Upload batch_13

batch_13/PMC2529298.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2529298",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2529298\nAUTHORS: Kazuyoshi Ikeda, Takatsugu Hirokawa, Junichi Higo, Kentaro Tomii\n\nABSTRACT:\nBackgroundMany studies have examined rules governing two aspects of protein structures: short segments and proteins' structural domains. Nevertheless, the organization and nature of the conformational space of segments with intermediate length between short segments and domains remain unclear. Conformational spaces of intermediate length segments probably differ from those of short segments. We investigated the identification and characterization of the boundary(s) between peptide-like (short segment) and protein-like (long segment) distributions. We generated ensembles embedded in globular proteins comprising segments 10–50 residues long. We explored the relationships between the conformational distribution of segments and their lengths, and also protein structural classes using principal component analysis based on the intra-segment Cα-Cα atomic distances.ResultsOur statistical analyses of segment conformations and length revealed critical dual transitions in their conformational distribution with segments derived from all four structural classes. Dual transitions were identified with the intermediate phase between the short segments and domains. Consequently, protein segment universes were categorized. i) Short segments (10–22 residues) showed a distribution with a high frequency of secondary structure clusters. ii) Medium segments (23–26 residues) showed a distribution corresponding to an intermediate state of transitions. iii) Long segments (27–50 residues) showed a distribution converging on one huge cluster containing compact conformations with a smaller radius of gyration. This distribution reflects the protein structures' organization and protein domains' origin. Three major conformational components (radius of gyration, structural symmetry with respect to the N-terminal and C-terminal halves, and single-turn/two-turn structure) well define most of the segment universes. Furthermore, we identified several conformational components that were unique to each structural class. Those characteristics suggest that protein segment conformation is described by compositions of the three common structural variables with large contributions and specific structural variables with small contributions.ConclusionThe present results of the analyses of four protein structural classes show the universal role of three major components as segment conformational descriptors. The obtained perspectives of distribution changes related to the segment lengths using the three key components suggest both the adequacy and the possibility of further progress on the prediction strategies used in the recent de novo structure-prediction methods.\n\nBODY:\nBackgroundVast amounts of three-dimensional (3D) protein data from structural genomic studies and other individual efforts have been added to our knowledge, thereby enhancing our understanding of protein structures. To date, only two extremes of protein structural data have been studied. One extreme includes local features of proteins: those of short protein segments, typically of 10 residues long or less. The other extreme includes global features of proteins: protein folds or structural domains.Regarding the short protein segments, abundant research examples exist partly because of the existence of variations of methods to analyze the local features of proteins. Various measures, such as RMSDs after structural superposition [1-3], Cα-Cα atomic distances coupled with the torsion angles [4,5], dihedral angles [6], and so on have been used to define the conformational similarity of protein segments. Different clustering techniques, such as k-means clustering [7,8], hierarchical methods [9], competitive learning [6,10], and other methods [11], have been used to describe the organization of the segments' conformational space. The abundance of research results in this area is also partly attributable to various applications of the clustering results of the short segments. A set of representatives from the resulting clusters are often called structural building blocks (SBBs). Even when using different procedures, clustering resolutions of SBBs can be categorized into only a few levels depending mainly on their respective applications, such as structural modeling, verification, comparison, and prediction [6,12]. The most dominant cluster of the short segments, which is common in all studies, corresponds to α-helices, whereas the variability of β-strands is observed at the high-resolution clustering. Regarding global features of proteins, understanding of their organization and analysis of the protein-fold (or structural domain) space studies are progressing well.As reviewed recently [13], both hierarchical and continuous aspects of fold space have been realized. Regarding hierarchical classification, widely used databases such as CATH [14] and SCOP [15] have been constructed. Other databases such as FSSP [16] and VAST [17] have been developed. They are based on continuous measurements of protein structural similarity. Several studies have provided insights into the nature of fold space. Holm and Sander first described the conformational distribution of protein folds in a fold universe with multi-dimensional scaling methods based on an all-on-all comparison using the Dali program [18]. Using the same measurement, Hou et al. [19] showed visual representations of the protein fold universe and identified three major components which characterize the fold space: secondary structure compositions, chain topologies, and the protein domain size.Compared to these two extremes, limited surveys have been done on the conformational space of medium size segments between protein short segments and folds. Specifically, supersecondary structures such as α-hairpin, βαβ-unit, and β-hairpin are typical structural motifs of medium size; those motifs have been analyzed. For example, Salem et al. reported that most superfolds contain a higher proportion of their α-helical or β-strand residues in one such supersecondary structure [20]. Szustakowski et al. built a dictionary of supersecondary structures [21]. Kurgan and Kedarisetti studied regularity among twilight zone protein structures at the level of the sequence segments that correspond to the secondary structure fragments of varying length [22]. However, the organization and statistical properties of the whole conformational space of medium-to-long segments remain unclear. Statistical and systematic analyses should be done on the 'segment universe' from short to long lengths to bridge this gap.Our previous study identified structural clusters and visualized the uneven distribution of short segments in the conformational spaces of 6–22 residues, where known and novel secondary-structure motifs are distributed as isolated clusters [23]. The general features of the segment distribution were consistent for these lengths. However, the question we sought to answer is: Do spaces of long segments differ from those of short segments? In this study, we explore the relationships between the conformational distribution of segments and their length: 10–50 residues, thereby providing a global view of a 'segment universe' and showing critical dual changes (i.e. dual transitions) of the distribution shape in the conformational space of short to long segments. The critical changes might reflect changes of the protein structures' organization. Therefore, the present results suggest the adequacy and the possibility of further progress of the hierarchical treatment used in the recent de novo structure prediction methods. Furthermore, by comparing conformational components among structural classes (i.e., all-α, all-β, α/β, and α+β), we demonstrate the specificity and generality of protein fold classes.ResultsTransitions of segment distribution: short, medium, and long segmentsThe coverage of segments in cluster(s) was calculated as described below. A densely populated region in the 3D principal component analysis (PCA) space was defined as a cluster [23]. Given a density threshold, the segments are classifiable into two groups: those in regions of a density larger than the threshold and those outside the regions. The coverage of segments in clusters is defined as a ratio of the segments in the regions to all the segments.Figure 1a portrays the coverage of segments versus the density threshold for the conformational spaces of 10, 20, 30, 40, and 50 residue lengths. The coverage curves exhibited a transition from concave shapes for short lengths (10 and 20 residues long) to convex ones for long lengths (30, 40, and 50 residues long). Notably, the differences of coverage at a density of 0.2 or less show a transition between the short and long segments. For instance, at a density of 0.1, the coverage is only 16.3% for 10 residues, although the coverage is greater than 50% for 30 residues. In addition, at a density of 0.01, the coverage for 10 residues is 45.6%, although coverage for 30 residues is 91.9%. These quantitatively indicate that the density gradient of the conformational space changes markedly with segment elongation.Figure 1Coverage versus the density threshold. Coverage of segments in clusters versus the density threshold for segment lengths of 10, 20, 30, 40, and 50 residues (A), and those of 21–30 residues (B). Density is presented on a logarithmic scale.Further analyses of the coverage graphs between the short and long segments were meaningful to discover the boundaries of distribution changes. Figure 1b shows coverage curves for lengths of 21–30 residues. The dual and critical transitions, with an intermediate phase for segment lengths of 23–26 residues, can be recognized clearly, as presented in Fig. 1b. The transitions at intermediate length are also characterized by the distributional alteration of the radius of gyration of segments in the populated region with density of 0.10–0.35 (Fig. 2). To adjust the effect of different segment lengths, we defined here the relative score (F_Rg) of the radius of gyration for a segment as (Rg(i,j) - Min Rg(j))/(Max Rg(j) - Min Rg(j)), where Rg(i,j) denotes the radius of gyration of a segment i with length j, and where Max Rg(j) and Min Rg(j) represent the maximal and minimal radius of gyration of the entire segment dataset with length j. Based on these observations, the segment length is categorized into the following three groups: short (10–22 residues), medium (23–26 residues), and long (27–50 residues). We were able to show that changes in the density gradient are associated with distributional alterations in the segment universe in subsequent analyses of visualizing the 3D PCA space. In fact, the difference in the coverage between lengths of 10 and 30 residues was attributable to the increase in the volume for the most populated region, as discussed