recalculation when the region between the centre of the HSP and its SARs overlaps with a previously reported HSP, in which case, the SARs are disallowed. Otherwise the DP is recomputed for SARs which contain part of an alignment which has been reported since the DP for the SAR was last calculated. Automated model generation As illustrated in the previous section, BSDP becomes quite complex and requires a large number of DP algorithms for computation of the alignment through the SARs. We have build a system to facilitate implementation of these models and the integration of the sub-alignments which they produce. To allow generalisation of the BSDP, everything must be defined in terms of the underlying alignment models. The alignment models are described as finite state machines, consisting of states and transitions, similar to those used in Dynamite [ 6 ]. Briefly, to convert these models into DP implementations, each state must correspond to a score in each cell of the DP matrix, and the scores for each cell are calculated by taking the maximum of the score from transitions arriving at each cell. In addition, a topological sort is required to satisfy dependency ordering for silent states. However, in addition to automated generation of code from alignment models, the generation of the models required for DP in the SARs is also automated, as described below. Building simple models Table 1 shows a few example alignment models which are generated by this system. The models are built in a modular fashion, allowing reuse of common components such as intron models and gap models. These models may be used for exhaustive alignment in quadratic time, but in order to use them for heuristic alignment, manipulations of the models are necessary to perform DP in the SARs, as detailed in the following sections. Building the heuristic model To enable DP in the SARs for the BSDP, a heuristic model is generated from the original alignment model. This model is not used directly for calculation of alignments, but a derived model is generated corresponding to each transition in the heuristic model. Each of these derived models correspond to a type of SAR used in the BSDP. The model is first annotated, labelling certain states as either portal states or span states. A portal defines a group of states which can share a set of HSPs (High-scoring Segment Pairs); these are the match states. A span is a state which has sequence independent looping transitions ( e.g. states for introns, or non-equivalenced regions). The heuristic model is build using states corresponding to each portal and span state, with transitions between these states in cases where there is a valid path between the corresponding states in the original model. An example model is shown in Figure 4 for alignment of ESTs to genomic DNA. In this example, there is a portal which corresponds to the match forward and match reverse states, and the intron forward and intron reverse states are span states. Building derived models Derived models are produced for the DP in SARs automatically from each transition in the heuristic model. The source and destination states from each transition in the heuristic model become the start and end states in each derived model. All reachable states and transitions from these states in the original model are recursively copied to the derived model. An example of this process is shown in Figure 5 . The relationships between the states and transitions in the derived model and the original model are tracked to allow conversion of the partial alignments from the derived models back to complete alignment in the original model. Terminal models (case A and E in Figure 2 ) are generated between portal states and a start or end state. Join models (case B in Figure 2 ) are generated between portal states, including from a portal state to itself. Span models (case C and D in Figure 2 ) are generated from a portal state to a span state, and vice versa. These allow incorporation of a large feature such as an intron into an alignment. The span models require a special end state in the model at the start of the span, and a special start start in the model at the end of a span, so that scores can be transferred from one DP matrix to the other via an intermediary score matrix. For some cases, such as between the match forward and match reverse states, shown in Figure 4 , there is no possible path, and no corresponding transition in the heuristic model, in which case, a derived model is not produced. Ten different derived models are generated from the model shown for cDNA to genomic alignment in Figure 4 , because there are two portal states and two span states in this model, and therefore, a derived model is generated for each of the five cases in Figure 2 for both the forward and reversed genes. Building models for calculation of upper bounds For each derived model, an additional model is created which is used for the pre-calculation of upper bounds for all possible sizes of SARs. For each transition in the model which has a position dependent score associated with it, the upper bound is also supplied. For example, in a match transition, the upper bound is the maximum value from the substitution matrix. A special model is created using these transition upper bounds, instead of the normal sequence-dependent transition weights. As this removes the sequence-dependent components of the algorithm, it allows pre-calculation of the upper bound for alignment score of any sequences up to the maximum permitted size of SARs. These bounds are then stored in a table for retrieval during the BSDP. Results The model used in Figure 4 was used with this system in the program exonerate for rapid comparison of cDNAs and genomic DNA. This model was used to compare a test set of genomic sequences to mRNAs extracted according to the Ensembl annotations. exonerate was compared to sim4 [ 15 ], another heuristic for comparing cDNAs to genomic DNA, and EST_GENOME [ 8 ], which is implements the exhaustive DP for essentially the same underlying model. As can be seen from the results of this comparison are shown in Table 2 , similar results are produced by this method, but in much less time. This system is used for alignment of ESTs within the Ensembl gene-building pipeline [ 25 ], and a prototype implementation of this system has been used in comparison mouse and human sequences [ 26 ]. In this analysis, 13 million raw shotgun mouse reads were aligned as part of the comparative analysis of chromosome 20. Figure 6 shows an example the SARs generated by BSDP. Only a small proportion of these SARs will require confirmation by DP. This approach scales well – in a larger example using Titin, 27346 candidate SARs are generated, of which only 236 are confirmed and 206 appear in the final alignment. In many cases, BSDP obviates the use of DP altogether as there will be no candidate alignment with a total upper bound score over the threshold. This is especially likely to be true in cases when a high score threshold is used, such as searching for an alignment covering a certain percentage of the query, or when a dynamic score threshold is used in the search for the best few hits from a particular query to a database. Such dynamic thresholds can actually cause the search to speed up as it progresses, once good matches to the query have been found. Another example is the alignment of the Collagen alpha 1(IX) precursor to a region of chromosome 6 containing the corresponding gene. This is a large gene, containing 38 exons over about a 90 Kb region of the genome. Using GeneWise, the alignnment required 4260 seconds (about 1 hour, 11 minutes) using exonerate to perform full DP alignment required 2700 seconds (about 45 minutes), and using exonerate with the heuristic BSDP approach required only 7 seconds, with all three methods generating identical alignments. Conclusion This approach represents an advance from previous methods for automatic implementation of sequence alignment algorithms [ 4 - 6 ], in that it is not just the generation of code from the models which is automated, but also the generation of many of the models themselves. This has allowed development of heuristics with sub-quadratic running times. This makes their application practical to a much larger set of problems, while retaining the much of the simplicity of their implementation. We recognise that this approach is limited to the subset of sequence comparison problems which can be represented by a regular grammar. For example, it is not possible to use this system to model stochastic context free grammars as used in some types of RNA secondary structure such as pseudo-knots [ 27 ]. This approach is also unsuitable for the types of DP algorithms used for determining optimal segmentation in gene finding algorithms [ 28 ], however for these types of problems running time of the algorithm is not an issue. This framework has already been used for many different alignment models ranging from simple models such as Smith-Waterman-Gotoh, to more complex models such as those for protein to genome alignment as used by GeneWise [ 12 ]. It is also well suited to problems requiring comparison of very long sequences, and we are currently extending this system to accommodate syntenic pairwise comparisons of genomic sequences of the type tackled by MUMmer [ 29 ]. Although this method is useful in many cases, for some types of distantly related sequences, the method can breakdown. For example, when long regions of the correct alignment contain many gaps with intervening sequences too short to be an alignment seed, the alignment cannot be extended beyond the boundary of the SAR, and the correct alignment may be missed. These cases can be avoided by increasing the size of the SARs, but this results in higher bound scores, so DP become necessary in more of the SARs. Such gap rich alignment are the type where the gapped HSP extension used by Gapped BLAST excels, however such gapped extension necessitates DP surrounding all HSPs, which becomes time consuming when allowing both short exons and long introns during alignments of cDNAs or proteins to genomic sequences. As this system separates the development of the underlying alignment model from the heuristics which are built on top of them, we expect that this framework will prove useful for evaluation of the quality of the heuristics, as comparison between the alignments from the two techniques can be automated, and training may be performed to select optimal parameter sets. Availability The implementation of the system described here is called C4 (to reflect the aim of producing something more powerful than Dynamite [ 6 ]), and is implemented in C programming language. It is available as part of the exonerate sequence alignment package available from , and in the exonerate module of the Ensembl CVS repository from . It is built using the glib portabilility library available from Both exonerate and glib are available under the GNU lesser general public license. The code has been tested extensively on Linux and OSF1/alpha but has been written to be portable to many UNIX systems. Code generation for the DP is performed to exploit compile time optimisations such as loop un- rolling and and efficient handling of edge conditions in the viterbi matrix, which is particularly beneficial for the small DP calculations required by SARs. This generated code is typically about five times faster than using a generic viterbi implementation, but produces about a million lines of code for the current set of models used by exonerate, and hence compilation takes longer.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553969.xml
554974	Socioeconomic factors and low birth weight in Mexico	Background Low birth weight (LBW) is a public health problem linked to lack of equity in populations. Despite efforts to decrease the proportion of newborns with LBW, success has been quite limited. In recent years, studies focused on explaining how social factors influence this problem have shown that populations with greater inequities have a greater proportion of newborns with LBW. Methods The objective was to describe socioeconomic factors related to LBW adjusted by demographic, reproductive and health services variables in Mexico City. A case-control study was carried out in three hospitals with gynaecological and obstetrics services in Mexico City during the first half of 1996. During the recruiting period all children with LBW (cases), defined as newborns weighing <2500 grams, were matched with children born on the same day to control for time of birth. Upon arrival at the hospital for delivery, women were interviewed to determine if they met our inclusion criteria. Women with a history of chronic conditions and those with twin or multiple pregnancies were excluded. Variables with clinical and statistical significance were included in a multivariate model (logistic regression). Results We found that low socioeconomic level was the most important risk factor for LBW and was independent of other factors, including those related to reproduction and nutrition, smoking, morbidity during pregnancy, accessibility to health services and prenatal care (OR 2.68; 95% CI 1.19, 6.03). Conclusion We found that socioeconomic factors are relevant to LBW. However further research should be done in different population groups as well as developing precise ways of measuring socioeconomic factors and their role in reproductive health.	Background On average, the worldwide incidence of low birthweight (LBW) is 17% per year, making LBW an important infant health problem in many populations [ 1 ]. The incidence of LBW varies among countries, ranging from 4% to 6% in Western countries like Sweden, France, United States and Canada (UNICEF 2003) and much higher in developing countries. In Latin America, the overall LBW rate varies according to geographical region. The Pan American Health Organization has estimated that the overall LBW rate is 8.27% in South America and Mexico and ranges from 6% in Peru to 10% in Bolivia and Venezuela. In Central America the overall LBW rate is 11.2%, ranging from 6% in Belize and El Salvador to 15% in Nicaragua [ 2 ]. The LBW rate can also vary within countries. For example, in Mexico, the incidence ranges from 8.2% to 12%, depending on geographic al region [ 3 ]. In addition to its impact on infant mortality, LBW has been associated with higher probabilities of infection, malnutrition and handicapping conditions during childhood, including cerebral palsy, mental deficiencies and problems related to behavior and learning during childhood [ 4 - 6 ]. Children who survive LBW have a higher incidence of diseases, retardation in cognitive development, and undernourishment. There is also evidence that LBW or its determinant factors are associated with a predisposition to higher rates of diabetes, cardiac diseases and other future chronic health problems [ 7 - 9 ]. The biological processes that affect the fetus in utero are related to the mother's physiology, including her nutrition (mother's weight before pregnancy and history of having newborns with LBW), exercise, infections, and consumption of tobacco, alcohol and other drugs [ 10 , 11 ]. During the fetal phase, growth depends on the nutritional condition of the mother, indicating that pregnant women should not only increase their weight but also consume essential nutrients. For many women in the developing world, however, economic, social and cultural factors make it difficult for them to obtain the necessary food and healthcare, which are closely interrelated [ 12 ]. Associations between poorer child health and poverty, inequity and social exclusion have been documented worldwide and have been shown to be independent of research methods, local culture, and available health care services [ 13 ]. While the relationship between socio-economic conditions and health have been of interest and concern for centuries, recent studies have sought to identify the social factors most relevant to health [ 14 ]. Some authors consider that health therefore may be an important determinant of opportunities in life and this process, termed "selection by health", and suggests that health "selects" people in different social strata [ 15 , 16 ]. Among the socio-economic factors are income, education, occupation, household leadership and gender differences related to roles within the family [ 17 , 18 ]. In Mexico, there are also differences in socio-economic conditions resulting from geographic area and political organization. These may affect various health parameters, including mortality, morbidity and reproductive behavior. Several studies have shown different results on whether socioeconomic factors affect pregnancy outcomes and newborn conditions [ 19 - 21 ]. The inconsistency of these findings may be due to poor clarification of the mechanisms by which socioeconomic status affects LBW. This is especially true in relationship to the mother's nutritional conditions, although low maternal weight before pregnancy and small weight gain during pregnancy have been shown to be associated with higher risks of preterm infants and LBW [ 22 - 25 ]. Other studies have questioned whether maternal occupation or educational level is associated with LBW, or whether the latter is related to a group of socioeconomic factors. In this study, we have analyzed the socio-economic factors related to birthweight adjusted by others known factors in the urban population of Mexico in three hospitals located in three different geographic areas of Mexico City. Methods A case-control study was carried out in three hospitals with gynaecological and obstetrics services in Mexico City during the first semester of 1996. We recruited 154 LBW newborns, defined as newborns weighing <2500 grams, from these three hospitals, as well as 474 controls chosen from births on the same day to control for time of birth. All newborns included in this study provided a statistical power of 80% when we assumed a = 0.05 (one-sided test), a case/control ratio 1:3, and a 3% difference among both groups in the low socioeconomic level [ 26 ]. Upon arrival at the hospital for delivery, women were interviewed to determine if they met our inclusion criteria. Women with a history of previous chronic conditions or those with twin or multiple pregnancies were not included. Following delivery, and after obtaining informed consent, each woman was administered a questionnaire by a trained interviewer to obtain information about socio-economic, reproductive, and nutritional factors. Socioeconomic factors included age, level of education, civil status, occupation, income and owning certain goods; reproductive factors included parity, history of preterm delivery and LBW, which were classified as positive or negative for all previous pregnancies; and nutritional factors included calcium and iron supplementation, pre-gestational weight, prenatal care, morbidity during pregnancy and tobacco exposure. Clinical records were also reviewed to verify information about each newborn. To create a socioeconomic level index, we used two variables that have been considered as proxies [ 27 ]. Ownership of goods was defined as to whether a woman owned her house or flat and if she had a car and whether the woman and her partner were employed. Using these two variables, we constructed a three category socio-economic level index, in which High indicated that the woman and her partner had jobs and goods; Medium indicated that the woman or her partner had jobs or goods; and Low indicated that the woman and/or her partner did not have jobs or goods. Statistical analysis was performed by describing sociodemographic, reproductive and prenatal care characteristics of the mother and newborn. Univariate analysis was used to evaluate the association between the independent variable (socioeconomic level) and covariates with LBW (outcome variable). Stratified Mantel-Haenszel analysis was performed to evaluate confounding and/or interaction (i.e. parity and age). To obtain the association magnitude of the socioeconomic level adjusted by the covariates, multivariate analysis was performed (logistic regression). Variables with clinical and statistical significance were included in multivariate modeling. We used three models to classify factors relevant to LBW into three groups as known, controversial and unknown risk factors for LBW. Model 1 included socioeconomic level index and maternal age, maternal education, marital status and accessibility to public services. In model 2 we included all variables of the model 1 besides, prenatal care, tobacco exposure, and morbidity during pregnancy were considered as adjusted variables. Finally, model 3 included all variables previously mentioned (model 1 and model 2 besides reproductive variables. Analyses were performed using StataCorp. 2002. Stata Statistical Software: Release 7.0 College Station, TX:Stata Corporation. Results The sociodemographic characteristics of the LBW and normal birthweight groups are shown in Table 1 . There was a high proportion of married women in both groups, and 90% of women in both groups had access to public services, including electricity, water and a sewage system. The effect of maternal education had the expected direction, although it was not statistically significant. Women in the lower socio-economic level had a higher risk for LBW (OR, 2.19; 95% CI, 1.18–4.07) than those in the medium and high socioeconomic levels. Table 1 Sociodemographic characteristics of mothers of low and normal birthweight infants Variable Birthweight OR 95% CI <2500 grs ( n = 158 ) ≥ 2500 grs ( n = 474 ) Socioeconomic level High 81 267 1.0 Medium 57 177 1.06 [0.71,1.56] Low 20 30 2.19* [1.18,4.07] Maternal age (years) <19 22 83 0.77 [0.46,1.29] 20–30 110 321 1.0 >30 26 70 1.08 [0.65,1.80] Maternal education (years) >12 13 50 1.0 10–12 56 143 1.50 [0.76,2.98] 7–9 52 165 1.21 [0.61,2.40] <7 37 116 1.22 [0.60,2.50] Marital Status Married 138 429 1.0 Unmarried 20 45 1.38 [0.78,2.42] Accessibility to public services No 12 51 0.68 [0.35,1.31] Yes 146 423 1.0 * p < 0.05 We observed no significant between group differences in tobacco exposure prior to or during pregnancy, or in the frequency of urinary tract infections (Table 2 ). However hypertension (OR, 1.53; 95% CI, 0.93–2.53) and calcium supplementation (OR, 1.86; 95% CI, 0.97–3.56) during pregnancy were marginally significant. Table 2 Tobacco exposure, morbidity during pregnancy and prenatal care characteristics of mothers of low and normal birthweight infants Variable Birthweight OR 95% CI <2500 grs ( n = 158 ) ≥ 2500 grs ( n = 474 ) Smoking before pregnancy No 126 384 1.0 Yes 32 90 1.09 [0.70,1.71] Smoking during pregnancy No 149 451 1.0 Yes 9 23 1.19 [0.53,2.66] Hypertension during pregnancy No 131 416 1.0 Yes 27 58 1.53 [0.93,2.53] Urinary tract infection during pregnancy No 100 318 1.0 Yes 58 156 1.22 [0.83,1.77] Calcium supplementation Yes 146 409 1.0 No 12 65 1.86 [0.97,3.56] Iron supplementation Yes 64 216 1.0 No 94 258 1.20 [0.83,1.74] Prenatal care Yes 150 454 1.0 No 8 20 1.07 [0.45,2.52] Reproductive characteristics, including maternal weight, previous preterm-birth, and previous LBW infants, were significantly different between mothers of LBW and normal birthweight infants (Table 3 ). Table 3 Reproductive characteristicsof mothers of low and normal birthweight infants Variable Birthweight OR 95% CI <2500 grs ( n = 158 ) ≥ 2500 grs ( n = 474 ) Parity Multiparous 92 284 1.0 Primiparous 66 190 1.04 [0.73,1.51] Maternal weight (kgs) <48 54 107 1.61* [1.09,3.84] 49–54 37 120 0.99 [0.58,1.67] 55–60 37 119 1.0 >60 30 128 0.75 [0.43,1.29] Previous a preterm birth No 58 213 1.0 Yes 34 71 1.75* [1.06,2.90] Previous low birth weight a No 66 246 1.0 Yes 26 38 2.55* [1.44,4.50] Previous abortion a No 67 222 1.0 Yes 25 62 1.33 [0.77,2.28] p <0.05 a Only applicable for women with more than one pregnancy Multivariate analysis showed that low socioeconomic level was the most important risk factor for LBW (Table 4 ). Using model 1, the OR was 2.57 (95% CI, 1.19–5.58) after adjustments for maternal age, education, marital status, and accessibility to public services. When adjustments for prenatal care, tobacco exposure, and morbidity during pregnancy were also included (model 2), we obtained an OR of 2.45 (95% CI, 1.13–5.36) for low socioeconomic level as a risk factor for LBW (Table 4 ). When we also included adjustments for reproductive variables (model 3), we obtained an OR of 2.68 (95% CI, 1.19–6.03) for low socioeconomic level as a risk factor for LBW (Table 4 ). Table 4 Adjusted odds ratios for LBW infants by socioeconomic level among women of Mexico City Variable Model 1 Model 2 Model 3 OR 95% CI OR 95% CI OR 95% CI Socioeconomic level Medium 1.14 [0.75,1.74] 1.16 [0.76,1.77] 1.12 [0.72,1.74] Low 2.57 [1.19,5.58] 2.45 [1.13,5.36] 2.68 [1.19,6.03] Maternal age (years) <19 0.74 [0.43,1.25] 0.74 [0.43,1.26] 0.57 [0.32,1.03] >30 1.17 [0.70,1.98] 1.09 [0.64,1.84] 1.34 [0.76,2.38] Maternal education (years) < 7 1.21 [0.58,2.56] 1.11 [0.52,2.40] 1.05 [0.47,2.34] 7–9 1.29 [0.63,2.62] 1.23 [0.59,2.55] 1.39 [0.66,2.94] 10–12 1.56 [0.78,3.13] 1.61 [0.79,3.27] 1.82 [0.87,3.77] Unmarried 0.99 [0.49,2.01] 1.08 [0.52,2.28] 0.96 [0.44,2.08] No accessibility to public services 0.68 [0.34,1.33] 0.58 [0.29,1.17] 0.56 [0.27,1.14] Smoking before pregnancy 0.99 [0.57,1.72] 1.07 [0.60,1.90] Smoking during pregnancy 1.09 [0.43,2.78] 1.14 [0.43,2.99] Hypertension 1.61 [0.95,2.73] 1.55 [0.90,2.68] Urinary tract infection 1.23 [0.83,1.83] 1.14 [0.76,1.71] No calcium supplementation 1.98 [1.01,3.87] 2.30 [1.14,4.63] No iron supplementation 1.23 [0.84,1.80] 1.32 [0.89,1.96] No prenatal care 1.15 [0.47,2.87] 1.27 [0.52,3.19] Primiparous 1.73 [1.05,2.83] Pre-gestational weight(kgs) <48 2.33 [1.33,4.08] 49–54 1.28 [0.71,2.31] >60 1.32 [0.74,2.36] Previous preterm birth a 2.95 [1.04,8.38]* Previous low birth weight a 2.61 [1.36,5.04] Previous abortion a 0.52 [0.17,1.63] a only applicable for women with more than one pregnancy. Discussion LBW is a public health problem linked to lack of equity in populations. Despite efforts to decrease the proportion of newborns with LBW, success has been quite limited, and the problem persists in both developing and developed countries [ 28 ]. In recent years, studies focused on explaining how social factors influence this problem have shown that populations with greater inequities have a greater proportion of newborns with LBW [ 25 ]. These inequities are caused by both social conditions of populations and gender differences. Although these differences are not explicit and conclusive, they are revealed by social indicators such as access to health care services, occupation, income, education, and social exclusion or isolation. Differences found in the results about the effect of socioeconomic factors on LBW are probably due to the use of different socioeconomic indicators. It should be pointed out, however, that obtaining information that accurately reflects social and economic characteristics can be difficult, leading to the generation of proxy variables. Thus, education has been used as a proxy variable of social class, and occupation has been used as a proxy of socioeconomic status [ 29 ]. In addition, studies performed in European countries have used education as a proxy for socioeconomic level [ 29 , 30 ]. A recent Mexican study, however, found no association between LBW and household infrastructure (including lack of indoor sanitation or water facilities and lack of electricity) when used as a socioeconomic indicator, although this study found that LBW was related mainly to access or utilization of prenatal services and disadvantageous maternal lifestyle behaviors [ 31 ]. In our study, we used ownership of goods and having a job as indicators of socioeconomic level in a Mexican urban population with social security. Using these indicators, we found that low socioeconomic level is the most important risk factor for LBW, independent of other factors such as reproductive and nutritional characteristics, smoking, morbidity during pregnancy, and accessibility to health services and prenatal care. These results are similar to those of other studies describing a positive relationship between socioeconomic condition and effects on health [ 32 - 34 ] It is important to mention that to diminish the possible error of misclassifying we used at least two indicators to construct the socioeconomic level variable (ownership of goods and occupation). However we thought that it will be a non differential error. Although we analyzed other proxies of socioeconomic level, we found that each of them was irrelevant. For example, although education has been used as a proxy of social status, we found that this indicator was not important in our population, perhaps because the women in our study were incorporated into the work force, making exposure to intermediate factors, such as occupational stress and load work, an influence on pregnancy outcomes. Family circumstances and biological processes may also be affected by a wider social context, including cultural and historic issues such as educational opportunities, parent's divorce, unemployment, risk of poverty risk, and risks factors for smoking and obesity. Although many socioeconomic factors related to LBW have been identified, the specific role of each of them is not known, limiting the ability to use preventive actions in exposed populations. Interventions aimed at reducing the number of LBW infants have had limited success on conditions of the newborn, although some showed benefits in pregnant women [ 35 ]. To decrease the incidence of LBW, it is important to consider health services interventions to get better quality of care for pregnant women. Investments in the health and development of the most vulnerable populations, such as pregnant women and children, are important in themselves, because they prepare the context and environment for a more productive and healthy life, with the full development and use of mental and physical human potential [ 36 ]. Conclusion Heterogeneity among different populations makes findings related to interventions in one population not applicable to others. Thus, it is necessary to design studies that account for the geographic, racial, cultural, social and economic context of each country and specific group. In the case of Mexico, we believe that our results provide information representing a closer approach to the effect of socioeconomic status on LBW. Our results provide a starting point in the search for better indicators for evaluating socioeconomic status in Mexican populations with other social conditions, including suburban, rural, and indigenous populations. Another aspect to consider within the context of each country is the availability of information. In Mexico, there is no way to obtain uniform information about the socioeconomic characteristics of the population. Information about conditions of different populations is important in designing programs aimed at solving existing inequities. It is a challenge to create a socioeconomic index that will reflect the real living conditions of pregnant women in Mexico. This effort will require including other variables, as well as refining proxy variables. Competing interests The author(s) declare that they have no competing interests. Authors' contributions LPTA, PCC, SFH contributed in the conception and design of the study and statistical analysis. JPVB, ERM reviewed for important intellectual content. All authors participated in the interpretation of data and read and approved the final version to be published. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554974.xml
514613	Insights into the pathogenesis of vein graft disease: lessons from intravascular ultrasound	The success of coronary artery bypass grafting (CABG) is limited by poor long-term graft patency. Saphenous vein is used in the vast majority of CABG operations, although 15% are occluded at one year with as many as 50% occluded at 10 years due to progressive graft atherosclerosis. Intravascular ultrasound (IVUS) has greatly increased our understanding of this process. IVUS studies have shown that early wall thickening and adaptive remodeling of vein grafts occurs within the first few weeks post implantation, with these changes stabilising in angiographically normal vein grafts after six months. Early changes predispose to later atherosclerosis with occlusive plaque detectable in vein grafts within the first year. Both expansive and constrictive remodelling is present in diseased vein grafts, where the latter contributes significantly to occlusive disease. These findings correlate closely with experimental and clinicopathological studies and help define the windows for prevention, intervention or plaque stabilisation strategies. IVUS is also the natural tool for evaluating the effectiveness of pharmacological and other treatments that may prevent or slow the progression of vein graft disease in clinical trials.	Introduction The success of coronary artery bypass grafting (CABG), although the gold standard for the treatment of multivessel coronary artery disease, is limited by poor long-term vein graft patency [ 1 ]. Early vein graft thrombosis (within 1 month) occurs in up to 15% of vein grafts due to graft spasm or technical error [ 2 , 3 ], whilst late vein graft failure occurs as a consequence of early neointimal hyperplasia with later superimposed atheroma, so called 'vein graft disease' [ 1 , 4 ] and as many as 50% of all vein grafts are occluded at 10 years post surgery [ 5 , 6 ]. Despite the superiority of arterial graft patency over that of vein grafts, the multivessel nature of coronary artery disease and ready availability of saphenous vein still result in its use in over 70% of CABG procedures [ 7 ]. The alternative treatment modality for multivessel coronary atheroma, percutaneous coronary artery angioplasty and stenting (PCI), has traditionally suffered from even worse long-term results compared to CABG, due to high early restenosis rates; over 30% within 1 year [ 8 ]. This results in more frequent and more rapid return of symptoms and major adverse cardiac events (MACE) with PCI compared to CABG, necessitating more repeat revascularisation procedures [ 8 ]. The apparent success of new drug eluting stents has challenged this paradigm however. Rapamycin (a macrolide antibiotic) and the taxane, paclitaxel, two agents with potent antiproliferative properties, eluted from intracoronary stents, have dramatically reduced restenosis rates, MACE and reintervention rates in clinical trials [ 9 , 10 ], to the point where the superiority of CABG is now being seriously challenged [ 11 ]. This represents the clinical application of intensive research into the mechanisms of atherosclerosis and restenosis and strategies for their prevention over the last decade. Conversely, CABG has suffered from its apparent success, and with some exceptions [ 12 ], there have been comparatively few attempts to prevent or inhibit the progression of vein graft disease in CABG patients, a condition that must change. Vein graft disease differs from arterial atherosclerosis in that its natural history is much shorter and the date of onset is clearly defined, i.e. graft implantation. This process is therefore potentially amenable to strategies that may inhibit its progression. Although the cellular and molecular mechanisms underlying vein graft disease have been systematically investigated, the time course and development of this process in patients after coronary bypass has only recently been defined as a consequence of the increasing use of intravascular ultrasound (IVUS). Quantitative coronary angiography, traditionally the predominant imaging modality used to assess the severity of vein graft disease underestimates the severity of vein graft remodeling and athermanous plaque development by measuring the vessel lumen in only two dimensions [ 13 , 14 ]. In contrast, the tomographic IVUS image enables visualization of the full circumference of the vessel wall [ 15 ] allowing measurement of wall thickening, vein remodeling and atherosclerotic plaque size, distribution, and composition [ 15 ]. This results in the detection of diffuse atherosclerotic plaque, compensatory vessel enlargement and preservation of the luminal diameter even in angiographically normal vessels [ 13 , 14 , 16 ]. IVUS findings in vein grafts also show good correlation with histological findings in clinicopathological studies [ 17 ]. The purpose of this review is to summarize our current understanding of the natural history of vein graft disease from IVUS studies, correlate this with the findings of experimental and clinicopathological studies, and, finally to consider how this knowledge, may be used to target prevention or treatment strategies. Early changes in saphenous vein bypass grafts; wall thickening and adaptive remodelling Glagovian remodeling was first described as a radial enlargement of the entire cross sectional area of a vessel in response to intramural atheroma [ 18 ]. First identified in humans in post-mortem studies it was only with the widespread use of IVUS that its central role in atherosclerosis, post angioplasty restenosis, transplant vasculopathy and vein graft disease was realised [ 15 ]. Currently, the term not only applies to vessel enlargement, but also shrinkage, where in the presence of underlying plaque it becomes an important determinant of lumen loss [ 19 ]. In vein grafts, early after implantation, increases in overall vessel cross sectional area preserve luminal size despite significant increases in wall thickness [ 16 ]. IVUS measured parameters of vessel remodeling and wall thickening in vein grafts pre, or early post implantation versus later periods are described in Table 1 . Vein graft dimensions within one month of implantation are remarkably similar to those in grafts prior to implantation [ 16 , 20 , 24 ], with significant wall thickening having occurred by six months, even in angiographically normal grafts [ 14 , 16 ]. Other studies have demonstrated wall thickening as early as 3 weeks to 3 months after CABG [ 23 , 24 ]. These changes are diffuse and concentric, and are observed from the aortic root to the coronary anastomosis [ 23 , 24 ]. Higuchi and colleagues compared IVUS measurements of 15 vein grafts performed within 1 month postoperatively with 14 vein grafts after 6 months postoperatively. This showed that significant wall thickening had occurred by 6-months, accompanied by compensatory enlargement, and preservation of the graft luminal diameter (Table 1 ), however wall thickening appeared to reach a plateau after 6 months with preservation of lumen area (Figure 1 ), suggesting that early remodeling responses may stabilize in the absence of atherosclerotic development [ 16 ]. Figure 1 Increases in wall thickness versus time after surgery. Wall area expressed as a percentage of total vessel area (%VWA) exceeded 40% and reached a plateau state after 6 months in angiographically normal vessels. Reproduced with permission from Higuchi et al, Heart Vessels 2002, 17:57–60, Springer – Verlag, Heidelberg, Germany [16]. Table 1 Early adaptive changes and neointima formation in saphenous vein grafts Study Reference Grafts / Patients Pre implantation to 1 month (mm 2 ) >12 months (mm 2 ) Lumen Wall Area Vessel CSA %wall area Lumen Plaque area Wall Area EEL area Vessel CSA % plaque area %wall area Nishioka et al 1996 [20] 43/42 16.5 ± 5.7 7.4 ± 2.1 23.9 ± 7.3 32.3 ± 7 8.9 ± 2.7 10.0 ± 5.3 15.2 ± 5.8 18.8 ± 7.5 24.0 ± 7.8 51 ± 10 63 ± 7 Ge et al 1999 [21] 43/43 12.6 ± 4.0 – 19.0 ± 9.7 64.5 ± 15.5 Hong et al 1999 [22] 104/93 12.0 ± 4.2 – 3.8 ± 1.9 7.2 ± 4.1 – 13.9 ± 4.9 10.0 ± 3.0 – 20.3 ± 6.5 16.7 ± .9 17.8 ± 6.1 20.8 ± 5.1 – 24.1 ± 7.8 30 ± 5 – 79 ± 9 45 ± 5 – 83 ± 7 Higuchi et al 2002τ [16] 47 16.2 ± 5.5 5.3 ± 2.0 21.6 ± 7.1 24.9 ± 5.0 12.8 ± 4.6 15.8 ± 5.2 28.8 ± 8.8 55.7 ± 6.8 **values represent range from reference segment to focal stenosis, τ angiographically normal vein. Vessel CSA, (cross sectional area) measured by tracing the outer border of the whole vein graft, Wall area , Vessel CSA minus lumen area. Percent wall area was calculated as the wall area divided by Vessel CSA. In situ veins do not have an external elastic membrane however; arterialized saphenous vein grafts develop a sonolucent zone, which has been reported to represent media. The EEL (external elastic membrane) area is measured by tracing the outer border of this sonolucent zone. Plaque area is calculated as external elastic membrane minus lumen area. Percent plaque area is calculated as plaque area divided by external elastic membrane area; this has also been called the plaque burden. Plaque burden and percent wall are closely correlated. IVUS changes correlate closely with the findings of experimental and clinical studies. In porcine saphenous vein bypass grafts, in the first week after grafting, adventitial medial and neointimal thickening occurs as a consequence of increased shear stress, surgical preparative injury and the subsequent activation of multiple growth factor and cytokine cascades. This is associated with the infiltration of inflammatory cells, medial smooth muscle cell proliferation and migration to form a neointima [ 25 ]. Adventitial myofibroblast proliferation and extracellular matrix deposition also results in the formation of a thick neoadventitia [ 26 ]. These myofibroblasts migrate through all the layers of the vessel wall, where subsequent extracellular matrix deposition contributes to overall wall thickening [ 27 ]. A similar distribution of cytoskeletal proteins characteristic of myofibroblasts is observed in explanted human saphenous vein grafts suggesting that similar mechanisms occur in man [ 27 ]. After the first week, wall thickening in porcine vein grafts occurs largely due to extracellular matrix deposition (fibrosis) and neointimal smooth muscle cell proliferation, however this thickening plateaus after one month [ 26 ]. The early changes seen in the vessel wall of vein grafts are similar to those seen during vessel remodeling in atherosclerotic coronary artery segments [ 19 ]. In normal arteries, remodeling is a homeostatic response to changes in flow and circumferential stretch, with compensatory enlargement and wall thickening normalizing shear stress and wall tension in response to higher blood pressures and flow velocities respectively. Outward remodeling in response to increased flow is largely dependent on shear-responsive endothelial production of nitric oxide and the gelatinase matrix metalloproteinases (MMPs) MMP-2 and MMP-9 [ 28 , 29 ]. MMPs are central to the turnover of the extracellular matrix, altering cell-cell interactions, modifying the extracellular milieu and permitting the movement and division of cells. Increased MMP production, with extracellular matrix degradation is a feature of the infiltration of inflammatory cells as well as the migration of smooth muscle cells and myofibroblasts [ 30 , 31 ], and this may also contribute to the remodeling process [ 19 ]. Late changes in vein grafts: atherosclerosis and pathological remodelling Early vein graft changes can be viewed as adaptive, however they also predispose the graft to later accelerated graft atherosclerosis [ 32 ]. Several components of the extracellular matrix that are abundant in diffuse fibrous intimal hyperplasia may increase the residence of atherogenic molecules, and promote the development of lipid-laden lesions [ 33 , 34 ]. Similarly, myofibroblasts are associated with contractile responses as part of wound healing [ 35 ] and it has been hypothesized that dissemination of these cells throughout all layers of the vein graft may be central to later inadequate or constrictive vessel remodeling [ 36 , 37 ]. Risk factors for, and the microscopic appearance of vein graft atherosclerosis are largely similar to those in coronary arteries and it is reasonable to suggest that similar pathological mechanisms are at work, however these occur over a much more rapid time course in vein grafts [ 1 , 4 ]. Atheromatous plaque is detected by IVUS as early as eight to ten months post grafting [ 38 ] in association with both expansive and constrictive remodelling [ 22 ] (Table 2 ). This is much earlier than originally suggested by angiography [ 5 , 6 ]. Early IVUS studies disagreed as to the nature of vein graft remodeling, with some studies reporting expansive remodeling [ 21 , 38 ] whilst others did not [ 20 ]. This confusion was most likely due to the small sample sizes in these early studies however (Table 2 ). Hong et al [ 22 ] used IVUS to assess the extent and direction of remodeling in 104 grafts in 93 patients, the largest analysis of diseased vein grafts published to date. In individual lesions, they defined remodeling by comparing the area within the EEL at the site of stenosis to that of a reference point. Positive remodeling was defined as a stenosis/ reference EEL area ratio >1.1, intermediate remodeling as a ratio 0.9 to 1.1, and negative remodeling as a ratio <0.9 [ 39 ]. All three processes were shown to occur, sometimes even within the same vessel. Overall plaque burden was greater in positively remodeled segments compared with intermediate or negative remodeling, whilst lumen area was preserved in all groups [ 22 ]. This is similar to changes that occur in atherosclerotic coronary arteries [ 40 ]. Mendellson and colleagues identified expansive remodeling in 98.5% of 24 vein graft lesions studied (Table 2 ). They showed that lumen area did not change with increasing percent area stenosis for vessels with ~30% of their area occupied by plaque (p = NS), however, for segments with >30% of the vessel area occupied by plaque, there was an inverse relation with the lumen area [ 38 ]. In this group (with >30% effective plaque area stenosis), the lumen area decreased as the percentage of vessel area occupied by plaque increased. Again these were the same changes as those noted in coronary arteries [ 41 ]. They suggest that "compensatory" enlargement mechanisms are very effective for early atherosclerotic lesions; but as lesions become more progressive, these mechanisms can no longer compensate, and lumen narrowing occurs. This does not explain constrictive or negative remodeling however and the mechanisms underlying this process remain unclear. In addition to the role of myofibroblasts, the formation of a dense neoadventitia with extensive collagen deposition has been implicated as a mechanical barrier to vessel enlargement, with later remodeling of this fibrous tissue ultimately resulting in vessel shrinkage [ 37 ]. In atherosclerotic arteries negative or inadequate remodeling is more common in insulin-using than non-insulin-using diabetics [ 42 ], more common in smokers compared with nonsmokers, and less frequent in patients with hypercholesterolaemia [ 43 ]. Mendellson et al showed that expansive vein graft remodeling was independent of graft age, insertion site, plaque eccentricity, patient age, or gender [ 38 ]. Altered local haemodynamics can also affect remodeling. Low shear predisposes the inner curves of tortuous segments to develop atheroma and may impair outward remodeling in a similar manner [ 44 , 45 ]. Alternatively, medial thinning as a consequence of atherosclerotic plaque development may result in bulging of the vessel wall due to diminished structural support at the site of the plaque [ 46 ]. Plaque volume often correlates with the level of inflammatory infiltrate, which again may contribute to expansion by promoting collagen lysis and deposition of loose myxoid extracellular matrix [ 47 ]. This is thought to underlie the propensity for such plaques to rupture and explain the association between expansive remodeling and unstable coronary symptoms [ 19 ]. Table 2 Late remodeling in atherosclerotic vein grafts Grafts / Patients Latency (years) Lumen area Intimal and medial(plaque) area External elastic lamina area Reference Stenosis Reference Stenosis Reference Stenosis Mendellson et al 1995 [38] 21/19 0.8–16 14.6 ± 7.5 7.1 ± 4.5 5.6 ± 3.4 18.3 ± 7.0 20.2 ± 8.5 25.4 ± 8.2 Nishioha et al 1996 [20] 43/42 3–12 15.7 ± 6.8 5.0 ± 1.5 3.2 ± 1.5 13.7 ± 6.3 18.9 ± 7.0 18.7 ± 7.3 Ge et al 1997 [21] 43/43 1–6 12.6 ± 4.0 19.0 ± 9.7 Hong et al 1999 [22] Negative remodeling 104/93 1.2–20.7 12.5 ± 4.0 3.6 ± 2.0 5.0 ± 2.4 - 10.9 ± 6.2 17.8 ± 7.9 14.2 ± 8.1 Positive remodeling 11.8 ± 3.5 4.0 ± 1.8 5.1 ± 1.4 15.3 ± 4.7 16.7 ± 4.6 19.4 ± 6.2 Remodeling can be defined as the ratio of the EEL at the site of the stenosis to that of the reference point. Positive remodeling is defined as a stenosis/ reference EEL area ratio >1.1, intermediate remodeling is defined as a ratio 0.9 to 1.1, and negative remodeling as a ratio <0.9. The future IVUS studies have clearly shown that early 'adaptive' or pathological changes occur within weeks of grafting and that occlusive atheroma, in susceptible individuals occurs within 1 year. IVUS studies have therefore defined the window in which strategies to inhibit vein graft disease might be effective. Furthermore, in addition to its advantages over coronary angiography, IVUS in vein grafts has been shown to be both accurate [ 12 , 16 ] and reproducible [ 20 , 22 ] making it the obvious investigational tool to explore the effectiveness of these strategies on in clinical trials. IVUS measurements correlate closely with clinicopathological findings [ 12 ] and can detect standardised differences of >1 in studies of saphenous vein grafts versus ungrafted vein as well as in vein grafts early post implantation (within 1 month) compared with later periods [ 16 , 20 ]. In addition, Nishioka and colleagues [ 20 ] demonstrated that for the measurement of lumen area, the mean difference between two observations by the same observer was 3.1%, with a range of 0% to 9.0%. Between two observers, the mean difference was 3.8%, with a range of 1.0% to 5.6%. This reproducibility facilitates not only accurate comparisons between groups of patients but also assessment of the effects of intervention on grafts in longitudinal studies. The ability to manipulate vein grafts ex vivo prior to implantation using pharmacological or other methods that may inhibit subsequent disease is a feature unique to vein graft disease. There are many examples of this being successfully achieved in experimental models. Pre-treatment with rapamycin [ 48 ], paclitaxel (GD Angelini, unpublished observations) and the intracellular calcium dependant ATPase inhibitor, thapsigargin [ 4 ], have been shown to significantly inhibit the progression of vein graft disease in experimental models in vivo. Oral agents, such as NO donating aspirins [ 49 ] and endothelin antagonists have also been shown to be effective in porcine vein grafts in vivo [ 50 ] as has the application of a porous external polyester stent which inhibits neointima formation and promotes expansive remodelling in the absence of vein wall thickening [ 51 ]. Targeted gene transfer is another attractive option, with viral transfer of E2F-decoy oligonucleotides inhibiting vein graft failure after peripheral arterial reconstruction in clinical trials [ 52 ]. Stabilisation of vein grafts that have undergone early adaptive changes, but have yet to develop a large plaque burden is also a possibility. Risk factor modification such as cessation of smoking and aggressive lipid lowering [62–64] has been shown to improve long-term graft patency on angiography. The effect of other interventions such as antiplatelet therapy on the progression of vein graft disease have not been evaluated in IVUS studies however. Identification of patients with a large plaque burden on IVUS, but otherwise angiographically normal vein grafts may also enable targeted plaque stabilisation therapy. Conclusion IVUS has significantly contributed to our understanding of vein graft failure. It also serves as the natural tool for the development of clinical strategies that may lead to significant improvements in vein graft patency and more importantly for better long-term quality of life and longevity for patients with coronary artery disease. The introduction of clinical trials to address this Achille's heel of coronary bypass surgery are long overdue. Competing interests None declared. Authors contributions G Murphy performed the literature review and prepared the manuscript. G Angelini conceived the idea for the manuscript and prepared the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514613.xml
535805	A comparison of biologically variable ventilation to recruitment manoeuvres in a porcine model of acute lung injury	Background Biologically variable ventilation (return of physiological variability in rate and tidal volume using a computer-controller) was compared to control mode ventilation with and without a recruitment manoeuvre – 40 cm H 2 O for 40 sec performed hourly; in a porcine oleic acid acute lung injury model. Methods We compared gas exchange, respiratory mechanics, and measured bronchoalveolar fluid for inflammatory cytokines, cell counts and surfactant function. Lung injury was scored by light microscopy. Pigs received mechanical ventilation (F I O 2 = 0.3; PEEP 5 cm H 2 O) in control mode until PaO 2 decreased to 60 mm Hg with oleic acid infusion (PaO 2 /F I O 2 <200 mm Hg). Additional PEEP to 10 cm H 2 O was added after injury. Animals were randomized to one of the 3 modes of ventilation and followed for 5 hr after injury. Results PaO 2 and respiratory system compliance was significantly greater with biologically variable ventilation compared to the other 2 groups. Mean and mean peak airway pressures were also lower. There were no differences in cell counts in bronchoalveolar fluid by flow cytometry, or interleukin-8 and -10 levels between groups. Lung injury scoring revealed no difference between groups in the regions examined. No differences in surfactant function were seen between groups by capillary surfactometry. Conclusions In this porcine model of acute lung injury, various indices to measure injury or inflammation did not differ between the 3 approaches to ventilation. However, when using a low tidal volume strategy with moderate levels of PEEP, sustained improvements in arterial oxygen tension and respiratory system compliance were only seen with BVV when compared to CMV or CMV with a recruitment manoeuvre.	Background A negative consequence of mechanical ventilation using lower tidal volumes (V T ) in patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) is alveolar collapse [ 1 - 3 ]. Numerous strategies to recruit these collapsed units have been advocated, but the efficacy of various recruitment manoeuvres for improving and sustaining gas exchange is controversial. Increased PEEP levels have been advocated to maintain patency of the recruited lung, but higher levels of PEEP can cause regional overinflation [ 4 ], potentially contributing to ventilator associated lung injury [ 5 ]. Moreover, recent evidence finds that when patients with ALI/ARDS are managed with a low tidal volume (V T ) approach the addition of higher PEEP levels offers no further improvement in outcome [ 6 ]. Thus, high levels of PEEP may no longer have the same relevance for ALI/ARDS management as before. Independent of increases in F I O 2 , it remains unclear how best to improve and sustain oxygenation, during low V T ventilation strategies for ALI/ARDS management. Buchman [ 7 ] and others [ 8 , 9 ] have highlighted how decreased physiological variability can negatively impact critically ill patients. When such patients require assisted ventilation, physiological variability or "noise" can be restored to the respiratory rate and V T through use of biologically variable ventilation (BVV), a unique computer-controlled version of control mode ventilation (CMV). With BVV, gas exchange and respiratory mechanics improved in animal models, with [ 10 ] and without PEEP [ 11 ], during low V T protocols using an ARDSNet algorithm [ 12 ] and in healthy lungs during prolonged ventilation under anaesthesia [ 13 ]. After deliberate collapse with one lung ventilation, recruitment was accelerated [ 14 ]. A recent clinical trial showed BVV improved gas exchange and respiratory mechanics in patients undergoing abdominal aortic aneurysmectomy [ 15 ]. Other investigators showed noisy ventilation increased surfactant phospholipid levels compared to CMV [ 16 ] and a mathematical model of how BVV can enhance recruitment and gas exchange has been advanced [ 17 ]. While previous work has indicated BVV results in superior gas exchange and respiratory mechanics compared to CMV with added sighs, this was a post-hoc comparison in a model of deliberate alveolar collapse [ 14 ], not a model of ALI/ARDS. As well, the sigh breaths were not equivalent to the larger sustained breaths customarily seen with a recruitment manoeuvre. Thus, it remains unknown if BVV is inferior, comparable or superior to conventional low V T ventilation with a recruitment manoeuvre in ALI/ARDS using a low V T approach. Therefore, in this study in pigs with oleic acid lung injury, we compared BVV to conventional CMV or CMV with a recruitment manoeuvre (CMV-RM) of 40 cm H 2 O of continuous positive airway pressure for 40 sec performed hourly for 5 hrs. This approach has been shown to improve oxygenation in patients with early ARDS who do not have any chest wall impairment [ 18 ]. A multimodal approach was used to compare the three ventilation strategies. We measured gas exchange and respiratory mechanics. Bronchoalveolar lavage (BAL) fluid was collected to determine cell counts, inflammatory mediators and surfactant function. Tissue was examined by light microscopy to assess lung injury with an established scoring system [ 19 , 20 ] at end experiment. Methods Experimental Preparation Pigs (weighing 20–30 kg) received 0.6 mg atropine, 15 mg midazolam, and 300 mg ketamine intramuscularly for sedation. Isoflurane 5% in 100% oxygen was delivered via facemask to induce anaesthesia. When sufficient depth of anaesthesia was achieved, the pigs were intubated orotracheally with a 6.0 mm cuffed endotracheal tube. Mechanical ventilation was instituted with an Ohio 7000 ventilator (Ohio Medical, Madison WI) with minute ventilation adjusted to maintain a PaCO 2 of 35–45 mm Hg. Anaesthesia was maintained with 2% isoflurane in 100% oxygen during surgical preparation. Intravenous rocuronium bromide (1 mg/kg/hr) was administered by continuous infusion for muscle relaxation. Lactated Ringer's solution was given intravenously during the surgical preparation and for the duration of the experiment. A thermodilution pulmonary artery catheter (7.5-Fr) was inserted and advanced into the right external jugular vein until a satisfactory pulmonary capillary wedge tracing was obtained. Temperature was measured from the tip of the pulmonary artery catheter. The right femoral artery was cannulated for continuous pressure transduction and arterial blood gas (ABG) analysis. A 5-Fr single lumen femoral venous catheter was advanced into the inferior vena cava for infusion of oleic acid. A surgical tracheotomy was performed and the animal was switched to an Esprit ® ventilator (Respironics Inc., Palo Alto CA) capable of delivering either CMV or BVV. The ventilator was set to deliver a square wave inspiratory flow pattern with an I:E ratio of 1:2. Isoflurane was discontinued, and a propofol/ketamine infusion at 10/2.5 mg/kg/hr substituted to maintain anaesthesia. After a 30 cm H 2 O recruitment manoeuvre for 30 sec, animals were ventilated with a V T of 10 ml/kg at an F I O 2 of 0.3, with 5 cm H 2 O of PEEP. Respiratory rate was adjusted to maintain PaCO 2 between 35–45 mm Hg. After 15 min to stabilize, baseline measurements were obtained. Haemodynamic measurements included mean arterial pressure (MAP), heart rate, central venous pressure (CVP), mean pulmonary artery pressure (MPAP), and pulmonary artery occlusion pressure (PAOP). All haemodynamic data were continuously recorded on a Gould 2600 Oscillograph (Gould, Cleveland, OH). Cardiac output (CO) was measured by thermodilution, in triplicate, at stated measurement periods. A pneumotachograph (Model 3700; Hans Rudolph, Kansas City, MO) with the sensor immediately proximal to the tracheotomy was used to measure airway pressures and V T intermittently; this data was recorded using an advanced CODAS (Dataq Instruments, Akron, OH) data acquisition system. Arterial and mixed venous gases were analyzed using a Radiometer ABL 500 (Copenhagen NV, Denmark). Arterial and mixed venous oxygen content, oxygen saturation, and haemoglobin concentrations were measured with a Radiometer OSM3 set for porcine blood. Static respiratory system compliance (Crs) was measured in triplicate by clamping the expiratory limb of the ventilator circuit at the end of inspiration for 1 sec to obtain a plateau pressure. The V T used was that which the animal was receiving at that time. Oleic Acid Lung Injury Oleic acid was infused via the 5-Fr femoral venous catheter at 0.2 ml/kg/hr until PaO 2 <60 mm Hg for two consecutive measurements (PaO 2 /F I O 2 <200). Dopamine was started at 5 μg/kg/min and was titrated to keep the MAP >60 mmHg. When the oxygenation target was achieved, the oleic acid infusion was stopped and the PEEP was increased to 10 cm H 2 O. Ten minutes after the increase in PEEP, an arterial blood gas sample was obtained to determine if PaO 2 had increased. PaO 2 had to be >75 mm Hg but <90 mm Hg. This was considered to represent adequate lung injury, but indicate that collapsed alveoli could be recruited with the additional PEEP. If the PaO 2 did not increase, the experiment was terminated. If the PaO 2 increased to >90 mm Hg, additional oleic acid was infused until the PaO 2 decreased to <60 mm Hg. Ventilation Protocol A low V T (7 ml/kg) protocol was initiated and the respiratory rate increased to 30 bpm. After 10–15 min arterial blood gas sampling was done to assess the stability of the PaO 2 . If the PaO 2 remained stable, the animals were then randomized into one of three groups: conventional ventilation with a V T of 7 ml/kg (CMV); conventional ventilation with V T of 7 ml/kg with a 40 sec, 40 cm H 2 O recruitment manoeuvre performed hourly (CMV-RM). The recruitment manoeuvre was performed at end-expiration with the PEEP level maintained at 10 cm H 2 O at an F I O 2 of 0.3; or biologically variable ventilation (BVV) with a mean V T of 7 ml/kg. Following stable oleic acid lung injury, haemodynamic, gas exchange and respiratory system compliance (Crs) measurements were recorded and obtained hourly thereafter. Measurements in the CMV-RM group were obtained 5 min after each recruitment manoeuvre (RM). An additional measure of gas exchange and Crs was made in the CMV-RM group 50 min after the RM, to ascertain the duration of effect of the RM. Wet:Dry Lung Weight Ratios At the end of the experiment, a sternotomy was performed. Animals were sacrificed with a lethal dose of thiopental. The trachea was then clamped at end inspiration and the heart and lungs were removed en bloc . Following removal, the lungs were suspended upside down for 10 min to collect bronchoalveolar (BAL) fluid. Samples of BAL fluid were collected in heparinized saline and then frozen immediately at -80 o C. Fluid was then sent for cytokine analysis, measures of surfactant function and flow cytometry. The lungs and previously collected BAL fluid were weighed and the lungs were suspended and aerated overnight. The following day, the lungs were placed in an oven to dry, and following a stable dry measurement, wet:dry weight ratio was calculated. Surfactant Function Assays Surfactant function was assessed on BAL fluid samples using a capillary surfactometer (Calmia Medical, Toronto, Canada) in the manner of Enhorning and colleagues [ 21 , 22 ]. Such an approach can delineate differences in surfactant function with lung inflammation [ 23 ]. Surfactometry was performed under two conditions: 1. Raw BAL fluid surfactant function analysis – centrifuging at 200 g for 5 min to rid large debris, then the supernatant spun at 10,000 g and pellet resuspended in 100 μL saline and analyzed using the surfactometer. 2. Surfactant resuspended in 100 μL of saline after chloroform/methanol extraction. Chloroform/methanol extraction permits the lipid and phospholipid fraction to dissolve in the organic nonpolar solvent (chloroform) and the solvent evaporated to dryness. Volume of the final ratio for chloroform:methanol:water was 1:1:0.9. Following extraction, the pellet was resuspended and analyzed using the surfactometer. The percentage of time that the capillary tube was open for 2 min was determined for each sample, in triplicate, then averaged. Standards were saline; 0% patent for a 2 min time period and bovine surfactant; >98% patent for a 2 min time period. Each sample was measured in a blinded fashion in both conditions. Flow Cytometry Analysis The BAL fluid sample was passed through a cell strainer and 100 μL was transferred into a 5 × 75 mm tissue culture tube. Red cells were lyses by the addition of 500 μL Optilyse C (Beckman Coulter, Mississauga, Canada) and following 15 min incubation at room temperature, the cell suspension was diluted by the addition of 500 μL Isoton II and 100 μL FlowCount Fluorospheres (Beckman Coulter). The cell suspension was immediately analyzed on a Beckman Coulter EPICS Altra cell sorter configured with a high-speed quartz flow cell tip and a water-cooled argon laser emitting 150 mW at 488 nm. Forward and side light scatter signals were employed to derive 2-parameter histograms which clearly defined the fluorescent beads and three populations of cells, subsequently defined as mononuclear cells, neutrophils and eosinophils. At least 20,000 cells were analyzed for each sample. The acquisition software provided with the instrument automatically calculated the concentrations of each cell population based on the number of events in each of the 4 analysis regions. Cytokine Assays for IL-8 and -10 ELISA The incubation times and washes were preformed as specified in each respective kit: (BioSource International, Camarillo, CA). Immunoassay kits for swine IL-8 and IL-10 were used. Sample incubation times were kept as constant as possible by pre-plating on a blank 96 well plate before transferring to the coated assay plate. ELISA plates were read at 450 nm by an SLT Rainbow plate reader (Lab Instruments, Research Triangle Park, NC). Standard curves and concentration calculations were performed according to kit directions. Histological Assessment For light microscopy, four blocks of tissue from the right lung – upper lobe, middle lobe, nondependent and dependent lower lobe, fixed in buffered formalin, processed and embedded in formalin. Sections were cut and stained with haematoxylin and eosin. Under light microscopy, a lung injury scoring system modified from Rotta et al. [ 19 , 20 ] was used, based on the following variables: alveolar and interstitial inflammation, alveolar and interstitial haemorrhage, oedema, atelectasis and necrosis. The severity of injury was graded for each of the seven variables: no injury = 0; injury to 25% of the field = 1; injury to 50% of the field = 2; injury to 75% of the field = 3; and diffuse injury = 4; for a maximal total score of 28. Statistical Analysis Parametric data were analyzed by repeated measures ANOVA using least squares means test matrices to identify differences within and between groups from group × time or group effects. Bonferroni's correction was applied where appropriate. Non-parametric comparison of the lung injury scores was by Kruskal-Wallis test. In all circumstances a p-value ≤0.05 corrected for multiple comparisons was considered significant. Results Twenty-five experiments were undertaken. One animal was discarded because oleic acid injury could not be established after 1.5 hrs of infusion. In the CMV-RM group, one animal died 2 hr into the experiment but data are included in analysis in this experiment up to time of the animal's demise. Complete experiments were done in twenty-three animals; (n = 8 for all three groups). Temperature and Haemodynamic Data Temperature and haemodynamic data are shown in Table 1: see Additional File 1 . All data reported as mean ± SD, unless stated. A significant group × time interaction was seen for temperature. A crossover in temperature was seen with greater temperature at baseline in the BVV group that decreased over time by a mean of 0.8°C by end experiment. In contrast in both of the CMV groups the temperature increased modestly. Least squares means test showed MAP lower in CMV-RM at one hr and beyond after oleic acid compared to the other 2 groups. MPAP was increased significantly following administration of oleic acid in all groups. Between groups MPAP was lower with BVV from 1 hr after oleic acid. No group × time interaction was seen for PAOP (p = 0.755) but by least squares means test at baseline, PAOP was lower in the BVV group at baseline and at one hr following oleic acid. In all groups CO decreased following oleic acid and remained depressed beyond (group × time interaction p = 0.521). Respiratory Gas and Derived Data Significant differences were seen between groups for PaO 2 over time (Figure 1a ). The group × time interaction was p = 0.0001 with greater PaO 2 seen with BVV, compared to either CMV or CMV-RM. There was no difference in PaO 2 between groups at baseline, following oleic acid administration or during the first hour of the experiment. A clear separation in PaO 2 levels became apparent after 2 hr following oleic acid injury in the BVV group. There was no difference between CMV and CMV-RM over time. Respiratory system compliance for the 3 groups is shown in Figure 1b . The group × time interaction was p = 0.089 between groups. Comparison of compliance between groups with least squares means showed significant differences between BVV with CMV and CMV-RM after 2 hr. Figure 1 1a and b – Arterial Oxygenation and Respiratory System Compliance. Arterial oxygen tension (PaO 2 ) over time for the 3 groups (BVV in red; CMV in blue and CMV-RM in green) (a) . The group × time interaction is p = 0.0001. No difference is seen between groups at baseline and following oleic acid infusion. By 2 hr the PaO 2 is greater with BVV (+). There is no difference between CMV and CMV-RM. Respiratory system compliance over time for the 3 groups (b) . The group × time interaction is p = 0.089. Least squares means tests revealed BVV had significantly greater compliance after 2 hr (+). There was no difference between CMV and CMV-RM. Table 2: see Additional File 2 , shows respiratory gas and derived data for each group. There was no group × time interaction for PaCO 2 (p = 0.660). In all groups, the PaCO 2 increased following oleic acid injury and was essentially stable at these elevated levels for the remainder of the experiment. The group × time for PvO 2 was significant at p = 0.002 with significantly lower PvO 2 over time in the CMV-RM group. Shunt fraction was lower in the BVV group (group × time interaction p = 0.003). Recruitment Effects on Oxygenation and Compliance PaO 2 tended to decrease immediately following a recruitment manoeuvre, but increased over time as measured at 50 min. PaO 2 significantly decreased in RM4 and 5 (-11.1 ± 5.2 mm Hg and -5.2 ± 4.4 mm Hg respectively). Following recruitment, PaO 2 increased slowly early in the experiment at RM2 and RM3 (10.1 ± 5.5 mm Hg and 11.8 ± 9.1 mm Hg respectively), but failed to do so in the later time periods. For respiratory system compliance, no statistically significant differences were seen at any time period over the course of the experiment in the CMV-RM group. Airway Pressure Data A group × time interaction for mean Paw was seen (p = 0.002) with Paw lowest with BVV (Paw at 5 hr was 14.3 ± 0.4 cm H 2 O with BVV; 15.1 ± 0.2 with CMV-RM and 15.1 ± 1.0 with CMV respectively). In all 3 groups, mean Paw increased following oleic acid. A more pronounced effect was seen with peak Paw (group × time; p = 0.0001) with mean peak pressures over time least with BVV (peak Paw at 5 hr was 23.2 ± 1.9 cm H 2 O with BVV; 28.8 ± 0.8 with CMV-RM and 28.6 ± 2.8 with CMV respectively). There was no group × time interaction for V T in ml/kg measured over time (p = 0.617). Lung Histology When the lung was scored in the upper, middle, nondependent and dependent lower lobes in the right lung, no differences were seen for any region by Kruskal-Wallis test between the 3 groups. In the dependent region of the lower lobes, injury was considerable with scores ranging between 9/28 and 22/28. Wet:dry weight ratios for the three groups were BVV; 8.5 ± 4.0, CMV; 8.9 ± 1.2, and CMV-RM; 11.5 ± 4.8; not significant between any group. Capillary Surfactometry Capillary surfactometry was done on raw BAL fluid and following chloroform/methanol extraction. There was no difference in surfactometry for raw BAL fluid between groups; the duration of capillary tube patency as an index of surfactant function was: 6.7 ± 14.3%, 0.7 ± 0.8% and 0.9 ± 1.7%; for BVV, CMV and CMV-RM respectively – not significant. Following chloroform/methanol extraction, surfactant function of the same fluid usually improved dramatically – see Figure 2 . Here, duration of capillary patency was 57.6 ± 40.1%, 47.4 ± 43.8% and 81.3 ± 33.8%; for BVV, CMV and CMV-RM respectively – not significant between any groups. In 4 animals, the chloroform/methanol extraction had no discernible effect on surfactant function, in all others surfactant function improved markedly. Figure 2 Capillary Surfactometry. Box and whisker plots for capillary surfactometry results from both raw (blue) and chloroform/methanol extracted (red) BAL fluid and patency time (%) over 2 min – log scale. In most circumstances the raw BAL fluid has minimal surface activity – low capillary patency. The one animal with a high raw value was from the BVV group. This patency markedly improved with extraction suggesting reactivation of surfactant with removal of BAL oedema fluid, proteins, including cytokines and cellular debris. In 4 animals there was essentially no change in function with extraction. There were 2 animals in the CMV group, and one each in the BVV and CMV-RM in this population. No statistically different behaviour in surfactant function was seen between groups, either for raw or extracted surfactant function. Flow Cytometry The flow cytometry of BAL fluid indicate no significant difference between cell counts either in terms of neutrophils, monocytes or eosinophils between groups. Neutrophils were the predominant cell type, at 74% ± 15% of the totals. The total BAL cell counts were highly variable however, with neutrophil counts varying from a low of 187 to a high of 26,189 cells/μL – a 140-fold difference. There was no group effect for neutrophils; p = 0.741. The mean BAL neutrophil counts were 8237 ± 8738/μL for BVV, 12024 ± 12328/μL for CMV and 6492 ± 4219/μL for CMV-RM. BAL fluid volume was also measured based on gravity drainage from the lung over a 10 min period. Volume was not significantly different between groups and was 13.6 ± 16.8 mL with BVV, 15.1 ± 10.6 mL with CMV and 28.4 ± 23.2 mL with CMV-RM. The total neutrophil count in the BAL fluid was calculated as neutrophils/μL × BAL volume in μL for each animal. Mean total neutrophil counts were 1.38 × 10 8 for BVV, 1.82 × 10 8 for CMV and 2.34 × 10 8 for CMV-RM. There was no significant difference between groups. ELISA Results and Cytometry Correlations The ELISA results to assess the cytokine IL-8 concentration in BAL fluid indicated no statistically significant difference for the 3 groups: 564 ± 551 pg/mL with BVV; 653 ± 639 pg/mL with CMV and 300 ± 235 pg/mL with CMV-RM. As in the measurement of neutrophil counts, there was significant variability with IL-8 levels ranging over a 385-fold concentration difference. Such a broad range of data in both neutrophil counts and IL-8 concentration suggested the possibility of power law behaviour and correlation between these variables. A very strong power law relationship was found between IL-8 pg/mL and absolute neutrophil count/mL for pooled data; y = 77609x 0.75 ; R 2 = 0.85, n = 13. This power law relationship exceeded the linear (R 2 = 0.60) and exponential (R 2 = 0.41) curve fits. The number of mononuclear cells from cell cytometry of BAL fluid was not different between groups: for BVV, 1346 ± 743 cells/μL; for CMV, 1767 ± 905 cells/μL; and for CMV-RM, 1068 ± 579 cells/μL. These cells presumably in large part represent alveolar macrophages. A power law relationship was seen between pooled data for a correlation between IL-8 concentration in pg/mL and monocyte count/mL – 1.0 × 10 -6 x 1.38 ; n = 13, R 2 = 0.83. In this situation the linear and exponential fits were R 2 = 0.61 and R 2 = 0.69 respectively. In this analysis, IL-8 concentration is presumed to be the dependent variable – being released by the monocytes (counts on the x-axis). The IL-10 results from BAL fluid, as well, indicate no statistically significant difference for the 3 groups: 119 ± 173 pg/mL for BVV; 124 ± 146 pg/mL for CMV; and 149 ± 168 pg/mL for CMV-RM. Unlike IL-8 concentrations the range of variation for IL-10 concentrations was significantly less – the maximum range differing by only 16-fold. There was no power-law relationship found between IL-10 and absolute monocyte count for pooled data: y = 7.5x 0.15 ; n = 13, R 2 = 0.07. Discussion The main finding in this study of acute lung injury in a porcine model is that BVV significantly improved oxygenation and respiratory mechanics with no difference in indices of lung injury, inflammation or surfactant function compared to the more conventional ventilation techniques – CMV or CMV with a standard recruitment manoeuvre. The improvements seen with BVV were sustained over the course of the experiment; an effect not seen with recruitment. Such improvement was obtained at similar measured mean V T but at lower peak and mean airway pressure and greater respiratory system compliance. These findings are similar to previously published results from our laboratory showing that BVV was superior to both CMV and CMV with sigh breaths. These sigh breaths were of no advantage in a model of healthy lung recruitment – lung reinflation after a controlled collapse for one hr with contralateral one-lung ventilation [ 14 ]. In that model, the sigh breaths were not as large or as sustained as the recruitment manoeuvre studied here. An important difference from our previous work is that the current experiment examined recruitment in a low V T model of ALI/ARDS; not in healthy lung. As well, we have confirmed the differences between BVV and CMV seen in other work from the laboratory using the same model [ 12 ]. No important differences were seen for gas exchange or respiratory mechanics between CMV and CMV-RM in the current study. Shunt fraction was significantly lower with BVV. Lynch et al. [ 24 ] and Sandoval et al. [ 25 ] showed that shunt fraction is directly related to cardiac output and PvO 2 . Thus, lower PvO 2 in the CMV-RM group should have minimized shunt but here the shunt fraction is greatest. The lowest shunt with BVV indicates that the numerator of the shunt equation is less in this group, indicating enhanced blood flow in aerated lung units. A multimodal approach to assess lung injury, lung inflammation and surfactant function indicated no discernable differences between the 3 approaches to ventilation: 1. Light microscopy studies demonstrated no significant differences between groups for lung injury in any region examined – from upper to dependent lower lobes. This finding indicates no difference between groups in lung injury as assessed by histology, with the caveat that only small areas of lung in each region were examined in a condition known to be heterogeneous. 