below. The typical global images of segment universes from the three categories are depicted in Fig. 3d. The segment universes here were generated by the first three principal components derived from the entire segment dataset: PCall1, PCall2, and PCall3 (see Methods).Figure 2Probability density function of the F_Rg score. Distribution of F_Rg score of segments in the region with density from 0.10 (magenta in Fig. 3) to 0.35 (red). Distributions for segment lengths of 21, 23, 25, 27, and 29 residues are shown.Figure 33D representation of segment universes. A 3D representation of short (A, 10 residues), medium (B, 26 residues), and long (C, 30 residues) segment universes. The 3D representations were generated using the first three PCall axes. They were expressed by four iso-density contours: 0.005 (blue), 0.01 (orange), 0.1 (magenta), and 0.35 (red). The PC axis numbers (1, 2, and 3) are given near the axes. Front and side views of images are shown, respectively, in the upper and lower areas of the figure. D: Schematic diagram of the relationship between properties of the cluster(s) in the segment universe and segment length. The properties of the cluster(s) in the segment conformational space changed according to increased segment length.Short length (10–22 residues long)The conformational space of short segments showed a distribution with an extreme density gradient that originated from secondary structure clusters: α-helix and β-strand clusters were discriminated using a density of 0.01 (shown in orange in Fig. 3a). Between the lengths of 10 and 20 residues, spatial arrangements of the segment distribution, especially for α-helical, β-strand, and β-hairpin clusters, were conserved in short conformational spaces. The highly populated core of the α-helix cluster exhibited a density of 0.1 (shown as magenta in Fig. 3a), consisting of completed α-helical segments. The surrounding area of the central region consisted of various types of helical conformations including helix-capping motifs [12]. The central region of the β-strand cluster consisted of fully extended segments that originated mainly from β-sheets and loop regions. The β-hairpin conformations were separated into several clusters at a density of 0.005. Then they were discriminated using the coordinate c2 along PCall2 (see Methods for the definition of c2). The β-hairpin clusters showed a symmetrical relationship related to the N-terminal and C-terminal halves. They were arranged symmetrically around an edge of segment universes of short length.Medium length (23–26 residues long)The segment distribution for medium lengths differed from that for short lengths. The distributional change from short to medium lengths is characterized using a diminishing β-strand cluster and a growing α-helix cluster. The overall distribution was shortened in the direction of PCall1, and enlarged in the direction of PCall2 and PCall3. In the segment universe of 26 residues, the α-helix cluster was discriminated using a density of 0.1 (magenta in Fig. 3b). Interestingly, the shape of the α-helix cluster was a ring (designated as a helix ring cluster). The helix ring cluster that is specific to the medium-length universe consisted not only of the extended α-helices but also of various α-helical conformations, as presented in the inset of Fig. 3b. This cluster included conformations that had originated mainly from all-α, α/β, and α +β proteins (Fig. 4a). The average content of the α-helical residues per segment in the helix ring cluster was about 50% (Fig. 4b); 24.9% of all segments were included within the helix ring cluster. The long-α-helical segments, whose conformation was not compact, were located near the origin of the conformational space (red in Fig. 3b). In contrast, the α-hairpin conformations with a small radius of gyration were located on the opposite side of the position on PCall1. The various α-hairpin conformations with the different turn positions were located symmetrically along PCall2. For medium lengths, the β-strand clusters were diminished because long extended β-strands are rarely found in proteins. The β-hairpin conformations were located symmetrically along PCall2, although the cluster separation of β-hairpins was not clear in medium lengths.Figure 4Relationships between segment properties and their length in the populated regions. Upper figures show that the percentages of structural classes from which segments in the clusters with density of 0.10–0.35 (A) and 0.35–1.0 (C) were derived. Lower figures show the percentages of structural properties of the segments in the clusters with density of 0.10–0.35 (B) and 0.35–1.0 (D). The magenta and green lines in lower figures represent the percentage of helical and strand segments in the cluster. A helical segment is defined as the segment with the rate of α-helical residues >= 0.5. A strand segment is defined as the segment with the rate of β-strand residues >= 0.5. The percentages of helix (H) and sheet (E), that are determined using the DSSP program [40], are also represented by the red and blue lines, respectively.Long length (27–50 residues long)Conformational spaces for the long lengths were further shortened in the direction of PCall1 and enlarged in that of PCall3. The segment distribution converged on a large populated region that exhibited a density of 0.1 (magenta in Fig. 3c) in the conformational space. With a length of 30 residues, there were two clusters consisting of compact segments and long α-helical segments, respectively, with densities of 0.35 (red in Fig. 3c) in the populated region. The emergence of the compact-segment cluster was attributable to an increase in various types of segments with a small radius of gyration (see inset of Fig. 3c). Various types of conformations are mixed up in the compact-segment cluster. The α-hairpins are derived mainly from all-α proteins. The compact β-sheet structures are derived mainly from all-β proteins. Compact conformations of other types are derived from α/β and α +β proteins (Fig. 4c). About 2% of all segments were included in the compact-segment cluster for 27-residue length. In contrast, long α-helical segments with a large radius of gyration were located on the opposite side of the cluster of the compact segments along the PCall1 axis. For lengths greater than 30 residues, the proportion of the conformations with a small radius of gyration in the compact-segment cluster increased rapidly to around 14% for 50-residue lengths. Those conformations were derived from various folds (Fig. 4c). The supersecondary structures, such as βαβ units and β-sheets, were included in the compact-segment cluster (Fig. 4d).Contribution ratios of principal axesDistributional alterations were observed associated with the changes of segment length. For principal component analyses, the contribution ratios (see Methods for the contribution ratios) of the principal components (i.e. PC axes) to the entire distribution indicate how well the PC axes can cover the variation in the original data. Figure 5 portrays contribution ratios of the first five PC axes (PCall1 – PCall5) for segment lengths of 10–50 residues. Even with a length of 43 residues, the cumulative contribution ratio of the first three PC axes, Q123 (= Q1 + Q2 + Q3), was greater than 60%, although Q123 decreased constantly with increased segment length. Each of Q4 and Q5 was always less than 8%. The contribution ratios for higher-order PC axes than PCall5 did not exceed 5% for the examined segment lengths. Therefore, it is sufficient to use only the first three PC axes (or the first five PC axes occasionally) to explain the original structural variation.Figure 5Contribution ratios for the PC axes. Contribution ratios for the first five PC axes for segment lengths of 10–50 residues are shown here. The cumulative contribution ratio for the first three PC axes (Q123, solid line with no symbol) and the individual contribution ratios of PCall1 (Q1, crosses), PCall2 (Q2, asterisks), PCall3 (Q3, squares), PCall4 (Q4, filled squares), and PCall5 (Q5, circles) are shown.With respect to the individual contribution ratios (Q1-Q3) of the first three PC axes, Q1 was overwhelmingly higher than those of the other PC axes up to 50-residue length (Fig. 5), which indicates that PCall1 is a meaningful and fundamental descriptor for segment conformation. Actually, Q1 decreased rapidly, and Q2 increased in the short segment lengths (i.e. 10–22 residues). Thereafter, both Q1 and Q2 decreased slowly. In addition, Q3 increased gradually with lengths up to 33 residues, with a maximum value of 11.5%.Investigation of structural properties of conformational axesAn eigenvector was analyzed for each PC axis with a triangle map to elucidate the physical and conformational meaning of the PC axes of the conformational space of the short to long segments. The eigenvector can be regarded as a collective variable to describe the segment conformation. Figure 6 shows triangle maps of the first five PC axes (PCall1 – PCall5) for short (10 residues), medium (26 residues), and long segments (30 residues). The triangle map clearly portrays residue pairs, with large or small deviations of Cα-Cα distances along each PC axis from the average distance <qi>. In the triangle map, positive (red) and negative (blue) areas correspond to residue pairs with mutually inverse deviations. The patterns of red and blue areas are conserved in the universes of short to long segments, indicating that conformational deviations related to the PC axes are conserved among the universes. Figure 7 depicts the conformational changes along the PC axis using colored arrows.Figure 6Distance matrices for the PC axes. Distance maps express eigenvectors for the first five PC axes for 10, 26, and 30 residue lengths. Each map shows deviations of Cα-Cα distances along each eigenvector from the mean Cα-Cα distances. The scale bar indicates the relative deviation. The color indicates whether in this particular mode the distances are increasing (red) with respect to the mean value or decreasing (blue). The eigenvector values are scaled by the square root of eigenvalue of the PC axis k, λk1/2. Residue numbers are displayed with horizontal and vertical sides of each triangle map. Two segment conformations were picked up from both ends on each PC axis; they are displayed under each triangle map.Figure 7Visualization of collective variables for the first five PC axes of 26 residue lengths. Collective variables (eigenvectors) for the first five PC axes of 26 residue lengths are visualized by the vectors onto Cα atoms of a segment. The vector for the Cα atom of residue i is calculated as ri = Σi ≠ jqijevkijλk1/2, where qij is the vector from residue i to j, evkij is the element of eigenvector of the PC axis k (vk) corresponding to the Cα-Cα pair between residue i and j, and λk is the eigenvalue of the PC axis k. Positive (blue) and negative values (red) are shown for elements of eigenvectors. The reference segment used in this figure is designed to clarify the difference between structural variables.Actually, PCall1 corresponds to the change of the radius of gyration (Rg). The triangle map for PCall1 has only one positive area, shown as red in Fig. 6, which is located near the residue pairs at the N-terminal and C-terminal sides. This single area indicates that the distant residue pairs in the sequence have a larger conformational deviation along PCall1. The correlation coefficient of the conformational deviation along PCall1 with Rg was greater than 0.9 in segment lengths of 10–50 residues (Fig. 8). The arrows in Fig. 7 point to the center of the segment, which indicates clearly that the conformational changes along PCall1 are involved with expansions or compressions of the conformation. For short lengths, PCall1 also shows a strong correlation with the changes of the segment end-to-end distance (Dend), which is defined as the Cα-Cα distance between the first and last residues of segments. Correlation between PCall1 and Dend slowly weakened with increased segment length: 0.91 for 10 residues, 0.79 for 26 residues, and 0.77 for 30 residues.Figure 8Correlation coefficients between conformational deviation along each PC axis and physical indicator. The radius of gyration (Rg) for PCall1, structural symmetry related to the N-terminal and C-terminal halves (Dsim) for PCall2, a simple indicator of β-hairpin (Dmn+mc) and Dmn+mc/end-to-end distance (Dend) for PCall3, and the radius of gyration (midRg) around the midpoint of the segment for PCall4 were used in these analyses. Correlation coefficients were calculated at every 10 residues of 10–50 residues.The PCall2 correlates to a degree of structural symmetry (Dsym) of a segment with respect to the N-terminal and C-terminal halves. The Dsym is defined as follows: Given a distance matrix for a segment, where element (i,j) is the distance (denoted as rij) between Cα atoms of residue i and j. Then, the degree of structural symmetry is defined as the sum of the squared differences of symmetric elements in a distance matrix for a segment: Dsym = Σ1 ≤ i < j ≤ n (rij - rn-(j-1)n-(i-1))2, where n is the segment length. The triangle map for PCall2 was separated into one positive area (red) and one negative area (blue). The correlation coefficient of the conformational deviation along PCall2 with structural symmetry, Dsym was greater than 0.90 in the segment lengths of 10–50 residues (Fig. 8). Both conformations displayed mirrored symmetry about a plane constructed by PCall1 and PCall3 when two conformations were picked from opposite positions along PCall2. The segment conformations picked up along PCall2 are shown in Figs. 3a–3c.The PCall3 correlated with a physical indicator that describes a conformational transition between structures with one turn and ones with two turns (PCall3 in Fig. 6). The picked conformations along PCall3 indicate that segregation of a β-hairpin structure exists along with conformational changes by PCall3. We defined the physical indicator (Dmn+mc) of the β-hairpin formation: Dmn+mc is the sum of the norms of two vectors, which were generated by the middle point of the segment for both the N-terminal and C-terminal residues: Dmn+mc = |dmn→+dmc→|, where dmn→ and dmc→ respectively denote the vectors from the midpoint to the N-terminal and C-terminal residues of the segment. Good correlation was found between PCall3 and Dmn+mc (Fig. 8). The correlation coefficient was greater than 0.7 for the 10–50 residues. The triangle map of PCall3 indicated a separation of one positive area (red) and two negative areas (blue). It is noteworthy that the triangle map of PCall3 for short segments differed slightly from those of medium and long segments. A positive area is visible near the residue pair of the N-terminal and C-terminal in the short map, suggesting that PCall3 has a (negative) correlation with Dend. For medium and long lengths, the positive area was close to the center of the triangle map. Therefore, the correlation between PCall3 and Dmn+mc/Dend was necessarily smaller in medium and long lengths.The triangle map of PCall4 had one negative area and one positive area. The positive area, located at the map center, suggests that PCall4 is correlated with the radius of gyration (midRg) of the middle region of the segment – except for both the N-terminal and C-terminal quarter portions – in the medium and long segments. The respective correlation coefficients for the 26 and 30 residue lengths were 0.73 and 0.72. The PCall4 also has a weak (negative) correlation with Dend. The respective correlation coefficients between PCall4 and Dend for the 26 and 30 residue lengths were -0.45 and -0.42.We identified no simple physical indicator for conformational changes along PCall5. However, visual inspection from conformations picked along PCall5 suggests that PCall5 is a conformational axis that represents segregated β-sheet structures. Conformations picked up from both ends on PCall5 are depicted in Fig. 6. In the triangle map for PCall5, two positive and two negative areas exist along the diagonal line, which might indicate that PCall5 segregates segment conformations with double turns. The PCβ5 contribution ratio, which was derived from all-β proteins, was higher than that derived from other structural classes, which suggests that PC5 is important for describing the structural variation of β-structures.Segment universes derived from different structural classesThe segment universes described above are those derived from proteins of the four structural classes. Therefore, decomposition of the universe into four classes is helpful to evaluate the influence of each structural class on the segment universe. To this end, a segment universe was constructed for each structural class separately, and compared the PC axes derived from each universe with those of all segments (i.e., PCall1-PCall3). The first three largest eigenvectors of each structural class were also compared respectively with PCall1, PCall2, and PCall3 to elucidate the structural properties of PC axes derived from each universe.Figure 9 depicts the contribution ratios of the first three PC axes, PCx1 -PCx3 (x = α, β, α +β, or α/β), in each structural class. The marks on the curves in Fig. 9 indicate that the correlation coefficient (vxi·valli) between PCx1 -PCx3 and PCall1 -PCall3 (i.e., i = 1, 2, 3) is greater than 0.7, which was used here as a threshold of conservation of structural properties. The properties of the first two PC axes corresponding to the PCall1 and PCall2 were highly conserved in all four structural classes. The characteristics of PCall3 were also conserved in all four structural classes, although exceptions were apparent for the 20-residue-long and 10–16-residue-long all-α and all-β classes. Therefore, it is confirmed that the first three PC axes (Rg, symmetry, and one/two turn(s)) are important in almost all cases to describe the conformation of segments embedded in globular proteins.Figure 9Contribution ratios for the PC axes for each structural class. Contribution ratios (Q1, crosses; Q2, asterisks; and Q3, squares) of PCx1-PCx3 vs. segment lengths of 10–50 residues for each structural class, where x = α, β, α+β, or α/β. The correlation coefficient between PCxi and PCalli (i = 1, 2, 3) is 0.7 or less if no mark is present at a segment length. For the all-β class, no axis exhibited a correlation coefficient greater than 0.7 up to PCall3 for segment lengths of 10–16 residues.However, the curves for the contribution ratios of both all-α and all-β classes (see two panels of Fig. 9) differ clearly from those of PCall1 – PCall3 (i.e. Q1 – Q3 in Fig. 5). The Qα1, contribution ratio was always higher than 40%, which indicates that the distribution of the all-α segments has a large deviation with respect to Rg. In contrast, the Qβ1 contribution ratio decreased rapidly with increasing segment length. The value of Qα2 increased moderately with increasing segment length. In contrast, the Qβ2 had a maximum value greater than 20% at a length of 22 residues. This rapid increase of Qβ2 might reflect a typical feature for β-sheet conformations. For PC3, the curves for the contribution ratios of the all-α and all-β classes also mutually differed. Although Qβ3 peaked at a length of 35 residues, Qα3 peaked with a short length, which indicates that the structural variable based on PCall3 is important for β-segments longer than 30 residues. In contrast, the behaviors of the contribution ratios for both α+β and α/β classes along with the segment length resembled each other. They were also similar to Q1-Q3 in Fig. 5 because those structural classes are mixtures of α-helices and β-sheets.Subsequently, PC axes that were specific for each structural class were examined. For this analysis, the PC axis was defined as a \"class-specific\" one when a PC axis from a structural class showed no similarity with the first 20 PC axes from the other three structural classes (see Methods). The first 10 PC axes of each class were investigated for the short (10 residues), medium (26 residues), and long (30 residues) segments. Ten class-specific conformational axes were identified and consisted of one (PCβ10) for the short length, eight for the medium, and one (PCα8) for the long. The eight class-specific axes for the medium-length segments are PCα5, PCα8, and PCα10 for all-α, PCβ10 for all-β, PCα+β9 and PCα+β10 for α+β, and PCα/β8 and PCα/β10 for α/β. Four examples out of eight are depicted in Fig. 10. A clear correlation of these PC axes is difficult to discern according to simple physical or structural quantities. Figure 10a shows that the PCα8 describes a structural change of three (both ends and the middle portion) parts of α-segments. The PCα/β8 is related to βαβ motifs, which is the most fundamental structural unit for α/β proteins.Figure 10Examples of class-specific conformational axes. Conformational axes specific to structural class for 26-residue segments are shown. Eigenvector maps and conformations picked up from both ends on each PC axis are shown: A for PCα8, B for PCα/β8, C for PCα10, and D for PCβ10. The color contrast of the maps is enhanced to aid viewing. The anti-phase of the blue color is shown in red. Residue numbers are displayed with horizontal and vertical sides of each triangle map. The respective contribution ratios of PCα 8, PCα/β8, PCα10, and PCβ10 were 1.3%, 1.4%, 0.9%, and 0.7%.DiscussionInvestigation of the protein segment universe is an important subject for bioinformatics. Results of this study show that the segment universe can be categorized naturally into three regimes: short, medium, and long. A main finding of this study is that the three regimes are clearly demarcated by critical changes in the shape of the segment distribution in the conformational space. Preceding studies demonstrated that the average length of α-helix is 14 residues [24] and that for β-strand is five residues [25]. Results of the present study show that transitional segment lengths (22 and 26 residues long) do not coincide with these average lengths. Therefore, a