2. No difference in surfactant function was seen between groups as assessed by capillary surfactometry. Following oleic acid lung injury, surfactant function was markedly depressed when compared to a surfactant standard – see Figure 2 . The raw surfactant from the BAL fluid was not usually active beyond a few percent of normal function, with no difference between groups. In contrast, following chloroform/methanol extraction, surfactant function improved toward normal in most circumstances, again with no difference between groups. It is well known that surfactant is inactivated in the presence of oedema fluid, inflammatory proteins and cellular debris, but that the surfactant can be reactivated in the right circumstances as seen here following chloroform/methanol extraction. No correlation was seen for surfactant function of pooled data for PaO 2 , wet:dry weight ratios, respiratory system compliance, or peak airway pressure. Arold et al. [ 16 ] have shown that noisy ventilation similar to BVV can result in greater surfactant phospholipid concentration over time in healthy lungs. We do not address this specific issue. 3. No difference between groups is seen for cell counts in the BAL fluid. There is considerable variation in this analysis and a full data set is not present. A significant number of studies of cell count could not be undertaken because of the nature of the BAL fluid. When thick with proteinaceous debris the samples often were unable to be prepared for meaningful cytometry analysis. Of the data represented, no difference in the proportion of neutrophils, monocytes or eosinophils is apparent for any of the 3 approaches to ventilation. Allen and Bates [ 26 ] have shown that the number of neutrophils in BAL fluid – a 500 fold difference between groups in their study – did not correlate with changes in respiratory mechanics as assessed by changes in elastance with deep inflation in mice. Their findings suggest that the non significant differences in neutrophils counts – 70% greater total counts with CMV-RM than BVV – should be of no influence on respiratory mechanics. 4. IL-8 is considered to be the major neutrophil chemoattractant cytokine in lung diseases like ARDS [ 27 ]. IL-10 markedly inhibits lymphocyte and phagocytic function, essential for an adequate immune response to invading microbes [ 28 ]. Initial IL-10 serum levels have been shown to be significantly higher in patients with ARDS who died as compared to survivors [ 29 ]. No differences were seen between groups in the levels of the inflammatory cytokines IL-8 and IL-10 in BAL fluid. This is not surprising given similar monocyte and neutrophil counts between groups. A large coefficient of variation for both cytokine concentrations is demonstrated, a common finding in most such studies [ 30 ]. Such large variation makes meaningful comparisons between groups problematic. A power law relationship between IL-8 concentrations and monocyte and neutrophil counts in BAL fluid is a new insight. We found a very positive correlation relating IL-8 concentration and cell counts. The markedly variable IL-8 concentrations and cell counts in BAL fluid under controlled experimental conditions, suggests a highly nonlinear process has been initiated following oleic acid administration. Despite the observed variation in cytokine and cell counts a strong linkage is suggested by the power-law correlation. This observation, in and of itself is not surprising, given that IL-8 is a known chemoattractant. But the power law descriptor suggests that both low and high concentrations of IL-8 and corresponding cell counts are more frequent than expected. These data also suggest that this interaction is "scalable" with no specific mean value to be anticipated [ 31 ]. In this circumstance, large coefficients of variation are to be expected, making meaningful standard comparisons between groups problematic. Looking for similar power law correlations may help to elucidate the nature of the inflammatory process in the future. We have seen differences in temperature between BVV and CMV in the past with greater IL-8 levels with CMV [ 12 ]. We cannot confirm that finding in this study. As well, we were unable to confirm a previous observation of an inverse correlation of IL-8 concentration and wet:dry weight ratios in this model (R 2 = 0.007 in this experiment, n = 24). Failure to reconfirm these findings may, in part relate to the nonlinear relationships highlighted above. Small changes in initial conditions may preclude similar findings at end experiment. The above observations, collectively, indicate no fundamental differences between the ventilation strategies studied in regards oleic acid or superimposed ventilator associated lung injury. No one technique seems clearly advantageous. However, BVV alone improves oxygenation significantly following acute lung injury over time, an effect that was sustained, suggesting it is the best technique of the 3 studied to recruit atelectatic lung as assessed by greater PaO 2 and respiratory system compliance. BVV has been shown to recruit in a pure model of lung collapse – re-expansion of lung following cessation of one lung ventilation [ 14 ]. With BVV, the addition of a noisy end-inspiratory pressure has been advanced as the mechanism to improve oxygenation in the face of the nonlinear characteristics of alveolar recruitment [ 17 ]. It could be argued that we have not chosen the optimal recruitment strategy to compare to the two other modes of mechanical ventilation. That said, we have chosen a well recognized approach to recruitment [ 18 ], one associated with improved patient outcome in a carefully conducted clinical trial. Amato and colleagues [ 32 ] did not state how often their chosen recruitment manoeuvre was used in their clinical study – just that it was used frequently – usually in association with suctioning the tracheobronchial tree. We have applied the Amato recruitment strategy in a comprehensive manner – hourly in an acute model of ARDS. But, in contrast to Amato and colleagues, where PEEP was increased from 5 cm H 2 O to 2 cm H 2 O above P flex in an attempt to prevent derecruitment, we maintained PEEP at 10 cm H 2 O throughout. In a porcine model of lung lavage, lung remained recruited if PEEP was at 10 cm H 2 O: the value chosen in this study [ 33 ]. This level of PEEP is that recommended in a recent review of mechanical ventilation in ARDS, especially in the context of patchy ARDS where the risk of overdistention of patent alveoli increases the potential for volutrauma [ 34 ]. Furthermore, in the ARDSNet study, the Amato recruitment manoeuvre failed to sustain an improved arterial oxygenation, so was abandoned [ 6 ]. Failure to demonstrate an increase in oxygen with recruitment suggests either redistribution of blood flow to poorly ventilated regions or potentially increased alveolar flooding with alveolar-capillary disruption. Such causes may account for our failure to demonstrate improved oxygenation at 5 min after individual recruitment manoeuvres. Positron emission tomography indicates that PaO 2 will not increase if the sustained inflation does not restore aeration to the atelectatic regions because a significant fraction of pulmonary blood flow is shunted to nonaerated regions [ 35 ]. Such may be the case in this study as shunt fraction was greatest in the CMV-RM group. Fujino and colleagues [ 36 ] have shown that a recruitment manoeuvre similar to the Amato manoeuvre – 40 cm H 2 O for 60 sec – was not beneficial in another animal model and that a more aggressive recruitment manoeuvre – PEEP 40 cm H 2 O and pressure control ventilation of 20 cm H 2 O, with a respiratory rate of 10, and I:E ratio 1:1 for 2 min – was successful only after the second hour. Allen et al. [ 37 ] have shown in a mouse model with deep inflation that PEEP at various, albeit low levels (1, 3, 6 cm H 2 O) did not influence the time constants for recovery in elastance both before and after lung lavage. They conclude that for a deep inflation to be beneficial in their model that it may be necessary to apply the inflation several times a minute. Thus, the appropriate PEEP level, with and following recruitment, remains controversial [ 38 , 39 ], but what is clear from the current study is that at the same PEEP level significantly better gas exchange and respiratory mechanics occurs with BVV over either mode of CMV – with or without recruitment, with no differences in lung injury, inflammation or surfactant function. The human variability file for respiratory rate used to program the ventilator is shown in Figure 3 . Mathematical analysis indicates that this file is not fundamentally different in characteristics from previous files used to program the ventilator. Such variability files have fractal time sequences and are associated with health. Whether or not "fractal noise" is superior to random or "white noise" is as yet unproven, however, such fractal variability is characteristic of human health. Suki et al. [ 17 ] showed the importance of noise to increase the volume recruited when alveoli are collapsed and provide a mathematical framework for how BVV can result in better gas exchange and respiratory mechanics in the context of alveolar collapse. This study confirms previous findings that BVV can deliver the same minute ventilation over time at lower mean peak airway pressure and greater respiratory compliance. Figure 3 Human Respiratory Rate Variability File. Human breathing variability file used for the study. This is the raw data from a spontaneously breathing healthy female volunteer. The mean rate was 13.4 ± 2.0 breaths/min (shown as the red line). There are 1587 breaths in this file. With BVV, the ventilator is configured as a volume divider at a fixed minute ventilation so that respiratory rate × tidal volume product is constant. Thus the breath-by-breath volume related to instantaneous respiratory rate obtained from sequentially reading the above file in any given experiment is obtained from the minute ventilation/[(instantaneous breath rate/13.4) × chosen mean rate]. Analysis reveals that these data have fractal characteristics. Conclusions This study shows that BVV with a human variability file was superior to either CMV or CMV with a recruitment manoeuvre (CMV-RM) for sustained improvement in gas exchange and respiratory mechanics in a porcine model of acute lung injury. Prior work has demonstrated an advantage of BVV in a clinical setting of atelectasis [ 15 ]. The current study suggests that BVV may be superior to more conventional approaches to improve oxygenation without increasing the risk of volutrauma when low V T strategies are used for management of ALI/ARDS. A well accepted recruitment strategy has been compared to BVV in this animal study. It is possible that other approaches for recruitment could be superior to the one chosen. However, all chosen modes of recruitment would increase airway pressures over and above those seen with BVV, potentially increasing the risk of volutrauma. Authors' contributions D.J.F. was responsible for conduct of the experiments as a fellow in the Anesthesia Laboratory. M.R.G. supervised conduct of the experiments, helped analyze data and helped write the paper. L.G.G. helped with the experiments, data retrieval and collation and table and figure production. J.A.T. did the histological assessments and analysis. B.M.M. supervised the cytokine assays and their interpretation and helped write the paper. E. K-Y. W. conducted the cytokine assays. C.H. did the surfactant assays while a fellow in the Oral Biology Laboratory. J.E.S. supervised and interpreted the results of the surfactant assays. W.A.C.M. conceived the study, analyzed and interpreted data and helped write the paper. Supplementary Material Additional File 1 Table 1: Temperature and Haemodynamics Click here for file Additional File 2 Table 2: Respiratory Gas and Derived Data Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535805.xml
535346	Mosquitoes and transmission of malaria parasites – not just vectors	The regional malaria epidemics of the early 1900s provided the basis for much of our current understanding of malaria epidemiology. Colonel Gill, an eminent malariologist of that time, suggested that the explosive nature of the regional epidemics was due to a sudden increased infectiousness of the adult population. His pertinent observations underlying this suggestion have, however, gone unheeded. Here, the literature on Plasmodium seasonal behaviour is reviewed and three historical data sets, concerning seasonal transmission of Plasmodium falciparum , are examined. It is proposed that the dramatic seasonal increase in the density of uninfected mosquito bites results in an increased infectiousness of the human reservoir of infection and, therefore, plays a key role in "kick-starting" malaria parasite transmission.	Introduction A strategic aim of the global commitment to Roll Back Malaria is the development of reliable and practicable forecasting methods to enable the containment of epidemics. Much of the basis of our knowledge about the causes and dynamics of malaria epidemics comes from the qualitative analyses of the large scale, regional epidemics occurring in the first third of the last century [ 1 - 3 ]. Such regional epidemics have been consistently associated with unusual seasonal climatic conditions, namely those favourable for malaria transmission, but which follow years of unfavourable conditions. Favourable conditions are those that promote both anopheline density and longevity, notably high atmospheric humidity [ 1 ]. Thus, epidemics occur when there are "unusually" favourable transmission conditions in an otherwise poorly immune population and that an epidemic reflects the loss of equilibrium between the degree of immunity and the force of infection [ 1 , 2 ]. The implementation of remote sensing technologies for use in early warning systems for predicting epidemics have been developed on the basis of these original observations [ 4 ] and the development of more precise measurement of key parameters, such as temperature and rainfall anomaly, is in progress. However, as pointed out, the application of such new technologies must be carried out within the context of the wealth of knowledge concerning the basic epidemiological processes of malaria [ 5 ]. In this spirit, recent advances in mathematical epidemiology have paid special interest to the formal analysis of the cyclical nature of epidemics and how the epidemiological system responds to demographic and environmental variability [ 4 - 8 ]. One conclusion from such analyses is that a deeper comprehension of longitudinal epidemiological patterns requires a more detailed appreciation of the intra-host parasite population dynamics, including the effects of co-infection with both multiple strains and species of Plasmodium [ 9 ]. Despite the extension by Macdonald of the basic models of malaria, pioneered by Ross and Lotka, to introduce some level of biological complexity, many points raised at the time of the great regional epidemics remain unexplored or attributed to the error inherent in estimating biological parameters [ 10 ]. In his reappraisal of epidemics in general [ 1 ] and of the 1934–5 Ceylon epidemic in particular [ 11 ], Gill not only emphasised the value of climatic features in predicting epidemics, but also made several observations that ran contrary to the opinion of malariologists of the day. Firstly, he noted that during the first five weeks of the epidemic, there was a complete absence of child morbidity. He subsequently suggested that the explosive nature of the Ceylon epidemic must have been due to a sudden increase in the infectivity of the adult population, and that, therefore, the beginning of the epidemic was characterised by relapses in the adult population that then generated the necessary human transmission population [ 11 , 12 ]. This was countered at the time by malariologists who believed, based on fever charts and the absence of gametocytes, that the initial wave of morbidity was due to novel infections [ 13 ]. However, to date there has been no satisfactory explanation to account for the age-structured nature of the morbidity rise, other than relapses (of both Plasmodium vivax and Plasmodium falciparum ) in the adult population. If Gill was indeed correct, then what caused the sudden increase in infectiousness of the adult population? Seasonal epidemic rises superimposed on endemic prevalence are probably characteristic of most regions with endemic malaria [ 2 ]; there is always a degree of seasonality in mosquito bionomics. If the unusually favourable conditions that generate epidemics are an extreme example of more common seasonal variations in transmission intensity, Gill's pertinent observations may throw light on a fundamental question in malaria epidemiology: How exactly does transmission restart upon return of the mosquitoes. Current wisdom suggests that the sheer number of mosquitoes under favourable conditions results in the rapid expansion of the parasite population from a few initial source infections – as formalized in the classic Ross-Macdonald model of malaria, Plasmodium ( falciparum ) has a very high R 0 (reproductive rate) that is strongly dependent on the mosquito biting rate. But where do the initial source infections come from? Several authors have noted the absence of gametocytes during inter-epidemic periods, but commented that a geometric rise in gametocyte carriers could generate the observed rapid increase in gametocyte rates during the course of an epidemic [ 14 - 16 ]. In an endemic field situation, chronic infections were found to produce gametocytes throughout the dry season in Sudan, although the transmissibility of such gametocytes was not assessed [ 17 , 18 ]. However, induced infection studies in naive individuals have shown that a P. falciparum infection can produce infective gametocytes throughout its infection duration [ 19 ]. Even this, however, cannot explain the explosive nature of the Ceylon epidemic [ 11 ], which strongly suggests that there may be underlying seasonal changes in intra-host parasite population dynamics that could have an effect on parasite transmission to mosquitoes. Inherent seasonality in some Plasmodium spp. (notably P. vivax ), is well known and yet rarely considered in models of malaria epidemiology. In this paper, the nature of seasonality in malaria parasites is addressed and by examination of three historical data sets from Africa and Asia, it is proposed that uninfected mosquito bites play a significant role in increasing the infectiousness of the human reservoir of infection to mosquitoes. Thus, the intense expansion of the mosquito population at the start of the seasonal transmission season (whether resulting in an epidemic or not) has a significant biological effect above and beyond the role of vector. Seasonal chronology of infections: relapses, recrudescences and gametocyte production (a) Temperate regions The vernal (late winter – early spring) rise in the incidence of P. vivax in temperate regions is legendary, and was discussed in depth by Swellengrebel and De Buck (1938) [ 20 ]. Such vernal fevers occur in the apparent absence of Anopheles . Korteweg proposed that such fevers were the result of infections incubating from the previous year. Induced infection studies, implementing P. vivax and P. falciparum to treat neurosyphylitic patients, by Swellengrebel and others confirmed Korteweg's hypothesis that vernal fevers were due (at least in part) to latent vivax infections from the previous autumn [ 20 - 22 ]. A variety of strains were used in the course of such treatment, leading to the conclusion that there are inherent differences in the seasonal behaviour of strains, depending on their geographical (latitudinal) origin. James (1931) [ 21 ], using the Madagascar strain, found that only 1.6% of the infections had extended incubation periods (200–317 days); similarly Boyd & Kitchen (1949) [ 22 ], using the McCoy (southern USA) strain, found only 1.1% with incubation periods in excess of 75 days. In contrast, Swellengrebel and de Buck (1938) [ 20 ], using the Dutch strain, found 38% with protracted incubation periods. Thus, temperate strains were considered to exhibit inherent seasonal behaviour, emerging when the insect vector commences its seasonal activity. Transmission to mosquitoes is achieved through the production of gametocytes. Ziemann (1914) [ 23 ] noted that a feature of the P. vivax spring relapses was a tendency for gametocyte production just before mosquitoes appeared. This is a feature of other haemosporidian parasites, most notably Leucocytozoan and Haemoproteus spp. that are widespread parasites of birds and lizards, the latter of which have recently been shown to be paraphyletic with Plasmodium species [ 24 ]. These parasites differ from Plasmodium spp. in maintaining the infection exclusively by exo-erythrocytic forms and only gametocytes are present in the circulating blood system; seasonal activity and the timing of gametocyte production are thus simultaneous in these genera. There is a rapid appearance of gametocytes in March and April, which is generally too early for novel infections by simulid black flies or ceratopogonid midges [ 25 ]; rather infection relapse appears to be related to altering hormone levels prior to the bird's reproductive season. Haemoproteus infections normally relapse during host breeding, peaking during egg laying and decreasing throughout the nestling rearing phase [ 26 ]. This reduction is attributable to an increase in immune function accompanied by a down-regulation of hormone levels [ 27 ]. An influence of sex hormones on gametocyte production is also suspected for Plasmodium infections in lizard malaria species. Gametocyte production is intensified during mid-late summer in Plasmodium mexicanum infections in lizards when similar seasonal fluctuations occur in testosterone levels. Although testosterone does not seem to affect overall parasite density and the course of infection, elevated testosterone significantly reduced variation in the timing of the onset of gametocyte production [ 28 ]. (b) Tropical regions and P. falciparum In contrast to temperate and subtropical zones, seasonal patterns of parasite relapse in parasites such as P. vivax and Hepatocystis spp. (tropical relatives of Haemoproteus infecting baboons) in the tropics is less evident and thought not to occur [ 29 ]. Unlike temperate zones, seasons in the tropics are based largely on rainfall and are, therefore, intimately linked to mosquito bionomics. Given this intimate link between season, mosquitoes and transmission, how can seasonality in P. falciparum be detected? As stated previously, current wisdom holds that there is a continual production of low numbers of gametocytes that are the source of infection once mosquito numbers expand. However, contrary to accepted belief, mosquitoes can be found at all times of the year, even under very hostile climatic conditions [ 30 ] and seasonal mosquito activity is, therefore, primarily one of greatly increased numbers. Thus, as an initial test of transmission seasonality, the occurrence of infected "out of season" anophelines is examined. The traditional explanation for the absence of infections in anophelines is that climatic conditions reduce the lifespan of the adult mosquitoes such that the probability of their surviving long enough to allow completion of sporogonic development is negligible – that is, the vectorial capacity [ 31 ] is considerably reduced by the decreasing vector longevity. However, by examining the mosquito for oocyst (midgut) as well as sporozoite (salivary gland) infections, the extent of human infectiousness can be established irrespective of whether there is actual transmission. Bentley (1911) [ 32 ] summarizes several studies of this nature where anopheline stomachs and salivary glands were examined on a seasonal basis (Table 1 ). Although admittedly very limited, these data argue against a persistent level of human infectiousness to mosquitoes. Table 1 Comparison of the oocyst and sporozoite rates in mosquitoes sampled during the non-transmission and transmission seasons. Data from Bentley (1911) [32]. Author Place Season N° anophelines dissected for oocyst/sporozoites Oocyst positive (%) Sporozoite positive (%) Bentley 1911 Bombay Nov-June 178 / 123 0 0 July-Oct 659 / 703 12.7 4.1 Gosio 1905 Tuscany Apr-June 318 0 0 July-Oct 512 27 4 Daniels British Central Africa Dry season 1500 0 0 Such anopheline data are one side of the coin and gametocytes are the other. If tropical seasons are defined most precisely by mosquito bionomics, it is conceivable that Plasmodium spp. would have evolved to respond to the mosquitoes themselves. Thus, in a manner akin to temperate parasite species that utilize vertebrate host seasonal cues to produce gametocytes at the optimal time, tropical species may respond to tropical cues – the mosquito bites. Is there any evidence that P. falciparum , for example, produces gametocytes in response to mosquito bites? As with P. vivax , P. falciparum gametocyte production has been found to peak at a time when the anopheles abundance was at a maximum [ 33 - 35 ]. However, this peak gametocyte rate occurred notably after the peak in the number of clinical cases and, thus, after the peak in transmission. Such an increase in gametocyte rate would be expected as a result of novel infections and, thus, characteristic of an endemic region during the transmission season. However, what evidence is there that, at the beginning of the transmission season, there is an increase in infectiousness of humans to mosquitoes that is not due to novel infections, but due to relapses in existing chronic infections. Identifying the human reservoir of infection and correlating gametocyte density with transmission success to mosquitoes is far from clear [ 36 ]. The human reservoir of infection (during the non-transmission season) will depend on the rate of recovery from infection, which in turn depends on the extent of previous exposure and the immune reaction to infection. Both these factors will alter with age for a given intensity of transmission. Therefore, in the absence of novel infections, the human reservoir of infection will have an age-specific distribution. If mosquito bites are having a gametocyte-promoting effect on existing chronic infections, then age-specific patterns of gametocyte production might be expected. However, although increasing gametocyte density tends to result in greater infectiousness to mosquitoes [ 19 , 37 , 38 ], it has been repeatedly demonstrated that high gametocyte densities do not guarantee high mosquito infection rates [ 19 , 38 - 40 ]. Cryptic infectors with no or very few apparent gametocytes, are capable of infecting mosquitoes and may contribute to a very significant proportion of the human transmission reservoir [ 39 - 41 ]. Moreover, gametocyte density varies greatly according to the region of study, and also with age (i.e. history of exposure) of the individual and tends to reflect the overall asexual parasite density [ 42 , 43 ]; consequently, infections in the younger and therefore less immune individuals tend to produce higher densities of gametocytes [ 44 - 46 ] than adults. Therefore gametocyte density itself may not be a sufficiently sensitive indicator. Rather, age-specific seasonal changes in intra-host parasite prevalence rates are preferred. This measure encompasses changes in both sexual and asexual prevalence rates and the tendency for an infection in a particular age group to produce gametocytes; it is, thus, a measure of parasite behaviour within a particular age-group during a certain season. To examine the potential role of mosquito bites on the seasonal nature of P. falciparum transmission, three historical data sets, addressing age-specific longitudinal patterns of P. falciparum asexual and sexual prevalence rates in relation to mosquito abundance patterns, are analyzed. One of the major shortcomings of these studies is that they did not have recourse to PCR (quantitative, reverse transcriptase) technology and thus will have underestimated parasite prevalence rates and can not distinguish a relapsed chronic infection from a novel infection. However, despite such limitations, the data sets reveal much on the seasonality of P. falciparum transmission. Study 1 : Barber & Olinger (Ref: 47.1931 – Urban, infant/mother, Lagos, Nigeria) For a period of 18 months, records of vaccine visits to the Lagos local health office of 3–4 month old babies with their mothers provided P. falciparum seasonal prevalence data. During the same period, intensive surveys of anopheline activity and mosquito infection prevalence rate (both salivary gland sporozoite and midgut zygote (oocyst) infection rates) were carried out throughout Lagos by indoor resting catches in from 200 to more than 400 rooms. Infants (three to four months) are no longer expected to be protected by maternal antibodies and prevalence rates in this very young "naïve" age group provide a good indication of the force of infection and hence the current transmission intensity. Mothers will include a distribution of older age groups that are expected to have developed some immunity to infection and disease. It should be noted that although Plasmodium infection characteristics in pregnant and recently post-partum women differ from non-gravid women of similar ages [ 48 , 49 ], the mothers here are three to four months post-partum and so likely to be representative of the adult population. One confounding factor of the data set is that urban populations are mobile and individuals may thus have acquired infections elsewhere. Although it is likely that mothers and infants would have travelled together and thus be exposed to identical mosquito biting rates, the absence of paired mother-infant data weaken any comparison of parasite rates. Using the mosquito data, the estimated monthly entomological inoculation rate (EIR) (mosquito density × mosquito biting rate × sporozoite rate) and the estimated human reservoir of infection (infectious gametocyte rate) are calculated, using the formula of Macdonald (1952) [ 50 ]. where s (sporozoite rate), (infective mosquito lifespan), a (human biting rate; here estimated as 0.33) and x is the proportion of human infections that are infectious to mosquitoes (hereon referred to as estimated gametocyte infectious rate. Note this is not necessarily equivalent to the gametocyte rate). p can be calculated from the measured total mosquito infection rates: sporozoite rate, s , and zygote rate, z . where n and m are the number of days required for the development of identifiable presence of sporozoites and oocysts respectively (i.e. the extrinsic incubation periods for sporozoites and oocysts). Here standard values of n = 12 and m = 3 [ 45 , 50 ] were taken using the Moshkovsky scale where and a mean annual temperature T fluctuating around 25°C. In this way, values for x can be estimated, thereby enabling examination of how the infection rates in the two directions, mosquito man (EIR) and man mosquito (infectious gametocyte rates), relate to the age-specific (three to four month old infants versus mothers) seasonal changes in parasite prevalence rates and the mosquito activity patterns (Figs. 1a & 1b ). Figure 1 Mother-infant seasonal P. falciparum prevalence rates during routine vaccination visits. Data adapted from Barber & Olinger (1931) [47], Lagos, Nigeria. (a) P. falciparum prevalence rates in three month old infants and their mothers. Anopheline mosquito inoculation parameters from selected houses throughout Lagos include the Entomological Inoculation Rate (EIR), which is the number of infected mosquito bites per person per month; the Uninfected biting rate, which is the number of uninfected mosquito bites per person per month; the sporozoite rate, which is the percentage of mosquitoes with sporozoites in their salivary glands. (b) Anopheline biting rates (as above) and the relative proportion of mosquitoes with zygote (oocyst stage infections) compared with sporozoite stage infections from which the estimated gametocyte infectious rates can be calculated (see text). Interpretation Start At the beginning of the transmission season, there is a dramatic increase in anopheline biting density and, although the sporozoite rates do not increase, EIR increases from one to nine infectious bites per month (Fig. 1a ). Paradoxically, however, although there is an increase in adult parasite prevalence rates, there is a decrease in infant prevalence rates (Fig. 1a ). This latter, more sensitive marker of the force of infection, would suggest that any increase in transmission is offset by a greater infant recovery rate or that the increased EIR is not yet manifest in this age group. The estimated gametocyte infectious rate and the proportion of zygote vs. sporozoite infection prevalences decrease (Fig. 1b ), further suggesting that the increase in EIR has little impact in generating novel or super infections that lead to gametocyte production. Peak By contrast, during the "Peak" of transmission, following the peak EIR (June), there is a parallel increase in prevalence of infection in both the young and older age groups (Fig. 1a ), a large increase in both estimated gametocyte infectious rate and zygote/sporozoite proportions (Fig. 1b ), prior to the decrease in mosquito numbers and longevity (August). Thus during this phase, the occurrence of novel/super infections is notable during the month after peak EIR (June) and is signalled not only by increases in prevalence of infection across all age groups (especially in the infant age class), but also by the increase in the estimated gametocyte infectious rate (Figs. 1a & 1b ). The EIR follows the sporozoite rate which increases in the absence of the emergence of new adult mosquitoes (note the drop in the mosquito activity). End In the final "End" phase, prevalence rates drop markedly and notably more rapidly in the younger age class (Fig. 1a ), indicative of a greater rate of recovery from infection. The estimated gametocyte infectious rate drops, the zygote/sporozoite proportions return to an "equilibrium" value expected from the low and relatively stable mosquito densities. Transmission is, thus, minimal as this point, as indicated by the very low infant rates and the low estimated gametocyte infectious rate. The data in the Peak and End phases follow the classic description of a malaria transmission season, most especially confirming the sensitivity of the infant versus the adult age groups in revealing changes in the transmission intensity and the relationship between transmission and the (estimated) patterns of gametocyte dynamics. However, the Start phase highlights a novel effect whereby there is little effective transmission (infant rates or estimated gametocyte rates) but an increase in adult prevalence rates, which notably parallels the dramatic increase in overall biting rate, predominantly by uninfected anophelines. This pattern suggests that the increase in mosquito activity per se may be having an effect on parasite prevalence rates in the adult population. In high endemicity areas, subpatent infections are common and likely to occur in older age groups with a previous history of exposure [ 51 ]. This contrasts with infants who respond to infections with strong fever and cytokine reactions, thus reducing the duration of infection. Such age-specific differences in immune response and duration of infections has been highlighted by molecular epidemiological studies of multiple clone infections by P. falciparum in highly endemic areas, which have demonstrated age dependence in both the multiplicity of infection and the relationships between this multiplicity and the risk of acute illness: in older children, a high multiplicity of infection is characteristic of low-level chronic parasitaemia [ 52 , 53 ]. In areas of high transmission, parasitaemias are likely to be determined mainly by the interaction of schizogony and anti-blood stage immunity, leading to periodic fluctuations in levels of parasitaemia [ 54 ]. Although the early increase in the parasite rates of the mothers in this reported study of Barber & Olinger (1931) [ 47 ] may be simply the result of stochastic variation in periodicity, the coherence of infant and mother parasite rates during the other phases does suggest that there may be a real biological cause underlying this apparently anomalous increase: that insect bites per se cause the re-emergence of existing infections in this adult population. This hypothesis is to some extent corroborated by observed patterns of parasite prevalence versus parasite density in an adult population in Liberia [ 55 ]. Here, at the start of the transmission season, the prevalence of infection decreased but the mean parasite density increased. Rather than invoking superinfection, which would be at odds with the reduced prevalence rates, mosquito bites could actually be resulting in an increase in asexual (and thus sexual) parasite densities. Such age-specific effects have also been noted by Muirhead-Thomson [ 56 ] during longitudinal studies in Jamaica (Fig. 2 ). Anopheles albimanus productivity from larval collections was at a maximum at the very time when sudden cold spells were producing a wave of relapses and a crop of gametocyte carriers. At that critical period, at the beginning of the malaria season, the increase in gametocyte rate was particularly marked in the group >7 years of age. In this age group, the gametocyte rate increases from 1.1% in the summer to 17.3% in the rainy season. Muirhead-Thomson [ 57 ] suggested that these age groups were more often bitten and that therefore age-specific differences were simply a result of differential exposure; that is the older age group were subject to a higher force of infection. Although such an explanation is possible, the differences in the seasonal prevalence rates are negligible. By contrast, the relative differences in gametocyte rates in the young and older age groups are marked: in the younger age groups the proportion of P. falciparum infections positive for gametocytes varied little with season (from 67 to 62%), whereas in the older age group the change was marked, increasing from 22 to 61%. Why should the two groups display very different changes in gametocyte production but similar increases in prevalence rates? Could mosquito biting be having an effect on infections in the older age group? Figure 2 Age-specific (less than or greater than seven years old) seasonal P. falciparum all stage ( Pf positive) prevalence rates; those with gametocytes and the proportion of P. falciparum positive individuals also with gametocytes. Data from Muirhead-Thomson, Jamaica (1952) [56]. Study 2 : Wilson (Ref. 44.1936 – Rural, active case detection, Gombero, Tanzania) The second data set comes from a longitudinal study in a rural village in Tanzania characterised by seasonally intense transmission. In this study both gametocyte and asexual prevalence rates were measured intermittently (every two or three months) and classified into three age groups : <5 years, 6–20 years and 20+ years. In addition, anopheline activity and sporozoite rates were measured on a monthly basis. Wilson noted distinct seasonal fluctuations in gametocyte rates, whereas overall parasite prevalence rates varied little. The longitudinal age-specific parasite and gametocyte prevalence rates are shown in Fig. 3a . How the proportion of parasite positive individuals that also have gametocytes changes over time and by age group is shown in Fig. 3b . This is calculated simply as the proportion at time t+1 divided by the proportion at time t (1 is then subtracted from this figure such that a value of 0 indicates no change). The mosquito activity patterns and sporozoite rates are given in Fig. 3c . Figure 3 Seasonal study of P. falciparum prevalence rates in a rural village in Tanzania. Data from Wilson (1936) [44]. (a) Age-specific parasite prevalence rates (all stages and those with gametocytes). Three age groups are considered: less than five, from six to 20 and greater than 20 years old. (b) Age-specific rate of change in the proportion of infected individuals that also have gametocytes (i.e. "(Proportion gametocyte positive at month 'm+1'/proportion at month 'm') - 1)". A value of 0, therefore, means no relative change in the proportion of infected individuals with gametocytes. (c) Entomological parameters measured in selected houses. Number of infected (EIR) and uninfected mosquito bites per person per month and the sporozoite rate. Interpretation The parasite prevalence rates change very little but do generally follow the monthly fluctuations in EIR; the absence of monthly data points, however, abnegates any rigorous comparison. The age-specific prevalence rates are characteristic of hyperendemic transmission intensity with peak rates in the very youngest age group. As discussed in the previous study [ 47 ], the occurrence of asymptomatic chronic infections would be expected to increase with age. Despite missing monthly data points for the human parasite prevalence data, comparison of the three graphs highlights several distinct points. Most notably, although overall gametocyte rates increase most significantly in the infant age group, the proportional changes are most significant in the older age groups (January) and notably in the adult age group (July). The January rise in gametocyte rates follows the October – December increase in EIR and thus could be indicative of novel/super infections. However, the prevalence rates scarcely change in any age group. Although prevalence rate is considered an insensitive marker of alterations in transmission intensity, here they do clearly decrease in March following the absence of transmission in February and markedly increase during the unusually sustained transmission season from March to October. Superimposed on this sustained seasonal transmission is a significant increase in the gametocyte production rate in the adult population (July). As shown in Figs. 3a,3b,3c , this coincides with large increases in mosquito abundance (June & July). Several features argue against this being due to novel/super infections : (i) the age-specific nature of this increase, (ii) the absence of significant EIR in June and (iii) the absence of any additional increase in adult prevalence rates over and above that following the general trend from March to October. By contrast, following this gametocyte rise in July, there is an increase in sporozoite rates and hence the EIR and subsequent increases in overall prevalence rates in October, i.e. the July gametocyte rise provides the source of infection. As observed in the previous study, there appears to be a distinct effect of mosquito biting per se on parasite prevalence rates in the adult population. Study 3 : Rosenberg (Refs. 58–60 1990 – Rural, Active case detection, Thailand) Rosenberg et al . (1990) [ 58 - 60 ] conducted a two year longitudinal study in a rural farming village with hyperendemic P. falciparum and P. vivax malaria in S.E. Thailand. During the two year study period, monthly human and mosquito data were collected. Despite the very different nature of malaria epidemiology in Thailand, when compared with the previous two African studies, Rosenberg et al . (1990) [ 60 ] noted very similar seasonal, mosquito associated fluctuations in gametocyte density to those observed in the Wilson study [ 44 ] – « Nonetheless, the pattern we observed in Thailand was strikingly similar to that which Wilson (1936) described for Tanzania 50 years earlier; the similarity is all the more remarkable as the 2 sites have virtually no features in common other than seasonally intense transmission. » Rosenberg et al . (1990) [ 60 ] discussed at length the seasonal fluctuations in gametocyte prevalence rates and noted that the perennial cycle of malaria incidence was more evident in the high trophozoite densities and the gametocyte prevalence rates than the gross prevalence rates. Such parasite fluctuations were interpreted as being the result of superinfections in 50% of the cases. The remaining 50% were considered to be the result of novel infections. However, they stated their uncertainty as to why such a considerable incidence of novel infections « ......did not inflate prevalence soon after transmission started » [ 60 ]. The pertinent data are displayed in Figs. 6a-c with the addition of the uninfected mosquito biting density [ 59 ]. In addition, from Rosenberg et al . [ 60 ] Table 1 and text, the following supplementary details are given: (i) 14+ age group harboured 49% of the gametocytes but 84.6% high gametocyte densities (<20/500 white blood cells) occurred in the group <14 years of age; (ii) dry season gametocyte prevalence was 2.6x higher than the wet season prevalence in <14 age group vs . 8x higher in the 14+ age group ; (iii) in year two, the EIR decreased by 67% and, whereas the gametocyte prevalence decreased in <14 age group, it increased in the 14+ group. Interpretation The younger age groups have higher gametocyte densities reflecting their generally higher parasite loads. Younger age groups also tend to have higher rates of acquisition and recovery (i.e. turnover) and, therefore, most gametocytes would be expected to be due to novel infections. This is supported by the finding that gametocyte prevalence decreased in the <14 group in year 2 when inoculation rate and force of infection was less. Gametocyte prevalence increased in the 14+ group despite a decrease in inoculation rate in year two; note, however, that despite the drop in EIR, the mosquito biting intensity was higher. Thus, the apparent paradox of there being considerable novel infections at the beginning of the transmission season, but with no influence on parasite prevalence, may be explained by the re-emergence of pre-existing chronic infections in the 14+ age group. This group is more likely to harbour such infections and is shown here to display the most significant seasonal changes in gametocyte rates [ 60 ]. The sequence of events in this study could thus be interpreted as follows : In year one, the mosquito biting rate increases dramatically in October, with a concomitant monthly rise in oocyst rates, but not sporozoites (data not shown here), and no rise in gametocyte rates. In November, the sporozoite rate and EIR increase as expected from the previous month's oocyst rates and the decrease in newly emerged mosquitoes (evident from the parous rate [ 59 ]). Despite the absence of high P. falciparum trophozoite density infections, suggesting an absence of novel infections, there is a small rise in gametocyte rates. The following month (December), this increased EIR leads to a dramatic increase in high P. falciparum trophozoite density infections and gametocyte prevalence rates. A similar pattern is seen in the second year, and notably, although the EIR and sporozoite rates were lower, resulting in reduced gametocyte prevalence rates in children, the peak mosquito biting density (which again preceded the increase in sporozoite rate) was higher, as was the gametocyte rate in adults. Whereas the large concomitant increases in high P. falciparum trophozoite density infections and gametocyte prevalence rates can be taken as evidence for novel/super infections, the increase in gametocyte rates in the absence of sporozoites at the start of transmission, coupled with the variation in age-specific gametocyte rates between the two years, suggests once more that mosquito biting itself may induce gametocyte production and subsequently augment mosquito infection rates. Conclusions Here, a case is presented for there being a role of uninfected mosquito bites in increasing the human reservoir of infection at the beginning of the transmission season. The available data is limited, but consistently suggest that the traditonal view that there is a rapid expansion of the parasite population from a few source infections following the seasonal increase in anophelines is overly simplistic. It must, however, be emphasised that in no way is it suggested that once transmission is under way, the epidemiology of malaria is not satisfactorily described by classic Ross-Macdonald models or variants thereof [ 10 ]. Rather, it is simply proposed that the image of intense parasite activity and dynamics during peak transmission detracts from a more subtle process occurring at the very beginning of the season. That is, Gill [ 11 , 12 ] may have been correct in suggesting that a subset of the human population becomes increasingly infectious to mosquitoes at the onset of mosquito activity [ 1 ] and that the factor responsible may be the dramatic increase in mosquito bites themselves. Moreover, it should be noted that mosquito bites do not need to be from anophelines. Nuisance culicine mosquitoes are often found to be well in excess of anophelines [ 61 - 63 ] and notably were so in the Rosenberg study [ 59 ]. Mosquito spp. differ in their seasonal bionomics and some culicines ( Aedes spp.), for example, have a capacity to increase more rapidly in numbers than anophelines; this is because of the ability of their eggs to resist dessication. Thus, the seed population of culicines is probably greater than that of anophelines, which must expand from pockets of aestivating females. Clearly this view is very preliminary, but the available information is encouraging and this hypothesis warrants closer examination in regions of endemic seasonal transmission. If, indeed, there is a role for mosquito bites, this could have far reaching possibilities, not only for novel intervention strategies, but also for strategies in combatting malaria in regions of seasonally intense transmission, including regions at risk of epidemics. If intervention measures can be taken to reduce the potential reservoir of infection, by targeted use of antimalarial drugs with gametocytocidal effects during the window of mosquito resurgence, for example, the increase in transmission intensity will be severely delayed. Moreover, reducing human-mosquito contact by the use of insecticide treated bednets (ITBNs) or repellents may have considerable impact at this crucial time. A mass action community level benefit of ITBNs has already been noted, where individuals not sleeping under nets in villages where bednets were implemented, had reduced mortality and morbidity rates [ 64 , 65 ]. ITBNs can clearly have several community level beneficial effects and it would be of considerable interest to see if ITBNs actually alter the transmissibility of the parasite as well as reducing the burden of disease. Authors' contributions RELP, MD and PTB were all involved in the conception of the article that arose out of lengthy discussions among the authors. The final draft was written by RELP with various sections added and/or removed by MD and PTB. Figure 4 Longitudinal study of P. falciparum prevalence rates in a rural village in Thailand. Data from Rosenberg et al (1990) [58-60]. (a) Entomological parameters measured in sentinel houses. Number of infected (EIR) and uninfected mosquito bites per person per month and the sporozoite rate. (b) P. falciparum high trophozoite density and gametocyte prevalence rates. (c) Age-specific P. falciparum prevalence rates. Two age groups are considered: Greater than and less than 10 years old.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535346.xml
535352	Density-dependence and within-host competition in a semelparous parasite of leaf-cutting ants	Background Parasite heterogeneity and within-host competition are thought to be important factors influencing the dynamics of host-parasite relationships. Yet, while there have been many theoretical investigations of how these factors may act, empirical data is more limited. We investigated the effects of parasite density and heterogeneity on parasite virulence and fitness using four strains of the entomopathogenic fungus, Metarhizium anisopliae var. anisopliae , and its leaf-cutting ant host Acromyrmex echinatior as the model system. Results The relationship between parasite density and infection was sigmoidal, with there being an invasion threshold for an infection to occur (an Allee effect). Although spore production was positively density-dependent, parasite fitness decreased with increasing parasite density, indicating within-host scramble competition. The dynamics differed little between the four strains tested. In mixed infections of three strains the infection-growth dynamics were unaffected by parasite heterogeneity. Conclusions The strength of within-host competition makes dispersal the best strategy for the parasite. Parasite heterogeneity may not have effected virulence or the infection dynamics either because the most virulent strain outcompeted the others, or because the interaction involved scramble competition that was impervious to parasite heterogeneity. The dynamics observed may be common for virulent parasites, such as Metarhizium , that produce aggregated transmission stages. Such parasites make useful models for investigating infection dynamics and the impact of parasite competition.	Background In most models of host-parasite dynamics, the parasites are considered as part of discrete infections, involving only a single parasite. However, as the transmission stages of parasites tend to be clustered, the majority of host-parasite interactions will rather involve multiple parasite individuals. This is particularly the case for those parasites that exhibit a semelparous life-history, releasing all their transmission propagules in a single event that normally coincides with host death [ 1 , 2 ]. In the same models, transmission is assumed to follow the mass action principle, βSI , where S and I are the densities of susceptible and infected individuals respectively and β is a constant describing the probability of infection [ 3 ]. Thus the probability of a host becoming infected per unit time will be directly proportional to the number of parasites that it encounters. Host-parasite interactions will also often involve multiple genotypes of parasites, further complicating their dynamics [ 4 , 5 ]. The co-occurrence of multiple parasites within a single host makes within-host competition between the parasites inevitable. Hosts represent limited resources and there is a carrying capacity for the total biomass of parasites that the host resources can support. The within-host growth of the parasites will thus normally be density-dependent, decreasing as the carrying capacity is approached [ 6 , 7 ]. Interactions between different parasite genotypes are often considered to result in increased virulence, but have also been predicted to lead to decreased virulence depending on the dynamics [ 4 , 8 - 11 ]. The outcome of interactions will depend critically on both the scale of competition and the relatedness of the parasites involved, with higher relatedness selecting for reduced competition, or even cooperation, and more local competition offsetting this [ 12 - 14 ]. The nature of the within-host competition has further been characterized as a continuum between two extremes: superinfections, which exhibit contest competition with the most virulent parasites eliminating those less virulent, and coinfections, where the parasites differ little in virulence and resources end up being shared amongst individuals via scramble competition (or in its purest form by parasites exploiting different within-host niches and not competing at all) [ 8 , 15 , 16 ]. The mechanism of competition can be: (1) exploitation, with parasites competing for resources, (2) interference, with parasites, for example, producing antagonistic compounds, or (3) apparent, being mediated by the host immune system [ 5 ]. Exploitation competition is inevitable whenever parasites do not have completely separate niches within the host, while interference competition via antagonistic compounds drives, for example, the interaction between strains of entomopathogenic bacteria that produce bacteriocins [ 17 ]. Apparent competition has been argued to be the most important type [ 4 ], and can occur even in invertebrates with their less complex immune systems. For example, malaria parasites suppress the immune response of their mosquito vector [ 18 ], and various entomopathogenic fungi have been shown to produce immunodepressant compounds [ 19 - 23 ]. In spite of the fundamental importance of interactions between different strains of parasites, empirical studies of mixed infections are relatively rare and a number of them have produced results that conflict with the theoretical predictions [ 5 , 24 ]. Most studies have found either that mixed infections are more virulent than single infections [ 25 - 29 ], or that virulence equalled that of the most virulent strain [ 23 , 25 , 30 - 32 ]. Interactions can be more complex though, and may also depend on environmental conditions, host genotype, or on the parasite genotypes involved [ 25 , 32 - 35 ]. For example, the virulence of bacteriocin-producing bacteria matches the winning strain only when one strain can kill the other, but is reduced compared to single infections when both strains can kill each other [ 17 ]. Furthermore, consideration of virulence only does not provide a full picture of the complexity of the interaction. The production of transmission stages may be increased [ 26 - 28 , 30 ] or decreased [ 31 ] during mixed infections, and may also depend upon the order of infection [ 32 ]. Although the finding that virulence matched that of the most virulent strain might suggest that that strain has outcompeted the less virulent strain, in some studies the transmission stages produced were from the both strains [ 29 , 31 , 33 ], or even entirely from the less virulent strain [ 23 ]. Here we examine the infection dynamics of the parasite M. anisopliae var. anisopliae (Metschnikoff) (Deuteromycotina: Hyphomycetes) in the leaf-cutting ant Acromyrmex echinatior Forel (Hymenoptera: Formicidae: Attini). M. anisopliae var. anisopliae is a generalist entomopathogenic fungus that is known to infect leaf-cutting ants [ 36 - 41 ], as well as many other insects. Infection may be from sporulating cadavers or from spores dispersed in the soil. Metarhizium spores have a tendency to remain attached to one another, making even dispersed spores likely to be clustered, and soil has been estimated to contain as many as 1,000 to 50,000 spores g -1 [ 41 ]. Spores germinate and penetrate directly through the host cuticle without first growing over the surface of it as some other fungi do. Host individuals will thus often be infected by multiple parasite propagules. In addition, the group-living life-style of leaf-cutting ants, as with other social insects, makes them especially prone to being exposed to multiple infections under natural conditions [ 42 ]. Inside the host, the parasite produces blastospores and then hyphal bodies that release immunodepressant and antibiotic toxins [ 19 , 21 , 22 ]. After a period of time the host dies by some combination of the depletion of its resources due to the parasite infection, direct invasion of tissues by hyphae or the action of the parasite's toxins [ 21 ], and the parasite sporulates shortly after this. Metarhizium thus has a semelparous life-history and is an 'obligate killer' [ 1 ], producing transmission stages only after host death. Such parasites represent excellent models for studying infection-growth dynamics because the rate of successful infections equates exactly to host mortality, the time of host death will relate to the number of hyphal bodies within the host and thus acts as a gauge of within-host growth, and because the lifetime reproductive output of the parasite is represented entirely by the spores produced upon host death in contrast to other parasites that produce transmission stages continually [ 1 ]. Yet while applied studies using Metarhizium are common, fundamental studies of its infection dynamics are rare. There is a diversity of M. anisopliae var. anisopliae strains near leaf-cutting ant nests in Panama [ 41 ], indicating the potential for within-host competition between multiple parasite genotypes. We established and compared the dynamics of within-host competition in this system at two levels. We first investigated competition between parasites of the same genotype by establishing whether the infection rate of M. anisopliae var. anisopliae adhered to the mass action principle and whether parasite growth and fitness was density-dependent. We then examined whether the infection dynamics were consistent between different strains of parasite, and how the dynamics were affected by intraspecific within-host competition involving multiple parasite genotypes. We used two Panamanian strains that had had the opportunity to coevolve with A. echinatior and one exotic strain that had never previously encountered the host. A number of studies suggest that exotic strains are less competitive during within-host interactions than are native strains that have coevolved with the host [ 43 - 48 ]. We define parasite virulence to be parasite-induced host mortality as measured by case mortality and time-of-death. This differs from the instantaneous mortality rate used in many models but has been argued to be a more suitable measure of virulence [ 49 ]. Note that because the ant hosts were adult individuals that do not grow any further, the potential advantage to an obligate killer parasite of delaying host death to allow further host growth before semelparous reproduction [ 1 ], will not apply in this system. Results Experiment 1: intra-strain competition The dose of Metarhizium anisopliae var. anisopliae spores applied had a significant effect on ant mortality (Wald = 131.1, d.f. = 10, P < 0.0001) (Figure 1 ). The colony of origin did not affect either the dose-response relationship (Wald = 13.2, d.f. = 10, P = 0.213) or mortality overall (Wald = 1.56, d.f. = 1, P = 0.212). The mortality caused by the lowest two doses did not differ significantly from that in the controls (Figure 1 ). There was also no significant difference in mortality between the highest three doses because mortality was close to 100% at all these doses. The dose-mortality relationship consequently followed a sigmoidal pattern (F 3,7 = 162.5, P < 0.0001; Figure 2a ). Figure 1 Survival of ants in Experiment 1 treated with serial doses (spores/ant) of M. anisopliae var. anisopliae isolate KVL 02–56 or a control solution of 0.05% Triton-X (n = 60). Different letters indicate doses whose survival distributions differed significantly. Figure 2 Dose relationships for ants in Experiment 1 treated with serial doses of M. anisopliae var. anisopliae isolate KVL 02–56 or a control solution of 0.05% Triton-X. (a) Mortality at end of experiment ( y = -0.021 x 3 + 0.144 x 2 - 0.054 x + 0.171, r 2 = 0.986). (b) Proportion of dead ants sporulating ( y = 0.113 x + 0.269, r 2 = 0.700). (c) Mean number of spores (± SE) produced per sporulating ant ( y = 0.3692 x + 0.6692, r 2 = 0.4081). (d) k -values for ants in Experiment 1 treated with serial doses of M. anisopliae var. anisopliae isolate KVL 02–56 ( y = 0.809 x + 0.468, r 2 = 0.976). The dashed line has a slope of 1 and is included for comparison. (e) Per capita fitness (dashed line) and probability of infection (solid line; calculated by multiplying the probability of death and of sporulation if death occurs). (f) Per capita fitness after adjusting for the probability of infection. There was no effect of the time between death and the assessment of spore production upon either the proportion of cadavers sporulating (Wald = 1.97, d.f. = 1, P = 0.16) nor the number of spores produced (F 1,127 = 1.13, P = 0.289). This supports the assumption that sporulation was complete at the time of assessment and also indicates that parasites that took longer to kill their host did not produce more spores. The proportions of cadavers sporulating were significantly less than expected based upon the number of ants estimated (from the control data) to have died from other causes (F 1,17 = 8.55, P = 0.0095). This indicates that certain ants killed by Metarhizium failed to sporulate. Both the proportion of ant cadavers sporulating (Wald = 19.4, d.f. = 9, P = 0.022) and the mean number of spores produced per sporulating cadaver (F 9,127 = 1.97, P = 0.048) increased significantly with dose (Figures 2b and 2c ). However the increase in spore production was relatively small, only doubling over the full range of doses examined. When the spore numbers produced were used to calculate k -values (which assess the density-dependence of growth, with a zero value indicating that spore production increases proportionally to dose and positive values indicating spore production is less than proportional to dose), it was found that k increased significantly with dose (F 1,9 = 231.1, P < 0.0001; Figure 2d ). The slope of the relationship (0.8) was significantly less than 1 (t = 2.42, d.f. = 8, P = 0.042). Correspondingly, the per capita fitness of the parasite decreased linearly with dose, while, in contrast, the probability of infection (calculated by multiplying the probability of death by the probability of a cadaver sporulating) increased sigmoidally (Figure 2e ). By multiplying these two variables, the overall fitness of the parasite can be calculated, and can be seen to decrease more or less linearly with dose (Figure 2f ). Experiment 2: inter-strain competition Ant survival was affected significantly by the concentration of M. anisopliae var. anisopliae spores (Wald = 186.0, d.f. = 5, P < 0.0001), the strain (Wald = 9.42, d.f. = 3, P = 0.024), and also the colony of origin (Wald = 13.5, d.f. = 4, P = 0.009). There were no significant interactions between strain and concentration (Wald = 11.8, d.f. = 15, P = 0.681). Mortality was generally positively correlated with dose, although the significance of pairwise differences between doses did vary somewhat between strains (Figure 3 ). In none of the strains was there any difference in survival between the lowest two doses, and, with the exception of strain 02–73, there was also no difference in survival between the highest two doses. Overall, ant survival was greatest when treated with the allopatric Ma275 strain, followed by the 02–73 strain (obtained from a soil sample at the collection site), from which it did not differ significantly (Breslow statistic = 2.45, d.f. = 1, P = 0.118). Survival of ants treated with Ma275 was significantly greater than of those treated with either the 02–72 strain (obtained from an Atta worker) (Breslow statistic = 8.36, d.f. = 1, P = 0.004) or the mixture of all three isolates (Breslow statistic = 5.00, d.f. = 1, P = 0.025) (Figure 3 ). The survival distributions of ants treated with the different strains did not otherwise differ significantly (02–72 vs. mixture: Breslow statistic = 0.84, d.f. = 1, P = 0.360; 02–73 vs. mixture: Breslow statistic = 0.26, d.f. = 1, P = 0.613; 02–72 vs. 02–73: Breslow statistic = 2.04, d.f. = 1, P = 0.153). Figure 3 Survival of ants in Experiment 2 treated with either Metarhizium anisopliae var. anisopliae isolate (a) KVL 02–73, (b) KVL 02–72, (c) Ma275 or (d) a mixture of all three. Different letters indicate doses whose survival distributions differed significantly. Although the proportion of ant cadavers sporulating was consistently high (>80%) in strain 02–73 and was positively correlated with dose in the other strains (Figure 4a ), the interaction between strain and dose was nonsignificant (Wald = 8.92, d.f. = 15, P = 0.882), as were both the main effects (strain: Wald = 2.81, d.f. = 3, P = 0.422; dose: Wald = 5.49, d.f. = 5, P = 0.359). Spore production from the sporulating cadavers was estimated by the ranking method described. It was found to be unrelated to the lag-time between death and ranking (F 1,270 = 0.489, P = 0.485) and to increase slightly with dose in all treatments (F 5,270 = 2.42, P = 0.036; Figure 4b ). The relationship between spore production and dose did not differ between strains (F 14,270 = 1.12, P = 0.342) and the strains also did not differ overall (F 3,270 = 0.19, P = 0.903). The k -values for all three strains were positively correlated with the number of spores applied (F 1,15 = 3612.9, P < 0.0001) and had similar slopes (F 3,15 = 0.78, P = 0.523; Figure 4c ). The values for Ma275, though, were significantly lower than for the other strains (F 3,15 = 13.71, P = 0.0001). Figure 4 Dose relationships and lines of best fit for ants in Experiment 2 treated with either Metarhizium anisopliae var. anisopliae isolate KVL 02–72 (circles; grey, dashed line), KVL 02–73 (triangles; black, dashed line), Ma275 (squares; grey, solid line) or a mixture of all three (diamonds; black, solid line). (a) Proportion of dead ants sporulating. Lines of best-fit are: KVL02–72: y = -0.0686 x 2 + 0.5434 x - 0.0014, r 2 = 0.8983, P = 0.032; KVL 02–73: y = 0.0292 x + 0.8339, r 2 = 0.5182, P = 0.107; Ma275: y = 0.1623 x + 0.0839, r 2 = 0.7699, P = 0.030; mixture: y = -0.0532 x 2 + 0.4286 x + 0.1476, r 2 = 0.9291, P = 0.019. (b) Spore ranks of sporulating ants. Lines of best-fit are: KVL02–72: y = -0.358 x 2 + 2.5553 x + 3.391, r 2 = 0.9148, P = 0.025; KVL 02–73: y = 0.411 x + 5.5392, r 2 = 0.7212, P = 0.032; Ma275: y = 0.222 x + 6.1677, r 2 = 0.116, P = 0.575; mixture: y = 0.1261 x 3 - 1.2959 x 2 + 3.7405 x + 4.265, r 2 = 0.9646, P = 0.833. (c) k -values. Lines of best-fit are: KVL02–72: y = 0.93 x - 1.18, r 2 = 0.992, P < 0.0001; KVL 02–73: y = 0.953 x - 0.982, r 2 = 0.999, P < 0.0001; Ma275: y = 0.972 x - 2.06, r 2 = 0.995, P = 0.0002; mixture: y = 0.991 x - 1.127, r 2 = 0.998, P < 0.0001. Discussion Parasite virulence, as measured by host mortality, was found to show a clear density-dependent pattern. This has also been recorded previously in Metarhizium [e.g. [ 50 - 53 ]] and other parasites [ 6 , 7 , 54 ]. Rather than being a linear relationship though, the pattern was sigmoidal. Mortality was not increased significantly by further increases in dose beyond 1 × 10 7 spores per ml. In addition, the mortality caused by doses of 1 × 10 4 spores per ml or less did not differ from that of ants treated with the control solution (Figure 2a ). All parasite-induced mortality was expected to have occurred by twenty days after treatment, even at the lower doses, and this assumption was supported by the levelling out of mortality seen in both experiments. It follows from this that the infection rate of Metarhizium is directly represented by the host mortality rate after twenty days. The lack of a difference in mortality between ants treated with the two lowest doses and the control ants suggests the occurrence of an Allee effect, with an invasion threshold for infection to be successful [ 54 ]. Such an effect has also been evidenced in some other studies of Metarhizium [e.g. [ 50 , 53 ]], and is a subtle effect that will only be detected when a sufficient range of doses is tested. Many models of host-parasite dynamics are based upon the mass-action principle, under which the infection rate is