single secondary structure element does not characterize the shape of the segment distribution. The appearance of the medium length regime segregates the segment fold universe into three. The combination of secondary-structure elements is important to characterize not only the medium-length segment universe but also the entire segment fold universe.Meanwhile, loops, which make up 30% of the protein structures [26], are also expected to take a larger role to form some unique conformations by connecting secondary-structure elements in the medium to the long-length segment universe than short one. The segments in the cluster of the medium to long-length universe tend to contain more loop regions than those of the short segment universe, as shown in Figs. 4b and 4d, and have a wider variety of origins (Figs. 4a and 4c). For example, the segments in the cluster with density of 0.35–1.0 of the universe of 30 residues length are derived from 461 proteins out of all 600 representatives used for this study (see Additional File 1). Longer loops that possess extended conformations are located on the opposite side of the compact-segment cluster along PCall1 in the medium to long segment universe (Figs. 3b and 3c). Instead of discrete clusters, they appear to constitute a rather continuous distribution. Some analyses examine short loops with respect to their completeness [27,28] and elaborate classification [26,29]. In the analysis of short segments, our method also captured some loop conformation classes, such as joint loops connecting two helices, and exposed and extended loops participated in protein-protein interactions [23].A natural boundary was identified, in this study, between the peptide-like and protein-like distributions between the lengths of 23 and 26 residues using actual conformations of protein segments. This observation with respect to the boundary is consistent with the results described by Shen et al. [30], even though they used a sphere-packing model to estimate a minimal domain size of about 20 residues. A recent study by Sawada and Honda [31] also identified a boundary at 10–20 residue length by calculating the structural diversity of segments. They discretized the conformational space using a single-pass clustering method. In contrast, we observed the density distribution to uncover differences of conformational space between short and long segments. The segment conformational space for lengths of 10–22 residues provided a distribution with an extreme density gradient towards the secondary structure, such as the α-helix, β-strand, and β-hairpin clusters, which are expected to belong to the peptide-like conformational regime. This conformational variation reflects that short segments embedded in globular proteins are mainly stabilized by the physicochemical property of the peptide. On the other hand, the segment conformational spaces for lengths of 27 residues or more have a distribution that is dominated by compact segments, which suggests a protein-like distribution (protein-like conformational regime). This distribution arises from the hydrophobic effect imparted by the solvent molecules, which is of great importance for structural stability in long segments derived from globular proteins. If this is the case, our observations support the de novo structure prediction methods, so-called fragment assembling methods, that have been developed recently [32-35]. These approaches are usually based on the prediction of local segment conformations followed by assembly of segments, and are generally used to separate criteria at each step; sequence similarity or secondary structural propensity for the prediction of segment conformations, and non-local energy terms for the assembling step. These strategies used in the de novo prediction methods seems to be consistent with the results shown here. Results of our analyses clearly show such a hierarchical organization of protein structures, and indicate that preparing segment libraries up to around 20 residues long would be helpful for such methods.These results indicate that the structural meanings for the conformational axes (i.e., the radius of gyration for PCall1, structural symmetry related to the N-terminal and C-terminal halves for PCall2, and a single-turn/two-turn structure for PCall3) are conserved in the different lengths and structural classes. This fact suggests that these conformational components are key structural variables for protein segments. On the other hand, when conformational axes among the four structural classes were compared, we were able to identify several conformational axes that were specific to each structural class, especially in the medium length range. In fact, a distribution change for medium lengths was observed, involving an increase in compact segments. Those segments included supersecondary structures such as α-hairpins, parts of the β-sheets, and βαβ units. These results might be related to the specificity of the structural class or fold of the contents of supersecondary structures [20]. Typical supersecondary structural motifs, α-hairpin, β-hairpin, and βαβ are, respectively, the basic structural units for the all-α, all-β, and α/β proteins. These motifs are often shared within the structural classes. Therefore, the contribution ratios observed for the class-specific conformational axes were high. Class-specific conformational axes were rarely observed in short and long lengths, probably because short segments are too nonspecific and are often shared over different structural classes; long segments are too specific and have very low contribution ratios for conformational axes that are specific for each structural class.The currently found class-specific conformational axes provide a hint to solve a difficulty in classifying diverse sets of protein structures. Both α/β and α+β classes are known to show a substantial overlap. In the CATH classification, α/β and α+β classes are treated as one structural class as α-β class. Classifying α/β and α+β proteins is sometimes a difficult problem, although several classification [19,36,37] and also prediction [38,39] schemes have been proposed. The present study showed that α/β and α+β classes have similar characteristics of universes, and also have unique ones at the same time. For example, our results show that PCα/β8, whose contribution ratio was 1.4%, was associated only with the βαβ motif. In the α+β class, no axis was strongly correlated with PCα/β8 (see Additional File 2), which is a clear example of the difference in structural variables between α+β and α/β classes originating from class-specific supersecondary structures. Consequently, projecting segments onto a conformational subspace using the axis PCα/β8 could be useful for objectively dividing protein domains of α-β class into α/β and α+β classes. A considerable localization of segments derived from α/β proteins in a PCA subspace is observed (see Additional Files 3 and 4).An effective method must be developed for conformational sampling for de novo prediction methods. The resulting structural variables analyzed in this study would be helpful for additional progress in de novo structure prediction. For example, testing the distribution of segments or models in terms of the degree of symmetry using the descriptor (Dsym) might be useful to verify the completeness of sampling of the conformational space. Using a filtering threshold or function (generally used in fragment assembling methods for selecting proper models) that is tolerant of the radius of gyration might be useful for improving the prediction of all-α proteins because the contribution ratio, Qα1, of PCα 1 corresponding to the radius of gyration (Rg) is larger than those of the other structural classes in the medium and long segments. Consequently, projecting segments of models onto a conformational subspace constructed by PCx (where x = α, β, α/β, α+β, or all) axes might be helpful for filtering out models and assigning a protein to a structural class.ConclusionIn this study, the dual critical transitions in the protein segment universe from short to long length are shown. Our observations are related to the transitions proposed by the significance of two-stage treatment in de novo structure prediction. Considering the hierarchical organization of a protein segment universe that we have shown, we suggest the efficacy of using the evaluation functions that is secondary-structure-directed for sampling local structures less than 23 residues long. We also suggest the suitability of evaluating protein-like features of models using another function (e.g. Rg) for longer segments. Changing the criteria of filtering for each structural class will enhance the effectiveness of the conformation sampling process. Through these analyses, we have demonstrated that our clustering methodology is useful to identify a distinctive distribution shift of conformational space between short and long segments and that distribution changes depend on structural classes.MethodsPreparing the segment librariesOne representative from each fold group of the SCOP database (ver. 1.63) [15] was chosen to obtain a segment library without a bias of usage of the folds. The representatives cover the four major structural classes (all-α, all-β, α+β, and α/β), because we are interested in and specifically examine characterization of the nature of segments embedded in usual size globular proteins. Small proteins of less than 50 residues and non-single chain proteins with less than 100 residues were excluded, as were membrane proteins. It is expected that those proteins possess different structural properties from those of usual size globular proteins and induce biased results. In all, 600 representatives were used for this study (all-α, 150; all-β, 116; α+β, 219; α/β, 115; see Additional File 5). Dividing the protein structures into segments with a sliding window by one residue along the sequence generated a segment library of arbitrary length. We prepared a segment library for each length of 10–50 residues to generate conformational spaces of short-to-long segments. In such cases, segments with incomplete coordinate data (e.g., having an unusual covalent-bond length or lacking main-chain atoms) were excluded. Furthermore, to elucidate differences among the conformational spaces derived from the four major structural classes, we generated a segment library for each class.Construction and visualization of conformational spaceWe previously reported a method for constructing and visualizing the conformational space of protein segments using principal component analysis based on intra-segment Cα-Cα atomic distances [23]. Briefly, atomic distances of all Cα-Cα pairs for each segment in a segment library of an arbitrary length were calculated first. A distance is designated as qi, where i is the index for the Cα-Cα pair, i = 1, ..., n(n - 1)/2, and n is the segment length, as expressed by the number of residues in the segment. Subsequently, a set of eigenvectors and eigenvalues were obtained by diagonalizing a variance-covariance matrix, C, that was calculated as Cij = (<(qi - <qi>)(qj - <qj>)> = <(qiqj - qi<qj>- <qi>qj + <qi><qj>)> = <qiqj>- <qi><qj>- <qi><qj> + <qi><qj> =) <qiqj>- <qi><qj>, where the average <...> is taken over the segments. Two equations, Cvi = λivi and vi·vj = δij, are satisfied. Eigenvectors with larger eigenvalues are more important in the study of the conformational varieties of the segments. Eigenvalues are arranged in descending order: λi > λj if i <j. The contribution ratio of the i-th PCA element (i.e. the i-th eigenvector) to the whole conformational distribution is given as Qi = λi/Σkall λk. The eigenvectors, which are called PCx1, PCx2, PCx3, ...etc., were used as conformational axes to construct a segment conformational space, a PCA space, in which x indicates a segment dataset: x = α, β, α/β, α +β, or all). The indicator \"x = all\" is given when conformational axes are generated by the whole segment dataset. The origin of the PCA space is set on the average Cα-Cα atomic distances: <q> = [<q1>, <q2>, <q3>, ..., <qn>]. This enables ready comparison of conformational distributions between constructed universes. Any position (i.e. any segment structure) in the PCA space can be expressed using a linear combination of eigenvectors as ck = Σnall (q - <q>)·vk λk1/2, where ck is a coordinate (i.e. projection of q) on the PC axis k. Using the first three eigenvectors (PCx1, PCx2, PCx3), a three-dimensional (3D) PCA space can be constructed.We defined a vector, r, to express the position of each segment in the 3D PCA space: r = [c1, c2, c3]. After projection of the segments on the 3D PCA space, the distribution of segments in the 3D PCA space was visualized using the following procedure. The 3D space was divided into N bins (total N3 cubes). The bin size was defined as (max [c1] - min [c1])/N, where N = 36, and max [c1] and min [c1] respectively signify the maximum and minimum of the coordinates of the segments along the first principal component axis. The number (i.e. frequency) of segments detected in a cube represents the density (i.e. probability) of segments to be found in the cube. The density of each cube, ρ was normalized by the maximum density, ρmax among the cubes so that the maximal value of normalized density (we call this density in the text) is set to 1 (refer to eq. (3) in [23]). Four levels of contour surfaces (i.e. iso-density surfaces) were depicted to visualize the 3D PCA space. The density values for those surfaces were set respectively as 0.005, 0.01, 0.1, and 0.35.We also separately constructed the universe for four structural classes to assess differences among their conformational spaces. For this study, we specifically examined the first 10 PC axes of each structural class because the 10 PC axes are more important than the other axes with respect to capturing the differences in the conformational axes. Although the eigenvectors from the same structural class are mutually uncorrelated (i.e., vxi·vxj = 0, where i ≠ j and x = α, β, α/β, or α+β), the eigenvectors from different structural classes might have some correlation (i.e., vxi·vyj ≠ 0, where x ≠ y). The PC axis is defined as the conformational component specific to the structural class when a PC axis from a structural class has no similarity to the first 20 PC axes from the other structural classes with a correlation coefficient > 0.8 (i.e. vxi·vyj > 0.8).Authors' contributionsThis study was conceived and carried out by KI, who also analyzed the results and drafted the manuscript. HT approved the study and participated in the discussion. JH participated in the design and coordination of the study. He also helped to write the manuscript. KT participated in the design and discussions of the study and wrote the manuscript. KI and JH developed the methodology. All authors read and approved the final manuscript.Supplementary MaterialAdditional file 1Origins of segments in the cluster of 30 residue length. Distributions of the origins of segments in the cluster of the universe of 30 residues length are shown.Click here for fileAdditional file 2Correlation with the first 10 PC axes of α/β class of the medium (26 residue) segments. Maximal correlation coefficients between the first 10 PC axes of α/β class and PC axes of the other three structural classes are shown.Click here for fileAdditional file 3Class-specific region for α/β segments on the PCα/β8-PCα/β3 plane. Distributions of segments of α/β structural class proteins for the medium length are shown.Click here for fileAdditional file 4Discrimination of segments from the α/β structural class in the PCα/β8-PCα/β3 plane. Specificity and coverage rates of segments of α/β structural class proteins in the PCα/β8-PCα/β3 plane are presented.Click here for fileAdditional file 5List of PDB ids used in this study. The PDB and SCOP IDs of proteins used in this study are listed.Click here for file\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2529327.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2529327",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2529327\nAUTHORS: Narayanan Venkatasubramani, Colin D Rudolph, Manu R Sood\n\nABSTRACT:\nBackgroundMotilin, a peptide hormone has a direct excitatory effect on circular smooth muscle strips derived from the human colon. Reduced plasma motilin concentration has been reported in adults with chronic constipation. Erythromycin, a non-peptide motilin receptor agonist, induces phase 3 of the migrating motor complex (MMC) in the antro-duodenum and also reduces oro-cecal transit time. A pediatric study has reported an improvement in clinical symptoms of constipation following erythromycin administration, but the effect on colon motility in children has not been formally evaluated. We used colon manometry to study the effect of intravenous erythromycin lactobionate at 1 mg/kg on colon motiltiy in ten children.MethodsWe selected patients with normal antroduodenal and colon manometry studies that were performed simultaneously. All studies were performed for clinically indicated reasons. We quantified the effect of erythromycin on colon contraction by calculating the area under the curve (AUC).ResultsThe mean (SE of mean) AUC in the colon during the fasting, post-erythromycin and postprandial phases of the study was 2.1 mmHg/sec (0.35), 0.99 mmHg/sec (0.17) and 3.05 mmHg/sec (0.70) respectively. The AUC following erythromycin was significantly less compared to the fasting phase of the study (p < 0.01).ConclusionErythromycin lacks colon prokinetic effect in children with chronic constipation evaluated by colon manometry.\n\nBODY:\nBackgroundConstipation accounts for 3% of general pediatric and 10–20% of pediatric gastroenterology outpatient visits [1]. A majority of these patients have functional constipation and the symptoms improve following behavioral modification and laxative treatment [2]. Almost 30% of children with chronic constipation will have persistent symptoms or relapses that can persist into adult life [3]. Children with chronic intractable constipation, who do not respond to conventional medical therapy, require manometric evaluation to exclude an underlying colon neuromuscular abnormality.Motilin is a 22-amino acid peptide hormone secreted by the enterochromaffin cells of the small intestine [4,5]. It exerts profound effect on gastric and small bowel motility by inducing the inter-digestive phase 3 of the migrating motor complex (MMC) [6]. Peak plasma concentration of motilin is associated with MMC, both in animals and humans [7-9].Conflicting results exists regarding the erythromycin effect on circular smooth muscle strips derived from the human colon. Few studies have reported the stimulatory effect of erythromycin in human smooth muscle contractions [10,11] and Nissan have reported lack of any excitatory effect on human colon [12]. Studies have shown that motilin receptor is expressed in the enteric neurons of the colon [10,13] and plasma motilin concentration is reduced in adults with chronic constipation [14]. Oral and intravenous erythromycin has no effect on distal colon contraction or transit in healthy human volunteers [15].Erythromycin is a non-peptide motilin receptor agonist which induces phase 3 of the migrating motor complex in the antro-duodenum. Several studies have reported that erythromycin is safe and effective in improving feeding intolerance in preterm infants and children [16-19]. The prokinetic effect of erythromycin has also been reported in older children with motility disorders [20]. The data regarding the colon prokinetic effect of erythromycin is controversial. Oral erythromycin has been shown to reduce the colonic transit time assessed using radio opaque markers in adults with chronic constipation [21]. However, another adult study using colon manometry reported no significant improvement in colon motility with erythromycin compared to a placebo [22]. To date, no studies have evaluated the effect of erythromycin on colon motility in children.The aim of our study was to evaluate the effect of intravenous erythromycin lactobionate (1 mg/kg) on colon motility in children with chronic constipation and fecal incontinence using colon manometry.MethodsWe retrospectively evaluated 10 simultaneously performed antro-duodenal manometry (ADM) and colon manometry (CM) studies performed at the Children's Hospital of Wisconsin between June 2000 and June 2005. These studies were performed to exclude an underlying small bowel and/or colon motility disorder. Only patients with normal antro-duodenal and colon motility studies were included. The presenting symptoms were chronic constipation in 8 patients, abdominal pain and fecal incontinence in 1 and abdominal distension and pain in 1 patient. Hirschsprung's disease was excluded either by anorectal manometry or rectal biopsies in all subjects. This retrospective chart review study was approved by the Children's Hospital of Wisconsin Human Research Review Board (Protocol number: CHW 05/187, GC 31). Informed consent was not obtained from the patients as this was a retrospective chart review study.All drugs known to affect the gastrointestinal motility were discontinued at least 72 hours before the motility studies. Patients fasted for at least 8 hours before the study. All children were anesthetized without a muscle relaxant. We waited for the child to recover completely from the effects of the drug before starting the motility tests [23]. Colonoscopy was performed to assist colon manometry catheter placement. The tip of the colon motility catheter was positioned in the cecum/ascending colon in all subjects and the position was confirmed by fluoroscopy. A water perfused manometry catheter with 8 recording sites was used for ADM and CM. Catheter position was checked using fluoroscopy. The catheter was perfused with 0.45% sodium chloride solution at the rate of 0.4 ml per minute per recording site, using a pneumo-hydraulic infusion system. The pressures were transmitted to a transducer and recorded on