a linear function of the density of parasites a host individual encounters [ 3 ]. The sigmoidal pattern observed in the Metarhizium - Acromyrmex system indicates that, here at least, this principle applies only at intermediate doses. The Allee effect most probably relates to the effectiveness of the host defences against parasites. Leaf-cutting ant defences, as with most insects, consist of 'first-line' defences involving grooming and the secretion of antibiotic compounds on to the cuticle, and 'second line' defences based upon the cellular and humoral immune responses [ 37 , 40 , 55 - 58 ]. Spores will interact independently with the first line defences, the immune response may be saturated if it has to defend against very high numbers of parasites. This effect will be exacerbated by the toxins that the hyphae of many Metarhizium strains produce to incapacitate the immune system [ 19 , 21 , 22 , 59 ]. Greater doses of infecting spores will more quickly produce a larger pool of mycelium, which will produce greater quantities of these toxins and thus make the immune system decreasingly capable of mounting an effective response. Neither the probability of a cadaver sporulating nor the number of spores it produced were related to the time of host death, indicating both that parasites did not gain increased spore production by taking longer to kill the host, and that spore production was complete at the time of assessment. Both spore production and the proportion of cadavers sporulating were positively correlated with the within-host density of the parasite. However, the efficiency of the conversion of host biomass into parasite propagules was negatively density-dependent, as demonstrated by the k -values. The number of spores produced was less than proportional to increases in the dose of parasites applied, indicating the occurrence of density-dependent parasite growth and within-host competition. Although this competition was between genetically identical parasites in the first experiment, it still resulted in a reduction in parasite fitness because of the less efficient use of host resources. Were spore production to be unaffected by, or even inversely related to dose, as found in a study with Beauveria [ 60 ], then this negative effect on fitness would be even greater. It is important to note that this is based only upon the cadavers that sporulated. As both the proportion of cadavers that sporulated and the number of ants that died increased with dose, there is a trade-off for the parasite (Figure 2e ). Higher doses will result in more hosts dying and producing parasite spores, but where cadavers do sporulate, the per capita spore production will decrease as dose increases. Interestingly, this trade-off resolves itself such that increases in dose always result in a decrease in fitness (Figure 2f ). Although low numbers of parasite spores have only a small chance of successfully infecting a host, the strong effect of within-host competition makes dispersal the best strategy. This is even without taking into account the probability of at least one spore encountering a host, which will also be substantially increased by spores dispersing (and thus having multiple chances to encounter a host per unit time), rather than staying in a single aggregation (and having only a single chance per unit time to encounter a host). In other studies of obligate killer parasites, even stronger effects of within-host competition between propagules of the same parasite clone have been found [ 7 , 60 ]. It therefore seems likely that these dynamics may be broadly similar for most semelparous, obligate killer parasites, and that maximum dispersal is the best strategy for these parasites, as well as any others that exhibit strong within-host competition. When the different strains of M. anisopliae var. anisopliae were compared in the second experiment, they were found to differ in their virulence. Interestingly, it was the exotic strain (Ma275), originating from a different geographical location and from a different host order (Lepidoptera) that had the lowest virulence, and the strain isolated from a leaf-cutting ant worker (KVL 02–72) that had the highest. Many studies have found Metarhizium strains to vary in virulence [e.g. [ 53 , 61 ]]. Differences in strain virulence may in part be due to the variation in the production of destruxins that occurs between strains, but other virulence factors are undoubtedly also important [ 59 ]. As virulence is defined here as parasite-induced host mortality, the differences between strains could also be due to differences in the proportions of spores germinating and penetrating into the host rather than differences in within-host growth [ 62 , 63 ]. Although all strains had similarly high germination rates on artificial media, the interaction with host cuticle is more complex and involves various antibiotic compounds such as those produced by the metapleural gland [ 56 , 64 , 65 ]. Differences between strains in their susceptibility to such compounds would seem quite likely. Aside from the differences in virulence, however, the strains did not otherwise differ in their infection dynamics. They showed similar spore production and identical density-dependent growth patterns. The dynamics described above therefore appear to be consistent across strains, at least for those tested here. A fundamental assumption of most models of host-parasite relationships is that within-host competition is more intense, and results in heightened virulence, when it involves more than one parasite genotype [ 8 - 10 ]. However, in experimental studies virulence has often been found to be unaffected by parasite heterogeneity [ 24 ]. This was also the case in the current study, in which the virulence of the mixed infection was the same as that of the most virulent strain in the infection. In addition, the parasite k -value vs. dose relationship was identical for the mixed and single infections, indicating that the density-dependent growth patterns were unaffected by host heterogeneity. There are two possible explanations for the results. The most parsimonious is that the most virulent strain in the mixed infections simply outcompeted the other strains and drove them to extinction within the host. Such superinfection dynamics have previously been suggested for Metarhizium and other entomopathogenic fungi [ 32 , 66 , 67 ] and would be in accord with some models [ 6 ]. However other studies have found mixed infections to produce transmission stages from more than one strain of parasite in spite of the overall virulence matching that of the most virulent strain [ 29 , 31 ]. It remains possible that the mixed infections in the current study did not involve any of the strains being competitively excluded and that the spores produced came from all the strains. To distinguish between these possibilities it would be necessary to isolate and sequence monospore cultures from the sporulating cadavers, something would be an interesting objective for future work. Importantly, the infection dynamics did not differ between the single and mixed infections. The occurrence of within-host competition, whether it involves the production of antagonistic compounds or is mediated by the host's immune system, might be expected to force parasites to divert some resources from growth and the production of transmission stages to producing or coping with competitive mechanisms. While there is some evidence for this [ 31 ], the production of transmission stages has been increased during mixed infections in many other studies [ 26 - 28 , 30 ]. Clearly the outcome will depend upon the particular genotypes involved. If different strains produce antagonistic compounds that are effective against one another, as in the study by Massey et al. [ 17 ], then reduced transmission stages can be expected. If they do not, and if they exploit different within-host niches, produce cooperative compounds (such as iron-binding agents [ 13 , 14 ]), or act synergistically to depress the host immune system, then increased transmission stages can be expected. The fact that spore production was the same in single and mixed infections in the current study therefore suggests either that the most virulent strain outcompeted the others without suffering any cost from the competitive interaction, or that the Metarhizium strains were engaged in scramble competition with dynamics that are impervious to interactions being inter- or intraclone. Conclusions The importance of within-host competition between parasites is well illustrated by the Metarhizium - Acromyrmex system studied here. Even though the probability of a successful infection was increased substantially by parasites occurring in aggregations, the effect of competition between parasite propagules of the same clone makes dispersal the best strategy. It seems likely that this is generally true for semelparous parasites. Further investigations of the impact of within-host competition on parasite fitness are needed and should endeavour to establish where on the superinfection-coinfection continuum the interaction lies by identifying which parasites produce transmission stages. The fact that the production of transmission stages of semelparous parasites, such as Metarhizium , is concentrated into a single bout, and thus that their lifetime fitness can be readily quantified, makes them excellent models for doing this. Methods General methodology Colonies of A. echinatior were collected from Gamboa, Panama, and maintained in the lab under controlled conditions (ca. 24°C, 70% RH) on a diet of bramble leaves ( Rubus fruticosus ) and rice grains. For the experimental replicates, large workers (head width 2.1 to 2.4 mm) were removed from their colonies and placed individually in plastic pots (diameter: 2.5 cm, height: 4 cm) where they were maintained at 24°C with an ad libitum supply of water and sugar water. A number of isolates of M. anisopliae var. anisopliae were collected from the vicinity of leaf-cutting ant nests at the same location in Gamboa, Panama, and were cultured as monospore isolates on Sabouraud dextrose agar [ 41 ]. Spore (conidia) suspensions were made by flooding agar plates with mature spores with a sterile solution of 0.05% Triton-X and scraping off the spores with a glass rod. The spores were centrifuged and washed three times with sterile 0.05% Triton-X solution with intervening centrifugation steps. The concentration of spores was then quantified using a haemocytometer and diluted to the required concentration. The viability of the spore suspensions was checked by spreading 100 μl of them on to Sabouraud dextrose agar plates and counting the proportions of spores that had germinated after 12–16 hours at 24°C. Spore viability was >95% in all cases. Ants were treated with M. anisopliae var. anisopliae by applying 0.5 μl of a spore suspension to their thorax using a micropipette. Spore suspensions were vortexed thoroughly immediately prior to application to ensure spores were fully dispersed. Control ants had 0.5 μl of a 0.05% Triton-X solution applied in the same way. Following application, ant mortality was assessed daily for a period of twenty days. Based on previous work [ 23 , 40 , 56 ], this time period was judged sufficient to ensure that all parasite-induced mortality had occurred by the end of the experiment. Dead ants were surface sterilised [ 68 ], and placed in a petri dish lined with damp filter paper. After the completion of the experiments, the cadavers were left for a further ten days in order to allow full sporulation of the parasite. The level of sporulation on the cadavers was then assessed by one of two methods. In Experiment 1, sporulation was quantified by directly counting the number of spores on the cadavers. The cadavers were placed in individual vials with 1 ml of 0.05% Triton-X solution and vortexed for 1 min to remove the spores into suspension. The concentration of spores in the suspension was then quantified with a haemocytometer. In Experiment 2, sporulation was estimated by examining the cadavers under a binocular microscope and giving each a rank of between 0 (no spores visible) and 10 (cadaver almost completely covered by spores) depending upon the level of sporulation. Based on data collected prior to these experiments, spore ranks estimated in this manner correlate well with the actual number of spores on the cadavers ( y = 0.159 x - 0.862, r 2 = 0.746, Spearman's r = 0.820, N = 73, P < 0.01) and thus provide a reliable estimate of spore production. Experiment 1: intra-strain competition The experiment involved M. anisopliae var. anisopliae isolate KVL 02–56, which had been isolated from the dump pile of an Atta colombica nest in Gamboa, Panama [ 41 ]. A spore suspension was made up and serially diluted five-fold to give concentrations from 1 × 10 8 to 5 × 10 3 spores ml -1 (equivalent to an average of 50,000 and 2.5 spores per ant respectively). Given that soil at the site in Panama has been estimated to contain as many as 1,000 to 50,000 spores g -1 [ 41 ], it seems likely that this range encompasses the natural doses that the ants are exposed to. Thirty ants from each of two colonies of A. echinatior (Ae47 and Ae109) were treated with each of these doses, or with the control solution, and their survival monitored for twenty days after application. The effect of dose and colony of origin on ant mortality was analysed with a Cox proportional hazard regression model to examine the effect of parasite density (dose) on virulence and infection rate. This incorporates both case mortality and the time of death. Pairwise comparisons of the doses were done using Kaplan-Meir survival analyses and the Breslow statistic. The ant cadavers were left for ten days after the end of the experimental period, in order to allow ample time for all cadavers to sporulate fully. The numbers of cadavers sporulating and the numbers of spores produced by these cadavers were assessed with binary logistic and general linear models respectively. These data were used to calculate k -values to assess whether parasite growth was density dependent, as done previously by Ebert et al. [ 7 ]: k i = log 10 (M 0 D i / M i D 0 ) where k i is the k -value at dose i , M is the number of spores produced, D is the number of spores applied, and M 0 and D 0 represent the number of spores produced and applied respectively at the lowest dose tested. A zero value for k indicates that the number of spores produced at dose i are exactly proportional to the increased dose (so, for example, a doubling of dose results in a doubling of the number of spores produced). The equation gives a positive value for k when the number of spores produced at dose i are less than proportional to the increase in dose. Zero values for k thus indicate that spore production is independent of parasite density, while positive values of k indicate that spore production is negatively density dependent. Experiment 2: inter-strain competition To establish whether the dynamics recorded in the first experiment were consistent for different strains of the parasite, three isolates of M. anisopliae var. anisopliae were compared. These were KVL 02–73 (isolated from soil at the field site from which the ant colonies were collected), KVL 02–72 (isolated from an Atta colombica leaf-cutting ant worker), and the strain Ma275 (isolated from Cydia pomonella (Lepidoptera: Tortricidae) in Germany). The strains therefore represented a range in terms of their potential coevolution with A. echinatior , with the former two being likely to have had some interaction while Ma275 would not previously have encountered A. echinatior . Each strain was diluted tenfold and tested at six doses from 5 × 10 8 to 5 × 10 3 spores ml -1 . In addition, a mixed spore suspension was made up to examine inter-strain competition. The suspension contained equal numbers of spores of each of the three strains. This was serially diluted and applied at the same doses as the individual isolates. Four ants from each of five colonies of A. echinatior (Ae48, 109, 143, 153 and 154) were treated with each dose of each isolate or with the control solution and their survival monitored for twenty days. The survival of the ants was analysed as in the preceding experiment to examine if strains differed in virulence and if the growth of each strain was density-dependent. Ten days after the end of the experimental period, spore production was assessed by ranking each cadaver for the amount of sporulation as described earlier. These ranks were used to estimate the actual number of spores produced and these values were then used to calculate k -values. A regression analysis was carried out on these data in order to assess if the growth of each strain was density-dependent and whether the density-growth dynamics differed between strains. Authors' contributions WOHH conceived the study, assisted with the experiments, analysed the results, and drafted and revised the manuscript. KSP, LVU and DP carried out Experiment 1 and KSP also participated in Experiment 2. MP assisted with both experiments. LT assisted with mycological aspects and JJB provided support throughout. All authors contributed to the writing of the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535352.xml
549528	Patient characteristics and clinical management of patients with shoulder pain in U.S. primary care settings: Secondary data analysis of the National Ambulatory Medical Care Survey	Background Although shoulder pain is a commonly encountered problem in primary care, there are few studies examining its presenting characteristics and clinical management in this setting. Methods We performed secondary data analysis of 692 office visits for shoulder pain collected through the National Ambulatory Medical Care Survey (Survey years 1993–2000). Information on demographic characteristics, history and place of injury, and clinical management (physician order of imaging, physiotherapy, and steroid intraarticular injection) were examined. Results Shoulder pain was associated with an injury in one third (33.2% (230/692)) of office visits in this population of US primary care physicians. Males, and younger adults (age ≤ 52) more often associated their shoulder pain with previous injury, but there were no racial differences in injury status. Injury-related shoulder pain was related to work in over one-fifth (21.3% (43/202)) of visits. An x-ray was performed in 29.0% (164/566) of office visits, a finding that did not differ by gender, race, or by age status. Other imaging (CT scan, MRI, or ultrasound) was infrequently performed (6.5%, 37/566). Physiotherapy was ordered in 23.9% (135/566) of visits for shoulder pain. Younger adults and patients with a history of injury more often had physiotherapy ordered, but there was no significant difference in the ordering of physiotherapy by gender or race. Examination of the use of intraarticular injection was not possible with this data set. Conclusion These data from the largest sample of patients with shoulder pain presenting to primary care settings offer insights into the presenting characteristics and clinical management of shoulder pain at the primary care level. The National Ambulatory Medical Care Survey is a useful resource for examining the clinical management of specific symptoms in U.S. primary care offices.	Background Shoulder pain is a common clinical problem in the ambulatory setting. The one year prevalence of shoulder pain is as high as 50% in the general population, and 50% of those afflicted consult a physician [ 1 , 2 ]. As many patients with shoulder pain miss work because of the condition, it should be no surprise that the costs associated with shoulder pain are high. Examining the clinical management of shoulder pain in primary care settings, where the vast majority of patients present, is essential to improving the quality of care and to understanding the associated costs [ 3 ]. However, studies of the clinical management of shoulder pain usually come from small select populations in orthopedic clinics. To our knowledge, there have been only three published studies of shoulder pain conducted in primary care settings, and none of them were carried out in the United States. The National Ambulatory Medical Care Survey offers a means of studying how common clinical conditions are managed by United States primary care physicians in a large, nationally representative sample. We sought to examine the presenting characteristics and clinical management of patients presenting to primary care physicians for evaluation of shoulder pain. In addition to offering insights on the clinical management of shoulder pain, this investigative strategy serves as a model for using this national data set to examine the quality of musculoskeletal care. Methods Data for this study comes from the National Ambulatory Medical Care Survey (NAMCS), 1993 to 2000. Conducted by the National Center for Health Statistics (Hyattsville, Maryland), the NAMCS survey uses a multistage probability sample design [ 4 ]. Using the master lists of all US physicians from the American Medical Association and American Osteopathic Association, a sample of patient care physicians is selected each year by random, stratified by geographic area and specialty. Among identified physicians, annual participation ranges from 74% in 1989 to 68% in 1998 (63% in 1999) [ 5 , 6 ]. For participating physicians, patient visits during a randomly selected week are sampled systematically. For each selected patient visit, the physician completes a visit form that details patient, physician, and clinical information. Patient information includes demographics, insurance status, and up to three reasons for the visit. Physician information includes self-selected specialty, geographic location, and if the practice is in a metropolitan area. Information on clinical management includes which diagnostic and therapeutic maneuvers took place at the time of the visit or were ordered as a result of the visit. Up to three reasons for the office visit were solicited by the survey. Using the NAMCS categorization scheme for reason for visit, we extracted all visits for which shoulder pain (number 14900–14950) was a reason for the visit [ 7 ]. We limited data to patients aged 18 or older, and to physicians who were self-reported practitioners of internal medicine, and family practice. We combined the data from the most recent years of the survey (1993–2000) to define a set of 3023 visits for shoulder pain for further analysis. Each visit is assigned a weight derived from the probability of being sampled, to account for regional and specialty sampling bias as well as nonresponse. Sampling weights are often used to produce national estimates based on the available sample. Because the weighting scheme of NAMCS was not based on symptom, and the decision to analyze data based on a single specific condition such as shoulder pain precludes the use of weights to produce national estimates, we present our other results as unweighted analyses (Korn). All analyses were conducted using JMP-SAS (version 5.10a, SAS Institute Inc, Cary, NC) and Stata (version Intercooled Stata 8.0, College Station, TX). This sample of patients with shoulder pain was characterized by age, race, and gender. We reported whether the shoulder pain was the result of an injury, whether the injury was work related, and how the injury states differed by demographic characteristics. Younger versus older adults were designated by using the median age of 52 years of age to separate age groups. ICD-9 diagnosis codes were reported for each patient (International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9-CM)). Clinical management of the shoulder pain was examined by analyzing the proportion of patients for which plain x-rays and more advanced imaging (CT scan, MRI, or ultrasound) and the proportion of patients for whom physiotherapy was ordered. Demographic differences for imaging and physiotherapy orders were examined. We attempted to characterize the treatment of shoulder pain by examining (1) whether physiotherapy was ordered by the physician and (2) whether the physician administered an intraarticular steroid injection. For the physiotherapy issue, specific question was part of the survey during survey years 1995–2000. However, determining whether a corticosteroid was administered was not possible. The survey solicits whether an office surgical procedure was performed during the office visit, but the performance of a minor procedure such as arthrocentesis is not usually considered a surgical procedure. The coding of up to six medications administered or prescribed during the office visit in the NAMCAS survey should allow identification of corticosteroid medications used in a intraarticular injection. However, the route of administration (intramuscular, intraarticular, topical, etc.) is not specified by the survey, and the availability of a given corticosteroid in various preparations makes for uncertainty in assigning as truly including determining whether an intraarticular injection took place during an office visit. Results Figure 1 shows how the cohort of office visits for shoulder pain was assembled. For adult patients (aged 18 or older), shoulder pain was given as a reason for the visit in 3023 office visits during the years 1993 through 2000 of the survey. Of these office visits for shoulder pain, 692 (22.8%, 692/3023) were to general internists (n = 327) or family practitioners (n = 365). The mean number of patient visits contributed by any one physician was 1.55 (maximum, 8; minimum, 1; median, 1) and 1.39 (maximum, 4; minimum, 1; median, 1), respectively. The ICD codes most commonly listed for shoulder pain included 72610 rotator cuff syndrome of the shoulder (9.6%, n = 24), 71941 shoulder pain (8.5%, n = 21), 176210 (6.0%, n = 15), 71590 osteoarthritis (5.6%, n = 14), 72690 tendinitis of an unspecified site (4.8%, n = 12), but 99 different codes 176210 (6.0%, n = 15), were used by this group of physicians. Figure 1 Cohort Assembly of Primary Care Office Visits for Shoulder Pain in the U.S. from National Ambulatory Medical Care Surveys (1993–2000). A cohort of patients ≥ 18 years of age with shoulder pain were assembled from office visits to self-reporting internal medicine and family practice physicians. Presenting characteristics of patients with shoulder pain The mean age of the patients seen at these visits was 53.3 ± 17.8 years, and 54.3% (376/692) of the visits were for female patients. White patients outnumbered blacks (84.6% (586/692) versus 12.1% (84/692) in this data set. The proportion of patients whose shoulder pain was a result of an injury was 33.2% (230/692). Males, and younger adults were more likely to have had an injury associated with the shoulder pain (males 36.7% versus females 30.3%, p = 0.01; age >52 19.3% versus age <52 46.9%, p < .0001), but there were no race differences in injury association (whites 33.8% versus blacks 28.6%, p = .50). For the survey years 1995–2000 when the specific question was asked, the proportion of office visits for shoulder pain from injury that were related to work was 7.6% (43/566) of all visits to these primary care physicians and 21.3% (43/202) of those associated with injury. Diagnostic imaging During the survey years 1995–2000 when the specific question was posed regarding the performance or ordering of an x-ray, 164/566 (29.0%) of the visits results in an x-ray order, a finding that did not differ by gender (29.0% for males versus 29.0% for females, p = 0.46), by race (29.8% for whites versus 24.3% for blacks, p = 0.65), or by age status (older adults 27.8% versus younger adults 30.0%, p = 0.43). Whether an x-ray was performed was not associated with a history of injury 27.2% versus 30.0%). Advanced imaging (CT scan, MRI, or ultrasound) was performed in 6.5% (37/566) of visits. Therapeutic interventions for shoulder pain Physiotherapy was ordered in the case of 23.9% (135/566) of visits. There was no significant difference in the ordering of physiotherapy by patient gender (female 27.0% versus male 20.1%, p = .052) or by patient race (white 25.1% versus African-American 17.1%, p = .070), but younger adults were more likely to have physiotherapy ordered (younger adults 18.7% versus older adults 28.7%, p = .005). Patients with a history of associated injury were more likely to receive an order for physiotherapy (injury 36.1% versus no injury 17.0%, p = <.001). As discussed above, it was not possible to determine whether intraarticular steroids were administered. Discussion The labeling of the years 2000–2010 as the "Bone and Joint Decade" is, in part, a worldwide plea for better understanding and management of common musculoskeletal conditions [ 8 ]. Although musculoskeletal complaints are among the most common reasons for physician consultation, clinical management of these conditions is not well understood. There are few published studies of the management of shoulder pain in the primary care setting where it most commonly presents. To our knowledge, the only published studies of shoulder pain in the primary care setting come from the Netherlands and the United Kingdom. Van Der Windt et al examined the characteristics and management of intrinsic shoulder disorders for 349 patients from eleven Dutch general practices during a one year period (1995) [ 9 ]. Croft et al reported a prospective cohort study of 166 patients consulting twelve British general practitioners for shoulder pain during the year [ 10 ]. More recently, Hay et al conducted a randomized controlled trial of corticosteroid injection versus physiotherapy in 207 patients from nine general practices Britain [ 11 ]. Clinical management of musculoskeletal disorders in the Untied States should be different, given the differences in health care delivery and reimbursement. However, we are unaware of any studies of shoulder pain in the United States primary care practices. Our findings represent the largest study to date to examine the characteristics and clinical management of shoulder pain at the primary care level in the United States. Although our data come from a cross sectional survey and the findings are primarily descriptive, these findings provide insights into how U.S. primary care practitioners experience and manage patients with shoulder pain. The demographic characteristics of patients of these primary care settings were similar to those of other studies of shoulder pain in primary care settings [ 9 , 11 ] with a female predominance (54.3% of all patients presenting with shoulder pain), and a wide age range of patients (53.3 years ± 17.8). However, studies of primary clinic populations differ from studies of shoulder pain in the general population where shoulder pain increases in prevalence with age, even to geriatric populations [ 12 ]. Our findings support the idea that older patients with shoulder pain do not seek or are not brought to medical attention as frequently as younger adults [ 9 ]. Racial characteristics of patients with shoulder pain have not been previously reported, but overall the demographic characteristics, reflect those of all patients presenting to U.S clinic settings with no salient differences. The proportion of patients whose shoulder pain was a result of an injury was 33.2% (230/692) in this study, higher than in the van der Windt study where 12% of patients gave a history of injury and 13% of strain/overuse with unusual activities [ 9 ]. For survey years 1995–2000 when the specific question was posed, the percent of shoulder pain related to work was 7.6% (43/566) of all visits for shoulder pain and 21.3% (43/202) of those associated with injury. While there are no comparable data from primary care settings regarding the circumstances of injury, studies from occupational settings show that many factors influence the occurrence of shoulder pain in work settings [ 13 ]. Imaging was performed in 29.0% of patients in this study of U.S office visits, a marked contrast to the British study of primary care management where only 2% of the patients presenting with shoulder pain underwent x-ray studies [ 9 ]. The value of radiographic plain films may be of limited value, but plain x-rays are still recommended as an early diagnostic step in primary care settings [ 14 ]. Because nearly half of all patients who present with shoulder pain have a prior history of that condition [ 9 ], and our study did not distinguish between incident and chronic shoulder pain, the high proportion of patients receiving x-rays in our study suggests overuse of this procedure. However, other imaging procedures such as Magnetic resonance imaging, computerized tomography, or ultrasound were infrequently ordered. Comparable studies from other primary care settings are not available, to our knowledge [ 15 ]. A specific question about the clinician's actions allowed us to determine that physiotherapy was ordered a rate of 24%, comparable to the rate of 30% reported for British general practitioners [ 9 ]. Interest in the comparative value of physiotherapy versus steroid injection led us to attempt the same comparison in this data set [ 11 ]. However, there was no reliable mechanism for determining whether the ordering or performance of corticosteroid injection took place during the office encounter. Several limitations of using this data set for investigation of shoulder pain deserve mention. A precise diagnosis for the shoulder pain would be desirable. The large number of ICD-9 diagnosis codes assigned by the clinician illustrates well the problems in defining and managing this syndrome. However, this problem of imprecise diagnosis of shoulder pain is well known and is not unique to this data set [ 9 ]. Second, as this is a cross sectional study with secondary data analysis, the amount of data available in the data set does not allow for absolute certainty in following the clinical reasoning process. As an example, while we can be certain that an imaging procedure or physiotherapy was ordered for a given office visit, albeit self-report by the clinician, we cannot be certain that the imaging procedure or physiotherapy was specific to the shoulder. Third, hypothesis testing and statistical inference is difficult with data derived from a multilevel sampling strategy [ 16 ]. In addition to seeking insights into the clinical management of shoulder pain, we were interested in exploring whether this data set based on a symptom complex was possible and meaningful. The NAMCS, a series of annual surveys conducted since 1990 in the United States, has been utilized to study health service utilization, patients with known diagnoses, and prescribing behaviour of office-based physicians. To our knowledge, there are no published studies focusing on specific symptoms. The advantages of using this data set for analyzing clinical management of symptoms includes a well organized and standardized classification system of symptoms, a large number of office visits, and the systematic sampling strategy. While our attempt to explore meaningful clinical issues was hampered by the nature of the data set, we nevertheless succeeded in offering insights into the presenting characteristics and clinical management of shoulder pain for this population of patients. Conclusions Shoulder pain was associated with an injury in one third of office visits in this population of US primary care physicians. Males, and younger adults were more likely to relate their shoulder pain to injury, but there were no racial differences in injury status. Shoulder pain from injury was related to work in over one-fifth of office visits. An x-ray was performed in nearly one third of office visits, a finding that did not differ by gender, race, or by age status. Other imaging (CT scan, MRI, or ultrasound) was infrequently performed. Physical therapy was ordered in one quarter of visits for shoulder pain. Younger adults and patients with a history of injury were more likely to have physiotherapy ordered but there was no significant difference in the ordering of physiotherapy by gender or race. Examination of the use of intrarticular injection was not possible with this data set. The National Ambulatory Medical Care Survey is a useful resource for examining the clinical management of specific symptoms in U.S. primary care offices. Competing interests The author(s) declare that they have no competing interests. Authors' contributions JW conceived of the study, and participated in its design and coordination, and drafted the manuscript. RM and RW participated in formulating the analysis strategy. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549528.xml
516776	Hepatitis B virus variants in an HIV-HBV co-infected patient at different periods of antiretroviral treatment with and without lamivudine	Background Lamivudine inhibits replication of both human immunodeficiency virus (HIV) and hepatitis B virus (HBV) and is commonly used as part of antiretroviral therapy. The main limitation in the use of lamivudine is resistant mutation selection. Most of these mutations affect the YMDD motif of the HBV DNA polymerase. The resistance occurs through M550V or M550I aminoacid replacements. The M550V variation may be accompanied by L526M mutation, notably in HIV-HBV co-infected patients. The aim of this study was to investigate mutations associated with lamivudine resistance in a hemodialysis patient chronically co-infected with HIV-1 and HBV, who was submitted to several antiretroviral treatments. Methods HBV isolates derived from three blood samples collected at different times of antiretroviral therapies with and without lamivudine, were titred and submitted to nucleotide sequencing. Results HBV isolate derived from a sample collected in 1999 during an antiretroviral treatment with lamivudine showed the lamivudine resistant double mutation (L526M, M550V). However, no mutation associated with lamivudine resistance was observed in the HBV genome derived from the sample collected during a period of treatment without lamivudine (2001). After reinstitution of lamivudine (2002), the predominant HBV population exhibited a rare triple mutation (V519L, L526M, M550V), which has previously been associated with an in vitro reduction of virus antigenicity (escape mutant). HBV DNA was detected at high levels (10 8 –10 9 copies/ml) in the three blood samples. Conclusions Reintroduction of lamivudine as part of antiretroviral treatment in a patient who had developed lamivudine resistant HBV strains favored the predominance of an HBV isolate with reduced antigenicity. The absence of hepatitis acute exacerbation in this patient may be correlated to the absence of significant variations of the viral load, which was independent of the presence of mutations in the HBV DNA polymerase.	Background Hepatitis B virus (HBV) and human immunodeficiency virus (HIV) share common routes of transmission, mainly sexual, parenteral and vertical. Therefore, the prevalence of HBV serological markers is higher among HIV infected patients than in non-HIV infected individuals [ 1 , 2 ]. Considerable variations in the prevalence of HBV markers have been observed in HIV patients according to the geographical region and risk of exposure [ 3 ]. Recently, a prevalence of 68% of antibodies against hepatitis B core antigen (anti-HBc) was observed in HIV infected patients living in Rio de Janeiro, Brazil [ 4 ]. Co-infection with HIV interferes with the natural history of HBV infection and is associated with higher HBV DNA levels [ 5 , 6 ]. A more common progression to cirrhosis, despite a milder histological necro-inflammatory activity, has also been observed in cases of HIV-HBV co-infection [ 5 ]. Lamivudine is a nucleoside analogue that inhibits the reverse transcriptase activity of both HIV and HBV [ 7 ] and is commonly used in the treatment of both viral infections [ 8 , 9 ]. The HBV DNA loss due to lamivudine treatment is usually accompanied by significant histological and biochemical improvement [ 10 ]. The major limitation in the use of lamivudine is the selection of resistant mutations that may arise and accumulate during therapy. Most of these mutations usually affect the YMDD motif of the HBV DNA polymerase, by replacement of a methionine residue at position 550 with either valine (M550V) or isoleucine (M550I). Such mutations have notably been reported in HIV-HBV co-infected patients [ 11 , 12 ] who develop lamivudine resistance at an annual rate of 20%, with a projected rate of 90% after four years of therapy [ 13 ]. The consequences of drug resistance mutations for the evolution of HBV induced liver disease are currently under study. Hepatitis B acute exacerbation has been described after withdrawal of lamivudine therapy [ 14 ]. Such an exacerbation has also been associated with the appearance of YMDD mutants [ 14 , 15 ] and with a rapid increase of viral load [ 16 ]. The present study was performed to characterize HBV variants and genetic patterns of lamivudine resistant HBV strains in a patient co-infected with HIV-1 at different periods of an antiretroviral treatment with and without lamivudine. Methods Patient characteristics and serological markers A 30-year-old male patient initiated a hemodialysis treatment in August 1995. At that time, he was negative for hepatitis B surface antigen (HBsAg) as well as for anti-HBc and anti-HIV antibodies. The patient received three blood transfusions during the first trimester of 1996 and developed a posttransfusional hepatitis. He became HBsAg and anti-HIV positive and remained positive for both markers in all routine tests performed between 1996 and 2002. Data of other tests, such as serum transaminases, CD4 levels, and HIV load, performed in external laboratories between 1995 and 2002, were available. Resistance to different HIV therapies was evaluated by the medical staff based on HIV load and CD4 levels. HIV treatment started in April 1996, with a combination of zidovudine (AZT) and didanosine (ddI). In November 1996, this treatment was modified to a combination of AZT and lamivudine which was maintained until March 1997, when treatment was once again changed to triple therapy with AZT, lamivudine and invirase. In May 1999, high HIV load (3.1 × 10 5 copies/ml) was observed, indicating HIV resistance. At that time the patient started an anti-tuberculosis treatment with rifampicin. However, due to interaction of antiretroviral drugs with rifampicin, the administration of antiretroviral drugs was discontinued. This resulted in a decrease of CD4 counts to 156 cells/ml, while elevated levels of HIV persisted at the end of 1999. Antiretroviral treatment resumed in January 2000, composed of four drugs, namely ddI, stavudine, nevirapine (NVP) and nelfinavir (NFV). As a consequence, CD4 cell counts increased noticeably, and HIV RNA became undetectable. Unfortunately, resistance to therapy was observed during the second semester of 2001, when HIV load reached 1.2 × 10 5 copies/ml and CD4 counts decreased to 120 cells/ml. The treatment was then replaced by another including lamivudine associated to AZT, NVP and NFV. No HIV resistance was observed until the end of the follow-up (November 2002). The patient died of renal failure at the end of 2003. DNA extraction and PCR assays for DNA sequencing Three serum samples collected in March 1999, August 2000 and November 2002, were available for HBV DNA analysis. Sera were submitted to DNA extraction by the phenol-chloroform method after treatment with proteinase K, as described previously [ 17 ]. HBV pre-S/S genomic region was amplified by PCR using sense primer PS1 (5'-CCATATTCTTGGGAACAAGA-3', nt 2826-2845) and a mix of antisense primers S2 (5'-GGGTTTAAATGTATACCCAAAGA-3', nt 841-819) and S22 (5'-GTATTTAAATGGATACCCACAGA-3', nt 841-819), able to amplify all HBV genotypes. PCR assays were performed under the following conditions: 94°C, 30 s; 52°C, 1 min; 72°C, 2 min; 35 cycles, followed by a final elongation of 7 min at 72°C. Amplification products (10 μL) were loaded on a 2% agarose gel, electrophoresed, stained with ethidium bromide, and visualized under UV light. Quantification of HBV DNA Quantification of HBV DNA was performed by endpoint dilution. DNA samples were diluted to a tenfold series up to 10 -8 dilution. Each dilution was submitted to PCR with oligonucleotide pairs PS1–PS2, designed in the pre-S region and C1–C2 (core region), as described previously [ 4 ]. Serial dilutions and PCR assays for HBV DNA quantification were performed in triplicate. The sensitivity of the method has been estimated to about 100 copies per PCR reaction [ 18 ], equivalent to 10 4 HBV genome copies per milliliter of serum. Quantitative results were estimated by dividing 10 4 copies/ml by the last positive dilution. Molecular cloning and nucleotide sequencing Pre-S/S PCR products were cloned into pCRII plasmid vector using TA cloning kit (Invitrogen, San Diego, CA). For nucleotide sequencing, recombinant plasmid DNAs were purified by a commercially available kit (Plasmid midi kit, Qiagen, Hilden, Germany). Nucleotide sequences were determined using the Cy5 auto read sequencing kit (Amersham Biosciences, Little Chalfont, UK) with M13 universal and reverse primers, as well as internal, HBV specific primers. For direct sequencing (S region), PCR products were extracted from low melting agarose gels (Qiaquick gel extraction kit, Qiagen). Sequencing reactions with HBV specific primers were done using the thermo sequenase Cy5 dye terminator sequencing kit and analyzed on an ALFexpress automated sequencer (Amersham Biosciences). Independent plus and minus strand sequencing was completed. Phylogenetic analysis Nucleotide sequences were aligned using PILEUP (Wisconsin Sequences Analysis Package GCG, Madison, WI). A phylogenetic tree was generated by neighbour-joining analysis of genetic distances, using the TREECON software package for Windows [ 19 ]. HBV sequences available from the GenBank database (accession numbers AY090458, M57663, U55220-U55222, J02201 and X51970) were used for the construction of a phylogenetic tree. Results Markers of HBV infection The patient under study developed an acute hepatitis and became anti-HIV positive in 1996, soon after receiving blood transfusions during the initial period of hemodialysis treatment. He became an HBV chronic carrier, with all blood samples collected between 1996 and 2002 being positive for HBsAg and HBeAg. During this period, the patient remained asymptomatic without clinical signs for HBV infection. Aspartate aminotransferase and alkaline phosphatase levels remained normal all during the follow-up. However, an increase of alanine aminotransferase (ALT) levels to 180 IU/L was observed in March 1999 (Figure 1 ). HIV treatment started in April 1996. Lamivudine was introduced as part of antiretroviral treatment in November 1996. HBV DNA was detected at high levels (10 8 copies/ml) in the blood sample collected in March 1999, during this first period of lamivudine treatment (Figure 1 ). In May 1999, lamivudine was discontinued, and the sample analyzed during the period of lamivudine interruption (August 2001) showed the highest (10 9 copies/ml) HBV DNA titre mesured in this study. HBV DNA was again detected at high levels (10 8 copies/ml) after reintroduction of lamivudine (November 2002, Figure 1 ). HBV variants Nucleotide sequences of HBVs (pre-S/S region) derived from HBsAg/HBeAg positive samples collected in March 1999, August 2001 and November 2002 were determined (three clones each). Phylogenetic analysis showed that all nine clones belonged to genotype A. This genotype has been subdivided into two subgenomic groups, designated A-A' (genotype A excluding A') and A' [ 20 ]. Recently, subgroups A-A' and A' were designated respectively as A1 and A2 [ 21 ] or Ae and Aa [ 22 ]. Isolates belonging to subgroup A' have been first identified in South Africa and circulate in a high proportion among HBV Brazilian isolates [ 23 ]. As can be observed on the phylogenetic tree represented in Figure 2 , all clones belonged to subgroup A' and were closely related to each other. The six clones obtained during the two periods of lamivudine treatment (1999 and 2002) clustered separately (with a bootstrap value of 86%) from those obtained in 2001, during the period of lamivudine interruption. Table 1 shows the main amino acid changes observed in the S region of the genome. Two substitutions, H359Y in the polymerase and Y100C in the surface antigen (small S), were observed in all HBV sequences. These two changes were natural variation, characteristics of the HBV strain infecting the patient under study. All three clones derived from the sample collected in 1999, during the first lamivudine treatment, showed two lamivudine resistant mutations (L526M and M550V), also detected in the major viral population by direct sequencing. One clone (1-B57) showed an additional mutation related to drug resistance (V519L) that was not detected by direct sequencing. None of these three mutations associated with lamivudine resistance was observed in sequences derived from the second sample (2001). In contrast, all three clones derived from the third sample (2002) showed the three lamivudine resistant mutations. However, by direct sequencing of PCR products of this last sample, the electropherogram could detect two nucleotides (A and G) at the same sequence position indicating a mixture of V519 and L519 residues. Curiously, all HBV sequences obtained during the period without lamivudine treatment (2001) displayed a unique G473E substitution in the polymerase gene. Stop mutations were also observed in the small S protein, which was truncated at position 182 in clone 2-B14 and at position 216 in clone 2-B62 (Table 1 ). Due to overlapping of polymerase and S genes on the HBV genome, mutations at positions 473, 519 and 550 of the polymerase were accompanied by mutations at positions 119, 164 and 195 of the small S protein (Table 1 ). Discussion Among lamivudine resistant mutations, those in the YMDD motif of the HBV DNA polymerase are the most common. The resistance occurs by replacement of a methionine residue at position 550 by either valine (M550V) or isoleucine (M550I). More rarely, HBV variant presenting M550S replacement may be selected during lamivudine treatment [ 24 ] The M550V variant may be accompanied by a mutation (leucine to methionine) at position 526 [ 25 ]. In the absence of HIV infection, the mutation at position 550 alone has been found in up to two-thirds of patients with lamivudine-resistant chronic hepatitis B. However, more than 90% of HIV-HBV co-infected patients display the double lamivudine resistant mutation at positions 526 and 550 [ 12 ]. Furthermore, the presence of L526M mutation in addition to mutation at position 550 seems to be associated with prolonged lamivudine treatment [ 26 ]. The double mutant has been shown to exhibit a 15-fold decrease of the in vitro susceptibility to lamivudine [ 27 ], since each mutation contributes to the loss of lamivudine sensitivity [ 28 ]. It is believed that the L526M mutation, when not accompanied with a mutation in the YMDD motif, does not confer lamivudine resistance [ 29 ]. Even so, the L526M mutation has been found alone in patients under lamivudine therapy [ 30 ]. Here, both L526M and M550V mutations were detected in all HBV sequences derived from the two blood samples collected during lamivudine treatment (1999 and 2002). Besides these common mutations, a third lamivudine resistant, rare mutation [ 11 , 31 ], namely V519L, was detected in one clone derived from the sample collected in 1999 as well as in all HBV sequences derived from the sample collected in 2002. This was in agreement with recent observations, showing that long lamivudine treatment may result in the predominance of this rare mutant [ 32 ]. The triple mutation V519L, L526M, M550V causes the concomitant amino acid substitutions E164D and I195M in the small S protein. It has been shown that this triple mutant has a reduced in vitro affinity to anti-HBs antibodies, similar to the hepatitis B vaccine escape mutant G145R [ 11 , 31 ]. The accumulation of mutations in the HBV genome should be monitored in order to guide patient management adequately. There is a general consensus that the lamivudine resistant single mutants in the YMDD motif (M550V/I) replicate substantially more slowly than the wild type. The addition of the mutation at position 526 may act as a compensatory change that partially restores the replication fitness of the virus [ 33 ]. Here, in agreement with this observation, HBV DNA was detected at high levels (about 10 8 copies/ml) in the blood sample collected in 1999, in which the double mutant represents the major viral population. Similar high HBV loads were observed in the sample containing the triple mutant (2002). This result is in agreement with a recent study, showing that V519L mutation also enhances viral replication [ 34 ]. The HBV population that emerged during the interruption of lamivudine treatment did not show mutations at polymerase positions 519, 526 and 550. However, another substitution, namely G473E, was observed that was associated to G119R substitution in the small S protein. Such a virus might be present as a minor population but not detected during the first lamivudine treatment. Although two out three clones of this population possessed stop codon mutations in S gene, HBV load (10 9 copies/ml) was moderately higher during the lamivudine interruption period than that found during lamivudine periods. HBV chronically infected patients submitted to lamivudine treatment may have acute exacerbation after withdrawal of drug therapy or when lamivudine resistance emerges [ 8 , 14 - 16 , 26 , 35 - 37 ]. In HIV-HBV co-infected patients, withdrawal of lamivudine may result in severe [ 8 , 35 , 37 ] or fulminant [ 37 ] hepatitis. Factors that trigger severe hepatitis in these cases are not well known. Both viral load and genome variations have been implicated in the pathogenesis of acute exacerbation. Here, the patient under study did not present clinical signs of HBV infection. The absence of acute exacerbation after the emergence of lamivudine resistant variants may be correlated to the absence of significant variations of the viral load. Indeed, HBV DNA was detected at high levels (10 8 –10 9 copies/ml) at different periods of antiretroviral treatment with and without lamivudine. Conclusions A rare HBV triple mutant, belonging to genotype A, subgroup A', appeared predominantly in a patient submitted to lamivudine as part of HIV treatment. This type of mutant, previously found in isolates belonging to genotypes A, D and G [ 15 , 36 ], may behave as a vaccine escape mutant. Understanding the circumstances leading to the appearance of such HBV strains may help to guide future therapies in HIV-HBV co-infected patients. A correlation may exist between acute exacerbation of hepatitis B and HBV load in lamivudine treated patients. Competing interests None declared. Authors' contributions EAS and MVS carried out cloning and sequencing of HBV DNA. JA was involved in clinical evaluation of the patient and supervised the antiretroviral treatment. SAG conceived the study, participated in its design and coordination, and wrote the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516776.xml
340964	Comparing the Networks that Power Bacterial Chemotaxis	null	When we think of foraging for food, we usually imagine animals wandering in the woods, poking behind bushes and trees, trying to find something tasty. Amazingly, even single-cell bacteria display a simplified version of this behavior. Many species of bacteria can respond to chemical or nutritional cues (chemoattractants) in the environment by moving toward locations with more favorable conditions, a process known as chemotaxis. The bacteria adjust their movements by rotating threadlike projections called flagella either clockwise or counterclockwise; these adjustments are made by a network of proteins in response to chemoattractants. Chemotaxis has been identified in many bacterial species, but two of the best-studied examples are in Escherichia coli and Bacillus subtilis . Computer modeling of chemotaxis in these species now reveals some important differences in the network architecture that controls this complex behavior. Most of the proteins involved in chemotaxis in E. coli and B. subtilis have been identified and well studied, but much remains to be learned about this biological process. As scientists have begun to understand how the proteins work together, they're discovering a network of interactions that operates a bit like an electronic circuit. Researchers have found that using the circuit as a model for protein networks has helped them to understand how complex system properties arise from seemingly simple interactions between proteins. These properties can be explored with the aid of computer simulations, whereby researchers can rapidly test a given system under many different situations and can tweak the properties of the proteins and their connections. The team, led by Adam Arkin of the University of California at Berkeley, has compared the system level properties of chemotaxis in the two bacterial species E. coli and B. subtilis . Not surprisingly, the proteins involved in the signaling pathway are conserved—that is, they have changed very little since they first evolved—even though these species are evolutionarily very distant. In many cases, a gene from one species can even substitute for the ortholog (a conserved gene that retains the same function even though two species have diverged) in the other. Despite these similarities, however, disrupting the function of orthologous genes in these two species often has different, even opposite, effects. This is surprising, especially given that the chemotactic behaviors of E. coli and B. subtilis are almost identical. In order to understand this puzzling observation, the researchers constructed a network model of the chemotaxis system in B. subtilis and used simulations to understand how the network properties differ from those of existing models of E. coli chemotaxis. The group found that despite the similarities in proteins and the nearly identical behavior between the two species, the mechanisms underlying the behavior are quite distinctive. When comparing the system properties of these two bacterial systems, the researchers also made an unusual observation. Though the two “circuits” have different wiring, the system properties underlying the behavior, called the control strategy, are very similar. The two species of bacteria therefore achieve the same chemotaxis behavior by using similar proteins, but in different ways. Arkin and colleagues draw two important conclusions from these results. First, these two systems have conserved proteins, but the proteins are wired together differently. This means that wiring of signaling networks cannot be inferred simply by identifying the conserved proteins in the network. Second, in these systems, conserved proteins use different mechanisms to accomplish the same overall control strategy. This raises the question of how such systems evolve. The authors suggest that the control strategy itself may be an evolutionarily conserved property. These conclusions will be important to keep in mind as researchers examine these systems in more detail and begin to examine more complex systems as well. Modeling chemotaxis	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340964.xml
555854	Competition in health research: the experience of the John Curtin School of Medical Research	Background In 2002 the Australian National Competitive Grants System was opened to the Institute of Advanced Studies at the Australian National University as part of a commitment to transparency, competitiveness, and collaboration in national research funding. Results The block grant to the John Curtin School of Medical Research had progressively eroded over many years. Access to the National Competitive Grants Schemes and associated infrastructure (through an agreed 'buy-in' price of 20% of block funding) has succeeded in its aims and in reversing this progressive effective decrease in funding. Conclusion Access to the National Competitive Grant Scheme has allowed the John Curtin School of Medical Research to contribute more broadly to Australia's health and medical research effort through increased collaboration, in a transparent and competitive funding environment.	Discussion Australia has developed a unique blend of medical research institutions, from universities, teaching hospitals, independent medical research institutes to a range of smaller centres. Universities with medical schools have been the primary forces in medical research and many of the university teaching hospitals fostered the development of research institutes within their grounds [ 1 ]. The spectacular development of medical research in Australia over the last 50 years was reviewed recently in recognition of Australia's Centenary of Federation celebrations [ 2 ]. The Australian National University (ANU) was established as a research university [ 1 ] with four founding schools. The John Curtin School of Medical Research (JCSMR) was funded for its first 50 years through a one-line Commonwealth Grant to the ANU. Although the School had always sought external funds (its first grant was £1500 to Frank Fenner from the Rural Credits Development Fund in 1952), it was ineligible for National Health and Medical Research Council (NHMRC), Australian Research Council (ARC) and other national competitive grants. The School unquestionably benefited from generous (by local standards) funding in its early years and its science profited accordingly. Fundamental discoveries have included Nobel Prize winning work on the elucidation of mechanisms of transmission of signals in the nervous system (Eccles) and the discovery of the role of the major histocompatibility complex (Doherty and Zinkernagel). Today, the School has a wide range of research in such areas as infection, neurosciences, genomics and molecular bioscience, and has spawned a major national research facility, the Australian Phenomics Facility. Some key discoveries in recent years are summarised in Table 1 . Table 1 Recent Research Achievements, John Curtin School of Medical Research Year Achievement Researchers(s) Comment Reference 1999 Heparanase cloned Hulett, Freeman and colleagues Inhibiting the enzyme is the basis of cancer treatments based on sulfated polysaccharides Nature Medicine 5:803-809, 1999 2001 First diabetes susceptibility gene identified Slattery and colleagues Serendipitous discovery providing a potential therapeutic target in type 1 diabetes Proc Nat Acad Sci, USA 98:11533-11538, 2001 2002 Amiloride derivatives block ion channel activity and enhancement of virus-like particle budding caused by an HIV-1 protein Gage and colleagues Research on compounds that block viral ion channels raises the possibility of inhibiting viruses that utilize ion channels Eur. Biophys. J. 31:26-35, 2002 2002 Antibody 'tail sequence' identified that has significant implications for immunological memory Martin and Goodnow Finding has wide implications for vaccination, allergy and autoimmunity Nature Immunology 3:182-188, 2002 2002 Phase II clinical trials of PI 88 anti-cancer drug Parish and colleagues Promising results in treating advanced melanoma Eur. J Cancer 38(S7):74, 2002 2003 New approach to vaccination against cancer Parish Approach is potentially less susceptible to immune evasion Immunology and Cell Biology 81, 106-113, 2003. It can be argued that JCSMR's position as one of the original block-funded research schools of the Institute of Advances Studies (IAS) has given it the opportunity to pursue long-term, independent medical research. Many staff and alumni of JCSMR, including its present Director, argue that the Nobel Prizes, the Albert Einstein World Award, the Japan Prize, the Copley Medal, and a host of other international awards attest to the success of this funding strategy over the past 50 years, and that the scientific achievements that those prizes signify could only have been made by research that could be conducted over long (10 to 20 year) time frames. But the funding climate has changed dramatically over the last few decades. Twenty five years ago many tenured Australian university or teaching hospital staff were able to undertake research even without external funding. Infrastructure was relatively generous, there was still funding for university technical research staff, hospitals provided drugs and consumables and even beds at no cost to the researcher, and research activities merged relatively seamlessly into those of teaching and patient care. Today research is increasingly separated from teaching and care, and the days of in-kind support are long gone. Research is an investment for the nation and the future, but in contemporary funding climates within universities and hospitals it is increasingly becoming an optional extra, undertaken only when researchers are able to attract sufficient external resources. The change in environment and the disappearance of cross subsidisation from patient care and/or teaching has both advantages and disadvantages. Today's research (and research funding) is more transparent: what is spent on research as opposed to care and teaching, at least at the micro level, is much better defined. And, in parallel, the loss of capacity to undertake research that is not peer reviewed has almost certainly raised research standards. The negatives are equally obvious. The University of Melbourne submission to the Wills review stated 'The danger is that if the importance of the nexus between research and learning is not visible to tertiary students because University research is allowed to run down through poor infrastructure, equipment or lack of opportunity of research scientists then research institutes and teaching hospitals will inevitably suffer from a lack of quality of graduates available to them'[ 3 ]. The JCSMR has not been exempted from these pressures, and over the last 20 years the value of the block grant has been progressively eroded. The JCSMR grant was a one-line grant to the Director, which provided flexibility, but inhibited collaboration because of restrictive rules around the National Competitive Grants Scheme. Accordingly, a desire for transparency, competitiveness and collaboration led to a decision in 2000 by the Commonwealth Government that the IAS of the ANU could enter the National Competitive Grants Scheme. The negotiations that led to partial entry of JCSMR into the NHMRC schemes in 2001 (for 2002) were long and complex. However the negotiations between JCSMR and NHMRC were paralleled by those between the ARC and ANU and led to a jointly agreed 'buy in' price of 20% of 2000 funding for the IAS to gain access to the various national funding, research training and related schemes. This was $1.7 million per annum for JCSMR access to the National Competitive Grant Schemes and a similar amount for access to the training and infrastructure schemes. It was determined initially that entry would be phased, but after one year the ARC determined that the phase-in was unnecessary and gave the IAS full entry from 2002 (for 2003), and the NHMRC followed suit. The JCSMR felt that, at last, it was able to redress the competitive disadvantage the School faced because of its essentially fixed funding over the last 10 to 15 years, at a time when government funding for the NHMRC system had doubled and then redoubled, from around $65 million in the late eighties to $176 million in 1999 to $381 million in 2004. The rest of the system waited, not without apprehension, for the outcome, but there had long been a wish within the research community that JCSMR be subject to the same forms of peer review as the broader medical research community. Removing barriers to collaboration, the outcome of these changes, is important for 21 st century approaches to improving health. The World Health Organisation has stated that the likely trends in global health in the 21 st century include widespread absolute and relative poverty, demographic changes, ageing, growth of cities, epidemiological changes, continuing high influence of infectious diseases, increasing incidence of non-communicable diseases, injuries and violence, global environmental threats to human survival, new technologies, information and telemedicine services, advances in biotechnology, evolving partnerships for health that include private and public sectors and civil society, and globalisation of trade, travel and the spread of values and ideas. Research to deal with global health problems will therefore necessarily be multidisciplinary, involving biomedical, clinical, public health and health services research, and include the social sciences, information sciences and engineering, physics, chemistry, ecology and environmental sciences and economics. As part of the IAS, the JCSMR is accordingly strongly positioned. Not that this need for a multidisciplinary approach is really a new concept – in 1902 Osler stated that the remit of medical research was 'to wrest from nature the secrets which have perplexed philosophers in all ages, to trace to their sources the causes of disease, to correlate the vast stores of knowledge, that they may be quickly available for the prevention and the cure of disease – these are our ambitions' [ 4 ]. In the first three years that JCSMR has been eligible to apply for research funding, their researchers have been awarded a total of $18 million (88% NHMRC, 12% ARC) in competitive funding for periods of up to five years. All three NHMRC program grants held by researchers at the JCSMR are collaborative with other Australian institutions and School researchers hold four ARC linkage grants. This collaboration bodes well for Australian health and for biotechnology growth. In a highly competitive world, Australian researchers need every opportunity to succeed.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555854.xml
340958	Activating p53 in Cancer Cells with Protein Therapy Shows Preclinical Promise	null	Late-stage cancers are notoriously unresponsive to treatment, making certain hard-to-detect cancers particularly insidious. Ovarian cancer, for example, most often escapes diagnosis until the tumor has already metastasized. At this stage, ovarian cancer is classified as peritoneal carcinomatosis, a terminal condition characterized by widespread tumor growth throughout the peritoneum, the large serous membrane that lines the abdominal cavity, pelvis, and associated organs. Advanced cases of peritoneal carcinomatosis are largely resistant to chemotherapy and account for the bleak 15%–20% survival rate of ovarian cancer. Biologists often view cancer as an evolutionary process in which cells that would normally cooperate with their neighbors begin to compete with them. Selective advantage for cancer cells often begins with mutations that inhibit tumor suppressor pathways. p53, like other tumor suppressor genes, arrests cell growth and induces apoptosis (programmed cell death) in response to cellular stress, such as chromosomal damage. Cells with p53 mutations often escape these constraints, leading to the uncontrolled growth characteristic of “immortal” cancer cells. Nearly all types of tumors have mutations in the p53 pathway. Treatments focused on restoring p53 function—which is likely to be defective only in cancer cells—should prove more effective than chemotherapies, which indiscriminately kill all dividing cells, healthy or cancerous. With the goal of developing targeted therapeutic strategies, Steven Dowdy and colleagues show that restoring p53 protein function in tumor cells not only dramatically increases lifespan in mice but also eliminates disease. While past efforts to restore tumor suppressor function in cancer cells have focused on gene therapy, Dowdy and colleagues introduced modified p53 peptides, or protein fragments, into cancer cells. p53 works as a “transcriptional” activator that binds to specific sequences of DNA and triggers apoptosis in response to DNA damage. Its biological function flows from this binding ability. One region of this large protein, called the C-terminal domain, facilitates effective binding. In cancer cells, synthesized peptides (called p53C′) derived from this region can induce apoptosis by activating p53—which is normally present in low levels in a biologically inactive form—and restoring function to p53 proteins with DNA-binding mutations. To get p53C′ peptides into cancer cells, Eric Snyder et al. used a technique pioneered by Dowdy that delivers large proteins into the cell interior. Since the cell membrane normally limits passage to only small molecules (larger molecules generally enter through surface receptors), this is no small feat. The technique exploits the ability of a small peptide region from the HIV TAT protein to smuggle macromolecules through cell membranes that normally prohibit entry to such large molecules. After synthesizing a structurally modified form of p53C′ less prone to degradation, the researchers first confirmed that the peptide was functional and then that it activates p53-specific genes in tumor cells, but not in normal cells. Testing the effectiveness of the peptide therapy on mouse strains used to model human metastatic disease, they found that mice treated with the TATp53C′ peptide showed a significant reduction in tumor growth and lived six times longer than both mice treated with a control peptide and untreated mice, with some mice remaining disease-free more than 200 days after treatment. This macromolecular delivery approach, Snyder et al. argue, works with greater specificity and avoids the tumor-generated neutralizing effects observed in small molecule strategies. Because a mutation in the p53 gene is one of the most common events in the development of cancer, these results could have implications for a wide variety of cancers. And by working with mouse models that approximate the physiological burdens metastatic cancer imposes on humans, Dowdy's team demonstrates the promise of developing targeted “intracellular biologic” therapeutics that treat the systemic pathology of cancer—inhibiting tumor growth as well as alleviating the lethal complications associated with the disease. TATp53C′ treatment extends survival of mice	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340958.xml