a computer with specialized motility software (Medical Measurements System, Amsterdam). We performed at least 2 hours fasting recording of both ADM and CM studies, following which we administered intravenous erythromycin lactobionate 1 mg/kg and performed another 60 minute of recording. Next, the patients ate a meal appropriate for their age (meal provided >30% of daily caloric requirement) and the recording was continued for another 60 minutes [24]. Bisacodyl (5–10 mg) was administered through the central lumen of the manometry catheter, directly into the colon to induce high amplitude colon contractions (HAPCs) and recording was continued for at least 30 minutes (Figure 1). This is a standard protocol for simultaneously performed ADM and CM studies performed at our center.Figure 1HAPCs in a patient with chronic constipation. High amplitude propagating contractions (HAPCs) following bisacodyl stimulation in a 10 year old boy with chronic constipation. The colon contractions are propagating from the cecum to the sigmoid colon.Antro-duodenal manometry was considered normal if i) phase 3 of the MMC, as defined by repetitive antral contractions occurring at a rate of 2–3 per minute and 10–12 per minute small bowel contractions, lasted for more than 2 minutes with normal antegrade propagation ii) there was normal stomach antrum and small bowel response to a meal [25]. Colon manometry studies were considered normal if there was a gastro-colonic response and spontaneous or bisacodyl stimulated antegrade propagating HAPCs [26]. We defined HAPCs as colon contractions with amplitude of at least 60 mmHg, and propagating over at least 30 cm of colon.All colon manometry studies were reviewed by an experienced gastroenterologist and artifact was removed. The area under the curve was calculated by measuring the area under the pressure line for one 60 minutes period during fasting, following erythromycin administration and after a meal. Increase in motility index following a meal is considered a gastro-colonic response. The CM recordings were also evaluated for HAPCs.We used SPSS software, version 10 for Windows (SPSS Inc., Chicago, Illinois, USA). We compared the mean (SE of mean) AUC using Student's t test. To compare the effect of erythromycin and bisacodyl on colon contractions, we used McNemar's test for disagreement.ResultsThe mean age of the patients at the time of the study was 9.6 years (range 4–12 years); there were five females. In all patients, we recorded phase 3 of the MMC following intravenous erythromycin and a normal postprandial antro-duodenal manometry.The mean (SE of mean) AUC in the colon during the fasting, post erythromycin and postprandial phases of the study was 2.1 mmHg/sec (0.35), 0.99 mmHg/sec (0.17) and 3.05 mmHg/sec (0.70), respectively (Table 1). The AUC following erythromycin was significantly less compared to the fasting phase of the study (p < 0.01), suggesting that colon motor activity was reduced following erythromycin administration.Table 1AUC and HAPCs during fasting, following a meal and erythromycinFasting PeriodPostprandial periodPost-erythromycin periodBisacodyl stimulationMean area under the curve in mm/sec (SE of mean)2.1(0.35)3.05 (0.70)0.9 (0.17)NAMean number of HAPCs (range)0.4 (0–1)0.00 (0)0.1 (0–1)8.9 (1–16)The mean area under the pressure line during fasting, following a meal and erythromycin lactobionate administration. The mean number of HAPCs recorded during each of these periods and following bisacodyl stimulation are also shownAll patients had a normal gastrocolonic response. Four patients had spontaneous HAPCs during the fasting period, one following erythromycin stimulation and none following the meal. All patients had HAPCs with bisacodyl stimulation, the mean number of HAPCs was 8.9 (range 1–16). McNemar's test for disagreement between erythromycin and bisacodyl induced HAPC was statistically significant (p = 0.004). The mean interval of first HAPC after bisacodyl was 7.18 minutes (range 1–10 minutes).DiscussionThis retrospective study evaluates the effect of erythromycin on colon motility in children with chronic constipation using colon manometry. All our patients had normal antro-duodenal and colon manometric studies. The two recognizable features of normal colon motility in children are an increase in colon contractions following a meal (gastrocolonic response) and HAPCs[26]. If spontaneous HAPCs are not recorded during fasting and postprandial period, bisacodyl is used to stimulate HAPCs [27]. All patients in our study had a normal gastrocolonic response and bisacodyl induced HAPCs. This suggests that none of our patients had intestinal pseudo-obstruction or colon neuromuscular abnormality. In our opinion, these patients had functional constipation and/or fecal incontinence.All patients showed normal phase 3 MMC activity following intravenous erythromycin lactobionate (1 mg/kg dose). This shows that the dose of erythromycin lactobionate used was adequate to stimulate the motilin receptors in the foregut. In the colon there was a significant decrease in the frequency and amplitude of contractions following erythromycin lactobionate when compared to the fasting period. This may be because the motilin receptors in the colon have a higher threshold of activation compared to small bowel. An adult study reported no prokinetic effect of erythromycin on colon motility as determined by colon manometry [22]. This suggests that unlike the foregut, erythromycin lactobionate in a dose of 1 mg/kg does not have a prokinetic effect on the colon and probably a higher concentration may be necessary to stimulate the motilin receptors in the colon.There is heterogeneity in the motilin receptor affinity for erythromycin in the gastrointestinal tract [28]. It is possible that the colon motilin receptors may have reduced affinity for erythromycin compared to the antral nerves or the expression of motilin receptor may be reduced in the colon. The limitation of our study is that we only evaluated the effect of a single dose of erythromycin (1 mg/kg) given intravenously. It is possible that a higher dose may be necessary to induce a prokinetic effect on the colon. An adult study using a higher oral dose of erythromycin (1 g/day), reported improvement in segmental and colon transit time assessed using radio-opaque markers [21].ConclusionOur study suggests that erythromycin lactobionate at 1 mg/kg does not have a colon prokinetic effect in children with chronic intractable constipation. Further studies are needed, using a higher dose of erythromycin, to evaluate the dose response curve and affinity of the colon motilin receptors to erythromycin.AbbreviationsMMC: migrating motor complex, AUC: area under the curve, ADM: antroduodenal manometry, CM: colon manometry, HAPCs: high amplitude colon contractionsCompeting interestsThe authors declare that they have no competing interests.Authors' contributionsAll authors read and approved the final manuscript.NVS: Data analysis and manuscript writing. CR: Helped with manuscript writing and providing patients. MS: Helped with manuscript writing, data analysis and providing patients.Pre-publication historyThe pre-publication history for this paper can be accessed here:\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2531245.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2531245",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2531245\nAUTHORS: Hind A. AL-Khayat, Edward P. Morris, Robert W. Kensler, John M. Squire\n\nABSTRACT:\nA number of cardiac myopathies (e.g. familial hypertrophic cardiomyopathy and dilated cardiomyopathy) are linked to mutations in cardiac muscle myosin filament proteins, including myosin and myosin binding protein C (MyBP-C). To understand the myopathies it is necessary to know the normal 3D structure of these filaments. We have carried out 3D single particle analysis of electron micrograph images of negatively stained isolated myosin filaments from rabbit cardiac muscle. Single filament images were aligned and divided into segments about 2 × 430 Å long, each of which was treated as an independent ‘particle’. The resulting 40 Å resolution 3D reconstruction showed both axial and azimuthal (no radial) myosin head perturbations within the 430 Å repeat, with successive crown rotations of approximately 60°, 60° and 0°, rather than the regular 40° for an unperturbed helix. However, it is shown that the projecting density peaks appear to start at low radius from origins closer to those expected for an unperturbed helical filament, and that the azimuthal perturbation especially increases with radius. The head arrangements in rabbit cardiac myosin filaments are very similar to those in fish skeletal muscle myosin filaments, suggesting a possible general structural theme for myosin filaments in all vertebrate striated muscles (skeletal and cardiac).\n\nBODY:\n1IntroductionMyosin filaments (also known as thick filaments), which interact with actin filaments in muscle to produce force and movement, are assemblies of myosin molecules and accessory proteins. Much is known about their 3D structure, but high resolution information is hard to obtain. Mutations in cardiac muscle myosin and its associated proteins (e.g. C-protein; myosin binding protein C (MyBP-C)) are known to be associated with a number of myopathies (e.g. familial hypertrophic cardiomyopathy and dilated cardiomyopathy) (Watkins et al., 1992, 1995; Seidman and Seidman, 2001; Tajsharghi et al., 2003). To understand how heart muscle normally works and how the known mutations affect contractility it is essential to understand the structure and properties of normal mammalian cardiac myosin filaments.Myosin molecules comprise two heavy chains and four light chains. Parts of the heavy chains twist together to form a 1500-Å long coiled-coil α-helical rod-shaped tail domain on one end of which are two elongated globular myosin heads (crossbridges) formed by the rest of the heavy chains together with the light chains. The way in which myosin molecules and accessory proteins associate to form a myosin filament can vary depending on the type of muscle. Generally the tails pack together, often in a helical arrangement, to form the filament backbone from which the heads project. The heads are ATPases that interact with actin filaments to produce force and movement (Huxley, 1963). In vertebrate striated muscle the myosin molecules pack together into bipolar myosin filaments with the rods forming a roughly cylindrical filament backbone and the myosin heads arranged in a quasi-helical array on the filament surface. Halfway along each filament, where the rod packing is anti-parallel, there is a head-free region known as the bare-zone or M-region (Huxley, 1963; Sjostrom and Squire, 1977a, 1977b). The first myosin heads in each half A-band are at the edges of the M-region. The backbone also has on its surface additional non-myosin protein components including C-protein (MyBP-C; Offer et al., 1973); or possibly its analogue X-protein (Bennett et al., 1986) and the A-band part of titin (Labeit and Kolmerer, 1995; Trinick, 1996).The head arrangement on vertebrate striated muscle myosin filaments approximates to a 3-stranded right-handed helix where each strand possesses nine subunits (pairs of myosin heads) per turn and a pitch of 1287 Å. Since the filament is 3-stranded, the axial repeat is reduced from 1287 Å to 430 Å. The pairs of heads from three myosin molecules project from the backbone at regular intervals to form a so-called ‘crown’ of heads, with successive crowns separated axially by roughly 143 Å (Squire, 1972). The angle between each of the three head pairs on one crown is 120°. Each 430 Å repeat contains three such crowns (Squire, 1972, 1973, 1974; Harford and Squire, 1986; Stewart and Kensler, 1986; Cantino and Squire, 1986; Kensler and Stewart, 1989). Vertebrate striated muscle myosin filaments are characterised by a systematic departure from perfect helical symmetry, known as a perturbation. This was originally identified by the presence of ‘forbidden’ meridional reflections (i.e. X-ray meridional reflections seen at orders of the 430 Å repeat other than multiples of 3) in the X-ray diffraction patterns from relaxed vertebrate skeletal muscles (Huxley and Brown, 1967; Squire et al., 1982; Harford and Squire, 1986).Electron microscopic analysis of vertebrate striated muscle myosin filaments has included 2D analysis of images of negatively stained myosin filaments from fish, frog, chicken, and rabbit skeletal muscles (Kensler and Stewart, 1983, 1989, 1993; Stewart and Kensler, 1986; Kensler and Woodhead, 1995), shadowed, freeze-fractured myosin filaments from frog skeletal muscle (Cantino and Squire, 1986), and myosin filaments from freeze-substituted frog skeletal muscle (Craig et al., 1992). These studies have shown the presence of the perturbation in 3-stranded crossbridge arrangement in all these species, consistent with the perturbation first described by Huxley and Brown (1967), but could not definitively demonstrate the nature of the perturbation. Myosin filaments in vertebrate skeletal muscles have also been studied in 3D (Stewart and Kensler, 1986; Eakins et al., 2002; AL-Khayat and Squire, 2006; AL-Khayat et al., 2006), but only recently has single particle analysis been used (AL-Khayat et al., 2004, 2006).Myosin filament 3D structure has also been extensively studied in several well-ordered invertebrate striated muscle tissues such as insect flight muscle, tarantula, Limulus, scorpion and scallop (Crowther et al., 1985; Stewart et al., 1985; Morris et al., 1991; Vibert, 1992; Offer et al., 2000; AL-Khayat et al., 2003). In these cases helical reconstruction by the Fourier-Bessel approach has usually been applied, since the myosin head arrays in these muscles appear to be helical. The exception is the ground-breaking study by Woodhead et al. (2005) who applied single particle analysis and helical averaging to cryo-EM images of tarantula myosin filaments and achieved a very detailed 3D reconstruction. We have also applied the single particle approach to other helically ordered myosin filaments from both insect and scallop invertebrate striated muscles (AL-Khayat et al., 2008a, b).Recent 3D-information about the structure of the vertebrate striated muscle myosin filament from fish skeletal muscle has been achieved both by modelling of the X-ray diffraction data (Hudson et al., 1997; AL-Khayat and Squire, 2006) and by single particle analysis of EM images (AL-Khayat et al., 2006). Both these studies have provided better structural detail than had been obtained previously (Eakins et al., 2002) and they started to quantify for the first time the perturbation of the crossbridge array in the fish myosin filament. This work, together with the earlier 3D reconstruction of isolated frog skeletal muscle myosin filaments (Stewart and Kensler, 1986), has provided significant information about the crossbridge arrangement in the vertebrate skeletal muscle myosin filament. However, less is known about the structure of myosin filaments in vertebrate heart muscles. Although 2D-analysis of isolated cardiac thick filaments has been performed (Kensler, 2002, 2005a, 2005b; Kensler and Harris, 2008), a 3D reconstruction of the structure of the cardiac muscle myosin filament has not been performed.In the present work, the 3D structure of myosin filaments isolated from normal mammalian rabbit cardiac muscle has been studied. Using the technique of single particle analysis of non-helical filamentous systems (Paul et al., 2004; AL-Khayat et al., 2004; AL-Khayat et al., 2006) we produced a 3D reconstruction at about 40 Å resolution of the head arrangement in the C-zone region of relaxed rabbit cardiac myosin filaments. The results are very similar to the published structure of the vertebrate skeletal (fish) muscle myosin filament from EM 3D reconstruction (AL-Khayat et al., 2006) and from X-ray diffraction modelling (AL-Khayat and Squire, 2006), possibly suggesting a common structural theme for vertebrate striated muscle myosin filaments (cardiac and skeletal). However, for the first time, we show evidence that at low radius from the filament axis, the myosin heads origins are closer to being on a perfect helix than at high radius and that especially the azimuthal perturbation becomes more marked as the radius increases. This new structure serves as a starting point from which to understand the effects of the mutations in myosin and C-protein (MyBP-C) associated with different cardiomyopathies.2Materials and methods2.1Electron microscopyMyosin filaments were isolated from rabbit ventricular muscle in relaxing solution as described by Kensler (2002, 2005a). The isolated myosin filaments were applied to a thin carbon film support on holey carbon grids and negatively stained. Only filaments lying over holes were used in the current analysis. Electron micrographs were collected under minimal dose conditions on a JEOL 1200 electron microscope at 40 K magnification.2.2Preparation of images for single particle analysisA total of 52 electron micrographs were digitised using a Nikon Super Coolscan 8000 16-bits per colour pixel scanner at a step size of 6.35 μm/pixel (equivalent to 1.58 Å sampling in the specimen). Digitised images for the whole micrographs were saved on the PC in TIFF format and were then transferred to a Linux UNIX operating system PC running the IMAGIC suite of programs (van Heel et al., 2000) to analyse them. Initially, they were converted to floating point IMAGIC format using EM2EM. The resulting images were then binned by a factor of 4 in both the x and y directions, resulting in a final step size of 6.35 Å/pixel and converted to MRC format for pre-processing using the MRC suite of programs (Crowther et al., 1996) and also using locally developed software. Regions were selected which contained intact half-filaments which were relatively straight, not overlapped by other actin and myosin filaments, and had readily identifiable bare-zones (Fig. 1(A)). Location of the bare-zone was essential to properly deduce the location of the C-protein stripes. The area selected on either side of each half-filament was also required to have as little background as possible (Fig. 1(B)), so as to reduce noise in the calculated Fourier transforms (Fig. 1(C)). Images of whole myosin filaments were cut into two halves with the whole bare-zone (M-band) included in each half-filament. In order to preserve polarity in the processing, half-filaments (i.e. from the M-band to the pointed end of the myosin filament) were then rotated to make each filament image vertical and oriented with its bare-zone (M-band) region at the bottom (Fig. 1(B)).From the 52 available micrographs and using the above selection criteria, 153 half-filaments were identified. Half-filament images were floated in 2048 square arrays and their Fourier transforms computed (Fig. 1(C)). The sixth order of the 430 Å repeat, the 71.5 Å meridional reflection, which was strong in most computed Fourier transforms, was used to calibrate the magnification and to adjust the sampling of each half-filament from all the different micrographs to be exactly 7.54 Å/pixel. The majority of the Fourier transforms for the filaments showed up to the 11th order of the 430 Å repeat corresponding to 39 Å resolution (the titin sub-repeat; Fig. 1(C)). The correctly scaled half-filament images, in MRC format and with the pixel size accurately scaled to 7.54 Å/pixel, were then read again into IMAGIC and converted back to IMAGIC format using the EM2EM command.All the further single particle image analysis was carried out within IMAGIC. The modified exact filter method for back-projection described in Paul et al. (2004) was used for calculating the 3D reconstruction. This allows the thickness of the central section to be adjusted taking into account the fact that the diameter of the filament is less than the size of the cube. 3D structures were visualised with both IMAGIC and PyMOL (DeLano, 2002).3Results3.1Selection of myosin filament segmentsFig. 1(A) shows a typical micrograph of negatively stained isolated rabbit cardiac myosin filaments that contain good detail and from which half length myosin filaments were selected as shown in Fig. 1(B). As previously reported (Kensler, 2002, 2005a), well-preserved rabbit cardiac muscle myosin filaments, which are bipolar, have regular myosin head arrays in each half-filament with clear bare-zones (M-regions) halfway along. M-band protein density was sometimes visible in the middle of the M-region. The filament Fourier transforms showed meridional peaks out to the 11th order of 430 Å at 39 Å (Fig. 1(C)). Our aim in this study was to produce a 3D reconstruction of the structure of the myosin filament from only within the C-zone area (Sjostrom and Squire, 1977a, 1977b). This should result in a closer representation of the C-protein distribution in the final 3D structure than has been achieved before (AL-Khayat et al., 2006). Previously particles were selected from the whole of the half-filaments and thus included data from the P-zone and D-zone regions of the A-band as well as the C-zone (Fig. 2(A)).3.2Locating C-protein along the filamentsTo locate the C-zone, 1D density profiles were calculated for each of the 153 individual half-filaments examined. These half-filaments ranged in length from 6000 to 7000 Å. Their 1D profiles were aligned together by cross-correlation using a program especially developed for this purpose, Cross-Corr (Knupp, C. unpublished). The Cross-Corr program was used to align and sum these 1D profiles to give averages in order to locate precisely the positions of C-protein. Seven stripes of higher density were observed. The first particle was taken to start at a distance 2040 Å from the middle of the M-band (Fig. 2(A)) and to include at its centre the first C-protein line, C3, reported by Sjostrom and Squire (1977a, 1977b) and Bennett et al. (1986; they called it C5). Beyond this, six further strong peaks were seen separated by approximately 430 Å. These