516038	Work-related musculoskeletal disorders : A survey of physical therapists in Izmir-Turkey	Background This study was planned to collect data about causes, prevalence and responses to work-related musculoskeletal disorders reported by physiotherapists employed in Izmir, Turkey. Method A two-page survey with closed ended questions was used as the data collected method. This survey was distributed to 205 physiotherapists working in Izmir, Turkey, and 120 physiotherapists answered. Questions included occupational history of physiotherapists and musculoskeletal symptoms, special areas, tasks, job-related risk factors, injury prevention strategies, and responses to injury. Results Eighty-five percent of the physiotherapists have had a musculoskeletal injury once or more in their lifetime. Injuries have been occurred mostly in low back (26 %), hand-wrist (18 %), shoulders (14 %) and neck (12 %). The highest risk factor in causing the injury was transferring the patient at 15%. Sixty-nine percent of physiotherapists visited a physician for their injury and sixty-seven percent of the respondents indicated that they had not limited their patient contact time as a result to their injury Conclusions According to the results of this study, the rate of musculoskeletal disorders in physiotherapists in Izmir-Turkey has been found to be high due to their profession. Respondents felt that a change in work habits was required in order to decrease the risk of another injury.	Background A work-related musculoskeletal disorder (WRMD) is defined as a musculoskeletal injury that results from a work-related event. This may result in lost work time, work restriction, or transfer to another job [ 1 - 4 ]. These types of injuries are common among physiotherapists [ 1 , 2 , 5 - 9 ]. This group has a moderately high prevalence of occupational low-back pain [ 2 , 9 - 12 ]. Physical therapists routinely perform manual therapy, such as soft-tissue mobilization, which means that the upper limb is also exposed to risk factors associated with musculoskeletal and neurovascular disorders [ 1 , 2 , 6 - 8 , 13 , 14 ]. In addition, these professionals routinely perform activities that involve transferring a patient (from exercise mat to chair, to parallel bar etc), assisting with activities on the exercise mat, and lifting and using cumbersome equipment [ 1 , 2 , 6 - 15 ]. These work tasks put therapists at risk for both acute and cumulative musculoskeletal pain. Understanding the issues related to musculoskeletal disorders in physical therapists requires some awareness of the context in which these professionals work. In Turkey, physiotherapists cannot treat patients directly. According to law, direct accesses to physiotherapists are not possible [ 16 ]. Further, physiotherapists are not allowed to plan patients' physiotherapy programs; this is done by a physiatrist. Currently, efforts are underway to enact legal amendments that will permit a doctor to direct a patient to a physiotherapist, with the physiotherapist creating the physiotherapy program him- or herself. The purpose of this study was to investigate the prevalence and features of WRMDs in a group of Turkish physiotherapists working in various capacities in a large Turkish city, and to compare the findings to reports from other countries. Methods The study was approved by the Ethics Committee of Dokuz Eylül University. The research population was selected physiotherapists who were employed in a broad spectrum of practice settings in İzmir, Turkey. All data were collected by a mailed questionnaire (Appendix 1 [see Additional file 1 ]) that consisted of 17 close-ended questions. Questionnaires were distributed to all physiotherapists (205 total, 157 females and 48 males) who were registered members of the Izmir Branch of the Turkish Physiotherapy Association. Each member was asked to complete the self-administered questionnaire if they had more than 2 years of experience in practice. One hundred and twenty questionnaires were returned and analyzed (total response rate 59%, for women 59%, for men 58%). The questionnaire was based on another published survey [ 1 , 6 ] and simply adapted and translated for the Turkish context. The questionnaire was composed of two parts, personal and occupational. The personal portion asked about general characteristics, including sex, age, weight, and height. The occupational portion inquired about years of experience, work setting, and number of hours of contact with patients per week. This section also asked whether the subject had experienced any WRMDs. If the answer was yes, the person was asked to state the type of injury, the body part affected, specific activities caused on occupational injury, the work setting in which the injury occurred, whether the injury was reported or a physician was consulted, and what sort of treatment was applied. They were also asked whether they lost work time as a result of the injury, what activities caused symptoms to recur, and whether the injury had caused the respondent to alter his or her work habits, reduce hours with patients, or change employment settings. Data were analyzed using SPSS 10.0 for Windows. Results for the general information items were expressed as mean ± standard deviation, and results for items in the occupational portion were expressed as percentages. χ 2 were used to analyze influence personal characteristics (sex, age, number of years in physiotherapy practice, number of hours per week in direct patient care) to WRMDs. Results The 120 respondents included 92 females and 28 males of mean age 30.4 ± 6.9 years (range, 22–55 years). Total response rate was 59%, for women 59%, for men 58%. General information about the group is given in Table 1 . The questionnaire answers indicated that the respondents spent an average of 38.8 ± 9.2 hours per week in direct patient care, and had 8.0 ± 6.0 years of work experience on average. Eighty-six of the physiotherapists worked in general hospitals, 24 worked at hydrotherapy centers, and 10 worked at pediatric rehabilitation centers. Frequency of WRMDs was not correlated with sex (χ 2 =.234, P=.629), age (χ 2 =.383, P=.536), number of years in physiotherapy practice (χ 2 =.067, P=.795), or number of hours per week in direct patient care (χ 2 =.151, P=.698) (Table 2 ). One hundred and two respondents (85%) reported that they had experienced work-related musculoskeletal pain or discomfort at some time in their occupational lives. Seventy-eight (65%) reported that they had sustained more than one WRMD. The lower back was the body part with the highest frequency of occupational injury (26%), and wrist-hand (18%), shoulder (14%), and neck (12%) were other sites frequently affected (Table 3 ). The main types of injury reported were tendinitis (21%), vertebral disk problems (16%), muscle strain (16%), ligament sprain (16%), degeneration (15%), synovitis (6%), tear (2%), dislocation (1%), fracture (1%) and other (6%). The factors that most frequently led to WRMD were transferring patients (15%), performing repetitive tasks (14%), and lifting (14%) (Table 4 ). The respondents who had experienced a WRMD indicated that lifting (18%), maintaining a position for prolonged period of time (17%), performing repetitive tasks (16%), and transferring patients (16%) were the activities that most often exacerbated their symptoms during clinical practice (Table 5 ). Improvements in body mechanics (21%), avoiding lifting (16%), and changing working positions frequently (14%) were the top three changes that the injured respondents said they made in their working habits (Table 6 ). Sixty-nine percent of the respondents with WRMDs had visited a physician for the problem, and 46% of whom stated that they had officially reported the injury to their employer. The respondents who had suffered WRMDs reported they used their own occupational knowledge, rest, medications, and exercise to treat the problem (Table 7 ). Sixty-seven percent of the respondents who had had WRMDs indicated that they had not permanently reduced their patient-contact time as a result of their injury, and 82% said they did not limit their areas of practice after sustaining the injury. Most of the physiotherapists (63%) who had had WRMDs indicated that they would not consider a job change because of their injury or due to the risk of sustaining another injury. Discussion Musculoskeletal system problems connected to occupational conditions are common among health care workers. The costs of these are substantial, both in terms of money and in terms of work time lost [ 1 , 7 , 9 , 17 , 18 ]. Research has shown that musculoskeletal problems are particularly common in health care workers who are in direct contact with patients [ 1 , 7 , 8 , 17 , 18 ]. Physiotherapists have a high prevalence of WRMDs [ 1 , 7 , 8 , 15 ]. The results from studies on WRMDs in physiotherapists have generally been similar, though some have differed according to country. Such variations are linked to level of development, the status of the profession of physiotherapy in a given country, psychosocial, and epidemiological factors [ 2 , 5 , 8 , 11 , 13 , 14 ]. In Turkey the law bars physiotherapists from providing primary care; each patient must be referred. Considering the differences in physiotherapy practice among countries and regions, we felt it would be valuable to investigate the prevalence and features of WRMDs in a group of Turkish physiotherapists working in various capacities, and to compare to findings in other countries. In this study, we collected demographic and WRMD data from 120 physiotherapists in Izmir, Turkey, and analyzed rates of injury, risk factors, injury types and sites, and post-injury management. We asked to complete the self-administered questionnaire if they had more than 2 years of experience in practice. Thus, response rate of questionnaire was low (58.5 %) in our study. The survey answers revealed that 85% of the respondents had experienced WRMDs. Cromie [ 8 ] reported that younger physiotherapists have a higher prevalence of musculoskeletal problems related to occupational conditions. Rugelj [ 9 ] investigated low-back pain in physiotherapists, and found an incidence of 66% in subjects between the ages of 20 and 40 years [ 1 ]. In line with this, a study of Australian physiotherapy students by Nyland [ 2 ] revealed that the 20- to 21-year age group had the highest frequency of low-back pain. The average age of the physiotherapists in our study was 30.4 years. This mean age corresponds with other findings in the literature, and confirms that physiotherapists tend to experience WRMDs at young age. Such injuries in younger physiotherapists may be associated with lack of professional experience, and the lower knowledge and skill levels people tend to have in the early years of this career. Concerning sites of musculoskeletal injury during professional activities, the highest incidence is in the low-back region. Biomechanical studies have shown that physical loading factors, such as body flexion, rotation and weight loading, play a role in this [ 18 ]. In a study that covered 25% of all physiotherapists working in Australia (824 total studied), Cromie [ 8 ] found that the rate of work-related low-back pain was 48%. Other authors have revealed various rates of this problem in physiotherapists: Bork [ 7 ] 45%, Holder [ 1 ] 62%, Molumphy [ 10 ] 29%, Scholey and Hair [ 12 ] 38%, Mierzejewski [ 11 ] 49.2%, Rugelj [ 9 ] 73.7%, and Nyland [ 2 ] 69%. Our survey of Turkish physiotherapists revealed a 25.5% incidence of low-back pain. Interestingly, this figure is lower than most other rates reported in the literature. In our study, when we categorized physiotherapists according to practice setting, low-back pain was the most common WRMD in all subgroups. Studies of WRMDs in health care of professionals have identified the lower back as the most commonly involved area of the body, followed by neck and upper extremities [ 3 , 13 , 19 - 21 ]. Investigations of physiotherapists have revealed similar results. Bork [ 7 ] and Holder and et al. [ 1 ] listed the regions most commonly involved musculoskeletal injuries as the lower back, hand-wrist, and neck, respectively among physiotherapists. On analyzing the different body parts injured in our subjects, we found the highest frequency of injuries in the lower back, followed by the hand-wrist (18.2%), shoulder (14.4%) and neck (11.8%). According to the literature the work-related activities that most commonly lead to injury in health professionals are lifting heavy equipment and patients, transferring patient, maintaining the same posture for a long period, manual therapy practices, responding to patients' sudden movements, and repeated movements [ 1 , 8 , 17 , 18 , 22 - 24 ]. Bork [ 7 ] identified the main causes of WRMDs in physiotherapists as staying in the same position for along time and continuing to work when tired; Molumphy [ 10 ] emphasized lifting and leaning downwards with sudden maximal effort; and West [ 13 ] highlighted maintaining the same posture for along time, manual therapy, repeated movements, and increased work load. In our survey, the main causes of WRMDs were patient transfer, repeated movements, lifting heavy equipment, patients and working when too physically tired. Holder et al. [ 1 ] identified three activities that aggravate the symptoms of former WRMDs in physiotherapists and assistant physiotherapists as staying in the same position for a long period, lifting and patient transfer. In our study, the respondents to our questionnaire noted lifting, staying in the same position for a long time, patient transfers, and repeated movements. In Turkey, most physiotherapists work in general hospitals. In these facilities, the daily treated patient number is too high, and this number far exceeds the number of physiotherapists on staff. In addition, the majority of patients in these hospitals are seriously ill [ 25 ]. People who suffer injuries on the job may be treated with medication, rest and exercise. Physiotherapists have fundamental knowledge about ergonomics and biomechanics, and using this knowledge may vary depending on professional knowledge and skills. In Turkey, physiotherapists are trained in ergonomic working principles at the undergraduate level. However, in many working environments the equipment used (treatment table, chair, etc .) is not ergonomic, and consequently the physiotherapist cannot follow ergonomic principles while doing his or her job. A study conducted by Cromie [ 6 ] in 2002 examined whether physiotherapists use their own knowledge to prevent WRMDs. The author found that this was true for most of the physiotherapists investigated. In our survey, of the physiotherapists who had suffered WRMDs, 27.6% said they used their professional knowledge and 26% said they used rest to manage the injury. Previous research has shown that physiotherapists who have suffered a WRMD tend to change their professional attitudes to avoid additional injuries. Holder et al. [ 1 ] found that 79% of physiotherapists and 81% of assistant physiotherapists who had suffered injuries on the job changed their professional attitudes to avoid other WRMDs. It has been reported that the most common strategies used by physiotherapists to avoid WRMDs are correction of body mechanics, and frequent postural changes. The respondents to our questionnaire said that, after suffering a WRMD, they paid more attention to correcting body mechanics, avoided lifting heavy equipment or patients, changed position frequently, and got other personnel to help them with tasks that involved lifting. Mierzejewski [ 11 ] reported that 13.7% of physiotherapists in Edmonton, gave up their career after suffering work-related low-back pain, and that 35.3% continued to work after this problem. Cromie [ 8 ] found that 17.7% of physiotherapists changed their field of practice in connection with WRMDs. Of the respondents in our Turkish survey who had suffered WRMDs, 67% did not change their field of practice after the injury, 82% did not restrict their time with patients, and 64% said they would not consider changing their field or department of practice due to their WRMD. This means that 33% did change field of practice after a WRMD, which is 1.9 times the rate of those who changed field in Cromie's study. As noted above, most physiotherapists in Turkey work in general hospitals, and it is not usually possible for such physiotherapists to change their field of work. Some professions have very high WRMD prevalence, and this has led to more intensive research in recent years. These types of injury have major consequences for society, workers, employers, and the insurance sector due to loss of labor force, long-term disability and delay in returning to work, decreased productivity, and psychological effects on employees. Therefore, minimizing and preventing WRMDs has significant potential social and economic benefits. It is important to invest in studies aimed at reducing these types of injuries in physiotherapists. Our study has three main limitations. A physiotherapist's working time in a day is not long, but he or she treats a very large number of patients during this time. Our survey would have been more informative if we collected data for the number of daily treated patients. We also did not inquire about the physiotherapists' activity levels. This would have been valuable information, as sports and recreational activity may affect WRMD frequency. Another limitation of our study was comparing various areas of practice and techniques used by physiotherapists and investigating their relationship to WRMDs. Conclusions Our survey reveals that the WRMDs in physiotherapists in Turkey are similar to rates reported in other countries. Physiotherapists in our country suffer similar work-related injuries as their counterparts elsewhere, despite different legal working conditions and cultural differences. This study provides data related to WRMDs in physiotherapists in Turkey. Further studies can be very useful if it research prevalence of WRMDs in physiotherapists who have employed different working conditions. Competing interests None declared. Authors' Contributions YS participated in the design, managed the data collection and performed statistical analysis. AO participated in designed the study protocol, managed the coordination. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Table 1, Table 2. Questionnaire on occupational injuries in physical therapists. Turkish version of questionnaire on occupational injuries in physical therapists Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516038.xml
544965	Managing emerging infectious diseases: Is a federal system an impediment to effective laws?	In the 1980's and 1990's HIV/AIDS was the emerging infectious disease. In 2003–2004 we saw the emergence of SARS, Avian influenza and Anthrax in a man made form used for bioterrorism. Emergency powers legislation in Australia is a patchwork of Commonwealth quarantine laws and State and Territory based emergency powers in public health legislation. It is time for a review of such legislation and time for consideration of the efficacy of such legislation from a country wide perspective in an age when we have to consider the possibility of mass outbreaks of communicable diseases which ignore jurisdictional boundaries.	The management of infectious diseases in an increasingly complex world of mass international travel, globalization and terrorism heightens challenges for Federal, State and Territory Governments in ensuring that Australia's laws are sufficiently flexible to address the types of problems that may emerge. In the 1980's and 1990's HIV/AIDS was the latest "emerging infectious disease". Considerable thought was put into the legislative response by a number of Australian jurisdictions. Particular attention had to be given to the unique features of the disease such as the method of transmission, the kinds of people who were at risk, and the protections needed by the community and the infected population to best manage the care of those infected and to minimize new infections. Health workers and researchers began to find that "the most effective strategies that we have so far found to help promote reduction of the spread of HIV involve the adoption of laws and policies which protect the rights of people most at risk of infection" [ 1 ]. A good example of a legislative response which adopts this approach is found in section 119 and 120 of the Victorian Health Act 1958. These sections emphasize the need to protect the privacy of the infected individual and to undertake a staged response which is proportional to the risk presented by the infected individual. The legislation has been very effective with HIV and has been praised for its progressive approach [ 2 ]. In 2003 the community has been faced with the emergence of two new infectious diseases, SARS and Anthrax. Whilst there were no cases of either disease in Australia, the threat of a possible outbreak had to be acknowledged and a response planned. Anthrax is not a new infectious disease. Humans can become infected with anthrax by handling products from infected animals or by breathing in anthrax spores from infected animal products (like wool, for example). People also can become infected with gastrointestinal anthrax by eating undercooked meat from infected animals. However, its manufacture and use as a weapon for bioterrorism forces us to rethink its management in a new context. These two infectious diseases have very different features from HIV which spreads only via transmission of infected bodily fluids such as blood or semen. SARS, by contrast is transmitted via droplets from infected cases which, as a result of coughing, carry the virus to close contacts [ 3 ] Thus, the infection profile of SARS requires planning for the possible overrun of Intensive Care Units and the likely infection of a number of ICU staff affecting both morale and capacity to cope. Anthrax raised different problems. These include the possible investigation of terrorist suspects alongside investigation of the outbreak of the infectious disease. Difficulties are also raised by likelihood of public panic, and the flooding of public health officials with reports of suspicious white powder. In early 2004 the media reported the spread of avian influenza across South East Asia. This disease has different features from HIV/AIDS and SARS and an approach to an Australian outbreak would also be different. The main difference is in the source of transmission of the virus, that is, from infected birds to humans. There is very little difference [from ordinary influenza] in the symptoms (though these may vary in severity) or treatment of the virus [ 4 ] It is too early to predict whether this may be the next "emerging infectious disease", but its current spread has given rise to concern about such a possibility [ 5 ] Australia is a federal system. There are two parallel sets of laws in operation. The Commonwealth Constitution sets out the legislative powers of the Commonwealth. Specific powers are listed in the Commonwealth constitution but State constitutions have broad powers covering matters such as peace, order and good governance. As the Commonwealth has no specific power to legislate with respect to health, other than the quarantine power, national legislative schemes in public health which rely upon a cooperative approach from all States and Territories are cumbersome and difficult. Without a specific head of power, the Commonwealth has limited ability to legislate with respect to health. "That is, the legislative powers of the Commonwealth are specified in the Constitution and do not include expressly most of the activities that together comprise the field of public health"[ 6 ] For this reason, there are no Commonwealth emergency health powers except quarantine powers. Quarantine powers are currently restricted to isolation at the border of the country of people, plants, and animals to prevent the spread of disease. There is a real possibility that quarantine laws could have a broader scope. It depends on how widely the High Court would interpret section 51(x) of the Commonwealth Constitution. A quarantine law could override state laws as long as it remained a law "with respect to quarantine". However, "the power is potentially a colossus so far as the expansion of legislative authority in the fields of public health is concerned". [ 6 ] The quarantine power would be the most likely candidate for a head of power on which to base development of commonwealth laws for the management of public health emergencies. Another possibility may be the external affairs power, if there was a relevant treaty or international agreement which could be given effect to in domestic law. However the legislation would have to be limited to laws giving effect to the treaty. States and territories have a range of emergency powers available to them in their existing public health legislation. Some are relatively old. For example, the Health Act 1911 (WA), Public Health Act 1952 (NT) based on an 1898 Ordinance (Both these Acts are currently under review). Health emergency powers vary from one jurisdiction to another, but include powers to support disease surveillance, contact tracing and orders to restrict behavior or movement of individuals with an infectious disease in certain circumstances. There are also powers to recall food, search premises and seize property, close buildings and a range of other substantial and intrusive powers. It is suggested that it is time to consider whether state and territory public health legislation contains sufficient measures to manage the outbreak of an infectious disease in a modern environment which includes mass travel, swift spread of infection and additional complexity raised by fears of bioterrorism. Currently, in a public health emergency caused by the spread of an emerging infectious disease, Australia could need to rely on a patchwork of legislative measures to assist it to cope. Commonwealth quarantine laws and State and Territory powers in public health legislation may all be needed to address the problem. If an outbreak occurred on a border, or in some area where jurisdiction may be in doubt such as airspace or offshore and a state or territory response was required in addition to any quarantine measures, there could be confusion over jurisdiction for the application of State and Territory powers. State and Territory public health acts do not adequately provide for interjurisdictional communication and cooperation. There could also be difficulties if an infectious disease caused overseas deaths of people from more than one State or Territory in circumstances where an Australian coronial investigation was considered desirable. In such a situation, the jurisdiction of more than one Australian coroner would be triggered. Several State and Territory coronial laws could apply and there could be different inquests under different laws undertaken by different coroners into the same incident. It is suggested that it is time to look at the efficiency of the emergency powers laws of Australia as a whole: to map the laws in each jurisdiction and the Commonwealth quarantine laws and to consider their effectiveness in the face of the outbreak of a fast moving, easily spread infectious disease. The efficacy of Australia's laws should also be considered in relation to bioterrorism. While there were no infections from anthrax in 2003 despite a great deal of media coverage and infections and deaths in the US, a responsible legislature ought to acknowledge the possibility and ensure that the law is ready to support a swift and effective response. It is not enough to consider whether the individual pieces of legislation are up to the task of managing outbreaks of newly emerging infectious diseases. Indeed many of the jurisdictions are currently reviewing their public health legislation and will no doubt give proper consideration to this issue as part of the review. But who is thinking about how the legislation of all jurisdictions and the Commonwealth quarantine fits together? What powers enable communication and cooperation between jurisdictions about the outbreak of infectious disease? What kind of opportunity is there for a coordinated response? Can public health orders made in one jurisdiction travel to another jurisdiction when the infected individual travels? What arrangements can be made if an outbreak occurs on or close to a interstate border? What if there is an outbreak on a bus carrying passengers from Victoria, through South Australia to the Northern Territory? It is encouraging to note that, even without specific legislation, there has been a mechanism to achieve communication and cooperation between jurisdictions through the Communicable Disease Network of Australia (CDNA). This Network has in fact been quite successful in fostering regular communication between the Communicable Disease Units across the country and has been involved in coordinated actions during a number of multistate outbreaks. Despite the existence of this network and other good working relationships between government officials and various agencies in different jurisdictions, a serious outbreak of communicable disease would require the existence of legislative powers. Public health emergencies generate confusion, even panic. Clarity of powers and the way those powers interact with each other would be crucial in an emergency. It became apparent after the Bali tragedy in 2002 that coroner's jurisdiction was triggered differently in different jurisdictions and some acts did not support communication and cooperation when inquests might be needed for deaths of people ordinarily resident in several jurisdictions. The time to find the shortcomings in the legislation is well before the crisis. A review of the efficacy of how these laws work together to protect the public health of all Australians should be undertaken. It has been possible to overcome the hangovers of federation for the betterment of all Australians in relation to corporations law. When doubts were recently raised about the constitutional basis of the corporations law scheme, the States and Territories were able to cooperate and refer the necessary powers to the Commonwealth to provide certainty about the laws which govern our corporations. Is our public health any less important than governance of our corporations? Could we cooperate to give ourselves certainty, flexibility and a consistent approach which protects the rights of those subject to some very broad powers? The States and Territories are generally reluctant to refer powers to the Commonwealth. It may be time to seriously discuss referral of powers in the context of health emergency powers. At the very least, it is time that the Commonwealth, States and Territories recognised the need for the laws to work as a set of laws to protect the whole country, not simply individual laws to protect individual jurisdictions. There has been work done internationally in this area. A model State Emergency Health Powers Act has been developed in the US in 2001 [ 7 ] In the preamble to this Act a rationale for its development is set out: "In the wake of the tragic events of September 11, 2001, our nation realizes that the Government's foremost responsibility is to protect the health, safety and wellbeing of its citizens. New and emerging dangers including emergent and resurgent infectious diseases and incidents of civilian mass casualties – pose serious and immediate threats to the population. A renewed focus on the prevention, detection, management and containment of public health emergencies is thus called for." The US, like Australia, is a Federal system. The model was intended to be taken up by those US states which wished to do so. To date, it has been passed in over half the US states. This bill would be an excellent starting point for development of an Australian model. There are a number of legislative mechanisms which could be used to support a nationally uniform approach to health emergency powers legislation in Australia. The development and adoption of the model food legislation provides a useful model of a cooperative uniform approach. A model act was developed in consultation with all jurisdictions. It covered areas agreed to be core areas of the Act which ought to be the subject of a national approach and other provisions which were considered to be administrative and were to be adopted at the discretion of each jurisdiction. An intergovernmental agreement was signed as a mechanism to protect the uniformity of the legislation. The agreement sets up a Ministerial Council, supported by a Food Regulation Standing Committee. The Council has responsibility for deciding on proposals to amend the model [ 8 ] If a decision is made in favor of amendment, States and Territories will use their best endeavors to submit to their respective Parliaments, legislation which gives effect to the amendment. The law is an important tool in supporting the management of the outbreak of infectious diseases. The existence of our Federal system has meant that we have a different approach in each State and Territory together with Commonwealth control of quarantine. Newly emerging infectious diseases creating real threats to public health in an era of easy mass travel, and the present threat of bioterrorism mean that it is time Australia examined all laws to contain and manage infectious disease outbreak. The laws should be examined both for their effectiveness in the areas they cover, and as part of a whole which ought enable a response which protects the health of all Australians, and crosses borders as easily as SARS or avian influenza.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544965.xml
544959	+874(T→A) single nucleotide gene polymorphism does not represent a risk factor for Alzheimer's disease	In the recent years, several cytokines have been associated with Alzheimer's disease (AD) development and progression and many studies have correlated this risk with polymorphisms in the genes encoding these molecules. Also the type 1 cytokine interferon (IFN)-γ belongs to a cytokine class that affects the immune function; in fact it plays a major role in defence against viruses and intracellular pathogens but also in the induction of the immune-mediated inflammatory response. The aim of this study was to evaluate the role of IFN-γ in AD by studying the association of +874T→A IFN-γ gene polymorphism with AD. We included in this study 115 AD patients (70 women, 45 men, mean age 80) and 90 sex and age-matched healthy controls (HC, 51 women, 39 men, mean age 82) from northern Italy. Genomic DNA was extracted with the salting-out method from whole blood of all subjects; the genotyping at IFN-γ loci was assessed with ARMS-PCR. The data obtained from the +874T→A IFN-γ gene polymorphism analysis of AD patients and HC lack of any statistically significant differences also when stratified according to gender. In conclusion these results confirm the previous shown lack of association between +874T→A IFN-γ gene polymorphism and the risk of AD. However, other polymorphisms have been demonstrated to influence IFN-γ transcription and since natural killer cells of AD patients show higher production of the cytokine, further analysis will be necessary to clarify the role of this gene in the pathogenesis of the disease.	In the human brain several cell types are responsible for initiating and amplifying a specific inflammatory response. In Alzheimer disease (AD) signs of an inflammatory activation of microglia and astroglia are present both inside and outside amyloid deposits. Cell cultures and animal models suggest an interactive relationship between inflammatory response activation, reduced neuronal functioning and amyloid deposition. Furthermore cells associated with extracellular plaques within AD brains can produce a variety of cytokines, chemokines and other related proteins that influence plaque and tangle formation [ 1 ]. For these reasons cytokines could play a critical role in the pathogenesis of AD. In the recent years, several cytokines have been associated with AD development and progression and many studies have correlated this risk with polymorphisms in the genes encoding these molecules [ 2 - 4 ]. Inside this research area we described that single nucleotide polymorphisms (SNP)s of the intereleukin(IL)-10 and IL-6 genes were associated with highest risk of AD with apparent interaction between these two genes [ 5 ]; these results supported the theory that the overall risk of developing AD may be governed by a 'susceptibility profile' and reflected the combined influence of inheriting multiple high-risk alleles. Also the type 1 cytokine interferon (IFN)-γ belongs to a cytokine class that affects the immune function; in fact it plays a major role in defence against viruses and intracellular pathogens but also in the induction of the immune-mediated inflammatory response [ 6 ]. It has been reported that the polymorphism +874(T→A) of the gene encoding IFN-γ is associated with a different production of this molecule, in particular the T allele correlates with increased levels of the cytokine [ 7 ]. In regards to the hypothesis that IFN-γ SNP may represent a genetic risk factor for AD, a recent study did not support this possibility [ 8 ]. Therefore with reference to this paper we try to confirm their hypothesis in a sample of patients, that were already genotyped for apolipoprotein E (ApoE), IL-10 and IL-6 [ 5 ]. We included in this study 115 AD patients (70 women, 45 men, mean age 80 ± 2) and 90 sex and age-matched healthy controls (HC, 51 women, 39 men, mean age 82 ± 2) from northern Italy. The clinical diagnosis of AD fulfilled the international criteria of the DMS IV and NINCDS-ADRDA; every patient had a recent brain magnetic resonance imaging (MRI)/computed tomography (CT) scan available. Cognitive performances were assessed according to the Mini-Mental State Evaluation (MMSE). Only AD and HC without clinical signs of inflammation (e.g. normal body temperature, no concomitant inflammatory condition) were eligible in order to minimize the risk of clinical or sub-clinical inflammatory processes. Blood chemistry tests were done and subjects with an abnormal red blood cell sedimentation rate or altered albumin and transferrin plasma levels were excluded. AD patients were further selected according to their C reactive protein (CRP) plasma levels and anyone with CRP above 5 mg/L (mean + 2 standard deviations of control values) were not eligible. Informed consent was obtained from all the subjects or their relatives. Genomic DNA was extracted with the salting-out method from whole blood of all subjects; the genotyping at IFN-γ loci was assessed with the same amplification method (ARMS-PCR) described by Scola et al. [ 8 ]. The genotype frequencies in the study groups were compared by the chi-square (χ) test in order to calculate significant different SNP distribution between AD patients and controls (Table 1 ). The data obtained from the +874T→A IFN-γ gene polymorphism analysis of AD patients and HC lack of any statistically significant differences also when stratified according to gender (data not shown). The percentage of the different genotypes was similar in AD compared with HC (T/T: 17.4% vs. 16.8% ; T/A: 50.4% vs. 55.1% ; A/A 32.2% vs. 28.1%) and consequently also the allele distribution shows no differences (T: 42.6% vs. 44.4% ; A: 57.4% vs. 55.6%). Our genotype distribution looks very similar to that found by Scola et al. [ 8 ]. In conclusion these results confirm the lake of association between +874T→A IFN-γ gene polymorphism and the risk of AD. Nevertheless other polymorphisms have been demonstrated to influence IFN-γ transcription and since natural killer cells of AD patients show higher production of the cytokine, further analysis will be necessary to clarify the role of this gene in the pathogenesis of the disease [ 9 , 10 ]. Table 1 IFN-γ genotype and allele distribution Genotype Allele T/T (H) T/A (I) A/A (L) T A AD 20(17,4%) 58(50,4%) 37(32,2%) 98(42,6%) 132(57,4%) HC 15(16,8%) 49(55,1%) 25(28,1%) 79(44,4%) 99(55,6%) Genotype χ 2 = 0.488, df = 2, p = 0.783 Allele χ 2 = 0.232, df = 1, p = 0.63 In brackets there are the corresponding phenotype high (H), intermediate (I) and low (L).	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544959.xml
546196	GATA: a graphic alignment tool for comparative sequence analysis	Background Several problems exist with current methods used to align DNA sequences for comparative sequence analysis. Most dynamic programming algorithms assume that conserved sequence elements are collinear. This assumption appears valid when comparing orthologous protein coding sequences. Functional constraints on proteins provide strong selective pressure against sequence inversions, and minimize sequence duplications and feature shuffling. For non-coding sequences this collinearity assumption is often invalid. For example, enhancers contain clusters of transcription factor binding sites that change in number, orientation, and spacing during evolution yet the enhancer retains its activity. Dot plot analysis is often used to estimate non-coding sequence relatedness. Yet dot plots do not actually align sequences and thus cannot account well for base insertions or deletions. Moreover, they lack an adequate statistical framework for comparing sequence relatedness and are limited to pairwise comparisons. Lastly, dot plots and dynamic programming text outputs fail to provide an intuitive means for visualizing DNA alignments. Results To address some of these issues, we created a stand alone, platform independent, graphic alignment tool for comparative sequence analysis (GATA ). GATA uses the NCBI-BLASTN program and extensive post-processing to identify all small sub-alignments above a low cut-off score. These are graphed as two shaded boxes, one for each sequence, connected by a line using the coordinate system of their parent sequence. Shading and colour are used to indicate score and orientation. A variety of options exist for querying, modifying and retrieving conserved sequence elements. Extensive gene annotation can be added to both sequences using a standardized General Feature Format (GFF) file. Conclusions GATA uses the NCBI-BLASTN program in conjunction with post-processing to exhaustively align two DNA sequences. It provides researchers with a fine-grained alignment and visualization tool aptly suited for non-coding, 0–200 kb, pairwise, sequence analysis. It functions independent of sequence feature ordering or orientation, and readily visualizes both large and small sequence inversions, duplications, and segment shuffling. Since the alignment is visual and does not contain gaps, gene annotation can be added to both sequences to create a thoroughly descriptive picture of DNA conservation that is well suited for comparative sequence analysis.	Background The most widely used methods for aligning DNA sequences rely on dynamic programming algorithms initially developed by Smith-Waterman and Needleman-Wunsch [ 1 , 2 ]. These algorithms create the mathematically best possible alignment of two sequences by inserting gaps in either sequence to maximize the score of base pair matches and minimize penalties for base pair mismatches and sequence gaps. Although these methods have proven invaluable in understanding sequence conservation and gene relatedness, they make several assumptions. One of their assumptions in generating the "best" alignment is that sequence features are collinear. For example, segments X, Y, Z in sequence one are also ordered as X, Y, and Z in sequence two. Another assumption is that short segments, like Y, have not become inverted or duplicated (e.g. X, Y, Y', Z). These rearrangement events are prone to be gapped out in dynamic programming and thus described as unrelated. Local alignment algorithms can be used to identify these rearrangements provided an exhaustive search is performed, but typically, only the highest scoring local alignments are considered valid and other, lower scoring local alignments are assumed to be spurious matches between unrelated sequences. When aligning protein coding sequences, dynamic programming works quite well. Evolution exerts significant functional constraint on protein coding sequences. When an inversion, duplication or segment-shuffling event occurs, the protein is often compromised by truncation due to the introduction of frame shifts and stop codons. These deleterious mutations are typically lost and not observed in the surviving population. When aligning this type of constrained sequence element, dynamic programming works quite well. Functional non-coding sequences do not appear to be as constrained in the ordering of elements as protein coding sequences [ 3 - 6 ]. Compact cis-regulatory modules, for example, enhance or suppress eukaryotic gene expression in response to external stimuli and play key roles in development and differentiation. One of the best characterized eukaryotic enhancers is the even-skipped stripe 2 element in Drosophila that controls transcription of the second transverse stripe of even-skipped mRNA during embryogenesis. Functional and comparative sequence analysis of stripe 2 clearly demonstrate that the enhancer maintains its specific activity across species yet displays significant small-scale insertions, deletions, and rearrangements of transcription factor binding sites within the module [ 7 , 8 ]. Tracing the evolutionary path of such non-coding elements is proving difficult with current alignment tools and may be assisted by a visual alignment program like GATA. Implementation GATA employs a two tiered architecture in aligning DNA sequences. GATAligner executes and processed BLASTN output. GATAPlotter displays the processed alignments and annotation from GATAligner. GATAligner The GATAligner application (figure 1 ) uses the NCBI bl2seq and BLASTN programs [ 9 , 10 ] to generate all possible local alignments between two input DNA sequences that score above a very low cut off (see Table 1 ). To avoid problems associated with visualizing both large and small local alignments, see Results/ Discussion, a sliding window is advanced at one base intervals across each local alignment. Windowed sequences scoring above a defined score are saved. To reduce the number of windowed sequences, each is compared to its neighbours and joined if they are of the same score and orientation. The score is not changed. These "sub-alignment" objects contain several features: a score, an orientation, a reference to the parental local alignment, the aligned sequences, and start and stop coordinates for each sequence. The sub-alignments are then saved to disk. This alignment and post-processing takes less than a minute for two 50 kb sequences using a window size of 24 and a score cut off of 25 bits on an 800 MHz PowerPC laptop computer. Our initial goal was to create a high resolution sequence alignment and visualization tool to use in identifying small sequence rearrangements, like those associated with evolving non-coding regulatory DNA. We initially divided the first sequence into overlapping windows offset by one base pair. A Smith-Waterman dynamic programming algorithm was then used to align each window against the entire second sequence. Windows were scored, merged, and saved as above. Although this method is more rigorous than using BLASTN, it took 20–50 times as long, and did not produce significantly different results (data not shown). It should be noted that BLASTN requires seven consecutive identical bases to align two sequences. Thus in rare cases, some windows will be missed, for example, GGGGGGcTTTTTTaCCCCCCgAAAAAA versus GGGGGGaTTTTTTgCCCCCCtAAAAAA. GATAPlotter The GATAPlotter application (figure 2 ) takes sub-alignment objects created by GATAligner and displays them graphically. Two boxes connected by a line are used to represent each sub-alignment. The boxes are plotted against horizontal representations of the input sequences with the reference sequence on top. The size of each box is determined by the start and stop positions in the sub-alignment. The shading of the boxes and connector line are scaled according to the sub-alignment score where solid black represents the highest score obtained, light grey the lowest. Lastly the colour of the connecting line is used to indicate the sub-alignment orientation, black for +/+, red for +/-. Where windows overlap, those with the highest score are displayed on top. Single clicking on overlapping windows retrieves all of the underlying windowed sequence alignment information. Double clicking fetches all of the associated local alignment information as parsed from BLAST. GATAPlotter also has the capability to display extensive gene annotation for one or both input sequences. The principle component of gene annotation rendering by GATA is the "GeneGroup" (figure 3 ). Each GeneGroup is drawn independent of other GeneGroups and is allowed to float within the panel to avoid overlap. A typical GeneGroup contains one DNA sequence from which one or more "TransGroups" are derived. Each TransGroup contains exons, an RNA transcript and possibly a protein translation. Each of these features are described using the Berkeley Drosophila Genome Project GFF format [ 11 ]. Coding and non-coding DNA sub features are only created in the presence of translation features and represent the most conservative estimation of what is protein coding sequence. If any translation predicts a larger coding region than the others, this is adopted for the entire GeneGroup. A point of confusion by many is that exons encode protein peptides. This is not necessarily true (i.e. 5' and 3' UTRs) and has lead to a variety of annotation rendering errors. When parsing a GFF file, GATAPlotter looks for the following GFF features: exon, translation, transcript, gene, rna, transpos, misc, where an * represents one or more wild cards. These wild card "genes" are interpreted as the closing feature on the GeneGroup from which all the proceeding TransGroups are derived. Annotation for both strands is drawn together; arrows are used to indicate orientation. Features not recognized by the parser are interpreted as novel user defined elements and rendered in their own tracks. GFF annotation examples, templates, and extensive descriptions are provided under the GATAPlotter "Documentation" menu. See table 2 and 3 for a complete listing of program options. Results and discussion To illustrate the types of rearrangements GATA can distinguish, examine figures 4 and 5 . Both contain alignments between Drosophila melanogaster and D. pseudoobscura . Figure 4 contains three highly similar genes. Figure 5 , inversions of putative enhancers. Annotation for each was obtained from whole_genome_annotation_dmel_RELEASE3-1.gff [ 11 ]. Orthologous sequences were isolated using the FlyCatcher program [ 12 ]. In cases where alignment windows overlap, the lowest scoring windows are drawn first and higher scoring windows placed on top. Both, connecting lines and their associated boxes are shaded according to score. Several related alignment and visualization tools have proven useful in comparative sequence analysis. Dot plot analysis can be used to identify duplications and inversions. Programs such as Dotter, JDotter, Dotlet, and Family Relations [ 13 - 16 ] generate graphical representations of sequence conservation by scoring identity between two perpendicular sequence representations. Although, mapping annotation to dot plots containing duplications and inversions is rather difficult and counter intuitive. Programs such as Artemis/ACT, LALNVIEW and to some extent, PLALIGN [ 17 - 20 ], utilize alignment information generated from dynamic programming algorithms to create box-line-box representations of each local alignment. These are similar to GATA but do not divide local alignments into window scored sub-alignments. This is unfortunate since window scoring enables a more detailed view of the actual sequence similarity within a large local alignment. Moreover, meaningful visualization using these browsers requires setting a high cut off score for the visualized local alignments. This effectively eliminates smaller, lower scoring local alignments that may provide alternative or even better inverted local alignments. GATA's windowed post-processing overcomes these associated problems. One program that is proving quite useful in avoiding the collinearity problem while still using a dynamic programming algorithm is Shuffle LAGAN [ 21 ]. Alignments generated by Shuffle LAGAN are combine with alignment annotation viewers such as VISTA [ 22 , 23 ] to align entire genomes. K-BROWSER/ MAVID and Mauve are two additional genome browser/ aligners that look equally promising [ 24 - 26 ]. Although, it should be noted, these programs are designed to provide genome wide alignments and identify large-scale rearrangements, GATA is best suited at interrogating non-coding DNA sequences between 0–200 kb in size for both large and small rearrangements. One of the major challenges facing bioinformaticians is the development of alignment and visualization tools for multi-species comparative sequence analysis. Within the fly community alone, 12 divergent species of diptera and hymenoptera will be sequenced within 3 years. A variety of higher eukaryotes including human, mouse, rat, dog, chimp, cow, chicken, opossum, and platypus have or are in the process of being completely sequenced. How can one visualize the alignment and species-specific annotation for 12 orthologs of a particular gene or a genomic segment? The GATA alignment paradigm is well suited to this challenge and will play a prominent role in GATA's development. Conclusions As comparative sequence analysis accelerates, scientists need more sophisticated alignment and visualization tools to define the evolutionary relationships and functional significance between particular orthologous sequences. This is especially true for regulatory, non-coding DNA that can show significant small-scale rearrangements. These new tools must incorporate detailed annotation alongside views of sequence conservation while providing easy access to the underlying sequence information. GATA provides one such solution. Availability and requirements Project name: GATA, graphic alignment tool for comparative sequence analysis. Project home pages: and Operating system(s): Platform independent Programming language: Java Other requirements: Java 1.4 or higher License: GNU GPL Any restrictions to use by non-academics: None Authors' contributions DAN designed and constructed the GATA programs with advice and supervision from MBE.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546196.xml
548144	PGF2alpha induced differential expression of genes involved in turnover of extracellular matrix in rat decidual cells	In the rat, the decidual tissue is an important component for maternal recognition of pregnancy. Decidualization can be induced by either the implantation of the blastocyst or by artificial stimuli. The process of decidua formation or decidualization, is characterized by growth and differentiation of endometrial stromal cells. Prostaglandin F2alpha (PGF2α) has been shown to be involved in inhibition of implantation, alteration of embryo development, induction of luteal regression, and the mediation of pregnancy loss induced by microorganism infections. In order to establish a direct role for PGF2α in decidual function, we have evaluated its effects on the expression of an extensive array of genes using primary decidual cell culture. Upon treatment with PGF2α sixty genes were significantly down-regulated whereas only six genes were up-regulated (from a total of 1176 genes studied). Interestingly, the majority of the genes inhibited by PGF2α are either directly or indirectly involved in the turnover of the extracellular matrix (ECM). Genes such as gelatinase A (MMP2), cathepsin L, tissue inhibitor metalloproteinases 2 (TIMP2) and 3 (TIMP3), plasminogen activator inhibitor1 (PAI1), tissue type plasminogen activator (tPA), urokinase plasminogen activator (tPA), endothelin 1, calponin, carboxypeptidase D and calponin acidic were down regulated. The opposite effect was observed for prostromelysin 53 kDa (proMMP3), plasma proteinase I alpha and alpha 1 antiproteinase, all of which were significantly up-regulated by PGF2α. The results strongly suggest that the abortificient role of elevated levels of PGF2α after implantation is due, in large part, to inhibition of genes involved in the normal turnover of the extracellular matrix necessary for decidual formation.	Background The establishment of successful pregnancy requires a profound reorganization of uterine tissues. Rapid growth and differentiation of the endometrial stroma is the earliest and most striking event in pregnancy. Differentiation of stromal cells leads to the formation of unique cells, termed decidual cells, which differ greatly from the original stromal cells [ 1 ]. The decidua is an important component in the maternal recognition of pregnancy and can be induced by either the implantation of the blastocyst or by artificial stimuli. An interesting feature is that – the growth and differentiation of the endometrial cells – occurs differently in different regions of the uterus [ 1 ]. Mesometrial decidual cells are formed after the antimesometrial decidua and their death takes place after antimesometrial cell degeneration. Regression of the decidual cells appears to be controlled by an intrinsic cell death program of apoptosis, which takes place after day 10 of pregnancy in the antimesometrial region first, followed by the mesometrial region [ 2 ]. Prostaglandin F 2α can induce different biological actions at the beginning and at the end of pregnancy. PGF 2α , Prostaglandin E 2 (PGE 2 ), and 6-keto-PGF 1α are produced by the pregnant uterus [ 3 ]. An increase of uterine PGE 2 and PGF 2α is observed on day 5 of pregnancy, allowing the decidualization process to take place. When embryo access to the uterus is impaired, production of prostaglandins (PGE 2 and PGF 2α ) is suppressed [ 4 ]. During postimplantation, PGE2 returns to the original preimplantation levels, but PGF 2α decreases [ 4 ]. Whereas PGF2α contributes to the process of decidualization, implantation and recognition of pregnancy. An increase in PGF 2α over certain values can terminate pregnancy [ 6 ]. A high level of PGF 2α is known to induce inhibition of implantation, alteration of embryo development, and induction of luteal regression [ 5 ]. During infection, an inflammatory response mediated by cytokines can be generated [ 7 ]. This release of PGF 2α can induce premature uterine contraction and premature labor [ 8 ]. Depending on which stage of pregnancy an increase of PGF 2α secretion occurs, it can alter implantation, induce abortion or even embriolethality [ 7 , 8 ]. Because an increase in PGF 2α levels after implantation can be detrimental for the progression of pregnancy, the aim of this investigation was to determine whether PGF 2α affects directly decidual cells and leads to disturbance in the expression of genes crucial for decidual survival. Methods Animal model Pseudopregnancy was induced by mating Holtzman Sprague Dawley female rats with vasectomized male rats. The day a vaginal plug was found was designated day 1 of pseudopregnancy. Decidualization of the uterine endometrium was induced by scratching the antimesometrial surface with a hooked needle on day 5 of pseudopregnancy under ether anaesthesia. Animals were housed in a controlled environment (22–24°C) and kept under controlled conditions (lights on; 0500–1900 hrs) with free access to standard rat chow and water. The University of Illinois at Chicago animal care and use committee approved the animal care and handling. Primary decidual cell culture and hormone treatment For each experiment 15 pseudopregnant rats were used. Decidual cells were isolated as previously described [ 9 ]. Cells were pooled and seeded into six-well plates (1.4 × 10 6 cells/ well). They were allowed to attach 3–4 hrs before washing in PBS to remove erythrocytes. Cells were incubated in RPMI 1640 without phenol red (Mediatech, Whashington DC), containing 1% CD-FBS (HyClone Laboratories Inc, Logan, UT), and treated with or without high levels (5 μM [ 10 , 11 ]) of PFG 2α (Sigma, St. Louis, MO) for 12 hr). After incubation, the cells were harvested in cold PBS, quick frozen in liquid nitrogen and kept at -80 C until RNA isolation. Gene identification by cDNA array Total RNA was isolated from primary decidual cells (PDC) by using a RNAII isolation kit (Clontech, Palo Alto, CA), following the manufacture's instructions. RNA isolated from wells treated with either PGF 2α or vehicle were pooled independently and subjected to cDNA array. cDNA probes were generated from 4 μg of total RNA in a reverse transcriptase reaction using a mix of dATP, dTTP and dGTP (Takara Biomedicals, Shiga, Japan) plus [α 32 P]-dCTP (Amersham, MO), and a mixture of primers specific to each gene present in the array. All probes used had 5 to 10 × 10 6 cpm and the difference between control and experimental probes in each assay was less than 10%. cDNA were hybridized to an Atlas Rat 1.2 Array (#7854-1) nylon membrane (Clontech, CA). Hybridization and post-hybridization washes were performed according to the manufacturer's protocol. The signal was scanned with a phosphorImager (Molecular Dynamics, CA) after 48 h of exposure. Control and experimental RNA were always processed in parallel. Data analysis Spot intensities from scanned membranes were analyzed using the AtlasImage 1.5 software (Clontech, CA). Grids were orientated manually and adjusted to ensure optimal spot recognition using AtlasImage's fine tuning tools, discarding spots with dust or locally high background. The software makes the analysis for background correction and normalization versus housekeeping genes. It also calculate the ratio and indicates which genes has a ratio > 2 (up regulated), or ratio < 0.5 (down regulated) or 0.5 < ratio < 2 (equal expression). Gene expression data were normalized using the Sum method included in the Atlas Image software. Data points where the expression was not greater than two standard deviations were discarded. For the final analysis, data points were the averages from triplicates and any non-reproducible data were discarded. The relative RNA expression with differences between control and treatment being higher or equal to 2 and lower than 0.5 were considered significantly different [ 12 ]. Results Effect of PGF 2α on gene expression in primary decidual cells From the total genes available in the rat 1.2 membrane arrays, only 20% of the genes were detected in rat primary decidual cells. Sixty genes were significantly down regulated, whereas only six genes were up regulated by PGF 2α (Figures 1 and 2 ). PGF 2α -receptor gene expression was similar in both control and treated groups. On the other hand, no differences in the expression of housekeeping genes such as ribosomal proteins and β-actin could be detectedin control and PGF 2α treated cells (data not shown). Figure 1 Representative cDNA expression array using mRNA obtained from rat primary decidual cells treated with either PGF 2α or vehicle. Figure 2 cDNA expression array of rat primary decidual cells treated with PGF 2α or vehicle. Effect of PGF 2α on the expression of genes related to the extracellular matrix (ECM) The majority of genes whose expression was affected by PGF 2α in primary decidual cells are genes involved in the regulation of the ECM. Genes such as gelatinase A (MMP2), cathepsin L, tissue inhibitor metalloproteinases 2 (TIMP2) and 3 (TIMP3), plasminogen activator inhibitor1 (PAI1), tissue type plasminogen activator (tPA), urokinase plasminogen activator (uPA), endothelin 1, calponin, carboxypeptidase D and calponin acidic were down regulated. The opposite effect was observed for prostromelysin 53 kDa (proMMP3), plasma proteinase I alpha and alpha 1 antiproteinase – all of which were significantly up regulated by PGF 2α (Table 1 ). Table 1 Effect of PGF 2α on the expression of genes related to the extracellular matrix (ECM). Gene Bank number Swiss protein number Name Function Ratio % decrease Fold increase U65656 P97581 Gelatinase A Metalloproteinase 0.04 ± 0.02 87.53 Y00697 P07154 Cathepsin L Cysteine protease 0.09 ± 0.05 91.09 L31884 P30121 Tissue inhibitor metalloproteinase 2 (TIMP2) Protease inhibitor 0.05 ± 0.01 95.54 U27201 P48032 Tissue inhibitor metalloproteinase 3 (TIMP3) Protease inhibitor 0.04 ± 0.01 91.95 X63434 P29598 Urokinase type Plasminogen activator (uPA) Serine protease 0.14 ± 0.09 94.86 M24067 P20961 Plasminogen activator inhibitor (PAI) Protease inhibitor 0.09 ± 0.05 91.68 M2367 P19367 Tissue type plasminogen activator (tPA) Serine protease 0.17 ± 0.02 89.90 X71071 Q08290 Calponin Cytoskeleton 0.08 ± 0.004 92.50 U06755 P37397 Calponin, acidic Cytoskeleton 0.24 ± 0.13 76.76 U62897 O35850 Carboxypeptidase D Carboxypeptidase 0.24 ± 0.14 76.48 M64711 P22388 Endothelin 1 Hormone 0.34 ± 0.2 77.48 X02601 P03957 Polypeptide, 53 kDa, growth factor induced Metalloproteinase 2.54 ± 1.04 3 J03552 P14046 Plasma proteinase inhibitor α1-inhibitor III Protease inhibitor 3.64 ± 0.64 4 M32247 P17475 α1 antiproteinase Protease inhibitor 2.45 ± 0.44 3 Ratio = net intensity of PGF 2α normalized/net intensity of control normalized, and represent the mean ± SEM of each ratio from separate experiments. Ratio > 2 = up regulation; ratio < 0.5 = down regulation; 0.5 < Ratio < 2 = equal expression. PGF 2α regulation of genes related to the proteasome protein component system Messenger RNA encoding proteasome Iota, proteasome component C2, proteasome subunit RC7-I, 26-S proteasome regulator and proteasome component C3 were down regulated by PGF 2α . Only proteasome component C9 was up regulated after PGF 2α treatment (Table 2 ). Table 2 PGF 2α regulation of genes related to the proteasome component system. Gene Bank number Swiss Protein number Name Function Ratio % decrease Fold increase D10755 P34062 Proteasome Iota Proteosomal protein 0.07 ± 0.04 93.64 M29859 P18420 Proteasome component C2 Proteosomal protein 0.31 ± 0.15 78.28 U77918 Q63569 26-proteasome component C3 Proteosomal protein 0.39 ± 0.08 61.94 J02897 P17220 Proteasome component C3 Proteosomal protein 0.31 ± 0.02 59.02 D21799 P40307 Proteasome subunit RC7-I Proteosomal protein 0.18 ± 0.01 82.62 X55986 P21670 Proteasome subunit C9 Proteosomal protein 2.32 ± 0.72 3 PGF 2α regulation of genes involved in trafficking and signal transduction The mRNAs encoding for 14-3-3ε and z/δ proteins, and Annexin IV were down regulated in PDC after PGF 2α treatment (Table 3 ). PGF 2α down regulated the expression of serum glucocorticoid kinase (SGK), casein kinase I, extracellular signal-regulated kinase (ERK-1), male germ cell-associated kinase (MAK), and Wee tyrosine kinase. PGF 2α also significantly inhibited Crk-associated kinase (CAS), calcium calmodulin kinase II (CAMKII) and IV (CAMKIV), phospholipase C1, and protein phosphatases such as phosphatase 2A and protein tyrosine phosphatase 1B (Table 3 ). Table 3 PGF 2α regulation of genes encoding proteins involved in trafficking and signal transduction. Gene Bank number Swiss protein number Name Function Ratio % decrease Fold increase D17615 P97286 14-3-3 z/δ protein Traficking protein 0.14 ± 0.07 86.07 M84416 P29360 14-3-3 Epsilon protein Traficking protein 0.29 ± 0.15 71.63 L01624 Q06226 Serum Glucocorticoid kinase (SGK) Kinase 0.25 ± 0.12 93.49 L07578 Q06486 Casein kinase I δ Kinase 0.29 ± 0.15 72.18 M61177 P21708 Extracellular signal-regulated kinase 1 (ERK1) Kinase 0.29 ± 0.13 66.67 D29766 Q63766 Crk-associated kinase Kinase 0.09 ± 0.06 87.16 M35862 P20793 Male germ cell-associated kinase (Mak) Kinase 0.20 ± 0.11 78.6 D31838 Q63802 Wee tyrosine kinase Cell cycle kinase 0.23 ± 0.12 69.47 M20637 P10688 Phospholipase C δ1 Phospholipase PI kinase 031 ± 0.15 68.24 J04503 P20650 Protein phosphatase 2C α Phosphatase 0.18 ± 0.03 58.87 L12385 P26438 Protein Phosphatase 2A Phosphatase 0.38 ± 0.13 65.91 L13408 P15791 Ca +2 /Calmodulin dependent kinase II δ subunit Kinase 0.34 ± 0.12 72.12 M63334 P13234 Ca +2 /Calmodulin dependent kinase IV Kinase 0.29 ± 0.15 70.90 M33962 P20417 Protein Tyrosine phosphatase 1B Phosphatase 0.21 ± 0.08 79.77 D38224 P55260 Annexin IV (ANX4) Exocytosis protein 0.14 ± 0.08 86.34 L24810 Q64572 Ca +2 /Calmodulin Phosphorylase B Kinase 5.06 ± 2.06 5 PGF2α effect on the expression of genes associated with G-proteins, growth factors, chemokines and cytokines The 5-hydroxytryptamine receptor 2A (HTR2A), adenosine A2A and A2B receptors (ADORA2A and ADORA2B respectively), guanine nucleotide-binding protein G α3 subunit (GN-BPG α 3 ), guanine nucleotide-binding protein α stimulating (GN-BP α stimul.), Rab 11A and guanine nucleotide-binding protein α12 subunit (GN-BP α 12 ) gene expression were down regulated by PGF 2α (Table 4 ). The mRNA expression of growth factors such as bone morphogenetic protein 4 (BMP4), ADP ribosylation factor 5 and 6 (ADPRF5 and 6 respectively), glia maturation factor β (GMFB), and transforming growth factor β I (TGF-βI) decreased after PGF 2α treatment as well (Table 1 ). Interferon inducible protein (Table 4 ) and leukocyte common antigen (data not shown) were the only cytokine and chemokine related genes found significantly down regulated in response to PGF 2α . Conversely, PGF 2α stimulated the expression of basic fibroblast growth receptor factor 1 (bFGRF1) (Table 4 ). Table 4 Effect of PGF 2α on the expression of genes associated with G-proteins, growth factors, chemokines and cytokines. Gene Bank number Swiss protein number Name Function Ratio % decrease Fold change X52498 P17246 Transforming Growth Factor β1 Growth factor 0.33 ± 0.2 72 M60921 P27049 Antiproliferative B-cell translocation gene 2 (BTG2) Growth factor 0.28 ± 0.15 51 X61381 P26376 Interferon-induced protein Chemokine 0.25 ± 0.14 56 L11586 Q64604 Leukocyte common antigen Intracell. phosphatase 0.07 ± 0.02 83 Z22607 Q06826 Bone Morphogenetic protein 4 (BMP4) Growth factor 0.34 ± 0.17 78 L12384 P26437 ADP Ribosylation factor 5 Intracellular transducer 0.02 ± 0.008 67 L12385 P26438 ADP Ribosylation factor 6 Intracellular transducer 0.18 ± 0.07 83 Z11558 Q63228 Glia maturation factor β (GMFB) Growth factor 0.17 ± 0.07 83 M30705 P14842 5-hydroxytryptamine receptor 2A (HTR2A) G proteins 0.06 ± 0.03 93 S47609 P30543 Adenosine A2A receptor (ADORA2A) G proteins 0.13 ± 0.08 87 M91466 P29276 Adenosine A2B receptor (ADORA2B) G proteins 0.17 ± 0.07 83 M20713 P08753 Guanine nucleotide-binding protein G α3 subunit G proteins 0.22 ± 0.14 78 M17525 P04894 Guanine nucleotide-binding protein α stimulating G proteins 0.24 ± 0.13 75 M75153 P24410 Rab-11A G proteins 0.18 ± 0.11 82. D85760 Q63210 Guanine nucleotide-binding protein α12 subunit G proteins 0.15 ± 0.09 85 D12498 Q04589 Basic Fibroblast Growth Receptor factor 1 Growth factor 2.32 ± 1.07 3 Discussion Levels of PGF 2α in the uterus need to be under tight control to avoid interference with the establishment and progression of pregnancy. Pathophysiological elevations in PGF 2α lead to excessive uterine contractions. Therefore, PGF 2α production must be avoided to maintain a quiescent uterus [ 13 ]. In this paper we present initial data demonstrating that elevated PGF 2α can target the decidua, an endocrine tissue whose integrity is fundamental for the success of implantation and for the progression of pregnancy. Our results demonstrate that elevated PGF 2α down regulate the expression of decidual genes related to the proteasome protein component system, those involved in trafficking and signal transduction and genes associated with G-proteins, growth factors, chemokines and cytokines. However, interestingly, the main effect of PGF 2α is related to the turnover and degradation of the ECM which provides mechanical strength to the tissue. Thus, in addition to inducing uterine contractions, PGF 2α can have a noxious impact on the success of implantation and progression of pregnancy, at least in part by deregulating the turnover of the ECM in the decidua. The largest group of genes down regulated by PGF 2α is directly or indirectly related with the turnover and degradation of the ECM. Genes such as gelatinase A (MMP-2), TIMP2 and 3 [ 14 , 15 ], cathepsin L [ 16 ], carboxypeptidase D are included in different groups of proteases, and can directly affect the degradation of the ECM. Carboxypeptidase D (CPD) or metallocarboxypeptidase D is a 180-kDa protein that contains almost three carboxypeptidase-like domains, a transmembrane domain, and a cytosolic tail. This gene participates in the processing of proteins transiting the secretory pathway [ 17 ]. TGF-β is a key factor that favours accumulation of collagen, laminin and fibronectin in the ECM [ 17 ]. TGF-β can stimulate PAI-1, inhibiting the protease degradation of the ECM. Other factors included in the plasminogen/plasmin and fibrinolytic system, such as tPA, uPA and PAI-1, can participate in the tissue remodelling of the ECM directly through the binding to specific receptors, and through the transformation of the plasminogen precursor bound to the cell surface to plasmin, which is an active serine protease. Plasmin is able to degrade most of the components of the ECM either directly or indirectly by the activation of MMPs. All of the plasminogen/plasmin factors mentioned previously participate in the process of decidualization [ 19 - 21 ]. Endothelin-1 is a peptide characterized as a potent endothelial cell-derived vasoconstrictor. It is synthesized as an inactive precursor (preproendothelin) and processed to a mature active form (endothelin) by zinc metalloproteinases. The active form, endothelin-1, promotes synthesis of collagen types I and II by fibroblasts, affects the ECM remodelling and promotes the proliferation of mesangial cells [ 21 ]. Endothelin-1 is also associated with neovascularization and regulation of blood flow [ 22 ]. This vasoactive factor is also involved in the genesis of endothelial cells behaviour [ 22 , 23 ]. Annexin IV (ANX4, also called Lipocortin) was differentially down regulated by PGF 2α . This protein belongs to a family of intracellular proteins that binds membrane phospholipids in a calcium-dependent manner. Thus, it can also inhibit phospholipase A2 (PLA2) [ 24 ]. ANX4 can also directly bind glycosaminoglycan (GAG) and can be localized not only to the cytoplasm but also the cell surface or the extracellular compartment [ 25 ]. It has been proposed that ANX4 could affect the mobilization of different substrates involved in the regulation of the ECM. Moreover, Annexin IV and other annexins can act as ligands for proteoglycans localized on the cell surface, in the ECM, or in secretory granules [ 24 , 25 ]. Calponin (CaP) has three genetic isoforms, h 1 , h 2 and acidic calponin, and can be identified by the individual C-terminal tail sequences. The c-terminal sequences regulate actin association and the cytoskeleton [ 26 , 27 ]. It is known that Cap or basic Cap inhibit the actomyosin ATPase in a calmodulin dependent manner [ 27 ]. Basic Cap, by affecting microtubules assembly, can modify the cytoskeleton of the cells, and indirectly, the associated ECM [ 27 , 28 ]. On the other hand, acidic Cap belongs to the family of actin-associated proteins. It can interact with F-actin but not with microtubules, desmin filaments, and tropomyosin as basic Cap does. These properties suggest that acidic Cap is functionally distinct from basic Cap and could affect the ECM in a different manner [ 29 ]. Pro-stromelysin, α 1 antiproteinase, plasma proteinase-Iα inhibitor, basic fibroblast growth receptor factor 1, phosphorylase B and proteasome component C9 were the only genes up regulated by PGF 2α in PDC. Pro-stromelysin is a 53 kDa peptide and a pro-MMP3 precursor and is directly related with the ECM regulation [ 30 ]. Plasma proteinase 1α inhibitor belongs to a major group of proteins that includes α2-macroglobulin [ 31 ], an important protein secreted by the decidual cells which has a critical role in the control of implantation [ 31 ]. Growth factor receptors, such as bFGRF, are part of a multigene family of structurally related factors (FGFs 1 to 9), some of which bind heparin sulphate proteoglycans that are components of the ECM [ 32 ]. bFGRFs as well as bFGF are also temporally and spatially present in the pregnant rat uterus [ 33 - 35 ]. Another member of the TGFβ superfamily expressed in the uterus during early implantation and down-regulated by PGF 2α was BMP4, a gene which appears to be involved in specific stages of embryo development and is [ 36 ]. Proteasome component system proteins, such as proteasome Iota, proteasome component C2, proteasome subunit RC-7 I, 26 S-proteasome and proteasome component C3 were down regulated by PGF 2α . Proteosomal component systems are the main non-lysosomal proteolytic structures of the cells that participate in the elimination of abnormal proteins, short half-life proteins, and proteins controlling cell cycle [ 37 ]. During the process of cell differentiation, the level of proteasome expression and its localization varies. The proteasomal proteins can be intermediaries of ECM by contributing to the modulation of the cell cycle through the induction of proteasomal degradation of cyclin dependent kinase 2. Cell attachment to ECM components, such as fibronectin (FN), does not affect p21 mRNA levels, but the stability of the p21 protein decreased [ 36 ]. Kinase activities such as ERKs, calcium/calmodulin kinase II and IV, SGK, and casein kinase can affect the ECM downstream or upstream [ 38 ]. On the other hand, ECM can regulate the availability of substrates, as well as factors or effectors downstream or upstream related with different cascades of signal transduction pathways. For example, proteins such as decorin are components of the ECM in many tissues and appear to be involved in matrix assembly [ 39 ]. Decorin can cause the rapid phosphorylation of EGF with the concurrent activation of mitogen-activated kinase protein kinase signal pathway. Via TGFβ, decorin can interact with the MAP kinases signal transduction pathways and cross talk with calcium /calmodulin-dependent kinase II [ 40 ], which in the cDNA array was down-regulated by PGF 2α . Crk-associated substrate (CAS) was down-regulated by PGF 2α and is another gene related to the ECM. Active CAS can modulate changes in cell motility and gene expression by the various MAP kinase cascades, and modify the organization of the actin cytoskeleton [ 39 ]. Another gene down-regulated by PGF 2α was SGK, which is a transcriptionally-regulated serine/threonine protein kinase with 45–55% homology to the catalytic domain of Akt/PKB protein kinase A [ 40 ]. Skg is expressed in decidual tissue and can be activated by the phosphoinositide 3-kinase pathway (PI3-Kinase) through PDK1-mediated phosphorylation [ 41 ]. Traficking proteins such as 14-3-3 z/δ and ε are involved in the regulation of genes related to the ECM because it can block the activation cascade of signal transduction pathways [ 42 ]. Protein phosphatase type 2A (PP2A) was down regulated by PGF 2α . The first and most important point of control of PP2A is at the transcriptional level. The increase of phosphatase activity corresponded with a decrease in the phosphorylation of cellular proteins in anchorage-dependent cells, but much lesser regulated in anchorage-independent cells [ 43 ]. It is well known that members of the proinflammatory cytokine family can induce MMP expression in numerous tissues [ 44 ]. Also, cytokines and chemokines, such as interferon inducible protein and leukocyte common antigen, are associated with the local induction of MMP expression in response to proinflammatory cytokines. The indirect action of these molecules through MMPs may aid to generate different changes in the endometrial stroma during maternal recognition of pregnancy [ 44 ] We have also compared the pattern of gene expression in rat decidual tissue in vivo on day 12 of pseudopregnancy (when the decidua undergoes regression) with that of the PGF 2α treatment in rat PDC in vitro (data not shown). Interestingly, we found a 49% coincidence in the genes that were down regulated in both experimental situations. This coincidence suggests a relationship between the physiological regression and reorganization of the decidual tissue that occurs on day 12 of pseudopregnancy, with the pattern of gene expression in rat PDC after PGF 2α treatment. On the other hand, many of the genes whose expression was affected by PGF 2α in the rat PDC, are expressed in the decidua during decidualization and implantation. This suggests a possible connection between the action of PGF 2α and the physiology of the decidual tissue. Moreover, the ECM is an important component of the decidualization and implantation process. Any modification in its turnover or degradation could affect the formation of the decidua, blastocyst invasion, and the timing of decidual regression and reorganization. It is known that the expression of MMPs/TIMPs plays an important role in the control of implantation. If PGF 2α silences or decreases the expression of genes related with the systems aforementioned, it also could affect tissue remodelling by directly modifying the proteases involved in this process, or indirectly by affecting the ECM turnover and degradation, important in the accumulation of a spongy mass of tissue around each embryo during decidualization [ 45 ]. Most probably, many of PGF 2α 's effects on the expression of genes involved in ECM turnover are indirect. Alterations of genes involved in different signalling pathways such as ERK-1, CaMCKII, PLCδ1 and G-protein subunits may impact the expression of a great number of transcripts. In summary, our results show, for the first time, that pathophysiological concentrations of PGF 2α have a severe impact on the expression of numerous genes associated with the turnover of the ECM in the rat decidua. Future investigation should corroborate the differentially regulated genes, at the level of the message, protein, and in some cases, such as for the metalloproteinases, at the level of activity of the proteins. These data will contribute to the design of future studies on a cluster of gene candidates as targets of PGF 2α action in this endocrine tissue. Abbreviations TIMP2 = tissue inhibitor metalloproteinases 2; TIMP3 = tissue inhibitor metalloproteinases 3; PAI 1 = plasminogen activator inhibitor 1; uPA = urokinase type; tPA = tissue type palsminogen activator; TGFβ = transforming growth factor β; SGK = serum glucocorticoid kinase; ERK-1 = extracellular signal-regulated kinase-1; MAK = male germ cell-associated kinase; CAS = crk-associated kinase kinase; CamKII = calcium calmodulin kinase II; CamKIV = calcium calmodulin kinase IV; PLC δ 1 = phospholipase C delta 1; BMP 4 = bone morphogenetic factor 4; ADPRF 5 = ADP ribosylation factor 5; bFGFRF1 = basic fibroblast growth factor receptor F1; HTR2A = 5-hydroxytryptamine receptor 2A; ADORA2A = adenosine receptor A2A; ADORA2B = adenosine receptor A2B; GN-BPG α 3 = guanine nucleotide-binding protein G α 3 ; GN-BP α stimul. = guanine nucleotide-binding protein α stimulating; GN-BPG α 12 sub. = guanine nucleotide-binding protein α 12 subunit Authors' contributions EC and SFG carried out the decidualization, and primary decidual cell culture. EC isolated the RNA isolation, performed the cDNA array assay, the analysis of the results, and drafted the manuscript. GG conceived the study and edited the manuscript. All authors read and approved the final manuscript. Figure 3 PGF 2α has major effects on extracellular matrix (ECM) regulation	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548144.xml
512289	'Perinatal outcome in preterm premature rupture of membranes with Amniotic fluid index < 5 (AFI < 5)	Background Our purpose was to determine whether AFI<5 cm after preterm premature rupture of the membranes (PPROM) is associated with an increased risk of perinatal morbidity. Methods We performed a prospective cohort study of 95 singleton pregnancies complicated by preterm premature rupture of the membranes (PPROM) with delivery between 26 and 34 weeks gestation. Patients were categorized in two groups on the basis of amniotic fluid index<5, (AFI<5 cm)(n = 26) or AFI ≥ 5 cm (n = 69). Categorical data were tested for significance with the χ 2 and Fisher exact tests. Continuous data were evaluated for normal distribution and tested for significance with the student t test. All 2-sided p values < 0.05 were considered significant. Results Both groups were similar with respect to selected demographics, gestational age at rupture of the membranes, gestational age at the delivery, birth weight. Both groups were similar with respect to selected variable, latency until delivery, early onset neonatal sepsis, RDS and neonatal death. Patients with AFI<5 cm demonstrated greater frequency of C/S delivery for non reassuring fetal tests (23%vs 2.8%) (p = 0.001). Our study demonstrated that patients in group I had a significant increase in the frequency of clinical chorioamnionitis (P < 0/001). Post partum infections were not seen in 2 groups. Conclusions An AFI<5 cm after PPROM between 26 and 34 weeks gestation is associated with an increased risk of maternal infections and frequency of C/S.	Background Preterm premature rupture of the membranes (PPROM) is one of the most common complications of the pregnancy. Preterm PROM is an important cause of perinatal morbidity and mortality, particularly because it is associated with brief latency from membrane rupture to deliver, perinatal infection, and umbilical cord compression due to oligohydramnios. PPROM is multifactor in nature. In any given patient, one or more path physiologic processes may be evident. Choriodecidual infection or inflammation appears to play an important role in etiology of preterm PROM, especially at early gestational ages. It has also been proposed that amniotic fluid posse's certain bacteriostatic properties that protect against potential infections processes and that a decrease in amniotic fluid volume may impair the gravid women ability to combat such infections, this latter hypothesis was tested by Vintziuleos, et al [ 2 ]. These investigations demonstrated that patients with olighhydramnios (AFI<5), were at greater risk of having chorioamnionitis and subsequent sepsis in the neonate [ 2 , 8 ]. Our purpose was to determine whether patients with PPROM and amniotic fluid index <5 cm (AFI< cm) are at an increased risk of having perinatal morbidity. Methods We performed a prospective cohort study of infants delivered between 26 weeks – 34 weeks gestation after preterm premature rupture of the membranes. This study was performed at Vali-e-Asr hospital at the medical university of Tehran between October 2000 and February 2002. All patients were at between 26 and 34 of weeks of pregnancy as best estimated by LMP, and confirmed by ultrasonography. In all patients rupture of the membrane was diagnosed by sterile speculum examination using pooled fluid, fern test and Nitrazine paper test. Patients with clinical chorioamnionitis, nonreassuring fetal status, obstetrical indication for immediate delivery, major congenital anomalies and advanced labor (cervical dilatation >3), and a growth-restricted fetus at initial admission were excluded. 95 singleton pregnancies have observed until 34 weeks of gestation. At admission, all patients had an ultrasonographic examination, which included confirmation of the estimated gestational age, and cumulative 4 guardant AFI measurements, as previously described by Phelan et al. [ 3 ] All patients received antibiotic prophylaxis at admission consisting of Ampicillin plus Erythromycin for 7 days following. In addition all patients received a single course of Betamethasone, consisting of two 12 mg Betamethasone injections during the first 24 hours after admission. Fetal surveillance incorporated daily non-stress testing. For fetuses with non-reassuring, non stress test results, biophysical profile assessments were performed. Indications for delivery included: labor, the diagnosis of clinical chorioamnionitis or non-reassuring fetal test results. Eligible patients were subsequently categorized into 2 groups on the basis of the admission AFI measurement. Patients in group 1 were those with an AFI<5 cm, whereas those in group 2 had AFI ≥ 5 cm. The 2 groups were compared for demographic characteristics, the estimated gestation age at both rupture of the membranes and delivery, latency until delivery, mode of delivery, birth weight, the development of clinical chorioamnionitis, postpartum endometritis, early onset neonatal sepsis and respiratory distress syndrome. Neonatal sepsis was diagnosed by positive blood, urine, or cerebrospinal fluid cultures. Possible neonatal sepsis was diagnosed when two or more of the following criteria were present: white blood cell count less than 5000/mm3, polymorphonuclear counts less than 1800/mm3, ratio of bands to total neutrophil counts greater than 0.2. Early onset neonatal sepsis was defined as sepsis in a neonate with positive culture results or possible sepsis within the first 48 hours of life and prior to the antibiotic administration. The clinical diagnosis of chorioamnionitis was made in presence of two or more of the following criteria: maternal fever greater than 38 C, maternal tachycardia (120 beats per minute or more), leukocytosis (greater than or equal to 20,000/mm3 white blood cell), fetal tachycardia (greater than 160 beats per minute), uterine tenderness, and foul-smelling amniotic fluid. Categorical data were tested for significance with the χ 2 and Fisher exact tests. Continuous data were evaluated for normal distribution and tested for significance with the student t test. All 2-sided p values <.05 were considered significant. Results A total of 95 patients with preterm premature rupture of membranes were included in the study; 26 were included in group I (AFI<5 cm) and 69 in group II (AFI ≥ 5 cm). The 2 groups were similar with respect to maternal age, Parity, and gestational age at admission (table I ). Gestational age at delivery and latency period until delivery, birth weight were not significantly different between the 2 groups. Both 2 groups had similar proportions of vaginal deliveries, however, in group I cesarean delivery was more likely to be performed because of non-reassuring fetal status (table I ). Table 1 Demographic variable between two groups with PPPROM Variable AFI<5 (n = 26) AFI ≥ 5 (n = 69) Statistical significance Maternal age 25,1 ± 5.2 26.3 ± 4.9 Ns Parity 3 ± 1,5 2.4 ± 1.2 Ns Gestational age at admission 31.5 ± 2.00 33.5 ± 1.8 Ns Gestational age at delivery 32.6 ± 4.0 34.5 ± 3.7 Ns Latency (Mean ± SD) 7.6 ± 4.0 6.6 ± 5.2 Ns C/S rate for fetal distress 6(%23) 2(28%) P = 0.001, s Birth weight (mean) 2120 gr 2445 gr Ns Ns = not significant P-value < 0.05 = significant Our study demonstrated that patients in group I had a significant increase in the frequency of clinical chorioamnionitis (P < 0/001). Post partum infections were not seen in 2 groups. Early onset neonatal sepsis, respiratory distress syndrome (RDS), neonatal deaths were not significantly different between the 2 groups. (Table II ) Table 2 Maternal and neonatal outcome comparison between two groups with PPROM Out come AFI<5(26) AFI ≥ 5 (69) Statistical significant Chorioamnionitis 5(19/2%) 2 (3%) p < 0/001 Early onset sepsis 7(30/4%) 19(27/9%) p = 0.819 RDS 6(26/1%) 8 (11/8%) p = 0.1 Neonatal death 4(17/4%) 5(7/4%) p = 0.163 P-value < 0.05 was considered significant Discussion Complications of preterm premature rupture of membranes count for approximately 25% to 33% of all preterm deliveries. Approximately, 75% of women will be delivered within 1 weeks of presentation [ 1 , 5 ]. More recent evidence suggests that membrane rupture is also related to biochemical processes, including disruption of collagen within the extracellular matrix of the amnion and the chorion and programmed death of cells in the fetal membranes. It has been proposed that the fetal membranes and the maternal uterine lining (decidua) respond to various stimuli, including membrane stretching and infection of the reproductive tract, by producing mediators, such as prostaglandins, cytokines, and protein hormones that govern the activities of matrix-degrading enzymes. When the fetal membranes rupture at term or before, the options are expectant management (with close observation for signs of labor, non reassuring fetal-heart-rate patterns, or intrauterine infection) or induction of labor [ 10 , 11 ]. Expectant management with antenatal antibiotics and corticosteroid administration are recommended the standard of care in the setting of PPROM at gestational ages of ≤ 34 [ 1 , 4 , 9 , 10 ]. Current evidence suggests adjunctive antibiotic therapy to reduced gestational age-dependent and infectious infant morbidity. Amniocentesis and amniotic fluid volume have been advocated as a useful adjunct for identifying these patients [ 9 ]Several studies have implicated oligohydramnios in patients with preterm premature rupture of the membranes as a significant risk factor for perinatal infection, and fetal distress, cesarean delivery, and neonatal death [ 5 - 8 , 10 ]. In our study the finding of an AFI<5 cm after preterm premature rapture of the membranes was associated with the development of chorioamnionits. However, patients in the group with AFI<5 did not have a shorter latency until delivery. Our study did not demonstrate an association between the development of chorioamnionitis and latency interval in patients with ruptured membranes (P = 0/783), because the latency period in our study were not significantly different between 2 groups. Other investigators have demonstrated an association between the development of chorioamnionitis and a shorter latency in patients with PPROM [ 5 - 8 ]. Post partum infections were not seen in our study. Perhaps, decreasing of post partum infections rates in our cases were the reason of using antibiotics after C/S. We were used intravenous Cephazolin for 48 h and then oral Cephalexin for 5 days after C/S. Our study demonstrated that the patients with oligohydramnios were more likely to undergo cesarean delivery because of non-reassuring fetal heart rate patterns and is consistent with the findings of these other studies [ 5 , 6 , 8 ]. This study didn't show an increased frequency of early onset sepsis in the group I (AFI<5 cm), because all newborn infants in the study were treated possible sepsis with clinical symptom and laboratory evidence. 7 of the included neonates had positive blood cultures or spinal fluid cultures; as a result, there was not sensitive mechanism for appropriately determining the diagnosis of early sepsis. The negative cultures in the neonates with possible sepsis may be related to inadequate culturing techniques or the inherent difficulty encountered by most laboratories in isolating anaerobic bacteria. Perhaps, Diagnosis of early onset neonatal sepsis and close observation for early signs of sepsis and more aggressive evaluation and early treatment for neonatal sepsis have decreased early onset sepsis in 2 groups. Preterm premature rupture of the membranes is associated with a significant decrease in the frequency of neonatal respiratory distress syndrome. In the Sims EJ study (2002), The frequency of respiratory distress syndrome in the neonate complicated with PPROM was (17%).11 This study evaluate the effect of AFI on the frequency of respiratory distress syndrome among two group that are complicated with PPROM. The frequency of respiratory distress syndrome in the neonate was not significantly lower in the group (II) than in the group I. (11/8% vs26/1%) (P < .01). The identification of oligohydramnios, defined as an AFI<5 cm, patients with preterm PPROM appear to indicate a significant risk of chorioamnionitis and early onset neonatal sepsis. These finding can aid in the counseling of patients with PPROM and may have several clinical application. Management of PPROM requires an accurate diagnosis as well as evaluation of costs and the risks and the benefits of continued pregnancy or expeditious delivery. It is important that the patient be well informed regarding the potential for subsequent maternal, fetal, and neonatal complications regardless of the management approach. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC512289.xml
406405	Dissecting the Complexities of Glucose Signaling in Yeast	xx	An organism's survival depends on developing effective strategies for identifying and adapting to available sources of food. Even organisms as small as the budding yeast Saccharomyces cerevisiae can respond in a complex way to the presence of different energy sources. While yeast can metabolize many different sugars, glucose provides the highest energy yield. To achieve this energy efficiency, yeast cells rely on specialized enzymes and metabolic states that mediate glucose metabolism by sensing and responding to this sugar. The presence of glucose triggers a rapid and dramatic change in the expression of about a quarter of S. cerevisiae 's 5,500 genes. Ras at the plasma membrane of S. cerevisiae Many of the cellular participants in the glucose “sense-and-response” pathway have been identified, but their exact relationships remain unknown. The gene expression response to glucose has been previously characterized through the use of a cDNA microarray, which allows simultaneous assessment of the transcriptional state of every gene. In research reported in this issue, James Broach and his colleagues at Princeton University have attempted to connect the individual components in this complex pathway by performing microarray analysis on a series of mutants (yeast strains with defects in specific proteins). They link certain portions of the response to known proteins and begin to understand how the pieces of this pathway fit together. The researchers initially focused on two proteins, called Ras2 and Gpa2. These proteins have been previously implicated in the transcriptional response to glucose and are members of a well-established family of signaling proteins. To investigate the roles of these proteins in the response, the researchers used mutants of Ras2 and Gpa2 that could be activated on demand and then performed microarray analysis to see which genes responded to activation of these proteins. They found that even in the absence of glucose, activation of either protein induced a transcriptional profile almost identical to the profile generated in yeast exposed to glucose. This shows that the complex and dramatic transcriptional response to glucose can be recapitulated by the activation of a single protein. The Ras2/Gpa2 pathway, however, is not the whole story. The group then went on to show that not all glucose-responsive genes are regulated in the same manner. They found that another pathway, independent of Ras2 and Gpa2, is able to elicit a portion of the transcriptional response to glucose. The partial redundancy of these pathways is a curious phenomenon and bears further investigation. The authors have achieved an initial step in mapping the topology of the intricate signaling pathway (or pathways) involved in the response to glucose. Furthermore, the approach—that of using microarray analysis of mutants in a pathway to deduce the mechanisms of regulation—will be useful in efforts to map other complex responses in yeast as well as in higher organisms.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC406405.xml
503390	Geographic correlation between deprivation and risk of meningococcal disease: an ecological study	Background Meningitis caused by Neisseria meningitidis is a serious infection which is most common in young children and adolescents. This study investigated the relationships between the incidence and age distribution of meningococcal disease, and socioeconomic environment. Methods An ecological design was used, including mapping using a Geographical Information System (GIS) at census ward level. Results Incidence of meningococcal disease was highest in the most deprived wards, with a relative risk of 1.97 (1.55 – 2.51). Mapping revealed geographical coincidence of deprivation and meningococcal disease, particularly in urban areas. Two-thirds of the increased incidence was due to cases in the under fives. Conclusions The results suggest that area deprivation is a risk factor for meningococcal disease, and that its effects are seen most in young children.	Background Infections caused by Neisseria meningitidis are still an important cause of morbidity and mortality in the United Kingdom, with 1548 cases notified in England and Wales in 2001[ 1 ], an incidence rate of 3.2 cases per 100,000 per year, and a case fatality rate of around 10%[ 2 ]. The infection is most common in infants and adolescents, with peaks of incidence at 0–5 years and 15–19 years[ 2 ]. Outbreaks cause high degrees of anxiety in local populations, though 95% of cases are sporadic[ 2 ]. It is not clear how the bacterium spreads through the population in time and space, or what determinants are most important in each area. Suggested environmental risk factors for meningococcal infection include passive smoking[ 3 ], overcrowding[ 4 ], and weather conditions[ 5 , 6 ]. It is accepted that many infectious diseases preferentially affect the most disadvantaged in society. This has been discussed at historical[ 7 ], global[ 8 ], and national[ 9 ] levels, but new geomapping techniques have shown that it holds true when comparing small areas such as postal or electoral districts. The environmental and social factors conferred upon an individual through their residence in a particular area may be as important as the individual 'risk factors' where communicable disease is concerned[ 10 ]. Mapping of socioeconomic status and certain infectious diseases, in particular sexually transmitted infections [ 10 - 13 ], has shown a relationship between area socioeconomic indicators and disease incidence. Geographical Information Systems (GIS) techniques are able to both test hypotheses about the geographical distribution of disease and to display environmental characteristics and disease incidence in a clear and interpretable way. This study was prompted by an impression, gained from the practice of a consultant in communicable disease control (CCDC), that infant cases of meningococcal disease came from families of low socioeconomic status, whilst older cases tended to be drawn from families of higher socioeconomic status. Studies in northeast Thames[ 14 ], southwest England[ 15 ], and Wales[ 16 ] have shown an increase risk of meningococcal disease in more deprived areas, particularly in the under five age group. This study aims to test these hypotheses concerning age, area socioeconomic status and meningococcal incidence, using a larger sample size and a Geographical Information System (GIS). GIS methods are used to display the information on deprivation and disease incidence in an informative way, enabling the viewer to formulate new hypotheses about disease transmission in the region. Methods The study was ecological in design, and used census-derived area data, map data, and individual case data as described below. The geographical unit of analysis was the 1991 census ward, and the study population was the entire Eastern region population of the UK (1999 estimate 5.3 million). The Eastern region of England is made up from the counties of Hertfordshire and Essex (both adjoining Greater London to the south), Suffolk and Norfolk along the east coast, Cambridgeshire centrally and Bedfordshire to the north. Much of the area is rural, although there are several medium-sized urban areas in Essex and Hertfordshire, with links to the capital. Away from the London area, the major urban centres include Cambridge and Norwich (both with large university populations), the ports of Ipswich and Harwich, and Peterborough and Luton. Data sources Data on cases came from the Eastern region Communicable Disease Surveillance Centre (CDSC Eastern) regional database of enhanced surveillance of meningococcal disease[ 17 ]. To be included in the surveillance data, a case had to fit the Public Health Laboratory Service (PHLS) case definitions[ 18 ] at a local level, and also had to have been included in the monthly returns sent to the regional level. The study used full postcode and age band information on all cases from 1999 and 2000, a subset of the enhanced surveillance data. Linked information on cases, such as the serogroup or date of notification, was not requested for this study. This also meant that clusters could not be excluded from the analysis. The regional population data were projections for 1999 based on 1991 census data, obtained from the compendium of clinical and health indicators 2001[ 19 ]. This source gave age-banded population data at 'synthetic ward' level (single or aggregated census wards, producing a population of over 5000). Each synthetic ward population was shared equally between all its constituent census wards. The deprivation index used was the Townsend score, which combines local measures of unemployment, car ownership, overcrowding, and housing tenure[ 20 ]. This measure was used as it has already been widely used in similar literature[ 11 , 14 , 15 , 21 ], does not include potential confounders such as percentage of under fives, and was available at census ward 1991 level. A higher Townsend score indicates a more deprived area. Ward level scores based on 1991 census data were obtained from the Manchester Information and Associated Services (MIMAS)[ 22 ]. The shape files used in the MapInfo (© MapInfo Corporation) and EpiMap2000 (in Epi Info™, Centres for Disease Control and Prevention) maps were digitised ward boundaries from the 1991 census, obtained from the EDINA (Edinburgh data and information access) UKBORDERS service[ 23 ]. Vector-based files of census ward boundaries in the counties of Bedfordshire, Cambridgeshire, Essex, Hertfordshire, Norfolk and Suffolk, which make up the UK Eastern region, were downloaded in combined form. Ward population density was calculated using the ward areas contained in the map files, and the population estimates described above. Ethical approval Ethical approval was obtained from the West Hertfordshire health authority local research ethics committee. An application form was submitted, along with a copy of the project protocol, and written approval was returned. Analyses Cases were mapped to census wards using their postcode georeference (NHS postcode database). Microsoft Access was used to assign the deprivation scores, to divide and manipulate ward data, and to link the geographical and attribute files. The age distribution of cases and incidences was calculated using Microsoft Excel, which was also used to create the charts. Statistical testing was performed using StatsDirect ( © StatsDirect Ltd), apart from the relative risk confidence intervals, which were calculated using Microsoft Excel[ 24 ]. Poisson confidence intervals for ward and deprivation group incidence rates were calculated by using the number of cases and the total person-years at risk (twice the ward or deprivation group population). Chi-squared tests for trend were performed for the successive incidence rates across the deprivation groups. The variation in the age distributions of incidence and case counts was compared using the non-parametric Friedman and Mann-Whitney tests respectively. Two Poisson regression models was constructed using StatsDirect, with incidence as the dependent variable and Townsend score and population density, and Townsend score alone, as predictors. The maps were produced with MapInfo, using the "range" function to colour each ward in shades depending on the level of deprivation or incidence of disease. The superimposed maps used a magnified version of the deprivation map in figure 3 , and a stick figure to represent the magnitude of ward disease incidence. Results Case data A total of 773 cases were reported to the CDSC (Eastern) enhanced meningococcal database during 1999 and 2000. Of these, 524 had some postcode details and 499 had the full postcode. 458 cases had a ward assigned, and, of these, 451 had a Townsend score. These data losses were due to postcodes not being recorded (the major factor) or incorrect, discrepancies within the NHS postcode file, and incomplete ward deprivation data (this only relates to the 524 – 451 cases, not the bulk 773-524 cases). Further analysis is therefore restricted to the 451 cases with Townsend score. Where incidence rates are given, they will generally be underestimates of the true incidence (on average 58% of the true value), due to the loss of case data. It was assumed that the data losses were random with respect to the variables of interest. The age structure was well preserved despite the loss of around 40% of the initial cases. Figure 1 compares cases included in the study (451) with the numbers expected if losses were uniform across the age groups. Chi squared testing confirmed that the losses were not related to age group (P = 0.8529). When broken down into the eight Health Authorities supplying case data, the percentage of cases that included postcode information was roughly similar, with around 2/3 of cases being postcoded, with the exception of one authority where only 38.6% of cases were postcoded. However, this authority only contributed 7% of the cases. Figure 1 Case losses: Comparison of numbers of cases included in analysis, by age group, with expected values if losses were equal across all age bands Sources Cases: Confirmed and probable cases of invasive meningococcal disease included in analysis, enhanced surveillance data from CDSC eastern Of the 1184 wards included in the analysis, 325 (27%) had at least one case of meningococcal disease in the two-year period. In these wards, the maximum number of cases was 6, and the median 1. Incidence of meningococcal disease The overall incidence for 1999 and 2000 was 7.4 cases per 100,000 per year. Within the wards containing more than one case, the median incidence was 12.9 per 100,000 per year, and ranged from 3.7 to 60.0. Given the small numbers involved, Poisson confidence intervals for these incidences are wide. In a high incidence ward (41.7 cases per 100,0000 per year), the confidence interval was 11.4 to 106.8 cases per 100,000 per year. In a low incidence ward (4.0 cases per 100,000 per year), the confidence interval was from 0.1 to 22.5 cases per 100,000 per year. The ward incidences will be underestimates due to the loss of cases described above. Figure 1 also shows the age distribution of the cases. This bimodal distribution shows that the peak incidence is in children under 5, with a second peak in the 15–19 group. Figure 2 Age-specific incidence rates: Comparison of cases in each Townsend score group (thirds, 1151 wards) Sources Cases : Confirmed and probable cases of invasive meningococcal disease included in analysis, enhanced surveillance data from CDSC eastern Denominator : 1999 population estimates, compendium of clinical and health indicators 2001, adjusted from synthetic wards (see methods) Map files : 1991 Census digitised boundary data Townsend scores : 1991 Census area statistics Incidence and deprivation Table 1 shows the summary figures for the wards analysed. The mean incidence of meningococcal disease in the most deprived wards is twice that in the least deprived wards (5.9 versus 3.0), with a relative risk of 1.97 (1.55 – 2.51). The intermediate wards have an overall incidence similar to that in the least deprived wards. A chi-squared test for trend shows a significant variation between the rates in each group (chi squared for linear trend = 39.0, p < 0.0001). Table 1 Meningococcal disease in the eastern region, 1999 and 2000: Incidence rates after division into thirds by ward Townsend score Least deprived wards Intermediate wards Most deprived wards Number of wards a 384 383 384 Range of ward Townsend scores -6.26 to -2.19 -2.2 to-0.33 0.34 to 8.44 Number of cases b (all ages) 88 97 267 Total population c (all ages) 1,473,272 1,506,359 2,271,294 Incidence [corrected] d (cases /100,000 /yr) 3.0 [5.1] (2.4 – 3.7) e 3.2 [5.5] (2.6 – 3.9) 5.9 [10.1) (5.2 – 6.6) Incidence relative risk compared to least deprived wards [CI] 1.00 1.07 [0.80–1.43] 1.97 [1.55–2.51] (a) Census wards 1991, divided by Townsend index, from 1991 census (b) All cases of meningococcal disease (confirmed and probable) from enhanced surveillance data, CDSC eastern, included in analysis (c) Population estimates for 1999, Compendium of Health and Clinical indicators 2001, adjusted as in methods (d) Corrected to allow for the loss of 42% of cases. Corrected incidence = uncorrected/0.583. Assumes losses are equal across deprivation groups (e) Poisson confidence intervals – see methods Poisson regression using population density and Townsend score as predictors revealed population density to be a non-significant contributor to the variation in incidence (p = 0.086). A second model, setting incidence against Townsend score alone, suggested that ward incidence rises by 12% (9 – 16%) for every unit increase in deprivation score. Age distribution and deprivation The age-specific incidence rates are shown in figure 2 . The most striking feature is the large excess of cases in the under ones and one to fours in the most deprived wards. The incidence is 1.9 times higher for the most deprived under fives (under one and one to four groups combined). The increased incidence in the under five age group accounts for 68% of the difference in overall incidence between the most and least deprived wards. The incidences in the 16 age groups used varied significantly between the three deprivation groups (Friedman test, p < 0.0001). Figure 3 Ward deprivation in the Eastern region, by Townsend score* Sources Map files : 1991 Census digitised boundary data Townsend scores: 1991 Census area statistics Software : MapInfo © Professional The Poisson confidence intervals for these incidences are narrower than for individual wards. Two were calculated: 21.3 to 35.1 for an incidence of 27.6 per 100,000 per year, and 0.3 to 4.9 for a lower incidence of 1.7 per 100,000 per year. Mapping deprivation and meningococcal disease Figure 3 shows ward deprivation by Townsend score using MapInfo. The wards are divided into eight bands by their ward score. Many of the most deprived wards correspond to urban areas, though there was a broad area of greater deprivation in the north of the region. Figure 4 shows a map of meningococcal incidence by ward in the region, with colours coded by six incidence ranges. Areas of high incidence often coincide with urban regions, though some rural areas also have high rates of disease. These include some of the deprived rural areas, including the area in north Norfolk noted above. Figure 4 Eastern Region, 1999 and 2000: Incidence of meningococcal disease by census ward 1991 Sources Cases : All cases of meningococcal disease (confirmed and probable) collected by CDSC Eastern for enhanced surveillance & included in analysis, 1999 and 2000 Denominator : 1999 population estimates from Compendium of Health and Clinical Indicators 2000, adjusted for true ward (see methods) Map files : 1991 Census digitised boundary data Software : MapInfo © Professional Figure 5 shows incidence rates superimposed on the deprivation map from figure 3 , and magnified to show local detail in the Hertfordshire/Essex region. It shows the relationship of incidence to deprivation, with high incidence wards being clustered within and around the 'foci' of deprivation. This is particularly marked in Harlow, represented by the cluster in the lower central part of the map. Figure 5 Meningococcal incidence and deprivation superimposed, Hertfordshire and west Essex *Sources Map files : 1991 Census digitised boundary data Townsend scores : 1991 Census area statistics Software : MapInfo © professional Cases : confirmed and probable cases of invasive meningococcal disease included in analysis, enhanced surveillance data from CDSC eastern Denominator : 1999 population estimates, compendium of clinical and health indicators 2001, adjusted for true ward (see methods) Discussion The results for the regional surveillance data support the theory that meningococcal disease is associated with socioeconomic deprivation. Compared to the Welsh study[ 16 ] this study included more cases (451 vs. 295) over a wider area, with a total population of 5.3 million. Also, both deprivation and incidence of meningococcal disease are mapped using GIS, showing the spatial relationship of disease foci to areas of deprivation, both urban and rural. The study, in common with that from southwest England[ 15 ], also shows that this relationship holds in an area containing many rural wards, despite the problems associated with deprivation indices in rural areas[ 25 , 26 ]. The analyses of the age distribution of cases and incidences suggest that there is variation between the deprivation groups. There was a significant difference between the age-specific incidence rates between the groups, and a median age difference of five years between the most and least deprived groups (non-significant). The relative risk of disease in the most deprived wards compared to the least deprived (1.97, CI 1.55 – 2.51), is greater than that observed in the northeast Thames study[ 14 ] (odds ratio 1.51 for N. meningitidis), and the southwest England study[ 15 ] (relative risk 1.76, between upper and lower quartiles), but less than that found in the Welsh study[ 16 ] (relative risk 2.4 between upper and lower quintiles). The incidence of other infections, such as infectious intestinal disease, has also been shown to vary with area socioeconomic conditions (relative risk 2.41, quintiles) [ 27 ]. All but two (25–29 years, 35–39 years) age groups experienced higher age-specific rates in the most deprived group, but the highest case numbers were seen in the under 5 age group. The increased incidence in the under fives in the most deprived group, compared to the least deprived, accounted for 68% of the total difference in age-specific incidence. This suggests that children under five are more vulnerable to meningococcal disease in the most deprived areas. The continuation of a similar pattern when case counts, percentage and age-specific incidence were considered improves the robustness of this conclusion. Several environmental factors might contribute to the increased risk of childhood meningococcal disease in more deprived areas. Childcare arrangements, necessitated by either personal or environmental circumstances, might expose a child in a more deprived area to more potential carriers. Smoking is also known to be more common in people from disadvantaged backgrounds and a number of studies have identified passive smoking as risk factors for both nasopharyngeal carriage and meningococcal disease [ 3 , 28 - 30 ]. An area with a relatively greater number of under fives might have a higher overall incidence rate, simply because there are more susceptible individuals. Analysis of the percentage age distribution of the populations in each deprivation group does not reveal any marked variations. There is a slight excess of 20–29 year olds in the most deprived wards compared to the other wards, but not an excess of under fives. Mapping of the incidence rates and deprivation indices by ward definitely added value to the routine surveillance data, emphasising the focal nature of disease, and the relationship of these foci to areas of higher deprivation. Mapping incidence at ward level, rather than pinpointing individual cases as in the Welsh study[ 16 ], both considers the population at risk, and avoids potential problems with case confidentiality. The maps also suggest that, in the Eastern region, invasive meningococcal disease is largely an urban problem. Although the higher population density of urban areas might explain this, the regression model suggested that the variation was better explained by ward deprivation (term for population density non-significant, p = 0.0866). Some of the highest rates were seen in deprived urban parts of Essex. Several rural wards containing cases corresponded to more deprived areas, particularly in North Norfolk. Figure 5 shows the relationship of incidence to deprivation in close-up, and also that disease is not confined to the more deprived areas. This is an ecological study, so the associations shown may not be valid at an individual level. Population density might still be a factor, as the regression is subject to some autocorrelation, the population variable occurring in the incidence (dependent) and population density (predictor) terms. A separate analysis, of urban and rural areas, might show whether this is the case. The paucity of cases in rural areas might make this difficult unless several more years' data were included. The north east Thames study[ 14 ] did, however, show the same relationship in an area of high population density, as have other urban studies of pertussis[ 31 ] and sexually transmitted infections[ 12 , 21 ]. Geographical bias in postcoded data might have contributed toward the results, as the health authority with the lowest rate of postcode inclusion, and therefore the greatest loss of data, includes many of the more affluent parts of the region. The cases supplied by this authority only accounted for 7.4% of all cases and 4.4% of postcoded cases, so this potential bias is unlikely to have had a major effect. Routine childhood immunisation with meningitis C conjugate vaccine started in November 1999[ 2 ], so should have had an effect on the year 2000 cases. The subset of enhanced surveillance data used for this study did not include the date of notification or the serogroup involved, so no comment can be made on the influence of the vaccine or of serogroup on the results in this paper. A breakdown of the serogroups was available for the years as a whole; of those where the group was known, 60% of cases were group B and 33% group C. Patterns of transmission may change as group C infections decline, and this might be seen with routine mapping of case data. Conclusions Mapping of deprivation indices and meningococcal cases is a useful tool in the analysis of routine surveillance data. Mapping of incidence rates revealed an association between areas of high incidence and areas of higher deprivation by Townsend score. High incidence and deprivation often coincided in urban areas. Mapping of deprivation indices also reveals areas of rural deprivation, such as the coastal band in north Norfolk. Yearly mapping of routine surveillance data can help to target control strategies for meningococcal disease locally. Analytic studies would be helpful in elucidating the mechanisms by which socioeconomic conditions influence the risk of meningococcal disease in the region. Along with the study described here, knowledge gained from such investigations could inform the work on health inequalities, and try to reduce such inequalities through health promotion and community infection control. Competing interests None declared. Authors' contributions CJW obtained and analysed the case and geographical data, performed the statistical analyses and mapping, and wrote the text of the paper. PH suggested the idea for the study and helped develop the methods by which the question was addressed. LW provided the data from the enhanced meningococcal surveillance database and advised on data quality and sources. IL provided expertise in geographical information systems and postcode geography. LW, PH and IL all advised on the design of the study, and the analysis and interpretation of the results. All authors contributed to drafting the paper and have read and approved the final draft. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC503390.xml
521694	Tissue Doppler echocardiography and biventricular pacing in heart failure: Patient selection, procedural guidance, follow-up, quantification of success	Asynchronous myocardial contraction in heart failure is associated with poor prognosis. Resynchronization can be achieved by biventricular pacing (BVP), which leads to clinical improvement and reverse remodeling. However, there is a substantial subset of patients with wide QRS complexes in the electrocardiogram that does not improve despite BVP. QRS width does not predict benefit of BVP and only correlates weakly with echocardiographically determined myocardial asynchrony. Determination of asynchrony by Tissue Doppler echocardiography seems to be the best predictor for improvement after BVP, although no consensus on the optimal method to assess asynchrony has been achieved yet. Our own preliminary results show the usefulness of Tissue Doppler Imaging and Tissue Synchronization Imaging to document acute and sustained improvement after BVP. To date, all studies evaluating Tissue Doppler in BVP were performed retrospectively and no prospective studies with patient selection for BVP according to echocardiographic criteria of asynchrony were published yet. We believe that these new echocardiographic tools will help to prospectively select patients for BVP, help to guide implantation and to optimize device programming.	Background Heart failure is among the most common chronic diseases in modern civilizations. The dilatation of the left ventricle frequently induces intracardiac conduction delays resulting in asynchronous left ventricular motion. This manifests as left bundle branch block in the surface ECG. Both QRS width and intraventricular asynchrony are predictors of hospitalization and severe cardiac events in patients with heart failure [ 1 - 3 ]. The mechanisms of myocardial asynchrony include a delayed left ventricular regional contraction and relaxation. The right ventricle contracts during left ventricular end-diastole, leading to a "bulging" of the septum into the left ventricle. The intra(left)ventricular delay of the systolic velocity induces the "delayed longitudinal contraction (DLC)". Furthermore, the delay of the contraction of the papillary muscles aggravates mitral regurgitation. This, in summary, leads to an increased oxygen demand of the myocardium [ 4 ]. Resynchronization of the intraventricular conduction can be achieved by introducing an additional lead through the coronary venous sinus to stimulate the left ventricle (biventricular pacing, BVP). The combination of BVP and a cardioverter-defibrillator (ICD) combines the clinical improvement by BVP and reduction in mortality [ 5 ]. Recent studies have shown an acute and sustained hemodynamic improvement, reversal of LV-remodeling, an increased quality of life, a reduction of symptoms of heart failure, and improvement of exercise tolerance after biventricular pacing. Markers of reverse remodeling were reduction of left ventricular volumes, increase in LVEF without an increase in oxygen consumption, reduction of mitral regurgitation [ 6 - 10 ]. However, a significant reduction of mortality after BVP alone could not be demonstrated. In the current guidelines, LBBB in the surface ECG and a reduced LVEF are the main indications for BVP [ 11 ]. However, about one third of patients in the large multicenter BVP studies did not improve – despite BVP [ 6 , 7 , 12 ]. There is increasing evidence, that there is only a weak correlation of electrical (QRS width) and mechanical asynchrony and the benefit of BVP. It seems that not all heart failure patients with LBBB have mechanical asynchrony [ 12 , 13 ]. Furthermore, asynchrony is common even in heart failure patients with narrow QRS complexes compared with healthy controls. A prospective study assessed left ventricular systolic and diastolic asynchrony in 67 patients with heart failure (LVEF < 50%) with normal QRS width and 45 patients with CHF and wide QRS complexes (>120 ms). 88 healthy control patients were included. Systolic (diastolic) asynchrony occurs in 51% (46%) of the heart failure patients with narrow QRS complexes and in 73% (69%) in the patients with wide QRS complexes. Systolic asynchrony was defined as the max difference in time-to-peak myocardial contraction of 12 myocardial segments. Diastolic asynchrony was defined as maximum difference of time-to-peak early diastolic relaxation. In summary, the authors state that asynchrony is common in patients with heart failure even without a wide QRS complex [ 14 ]. This is confirmed in a study with 158 heart failure patients (LVEF < 35%), that were divided in three subgroups: Group 1 with no (QRS < 120 ms), group 2 with mild (120–150 ms) and group 3 with severe LBBB (>150 ms). Interventricular asynchrony was defined by TDI as IVMD >40 ms and the intraventricular delay as maximum pre-ejection period of >50 ms in one or more myocardial segments. Asynchrony was seen in all three subgroups, however, there was no correlation between interventricular and intraventricular asynchrony [ 15 ]. A recent study demonstrated that successful BVP can be achieved in patients with a normal QRS duration and asynchrony [ 16 , 17 ]. These controversial data indicate the need for a more careful patient selection for BVP. Newer echocardiographic techniques, such as Tissue Doppler Imaging and Tissue Synchronization Imaging could potentially improve patient selection and guidance of implantation and programming of the devices for BVP. The risks of pacemaker implantation and expenses in non-responders to BVP could be avoided. Furthermore, the cost-effectiveness of BVP would be augmented. Definitions of asynchrony Regarding the nomenclature, the term "asynchrony" is used synonymously to "dyssynchrony" in this article. There is a variety of methods to determine asynchrony. In table 1 , the different approaches to asynchrony are listed concisely. The QRS width (LBBB > 120, 130, 150 ms) is the simplest method, but the sensitivity to predict benefit from BVP is rather low [ 18 , 19 ]. Magnetic resonance imaging can also detect areas of asynchrony but this technique can not be repeated for follow-up after device implantation. Table 1 Concise summary of the different approaches to echocardiographic measurement of asynchrony Assessment of asynchrony with: Ref. Criteria Segments Limitations Analysis time Prediction of benefit I. Global ventricular asynchrony ECG 4, 44 QRS width >120 ms Global assessment LBBB after myocardial infarction Short Low (30% non-responder) M-mode 21 Septal-to-posterior wall motion delay >130 ms septal and posterior scar tissue, only septal or posterior Short low pw-TDI 25 Cumulative asynchrony (EMD) >102 ms Intra LV (5 basal segments) and interventricular (vs. RV lateral segment) Low spatial resolution Long Good prediction of acute response (AUC in ROC 0,84) II. Interventricular asynchrony pw-Doppler echocardiography 47 Interventricular mechanical delay (IVMD) >40 ms Aortic and pulmonary outflow tract Not simultaneous Short No III. Intraventricular asynchrony 2D-TDI 27 Ts-SD: intraventricular systolic asynchrony index: >33 ms 12 segments complex (post-processing) Long Acute response (3 months) 4, 32 Difference in septal-lateral time-to-peak TDI, cut-off >60 ms 12 segments Complex Long EF increase after BVP 40 Mean regional myocardial performance index: Difference between regional Q-wave-to.peak systolic displacement times 12 segments 4 segments Complex Long, offline Acute response 33 Ts-SD: cut-off: 31,4 ms 12 segments Complex Long 3 months response, reverse remodeling Strain and strain rate 15, 34, 33 Myocardial deformation in systole, presence of post-systolic shortening 12 segments Complex, time consuming, in dilated ventricles low spatial resolution) Long Controversial data Tissue Tracking 28 DLC in >2 basal segments 12 basal segments in apical four chamber view. Requires correct timing of LV events Short Acute response TSI 36 Color-coded time-to-peak tissue Doppler velocities (cut-off >65 ms in anteroseptum and posterior wall in apical long axis view) 16 segments except apex Only velocity data Short Acute response (Sensitivity 87% Specificity 100%) 3-D-echo 26 No quantitative criteria defined All segments Reduced spatial resolution Time consuming, off-line analysis No systematic data Automated endocardial border detection (ABD) 26 Septal-lateral phase angle difference 100 segments. apical-four-chamber view (septal-lateral) High complexity, single imaging plane Long, only off-line Acute response ABD + Contrast 46 Echo-contrast cardiac variability imaging: displacement maps apical four chamber High complexity, single imaging plane Long Acute response Echocardiographic tools include 2D, Doppler and Tissue Doppler methods. Up to date, there is no consensus on the definition of echocardiographically measured myocardial asynchrony. The determination of asynchrony by M-mode echocardiography is limited to septal and inferior segments in parasternal long-axis and is not performed routinely in current studies [ 20 , 21 ]. Earlier echocardiographic approaches to asynchrony included the delayed long axis shortening that was found to suppress early diastolic transmitral flow and subsequently leading to decreased leftventricular function [ 22 ]. Tissue Doppler imaging (TDI) measures regional wall motion velocities. TDI can accurately quantify regional left ventricular function [ 23 ]. Pulsed wave TDI does not allow simultaneous comparison of regional timing in different segments within one cardiac cycle. Color-coded TDI reduces beat-to-beat variability and examination time. Color coded TDI has a very high time resolution of 10 ms. TDI technology includes tissue tracking and strain rate imaging. Tissue tracking allows the measurement and visualization of longitudinal motion in each myocardial segment during the different phases of the heart cycle. Strain measures compression and distension of myocardial segments ("deformation imaging") and strain rate imaging expresses strain changes per time interval. Post-systolic movement diagnosed with velocity or tissue tracking can be differentiated into passive or active motion (=contraction, then defined as PSS). But in ischemic cardiomyopathy PSS was not an useful criterion for response to BVP because this phenomenon is not only a sign of asynchrony but also a marker for ischemia and/or viability of severe hypo/akinetic segments [ 24 ]. Tissue synchronization imaging (TSI) is a newer technique that utilizes color-coded time-to-peak tissue Doppler velocities and visualizes segments of dyssynchrony in real-time by superimposing these temporal motion data on 2D echo images. TSI analysis is possible in all myocardial regions except the apex. The color-coding is green (normal time-to-peak velocity: 20–150 ms), yellow (150–300 ms) and red (300–500 ms) [ 25 ]. Online 3D echocardiography and automated border detection (ABD) might be future diagnostic tools to diagnose asynchrony but need evaluation in larger studies [ 26 ]. Myocardial asynchrony includes inter- and intraventricular asynchrony. Interventricular asynchrony can be assessed by comparing pw-Doppler signals in the right and left ventricular outflow tracts. A delay of >60 ms is considered to demonstrate interventricular asynchrony. These measurements in the outflow tracts cannot be performed simultaneously and, therefore, there is a high inter-measurement variability and dependence on cardiac workload. In addition, interventricular asynchrony can measured as the difference of the electromechanical delays in the basal LV segments and the lateral RV segments [ 35 ]. Intra(left)ventricular asynchrony is considered to be the most important aspect of the electromechanical delay (EMD). It can be measured by a variety of methods. EMD is defined as the delay between the onset of the QRS complex on the surface ECG and the onset of the systolic TDI wave in corresponding myocardial segments. Recently, the systolic synchronicity index has been introduced [ 27 ]. It is defined as the standard deviation (SD) of the EMD in 12 LV segments (6 basal, 6 mid-segmental model). Intraventricular asynchrony can also be demonstrated by tissue tracking with diastolic color-coded areas called DLC. This is the amount of post-systolic contraction after the closure of the aortic valve (i.e. post systolic shortening = PSS) which was confirmed by strain and strain rate in this study [ 28 ]. Intra-left ventricular asynchrony is not only of diagnostic value for selecting patients for BVP, but has prognostic value as well. Bader et al. [ 3 ] examined inter- and intraventricular asynchrony as an independent predictor of heart failure worsening: 104 patients with chronic stable heart failure without previous myocardial infarction (LVEF < 45%) were included, follow-up echocardiography was performed after one year. Study endpoint of heart failure worsening was hospitalization for cardiac decompensation. Intra-left ventricular asynchrony is an independent predictor of severe cardiac events. Only a weak correlation of intra/inter-ventricular EMD and QRS width could be demonstrated. In figures 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , the different approaches to assess asynchrony as well as echocardiographic examples of successful biventricular pacing are illustrated. Figure 1 Measurement of interventricular mechanical (IMD) delay by PW Doppler : A) PW Doppler in aortic outflow tract: Measurement from onset of QRS to the onset of PW curve in the aortic outflow tract. This time is also called the aortic pre-ejection time and is a marker for intra(left)ventricular asynchrony. B) PW Doppler in pulmonary outflow tract: Measurement from onset of QRS to the onset of PW curve in the pulmonary outflow tract. The IMD is the difference between the time of a) and b). Figure 2 Assessment of asynchrony in parasternal long axis view by M-mode : Time difference between peak of septal and inferior myocardial contraction. Figure 3 Tissue Doppler velocity data for the quantification of asynchrony from apical four chamber view. Sample volumes are in the basal lateral and basal septal segment. A) Normal control patient. There is a synchronous myocardial velocity in the septal (=yellow) and the lateral (=green curve) segment. IVC = isovolumetric contraction, IVR = isovolumetric relaxation, S = peak systolic velocity; E = early diastolic filling, A = late (atrial) diastolic filling. B) There is asynchronous myocardial velocity in the septal (=yellow) and the lateral (=green curve) segment. Figure 4 Assessment of asynchrony by strain from the apical four chamber view. The sample volumes are in the basal septal and the basal lateral segments. A) Normal strain curve in a control patient. ICT = isovolumetric contraction time. B) Strain curve with asynchronous myocardial velocity in the septal (=yellow) and the lateral (=green curve) segment. Figure 5 Tissue Tracking allows the visualization of longitudinal motion in each myocardial segment. Images are from the apical four chamber view. A) Normal control patient. There are normal colour-coded displacement values in the lateral and septal segments, with physiologically higher values in the more basal segments and lower values towards the apex. B) Tissue Tracking in a patient with dilated cardiomyopathy. There is a dilated left ventricle with "baseball shape" and reduced displacement values and no basal-apical gradient (max displacement = 8 mm) in the septal segments and DLC in the lateral wall (no colour-coding) indicating asynchrony of the lateral wall. Figure 6 Tissue Synchronization Imaging displays colour-coded time-to-peak tissue Doppler velocities. The colour-coding is green (normal time-to-peak velocity: 20–150 ms), yellow (150–300 ms) and red (300–500 ms) Apical four chamber view. A) TSI in a control patient (only green colour coding indicating synchronous contraction) B) TSI in a patient with LBBB: The basal and mid-septal segments show a delayed time-to-peak velocity (red colour). Figure 7 Delayed longitudinal contraction (=DLC) as a marker for asynchrony can be visualized by analysis of systolic and diastolic Tissue Tracking . Systolic tracking analyzes the systolic displacement i.e. tracking interval between the onset of QRS-complex and the end of the T-wave. Diastolic tracking can demonstrate DLC with colour coding (end of T until begin of R). Images from apical two chamber view A) Systolic Tracking: The inferior segments (=grey area) show DLC with no systolic motion B) Diastolic Tracking: The inferior segments (=colour coded area) show DLC with diastolic movement. Figure 8 Demonstration of successful BVP by Tissue Tracking in apical four chamber view in a patient with dilated cardiomyopathy. Images from apical four chamber view. A) Before BVP, there is a dilated ventricle ("baseball shape") with reduced systolic displacement (max displacement = 8 mm) in the septum and DLC in the lateral wall (no colour-coding) indicating asynchrony of the lateral wall. B) After three months of BVP, there is a reduction of left ventricular dilatation (reverse remodelling, "American football shape" of the left ventricle), increased tracking values and no DLC regions anymore. Figure 9 Successful BVP documented by Tissue Tracking in apical two chamber view. A) Before BVP, there is a dilated ventricle with reduced systolic displacement (max displacement = 8 mm) in the septum and DLC in the inferior wall (no colour-coding) B) After three months of BVP, there is a reduction of left ventricular dilatation (reverse remodelling), increased tracking values, a basal-apical gradient and no DLC regions anymore. Patient selection for BVP Only limited data are published concerning prospective echocardiography based patient selection for BVP. Bordachar et al. [ 29 ] performed a prospective study to identify TDI parameters that would predict the benefit of upgrading right ventricular pacing to BVP. 26 patients with normal LVEF and RVP and 16 patients with CHF and RVP were included. EMD was defined as the interval between the stimulation spike and the onset of the S wave. An intra-ventricular EMD of >50 ms identifies patients with significant asynchrony. No correlation between asynchrony and QRS width was seen in the heart failure patients. ECG criteria would have misclassified 44% of the patients for mechanical ventricular asynchrony. This study has defined relevant asynchrony but did not assess the hemodynamic or electromechanical effects after upgrade to BVP nor effects on morbidity and mortality. Retrospective analysis after BVP Several studies were performed to retrospectively correlate markers of asynchrony to benefit from BVP. Lafitte [ 30 ] has included 15 patients with idiopathic DCM and a QRS of more than 140 ms (NYHA III-IV, LVEF < 35%, LVEDD > 60 mm) for BVP. Measurement of EMD was performed at baseline and after one month. This study has found that BVP reduces EMD in the lateral left ventricular wall. In another study [ 25 ], 49 patients with heart failure (QRS > 130 ms, LVEF < 35%, NYHA II-IV) were included. Retrospectively, intra- and interventricular and the combined index of asynchrony (=the sum of left and right ventricular asynchrony) were assessed at baseline and after 6 months of BVP by pulsed wave TDI. The cut-off-values for LV-asynchrony was 60 ms (56 ms for RV-LV-asynchrony and 102 ms for the "sum-asynchrony"). By definition, patients with a relative increase in LVEF of more than 25% were classified as responders to BVP. Receiver-operating characteristics (ROC) analysis showed that the degree of echocardiographic asynchrony is superior to QRS width in predicting hemodynamic and clinical improvement after BVP compared to QRS duration or conventional echo data. In 82% of the patients, the benefit of BVP could have been predicted echocardiographically. The role of TDI and 3D echo on the long term (1 year) outcome after BVP was evaluated in 25 patients [ 19 ]. The extent of DLC in the basal segments at baseline predicted the long-term efficacy of BVP. The LV base DLC was reduced from 18,7% to 8.1% after BVP. In concordance with other studies, the QRS duration failed to predict BVP efficacy [ 28 ]. The myocardial segments with the best resynchronization after BVP were studied in 18 patients with an LVEF <35% and a QRS width of >120 ms (NYHA III-IV). Color tissue Doppler velocity imaging was performed from the apical four chamber view at baseline and one month of follow-up after BVP [ 31 ]. Peak velocities and regional time differences in basal and mid septal segments were compared to the corresponding lateral segments. At baseline, a regional asynchrony of 42 ms in the basal sites (only 14 ms in the mid left ventricular site) was measured. After one month of BVP, a reduction of asynchrony was seen in only the basal segments but not in the mid segments. In conclusion, it was suggested that hemodynamic improvement is mainly in basal sites. Reverse remodeling and improved synchrony after 3 months of BVP was evaluated in 25 patients [ 32 ]. Asynchrony was assessed as time-to-peak regional sustained systolic contraction (=Ts). After three months, a homogenous left ventricular delay of Ts, improved interventricular synchrony and a reduced isovolumic contraction time and increased diastolic filling time were documented. These beneficial effects were reversible after withholding BVP. In a univariate analysis, systolic dyssynchrony was the only independent predictor of reverse remodeling after three months [ 33 ]. One recent study has compared the value of TDI and SRI and post-systolic shortening in the prediction of reverse remodeling after BVP: The previously introduced asynchrony index (=Ts-SD) based on Tissue Doppler velocity data has the highest predictive value of reverse remodeling after BVP. PSS has predictive power only in non-ischemic heart failure. In ischemic heart failure, PSS seems not to be a marker for reverse remodeling but rather reflects viability and is therefore not altered by BVP. SRI imaging techniques did not predict reverse remodeling after three months of BVP [ 33 ]. This is in contrast to previously published data [ 34 ]. Kanzaki has introduced the synchrony index, which is defined as the correlation coefficient of linear regression of velocity of septal and lateral mitral annular region. This index showed an increase after 6 months of BVP paralleled by increased LV contractility [ 35 ]. One study [ 36 ] has retrospectively evaluated the use of TSI to predict the acute response to BVP in 29 patients. The acute benefit to BVP was defined as a >15% increase in echocardiographically measured stroke volume 48 h after device implantation. A difference of >65 ms in time-to-peak velocity in anteroseptal and posterior segments in the apical long axis view was associated with acute improvement after BVP. However, the ability of TSI to predict long-term improvement after BVP needs further evaluation. Guidance for implantation TDI could play a role in identifying patients during catheterization procedures that will profit from BVP. Catheterization studies have shown that the beneficial effects of BVP begin almost immediately [ 37 , 38 ]. But systematic evaluation with TDI-technique is currently ongoing. Furthermore, TDI can assist in finding the optimal pacing site for the coronary sinus lead. In 31 patients, it was documented that LV-stimulation on the site of longest EMD had the best benefit of BVP. The regional asynchrony was assessed by pw-TDI and the pacing site was determined fluoroscopically [ 39 ]. Lateral and postero-lateral LV lead positions were retrospectively found to improve left ventricular hemodynamics [ 40 ]. Optimal programming of biventricular device after implantation AV-time programming An AV time is considered to be optimal when the end of the A wave coincides with the complete closure of the mitral valve [ 41 ]. An optimal AV time setting of the pacemaker can improve systolic function [ 42 ]. However, there is only limited published data assessing the optimal AV time in patients with BVP. Optimization of the interventricular delay The optimal delay between the right ventricular and the coronary sinus stimulation is yet unknown. One study compared simultaneous versus sequential BVP in 29 patients. The optimum interventricular delay was found by maximum reduction of DLC as measured by Tissue Doppler and Tissue Tracking. An optimum sequential BVP could significantly reduce the extent of DLC compared to simultaneous pacing [ 43 ]. Patients with atrial fibrillation About one third of patients with heart failure have atrial fibrillation. The large trials, however, have only included patients in sinus rhythm. Only small studies with controversial results were performed in patients with atrial fibrillation and LBBB. Leclercq has performed one study in 59 NYHA III patients with chronic atrial fibrillation, a slow ventricular rate and the need for permanent pacing (VVI-paced QRS width of >200 ms). Due to a high drop out rate, the results did not show a significant increase in 6-min-walk distance after BVP [ 44 ]. Larger trials are needed to evaluate BVP for patients with atrial fibrillation. Preliminary own results We have performed a double-blind cross-over study in our clinic to assess the use of new echocardiographic techniques in BVP. Patients (n = 40) with a QRS >140 ms and a LVEF <35% received an InSyncICD 7272 (Medtronic, Minneapolis, Minnesota, USA). Preliminary results (n = 8) after two years demonstrate a reduction of the septal-posterior delay from 264 (±23) msec to 234 (±34) msec (p < 0,05) and a stabilization of clinical (NYHA class improvement) and hemodynamic status (EF and LV volumes). The study is ongoing. The following video loops underline the utility of TSI and Tissue Tracking to document improvement of synchronicity after BVP. In additional file 1 shows asynchrony before BVP implantation in apical four chamber view by TSI. In additional file 2 the effect of BVP is shown. in additional file 4 shows the acute changes of BVP as documented in this video loop by Tissue Tracking from apical four chamber view compared to baseline ( additional file 3 ). The long-term effect of BVP after six months is illustrated in additional file 5 (baseline) and additional file 6 (after 6 months). Conclusion and future perspective Many controlled and uncontrolled studies have demonstrated that new echocardiographic tools to determine myocardial asynchrony in heart failure patients will help to select patients for BVP help guidance of implantation and optimize device programming. To date, all studies employing tissue Doppler date were performed retrospectively. No prospective studies that have selected patients for BVP according to echocardiographic evaluation of asynchrony were performed yet. The ongoing CARE-HF study incorporates echocardiographic criteria of asynchrony in a subset of patients with a QRS of 120–150 ms [ 45 ]; results are not expected until 2005. The criteria of asynchrony in this study are (1) aortic pre-ejection delay >140 ms, (2) the mechanical interventricular (pw aortic valve vs. pulmonary valve) delay >40 ms and (3) the demonstration of left ventricular post-systolic contraction by M-mode and/or Tissue Doppler. Unresolved issues include different opinions regarding the various elements of asynchrony and their contribution to the pathophysiology and progression of heart failure. There is a lack of consensus about the best asynchrony marker for patient selection. There is evidence that ischemic and dilated cardiomyopathy might have different selection parameters for BVP. The practical consequences for patient selection and/or implantation site of the lead are currently under investigation. There are only limited echocardiographic data regarding the programming of the optimal interventricular (V-V) delay. There are no data concerning the long-term effect (i.e. years) of BVP on hemodynamics, amelioration of mitral regurgitation, reverse remodeling and mortality. Another area of uncertainty is the selection of patients for BVP without electrical (QRS < 120 ms) but with mechanical asynchrony. Abbreviations BVP Biventricular Pacing DCM Dilated Cardiomyopathy DLC Delayed longitudinal Contraction EMD Electromechanical Delay IVMD Interventricular Mechanical Delay LBBB Left Bundle Branch Block PSS Post-Systolic Shortening SRI Strain Rate Imaging TDI Tissue Doppler Imaging Ts Time-to-peak myocardial contraction TSI Tissue Synchronization Imaging Ts-SD Standard deviation of time-to-peak myocardial contraction Authors contributions F Knebel and AC Borges have performed the literature review and have prepared the manuscript. RK Reibis, have performed echocardiographic examinations for this article. HJ Bondke and J Witte and G Baumann have selected patients for BVP. HJ Bondke and J Witte have implanted the biventricular pacing devices. All authors have read and approved the final version of the manuscript. Supplementary Material Additional File 1 TSI in a patient with LBBB before BVP: The lateral segments show a delayed time-to-peak velocity (red colour). Apical four chamber view. Click here for file Additional File 2 TSI post-implantation: There is only green colour coding indicating synchronous contraction of all segments from apical four chamber view. Click here for file Additional File 3 Tissue Tracking without BVP : There are reduced displacement values and no basal-apical gradient in the septal segments and DLC in the lateral wall (no colour-coding) indicating asynchrony of the lateral wall. Apical four chamber view. Click here for file Additional File 4 Acute effect with BVP "on" (Tissue Tracking): There are increased displacement values, a basal-apical gradient. Apical four chamber view. Click here for file Additional File 5 Tissue Tracking before BVP : There are reduced displacement values and no basal-apical gradient in the septal segments and DLC in the lateral wall. Apical four chamber view. Click here for file Additional File 6 Long-term effect after 6 months of BVP: Reduction of left ventricular dilatation, increased displacement values, a basal-apical gradient. Apical four chamber view. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521694.xml
518968	B-Raf specific antibody responses in melanoma patients	Background Mutations of the BRAF gene are the most common genetic alteration in melanoma. Moreover, BRAF mutations are already present in benign nevi. Being overexpressed and mutated, B-Raf is a potential target for the immune system and as this mutation seems to be an early event, a humoral immune response against this antigen might serve as a diagnostic tool for detection of high risk patients. Methods 372 sera of 148 stage IV melanoma patients and 119 sera of non-melanoma patients were screened for B-Raf, B-Raf V599E and C-Raf specific antibodies by an ELISA assay. Sera were screened for specific total Ig and for IgG. Serum titers were compared with a two tailed Mann-Whitney U test. Sera with titers of 1:300 or higher were termed positive and groups were compared with a two tailed Fisher's exact test. Results B-Raf specific antibodies recognizing both B-Raf and B-Raf V599E were detected in 8.9% of the sera of melanoma patients and in 2,5% of the control group. Raf specific IgG was detected in some patients at very low levels. B-Raf specific antibody responses did not correlate with clinical parameters but in some cases, B-Raf antibodies emerged during disease progression. Conclusion These findings imply that B-Raf is immunogenic in melanoma patients and that it might serve as a potential target for immunotherapy. However, B-Raf specific antibodies emerge at rather late stages of melanoma progression and are present only with a low frequency indicating that spontaneous B-Raf specific antibodies are not an early marker for melanoma, but rather may serve as a therapeutic target.	Background Cutaneous malignant melanoma is responsible for 1% of all malignant tumors with a rising incidence in the Caucasian population [ 1 ]. Initial diagnosis is based on asymmetry, border regularity, multiple colours, diameter as well as elevation of the pigmented lesion. However, it is sometimes difficult to differentiate between irregular dysplastic nevi and a melanoma without histological analysis. Hitherto risk groups for the development of melanoma are characterized by fair skin, multiple and/or dysplastic nevi and the history of sunburns in childhood [ 2 ]. Invasive melanomas have a rapid tendency to metastasize. In these stages of disease, therapy is very difficult and the 5-year survival rate of stage IV patients is below 20% [ 3 ]. On the molecular level, melanoma is associated with several genetic changes, including mutations or transcriptional variations in tumor suppressors like p53, CDKN2A/p16, CDKN1A/p21 or in oncogenes like N-Ras [ 4 ]. Recently, it has been discovered by Davies et al. that 66% of melanoma have a mutated BRAF gene which results in higher kinase activity due to a single amino acid exchange (B-Raf V599E) occurring in almost 90% of the mutations [ 5 ]. Most interestingly, this mutation is somatic [ 6 , 7 ] and some authors describe the presence of the mutation already in benign nevi [ 8 ], whereas others fail to reproduce the high frequencies in early stages and speculate of mutated B-Raf being relevant for progression rather than initiation of melanoma [ 9 ]. Nevertheless, the high incidence of the mutations in melanoma qualify the B-Raf protein as a potential target for tumor therapy and preliminary results of phase II clinical trials with Raf kinase inhibitors suggest protective activity [ 10 ]. Promising results with new approaches for melanoma therapy have been obtained with active immunizations against tumor associated antigens (TAA) like MAGE-3 [ 11 ], MART-1 [ 12 - 14 ] tyrosinase [ 15 ] or survivin [ 16 ]. Regarding the high incidence of B-Raf mutations and the increased expression level in tumors, B-Raf should be an attractive target for immunotherapy [ 17 ] and most recently, independent findings demonstrated mutation specific CD4 + T-cell responses in melanoma patients [ 18 ]. Moreover, we have recently demonstrated CD8 + T-cells in melanoma patients reactive against an HLA B27 restricted B-Raf V599E epitope encompassing the mutation identified using computer assisted algorithms [ 19 ]. However, assessing CD4 + or CD8 + T-cell responses is not suitable for screening larger numbers of patients to estimate responder frequencies. To determine the frequency of B-Raf specific responses in melanoma patients and to evaluate whether B-Raf specific immune responses could serve as a melanoma marker we examined the Raf specific humoral response using an ELISA assay with purified recombinant B-Raf, B-Raf V599E or C-Raf protein. Methods Participants Patient sera were obtained from frozen stocks and were collected over a period of 5 years. All patients gave informed consent to use their sera for scientific analysis. Control sera were obtained from patients of the dermatology department without signs of melanoma. Control patients were fully anonymized and no further information is available. Antigens for ELISA Recombinant wild type B-Raf, B-Raf V599E and C-Raf proteins were expressed in Sf9 cells and purified as GST fusions (B-Raf V599E, C-Raf) or His-tagged proteins (B-Raf) as described [ 20 ]. Purity of the Raf kinase preparations was controlled by SDS-polyacrylamide gel electrophoresis and staining with Coomassie Blue. ELISA 372 sera of melanoma patients were analyzed for B-Raf V599E specific response, 271 sera were analyzed for B-Raf wt and C-Raf specific responses and 119 sera of non melanoma control patients of the dermatology department were used and analyzed for B-Raf V599E, B-Raf wt and C-Raf specific responses. For determination of serum levels of total Ig or IgG, an ELISA assay was developed. 1 μg/ml of purified B-Raf, B-Raf V599E or C-Raf protein in 100 μl coating-buffer or buffer alone as background control (carbonate buffer, pH 9,6) was coated on NUNC 96-Well MaxiSorp plates at 4°C overnight. Plates were washed twice with washing buffer (0,05% Tween (Sigma) in PBS) and blocked with 1% BSA (Sigma) in PBS. After washing twice, serial dilutions of human sera, starting at 1:100, in 100 μl conjugate buffer (1% BSA, 0,05% Tween in PBS) were incubated for 1.5 hours at 37°C. After four wash steps, alkaline phosphatase coupled goat anti human Ig (IgM + IgA + IgG, Dianova) or goat anti human IgG (Dianova) diluted 1:2000 in 100 μl conjugate buffer was added. After 1 h at 37°C and two washing steps, 50 μl of pNPP (Sigma) substrate in buffer was added. The reaction was incubated at room temperature and stopped after 20 min by 50 μl 1 M NaOH. Optical density was read at a wavelength of 405 nm (OD 405 ) in a TECAN Spectra Thermo microplate reader. For determination of specificity, a similar protocol was used, but plates were coated with different concentrations of antigen (0, 0.1, 1.0 and 2.0 μg/ml antigen in coating buffer). Each concentration was determined in duplicate. In addition, specificity was further confirmed using an unrelated antigen, prostate specific antigen (PSA, Sigma), at concentrations of 1 μg/ml for ELISA analysis. Two patients exhibited non-specific serum responses and were excluded from further analysis. For competition ELISA performed with some positive sera, a similar protocol was engaged, using a fixed dilution of the sera and B-Raf or B-Raf V599E antigen for coating. Competition was performed by adding different concentrations of B-Raf or B-Raf V599E (0.0, 0.1, 1 or 2 μg/ml) to the human sera at the corresponding incubation step. Data analysis For the analysis of the ELISA data, the titer of the sera was defined as the last serial dilution with an OD 405 value exceeding the Cut Off 0.275, corresponding to the mean + 1.5 SD of the background values of positive sera at serum dilutions of 1:50. Raf specific titers were compared to the background values using a two tailed Mann-Whitney U test with a 95% confidence interval. The amount of positive sera within the two groups were defined as sera exceeding a given titer; the final definition of positive sera were sera with titers of 1:300 or higher. Frequencies of positive sera were compared using two sided Fisher's exact test with 95% confidence interval. Results and discussion 372 sera of 148 melanoma patients and sera of 119 control patients were screened for Raf specific antibody responses by an ELISA assay using a secondary antibody directed against IgG, IgA and IgM. Serial dilutions of the sera were analysed for their OD 405 and titers were attributed as described using a cut off OD 405 of 0.275 (fig. 1A and 1B ). Positive sera reacted against B-Raf or B-Raf V599E, but not against an unrelated antigen (fig. 1A ). Specificity was further confirmed by coating the ELISA plate using different concentration of the antigen (fig. 1C ) and WESTERN Blot analysis (additional figure 1 [see additional file 1 ]). In some positive sera analysed for IgG antibodies, very weak B-Raf specific titers could be detected (fig. 1D ), but most sera were negative for specific IgG antibodies (data not shown). Some positive sera were additionally tested for specificity for the mutated epitope using a competition ELISA by coating B-Raf or B-Raf V599E and competing Raf specific human antibodies by titrating in unbound B-Raf or B-Raf V599E (fig. 2A,2B,2C ). In all cases tested, competition was achieved in every combination at corresponding antigen concentrations, whereas a control serum with non-specific reactivity (fig. 2D ) showed no competition as expected. Taken together these results demonstrate the presence of Raf specific antibodies. In most cases, titers against B-Raf and B-Raf V599E were consistently higher compared to titers for C-Raf (fig. 1B ); only some weakly positive sera showed a response dominated by C-Raf antibodies and in a minority of positive sera the specific response was equilibrated. Even though antigen preparations were not checked for biological activity and correct refolding and therefore a direct comparison between C-Raf and B-Raf specific antibody levels based on these assays is difficult, the data suggest that the Raf specific antibody response in patients is mainly directed against B-Raf. The moderate cross reactivity against C-Raf but not against an unrelated antigen might be explained by the high conformational and sequence similarity between Raf family members. In contrast, the differences observed between the reactivity against B-Raf or B-Raf V599E are weak (fig. 1B ) and suggest that B-Raf and B-Raf V599E are recognized to a similar extend. This notion is further strengthened by the results of the competition ELISA which failed to detect any differences in the pattern of competition for a given serum regardless of the antigen combination used (figure 2 ). Therefore we conclude that the antibodies can not discriminate between the naïve or the mutated form of B-Raf. Comparison between Raf specific antibody responses in sera from melanoma patients and control patients reveals apparent differences (figure 3 ). Raf specific responses are significantly different from background values for melanoma patients and all Raf variants tested (two tailed Mann-Whitney U test). In contrast, the difference observed within the control group is not significant (Mann-Whitney U test). However, reactivity against C-Raf is rather weak in all sera tested and only 5 of 271 tested sera from melanoma patients reach C-Raf specific titers of 1:300. A comparison between the percentage of patients with B-Raf specific antibody levels above a given titer reveals that sera derived from melanoma patients show consistently higher values compared to the control group (fig. 4A ). In contrast, the frequency of C-Raf specific responses are only marginally higher in the melanoma population and not detectable at titers above 1:300. If positive response is defined as sera with titers of 1:300 or higher, 2.5% of the control group is positive for B-Raf and 1.68% for B-Raf V599E antibodies (fig. 4B ). However, 5.41% (P = 0.12, Fisher's exact test) and 8.86% (P = 0.028, Fisher's exact test) of sera derived from melanoma patients were positive for B-Raf V599E and B-Raf respectively (fig. 4B ). Using this cut off, no control patients could be identified with C-Raf specific antibodies, and only 1.85% (P = 0.329, Fisher's exact test) of the melanoma patients had detectable C-Raf specific antibodies (fig. 4B ). These results suggest that melanoma is associated with higher rates of patients with detectable B-Raf specific antibody responses. The difference between the values obtained for B-Raf V599E and B-Raf specific antibodies, which is expressed in lower scores for B-Raf V599E, is most probably due to differences in the antigen preparation and is less likely due to different specificity as in most cases the scores against B-Raf V599E are in the same range as the scores against B-Raf (see figure 1B ), whereas the C-Raf response consistently shows much lower titers in positive cases and a different shape of the curve (figure 1B ). Taken together, these data strongly suggest that 8.9% of melanoma patients have a B-Raf specific antibody response and 2.5% of the control patients. At this stage, it is far too early to foresee the consequences of this finding for diagnostics or therapy of melanoma. All melanoma patients were in a very advanced, rapidly progressing stage with high tumor loads (stage IV) and consequently the presence of antibodies was not correlated with survival (data not shown). However, at least some patients had low but detectable levels of B-Raf specific IgG antibodies which is in line with the independent finding of B-Raf V599E specific T-cells in melanoma patients [ 18 ]. Another interesting point to note is the small, although not significant, number of control patients with detectable levels of Raf specific antibodies. At least two out of 119 control patients showed a very strong B-Raf specific response in the same range as we have observed for positive melanoma patients (figure 3 ). Due to the design of the study, no information is available for this patient group. Serum was sampled during diagnostic procedures, therefore other underlying diseases as a cause for Raf specific antibodies cannot be excluded, including autoimmune diseases or neoplastic diseases of different origin. Furthermore, as the control group has been sampled in the dermatology department, this group might contain patients with high numbers or irregular melanocytic nevi or even unrecognised melanoma. For different extracellular tumor antigens, including gangliosides, in melanoma, positive associations between the induction of IgM and IgG antibodies and survival time have been reported [ 21 ] and the protective effect was attributed to antibody dependent cellular toxicity [ 22 ]. Therefore, the induction of Raf antibodies might be a favourable goal for the design of melanoma vaccines. However it is questionable whether humoral responses against an intracellular antigen like Raf will have an effect. In this respect it is interesting to ask why responses against intracellular antigens like Raf or survivin [ 23 ] can be observed at all. The most straightforward explanation, which is particularly likely for advanced stage IV melanoma patients, relies on the fact that the partial necrosis of big tumor masses allows the crosspresentation of intracellular proteins. The longitudinal analysis of our patients, e.g. the one depicted in figure 1 , illustrates, that the occurrence of humoral immune responses was correlated rather to tumor burden than therapeutic measures, since the conversion occurred during the period when no therapy was applied. As oncogenic mutation of B-Raf and even transformation by other oncogenic events is frequently accompanied by Raf overexpression, the induction of an antibody response is not necessarily due to the presence of the mutation. Even though a polyclonal IgM response is very unlikely to be specific for a point mutation, this notion would explain the lack of specificity for the mutational epitope. However, it is also possible that a cellular immune response prior to the induction of Raf specific antibodies has occurred. Such a T-cell response directed against Raf or other tumor antigens, would also result in the lysis of tumor cells and the liberation of intracellular antigens. It is important to mention that the lysis leading to the formation of antibodies against intracellular antigens can be caused by T-cells other than the already observed B-Raf V599E specific CD4 + or CD8 + T-cells. This question still remains to be clarified as up to now only advanced stage patients have been examined. It is obvious that at this stage the occurrence of antibodies is not correlated with prognosis and that it is very unlikely that the antibody response itself will have an influence on the disease. However, as it is rather unlikely that the frequency of a B-Raf specific T-cell response exceeds the frequency of antigen responses, our data can help to set the upper limit for the expected frequency of B-Raf or B-Raf V599E specific T-cells as 8.5% which allows the rational design of search strategies especially for Raf specific CD8 + T-cells. Even though the detection of B-Raf specific humoral and T-cell responses suggest that B-Raf/B-Raf V599E is immunogenic and that a response could be induced in at least a part of the patients, this notion does not allow conclusions on its suitability as a target for immunotherapy per se. In theory, an optimal tumor antigen is exclusively expressed in the tumor tissue and essential for tumor cell growth and survival to avoid the emergence of escape mutants or antigenic loss, whereas the immunological attack of the tumor cell is independent of the function of the antigen [ 17 ]. To date, many TAA used for the immunotherapy of melanoma including MAGE [ 24 ] or tyrosinase provide no obvious advantage to the tumor cell and therefore there is no pressure to retain the antigen. In contrast, mutated B-Raf has a high prevalence in melanoma and the consequences of its activation, including induction of proliferation and block of apoptosis are well known. However, it is still a matter of debate whether B-Raf activity mainly drives melanoma initiation or whether its function is also relevant for later stages. The first conclusion is supported by studies demonstrating that the mutations already occur in melanocytic nevi at frequencies comparable to late stage melanoma [ 8 ], are not correlated with clinical outcome [ 25 ] and finally the Raf inhibitor Bay 43-9006 seems to have only moderate efficiency in advanced melanoma as reported by Ahmed et al. during the ASCO meeting 2004 [ 26 ]. However, other studies correlated B-Raf mutations with progression rather then initiation [ 9 ] and we and others suggested a role for B-Raf as a negative prognostic factor in metastatic melanoma [ 25 , 27 ]. Despite these conflicting data on the clinical relevance of mutated B-Raf at late stages, the high prevalence would already justify to evaluate B-Raf/B-Raf V599E as a target for immunotherapy, as this prevalence is in the same range as for the TAA currently used without any evident advantage for tumor cell growth. Conclusions Taken together we have demonstrated that B-Raf/B-Raf V599E specific antibodies are detectable in 8.9% of advanced stage melanoma patients and B-Raf V599E might therefore be a valuable target for immune therapy. Combined with the independently described B-Raf V599E specific CD4 + and our earlier demonstration of B-Raf V599E specific CD8 + T-cell responses this study makes the screening for novel B-Raf V599E MHC class I epitopes for vaccination approaches a promising task. Competing interests None declared. Authors' contributions JF and JCB designed the study and performed the data and statistical analysis. JF set up and supervised the ELISA and wrote the report. JCB set up and supervised the sample collection and was involved in writing the report. TP and VH were performing the ELISA assays, CSV was responsible for serum collection, transport and maintenance. EBB and URR were involved in providing the conceptual framework for this study. All authors approved the final version of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Supplementary figure with WESTERN Blot analysis of positive sera confirming specificity. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC518968.xml
548150	Reversal of a full-length mutant huntingtin neuronal cell phenotype by chemical inhibitors of polyglutamine-mediated aggregation	Background Huntington's disease (HD) is an inherited neurodegenerative disorder triggered by an expanded polyglutamine tract in huntingtin that is thought to confer a new conformational property on this large protein. The propensity of small amino-terminal fragments with mutant, but not wild-type, glutamine tracts to self-aggregate is consistent with an altered conformation but such fragments occur relatively late in the disease process in human patients and mouse models expressing full-length mutant protein. This suggests that the altered conformational property may act within the full-length mutant huntingtin to initially trigger pathogenesis. Indeed, genotype-phenotype studies in HD have defined genetic criteria for the disease initiating mechanism, and these are all fulfilled by phenotypes associated with expression of full-length mutant huntingtin, but not amino-terminal fragment, in mouse models. As the in vitro aggregation of amino-terminal mutant huntingtin fragment offers a ready assay to identify small compounds that interfere with the conformation of the polyglutamine tract, we have identified a number of aggregation inhibitors, and tested whether these are also capable of reversing a phenotype caused by endogenous expression of mutant huntingtin in a striatal cell line from the Hdh Q111/Q111 knock-in mouse. Results We screened the NINDS Custom Collection of 1,040 FDA approved drugs and bioactive compounds for their ability to prevent in vitro aggregation of Q58-htn 1–171 amino terminal fragment. Ten compounds were identified that inhibited aggregation with IC 50 < 15 μM, including gossypol, gambogic acid, juglone, celastrol, sanguinarine and anthralin. Of these, both juglone and celastrol were effective in reversing the abnormal cellular localization of full-length mutant huntingtin observed in mutant Hdh Q111/Q111 striatal cells. Conclusions At least some compounds identified as aggregation inhibitors also prevent a neuronal cellular phenotype caused by full-length mutant huntingtin, suggesting that in vitro fragment aggregation can act as a proxy for monitoring the disease-producing conformational property in HD. Thus, identification and testing of compounds that alter in vitro aggregation is a viable approach for defining potential therapeutic compounds that may act on the deleterious conformational property of full-length mutant huntingtin.	Background Huntington's disease (HD) is a severe, dominantly inherited neurodegenerative disorder that typically has its onset in mid-life, though it may occur in the juvenile years or in the elderly, and that produces an inexorable decline to death 10–20 years later [ 1 ]. Its cardinal clinical feature is a characteristic motor disturbance involving progressive choreoathetosis, but the disorder also involves psychological changes and cognitive decline. The neuropathological hallmark of HD is the loss of medium spiny striatal projection neurons in a dorso-ventral/medio-lateral gradient that eventually decimates the caudate nucleus, but considerable neuronal loss also occurs in other parts of the basal ganglia and in the cortex [ 2 ]. The pathogenic process of HD is initially triggered by an expanded polyglutamine segment near the amino terminus of huntingtin, an ~350 kDa protein whose precise physiological function is uncertain [ 3 ]. Huntingtin is required for normal embryonic development and neurogenesis, based on the lethal consequences of mutational inactivation in the mouse [ 4 - 6 ]. By contrast, the HD mutation itself does not impair this developmental activity but rather produces a "gain-of-function" that acts to cause the disorder [ 7 ]. Genotype-phenotype studies of HD patients, in comparison with other polyglutamine neurodegenerative disorders, have delineated a number of genetic criteria for the mechanism that triggers HD pathogenesis: 1) a threshold polyglutamine length (within a normal human lifespan); 2) progressive severity with increasing polyglutamine length above the threshold; 3) complete dominance over the wild-type protein; 4) greater dependence on polyglutamine length than on huntingtin concentration (within a physiological range) and 5) striatal selectivity, due to the huntingtin protein context in which the polyglutamine tract is presented [ 8 , 9 ]. The "gain-of-function" due to the HD mutation is thought to lie in a novel conformational property conferred on mutant huntingtin by the expanded polyglutamine tract [ 10 ]. This has been supported by in vitro studies of a small amino-terminal huntingtin fragment, where an expanded polyglutamine tract promotes self-aggregation in a manner that conforms to the first four genetic criteria [ 10 - 12 ]. The in vitro aggregation involves a conformational change of the polyglutamine segment from a random coil to an amyloid structure and is paralleled in cell culture in some ways by the formation of cytoplasmic and nuclear inclusions that also incorporate other proteins [ 13 ]. Neuronal inclusions containing amino-terminal fragment have also been detected in HD brain, though their role in pathogenesis remains a matter of debate, as they may occur late in the pathogenic process as a consequence of huntingtin degradation [ 14 ]. Precise genetic modeling of HD in the mouse supports the view that in vivo , the "gain-of-function" property conferred by the expanded polyglutamine acts within full-length huntingtin to cause abnormalities that do not initially involve formation of an insoluble aggregate [ 15 , 16 ]. Knock-in mice in which the HD mutation has been introduced into Hdh , the mouse orthologue, display early biochemical and histological phenotypes that are associated with expression of full-length mutant huntingtin at normal physiological levels and in a normal developmental pattern [ 7 , 15 - 20 ]. Indeed, the phenotypes associated with expression of full-length mutant huntingtin in these mice, and in neuronal progenitor cells derived from them, also fulfill the genetic criteria for the mechanism triggering HD pathogenesis [ 15 , 20 - 22 ] One of the earliest phenotypes is the nuclear localization of full-length mutant huntingtin in the nucleus of striatal neurons [ 16 ]. Together, the knock-in mouse data suggest that the process of pathogenesis is triggered by the presence of expanded polyglutamine in full-length huntingtin and leads only after many months to the formation of amino-terminal huntingtin fragment and inclusion formation [ 15 ]. We have postulated that the same conformational property that promotes aggregation in the context of a small fragment may also act with the context of full-length huntingtin to trigger pathogenesis, possibly by altering huntingtin's interaction with another cellular element. Consequently, we have identified small molecules from the NINDS Custom Collection of bioactive compounds that inhibit in vitro aggregation of amino-terminal mutant huntingtin [ 23 ]. These have been tested for their ability to reverse the huntingtin localization phenotype associated with full-length mutant huntingtin in cultured striatal progenitor cells from Hdh knock-in mice. Our findings indicate that some of these compounds reverse the effects of the expanded polyglutamine in both assays and support the view that some inhibitors of polyglutamine aggregation may lead to viable therapeutics targeted at full-length mutant huntingtin, early in the disease process. Results Screening for inhibitors of aggregation We have previously demonstrated that, when released from the protection of a GST fusion protein, the amino terminal fragment 1–171 of mutant huntingtin, forms aggregates in a manner consistent with the genetic criteria for the mechanism of HD pathogenesis [ 10 ]. We used a modified version of this assay, implemented using a 96-well format ELIFA dot blot apparatus, to screen the NINDS Custom Collection (NCC) which consists of 1040 small bioactive compounds, both FDA-approved drugs and natural products (Figure 1A ). The screening was carried out in a blinded fashion as part of the NINDS Neurodegeneration Drug Screening Consortium, with the identities of compounds in the NCC only being made available after completion of the screens [ 23 ]. In our primary screen (Figure 1A ), GST-Q58-Htn (20 μg/ml) was mixed with thrombin (0.5 unit/μg GST-Q58-Htn) and immediately dispensed into a 96-well PCR plate containing compounds diluted to a final concentration of 100 μM. Incubation was continued for 24 hours at room temperature to allow aggregate formation. The aggregation was stopped by 2% SDS/10 mM 2-mercaptoethanol followed by boiling for 5 minutes. The mixture was filtered through a cellulose acetate membrane by using a 96-well ELIFA dot blot apparatus. The aggregates retained on the membrane were detected and quantified by immunoblotting and subsequent image analysis. A typical immunoblot result is shown in Figure 1B . Congo Red, a known huntingtin aggregation inhibitor, was used as the positive control [ 24 ]. 10 μM Congo Red can completely inhibit the Q58-Htn aggregation. DMSO, used for the negative control, had no impact on Q58-Htn aggregation. Potential inhibitors were distributed evenly cross the whole NCC library (Figure 2 ). Sixty compounds that showed more than 50% inhibitory effect were selected to be retested in a second screen at a lower concentration of 10 μM. The 8 compounds in column 5 of plate 9, were missed in the primary screening at 100 μM, and were therefore also tested in the second screening at 10 μM. In the primary screening, a "hit" could have resulted either from direct inhibition of polyglutamine-induced aggregation or indirectly, by inhibition of the thrombin and consequent failure to cleave GST-Q58-Htn, which does not by itself aggregate. Consequently, the second screening at 10 μM was carried out after thrombin digestion, to eliminate thrombin inhibitors. Western blotting showed that more than 95% of GST-Q58-Htn is cleaved by thrombin (at ratio of 0.5 unit/1 μg protein) within 30 minutes (data not shown). Consequently, the mixture of GST-Q58-Htn and thrombin was preincubated for 45 minutes, followed by centrifugation to remove any aggregates already formed, before adding the test compounds. Nineteen of the compounds tested at 10 μM, showed significant direct inhibitory effects on aggregation. The 10 most potent compounds, corresponding to a 'hit' rate of 1%, are shown in Figure 3 . Characteristics of aggregation inhibitors To determine the potency of each inhibitor, we performed dose response assays at concentrations ranging from 0.01 μM to 500 μM. Representative curves for the 6 most potent compounds are shown in Figure 4 . Gambogic acid and celastrol showed strong but incomplete inhibition even at the maximum concentration, permitting approximately 20% residual aggregate to form. The average IC 50 (half-maximal inhibition) values for the most potent 10 compounds, which range from 0.7 to 15 μM, were obtained from at least two independent experiments each (Figure 3 ). The most effective aggregation inhibitor was gossypol-acetic acid complex, followed closely by gambogic acid, and then juglone, celastrol, and sanguinarine nitrate, which all had IC 50 values less than 6 μM. Effect on striatal cells expressing endogenous full-length mutant huntingtin To test the hypothesis that compounds, which inhibit the aggregation-promoting property of amino-terminal mutant huntingtin will also rescue effects of full-length mutant huntingtin, we tested the top six inhibitors in a striatal cell-based assay. Mutant Hdh Q111/Q111 and wild-type Hdh Q7/Q7 striatal cell lines, ST7/7 and ST111/111, respectively, which have been prepared by transformation with a tsSV40 vector, can be propagated in culture and used for cytological and biochemical comparisons [ 25 ]. These cells express full-length mutant or wild-type huntingtin, respectively, with no evidence of truncated amino-terminal fragments, no formation of polyglutamine aggregates and no cell death-producing toxicity. However, like the striatal neurons of Hdh Q111/Q111 knock-in mice, the ST111/111 cells show nuclear staining of huntingtin when tested with an amino-terminal huntingtin antibody that is sensitive to the conformation of the full-length protein (Figure 5A ). By contrast, ST7/7 cells expressing wild type huntingtin show both nuclear and cytoplasmic immunostaining with the same huntingtin antibody (Figure 5A ). This differential localization phenotype occurs early in the cascade of events detected during the lifespan of Hdh knock-in mice, months before the appearance of huntingtin amino-terminal fragment, and fulfills the genetic criteria from genotype-phenotype studies in HD patients, including polyglutamine length progressiveness and striatal specificity, suggesting that it follows from the same property that triggers HD pathogenesis. This huntingtin localization phenotype was used to monitor the effect of inhibitors in the mutant cells (Figure 5B ), and the results are shown in Table 1 . About 89% of ST7/7 striatal cells showed both nuclear and cytoplasmic immunostaining signal, while ST111/111 striatal cells showed only nuclear signal (99%). Of the six compounds tested, celastrol and juglone both reversed the mutant phenotype in a dose-dependent manner. Juglone showed no evident cell toxicity up to 10 μM, where 68% of mutant cells had reverted to wild-type phenotype. Celastrol reverted up to 81% of the cells but showed toxicity, killing ~4% of cells at 10 μM. Gossypol acetic acid complex was less effective, but showed no toxicity. At 50 μM, only 15% of the mutant cells displayed the wild-type phenotype. Gambogic acid showed a comparable small effect at 10 μM but was very toxic at high concentration, as all cells were killed at 50 μM. Sanguinarine nitrate and anthralin showed little effect on the huntingtin localization phenotype in this assay. Discussion A dramatic marker of pathology in many neurodegenerative disorders is the appearance of intracellular inclusions in some surviving neurons [ 13 ]. In HD, these inclusions stain positively for huntingtin, ubiquitin and a number of other proteins, but are thought to be initiated by the aggregation of an amino-terminal fragment of mutant huntingtin, due to its expanded polyglutamine tract [ 14 , 26 - 28 ]. A number of model systems have been developed to investigate the polyglutamine-driven aggregation process and its consequences both in vitro and in vivo , but it remains unclear in HD whether the formation of aggregates plays an essential role in the pathway of pathogenesis or is a downstream by-product of neuronal dysfunction induced by full-length mutant huntingtin [ 29 ]. In either event, the search for drugs that alter in vitro aggregation of amino-terminal huntingtin fragment is attractive, since the aggregation-promoting physical property exhibits characteristics comparable to the disease-producing property of corresponding human alleles, as defined from genotype-phenotype studies of HD patients. For example, if accumulation of huntingtin inclusions is the proximate cause of neuronal death, compounds that inhibit aggregation would have therapeutic potential. Conversely, if the inclusions are only a downstream marker of the pathogenic process, the same drugs may still have therapeutic potential if they act on the property of full-length mutant huntingtin that triggers pathogenesis. It was with a view to testing whether the in vitro aggregation assay could act as a proxy for monitoring the disease-producing property of mutant huntingtin that we undertook this study. A comparable screening assay to the one used here has been employed to screen a large chemical library and has demonstrated the feasibility of identifying small molecule inhibitors of polyglutamine aggregation, including a family of benzothiazole-related compounds [ 30 ]. However, the long, arduous and expensive process of developing compounds for use as drugs in humans prompted us to screen a smaller chemical library biased toward drugs already approved by the U.S. Food and Drug Administration. The aggregation-inhibiting compounds that we identified came from a collection of mostly FDA-approved compounds and bioactive natural products that were specifically assembled for a neurodegenerative disease drug screening consortium supported by the National Institute for Neurological Disorders and Stroke and several disease foundations, including the Huntington's Disease Society of America [ 23 ]. The drugs were distributed to 27 different labs for blinded testing in assays of potential relevance to neurodegenerative disease. Unfortunately, despite the preponderance of FDA approved drugs in the collection, none of our top hits (IC 50 < 10 μM) is a compound approved for internal use in humans. The most effective inhibitor of aggregation was gossypol-acetic acid complex. Gossypol, a polyphenol found in cottonseed, has been studied extensively as a male contraceptive in China but the World Health Organization has argued against its use because of induced hypokalemia, high toxicity and the risk of permanent sterility [ 31 ]. Almost as potent, gambogic acid is a complex ring-structured natural product that is the main active ingredient of gamboge resin from the Garcinia hanburyi tree. It has long been used as a pigment for painting and in traditional medicine as a potent purgative. Recently, it was identified in a high-throughput screen as an apoptosis inducer with potential for development as an anti-cancer agent [ 32 ]. Juglone is a napthoquinone found in the bark of the black walnut ( Juglans nigra ), which has been used as an herbal medicine for its antihaemorrhagic and antifungal properties [ 33 ]. Celastrol is a triterpenoid from the vine Tripterygium wilfordii used as an alternative medicine for rheumatoid arthritis that has anti-oxidant, anti-inflammatory activity, immunosuppressive and anti-angiogenic activities. It has been proposed as worthy of exploration as a therapeutic in Alzheimer disease [ 34 ]. Sanguinarine, a benzophenanthridine alkaloid from bloodroot ( Sanguinaria canadensis ) has broad antimicrobial and anti-inflammatory, as well as potential anti-angiogenic activity [ 35 ]. It has been used as an oral rinse and as a potential antigingivitis/antiplaque agent in toothpaste [ 36 ]. Anthralin is a synthetic derivative of chrysarobin, a traditional remedy for various skin ailments from Andira araroba , that has been widely used as topical treatment for psoriasis and alopecia areata [ 37 , 38 ]. Although none of these bioactive compounds is a candidate for immediate human trials, they provided a means to test whether compounds that inhibit polyglutamine aggregation might also block the neuronal phenotype caused by an elongated polyglutamine tract in full-length huntingtin. While we do not have a direct physical measure of huntingtin conformation and it remains possible that compounds could reverse the cellular phenotype by different pathways, our finding that juglone and celastrol, two drugs of different structure selected as hits in our primary aggregation assay, are both effective at restoring cytoplasmic huntingtin staining in the striatal cell assay suggests that they both act via a conformational property of mutant huntingtin. That gossypol, gambogic acid, sanguinarine and anthralin were not effective could be due to any of a number of reasons, including cellular uptake, toxicity, interaction with other cellular components, etc. However, it may also indicate that these compounds do not directly modify the conformational property of mutant huntingtin, but instead block a different step in the in vitro aggregation process or that they do so by a different physical effect than juglone and celastrol. A detailed analysis of structure-activity relationships using structurally-related compounds and testing in vivo in knock-in mice for their ability to reverse the cascade of mutant huntingtin-associated phenotypes will be needed to adequately assess the potential of any of these different types of compounds for testing in HD clinical trials. The remaining compounds identified as weaker aggregation blockers in our primary screen include selamectin, a veterinary anti-parasitic [ 39 ], pararosaniline pamoate, a treatment for schistosomiasis [ 40 ], tyrothricin, a cyclic peptide antibiotic, and meclocycline, a tetracycline-related antibiotic. The latter is of particular interest since it was the most potent of several tetracycline-related antibiotics present among the NCC compounds, including tetracycline, chlortetracycline, demeclocycline, doxycycline, methacycline, oxytetracycline and notably, minocycline, which has been proposed as a therapeutic in HD and other neurodegenerative disorders. Minocycline is an FDA approved antibiotic used for a variety of infections that has variably been reported to improve symptoms in the R6/2 exon 1 overexpression HD model [ 41 - 44 ]. It has anti-inflammatory and anti-apoptotic activity that has been proposed to involve several potential mechanisms of action. In our hands, minocycline is a weak inhibitor of polyglutamine aggregation with an IC 50 of 43 μM (unpublished data). This is consistent with the inhibitory effect on huntingtin exon 1 aggregation reported previously at 30 μM in long-term hippocampal slice cultures from the R6/2 mouse [ 44 ]. Two safety and tolerability studies of minocycline in human HD are completed [ 45 , 46 ] and can be expected to lead to efficacy trials earlier than any trials for strong aggregation inhibitors. However, it is conceivable that long-term, low level inhibition of mutant huntingtin's aggregation-promoting conformational property, independent of minocycline's anti-apoptotic activity, may be sufficient to alter detectably the timing of disease onset or early progression. If the hoped for positive results are obtained in minocycline HD trials, this alternative mechanism should be considered since it would have implications for testing of meclocycline and for assessing the potential trade-off between potency and toxicity in choosing other aggregation inhibitors as potential long-term therapeutics. Interestingly, the same set of 1040 NCC compounds were screened for their ability to block toxicity in a PC12 cellular assay where induced expression of huntingtin exon 1 encoding 103 glutamines leads to the accumulation of aggregates and rapid cell death [ 47 ]. Although eighteen compounds were found to be completely protective, none was among our hits, suggesting that the mechanism of polyglutamine toxicity in the PC12 cells is fundamentally different than the mechanism(s) involved in the in vitro aggregation assay. Among a secondary class of partially protective compounds in the PC12 assay, only celastrol overlapped with our hits. The NCC compounds were also screened in a cellular assay in HEK 293T cells expressing androgen receptor with 112 glutamines [ 48 ]. In this model for spinal bulbar and muscular atrophy, accumulation of intracellular inclusions, accompanied by caspase 3 activation, is followed by cell death within 72 hours. Twenty compounds that blocked caspase 3 activation included celastrol, gambogic acid, sanguinarine and tyrothricin, though all but sanguinarine showed toxicity. The major finding from this assay was that several cardiac glycosides were protective, presumably by a different mechanism than our hits. Indeed, although most of the assays in the NINDS consortium involved disorders associated with protein aggregation, including various polyglutamine disorders, amyotrophic lateral sclerosis and Parkinson disease, there was a remarkable lack of overlap in hits suggesting that the individual assays targeted fundamentally different mechanisms. A possible exception was celastrol, which was found as a hit in our aggregation assay, the two assays noted above, and other assays which will be discussed in a summary article describing the consortium. Conclusions The identification and further characterization of chemical inhibitors of in vitro aggregation of an amino-terminal fragment of mutant huntingtin offer promise for the development of potentially therapeutic compounds that also target the deleterious conformational property of full-length mutant huntingtin. Methods Chemical library, enzymes and antibodies A library containing 1040 small chemical compounds consisting of FDA-approved drugs and bioactive natural products, the National Institute of Neurodegenerative Diseases and Stroke Custom Collection (NCC), was provided in thirteen 96-well plates by MicroSource Discovery Systems, Inc (Gaylordsville CT). The complete list of NCC compounds is available [ 49 ]. All compounds were dissolved in 100% DMSO at a concentration of 10 mM. Thrombin was purchased from Amersham Pharmacia Biotech (Piscataway NJ). Anti-huntingtin antibody HP1, used in the screening assay was described by Persichetti et al. [ 50 ]. AP229 used in the cell-based assay was previously described and was a gift of Dr. A.H. Sharp [ 25 ]. GST-huntingtin construct and expression A recombinant pGEX-2TK expression vector with cDNA fragment encoding amino terminal 171 amino acids of human huntingtin with a polyglutamine tract of 58 was used to prepare the GST-Q58-htn fusion protein. The GST-Q58-htn was overexpressed in BL21 cells and purified by affinity chromatography over glutathione-sepharose 4B beads (Amersham Pharmacia Biotech). The purified proteins were stored at concentration of 2.0 mg/ml at -80°C. Aggregation assay For primary screening of the chemical library, 1 μl of 4.0 mM small compound stock, diluted from the original plates, was placed in wells of 96-well plates. The fusion protein, GST-Q58-Htn was mixed with thrombin (0.5 unit/1 μg protein) at a concentration of 20 μg/ml in a buffer of 50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 2.5 mM CaCl 2 , 1.0 mM EDTA. The mixture was immediately distributed into the 96-well plates containing diluted compounds at 40 μl/well and mixed well. The final concentration of the small compounds was 100 μM. After 24 hours incubation at room temperature, the reaction was stopped by adding 10 μl of 10% SDS/50 mM 2-mercaptoethanol to each well followed by boiling in a PCR machine for 5 minutes. The mixture from each well was filtered through a cellulose acetate membrane ((0.2 μm, GE Osmonics labstore, Minnetonka MN) by using a 96-well ELIFA (Pierce Biotech). The aggregates retained on the membrane were detected by a specific anti-huntingtin antibody, HP1 (diluted 1:1000), followed by incubation with peroxidase conjugated anti-rabbit antibody (diluted 1:10,000, Sigma). Signals from SDS insoluble aggregates were scanned and quantified by using ImageMaster Totalab image analysis software (Amersham Pharmacia Biotech). In the secondary screening, all steps were the same except the final concentration of compound in each well was reduced to 10 μM and the GST-Q58-Htn/thrombin mixture was preincubated for 45-minutes at room temperature and clarified by centrifugation at 28,000 × g for 30 minutes before being added to the test wells. For IC 50 determinations, the in vitro aggregation assay and signal quantification were performed as in the second screening but varying the final concentration of input drug. The data for each inhibitor were obtained from at least two independent experiments in which every sample was analyzed in triplicate with Prism 3.0 software (Graphpad Software, Inc., San Diego, CA). Striatal cell line assay Conditionally immortalized wild-type Hdh Q7/Q7 striatal neuronal progenitor cells (ST7/7) expressing endogenous normal huntingtin, and homozygous Hdh Q111/Q111 striatal neuronal progenitor cells (ST111/111), expressing endogenous mutant huntingtin with 111-glutamines, have been described previously [ 25 ]. The striatal cell lines were grown at 33°C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal bovine serum (FBS), 1% nonessential aminoacids, 2 mM L-glutamine and 400 μg/ml G418 (Geneticin (GIBCO-BRL, Life Technologies, Gaithersburg, MD). In the immunofluorescence and confocal analysis experiment, wild-type ST7/7 and homozygous mutant ST111/111 cells, grown to confluence on glass coverslips, were treated with 6 different drugs at four different concentrations (0.5, 5, 10 or 50 μM) for 30 min at 33°C. After the treatment the drugs were removed and the cells washed twice in PBS. The cells were then fixed by incubation in 4% formaldehyde for 15 min, permeabilized by 0.1% Triton X-100 for 5 min and incubated for 30 min with blocking solution (1% bovine albumin in PBS). Coverslips were then incubated in primary antibody (AP229 1:500 dilution) for 2 h at room temperature and washed three times in PBS before a further 1 h in blocking solution containing the secondary antibody (goat anti-rabbit Cy3, Jackson ImmunoResearch, West Grove, PA. USA). After three washes in PBS coverslips were mounted onto glass slides with Vectashield (Burlingame, CA. USA) and the images were analyzed with a laser confocal microscope (Bio-Rad, Hercules, CA. USA) using 20 × objective. Cell death was quantified by scoring the percentage of cells with apoptotic nuclear morphology, i.e. condensed or fragmented nuclei, under the confocal microscope. In each case five to ten randomly selected fields were counted, comprising at least 200 cells, and each experiment was repeated 3 times. Authors' contributions JW carried out the library screen and aggregation assays. SG carried out the cell-based assays. MEM and JFG contributed to the conception of these studies. JW and JFG drafted the manuscript and MEM and SG contributed to its final version. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548150.xml
517497	Effect of green tea on blood glucose levels and serum proteomic patterns in diabetic (db/db) mice and on glucose metabolism in healthy humans	Background Green tea is widely consumed in Asian countries and is becoming increasingly popular in Western countries. Epidemiologically, it has been suggested that green tea consumption prevents type 2 diabetes. The present study was aimed at providing evidence of improvement in glucose metabolism in diabetic mice and healthy humans upon green tea consumption. Results Green tea promoted glucose metabolism in healthy human volunteers at 1.5 g/body in oral glucose tolerance tests. Green tea also lowered blood glucose levels in diabetic db+/db+ mice and streptozotocin-diabetic mice 2–6 h after administration at 300 mg/kg without affecting serum insulin level, whereas no effect was observed in control mice (+m/+m and normal ddY mice). The serum protein profiles of db+/db+ and +m/+m mice were analyzed for the first time by SELDI (surface-enhanced laser desorption/ionization)-TOF (time-of-flight)-MS (mass spectrometry), and then compared to investigate any effects of oral green tea administration on serum proteins. The protein profiles in db+/db+ mice showed that the spectral peak intensities at the mass/charge ratios (m/z) of 4119, 4203, 4206, 4211, 4579, 9311 and 18691 were >3 times lower, and those of 13075, 17406, 17407, 17418, 17622, 18431 and 26100 were >3 times higher than respective peak intensities in +m/+m mice. When green tea was administered to db+/db+ mice, the peak intensities were markedly decreased at m/z 11651 and 11863, and slightly decreased at m/z 4212. The peak intensities at 7495, 7595, 7808, 14983, 15614, 31204 were markedly increased after the administration. Conclusion The present study provides evidence that green tea has an antidiabetic effect. Although we could not find simple reversed effect of green tea on the diabetes-induced modifications of the levels of several serum proteins, we found that the 4211 (4212) Da protein level that was decreased in the diabetic state was further decreased after green tea administration. This is the first report demonstrating that a certain serum protein may be involved in the antihyperglycemic effect of green tea. The contribution of this protein should be further studied.	Background Green tea (leaves of Camellia sinensis , Theaceae) is a popular beverage in East Asia, and also used as a herbal remedy in Europe and North America. Green tea is considered to be antiinflammatory, antioxidative, antimutagenic, and anticarcinogenic [ 1 , 2 ], and can prevent cardiac disorders. Epidemiologically, it has been suggested that green tea consumption prevents type 2 diabetes. The amelioration of insulin resistance by green tea is associated with the increased expression level of glucose transporter IV in a fructose-fed rat [ 3 ]. Green tea extract contains polyphenols (e.g., catechin, epicatechin, epigallocatechin, and their gallates), teanin and caffeine. The extract also includes pyrroloquinoline quinone, a newly discovered vitamin [ 4 ]. Some constituent components have been shown to enhance the basal and insulin-stimulated glucose uptake of rat adipocytes [ 5 ], to inhibit intestinal glucose uptake by inhibiting the sodium-dependent glucose transporter of rabbit intestinal epithelial cells [ 6 ], and to reduce serum glucose level in alloxan-diabetic rats [ 7 ]. Controversially, caffeine acutely lowers insulin sensitivity in humans [ 8 ]. The present study was aimed at providing evidence of the improvement in glucose metabolism in humans and diabetic mice upon green tea consumption. Furthermore, to investigate whether some serum proteins mediate the effects of green tea on hyperglycemia in the diabetic state, proteomic analyses were performed using SELDI (surface enhanced laser desorption/ionization)-TOF (time-of-flight)-MS (mass spectrometry), because the SELDI-TOF-MS enables the measurement of the relative abundance of proteins of various sizes in a blood sample. Here, we compared the proteomic patterns of sera from diabetic db+/db+ and wild-type mice, and further investigated the influence of green tea administration on the proteomic patterns of serum in diabetic mice. If the diabetic modifications of serum protein levels are reversed by green tea administration, these affected proteins can be used as therapeutic targets in diabetes. Results Oral glucose tolerance test (OGTT) We investigated whether oral glucose tolerance is improved in healthy volunteers by drinking a suspension of green tea powder. Blood glucose levels were measured before and 30, 60, and 120 min after drinking 1.5 g of green tea. The effects of hot water without tea powder on blood glucose level in the same participants were also investigated. Fig. 1 shows that glucose tolerance was substantially improved with tea administration compared with hot water administration. In detail, glucose metabolism increased in 14 participants, remained unchanged in 3 participants, and worsened in 5 participants. Basal blood glucose levels (BGLs) did not significantly change in all participants (Fig. 1 ). Figure 1 Increase in glucose metabolism in healthy humans administered with green tea. Oral glucose tolerance curves (OGTT, n = 22) and basal blood glucose levels (Basal, n = 25) before and 30, 60 and 120 min after administration with either a suspension of green tea powder (●) or hot water (○). All participants were fasted since the last supper before starting the experiment at 9:30 a.m. In the glucose tolerance test, 225 ml of Trelan-G75 containing 75 g glucose was perorally administered 10 min after drinking a suspension of green tea powder (1.5 g/150 ml hot water) or hot water (150 ml). Effect of green tea or hot water on basal blood glucose levels was examined without administrating Trelan-G75. Blood glucose levels (mg/dl) measured were averaged in each treatment group, and compared by cross-over test. Values represent the means ± S.E.M. P < 0.05, P < 0.01; significantly different from the vehicle control at each time-point, determined by one-way ANOVA, followed by Scheffé's test. Antihyperglycemic effects of green tea in diabetic mice The crude extract of fresh tea leaves picked from either No.12 or No. 13 cultivar at 300 mg/kg significantly lowered BGL 2–6 h after peroral administration in streptozotocin (STZ)-diabetic ddY mice in the fasting state (Fig. 2 ). In the saline-treated control group (n = 5), BGL slightly changed from 273 ± 28 mg/dl to 241 ± 28 mg/dl after 6 h, whereas in the green-tea (No. 12)-treated group (n = 3), BGL markedly changed from 235 ± 15 mg/dl to 116 ± 12 mg/dl. Similarly, BGL markedly changed from 201 ± 10 mg/dl to106 ± 8 mg/dl after 6 h in the green-tea (No. 13)-treated group (n = 3). Figure 2 Antihyperglycemic effects of green tea leaf extracts in the fasting STZ-diabetic ddY mice. The dried extracts of fresh tea leaves prepared at the Taiwan Tea Experiment Station [TTES, Cultivar No. 12 (Chinhsuan) and No.13 (Tzuiyu)] were dissolved in saline and perorally administered at 300 mg/kg into STZ-diabetic ddY mice in the fasting state. Blood glucose levels 2 h, 4 h, and 6 h after administration of green tea leaf extracts are expressed as percentages of the glucose levels at 0 h (before the administration) in each mouse, and then averaged. Values represent the means ± S.E.M. (n = 3–5). P < 0.05; significantly different from the blood glucose levels of STZ-diabetic mice administered with saline (control) at each time-point, determined by one-way ANOVA, followed by Scheffé's test. The antihyperglycemic effects of green tea powder suspension at 30, 150, and 300 mg/kg were examined 2 h after administration to the STZ-diabetic mice in the fasting state. As shown in Fig. 3 , green tea tended to lower BGL at 150 mg/kg, and significantly lowered BGL at 300 mg/kg. Furthermore, the effect of green tea was compared among different groups of mice, i.e., STZ-diabetic ddY mice, normal ddY mice, diabetic db+/db+ mice and +m/+m mice (Fig. 4 ). BGLs were significantly lowered in both STZ-diabetic mice and db+/db+ mice 2 h after administration of green tea powder suspension (300 mg/kg), whereas no significant changes were observed in normal ddY mice and +m/+m mice. Figure 3 Antihyperglycemic effects of green tea powder suspension in the fasting STZ-diabetic ddY mice. STZ-diabetic ddY mice were perorally administered with a suspension of green tea powder at 30, 150, and 300 mg/kg in the fasting state. Blood glucose levels 2 h after the administration of green tea are expressed as percentages of the glucose level 0 h before the administration in each mouse, and then averaged. Values represent the means ± S.E.M. (n = 6 – 20). *P < 0.01; significantly different from the blood glucose levels of STZ-diabetic mice administered with saline (control, 0 mg/kg of green tea), by one-way ANOVA, followed by Scheffé's test. Figure 4 Blood-glucose-lowering effects of green tea in diabetic mice but not in normal mice. STZ-diabetic ddY, normal ddY, diabetic db+/db+, and wild-type (+m/+m) mice were perorally administered with either a suspension of green tea powder at 300 mg/kg or saline in the fasting state. Blood glucose levels (mg/dl) were measured before and 2 h after the administration, and then averaged. Values represent the means ± S.E.M. (n = 5 – 21). P < 0.05, P < 0.01; significantly different from values at 0 h (before the administration), by one-way ANOVA, followed by Scheffé's test. No elevation of blood insulin levels was observed during the decrease in BGL either by tea leaf extract or by green tea suspension in STZ-diabetic mice (data not shown). These results suggest that both fresh green tea leaf extract and green tea suspension lower BGL without changing serum insulin concentration. Serum protein profiles altered in diabetic db+/db+ mice To explore the mechanisms underlying the antihyperglycemic effect of green tea, proteomic patterns of sera from diabetic (db+/db+) and wild-type (+m/+m) mice were first investigated and compared to identify peaks specific to the diabetic state using SELDI-TOF-MS. Samples of sera from fasting +m/+m mice were loaded onto CM10 (a cationic exchanger, pH4) ProteinChip arrays (Fig. 5 ). The other types of ProteinChip Array, i.e., CM10 (a cationic exchanger, pH7), Q10 (an anionic exchanger, pH8) and IMAC30-Cu 2+ (immobilized metal affinity chromatography, pH7), were also used to fractionate proteins in the serum. Then, these arrays were analyzed using a ProteinChip system, and the amount of individual serum proteins was estimated from the peak intensity of the mass spectral signal and the mass/charge ratio (m/z) equivalent to the molecular weight of each protein. All the spectral peaks of serum proteins from diabetic mice, which were significantly different (P < 0.05, by unpaired t -test) in intensities from those of serum proteins from wild-type mice, are shown in Fig. 6A : the peak intensities of proteins at m/z 4119, 4203, 4206, 4211, 4579, 9311 and 18691 were >3 times lower, and those at m/z 13075, 17406, 17407, 17418, 17622, 18431, 26100 were >3 times higher than respective peak intensities of proteins of sera from +m/+m mice. Typical spectra are shown in Figs. 6B and 6C (see additional file 1 concerning the other peaks). Figure 5 Analysis of proteomic profiles in mouse serum using SELDI-TOF-MS. Serum samples from fasting wild-type (+m/+m) mice were loaded onto the spots of CM10 (a cationic exchanger, pH4) ProteinChip arrays. The spots were analyzed using the SELDI ProteinChip system on three different ranges: m/z 3000 – 10000, m/z 10000 – 20000 and m/z 20000 – 100000. The y-axis of the spectra indicates the mass-to-charge ratio (m/z) of protonated proteins, and the x-axis indicates the relative intensities of mass spectral signals. Note: The intensity of each peak is directly proportional to the amount of protein, but the peak intensities could not be comparable among the different proteins because of the difference in ionization. Figure 6 Proteomic analyses demonstrating the differences in protein profiles of sera from diabetic and wild-type mice. Serum samples from fasting diabetic (db+/db+) mice and fasting wild-type (+m/+m) mice were loaded onto ProteinChip arrays. (A) List of peaks of proteins and/or peptides with indicated m/z values, the peak intensities of which were significantly changed in the diabetic state. Relative peak intensities were averaged (n = 8). The fold changes are presented as ratios of the peak intensities at indicated m/z values in db+/db+ to those in the +m/+m mice. Peaks at m/z 4203, 4576, 8515, 9291, 17406 and 18678 were obtained using CM10 ProteinChip (pH4). Peaks at 8733 and 9311 m/z were obtained using CM10 ProteinChip (pH7). Peaks at m/z 3933, 4119, 4206, 4369, 4566, 4579, 4637, 8523, 8827, 8915, 9283, 13075, 17407, 17418, 17622, 18431, 18691, 22334 and 26100 were obtained using Q10 ProteinChip. Peak at m/z 4211 was obtained using IMAC30. The chips were analyzed by SELDI-TOF-MS. (B) and (C) Typical data of relative peak intensities in +m/+m and db+/db+ mouse sera (upper, representative of 4–8 independent observations) and the peak intensity averages at m/z 4206 and 26100 (lower, n = 4 for +m/+m, n = 8 for db+/db+). The analyzed peak is indicated by arrows in the data of mass spectral signals. P < 0.01; significantly different from the peak in wild-type mice, by unpaired t -test. Modification of serum protein profiles of diabetic db+/db+ mice by green tea Serum protein profiles of db+/db+ mice were investigated 2 h after green tea administration, and compared with those administered with saline (control). Again, we confirmed a significant decrease in BGL by green tea in the blood samples used in these proteomic analyses (data not shown). All the protein signals that were significantly changed (P < 0.05, by unpaired t -test) in terms of peak intensity by the green tea administration, but not by the saline administration, are described in Fig. 7A . The serum proteomic patterns of the green-tea-treated group demonstrate that the peak intensities of proteins at m/z 11651 and 11863 decreased to less than one third after 2 h of tea administration, whereas no significant changes were observed in the control group. Typical spectra are shown in Fig. 7B (see additional file 2 for the change at m/z 11651). When the results in Fig. 7A are compared with those in Fig. 6A to identify protein peaks specific for diabetes and which are sensitive to green tea administration, a peak at m/z 4211 (4212) was found (Fig. 8 ). The intensity of this peak was significantly lowered in the diabetic state (Fig. 8A ), and was significantly decreased 2 h after green tea administration (Fig. 8B ). Figure 7 Changes in serum protein profiles in db+/db+ mice by administration of green tea. Serum samples from fasting diabetic (db+/db+) mice 2 h after administration of green tea suspension were loaded onto ProteinChip Arrays. The chips were analyzed using SELDI-TOF-MS. (A) List of peaks of proteins and/or peptides with indicated m/z values, the peak intensities of which were significantly changed by the green tea administration. Relative peak intensities were averaged (n = 4). The fold changes are presented as ratios of the peak intensities at indicated m/z values 2 h after to before green tea administration. Peaks at m/z 7495, 7595, 7808, 7920, 14983, 15612 and 15614 were obtained using CM10 ProteinChip (pH4), whereas those at m/z 7503, 7611, 7823, 7926, 11651, 11664, 11863, 15004 and 15638 were obtained using CM10 ProteinChip (pH7). Peaks at m/z 4212, 4226, 7499, 11637, 11846, 13711, 13831, 14974, 15180, 31204 and 65906 were obtained using IMAC30 ProteinChip. (B) Typical data of relative peak intensities (upper, representative of 4 independent observations) and the peak intensity average at m/z 11863 (lower, n = 4) after green tea administration and saline control. The analyzed peak is indicated by arrows in the data of mass spectral signals. P < 0.01; significantly different from the peak obtained before the administration, by unpaired t -test. Figure 8 Changes in peak intensity at m/z 4211 in diabetic state and after green tea administration. (A) Decrease in the peak intensity at m/z 4211 in serum samples from fasting diabetic (db+/db+) mice, compared with that of the fasting wild-type (+m/+m) mice. Typical data of relative peak intensities in +m/+m and db+/db+ mouse sera (upper, representative of 4–8 independent observations) and the averaged intensities of the peak at m/z 4211 (lower, n = 4 for +m/+m, n = 8 for db+/db+) indicated above by arrows. P < 0.01; significantly different from the peak in wild-type mice, by unpaired t -test. (B) Decrease in the peak intensity at m/z 4212 in serum samples from fasting db+/db+ mice 2 h after green tea administration. Typical data of relative peak intensity (upper, representative of 4 independent observations) and the averaged intensities of the peak at m/z 4212 (lower, n = 4) after administration with either green tea or saline. P < 0.01; significantly different from the peak obtained before the administration, by unpaired t -test. The relative peak intensities of hemoglobin-related multisignals ranging from m/z 14974 to 15638 [ 9 ] were increased by green tea administration (Fig. 7A ). Parallel changes were observed at half and double these m/z values (Fig. 7A , 9 ) (see additional file 2 concerning these changes). Since multicharged protein ions are apparently observed as proteins of different sizes in the mass spectrometry, these changes seem to reflect the modification of one group of proteins (hemoglobin-related proteins) by green tea intake. Figure 9 Hemoglobin-related multiple SELDI-TOF-MS signals in db+/db+ mice. Samples of sera from nontreated db+/db+ mice in the fasting state were loaded onto ProteinChip arrays. Left panel: hemoglobin-related multi-MS signals. From the average of 8 data, the signals were detected from m/z 14974 to 15638 (+H) using CM10 ProteinChip (pH7). The double-charged m/z values appear to be observed from m/z 7495 to 7823 (+2H). These values may correspond to hemoglobin α- and β-chains, and other hemoglobin-related proteins. Right panel: multi-MS signals observed at approximately double the single-charged m/z values for the hemoglobin-related signals, using IMAC ProteinChip. These signals may correspond to the dimers of hemoglobin α- and β-chains. Discussion Green tea is widely consumed in Asian countries, while black tea is most popular in Western countries. The manufacturing process of green tea differs from that of black tea because freshly picked young leaves of the tea are immediately steamed. This process destroys the enzymes responsible for breaking down the color pigments in the leaves and allows the tea to maintain its green color during the subsequent rolling and drying processes. The amounts of constituent compounds are slightly different from those of black tea. Pharmacological studies using constituent compounds in green tea have been recently reviewed by Kaszkin et al. [ 10 ]. Green tea extracts are more stable than pure epigallocatechin gallate, the major constituents of green tea, because of the presence of other antioxidant constituents in the extract [ 10 ]. In general, herbal medicines are complex mixtures of different compounds that often act in a synergistic fashion and exert their full beneficial effect as total extracts [ 11 ]. In the present study, we demonstrated that green tea produces an antihyperglycemic effect without affecting insulin secretion in STZ-diabetic mice. We therefore explore the mechanism underlying the green tea effect by investigating the serum protein profiles of db+/db+ mice, a genetic model of type 2 diabetes, using SELDI-TOF-MS. First, we performed a preliminary analysis to determine the peaks (molecular weights) of biomarker proteins that were abnormally contained in the serum of diabetic mice, without identifying individual serum proteins. As a result, we found that the levels of several serum proteins were significantly altered in the diabetic state. Secondly, we investigated which marker proteins are affected by green tea administration. Despite the changes in the levels of several serum proteins after green tea administration, none of the protein peaks specific for diabetes were sensitive to the administration, except for a peak at m/z 4211(4212). The level of this 4211(4212) Da protein was reduced both in the diabetic state and by green tea administration. Thus, no simple reversed effect of green tea on the diabetes-induced modifications of serum protein levels was observed. The 4211(4212) Da protein has not yet been identified, but only two candidate groups of vertebrate proteins are found by scanning Swiss-Prot database (molecular weight, 4211; molecular weight ranges, 0.1% according to the standard errors of the current MS analysis; pI, 6; pI ranges, 10): pancreatic polypeptide (PP) (primary accession numbers P13083, P37999, P41519, P38000, P11967) and antibacterial peptides [β-defensin C7 precursor (018815), 4 kDa defensin (P56686), cryptidine-5 precursor (P28312), antibacterial peptide BMAP-34 precursor (P56425)]. PP is primarily expressed in the endocrine cells of the pancreas, and the plasma PP concentrations are elevated by food intake [ 12 ]. Peripheral PP administration results in a reduction in food intake and an increase in energy expenditure [ 12 ]. Interestingly, a previous study revealed that PP-containing cell populations in the islets of Langerhans are reduced in db+/db+ mice with mild hyperinsulinemic diabetes [ 13 ], which is consistent with the observed reduction in the intensity of the peak at m/z 4211(4212) in the diabetic state. We speculate that the reduction of BGL by green tea causes a decrease in serum PP concentrations as a means of maintaining energy homeostasis, although more precise studies are required. The changes in serum protein profiles by green tea also demonstrate the increase in the peak intensities of hemoglobin-related multi-MS signals, suggesting the adverse side effects of green tea, although blood samples from db+/db+ mice tended to exhibit a hemolytic feature compared with those from wild-type mice (data not shown). Interestingly, the hemoglobin-related multi-MS signals shown in Fig. 9 may include hemoglobin α-chains and β-chains [ 9 ], their dimers, and possibly variously glycated hemoglobins. Hemoglobin A1c is used as a marker of diabetes in clinical diagnostic tests. Until now, however, the extent of glycation per hemoglobin that practically occurred is not clear. More detailed analyses of serum protein profiles using SELDI-TOF-MS will provide a more useful clinical index of the diabetic state. We observed that green tea improved oral glucose tolerance in humans. It is therefore likely that green tea is prophylactic against diabetes and ameliorates diabetic hyperglycemia. Green tea consumption at moderate doses may be associated with a reduced risk of type 2 diabetes in apparently healthy individuals by controlling postprandial hyperglycemia. Conclusions The control of postprandial hyperglycemia by green tea can help reduce the risk of type 2 diabetes. In the present study, we provide evidence showing that green tea promotes glucose metabolism in healthy humans, and produces an antihyperglycemic effect in diabetic mice. In addition, we analyzed the serum protein profiles of db+/db+ and +m/+m mice for the first time using SELDI-TOF-MS, and further investigated its association with any effects of oral green tea administration on serum proteins. Among the several proteins that were significantly lowered in the serum of diabetic mice, the 4211(4212) Da protein was significantly decreased after green tea administration. This is the first report demonstrating that a certain serum protein is involved in the antihyperglycemic effect of green tea. The contribution of this protein, therefore, should be further investigated in a future study. Moreover, we speculate that the observed effects of green tea on BGL are primarily due to the promotion of insulin action in peripheral tissues, such as skeletal muscles and adipocytes. Indeed, a recent paper showed that green tea supplementation for 12 weeks ameliorates insulin resistance and increases glucose transporter IV content in a fructose-fed rat model resembling the human type 2 diabetes mellitus [ 3 ]. Since the administration of green tea produced an acute antihyperglycemic effect on BGL in diabetic mice in the present study, additional mechanisms, such as changes in amelioration or enhancement of insulin action, should be clarified in a future study. To elucidate whether new protein synthesis is required for green tea action, it would be useful to examine the influence of protein synthesis inhibitors on the acute antihyperglycemic effect of green tea. Methods Housing and care of animals C57BLKS/J db+/db+ mice (male, 9 – 11 weeks old, 36.2 – 46.3 g, BGL: 243–411 mg/dl) and its age-matched control C57BLKS/J +m/+m mice (male, 20.6–23.9 g, BGL: 110–185 mg/dl) were purchased from SLC (Shizuoka, Japan) and used as the type 2 diabetic mouse model. Male ddY mice (4 weeks old, purchased from SLC, Shizuoka, Japan) were singly injected with STZ (150 mg/kg, i.v.), and then used 4–6 weeks after the injection (28.8–38.5 g, BGL: 247–600 mg/dl). The age-matched normal ddY mice (34.7–38.6 g, BGL: 129–197 mg/dl) were also used. The animals were housed (3–5 per cage) under a daily cycle of 12 h light and 12 h darkness, with free access to food and water. Animals were treated as approved by the Toyama Medical and Pharmaceutical University Animal Research Committee, and according to the guidelines for animal experiments established by the Japanese Pharmacological Society. Sampling of blood Mice were deprived of food for 11–14 h. Blood samples (20 μl) were collected from mouse tail veins under ether anesthesia. In experiments for proteomic analyses and serum insulin measurements, blood samples (100 μl) were collected from the orbital venous plexus of mice under ether anesthesia. The food deprivation was continued throughout the measurement of blood glucose levels (until 6 h after the administration with green tea). In the oral glucose tolerance test as described below, blood samples (0.3 μl) were obtained from human skin microvessels using FreeStyle (Nipro, Osaka, Japan). Oral glucose tolerance test (OGTT) Healthy human volunteers (18–24 years old) were fasted 12 h before the starting point of experiments. The participants were perorally administered with either a suspension of green tea powder or hot water at 9:20 a.m. Ten min after the administration (at 9:30 a.m.), the participants were perorally administered with 225 ml of Trelan-G75 (Shimizu Pharmaceuticals Co., Shimizu, Shizuoka) containing 75 g glucose. BGLs were measured before and 30, 60, and 120 min after the administration with Trelan-G75 (glucose). All the subjects enrolled in this study were ethnic Japanese. Before participation, the purpose and risks of the study were carefully explained, and written informed consent was obtained from all the participants. The protocol was approved by the Toyama Medical and Pharmaceutical University Ethics Committee regulating human research. Measurement of blood glucose levels and serum insulin levels BGLs were measured using ANTSENSE II (Horiba, Kyoto, Japan) in the mouse study and using FreeStyle (Nipro, Osaka, Japan) in the human study. Serum insulin levels were measured using an insulin-ELISA kit (Morinaga Seikagaku, Tokyo, Japan). To prepare the serum samples, blood samples from mice (100 μl) were kept on ice for 2 h, centrifuged at 16,000 × g for 1 min at 4°C, and its supernatant was immediately separated from the pellet. SELDI ProteinChip Analysis Quantitative serum proteomic profiles were measured with SELDI-TOF-MS (Ciphergen Biosystems, Yokohama, Japan). The mouse serum was prepared as described above. Several types of ProteinChip Array, i.e., Q10 (an anionic exchanger, pH8), CM10 (a cationic exchanger, pH4 and pH7) and IMAC30-Cu 2+ (immobilized metal affinity chromatography, pH7) (Ciphergen Biosystems), were used to fractionate proteins in serum. For the proteomic analyses with Q10 ProteinChip array, samples were 10-times diluted with denaturation buffer (7 M urea/ 2 M Thiourea/ 4% CHAPS/ 1% dithiothreitol/ 2% ampholine) and incubated on ice for 10 min, and then 10-times more diluted with buffer of 50 mM Tris-HCl (pH8). When CM10 ProteinChip array was used, samples were 10-times diluted with the denaturation buffer and incubated on ice for 10 min, and then 10-times more diluted with buffer of 100 mM NaOAc (pH4) or 50 mM HEPES (pH7). When IMAC30 ProteinChip array was used, samples were 100-times diluted with phophate-buffered saline. Spots on the different arrays were equilibrated with the buffer used for sample dilution, e.g., Tris-HCl buffer for spots on Q10 array, and each sample solution (70 μl) was loaded onto two separate spots on the arrays. After incubation for 20 min with rotation spots were rinsed with water and air-dried completely. The spots were analyzed using the SELDI ProteinChip system (PBS-IIc, Ciphergen Biosystems). In each sample, data from the two spots were averaged (duplicate assay). If a signal/noise ratio was larger than 2, the peak was considered to reflect the amount of a protein. Quantitative nature of the instrument was confirmed as previously described [ 14 ]. The m/z value of each protein peak (the molecular weight of each detected protein) was corrected, based on the m/z values of the external standards of peptides or proteins, the molecular weights of which are known. Standard errors in the estimated molecular weights were less than 0.1%. The axis of abscissa in spectra (Figs 6B , 7B , 8 and 9 ) indicates the mass-to-charge ratio (m/z) of protonated proteins. Reagents Fresh raw leaves of tea ( Camellia sinensis , Theaceae) (Cultivar No.12, Chinhsuan and No.13, Tzuiyu) were prepared at the Taiwan Tea Experiment Station, picked in May at Nan-Tou County, and immediately used without drying. The leaves were steeped in hot water (95°C) for 30 min and the filtrate was condensed under reduced pressure. The dry weight of the extracts was determined, and the extracts were dissolved in saline for administration into mice. On the other hand, green tea powder was donated by Fukuju-en (Kyoto, Japan). The particle size (median diameter) of the green tea powder was 2.9 μm, according to the measurement using a Centrifugal Particle Size Analyzer (Shimadzu, Kyoto, Japan). In the human study, the tea powder (1.5 g) was added to 150 ml of hot water (80°C) and then whipped with a bamboo whisk. In the mouse study, the tea powder was suspended in saline at room temperature using a sonicator. The contents of caffeine and catechins in these tea samples were determined by HPLC: the samples were injected into an HPLC column (TSK-GEL, ODS-80T M , Tosoh, Tokyo, Japan); the eluent was 10 mM phosphate buffer (pH2.6)/MeCN (gradient: 5 to 15%) at a flow rate of 1.3 ml/min. The data are shown in Table 1 . Table 1 Contents of catechins and caffeine in green tea samples (grams per 100 g) Green tea powder No. 12 a No. 13 b catechins C 0.1 1.3 0 EC 0.6 1.9 1.0 ECg 0.6 1.6 0.2 EGC 0.3 0.9 5.8 EGCg 5.6 10.6 3.5 total 7.2 16.3 10.5 caffeine 4.3 4.5 1.7 C: (+)-catechin, EC: (-)-epicatechin, ECg: (-)-epicatechin gallate, EGC: (-)-epigallocatechin, EGCg: (-)-epigallocatechin gallate, total: total catechins. a,b The dried extract of fresh raw leaves of tea ( Camellia sinensis , Theaceae) prepared at Cultivar No.12 (a, Chinhsuan) and No. 13 (b, Tzuiyu). Statistical analysis The significance of differences between two groups was assessed by Student's t -test, and the differences between multiple groups were assessed by one-way analysis of variance (ANOVA) followed by the Scheffé's multiple range test. Values of P less than 0.05 were considered to be significant. Especially, to determine the significance of the time-dependent effects of green tea (Fig. 1 and 2 ) on the blood glucose levels, repeated measures ANOVA was performed, and then, if the statistical significance was detected by this analysis, further statistical comparison at each measurement time between the groups was conducted by one-way ANOVA, followed by the Scheffé test. Authors' Contributions IK had the original idea for the study, designed and coordinated the experiments, and wrote the manuscript. JW prepared fresh young leaves of green tea and the hot water extract. MT assisted with measurement of blood glucose and insulin level. MI, HT and TS were involved in study design, data analysis, data interpretation and writing of the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Differences in serum protein profiles between diabetic and wild-type mice Typical data of relative peak intensities in +m/+m and db+/db+ mouse sera (left 2 panels, representative of 4–8 independent observations) and the peak intensity averages at m/z indicated (right panels; +m/+m: open column, n = 4; db+/db+: closed column, n = 8). The analyzed peak is indicated by arrows in the data of mass spectral signals. P < 0.01; significantly different from the peak in wild-type mice, by unpaired t -test. Types of ProteinChip used were described in the Fig. 6 legend. Click here for file Additional File 2 Changes in serum protein profiles of db+/db+ mice after green tea administration MS spectra shows typical changes in the serum protein profiles of db+/db+ mice administered with saline (left) or green tea (right). Graph shows the peak intensity averages at m/z indicated, before (open column) and after (closed column) administration with saline (n = 4) or green tea (n = 4). The analyzed peak is indicated by arrows in the MS spectra. **P < 0.01; significantly different from the peak obtained before the administration, by unpaired t -test. Types of ProteinChip used were described in the Fig. 7 legend. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517497.xml
517508	Tuberculosis presenting as immune thrombocytopenic purpura	Background Although various hematologic abnormalities are seen in tuberculosis, immune thrombocytopenic purpura is a rare event. Case Presentation We report a case of a 29 year-old male who was presented with immune thrombocytopenia-induced hemoptysis, macroscopic hematuria and generalized petechiae. The patient was found to have clinical, microbiological and radiological evidence of active pulmonary tuberculosis. The immune thrombocytopenic purpura was successfully treated with anti-tuberculous drugs combined with corticosteroids and high dose immune globulin therapy. Conclusion Immune thrombocytopenic purpura can be one of the hematological manifestations of tuberculosis which has a global prevalence with increasing incidence secondary to HIV infection.	Background During the past 2 decades, tuberculosis -both pulmonary and extrapulmonary- has re-emerged as a major health problem worldwide. Hematologic abnormalities have been described in association with mycobacterial infections for almost 100 years. Patients with both pulmonary and extrapulmonary tuberculosis (TB) may demonstrate peripheral blood abnormalities and findings may be minimal or profound [ 1 , 2 ]. A comprehensive review of the literature reveals a few case reports documenting tuberculosis as a cause of severe hematologic conditions such as hemophagocytic syndrome, thrombotic thrombocytopenic purpura and immune hemolytic anemia. Immune thrombocytopenic purpura (immune TP) associated with tuberculosis is exceedingly rare event. We report the case of immune TP associated with tuberculosis that was presented with severe hemorrhagic diathesis. Case Presentation A 29-year-old previously healthy immigrant male patient from Kazakhstan was admitted to hospital with new-onset severe hemoptysis, macroscopic hematuria and extensive cutaneous petechiae on lower extremities. He appeared ill and poorly nourished. The patient was oriented and well cooperated, and there was no previous history of hematologic or liver or another disease and recent medication. He presented with unexplained weight loss of 2 months duration along with intermittent fever, night sweats and cough. The physical examination revealed a blood pressure of 100/70 mm/Hg, pulse 100/min, a temperature 37.2°C, extensive cutaneous petechiae on lower extremities, hemorrhagic bulla on tongue and on mucosa of oral cavity, and amphoric soufflé on apex of right chest. No organomegaly or lymphadenomegaly or evidence of another disease such as chronic liver disease was detectable. The initial complete blood count revealed a white blood cell 25.1 × 10 9 /l (58% neutrophils, 29% bants, 9% lymphocytes and 4% monocytes), hemoglobin 11.2 gr/dl, hematocrit 36%, MCV 84 fl, reticulocytes 1% and platelet count 7.6 × 10 9 /l. Erythrocyte sedimentation rate was 110 mm/h. A peripheral smear was remarkable for a paucity of platelets. Coagulation profile [prothrombin time (PT), activated-partial thromboplastin time (aPTT), fibrin degradation products (FDP)] were normal. A bone marrow aspiration demonstrated hypercellularity of all cell lines with normal maturation of myeloid and erythroid precursors. Megakaryocytes were increased in number with normal morphology. On bone marrow aspiration hemophagocytosis was not observed. A chest X-ray (Figure 1 ) and computed tomography (CT) (Figure 2 ) demonstrated bilateral patchy infiltrates and walled cavities on left and right upper lobes. Acid-fast bacilli were strongly positive in sputum (Figure 3 ). Bone marrow aspirate and urine for acid-fast bacilli were negative. Sputum culture yielded mycobacterium tuberculosis complex. The following laboratory studies were normal or negative: biochemical tests (glucose, urea, creatinine, uric acid, sodium, potassium, calcium, chloride, phosphorus, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, lactate dehydrogenase, bilirubin, total protein and albumin), rheumatoid factor, anti-nuclear antibody, anti-platelet specific antibodies, Coomb's tests, HIV, hepatitis B and C virus, blood culture, bone marrow aspirate culture and abdominopelvic CT. No granuloma or hemophagocytosis was detected on bone marrow biopsy. The patient was started on rifampin 10 mg/kg/d, isoniazid 5 mg/kg/d, ethambutol 25 mg/kg/d, pyrazinamide 40 mg/kg/d, pyridoxine 75 mg/d and intravenous immune globulin (IVIg) 1 g/kg/d given for 2 days. On day 2 of hospitalization and treatment of anti-tuberculous therapy prednisolone 1 mg/kg/d was added. On day 8, platelet counts started to increase and on day 12 of the treatment it reached to 187 × 10 9 /l level. Patient improved on day 10 and he did not have any complaints on day 14; at time of discharge. He received a total of 6 red blood cell (RBC) units throughout hospitalization. During his hospitalization findings of hemolysis or gastrointestinal bleeding and massive bleeding in another site except hematuria and hemoptysis were not established. A complete blood count at discharge demonstrated a WBC 17 × 10 9 /l, Hb 9.6 g/dl, and platelet count 310 × 10 9 /l (Table 1 ). Corticosteroids were discontinued on day 14 of therapy and the patient was discharged and recurrent thrombocytopenia was not established after withdrawal of corticosteroid therapy. Ninety days after discharge, the patient was well with a platelet count of 300 × 10 9 /l (Table 1 ) and he had no side effect thought to be secondary to anti-tuberculous drugs. Figure 1 Anteroposterior chest radiograph showing cavitary lesions in both lungs Figure 2 CT showing cavitary lesions in both lungs Figure 3 Sputum showing positivity of acid-fast bacilli Table 1 Characteristics of patient Day Hb (g/dl) Hct (%) WBC (×10 9 ) Plt (×10 9 ) Treatment 1 11.2 36 25.1 7.6 Anti-tbc drugs + IVIg 2 8 25 23.1 9.6 Anti-tbc drugs + IVIg+Pred+1 RBC Susp 3 8.1 25 22.7 17.7 Anti-tbc drugs+ Pred+ 2 RBC Susp 4 10.3 33 20.6 15.1 Anti-tbc drugs+ Pred 5 8.6 27 24.5 12.3 Anti-tbc drugs+ Pred+ 1 RBC Susp 6 8.2 24 25.4 4.0 Anti-tbc drugs + Pred + 2 RBC Susp 7 10.1 29 20.6 4.7 Anti-tbc drugs+ Pred 8 10.5 32 17.7 20.2 Anti-tbc drugs+ Pred 9 9.9 31 15.8 36.6 Anti-tbc drugs+ Pred 10 10.3 33 18 59 Anti-tbc drugs+ Pred 11 9.4 28 14 100 Anti-tbc drugs+ Pred 12 9.7 28 18.1 187 Anti-tbc drugs+ Pred 13 9.1 26 15.2 324 Anti-tbc drugs+ Pred 14 9.6 28 17 310 Anti-tbc drugs 45 12.3 37 15.5 304 Anti-tbc drugs 90 14 43 11 300 Anti-tbc drugs Abbreviations: Anti-tbc drugs: anti-tuberculous drugs, Pred: Prednisolone, Susp: Suspension Discussion Tuberculosis differs from many other infectious maladies in having particular social and geographic distributions. The disease was under control in developed nations and getting under control in developing nations, until the emergence of HIV infection and the advent of multidrug resistant strains of mycobacteria [ 3 - 5 ]. Various hematologic abnormalities such as anemia, leukocytosis, monocytosis, lymphopenia, leucopenia, thrombocytopenia, thrombocytosis, leukemoid reactions and pancytopenia have been seen in tuberculosis [ 1 , 2 ], but severe thrombocytopenia and presenting of tuberculosis as immune thrombocytopenic purpura is extremely rare and there are a few reports about tuberculosis induced immune thrombocytopenic purpura published in the world literature [ 6 - 10 ]. The case that we report could be confused by coincidental presentation of adult idiopathic thrombocytopenic purpura and tuberculosis, by drug-induced thrombocytopenia, thrombotic thrombocytopenic purpura (TTP)-hemolytic uremic syndrome (HUS), hemophagocytic syndrome and disseminated intravascular coagulation (DIC) associated with TB. Idiopathic thrombocytopenic purpura (ITP, also known as primary immune thrombocytopenic purpura) is an acquired disease of children and adults defined as isolated thrombocytopenia with no clinically apparent associated conditions or other causes of thrombocytopenia. Adult ITP typically has an insidious onset with long-lasting histories of purpura (thrombocytopenia for >6 months) and spontaneous remission is uncommon and is likely to be incomplete [ 11 - 13 ]. Steroids are the conventional first line therapy for adult ITP. Platelets counts increase within one week in responding patients and usually reach peak values by two to four weeks. However, in most patients, thrombocytopenia recurs when steroids are tapered or discontinued. Also in adult, IVIg is used when clinical situations require a transient increase of the platelet count and a typical response is an increase in platelet count several days after the infusions are initiated and return to the pretreatment level within several weeks [ 11 , 13 ]. In our case, we excluded the adult ITP not only by basing on standard criteria [ 12 ], but with response to steroids and IVIg therapy since thrombocytopenia did not recur after withdrawal of prednisone and IVIg therapy. Also we excluded other causes of thrombocytopenia such as hemophagocytic syndrome, TTP, combined autoimmune cytopenias with history, clinical and laboratory findings, and examination of bone marrow aspiration and biopsy that were described in case presentation. Several factors are known to cause bleeding in association with infections, of which thrombocytopenia is the most common. The etiology of thrombocytopenia in most cases appears to be increased destruction of platelets such as due to DIC or septicemia without evidence of DIC or platelet adherence to damaged vascular surfaces or direct platelet toxicity caused by the microorganism or involvement of bone marrow. Adult acute immune thrombocytopenic purpura is defined as a bleeding disorder in otherwise healthy person caused by transient destruction of platelets. Although the most important therapy for infection-related thrombocytopenia is that directed at the underlying infection, treatment decisions for immune thrombocytopenic purpura remain controversial and may include single or combination therapy with corticosteroids, intravenous immunoglobulin (IVIg) according to degree of thrombocytopenia or hemorrhage [ 11 , 14 ]. The case that we reported was presented with symptoms of phthisis lasting for more than 2 months and severe hemoptysis, macroscopic hematuria and extensive cutaneous bleeding findings lasting for one week. Based on the clinical, radiological (X-ray and CT of chest) findings, demonstration of positivity of acid-fast bacilli in sputum and with exclusion of other causes of thrombocytopenia, immune thrombocytopenic purpura due to pulmonary tuberculosis was diagnosed. We applied anti-tuberculous therapy combined with corticosteroids and IVIg because of severe thrombocytopenia and severe hemorrhagic diathesis. Clinically, steroids are known to ameliorate the purpuric bleeding in patients before the platelet count actually increases. The early effect is due to decrease of vascular permeability. The effect of the steroids in the thrombocytopenia is probably complex and it is late effect. The mechanism of action of IVIg is unclear, but studies suggest blockage of the Fc receptors of the reticuloendothelial cells and suppression of antibody production and binding which may be a result of anti-idiotype antibodies that bind antiplatelet antibodies and modulate the immune response [ 14 ]. In our case, corticosteroids were discontinued on day 14 of therapy and the patient was discharged and recurrent thrombocytopenia was not established after withdrawal of corticosteroid therapy. These observations suggest that tuberculosis is the cause of thrombocytopenia in our patient. He received a total of 6 RBC units throughout hospitalization. He did not have findings of Coomb's positive or negative hemolytic anemia or microangiopathic hemolytic anemia, hemophagocytic syndrome, gastrointestinal bleeding and massive bleeding to thoracic cavity. But he had severe hemoptysis and macroscopic hematuria. Ninety days after discharge, the patient was in well health with a platelet count of 300 × 10 9 /l and he had no side effect thought to be secondary to anti-tuberculous drugs [ 15 , 16 ]. The patient is still in our follow-up without relapsing of thrombocytopenia. In conclusion, since the incidence of tuberculosis is currently increasing in worldwide countries and it may present with different hematologic manifestations, in case of immune thrombocytopenic purpura tuberculosis should also be recalled. Finally, further studies are needed in order to fully characterize the pathophysiology and immunological abnormalities in tuberculosis-related immune thrombocytopenic purpura.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517508.xml
314474	Long-Lasting Novelty-Induced Neuronal Reverberation during Slow-Wave Sleep in Multiple Forebrain Areas	The discovery of experience-dependent brain reactivation during both slow-wave (SW) and rapid eye-movement (REM) sleep led to the notion that the consolidation of recently acquired memory traces requires neural replay during sleep. To date, however, several observations continue to undermine this hypothesis. To address some of these objections, we investigated the effects of a transient novel experience on the long-term evolution of ongoing neuronal activity in the rat forebrain. We observed that spatiotemporal patterns of neuronal ensemble activity originally produced by the tactile exploration of novel objects recurred for up to 48 h in the cerebral cortex, hippocampus, putamen, and thalamus. This novelty-induced recurrence was characterized by low but significant correlations values. Nearly identical results were found for neuronal activity sampled when animals were moving between objects without touching them. In contrast, negligible recurrence was observed for neuronal patterns obtained when animals explored a familiar environment. While the reverberation of past patterns of neuronal activity was strongest during SW sleep, waking was correlated with a decrease of neuronal reverberation. REM sleep showed more variable results across animals. In contrast with data from hippocampal place cells, we found no evidence of time compression or expansion of neuronal reverberation in any of the sampled forebrain areas. Our results indicate that persistent experience-dependent neuronal reverberation is a general property of multiple forebrain structures. It does not consist of an exact replay of previous activity, but instead it defines a mild and consistent bias towards salient neural ensemble firing patterns. These results are compatible with a slow and progressive process of memory consolidation, reflecting novelty-related neuronal ensemble relationships that seem to be context- rather than stimulus-specific. Based on our current and previous results, we propose that the two major phases of sleep play distinct and complementary roles in memory consolidation: pretranscriptional recall during SW sleep and transcriptional storage during REM sleep.	Introduction Sleep is important for the consolidation of newly acquired memories ( Jenkins and Dallenbach 1924 ; Fishbein 1971 ; Pearlman and Becker 1974 ; Smith and Butler 1982 ; Smith and Kelly 1988 ; Karni et al. 1994 ; Stickgold et al. 2000 ; Laureys et al. 2002 ; Fenn et al. 2003 ). The discovery of experience-dependent neuronal reactivation during sleep ( Pavlides and Winson 1989 ) corroborated the notion that novel memory traces, after successful encoding, must be replayed in their supporting neuronal networks until synaptic plasticity can effect trace consolidation ( Hebb 1949 ; Gutwein et al. 1980 ; Winson 1985 ; Ribeiro et al. 1999 ). Postacquisition neuronal reactivation during sleep or quiet waking (WK) was found to preserve the temporal relationships of alert, exploratory WK in the hippocampus (HP) ( Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Nadasdy et al. 1999 ; Poe et al. 2000 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ) and the cerebral cortex (CX) ( Qin et al. 1997 ; Hoffman and McNaughton 2002 ), causing a correlated replay of activity patterns across two-neuron ( Wilson and McNaughton 1994 ) or many-neuron ( Louie and Wilson 2001 ) ensembles. To date, experience-dependent brain reactivation during sleep has been observed in rodents ( Pavlides and Winson 1989 ; Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Qin et al. 1997 ; Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ), nonhuman primates ( Hoffman and McNaughton 2002 ), humans ( Maquet et al. 2000 ), and even songbirds ( Dave and Margoliash 2000 ), pointing to a very general biological phenomenon. Importantly, postacquisition brain reactivation during sleep has been shown to be proportional to memory acquisition in rats ( Gerrard 2002 ) and humans ( Peigneux et al. 2003 ). In spite of the positive evidence, the brain reactivation hypothesis for memory consolidation during sleep faces several objections. First, the neocortical reactivation detected to date is extremely subtle and decays rapidly within less than 1 h of memory trace formation ( Qin et al. 1997 ; Hoffman and McNaughton 2002 ). Such transient reactivation falls short of explaining the disruption of memory traces by sleep deprivation several hours and even days after initial acquisition ( Fishbein 1971 ; Pearlman and Becker 1974 ; Smith and Butler 1982 ; Smith and Kelly 1988 ; Karni et al. 1994 ; Stickgold et al. 2000 ; Fenn et al. 2003 ). Second, strictu sensu neuronal reactivation during sleep in mammals has only been investigated in the hippocampocortical loop ( Pavlides and Winson 1989 ; Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Qin et al. 1997 ; Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Hoffman and McNaughton 2002 ; Lee and Wilson 2002 ), making it difficult to ascertain whether the phenomenon is particular to this neural circuit or whether it represents global experience-dependent changes in the brain. Third, brain reactivation has mostly been observed in highly trained animal subjects ( Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Qin et al. 1997 ; Nadasdy et al. 1999 ; Dave and Margoliash 2000 ; Louie and Wilson 2001 ; Hoffman and McNaughton 2002 ; Lee and Wilson 2002 ), raising skepticism about its relevance for the acquisition and consolidation of novel information ( Kudrimoti et al. 1999 ). Finally, experience-dependent brain reactivation has been reported to occur in all behavioral states ( Pavlides and Winson 1989 ; Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Qin et al. 1997 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ), including WK ( Nadasdy et al. 1999 ; Hoffman and McNaughton 2002 ). Although the first finding in this regard has hinted at a possible predominance of reactivation during slow-wave (SW) sleep ( Pavlides and Winson 1989 ), a comprehensive comparison of the relative contributions of WK, SW, and rapid eye-movement (REM) sleep for brain reactivation is still missing. To further complicate the issue, recent studies have raised the possibility that neuronal processing may occur at either slower or faster speed than normal physiological rates during REM ( Louie and Wilson 2001 ) and SW ( Nadasdy et al. 1999 ; Lee and Wilson 2002 ) sleep, respectively. Thus, it is uncertain at the moment how brain reactivation relates to different behavioral states. In order to address these objections, we set out to investigate the effects of a transient novel tactile experience on the long-term evolution of ongoing brain activity across the major behavioral states of the rat. In each of the five animals studied, extracellular activity of 59–159 neurons per animal and local field potentials (LFPs) representing larger-scale neural rhythms were simultaneously recorded from four different brain regions: HP, primary somatosensory “barrel field” CX, ventral posteromedial thalamic nucleus (TH), and putamen (PU) ( Figure 1 A and 1 B; see also Figures S6 and S7 ). These brain regions were chosen because they comprise three major forebrain circuit loops essential for rodent species-specific behaviors. Rats are nocturnal gatherers that exhibit a variety of exploratory behaviors during the night, sleeping intermittently and mostly during the day ( Timo-Iaria et al. 1970 ) (upper panel in Figure 1 C). In the wild, rats rely on spatial navigation and superb whisker-based tactile discrimination to explore new territories in search of food (Nowak 1999). The corticothalamic, corticohippocampal, and corticostriatal loops probed in this study have been implicated in tactile information processing ( Simons 1978 ; Ghazanfar et al. 2000 ), spatial navigation and memory formation ( O'Keefe 1976 ; Squire 1992 ), and the execution of complex motor sequences ( Dieckmann and Hasser 1968 ; Jog et al. 1999 ). Figure 1 Methodology (A) Neuroanatomical location of multielectrode implants, indicated on a schematic parasaggital section based on Paxinos and Watson (1997 ). Indicated are the cerebral cortex (CX), the hippocampus (HP), the thalamus (TH), and the putamen (PU). (B) Top view of a rat implanted with several multielectrode arrays. (C) Experimental design. The upper panel shows a representative example of the strong circadian dynamics of the rat sleep–wake cycle (rat 5). Gray bands indicate lights-off; white bands indicate lights-on. Notice the fixed 12-h periods of darkness and light. The lower panels show animals continuously recorded for up to 96 h that were kept undisturbed except for a 1-h period of novel CSS (white segment) produced by the tactile exploration of four distinct novel objects placed at the corners of the recording box. Neural data from pre- and postnovelty periods (black and red segments, respectively, in the middle panel) were compared. (D) Neuronal ensemble correlation method. Neuronal activity templates (red boxes) were compared with extensive recordings of neuronal action potentials (green ticks in upper panel) by way of an offline template-matching algorithm ( Louie and Wilson 2001 ) that generalizes the notion of pairwise correlations to neuronal ensembles of any size. Templates and targets (white boxes) were binned, firing-rate normalized, and correlated (middle panel). This procedure yields a time series of neuronal ensemble correlations for each template–target sliding match (lower panel). (E) Templates of interest (red boxes) were sampled around the origin of pre- and postnovelty periods during alert WK and slid against their corresponding neuronal targets so as to sample neuronal correlations every 30 s for up to 48 h. In our studies, neural signals were continuously recorded across the natural sleep–wake cycle for 48–96 h, with a single 1-h exposure to four complex objects placed in the four corners of the recording box ( Figure 1 C). All objects were strictly novel to the subjects and were designed to maximize shape, texture, and behavioral value differences ( Figure S1 ). Objects were presented half-way through the recording time (01:16 a.m. ± 00:51, mean ± SD) when lights were off and WK peaked (see Figure 1 C), so as to maximize the drive for whisker-based tactile exploration of the environment. The experiment, therefore, consisted of a naturalistic behavioral paradigm involving multiple novel sensory and spatial cues; it was designed to maximize changes induced by exposure to novel objects, as opposed to changes related to repeated behavioral training. As expected, this paradigm increased WK relative to sleep during the exposure time ( Figure S2 ), leading to novel complex sensory stimulation (CSS). Other than novel stimulation and the periodic removal of waste and introduction of food pellets and water, animals were kept undisturbed in the same environment throughout the recordings. Our paradigm produced marked and acute exploratory behavior ( Figure S2 ) without disrupting the large-scale sleep–wake structure across the many hours of recording (see upper panel in Figure 1 C). In order to investigate the long-term effects of novel stimulation on the spatiotemporal evolution of ongoing neuronal activity, we took advantage of a neuronal ensemble correlation method previously shown to detect experience-dependent reactivation of rodent hippocampal ensembles during SW and REM sleep ( Louie and Wilson 2001 ). This method generalizes the concept of pairwise neuronal correlations ( Qin et al. 1997 ; Hoffman and McNaughton 2002 ) to an arbitrarily large number of neurons, quantifying the degree of similarity between spatiotemporal patterns of neuronal activity by way of a firing rate-normalized template-matching algorithm (see Figure 1 D). Templates of alert WK neuronal ensemble activity were selected from moments when animals made whisker contact with the novel objects ( n = five templates per animal). Control templates were selected from epochs of alert WK 24 h or 48 h before novel stimulation (24 h for rats 1–3, 48 h for rats 4 and 5; n = five templates per animal), during which familiar tactile stimulation was produced by the contact of whiskers with the smooth walls of the recording box, to which animals had been habituated. Templates were matched against the entire record of neuronal activity using the neuronal ensemble correlation method (see Figure 1 E). The resulting correlation temporal profiles were averaged for each template set, aligned with reference to the light–darkness cycle to control for possible circadian effects, and compared. Results Novelty-Induced Patterns of Neuronal Activity Reverberate for up to 48 h First, we tested whether the neuronal ensemble correlation method could detect any trace of neuronal reverberation lasting at least more than 1 h after exposure to novel stimulation. For this, we examined correlation profiles obtained for all recorded neurons (three to four brain areas pooled together) in each animal. As shown in Figure 2 A, postnovelty average correlation distributions were significantly right-shifted relative to prenovelty distributions (ANOVA of mean pre- and postnovelty correlations over 24 h or 48 h; n = five animals, F = 9.5, d.f. 1, p = 0.016). This indicates that the neuronal firing patterns concomitant with novel stimulation persisted significantly more during the ensuing time than patterns sampled 24 h or 48 h before novel stimulation, when animals were in the same behavioral state (alert WK), but without novel objects to explore. The effect was independently observed, to a variable degree, in all the five animals studied (Bonferroni comparison, p < 0.01). Figure 2 B shows the temporal evolution of neuronal ensemble correlations for 24 h (upper three panels in Figure 2 B) and 48 h (lower two panels in Figure 2 B). Figure 2 Neuronal Ensemble Correlations Including up to Four Forebrain Regions Reveal Long-Lasting Reverberation (A) Postnovelty neuronal correlations were significantly larger than prenovelty correlations in all animals studied. (B) Temporal profiles of neuronal ensemble correlations. Gray bands indicate lights-off; white bands indicate lights-on. (C) Temporal evolution of p values associated with pre- and postnovelty Bonferroni comparisons performed in intervals of 1 h (rats 1–3) or 2 h (rats 4 and 5). Significant experience-dependent neuronal reverberation was detected up to 48 h after novel stimulation. Color bar in linear scale; black denotes p > 0.05. The minimum p values (MIN P) were, respectively, 10 −14 , 10 −3 , 10 −12 , 10 −23 , and 10 −22 . Despite the marked interanimal variability in the shapes and magnitudes of these profiles, a significant increase of neuronal correlations after exposure to novel stimulation was observed in most recording sites. Importantly, these increases lasted well above 1 h, as revealed by the temporal evolution of the p values generated by Bonferroni comparisons between post- and prenovelty correlation distributions ( Figure 2 C). These results indicate that significant experience-dependent neuronal reverberation could be detected in the forebrain up to 48 h after exposure to novel stimulation. In order to assess the contributions of different neurons to total ensemble correlations, we ran the correlation analysis on a neuron-by-neuron basis. We found that no one subset of neurons was particularly responsible for the reverberation effect, since the contribution of individual neurons was highly variable in time (data not shown). This indicates that the neuronal changes associated with exposure to novel stimuli were highly distributed through the neuronal populations sampled. Furthermore, judging by the maximum neuronal ensemble correlations observed (rat 3, peak at 0.35; Figure 2 B), one would conclude that novel stimulation-specific neuronal activity was not perfectly repeated, but was rather loosely reverberated for several hours. Neuronal Reverberation Occurs in Multiple Forebrain Areas In order to assess the anatomical distribution of experience-dependent neuronal reverberation, we performed the neuronal ensemble correlation analysis for each area separately ( Figure 3 A). At first glance, differences between pre- and postnovelty correlation profiles were noticeable in all animals, with predominant effects in a different subset of areas for each animal. For example, rat 4 showed marked reverberation in the HP, but small changes in the CX, while rat 5 showed just the opposite. In the majority of the recorded sites, postnovelty traces (red in Figure 3 A) run above prenovelty traces (black in Figure 3 A), but the reverse also occurs, suggesting some sort of antireverberation. The most widespread reverberation was seen in rat 1, which showed sustained reverberation in the CX and decaying reverberation in the HP and TH. A somewhat similar pattern was seen in rat 5, while rat 3 showed strong reverberation only in the PU, and rat 4 in the HP and TH. Rat 2 showed the least reverberation of all, with somewhat stronger effects in the PU. Figure 3 Long-Lasting Neuronal Reverberation Occurs in the CX, HP, PU, and TH (A) Temporal profiles of neuronal ensemble correlations in all recording sites. Gray bands indicate lights-off; white bands indicate lights-on. Red and black traces indicate post- and prenovelty correlations, respectively. (B) Temporal evolution of p values associated with pre- and postnovelty Bonferroni comparisons for individual brain areas, calculated as in Figure 2 C. Color bar in linear scale; black denotes p > 0.05. Minimum p values (MIN P) in crescent “rat number” order, as follows: CX: 10 −10 , 10 −4 , 10 −5 , 10 −9 , 10 −9 ; HP: 10 −12 , not significant, 10 −22 , 10 −3 (both red and blue scales); PU: 10 −2 , 10 −12 , 10 −15 (blue scale) and 10 −17 (red scale), 10 −16 (red scale) and 10 −34 (blue scale); TH: 10 −18 , 10 −3 (blue scale), not significant, 10 −30 , 10 −16 . (C) Neuronal ensemble correlations for no-contact templates (taken from epochs within the novel stimulation period in which animals had no sensory contact with the novel objects) also show enhanced neuronal reverberation. Despite the interanimal and interarea differences in the magnitude and shape of correlation profiles, significant changes between pre- and postnovelty correlations were observed in all areas studied (CX, five of five rats; HP, three of four rats; PU, four of four rats; and TH, four of five rats; Bonferroni comparison, p < 0.05). Indeed, experience-dependent changes were not statistically different across different forebrain areas (ANOVA, F = 0.24, d.f. 3, p = 0.86). The temporal evolution of p values (Bonferroni comparison) associated with single-area correlation profiles shows that significant reverberation was present in 16 of 18 recording sites for several hours after exposure to novel stimulation ( Figure 3 B). It also confirms that neuronal ensemble reverberation (post-/prenovelty correlations) is not the only kind of experience-dependent change possible. Some animals showed significant long-lasting antireverberation (pre-/postnovelty correlations), i.e., patterns of activity that were statistically more dissimilar from novel stimulation templates than expected by chance. Antireverberation (indicated by blue hues in Figure 3 B) occurred in the HP (one of four rats), PU (two of four rats), and TH (one of five rats), but not in the CX. Single-area postnovelty average correlations showed peaks of the order of 0.4 (rat 5, PU; Figure 3 A), but typically ranged from 0.1 to 0.2. Therefore, high-fidelity replay of neuronal firing patterns was not observed even when single areas were considered. An intriguing observation came from the scrutiny of no-contact templates of neuronal activity, sampled within the novel stimulation 1-h period during alert WK, but excluding moments of contact between whiskers and objects. Surprisingly, no-contact templates yielded correlation profiles that were almost indistinguishable from those obtained when animals had tactile contact with the novel objects ( Figure 3 C). Therefore, the exploration of the novel environment enhanced the reverberation of all the neuronal activity patterns concomitant with the experience and not just of those corresponding to moments in which animals received sensory inputs from the objects. This rules out the possibility that stimulus complexity, rather than novelty, is the underlying cause of the enhanced neuronal reverberation observed after exploration of the objects. Neuronal Reverberation Peaks during SW Sleep Single-area results indicate that neuronal ensemble correlations often peak during discrete epochs that last a few hours. We also noticed marked oscillations of the correlation trace in several recorded sites (e.g., rat 5, CX). These observations suggest that some underlying biological process, with slow evolution but with sharp phase transitions, governs the long-term reverberation of neuronal firing patterns. To test whether transitions in the wake–sleep cycle could amount for these effects, we investigated how experience-dependent changes in neuronal correlations varied across the three major rat behavioral states: WK, SW sleep, and REM sleep. A comparison across states of post-/prenovelty correlation ratios calculated from averages of entire recordings indicated a significant state-specific effect (ANOVA, F = 9.289, d.f. 2, p = 0.0004), with SW ratios being significantly higher than those of both WK (Bonferroni comparison, p < 0.05) and REM (Bonferroni comparison, p < 0.003). Indeed, significant state-specific differences in post-/prenovelty correlation ratios were individually detected in four of five animals (ANOVAs, d.f. 2: rat 2, PU, F = 4.13, p = 0.039; rat 3, CX, F = 6.45, p = 0.026; rat 4, HP, F = 3.99, p = 0.029; rat 5, CX, F = 13.81, p < 0.0001). The mean correlation values found in those recording sites for the three behavioral states reveal that SW correlations were systematically larger than WK correlations ( Figure 4 A). Several other recorded sites displayed similar but nonsignificant trends. Meanwhile, the REM correlations mea-sured were variable and could not be consistently ranked in relation to WK and SW sleep. Comparable neuronal reverberation between SW and REM sleep was observed in only one animal (rat 5, CX). A major effect of SW sleep on neuronal reverberation was corroborated by the temporal evolution of successive state-specific Bonferroni comparison p values calculated for pre- and postnovelty 4-h average correlations across all animals and brain areas studied ( Figure 4 B). The strongest contrast between pre- and postnovelty neuronal correlations was clearly seen during SW sleep, with less effect seen in WK and even less in REM sleep. Figure 4 C depicts the state-sorted pre- and postnovelty correlations for rat 5, CX, illustrating both the general SW effect and the much less-prevalent REM sleep changes. Figure 4 Neuronal Reverberation Depends on Behavioral State (A) Histograms (mean ± SEM) of post- and prenovelty correlation ratios in the recording sites where significant state-related differences in neuronal ensemble correlations were detected. SW sleep post-/prenovelty correlation ratios were significantly higher than WK in all four cases (Bonferroni comparison p values as follows: rat 2, PU, SW>WK 0.013; rat 3, CX, SW>REM 0.017 and SW>WK 0.022; rat 4, HP, SW>REM 0.013 and SW>WK 0.039; rat 5, CX, SW>WK 0.0001 and REM>WK 0.0002). (B) Bonferroni comparison p values for post-/prenovelty comparisons in all animals according to behavioral state and brain area, calculated in intervals of 4 h. Animal order and time as in Figure 3 B. Color bar in linear scale. (C) Neuronal ensemble correlations sorted by state for rat 5 CX. In comparison with WK, there is a clear increase in the contrast between pre- and postnovelty correlations during SW sleep. In this particular animal and brain area, increased correlations were also seen for REM sleep, but this was not the case in other animals (A). Furthermore, this REM effect was substantially weakened when expressed in p values (B), due to the very short duration of REM sleep episodes. In this respect, notice that REM sleep has much fewer datapoints, reflecting the short duration of this state relative to WK and SW sleep. Thus, even in a site where REM sleep showed results similar to SW sleep, the cumulative neuronal reverberation that takes place during REM is necessarily less than that of SW. (D) Statistical comparison of matches between templates of neuronal activity sampled at WK normal speed with a range of targets spanning different temporal scales. Plotted are Bonferroni comparison p values for post- and prenovelty comparisons in all animals according to behavioral state and brain area, calculated in intervals of 4 h for speed factors ranging from 20 times faster to two times slower than the WK normal rate. No evidence for optimization at speeds different from the WK physiological rate (1×) was found. Color scale as in (B). Next, we tested the possibility reported in hippocampal place cells ( Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ) that experience-dependent replay of neuronal firing patterns during sleep can be slower (REM) or faster (SW) than during WK. Template-to-target matches at different speed factors were obtained by comparing 250 ms-binned templates with targets binned within a range of different bin sizes (from 12.5 ms to 500 ms). By temporally compressing and expanding “target” spike records before matching them to templates, we determined the magnitude of neuronal ensemble correlations for speed factors ranging from 0.5 to 20 times the physiological WK processing speed, which covers the reported optimum speed ranges for SW ( Lee and Wilson 2002 ) and REM sleep ( Louie and Wilson 2001 ). A predominance of neuronal reverberation during SW sleep was seen for all speed factors, as indicated by Bonferroni comparisons ( Figure 4 D). However, no significant differences were seen when post- and prenovelty correlation ratios (calculated from averages of entire recordings) were compared across different speed factors (ANOVA, F = 1.496, d.f. 5, p = 0.19). Figure 4 D shows that within any given state or area, neuronal reverberation did not vary systematically with speed factor, and the temporal distribution of correlation hot-and-cold spots was largely insensitive to speed factor. Thus, we found no evidence that forebrain neuronal reverberation can be optimized assuming replay speeds different from the WK normal rate. Indeed, a subtle but consistent decrease of p values can be observed for speed factors 10 times and 20 times faster than normal WK rates, while speed factors near the physiological range (2×–0.5×) show stronger and similar effects. This was the case even in the HP, in contrast with previous findings in hippocampal place cells recorded in highly trained animals performing a spatial navigation task ( Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ). At present, it is unclear whether this discrepancy reflects differences in stimulus familiarity (novel versus habitual), stimulation modality (tactile exploration versus spatial navigation), the very low representation of place cells in our hippocampal samples (less than 5%), or possible differences in the analyses used in previous studies, based on the statistical boot-strapping of relatively small datasets ( Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ). All together, our results indicate that neuronal reverberation was consistently stronger during SW sleep, decreasing during WK. This is remarkably well-illustrated by a superimposition of behavioral state classification and neuronal ensemble correlations (middle panel in Figure 5 A), which reveals an exquisite long-term temporal match between epochs of increased neuronal correlations and SW episodes (red in Figure 5 A). Likewise, neuronal correlation troughs show a tight correspondence with WK episodes (blue in Figure 5 A). This characteristic state-dependency persisted throughout the 45 h of postnovelty recording (see Figure 4 B). Notice that REM sleep only showed SW-like results in one out of five animals (rat 5, CX; depicted in Figures 4 and 5 ). In the remaining animals, REM correlations were either closer to WK levels than to SW levels or in between (see Figure 4 A). Given this marked variability and the very short duration of total REM sleep in comparison with total SW sleep (WK 52%, SW 40%, and REM 8% of total recording time for five animals), this indicates that REM sleep plays a minor role in neuronal reverberation. Figure 5 Neuronal Reverberation Is Strongest during SW Sleep (A) Rat 5 (CX) dramatically illustrates the state dependency of neuronal ensemble correlations, which are strongly increased by SW sleep but readily decreased by WK. The upper panel shows the firing rates of 38 cortical neurons for approximately 45 h after exposure to novel stimulation (indicated by an asterisk). The middle panel shows the superimposition of successive neuronal ensemble correlations and concurrent behavioral states. Nearly all correlation peaks correspond to SW episodes, while almost all troughs match WK epochs. The lower panel represents pooled LFP forebrain coherence ( Amjad et al. 1997 ) over time, useful to discriminate between WK (strong coherence above 25 Hz and weak coherence under 5 Hz) and SW sleep (the opposite). Notice that in this particular example (rat 5, CX), REM episodes show correlations similar to those of SW sleep, but, as shown in Figure 4 , this was the exception and not the rule across several animals. (B) State-dependent neuronal reverberation was sustained throughout the recording period, as shown by segments representing the beginning (3,200–3,300 min), middle (4,700–4,800 min), and end (5,200–5,250 min) of the experimental record. In the upper panel, notice the progressive increase of neuronal correlations across single SW sleep episodes. (C) Blow-up of two selected data segments indicated by asterisks in (A). Despite having being sampled from moments of high neuronal firing rates (asterisk), novel stimulation templates reverberate most strongly during SW sleep when firing rates are low (single asterisk and double asterisks). The high firing rates that characterize WK correspond to decreased neuronal reverberation, probably due to sensory interference. Interestingly, a comparison of the correlation temporal profile with the concurrent neuronal firing record (upper panel in Figure 5 A) reveals that SW correlation peaks correspond to periods of decreased firing rate, while WK correlation troughs match epochs of increased neuronal activity. This is better shown in Figure 5 C, which depicts data segments approximately 2-h long, comprising the three major behavioral states studied. The first segment (shown by a single asterisk in Figure 5 A) corresponds to 60 min of novel stimulation and the immediately ensuing sleep–wake cycles, while the second segment (shown by double asterisks in Figure 5 A) illustrates sleep–wake episodes occurring approximately 15 h after the original experience. Thus, although novel stimulation templates of neuronal activity were selected from WK episodes characterized by high firing rates, ensuing reverberation of these neuronal firing patterns was most pronounced during SW sleep, under lower firing rates. Discussion In order to assess several objections to the replay hypothesis for memory consolidation during sleep, we conducted long-term continuous neuronal recordings on animals subjected to a naturalistic behavioral paradigm, which involved multiple novel sensory and spatial cues. Our results indicate that large-scale neuronal firing patterns generated during the exploration of novel objects can recur for several hours after the reference experience throughout most of the forebrain, while firing patterns associated with familiar stimulation (i.e., the walls of the recording box) are substantially less detectable over time. Our results fend off three major objections to the notion that neuronal reverberation during sleep may underlie memory consolidation. First, significant experience-depen-dent changes in neuronal ensemble correlations can be tracked as late as 48 h after the reference novel experience, being therefore compatible with memory impairment effects of sleep deprivation applied hours or days after training ( Fishbein 1971 ; Pearlman and Becker 1974 ; Smith and Butler 1982 ; Smith and Kelly 1988 ; Karni et al. 1994 ; Stickgold et al. 2000 ; Fenn et al. 2003 ). Second, these effects were observed in rats completely naïve with respect to the reference stimuli, ruling out the possibility that only the performance of highly trained behaviors would be followed by neuronal reverberation. Third, neuronal ensemble correlations were significantly enhanced during SW sleep and decreased during WK, while REM sleep produced variable results. The data indicate that novel experience caused sustained neuronal reverberation ( Hebb 1949 ) rather than discrete reactivation ( Wilson and McNaughton 1994 ; Kudrimoti et al. 1999 ), in the sense that reverberation decreased, but did not disappear, during WK (see Figure 4 C). The consistent increase in neuronal reverberation during SW sleep, the high interanimal variability of neuronal reverberation during REM sleep, and the small contribution of REM sleep to total sleep time indicate that the cognitive effects of experience-dependent neuronal reverberation ( Gerrard 2002 ; Peigneux et al. 2003 ) must be largely attributed to SW sleep. Therefore, our results suggest a major role for SW sleep in the reverberation of new memory traces. An important result of the present study is the inverse correlation between neuronal correlations and concurrent firing rates. Although all neuronal activity templates were taken from epochs of high arousal WK when firing rates were generally high, their neuronal reverberation during subsequent WK was not very prominent (see Figures 4 and 5 ). In contrast, reverberation of the same activity templates peaked during SW sleep, when the firing rates of forebrain neurons are generally low (see Figures 4 and 5 ). This suggests that reverberating patterns of neuronal activity associated with past novel experience are largely—but not completely—masked during WK by incoming sensory inputs unrelated to the reference experience. By the same token, peak neuronal correlations arise during SW sleep, when sensory interference ceases. Taken together, these observations corroborate the notion that the importance of sleep for memory consolidation stems from the offline processing of memory traces, i.e., from the absence of sensory interference ( Jenkins and Dallenbach 1924 ; Melton and Irwin 1940 ; Winson 1985 ). Were there differences in average firing rate before and after exposure to the novel environment, and could such differences account for the effects seen in neuronal correlations? The neuronal ensemble correlation method ( Louie and Wilson 2001 ) involves a normalization of firing rates after binning, and therefore it is insensitive to moderate changes in the mean firing rate. In our experiment, firing rates for individual neurons varied substantially during exposure to novelty, with some neurons firing more and other neurons firing less than before exposure. This variability caused a moderate but nonsignificant increase in the average activity of the cells so that the firing rates within novelty templates were on average approximately 10% higher than the neuronal firing rates of preexposure templates. Neuronal firing rates increased during exposure to novel objects and persisted elevated for up to 1 h after removal of the objects, returning to baseline afterwards. This contrasts with the timecourse of neuronal correlations changes, which were increased for up to 48 h. Finally, as explained above, neuronal reverberation was inversely correlated with firing rates. Thus, mean firing-rate differences were not responsible for the reverberation effect, which should rather be attributed to specific firing-rate relationships across multiple neurons. Our results impose some clear constraints on future sleep and learning theories. First, no sign of neuroanatomical specificity was found in the correlations measured, and in particular no significant differences between hippocampal and extrahippocampal areas could be detected. Despite considerable interanimal variability in the magnitude of the correlations observed in the different brain structures, statistically significant neuronal reverberation produced by novel stimulation was observed in 16 out of 18 recorded brain sites, comprising the CX, HP, PU, and TH. This broad forebrain reverberation was related to the free exploration of four novel and complex objects, placed in four well-separated places and including the presence of novel food. Thus, novel experience involving tactile, gustatory, olfactory, spatial, and motor components is able to engage multiple forebrain structures, all similarly capable of reverberating neuronal patterns of activity after novel stimulation. Second, neuronal ensemble correlations measured across the forebrain were typically small (on the order of 0.1–0.3), agreeing with values previously reported for pairwise ( Wilson and McNaughton 1994 ; Skaggs and McNaughton 1996 ; Qin et al. 1997 ; Hoffman and McNaughton 2002 ) or many-neuron ( Louie and Wilson 2001 ) correlations. Qualitatively similar results were observed for bins ranging from 5 ms to 1,000 ms, with higher correlation values for larger bin sizes. This suggests that neurons of multiple forebrain areas, once exposed to novel experience, do not accurately replay prior WK activity patterns longer than 5 ms. Instead, they show a mild but long-lasting bias towards (or against) the reference activity patterns. Indeed, not a single template-to-target match (out of 979,200 matches sampled) yielded correlation values higher than 0.45, indicating that novelty-induced neuronal reverberation occurs at low fidelity. It has been proposed that a high-fidelity replay of neuronal firing patterns during sleep may be achieved assuming that replayed patterns can undergo time compression and expansion ( Nadasdy et al. 1999 ; Louie and Wilson 2001 ; Lee and Wilson 2002 ). We assessed this possibility thoroughly, but found no evidence of such effects in any of the forebrain sites recorded. Thus, in the face of consistently low neuronal correlation values, the “high-fidelity replay hypothesis” for timeperiods larger than 5 ms should be rejected, at least in mammals ( Dave and Margoliash 2000 ). It remains to be seen whether more precise patterns of spike-to-spike correlations may reverberate in intervals smaller than 5 ms. A third important point regards the observation that neural activity sampled when animals were aroused, but not touching the objects, yielded neuronal reverberation that was nearly identical to that obtained when animals made sensory contact with the objects. This indicates that the kind of experience-dependent neuronal reverberation detected by the neuronal ensemble correlation method ( Louie and Wilson 2001 ) does not reflect the specific features of the stimuli, but is related to the overall behavioral salience of the novel stimulation period, i.e., is context- rather than stimulus-specific. In principle, these results are compatible with a slow and progressive process of memory consolidation ( Bryson and Schacher 1969 ), proportional to the novelty of the experience, and able to bind together a multitude of contextual cues related to its core sensory elements ( Kohler 1947 ). It has been suggested that the neuronal reverberation of newly acquired synaptic changes during SW sleep may lead to the recall and storage of new memories by way of “calcium-mediated intracellular cascades” capable of opening the “molecular gates to plasticity” ( Sejnowski and Destexhe 2000 ). This hypothesis is partially contradicted by evidence that calcium-dependent gene expression related to synaptic plasticity is upregulated during REM sleep ( Ribeiro et al. 1999 , 2002 ), but not during SW sleep ( Pompeiano et al. 1994 ). The present findings and the current literature suggest instead that SW and REM sleep play separate roles on memory consolidation, with memory recall occurring during SW sleep and memory storage taking place during REM sleep. According to this view, the deleterious effects of sleep deprivation on memory consolidation would be a consequence of the disruption of the underlying neuronal reverberation and gene expression during SW and REM sleep, respectively. The fact that neuronal reverberation is sustained for long epochs during SW sleep suggests that unconsolidated synaptic changes may not only be recalled, but also amplified over time during SW sleep. Indeed, a progressive increase of neuronal correlations across single SW sleep episodes was often observed (upper panel in Figure 5 B). A model of how such amplification may arise is presented in Figure 6 . We have recently proposed that the cyclical reiteration of trace amplification during SW sleep and trace storage during REM sleep promotes the postsynaptic propagation of memory traces ( Pavlides and Ribeiro 2003 ), as suggested by the hippocampofugal pattern of gene expression during REM sleep ( Ribeiro et al. 2002 ). Potentially, this propagation could cause memory traces to progressively reach farther and farther away from the original synaptic trajectory activated at initial encoding. Over time, this sleep-dependent propagation could lead to deeper encoding within the CX ( Craik and Lockhart 1972 ; Cermak and Craik 1979 ), as well as hippocampal disengagement ( Scoville and Milner 1957 ; Mishkin 1978 ; Kesner and Novak 1982 ; Squire 1992 ; Izquierdo and Medina 1997 ; Bontempi et al. 1999 ). The notion that the two major phases of sleep play distinct and complementary roles in memory consolidation is in line with evidence that SW and REM sleep have synergistic effects on human procedural learning ( Mednick et al. 2003 ). Put in historical perspective, our model argues that sleep separately harbors both mechanisms postulated in the past ( Hebb 1949 ) to be necessary for memory consolidation: postacquisition neuronal reverberation and structural synaptic plasticity. In conclusion, sustained neuronal reverberation during SW sleep, immediately followed by plasticity-related gene expression during REM sleep, may be sufficient to explain the beneficial role of sleep on the consolidation of new memories. Figure 6 Conceptual Model of the Role of Sleep for Memory Consolidation Arrows indicate pathway activation by sensory inputs during WK, as well as intrinsic brain activity such as pontine waves during sleep ( Datta 2000 ); different arrow sizes indicate different magnitudes of pathway activation. Red indicates calcium-dependent pretranscriptional processes, with different hue intensities representing the progressive amplification of recently acquired synaptic changes. Green indicates plasticity-related transcriptional regulation. The initial state of the model (data not shown) consists of environmental habituation, during which ongoing activity patterns only repeat themselves by chance. (First panel) A novel WK experience encodes a memory trace across multiple forebrain areas, selectively activating functionally related synapses. This triggers calcium-dependent pretranscriptional cascades (red) and plasticity-related gene expression (green) that lead to the common strengthening of the activated synapses. (Second panel) The continuation of WK involves a succession of unrelated sensory experiences capable of producing interference, i.e., a progressive weakening of recently encoded synaptic changes. (Third panel) Upon entering SW sleep, intrinsic brain activation is biased towards previously potentiated synapses, causing neuronal firing patterns originally produced during the novel WK experience to reverberate significantly above chance levels. (Fourth panel) The periodic activation of calcium-dependent second-messenger cascades by large-amplitude SW oscillations may result in the progressive amplification of the synaptic changes that encode the novel memory trace. (Fifth panel) SW-amplified synaptic changes are stored during REM sleep by way of plasticity-related transcriptional regulation. Materials and Methods Chronic neuronal recordings Multielectrode arrays ( Figure S3 ) were surgically implanted according to National Institutes of Health (NIH) guidelines in five adult male Long–Evans rats (250–300 g). The following coordinates in millimeters relative to Bregma ( Paxinos and Watson 1997 ) were used to center the arrays: HP (+2.8 anteroposterior [AP], +1.5 mediolateral [ML], −3.3 dorsoventral [DV]), CX (+3.0 AP, +5.5 ML, −1.5 DV), PU (−1.0 AP, +2.5 ML, −5.0 DV), and TH (+3.0 AP, +3.0 ML, −5.0 DV). Hippocampal data pool together neurons recorded with staggered electrodes from the CA1 field and the dentate gyrus. Locations of implants were histologically verified by comparing cresyl-stained frontal brain sections with reference anatomical planes ( Paxinos and Watson 1997 ) ( Figure S4 ). A Multineuron Acquisition Processor (128 channels; Plexon Inc., Dallas, Texas, United States) was used to perform recordings of neuronal spikes and LFPs, as previously described ( Nicolelis et al. 1999 ) (Figures S5 and S6 ). A waveform-tracking technique involving periodic template adjustment was employed for the continuous recording of individual units (see Figure S5 ). Units that showed nonstationary waveforms, unstable firing-rate profiles, or both were discarded. The stability of firing rates within each behavioral state can be appreciated in the upper panel in Figure 5 . Behavior Before the beginning of the experiment, animals were individually habituated to an empty recording box for 5–7 entire days (12 h:12 h light:dark schedule, lights on at 06:00 a.m.) so as to reach steady-state behavioral activity and baseline wake–sleep cycles. Behaviors were continuously recorded by way of two infrared-sensitive CCD video-cameras; infrared illumination was used to monitor behavior when visible lights were off. Behavioral states were coded as WK, SW sleep, or REM sleep based on a spectral analysis of LFPs and visual inspection of videotaped behaviors, according to previously described criteria ( Ribeiro et al. 1999 ) ( Figure S7 ). Data analysis Data were processed and analyzed by custom-made MATLAB TM (MathWorks, Natick, Massachusetts, United States) code running in a computer cluster comprising 32 CPU (Evolocity TM , LNXI, Sandy, Utah, United States). Neuronal ensemble correlations were calculated following the method in Louie and Wilson (2001 ). In brief, a stretch of data (target) is scanned for similarity to certain multineuron temporal patterns of activity (templates). Templates consist of CxN matrices corresponding to simultaneously recorded spike trains of C neurons binned into N intervals of a certain length. Our results were obtained using 9 s-long templates binned with 250 ms-long bins with as many neurons as were available for the brain areas in question. Hence, each template pattern yielded a Cx36 matrix. Target data comprising 24 h or 48 h were sparsely sampled (every 30 s) and binned into matrices of the same dimensions. For a C × N template data matrix x and a target data matrix y sampled at time t , the ensemble correlation index C t is obtained by: where (the same applies to Y c , y¯, and σ y ). Templates were selected by careful scrutiny of behavior recorded on videotapes and of the corresponding spectral characteristics of hippocampal LFPs, so as to assure sampling during alert WK ( Figure S7 ). Particular care was used to select template sets with a comparable prevalence of θ (5–8 Hz) over δ (2–4 Hz) frequencies (θ/δ hippocampal spectral ratios, ≥10). All templates and targets were individually sampled without overlap. Because binning may destroy higher-frequency phenomena, control analyses were run to assess the consequences of binning the data with bins of different sizes. Results obtained with bin sizes in the interval of 5–1,000 ms are qualitatively equivalent to the ones obtained using 250 ms bins ( Figure S8 ). In order to detect temporally expanded or contracted repetitions of template patterns, a range of larger- or smaller-than-template bin sizes, respectively, was used when binning each target sample (12.5 ms, 25 ms, 125 ms, 250 ms, 375 ms, and 500 ms). This allowed us to compare targets spanning different temporal scales with templates of neuronal activity sampled at WK speed. The bin size range used allowed the detection of patterns temporally compressed by factors of 20, 10, and 2 (bin sizes of 12.5 ms, 25 ms, and 125 ms), temporally expanded by factors of 1.5 and 2 (bin sizes of 375 ms and 500 ms) as well as replayed at the same speed (250 ms). StatView TM (SAS, Cary, North Carolina, United States) and MATLAB TM software were used for descriptive statistics and hypothesis testing. Supporting Information Figure S1 Objects Four different objects were used to produce CSS. (9.7 MB PPT). Click here for additional data file. Figure S2 Behavior All animals were highly habituated to the recording box, so that exposure to novel complex objects caused a general increase in the animals' arousal. Four of five animals showed an increase in time spent in WK with respect to SW and REM sleep during CSS (A), as compared to adjacent pre- and postnovelty periods of equal length (60 min). The only exception was rat 1, which showed nevertheless a marked exploratory drive, spending nearly 20% of the exposure period in direct whisker-contact with the objects (B). Individual object preferences were moderately varied, as indicated in (C). (746 KB PPT). Click here for additional data file. Figure S3 Multielectrode Arrays Teflon-coated tungsten wires (50 μm diameter, 300 μm between wires, 1.0–1.2 MΩ at 1 KHz; California Fine Wire Company, Grover Beach, California, United States) were assembled in multielectrode arrays shaped to fit different neuroanatomical targets. (282 KB PPT). Click here for additional data file. Figure S4 Location of Implants Frontal brain sections stained for cresyl-violet were used to determine the sites of electrode placement. Electrode tracks, tissue scars, and reference electrolytic lesions performed a few days before sacrifice were used to delimit the implant sites, indicated in red in the figure below. Numbers on the right represent standard AP coordinates ( Paxinos and Watson 1997 ) in millimeters from Bregma. (6.3 MB PPT). Click here for additional data file. Figure S5 Neuronal Recordings To record neuronal activity, differentiated neural signal was preamplified (2,000×–32,000×) and digitized at 40 KHz. Up to four neuronal action potentials per recording channel were sorted online (SortClient 2002, Plexon Inc.) and validated by offline analysis (Offline Sorter 2.3, Plexon Inc.) according to the following cumulative criteria: voltage thresholds greater than two standard deviations of amplitude distributions; signal-to-noise ratio greater than 2.5 (as verified on the oscilloscope screen); less than 1% of interspike intervals smaller than 1.2 ms; and stereotypy of waveform shapes, as determined by a waveform template algorithm and principal component analysis. In order to continuously record individual neurons for up to 96 h, we used an adaptive algorithm (available on SortClient 2002, Plexon Inc.) that adjusts waveform templates based on the recent accumulated mean shapes (1% of midline every 20 min). This allows for the same neuron to be tracked across consecutive days, as verified by the superimposition of waveforms acquired thoughout the experiment (Wavetracker software, Plexon Inc.). (3 MB PPT). Click here for additional data file. Figure S6 Neuronal Yield Up to 159 neurons were recorded from three to four different brain areas. (1.4 MB PPT). Click here for additional data file. Figure S7 Recording LFPs and Behaviors LFPs were recorded in parallel with spikes from the same electrodes. Neural signals were split, preamplified (1,000×), and filtered (0.5–400 Hz) by way of a Plexon LFP board. Signals were then fed to the MAP acquisition principal component through a NIDAQ card and digitized at 500 Hz. Behaviors were constantly recorded in videotape by two diametrically opposed infrared-sensitive CCD cameras (model WV-BP332, Panasonic, Laguna, Philippines). A millisecond-precision timer (model VTG-55, For-A Company, Tokyo, Japan) was used to synchronize the acquisition of spikes, LFPs, and videotape records. Behavioral states were identified by the combined inspection of videotapes and the spectral content (1–20 Hz) of LFPs. Behaviors were classified according to the following criteria: (1) alert WK: active exploration with whisking, plus strong hippocampal θ rhythm; (2) quiet WK: stillness or grooming, with eyes open and low-power hippocampal θ rhythm; (3) SW sleep: stillness with eyes closed, plus large-amplitude hippocampal δ rhythm; (4) REM sleep: overall stillness with intermittent whisking, eyes closed, strong hippocampal θ rhythm. The inspection of videotape records readily separates alert and quiet WK from sleep states, but the separation of SW and REM sleep relies strongly on LFP analysis. Hippocampal LFP is particularly useful to disambiguate SW and REM sleep: SW sleep has a strong δ band (2–4 Hz), while REM sleep shows increased θ band (5–8 Hz). The distinction between alert and quiet WK was used only for template selection (all templates taken from alert WK). For all other purposes, alert and quiet WK data were combined into a single WK category. The graphs depict the θ/δ hippocampal spectral ratios (mean ± SEM) of the three major behavioral states for rat 5 (entire recording). (2.1 MB PPT). Click here for additional data file. Figure S8 Bin Size Exploration The figure shows the effect of using different bin sizes to calculate neuronal ensemble correlations. We observed quantitative differences (the larger the bin size, the larger the correlations), but qualitatively the correlations profiles are equivalent, i.e., have very similar shapes. (404 KB PPT). Click here for additional data file.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314474.xml
535811	The rehydration transcriptome of the desiccation-tolerant bryophyte Tortula ruralis: transcript classification and analysis	Background The cellular response of plants to water-deficits has both economic and evolutionary importance directly affecting plant productivity in agriculture and plant survival in the natural environment. Genes induced by water-deficit stress have been successfully enumerated in plants that are relatively sensitive to cellular dehydration, however we have little knowledge as to the adaptive role of these genes in establishing tolerance to water loss at the cellular level. Our approach to address this problem has been to investigate the genetic responses of plants that are capable of tolerating extremes of dehydration, in particular the desiccation-tolerant bryophyte, Tortula ruralis . To establish a sound basis for characterizing the Tortula genome in regards to desiccation tolerance, we analyzed 10,368 expressed sequence tags (ESTs) from rehydrated rapid-dried Tortula gametophytes, a stage previously determined to exhibit the maximum stress induced change in gene expression. Results The 10, 368 ESTs formed 5,563 EST clusters (contig groups representing individual genes) of which 3,321 (59.7%) exhibited similarity to genes present in the public databases and 2,242 were categorized as unknowns based on protein homology scores. The 3,321 clusters were classified by function using the Gene Ontology (GO) hierarchy and the KEGG database. The results indicate that the transcriptome contains a diverse population of transcripts that reflects, as expected, a period of metabolic upheaval in the gametophyte cells. Much of the emphasis within the transcriptome is centered on the protein synthetic machinery, ion and metabolite transport, and membrane biosynthesis and repair. Rehydrating gametophytes also have an abundance of transcripts that code for enzymes involved in oxidative stress metabolism and phosphorylating activities. The functional classifications reflect a remarkable consistency with what we have previously established with regards to the metabolic activities that are important in the recovery of the gametophytes from desiccation. A comparison of the GO distribution of Tortula clusters with an identical analysis of 9,981 clusters from the desiccation sensitive bryophyte species Physcomitrella patens , revealed, and accentuated, the differences between stressed and unstressed transcriptomes. Cross species sequence comparisons indicated that on the whole the Tortula clusters were more closely related to those from Physcomitrella than Arabidopsis (complete genome BLASTx comparison) although because of the differences in the databases there were more high scoring matches to the Arabidopsis sequences. The most abundant transcripts contained within the Tortula ESTs encode Late Embryogenesis Abundant (LEA) proteins that are normally associated with drying plant tissues. This suggests that LEAs may also play a role in recovery from desiccation when water is reintroduced into a dried tissue. Conclusion The establishment of a rehydration EST collection for Tortula ruralis , an important plant model for plant stress responses and vegetative desiccation tolerance, is an important step in understanding the genome level response to cellular dehydration. The type of transcript analysis performed here has laid the foundation for more detailed functional and genome level analyses of the genes involved in desiccation tolerance in plants.	Background The cellular response of plants to water deficits has both economic and evolutionary importance directly affecting plant productivity in agriculture and plant survival in the natural environment. Ramanathan [ 1 ] has argued, based on predictions of global environmental changes, that developing crops which are more tolerant to water deficits while maintaining productivity, will become a critical requirement in the early part of the 21 st century. Understanding how plant cells tolerate water loss is a vital prerequisite for developing strategies for improving tolerance of, and biomass/seed production under drought conditions. In the last decade the genes induced by water-deficit stress have been successfully enumerated in plants that are relatively sensitive to cellular dehydration, in particular Arabidopsis thaliana [ 2 - 6 ]. In addition, the mechanisms by which the genetic response to water deficit is controlled by abscisic acid (ABA)-dependent and independent pathways have also been extensively elucidated [ 6 , 7 ]. However, even with the recent addition of in-depth examination of gene expression patterns using Arabidopsis microarrays [ 8 , 9 ] we have little functional knowledge of the genes that respond to water deficits. Of critical importance is the question of which of the genes identified as responding to water deficits actually have an adaptive role in establishing tolerance, in particular tolerance of cellular dehydration, and which are genes that are only responding to the injury incurred by the imposition of the stress. Injury may induce, or repress, specific genes that are not involved in promoting adaptation to cellular dehydration. An indication that at least some of the genes have an adaptive function in dehydration tolerance derives from the observation that they are expressed in tissues that acquire desiccation tolerance, the extreme manifestation of dehydration tolerance, such as in maturing seeds and in leaves of desiccation-tolerant plants during drying [ 3 , 10 , 11 ]. However, in order to fully address the question of the adaptive importance of genes involved in responses to cellular dehydration it is necessary to gain an evolutionary perspective of the involvement of a gene in dehydration tolerance mechanisms. To this end we have established an ongoing comparative genomics program to study the genetic responses to dehydration in species that span the evolution of dehydration (desiccation) tolerance mechanisms within the land plants. Here we present an analysis of ESTs derived from the desiccation-tolerant bryophyte Tortula ruralis [Hedw.] Gaertn. Meyer & Scherb., that is representative of the primitive genetic strategy for the acquisition of desiccation-tolerance [see [ 12 , 13 ]]. Desiccation tolerance, the ability to recover from the almost complete loss (90%) of protoplasmic water, is a phenomenon common in the reproductive structures of green plants: pollen, spores and seeds. However, the ability to survive vegetative desiccation is a demonstrable but uncommon occurrence in the plant kingdom [ 13 - 18 ]. Within the flowering plants there are only approximately 300 species of flowering plants that are known to tolerate vegetative desiccation [ 16 , 17 ]. Recent physiological phylogenetic analyses indicate that vegetative desiccation tolerance was primitively present in the bryophytes (the basal-most living clades of land plants), but was lost in the evolution of tracheophytes. Desiccation-tolerant bryophytes are found worldwide and occupy a variety of ecological niches, most of which could, during some period of the year, be considered as extreme either on a macro or microhabitat level. In most cases the extremes that these plants experience are both in water availability and temperature [ 19 - 23 ]. Desiccation-tolerant bryophytes, because of their simple architecture, have few, if any, morphological (or indeed physiological) characteristics or adaptations that can limit water loss or regulate plant temperature. As a result of this, the internal water content of their photosynthetic tissues rapidly equilibrates to the water potential of the environment once free water is lost from the surface of the plant. This in turn means that these plants experience drying rates that are much faster than those experienced by their more complex pteridophyte or angiosperm counterparts. In fact, the drying rates that desiccation-tolerant bryophytes experience are generally lethal to desiccation-tolerant ferns and flowering plants [ 14 ]. The rapid equilibration of protoplasmic water potential with that of the environment in bryophyte tissues appears to demand a type of desiccation tolerance that is significantly different from that exhibited by desiccation tolerant angiosperms [ 15 ]. Rather than acquiring desiccation tolerance in response to a dehydration event as seen in Craterostigma plantagineum , Sporobolus stapfianus , and other desiccation-tolerant angiosperms, desiccation-tolerant bryophytes appear to express this trait constitutively [ 15 , 24 ]. This form of desiccation tolerance is considered the most primitive of those that have received attention so far [ 13 ]. In this type of tolerance the primary response to a desiccation event, at least at the level of gene expression, occurs after the fact, during the first hour or two following rehydration. This has led to the suggestion that a major component of the mechanism of desiccation tolerance in bryophytes is a rehydration-induced cellular repair response [ 15 , 24 ]. The implication is that although cellular protection and hence desiccation tolerance is constitutive, it is not sufficient to prevent some damage from occurring (or being manifested) upon rehydration, and thus repair processes are needed and induced when water returns to the protoplasm of the cells. The repair aspect of the mechanism of desiccation tolerance in these plants, although demonstrated to be a major component of tolerance, is difficult to detail and characterize. Most work has focused on the proteins whose synthesis is induced immediately upon rehydration of desiccated gametophytic tissue. Early work [ 25 ] established the ability of T. ruralis and other mosses to rapidly recover synthetic metabolism when rehydrated. The speed of this recovery was inversely dependent upon the rate of prior desiccation: the faster the rate of desiccation, the slower the recovery. In addition, although the pattern of protein synthesis in the first two hours of rehydration of T. ruralis is distinctly different from that of hydrated controls, novel transcripts were not made in response to desiccation [ 26 ]. Hence it was suggested that T. ruralis responds to desiccation by an alteration in protein synthesis upon rehydration that is in large measure the result of a change in translational control. Changes in transcriptional activity were observed for nearly all transcripts studied [ 27 ] but did not result in a qualitative change in the transcript population during desiccation or rehydration. It thus appears that T. ruralis relies more upon the activation of pre-existing repair mechanisms for desiccation tolerance than it does on either pre-established or activated protection systems. In a detailed study of the changes in protein synthesis initiated by rehydration in T. ruralis , Oliver [ 26 ] demonstrated that during the first two hours of hydration the synthesis of 25 proteins is terminated, or substantially decreased, and the synthesis of 74 proteins is initiated, or substantially increased. Controls over changes in synthesis of these two groups of proteins, the former termed hydrins and the latter rehydrins, are not mechanistically linked. It takes a certain amount of prior water loss to fully activate the synthesis of rehydrins upon rehydration. RNA blots revealed that several rehydrin transcripts accumulate during slow drying [ 28 , 29 ] at a time when it is assumed that transcriptional activity is rapidly declining. These transcripts do not accumulate during rapid desiccation, nor is their accumulation during slow drying associated with an increase in endogenous ABA accumulation. ABA is undetectable in this moss [[ 30 ], M. J. Oliver, unpubl data], and T. ruralis does not synthesize specific proteins in response to applied ABA. The accumulation of these transcripts was postulated to be the result of an increase in mRNA stability brought about by the removal of water from the cells [ 27 ]. Recent studies clearly demonstrate that these transcripts are sequestered in the dried gametophytes in mRNP particles [ 29 ] and that this results in the change in their stability. The implication from this work is that the sequestration of mRNAs required for recovery hastens the repair of damage induced by desiccation or rehydration and thus minimizes the time needed to restart growth upon rehydration. The major question arising from these studies concerns the identity of rehydrins and what possible functions and roles they may play in regards to the response to desiccation and rehydration and to desiccation tolerance per se . We have some limited knowledge of the functions (or postulated functions) of a few of the rehydrins from classical molecular analyses, [ 31 , 32 ] and from a small-scale EST collection analysis [ 33 ]. However, there is an obvious need to extend this base and to develop testable hypotheses that will help us to elucidate the metabolic and genetic mechanisms that control the recovery and repair of dried plant cells and their role in the development and evolution of desiccation tolerance. To gain an appreciation of the number of possible rehydrin genes and the range of possible functions encompassed by their expression we have initiated a genomics level analysis of gene expression during the recovery of Tortula gametophytes from the desiccated state. The first step in this process was to establish an EST collection that is representative of the transcripts available to the moss during the first few hours following rehydration. In this report we present a bioinformatic analysis of 10, 368 ESTs from the early phases of recovery following rehydration of rapidly-dried Tortula ruralis gametophytes; rapid dried gametophytes were chosen in order to maximize the recovery and repair response upon rehydration. The bioinformatics approach we have taken is based on that used by McCarter et al., [ 34 ] to conduct a comprehensive analysis of 5,700 Meloidogyne incognita L2 ESTs, and includes cluster analyses, transcript abundancy estimations, and functional classifications based on InterPror domains, Gene Ontology hierarchy, and KEGG biochemical classifications. The overall goal was to gain an appreciation of the Tortula transcriptome during the time period following rehydration when the desiccation driven alteration in gene expression is at its peak. During this period we hypothesize the processes of cellular repair and recovery are the main focus of the metabolism of the gametophytic cells. Results and discussion Ten thousand three hundred and sixty eight individual cDNA clones were selected from a Tortula ruralis rehydration library and subjected to single-pass 5' directional sequencing to generate 10,368 primary ESTs of which 9,159 (88%) passed through quality control, vector trimming, E. coli contamination, and cloning artifact removal. The 9,159 ESTs averaged 648 nucleotides in length and totaled 5.93 million nucleotides submitted to Genbank. These submitted ESTs form the basis of the subsequent transcriptome analysis utilizing the High Throughput-Gene Ontology-Genome Annotation Toolkit (HT-GO-GAT), a software developed by S.E. Dowd (unpublished). Cluster analysis Utilizing the assembly algorithms incorporated in the SeqManII software (part of the DNASTAR suite from DNASTAR Inc, Madison WI), the 9,159 ESTs were grouped into contigs and clusters by establishing assembly stringencies that generated groupings defined by the analysis protocols of McCarter et al., 2003 [ 34 ]. Contigs contain EST members that appear to originate from single transcripts whereas clusters are assemblies of ESTs that could represent transcripts from the same gene but alternate splice isoforms, or in the case of Tortula ESTs, which derive from a population of individual gametophytes, alleles of the same gene. The 9,159 ESTs formed 7,272 contigs and 5,563 clusters, both of which exhibit an average size of 669 nucleotides. However, the longest sequence increased from 1,575 nucleotides for contigs to 2,229 nucleotides for clusters. Clusters varied in content from a single EST (singletons) in 4,362 cases (78%) to 48 ESTs for a single cluster (Figure 1 ). The elimination of redundancy during contig building and cluster formation reduced the total number of nucleotides for further analyses from 5.93 million to 4.87 million (contigs) and 3.71 million (clusters). Overall the 9,159 ESTs potentially represent 5,563 genes, a discovery rate of 60%, with 47.6% of the ESTs as singletons. This is an overestimation of gene discovery since several non-overlapping clusters can represent a single gene and from our blast search data this appears to be a possibility, at least in the case of clusters 121 and 204 that appear to independently represent a gene that has a weak similarity to the LEA protein of Caenorhabditis elegans . Even though the ESTs derive from a non-normalized library 96.5% of the clusters still have 5 or fewer EST members. Figure 1 Histogram of distribution of ESTs by cluster size. Transcript abundance The consensus Cluster sequences were subjected to a BLASTx style search of a custom curated non-redundant database derived from UNIPROT and annotated according to the degree of similarity of the cluster to the highest scoring match in the database of known gene sequences. The degree of similarity was based on the quality of the BLASTx statistical outputs as well as a visual inspection of the aligned sequences between the query and the target. Clusters that generated an HSP-bit score below 70 and or E values higher than 10 -7 were assessed manually for all possible alignments taking into account the alignment length, number of identical matches, gaps, and positive replacements. Utilizing these criteria we were able to annotate 3,321 (59.7 %) of the 5,563 clusters by their similarity to known genes within the database. This also meant that 2,242 of the clusters, or 40.3%, represent sequences that have no known counterpart in the public databases and that we categorize as unknowns. Table 1 lists the 30 most abundant EST clusters derived from the Tortula rehydration EST collection. These transcripts only account for 8.4% of the generated ESTs, however. Ten of the most abundantly represented transcripts encode proteins that do not match any of the sequences in the databases searched in this study and are designated as unknowns. Seven abundant transcripts appear to encode proteins that belong to a class of proteins known as Late Embryogenesis Abundant (LEA) proteins, although the relatively low HSP bit scores and high E-values for most of these BLASTx matches has to be considered as a caveat in this assessment. LEA proteins have long been ascribed a protective role for cells that are experiencing dehydration [ 11 , 35 ]. This is a conclusion drawn on the strong correlation between LEA transcript accumulation and water loss, rapid decline in transcript levels upon rehydration, and especially as LEA gene expression relates to the programmed desiccation stage of seed maturation. If one can conclude that LEA protein transcripts are abundant in the rehydration transcriptome of Tortula ruralis , which is consistent with some of our earlier findings [ 32 ], then it is also possible that these proteins are involved in either protection of cellular integrity during the initial phases following rehydration when cell disruption is apparent in bryophytes [ 36 ], or are actively involved in the restoration of cells damaged by a desiccation event. This difference in the response of LEA gene expression between Tortula and what has been reported for angiosperms may also be a reflection of what we believe to be a more primitive mechanism of dehydration tolerance, and perhaps a more primitive role for and control of LEA gene expression. Other abundant transcripts appear to fall into the membrane transport (aquaporin, cysteine rich proteins, channel and pore proteins) and proteins that can be associated with plant stress events (metallothionine, esterase, and rubredoxin). The Early Light Inducible Protein A (ELIP-A) transcript is the only member of the abundant transcripts that we have previously reported [ 33 ] as belonging to the group of proteins we have termed, rehydrins [ 31 ]. These proteins have been suggested to be synthesized in response to stress-induced photo-damage within the Tortula chloroplast and may play a protective or repair function for the photosynthetic apparatus [ 37 ]. Table 1 Thirty most abundant transcripts in the Tortula rehydration library. Length #ESTs in Cluster Description TrEMBL Accession # Bit Score E-value 962.0 48 Unknown 1102.0 40 ABA-inducible protein WRAB1 (Cold-responsive LEA/RAB-related COR protein) Q9XFD0 63.93 2.88E-09 1000.0 37 Caenorhabditis elegans CE-LEA O16527 75.87 6.51E-13 889.0 35 Unknown 1000.0 34 TspO tryptophan rich sensory protein homologue Q92ZA7 127.87 1.43E-28 994.0 34 Unknown 975.0 32 Hydrophobic LEA-like protein (Oryza sativa) Q9ZRF8 100.91 1.84E-20 964.0 32 Unknown 1201.0 31 Cysteine rich protein (ion transport) Q24977 50.45 3.86E-05 1051.0 28 Late embryogenesis abundant (LEA) protein 76 (Brassica napus) P13934 51.60 1.42E-05 683.0 28 Cysteine-rich non-metallothionein-protein Q24774 47.75 1.05E-04 1078.0 26 Early light-inducible protein ELIPA (Tortula ruralis) Q8RYB6 404.06 1.24E-111 956.0 25 Rb7 (Fragment)-MIP/Aquaporin O04179 215.70 3.62E-55 732.0 25 Stress-inducible membrane pore protein Q93Z87 88.58 6.21E-17 1043.0 24 Ribosomal-protein-alanine acetyltransferase Q8IAE2 116.32 4.73E-25 942.0 24 Unknown 902.0 22 Caenorhabditis elegans CE-LEA O16527 50.83 2.02E-05 1043.0 21 Protein DR1172-LEA type 1 family Q9RV58 55.45 1.01E-06 984.0 20 Unknown 661.0 20 Core protein (Pisum sativum) amino acid-selective channel protein Q41050 79.34 3.13E-14 514.0 20 Pyrus pyrifolia Metallothionein-like protein Q9LUX2 57.38 7.04E-08 1316.0 19 Hypothetical protein K08H10.2a Q9XTH4 92.82 8.12E-18 996.0 19 Unknown 982.0 18 Putative late embryogenesis abundant protein (Arabidopsis) Q9LF88 53.91 2.63E-06 910.0 18 Unknown 790.0 18 Gb\|AAF26109.1 (Hypothetical protein) Arabidopsis Q9FFJ0 125.56 5.31E-28 1717.0 17 Lanatoside 15'-O-acetylesterase precursor (Foxglove) O82681 165.62 1.39E-39 1000.0 17 Unknown 1000.0 17 P0577B11.21 protein (Oryza sativa) Rubridoxin-like Q84SC5 60.08 3.85E-08 957.0 17 Unknown Although transcript abundance may reflect the metabolic or physiological needs of the moss during the rehydration phase of a wet/dry/wet cycle it would be more desirable to know how these transcripts are recruited and utilized by the translational machinery to make the proteins that actually contribute to the recovery of the gametophytes following rehydration. Such information is beyond the scope of this analysis but the identification and isolation of the Clusters that represent the Tortula rehydration transcriptome does represent the first major step for such a pursuit. Functional classification of transcripts Physiological, biochemical, and molecular data all point towards an active period of cellular activity, presumably for repair and recovery from desiccation induced damage, during the first two hours following rehydration of dried gametophytes [ 13 , 15 ]. The identity of the more abundant transcripts (Table 1 .) does provide some insight into the nature of the metabolic activity that is associated with rehydration, at least it gives an indication of what metabolic processes may be of prime concern as the plant recovers from desiccation. However, the functional classification of the Tortula transcripts present during the initial phases following rehydration using the Gene Ontology (GO) classification system paints a broader view of the possible metabolic activity of the gametophytic cells at this time. This does, of course, come with the understanding that these are transcript based analyses and do not directly reflect protein levels which would offer a more definitive assessment of the metabolic capability of the cells during rehydration. Functional classification of the Tortula rehydration transcripts was achieved by matching the Tortula clusters to characterized protein domains in a combined protein database, using HT-GO-GAT (Materials and Methods), which allowed us to assign GO terms to each cluster. The assignment of GO terms made it possible to place the clusters into the GO hierarchy which can be viewed by use of an AmiGo browser. Of the 3, 321 clusters that exhibited significant similarity to known genes in the public databases, 2,203 (66% of annotated or 40% of all clusters) represent genes that contain conserved protein domains that have known biochemical and physiological functions in other organisms and map to the GO hierarchy. The GO representations for the Tortula rehydration clusters are presented in Tables 2 through 4 . The representations are segregated into the three main organizing principles of GO: biological process (Table 2 ), cellular component (Table 3 ), and molecular function (Table 4 ). A complete listing of the GO mappings is available on our website at . Table 2 Gene ontology (GO) mappings : Biological processes 1673 clusters Categories and subcategories Representation % of Total Cellular Processes 409 24.5 Cell Communication 75 4.5 Signal Transduction 68 4 Cell adhesion 6 0.5 Cell Death 13 1 Cell Growth & Maintenance 334 20 Organization & Biogen 31 2 Cell proliferation 21 1.5 Transport 286 17 Development 25 2 Physiological Processes 1641 98 Cell Growth & Maintenance 334 20 Death 13 1 Metabolism 1345 80 Alcohol metabolism 61 4 Amine metabolism 83 5 Amino Acid metabolism 91 5.5 Aromatic cpd metabolism 51 3 Biosynthesis 498 30 Carbohydrate metabolism 154 9 Carbon utilization 19 1 Catabolism 160 9.5 Coenzyme & Prosthetic grp 38 2 Electron transport 162 9.5 Energy pathways 115 7 Heterocycle metabolism 19 1 Lipid metabolism 55 3 Nitrogen metabolism 8 0.5 Nucleic acid metabolism 174 10 One carbon cpd metabolism 10 0.5 Organic acid metabolism 102 6 Oxidative phosphorylation 13 1 Oxygen and ROS metabolism 26 1.5 Phosphorous metabolism 111 7 Pigment metabolism 7 0.5 Protein metabolism 536 32 Secondary metabolism 16 1 Sulfur metabolism 10 0.5 Vitamin metabolism 8 0.5 Photosynthesis 73 4 Response to Endogenous Stimuli 16 1 Response to External Stimuli 74 4 Perception 14 1 Response to abiotic 22 1.5 Response to biotic 45 2.5 Response to Stress 61 3.5 Response to DNA damage 16 1 Response to oxidative stress 23 1.5 Response to pest/pathogen 8 0.5 Response to water deprivation 6 0.5 Secretion 3 Table 3 Gene ontology (GO) mappings: Cellular component 982 clusters Categories and subcategories Representation % of Total Cell 963 98 Intracellular 711 72 Cell Cortex 6 0.5 Chromosome 9 1 Cytoplasm 563 57 Cytoskeleton 34 3.5 Cytosol 27 2.5 Endoplasmic reticulum 29 3 Translation elongation cplx 6 0.5 Golgi apparatus 12 1 Mitochondrion 70 7 Plastid (Chloroplast) 122 12.5 Ribosome 224 23 Nucleus 109 11 Ribonucleoprotein complex 233 24 Ribosome 224 23 Membrane 368 37.5 Endomembrane system 5 0.5 Organelle Inner membrane 40 4 Integral to membrane 159 16 Unclassified 149 15 H+-transporting ATPase 10 1 Mitochondrial membrane 43 4 Organelle Outer membrane 11 1 Plasma membrane 5 0.5 Extracellular 7 0.5 Table 4 Gene ontology (GO) mappings: Molecular function 1992 clusters Categories and subcategories Representation % of Total Antioxidant activity 6 0.5 Binding 758 38 Amino acid binding 6 0.5 Carbohydrate binding 6 0.5 Lipid binding 6 0.5 Metal ion binding 161 8 Calcium 43 2 Magnesium 18 1 Transition metal 78 4 Nucleic Acid binding 296 15 DNA binding 125 6 Nuclease activity 15 1 RNA binding 58 3 Translation factor, nucleic acid 54 3 Nucleotide binding 335 17 Purine nucleotide (adenyl/guanyl) 334 17 Protein binding 25 1.5 Catalytic Activity 1162 58.5 Helicase activity 25 1.5 Hydrolase activity 374 19 acting on acid anhydrides 108 5.5 acting on carbon-nitrogen (not peptide) 5 0.5 acting on ester bonds 71 3.5 acting on ether bonds 7 0.5 acting on glycosyl bonds 39 2 Peptidase activity 97 5 Isomerase activity 48 2.5 Kinase activity 137 7 Protein kinase activity 90 4.5 Lipase activity 55 3 Lyase activity 93 4.5 Oxireductase (Ored) activity 264 13.5 disulfide Ored activity 17 1 Monooxygenase activity 28 1.5 Ored activity – CH-OH donors 55 3 Ored activity – acting on NADH/NADHP 13 0.5 Ored activity-acting on paired donors 13 0.5 Ored activity – peroxide acceptor 27 1.5 Ored activity – single donor 17 1 Ored activity – sulfur group donors 16 1 Oredactivity – CH-NH 2 group donors 9 0.5 Small protein conjugating enzyme activity 11 0.5 Transferase activity 310 15.5 transferring acyl groups 30 1.5 transferring alkyl or aryl groups 11 0.5 transferring glycosyl groups 35 2 transferring nitrogenous groups 12 0.5 transferring one–carbon groups 38 2 transferring phosphorous containing grps 154 7.5 Chaparone activity 45 2.5 Defense protein activity 7 0.5 Enzyme activator activity 6 0.5 Unknown molecular function 50 2.5 Nutrient reserve activity 8 0.5 Signal Transducer activity 58 3 Receptor activity 41 2 Two-component response regulator 9 0.5 Two-component response sensor molecule 18 1 Structural Molecule activity 261 13 Structural component of ribosome 239 12 Transcription Regulator activity 41 2 Transcription factor 27 1.5 Two-component response regulator 9 0.5 Translation Regulator activity 54 3 Nucleic acid binding 54 3 Transporter activity 286 14 amine/polyamine transport 8 0.5 carbohydrate transport 11 0.5 Carrier activity 109 5.5 electrochemical potential driven 29 1.5 primary active transporter 81 4 Channel/pore class transporter 25 1 Electron transport 61 3 Ion transport 57 3 anion 5 0.5 cation 49 2.5 metal ion 16 1 Organic acid transport 11 0.5 Protein transport 46 2.5 Of the 2,203 clusters, 1673 (76%) map into the Biological Processes classification, 1641 (98%) of these fall into the Physiological Processes category and 409 (24.5%) cross-map into the Cellular Processes category (Table 2 ). Within physiological processes 80% of the clusters were associated with metabolism and 20% cell growth and maintenance (it is within this group that most of the overlap occurs with the Cellular Processes category). The distribution is not surprising for ESTs (clusters) derived from a tissue that is harvested at a time of metabolic upheaval such as rehydration and recovery from the desiccated state. In support of this notion are the almost identical distributions of ESTs observed for cDNA collections from protonemal tissues of the moss Physcomitrella patens following various hormonal treatments designed to illicit developmental perturbations and metabolic switching (Table 5 , and from data reported by Nishiyama et al., 2003 [ 38 ]). Table 5 Comparison of GO mappings for Tortula ruralis and Phycomitrella patens Relationship level Gene ontology ISC P.pat ISC T.rur T:P 1 9981 % 2203 % 2 biological process 7119 71.33 1673 75.94 1.1 3 physiological processes 6901 69.14 1641 74.49 1.1 4 metabolism 5601 56.12 1345 61.05 1.1 5 biosynthesis 1672 16.75 498 22.61 1.3 5 carbohydrate metabolism 642 6.43 154 6.99 1.1 5 carbon utilization 44 0.44 19 0.86 2.0 5 catabolism 670 6.71 160 7.26 1.1 5 energy pathways 306 3.07 115 5.22 1.7 5 oxidative phosphorylation 28 0.28 13 0.59 2.1 5 oxygen and ROS metabolism 76 0.76 26 1.18 1.6 4 photosynthesis 138 1.38 73 3.31 2.4 5 photosynthesis, dark reaction 21 0.21 9 0.41 2.0 5 photosynthesis, light reaction 56 0.56 41 1.86 3.3 4 response to external stimulus 232 2.32 74 3.36 1.4 5 perception of external stimulus 37 0.37 14 0.64 1.7 5 response to abiotic stimulus 75 0.75 22 1.00 1.3 5 response to biotic stimulus 136 1.36 45 2.04 1.5 4 response to stress 231 2.31 61 2.77 1.2 5 response to oxidative stress 57 0.57 23 1.04 1.8 5 response to water deprivation 8 0.08 6 0.27 3.4 4 cell growth and/or maintenance 1536 15.39 334 15.16 1.0 3 cellular process 1845 18.49 409 18.57 1.0 4 cell communication 310 3.11 75 3.40 1.1 5 signal transduction 254 2.54 68 3.09 1.2 4 cell growth and/or maintenance 1536 15.39 334 15.16 1.0 5 cell organization and biogenesis 222 2.22 31 1.41 0.6 5 cell proliferation 219 2.19 21 0.95 0.4 5 transport 1131 11.33 286 12.98 1.1 3 development 119 1.19 25 1.13 0.9 2 cellular component 4290 42.98 982 44.58 1.0 3 cell 4137 41.45 963 43.71 1.1 5 cell wall 73 0.73 7 0.32 0.4 4 intracellular 2909 29.15 711 32.27 1.1 5 cytoplasm 2041 20.45 563 25.56 1.2 5 extrachromosomal DNA 37 0.37 15 0.68 1.8 5 nucleus 815 8.17 109 4.95 0.6 5 ribonucleoprotein complex 590 5.91 233 10.58 1.8 5 thylakoid 116 1.16 46 2.09 1.8 4 membrane 1628 16.31 368 16.70 1.0 5 mitochondrial membrane 104 1.04 43 1.95 1.9 5 inner membrane 100 1.00 40 1.82 1.8 5 outer membrane 28 0.28 11 0.50 1.8 2 molecular function 8755 87.72 1992 90.42 1.0 3 binding 3530 35.37 758 34.41 1.0 4 nucleic acid binding 1389 13.92 296 13.44 1.0 5 DNA binding 649 6.50 125 5.67 0.9 5 nuclease activity 89 0.89 15 0.68 0.8 5 RNA binding 330 3.31 58 2.63 0.8 5 translation factor activity 176 1.76 54 2.45 1.4 4 nucleotide binding 1478 14.81 335 15.21 1.0 4 protein binding 221 2.21 25 1.13 0.5 4 metal ion binding 616 6.17 161 7.31 1.2 3 chaperone activity 202 2.02 45 2.04 1.0 4 heat shock protein activity 57 0.57 17 0.77 1.4 3 signal transducer activity 233 2.33 58 2.63 1.1 4 two-component sensor molecule activity 38 0.38 18 0.82 2.2 4 receptor activity 167 1.67 41 1.86 1.1 5 transmembrane receptor activity 30 0.30 17 0.77 2.6 4 structural constituent of ribosome 553 5.54 239 10.85 2.0 3 transcription regulator activity 234 2.34 41 1.86 0.8 4 transcription factor activity 150 1.50 27 1.23 0.8 3 translation regulator activity 176 1.76 54 2.45 1.4 4 translation factor, nucleic acid binding 176 1.76 54 2.45 1.4 5 translation elongation factor activity 67 0.67 33 1.50 2.2 5 translation initiation factor activity 97 0.97 20 0.91 0.9 3 transporter activity 1115 11.17 286 12.98 1.2 4 carbohydrate transporter activity 49 0.49 11 0.50 1.0 4 carrier activity 414 4.15 109 4.95 1.2 4 channel/pore class transporter activity 68 0.68 25 1.13 1.7 5 alpha-type channel activity 59 0.59 23 1.04 1.8 3 catalytic activity 5294 53.04 1162 52.75 1.0 4 isomerase activity 183 1.83 48 2.18 1.2 5 intramolecular isomerase activity 41 0.41 10 0.45 1.1 4 kinase activity 735 7.36 137 6.22 0.8 4 lyase activity 352 3.53 93 4.22 1.2 4 oxidoreductase activity 960 9.62 264 11.98 1.2 4 transferase activity 1557 15.60 310 14.07 0.9 4 hydrolase activity 1764 17.67 374 16.98 1.0 The subcategory distributions within the Physiological and Cellular processes also seem to reflect the nature of the cellular disturbances that result from a desiccation-rehydration event. Processes involved in metabolite and ion transport within and between cells are represented by 17% of the clusters that map to Biological Processes, and almost 86% of those that map to the Cell Growth and Maintenance subcategory. Within the Metabolism subcategory of Physiological processes 40% (32% of total) of the clusters map to protein metabolism (synthesis) and 37% (30% of total) map to biosynthetic processes. The considerable representation within the aforementioned three subcategories is consistent with much of our biochemical and physiological evidence concerning the metabolic activity and emphasis in gametophytic cells during rehydration [ 15 , 24 , 25 ]. In particular, following rehydration the protein synthetic machinery is rapidly reconstituted, having been dismantled during drying, to direct the synthesis of pattern of proteins termed rehydrins that appear to be a crucial aspect of the desiccation tolerance mechanism of Tortula ruralis (as described above). The importance of the protein synthetic machinery and its re-establishment following rehydration is also highlighted by the preponderance of clusters associated with the Ribosomal subcategory of the Cellular Component Classification; 40% of the Cytoplasm category (23% overall). Only 982 clusters (45% of the 2,203 that constitute the mapped population) map within the Cellular Component Classification and are presumably associated with structural functions (Table 3 ). Of these 982 clusters, almost all map to the Cell classification within which 72% map to Intracellular components and 38% to the Membrane category. Within these categories, representation is most significant in the ribosomal, integral membrane protein, and plastid subcategories. As discussed above the importance of protein synthesis during recovery from desiccation tolerance may explain the preponderance of clusters associated with ribosomal structural components as well as the number of clusters that are associated with the membrane and plastid subcategories. Ultrastructural studies of dried and rehydrated gametophytes clearly indicate that the inrush of water during rehydration disrupts membranes and causes a disorganization of the internal granal structures and swelling of the large chloroplasts of the Tortula leaf cells [ 36 , 39 ]. Of the clusters that can be mapped into GO hierarchies, 90% can be ascribed molecular functions (Table 4 ). Of the major categories, Binding activity (38%), Catalytic activity (58.5%), Structural Molecule activity (13%), and Transporter activity (14%) are best represented in the cluster collection. Metal ion, nucleic acid, and nucleotide binding are the most represented subcategories within the Binding activity category perhaps reflective of the need for biosynthetic and repair activity associated with rehydration of moss cells. Within the Catalytic activity category the majority of the clusters are associated with Hydrolase (19%), Transferase (15.5%), Oxireductase (13.5%), and Kinase (7%) activities. Almost half of the clusters associated with the Transferase activity subcategory map as transferring phosphate-containing groups. Each one of these subcategories represent catalytic activities that could be argued as important for a cell to recover from a major metabolic perturbation such as that seen during rehydration. The significant representation within the Kinase and phosphate transfer categories also suggests an active metabolic control "program" occurs when the desiccated cells receive water and attempt to recover from the damage. It is also intriguing that there is a significant representation within the Transporter subcategory as little is known of this group within the context of desiccation tolerance in bryophytes, although solute (osmolytes) and sugar transport in and out of the vacuoles of desiccation tolerant Angiosperms [ 3 ] do occur during desiccation and rehydration. Of particular interest to this study are clusters that represent gene expression control factors both at the transcriptional (41 clusters) and translational level (54 clusters). These clusters, along with those that represent biochemical control mechanisms for signaling and gene expression at the protein level, such as kinases (90 clusters) and phosphate-transfer activities (154 clusters), may represent critical elements in the activation and execution of the cellular recovery processes necessary for the mechanism of desiccation tolerance exhibited by Tortula ruralis . This is of importance because of the key position of bryophytes in the evolution of desiccation tolerance in plants. Our main hypothesis is that an elucidation of the signaling and activation pathways for the rehydration response in this assumed primitive tolerance mechanism could have major implications for the study of stress tolerance mechanisms in all plants and thus these clusters represent important targets for further study at the molecular and biochemical levels. GO based comparison with Physcomitrella patens EST collections The representation of the Tortula rehydration clusters throughout the GO mapping system is indicative of the emphases on, but not expression levels of, particular cellular activities represented in the moss gametophytes during this period of a wet and dry cycle. In an attempt to assess if the accents on particular cellular activities indicated for rehydrated gametophytes are characteristic of the rehydration induced metabolic state or are simply indicative of processes associated with normally active bryophyte cells, we compared the representation of the Tortula rehydration clusters within the GO categories with similar "clusters" from Physcomitrella patens , the only other bryophyte that has similar genomic level information. The majority of the Physcomitrella "clusters" are derived from a large EST collection representing transcripts from both untreated and hormone induced (to switch developmental pathways) cells of protonemal cultures. The Physcomitrella ESTs are described by Nishiyama et al., [ 38 ] and were obtained from Physcobase as assembled contigs (assembled in an identical fashion to what we designate as clusters). In total 22,885 Physcomitrella contigs, derived from 102553 ESTs obtained from Physcobase and Genbank, were subjected to a BLASTx search, as described for the Tortula clusters using HT-GO-GAT. Of the 22,885 contigs 9,981 (43.6%) represent genes that contain conserved protein domains that have known biochemical and physiological functions in other organisms and map to the GO hierarchy. There are two caveats for this comparison; 1) differences in representation may simply reflect species differences in the emphasis on individual classes of cellular activities between Tortula and Physcomitrella , or 2) differences in representation may reflect differences in the emphasis on individual classes of cellular activities between mature gametophytes ( Tortula ) and protonema ( Physcomitrella ). Until a comparison of GO mapping distributions can be made directly between clusters derived from an EST collection from hydrated control gametophytes from Tortula with those from the rehydration collection these caveats remain important. Nevertheless, even with these difficulties and limitations the comparison is still useful for developing new hypotheses as to what cellular processes might be important in the recovery of moss cells from desiccation. In general the distribution of the Tortula clusters within the GO mappings are similar if not identical to the distribution of the Phycomitrella contigs gleaned from Physcobase. This can be seen especially at the GO map 2 and 3 relationship-level categories (Table 1 and supplemental material) using the multi-species GO browser (DrZOOview2.0 ). However, there are several notable differences and in general the differences are consistent with what has been determined, in earlier studies, to be important in the recovery of the moss from desiccation following rehydration [ 4 , 23 ] and what is known about plant responses to abiotic stress in general [ 8 , 13 ]. The differences evident in the comparison are presented in Table 5 , where the extent of the similarity in the distribution of the clusters are expressed as the ratio of representation for Tortula to representation for Physcomitrella (T:P). In the Biological category the most striking differences in GO representation of Tortula clusters occurs in categories where the percentage representations are relatively low but the numbers of clusters are substantial. In the level 4 relationship, Photosynthesis, the representation for Tortula is 3.31% compared to 1.38% for Physcomitrella , a 2.4 fold difference, the majority of which is accounted for by the difference in the representation levels for the light reaction category. The rapid recovery of photosynthesis is critical for the recovery of bryophyte cells, particularly in regards to the production of energy and reducing power for the metabolic activity associated with repair and reconstitution of the gametophytic cells. Chloroplast structure in Tortula is severely disrupted, especially if desiccation occurred rapidly, in the first few hours following rehydration [ 36 , 40 ] but recovers quickly. The difference in representation in this category is consistent with these observations and thus may reflect the greater need for a supply of a diversity of photosynthetic components in rehydrated Tortula than in Physcomitrella cells that have not experienced a disruption in the photosynthetic apparatus. Similar inferences can be made concerning the differences in representation observed for carbon utilization (T:P of 2.0) and oxidative phosphorylation (T:P of 2.1) for Tortula in that mitochondrial activity and integrity are also compromised during rehydration [ 25 , 36 ]. Other differences in representation between Tortula and Physcomitrella mappings relate to a higher representation for Tortula in the categories that relate to responses to external stimuli and responses to stress both of which would seem consistent with the emphasis that cellular activity for Tortula would have in comparison to unstressed Physcomitrella cells. In particular the increased representation within the response to oxidative stress is of note, as an elevated protection of cellular integrity from the damaging reactive oxygen species (ROS) typically associated with a desiccation rehydration event is a distinctive component of desiccation tolerant bryophytes when compared to their desiccation sensitive relatives such as Physcomitrella [ 41 ]. The above observations concerning the more extensive representation of clusters within the GO mappings related to organelle function in Tortula are mirrored in the Cellular Component Level 2 classification mappings. In the Cellular Component classification Tortula exhibits an almost two fold difference in representation within categories associated with either the chloroplast or mitochondria, such as Extrachromsomal DNA (T:P of 1.8), Thylakoid (T:P of 1.8), Mitochondrial membrane (T:P of 1.9), and both Inner and Outer membranes (T:P of 1.8). In this case the categories are generally related to genes representing membrane components and since it is the organelle membranes that exhibit the majority of the damage during desiccation and rehydration it is consistent that these categories would be better represented in the Tortula cluster mappings than those for Physcomitrella . In addition to the organellar related classifications the Tortula clusters also exhibit a higher representation within the Ribonucleoprotein complex category (T:P of 1.8) which in all likelihood reflects an emphasis on ribosomal components since a similar difference in representation is seen in the Structural Constituent of the Ribosome category (T:P of 2.0) within the Molecular Function Level 2 classification. Again such differences in representation in the comparison between Tortula and Physcomitrella GO mappings are consistent with our previous studies on the responses of Tortula gametophytes to desiccation and rehydration and comparisons to non-stressed bryophyte tissues. Protein synthesis is critical to the recovery of Tortula cells following a desiccation event [ 11 , 13 , 15 ] not only for the synthesis of proteins damaged by the stress of desiccation but also directing the response to the stress at the level of gene expression [ 26 , 29 ]. Early studies determined that the speed at which desiccation occurred has a marked effect on both the rate of recovery of protein synthesis and the rate at which either new ribosomes are formed or pre-existing ones are repaired [ 42 , 43 ], rapid desiccation results in a more prolonged recovery of normal protein synthetic levels and also slows the reconstitution of ribosomes upon rehydration. Since the Tortula clusters are derived from ESTs of rehydrated moss that was dried rapidly, the greater representation in ribosome related GO mappings for this collection compared to the Physcomitrella clusters, that represent transcripts from cells where presumably normal ribosomal turnover and synthetic rates are prevalent, is consistent with the biological state of the Tortula cells. The emphasis on protein synthesis in rehydrated Tortula cells compared to those of Physcomitrella is also evident in the comparison of representations within the Translation regulator, Translation factor, and Translation elongation factor activity GO mappings. Other differences seen in the Molecular function classification, such as the greater representation within the Tortula collection of clusters involved in Two-component Sensor or Channel/pore class Transporter activity designated mappings, offer novel possibilities for investigation into rehydration metabolism that have not been indicated as important until now. The individual identity of the clusters that map to these categories should offer possible hypotheses that can be tested in our future research. Functional classification based on KEGG analysis An alternative classification of clusters, based on biochemical function, involves the use of HT-GO-GAT to assign clusters to individual Kyoto Encyclopedia of Genes and Genomes (KEGG ) metabolic pathways. Of the 2203 clusters that map within the GO hierarchy only 642 clusters had assigned EC numbers generating 325 unique mappings (see additional file 1 : KEGG biochemical pathway mappings for Tortula rehydration clusters). The paucity of EC assignments limits this aspect of the analysis but the mapping of clusters to the KEGG metabolic pathways still presents some useful perspectives on the metabolic emphasis of the rehydrated gametophytic cells. Eighty-six of the 123 pathways contained within the Metabolism category (metabolic pathways), were represented by 84.6% of the 642 Tortula clusters. The KEGG metabolic pathways that are well represented by Tortula clusters are Carbohydrate Metabolism (83 enzymes represented), Amino Acid Metabolism (72 enzymes), Energy Metabolism (40 enzymes), Lipid Metabolism (27 enzymes), and Metabolism of Cofactors (24 enzymes). All of these pathways have been previously associated with the cellular recovery processes associated with rehydrated moss gametophytes [ 14 , 15 ]. Within the metabolic activities not represented by the Tortula clusters only the lack of ascorbate metabolizing enzymes appears unusual as ascorbate has been well documented as an important metabolite in the protection of moss gametophytes from oxidative damage during stress. This however appears to be the result of a limitation in the assignment of EC numbers since several Tortula clusters show significant similarities to enzymes involved in ascorbate metabolism in the original BLASTx search used to generate the GO mappings. The limitation of the KEGG based classification can also been seen in the poor representation of Tortula clusters in the other KEGG pathways (Genetic Information Processing, Environmental Information processing, and Cellular Processing) which is surprising given the number of clusters that were identified in the BLASTx search and in the GO databases as being contained within these classifications. As an example, none of the 239 clusters that map in the GO hierarchy as structural components of the ribosome (Table 4 ) or any of the clusters that mapped to transcriptional and translational components were contained in the corresponding KEGG pathways. Thus although useful information can be gleaned from the KEGG classification system and metabolic pathway mappings, especially in practical terms for functional studies using individual clusters, it has some major limitations for drawing any broad based hypotheses from the representation of Tortula clusters within each pathway. ORF based assessment of Tortula clusters Of the 5,563 consensus cluster sequences used in the BLASTx search of our database (see above) 40.3% failed to exhibit sufficient similarity (did not meet set criteria, see above) with known sequences to allow for an accurate annotation of the cluster. It is possible that these contigs, rather than containing novel amino-acid coding regions, contain mainly 3' or 5' untranslated regions (UTRs) or coding regions that are so short as to render them incapable of generating a significant similarity score. In order to investigate this possibility we examined the three classes of contigs, those with significant similarity scores; "good hits", those that generated poor similarity scores; "false hits", and those that failed to generate any scored similarity; "no hits", to determine the longest open reading frame (ORF). We limited the ORF determination to those clusters that contain an AUG codon in the 5' to 3' direction of the clone in any one of the three possible reading frames (the cDNAs were directionally cloned). The results of this analysis are shown in Figure 2 . Of the 5,563 clusters generated in the study, 4,789 generated ORFs under the limitations imposed by the analysis. Of these 2,983 were classified as "good hits", 1,564 as "false hits" and 242 as "no hits". Those clusters that are classified as "good hits" exhibit ORFs that are in general evenly distributed from 20–40 amino acids long to 220–240 amino acids long. The clusters that are classified as "false hits" from the BLASTx search do have a relatively larger proportion of shorter ORFs, in the 20–40 and 40–60 amino acid range but also a substantial proportion that are much longer. The distribution of ORFs in this category ('false hits") does not appear to be sufficiently skewed from that for the "good hit" clusters to render them incapable of generating similarity scores in the BLASTx search. This would suggest that these clusters do contain novel amino acid sequences that are not represented in the public protein databases by sequences sufficiently similar to generate significant HSP Bit scores. In addition, the distribution of ORFs in the "no hit" cluster category are distributed in a similar manner to those of the "good hit" classification and so are also likely to represent clusters encoding proteins with novel amino acid sequences. The clusters contained within the "false hit" and "no hit" categories are of particular interest in our search for novel genes and pathways that are associated with the ability of certain plants to acquire vegetative desiccation tolerance. Figure 2 Distribution of clusters by size of longest ORF (by number of amino acids) . Solid line , clusters that have significant similarity with a known sequence in the database by BLASTX. Dotted Line , clusters that showed weak similarity with a known sequence in the database. Dashed Line , clusters without similarity to known sequences in the database. Conserved gene comparison to Physcomitrella and Arabidopsis In this analysis the Physcomitrella and Arabidopsis databases were independently used in a BLASTx search using the 5,563 Tortula clusters as query sequences. Even though we used the entire 5,563 cluster sequences as individual queries only 3,321 actually represent sequences that could be annotated using the criteria for similarity discussed previously. When the Tortula clusters were used as queries against the Physcomitrella database, 5,554 generated HSP Bit scores however only 282 exhibited a level of conservation of sequence that passed the criteria established for annotation of our Tortula clusters (HSP Bits score above 70 and or E values of 10 -7 or less, see above). Against the Arabidopsis database only 927 Tortula clusters generated HSP Bit scores but 612 had a level of conservation sufficient to be considered reliable matches to Arabidopsis genes. These observations are difficult to rationalize and may simply reflect the inequality of the target databases. The Physcomitrella database, which is generated from a relatively small EST collection, is much smaller than that for Arabidopsis , which is gleaned from the full genomic sequence. This may explain why although most Tortula clusters generated HSP Bit scores only a relative few were capable of producing scores sufficiently high enough, and with low enough error probabilities, for confident assignment of co-identity with a Physcomitrella contig. Obviously Tortula ruralis is more closely related to Physcomitrella patens , as they are both bryophytes, than to Arabidopsis thaliana but the phylogenetic distance between the two bryophytes is still substantial, which may also help explain the paucity of high value matches and the large number of low value hits. The Arabidopsis database is much more comprehensive and would be expected to generate more significant matches as observed and perhaps because of the evolutionary distance between the two plants more queries that do not generate an HSP. The twenty gene products that exhibit the highest level of conservation for both comparisons, Physcomitrella (E values of 0 to -39) and Arabidopsis (E values of 0 to -55) are presented in Table 6 . Several common highly conserved genes are present in both comparisons, including genes involved in cell structure (tubulin and actin), protein synthesis (ribosomal proteins and elongation factors), protein-turnover (polyubiquitins), stress proteins (heat shock), chromosomal proteins (histones), signal transduction (ADP-ribosylation factor), and binding proteins (calmodulin). Interestingly there are differences between the two lists, which may reflect the relative phylogenetic distances between the three species. The three most conserved proteins between Tortula and Physcomitrella did not register as highly conserved between Tortula and Arabidopsis , the most conserved protein between Tortula and Physcomitrella , the rRNA intron-encoded homing endonuclease, did not generate a match at all with an Arabidopsis counterpart. Interestingly, the most conserved photosynthesis related protein between Tortula and Physcomitrella is a chlorophyll a/b-binding protein which generates a HSP Bit score of 205 and E value of 3.0E -53 , the Arabidopsis counterpart generates a HSP Bit score of 82 and E value of 8.9E-15. In this case the level of conservation of this protein within the comparisons appears to reflect the phylogenetic relationships that exist between the three plants. However, the most conserved photosynthesis related protein between Tortula and Arabidopsis is a photosystem I P700 apoprotein A2 which generates a HSP Bit score of 1120 and E value of 0, the Physcomitrella counterpart generates a HSP Bit score of 34 and E value of 1.25E -1 , which appears contrary to what is expected at what is seen for the chlorophyll a/b-binding protein. This type of relationship is also seen for the Ribulose 1 protein (Cluster 2170) that is conserved between Tortula and Arabidopsis but only to a low degree with the Physcomitrella counterpart. If such relationships can be substantiated with further study and complete sequences (both the Tortula and Physcomitrella clusters are based on ESTs) then some interesting evolutionary questions can be raised concerning both gene function and selection pressures on the individual plants and metabolic processes. Table 6 The twenty most conserved Tortula gene clusters between Physcomitrella and Arabidopsis Tortula Cluster Physcomitrella Identity Bit Score e-value Contig_2145 PPCContig_497 Q9AY32 (Q9AY32) RRNA intron-encoded homing endonuclease 813 0.0E+00 Contig_1352 PPCContig_406 Q9AVH2 (Q9AVH2) Putative senescence-associated protein (Fra 630 0.0E+00 Contig_3546 PPCContig_1008 O04892 (O04892) Cytochrome P450 like_TBP (EC 1.14.14.1) 448 1.8E-126 Contig_602 PPCContig_359 polyubiquitin UBQ10/SEN3 413 1.2E-115 Contig_480 PPCContig_12 Q9SPA1 (Q9SPA1) Elongation factor-1 alpha 3 306 1.4E-83 Contig_966 PPCContig_1903 EF1A_VICFA (O24534) Elongation factor 1-alpha (EF-1-alpha) 284 5.8E-77 Contig_1126 PPCContig_2103 Q41067 (Q41067) Polyubiquitin 250 5.9E-67 Contig_1901 PPCContig_1954 Q8H932 (Q8H932) Alpha tubulin 241 7.4E-64 Contig_1538 PPCContig_4431 heat shock protein hsc70-1 (hsp70-1) (hsc70.1) 233 1.6E-61 Contig_1444 PPCContig_49 ADP-ribosylation factor 1 (ARF1), putative 208 4.6E-54 Contig_919 PPCContig_1049 Q9SXW8 (Q9SXW8) Chlorophyll a/b-binding protein 205 3.0E-53 Contig_1381 PPCContig_2023 Q8S173 (Q8S173) Putative 60S ribosomal protein L37a 197 5.3E-51 Contig_5058 PPCContig_951 Q9SWW8 (Q9SWW8) Actin (Fragment) 195 3.3E-50 Contig_1676 PPCContig_2347 calmodulin 193 8.7E-50 Contig_1366 PPCContig_846 Q84NX8 (Q84NX8) Putative ribosomal protein L19 174 8.0E-44 Contig_2369 PPCContig_2689 O04664 (O04664) Small RAS-like GTP-binding protein (AT5G551 173 9.9E-44 Contig_1934 PPCContig_4083 Q9SJB9 (Q9SJB9) Putative translation initiation factor eIF- 171 3.3E-43 Contig_3793 PPCContig_3975 Q9LFN6 (Q9LFN6) DEAD box RNA helicase RH15 169 1.9E-42 Contig_1721 PPCContig_2679 H33_ARATH (P59169) Histone H3.3 165 2.1E-41 Contig_3462 PPCContig_970 Q43821 (Q43821) Ubiquitin conjugating enzyme (EC 6.3.2.19) 160 1.2E-39 Tortula Cluster Arabidopsis Identity Bit Score e-value Contig_3469 ATCG00340 (PSAB) photosystem I P700 apoprotein A2 1120 0.0E+00 Contig_882 AT1G07930. elongation factor 1-alpha (EF-1-alpha) 666 0.0E+00 Contig_602 AT4G05320.4 polyubiquitin UBQ10/SEN3 617 1.0E-175 Contig_2170 ATCG00490 (RBCL) riblose 1 435 5.0E-121 Contig_1984 AT5G02960.1 40S ribosomal protein S23 (RPS23B) 431 5.2E-120 Contig_1538 AT5G02500.1 heat shock protein hsc70-1 (hsp70-1) (hsc70.1) 366 3.2E-100 Contig_1124 AT4G05050.1 polyubiquitin UBQ11 341 5.8E-93 Contig_1444 AT1G23490.1 ADP-ribosylation factor 1 (ARF1), putative 307 1.2E-82 Contig_1901 AT1G50010.1 tubulin alpha-2/alpha-4 chain (TUA2) 303 3.7E-81 Contig_52 AT3G12580.1 heat shock protein hsp70 291 1.3E-77 Contig_1060 AT4G38510.1 probable H+-transporting ATPase 285 4.1E-76 Contig_1125 AT4G02890.2 polyubiquitin (UBQ14) 274 8.3E-73 Contig_3069 AT4G40040.1 histone H3.2 270 1.4E-71 Contig_1935 AT5G08290.1 YLS8, Dim1 homolog 264 1.1E-69 Contig_2698 AT5G45775.2 60S ribosomal protein L11 (RPL11D) 260 1.2E-68 Contig_5058 AT5G09810.1 ACTIN 2/7 (sp\|P53492) 258 6.8E-68 Contig_1676 AT3G43810.1 calmodulin 252 2.8E-66 Contig_3867 AT3G05530.1 26S proteasome AAA-ATPase subunit RPT5a 232 3.4E-60 Contig_347 AT3G11940.2 40S ribosomal protein S5 (RPS5B) 230 2.2E-59 Contig_1052 AT2G29550.1 tubulin beta-7 chain (TUB7) 218 1.3E-55 Conclusions Bryophytes have an important and underestimated place in the study of plant responses to water deficits, in particular desiccation tolerance. Bryophytes occupy what we believe to be one of the most primitive states, along with algae, in the evolution of desiccation tolerance and represent, in all probability, the stage in the emergence of plants from a fresh water environment to occupy the various niches available on dry land [ 13 ]. Unfortunately, it is only with the advent of the development of Physcomitrella patens as a model plant for molecular genetic studies, fired by its particular ability to perform efficient and homologous recombination in vitro, that bryophyte genomics has become a topic of some interest. Physcomitrella is rapidly becoming the model of choice for developmental and transgenic studies [ 44 ] because of its ease of manipulation and indeed there are plans in place to sequence its genome [ 45 ]. Tortula ruralis on the other hand has long been established as an attractive model for the analysis of environmental stress tolerance, in particular desiccation tolerance, It has been a very useful model in assessing structural, physiological, biochemical and genetic (gene expression) aspects of severe dehydration of plant cells and mechanisms by which primitive plants respond to and survive protoplasmic water loss [ 15 , 25 , 46 ]. The progression of the Tortula model into genomics is a critical aspect in the development, along with a transformation system and assessment of its ability for efficient homologous recombination, into a more useful and manipulable model for understanding desiccation tolerance and the nature of extremophiles. In addition to the importance of Tortula for stress biology, the establishment of a second and contrasting bryophyte model, especially with regards to genomics, is essential for the validation and usefulness of the information gained from the analysis of the Physcomitrella genome and the general principles gleaned from its use as a plant model. It is to these ends that we have initiated this study into the transcriptome of Tortula gametophytes as they respond to a major stress event, in this case a combination of desiccation and rehydration. The rehydration transcriptome of Tortula , as defined by the clusters presented herein, is remarkably consistent with what we know about the desiccation response for this bryophyte and its metabolic activity during the first two hours following rehydration. The GO mapping of the Tortula clusters enabled a broad look at what cellular activities appear to be emphasized in the rehydrated gametophytes and in agreement with our previous biochemical analyses highlighted the prominence of the protein synthetic machinery, both in structure and control, membrane structure and metabolism, and the need to reestablish plastid integrity. These observations were bolstered by the comparative GO analysis using the extensive EST collection generated for the desiccation sensitive moss Physcomitrella patens . The GO analysis has also provided fuel for new investigations and hypotheses into the role of other cellular processes, such as membrane transport, phosphorylation and signal transduction, in the mechanisms that enable desiccation tolerance in plants. Signal transduction is especially intriguing with regards to desiccation tolerance in this bryophyte as it appears to rely on alterations in translational control to effect a response to desiccation in contrast to the well characterized transcriptional responses, and associated signaling pathways, associated with abiotic stress in the Angiosperms [ 5 , 7 ]. In addition to the functional based analyses, the simple abundance estimates has also correlated well with previous work and has given further credence to the notion that LEA proteins may also play a role in maintaining cellular integrity when water is reintroduced into desiccated plant tissues. The strong correlation between what is known about the mechanism for desiccation tolerance employed by Tortula ruralis and what can be inferred from the analyses of the Tortula rehydration EST collection gives a measure of confidence not only in the value of a bioinformatics approach to gain a view of a particular transcriptome but also in the basis for new hypotheses and research directions that are generated from them. Although the type of analyses presented here are extremely useful in assessing the types of transcripts present in a particular tissue at a particular time and in response to some perturbation, either external or internal in origin, and generating hypotheses concerning the functional aspects of the transcriptome they are intrinsically correlative in nature. In order to gain a more direct picture of the transcriptome and more importantly, with regards to functional assessments, the "translatome", the bioinformatics must be linked to a detailed expression profile of the transcripts represented within the ESTs generated to investigate the particular biological response or mechanism of interest. To this end the clusters described in this report form the basis of a Tortula gametophyte microarray designed for the profiling of both the extant transcriptome and the associated translatome and their response to desiccation and rehydration under various conditions. The array and the experimental design of the expression profiling will allow us to generate a more accurate assessment of both transcript levels and transcript recruitment into the protein synthetic machinery during, and recovery from, desiccation in Tortula ruralis . In combination with the bioinformatics analyses presented here, the expression profiling will allow us to generate a more complete picture of the cellular response of a tolerant plant species, an extremophile, to an extreme abiotic stress event. Methods Source material Tortula ruralis ([Hedw.] Gaertn, Meyer and Scherb), also classified as Syntrichia ruralis , gemetophytes were collected, harvested, and stored as described previously [ 27 ]. For experimental purposes, gametophyte tissue was hydrated for 48 h to fully recover from dried storage and trimmed to remove stem material. Rapid-dried moss was prepared by placing the cropped gametophytes in a closed atmosphere of 0% relative humidity (RH) on 3-mm filter paper over activated silica gel in a Petri dish. This drying regime resulted in the attainment of the air-dried state within 30 min. The gametophytes remained in this atmosphere overnight to ensure desiccation and prior to library construction were rehydrated for 2 h in deionized water at 18°C in the light. Library construction Total RNA was isolated from the rehydrated gametophytes by a series of phenol extractions as described by Lane and Tumiatis Kennedy [ 47 ]. PolyA RNA was isolated from the total RNA fraction by oligo-(dT) chromatography [ 48 ], using DynaBeads oligo-(dT25) (Dynal, Inc., Lake Success, NY, USA), through two rounds of selection according to the manufacturer's instruction. The purified polyA fraction was used as a template for double-stranded cDNA synthesis using the Superscript Plasmid System (Invitrogen Life Sciences, Carlsbad, CA, USA). The resultant cDNA population was cloned into the pSPORT1 vector, according to manufacturer's instructions, to construct the unidirectional rehydration cDNA library. Small scale sequencing of 384 random clones confirmed the directional aspect of the inserts, the plant nature of the source cDNAs, and the frequency of positive clones. The average insert length for the library was assessed at 1.2 Kb. A subset of 10, 368 randomly picked positive clones (white in a blue-white X-Gal/IPTG based screen) were transferred to individual wells in 384 well plates containing suitable growth medium for storage, replication, and sequencing. Sequence analysis High throughput sequencing of the inserts contained in the 10,368 individual clones was performed using "rolling circle amplification" of the individual plasmids to generate suitable sequencing templates at the Joint Genome Institute, Walnut Creek, CA, U.S.A. Clones were sequenced using primers specific for vector sequence upstream of the multiple cloning site and at the 5' end of the cDNA insert. The sequences were delivered as primary binary files (raw trace files), which were then processed through our sequencing, pipeline as described below. Sequence preparation Trace files were entered into SeqMan II and quality screening performed with medium stringency corresponding to a phred threshold value of 12. Vector searching was performed using the pSPORT™ vector both in forward and reverse orientations with minimum match length of 7, connect distance of 3, Maximum register shift of 10, Minimum NW percent match of 90, gap weight of 0 and length weight of 2. Contaminant screening was performed with minimum match of 25. Following the quality screening the sequences were exported as a single FASTA file and polyA tails were removed using a custom PERL script. Several hundred sequences that contained internal polyA stretches were manually identified. These sequences were manually edited to remove polyA and trailing sequences. Contigs Edited sequences were then entered into SeqMan for assembly into contigs. Assembly for contigs was performed with match size of 50, minimum match percentage of 97, minimum sequence length of 100. Clusters Consensus contig sequences were exported from SeqMan as individual files and entered into a separate SeqMan project and reassembled using the same parameters used for contigs but with a minimum match percentage of 90. These were considered clusters. EST submission EST submission to GenBank was performed using the USDA-ARS Livestock Issues Research Unit's (LIRU) High Throughput-Gene Ontology-Genome Annotation Toolkit (HT-GO-GAT). HT-GO-GAT can be obtained from the LIRU website at . A total of 9159 EST sequences were submitted to GenBank and assigned accession numbers CN200321-CN209479 Functional genetics Functional annotations, Enzyme commission numbers, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway assignments were assigned to cluster sequences using HT-GO-GAT. Consensus sequences derived from Clusters were entered into the software that utilizes custom BLASTx, RPS-BLAST, and relational mySQL databases to identify potential functional assignments based upon sequence and functional domain similarity matching. The software was set to identify high stringency matches and the resulting data manually curated using the high throughput results viewer interface of HT-GO-GAT. HT-GO-GAT produces various reports related to given datasets including ORF analysis, KEGG pathway reports and visualization, and also exports gene ontology association files. These association files were imported into a mySQL database and visualized using the USDA-ARS-LIRU's DrZooView2.0 GO database browser . Authors' contributions MJO conceived of the study, generated the libraries and ESTs, annotated the EST database, performed much of the analyses, interpreted the results, and drafted the manuscript. SED designed the bioinformatics software and database, performed the sequence alignments, performed the ORF analysis, and generated the Bryobase website. JZ coded the HT-GO-GAT software and created the mySQL database. SAM participated in the design of the bioinformatics and design of the sequence analysis. PRP participated in the design of the database and sequence analysis, and participated in the annotation of the individual EST sequences. Supplementary Material Additional File 1 KEGG biochemical pathway mappings for Tortula rehydration clusters . An alternative functional classification of clusters based on established biochemical activities of gene products. The classification is achieved by the use of HT-GO-GAT to assign clusters to individual Kyoto Encyclopedia of Genes and Genomes (KEGG ) metabolic pathways. This file contains the tabulated results of this classification. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535811.xml

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.