are referred to as C6, C9, C12, C15, C18 and C21 as in Sjostrom and Squire (1977a, 1977b).3.3Particle selection, alignment, classification and 3D reconstructionFrom a dataset of 153 half-filaments, 802 boxed square segments, each of length 128 pixels (equivalent to 7.54 Å/pixel × 128 pixels = 965 Å) containing just over 2 × 430 Å repeats (just over six 143 Å-myosin crown levels) were selected by stepping the box along the half-filaments at intervals of 430 Å (Fig. 2(A)). The chosen box length was used as a compromise between the easier alignment of a longer particle and a better compensation of the effects of any filament bending or distortion obtained with a shorter particle. Up to seven particles were selected from each half-filament to ensure that they only include the C-zone area.Since the three crown levels within the 430 Å repeats were different, it was essential to make sure that the selected particles were in proper axial register. The aim was to align all the segments and classify them so that each class will be the same structure but rotated about the filament long axis. The simplifying assumption was made that in the main part of the bridge regions of vertebrate myosin filaments within the C-zone area, the 430 Å repeats are all the same. In order not to mix the different crowns within a 430 Å repeat, the alignment was checked thoroughly by calculating the 1D density profiles for the sum of the particles that came from each of the 153 individual half-filaments. A typical example of the sum of particles within a half-filament is shown in Fig. 2(D) and its corresponding 1D profile is similar to that shown in Fig. 3(E). The 1D density profiles for the sum of particles of each of the 153 half-filaments were then aligned together and the positions of the peaks calculated. The shift values needed to superimpose similar peaks were found for each half-filament. The sums of the shifted particles within each individual half-filament, together with their corresponding 1D profiles, were then re-calculated to check that they were all properly aligned.A circular mask was applied to the dataset of 802 images and these were then rotationally and translationally aligned to an initial reference corresponding to the average sum of all the particles. Since the selected filament images were already quite accurately rotationally and translationally aligned, the angular range in the alignment was restricted to ±15° in order to preserve the polarity of the filaments. Also, the axial shift along the filament long axis was restricted to be between −50 Å and +50 Å to prevent the crown levels getting out of step. The aligned images were then classified by Multi-variate Statistical Analysis (MSA) (van Heel et al., 2000) to group those particles corresponding to the same view.3.4Angular assignment and class refinementBecause the aligned filament segments are likely to be related by rotation about a single axis (the filament long axis) with little out-of-plane rotation normal to this axis, ab initio angular assignment by standard angular reconstitution is not possible (Paul et al., 2004). Therefore the approach adopted here (as suggested by Paul et al. (2004)) was based on using a starting reference 3D model. The starting model was used to create an ‘anchor set’ of 2D re-projections as a reference from which to assign projection angles to class averages derived from the individual segments. By convention, these projection angles are defined by the three Euler angles α, β and γ. α is the rotation angle in the plane of the image (i.e. the plane of the EM grid), β is the out-of-plane tilt angle (to allow for the fact that the filaments may not lie perfectly flat on the grid or the grid itself may not be flat) and γ is the rotation angle around the filament long axis. We had already developed and implemented the same procedure in the previous study on the fish skeletal myosin filaments (AL-Khayat et al., 2006). In that work a number of strategies were used to test the uniqueness of the final 3D reconstruction. These included using various different starting 3D models to generate the anchor sets so as to avoid any model bias, using a totally independent angular assignment method where no starting model was required (Patwardhan et al., 2004), and using different software, EMAN (Ludtke et al., 1999). All these different methods led to essentially the same final 3D reconstruction, giving us confidence in the methodology that we were using. In the present study, we therefore used a single first reference 3D model to generate an anchor set of 2D projections to use for angle assignment (Euler angles) for each of the class averages. This reference model was generated by calculating a 3D reconstruction of a single class average with Euler angles α, β and γ of 0°, 90°, and 0°, respectively, assigned to it as its angles of view and imposing 3-fold symmetry around the long axis. The anchor set produced from this reference 3D map was used to provide a first set of Euler angles for all the class averages. An improved 3D reconstruction could then be calculated. 2D re-projections of this new 3D map were used to make a new set of reference images with which to re-align the original raw image dataset of the 802 particles using Multi-Reference Alignment (MRA), to re-classify them and to carry out a further process of angular reconstitution/assignment. This procedure was repeated with the class images gradually improving, and refinement was stopped when no further significant change was observed and the assigned angles became stable.3.5The final 3D reconstructionThe final 3D reconstruction (Fig. 3(A–D); M-band direction downwards) was calculated with imposed C3 symmetry using the best 17 class averages containing a total of 301 particles and using the modified weighted back-projection method of Paul et al. (2004). According to the Fourier Shell Correlation (Fig. 3(H)), the resolution of the reconstruction lies between 40 Å (0.5 bit criterion, van Heel and Schatz, 2005) and 48 Å (0.5 correlation coefficient criterion), and the 3D image was filtered to include frequencies from 300 Å up to 40 Å resolution. The contour threshold used to show these images corresponded to the theoretical volume of the myosin molecules within the length represented (9 whole molecules/430 Å repeat) plus an additional 15% to account for non-myosin proteins (titin and C-protein) present in the myosin filament. The reconstruction shows slightly more than two full 430 Å repeats of the myosin filament and the two successive 430 Å repeats are substantially similar to each other. Although these regions contain some overlapping image data, they were not averaged together during the analysis and the structural similarity of the successive 430 Å repeats can be taken as support for the validity of the reconstruction.Four views related by a 30° rotation around the filament axis are included in Fig. 3(A–D) to illustrate the different crown structures as well as their angular perturbation from helical symmetry. Surface density features in the 3D map, presumably mainly myosin heads, can be seen following a perturbed helical path. There are three levels of myosin heads within each 430 Å repeat, but they are not identical. On levels labelled 1 and 3, the projecting masses appear to have approximately the same azimuthal positions, whereas those on levels labelled 2 are at quite a different azimuth. The overall appearance suggests that there is on average a large departure from helicity in the azimuthal positions of the crown densities from the constant 40° angular rotation expected from ideal helical symmetry.The azimuthal and axial perturbations from helical symmetry are further demonstrated in Fig. 3(G) which shows the density in circumferential sections (like a radial net) through the 3D reconstruction at radii of 90, 110, 130 and 150 Å from the filament axis. For comparison, the circumferential section at a radius of 110 Å from the filament axis in the EM reconstruction of fish skeletal muscle myosin filaments (AL-Khayat et al., 2006) is shown as Fig. 3(F) where the same leftwards shift of mass on level 1 is seen.Analysis of the positions of centres of mass on each crown in Fig. 3(G) reveals an interesting feature. Fig. 4(A and B) show how the azimuthal and axial perturbations vary with radius. The azimuthal and axial displacements of density peaks on crowns 1 and 3 relative to those on crown 2 are plotted as a function of radius between 90 and 150 Å. Fig. 4(A) shows that there is a very clear trend for the angular separation of peaks on crowns 1 and 3 relative to crown 2 to tend towards the 40° expected for a perfect helix as the radius reduces. In Fig. 4(B) for the axial perturbations it is also conceivable that the axial shift is tending towards 143 Å at low radius between levels 2 and 3 as indicated by the extrapolated yellow lines. However, what happens between levels 1 and 2 is less clear cut.In the present reconstruction, as in the fish muscle myosin filament (Fig. 3(F)), additional density which may be C-protein is associated with crown 1. However, there also appears to be extra density running from crown 1 down to the adjacent crown 3 (Fig. 3(G)) which could also correspond to C-protein, titin, or density associated with the myosin rods in the backbone.4Discussion4.1Single particle analysis of vertebrate mammalian cardiac myosin filament structureThe technique of single particle analysis of non-helical filamentous systems involving the creation of particles by dividing the myosin filaments into segments has already been applied to produce a 3D reconstruction of the myosin head arrangement in relaxed vertebrate striated (fish) skeletal muscle (AL-Khayat et al., 2004, 2006). These filaments deviate from ideal helical symmetry and their structure can not be accurately determined by traditional Fourier-based helical 3D reconstruction. Our new approach has allowed the 3D analysis of EM images of these non-helical filaments without invoking helical symmetry (AL-Khayat et al., 2004, 2006). The method can preserve perturbations in the myosin head array within the 430 Å repeat length which are otherwise averaged out in helical reconstructions.The new approach was originally applied to the 3D structure of myosin filaments in vertebrate skeletal muscle, but until the present work it had not been applied to the medically important 3D structure of myosin filaments in vertebrate heart muscles. The current analysis not only had the benefit of extending the work from vertebrate skeletal to cardiac myosin filaments, but was also an improvement in that the analysis was carried out using only the C-zone part of the filament. Studying cardiac myosin filament structure in the normal state is an essential starting point from which to understand the mechanisms of the diseased system, in particular the effects of the mutations in myosin and its accessory proteins such as C-protein (MyBP-C) associated with different cardiomyopathies.Good evidence that our current 3D map is a reliable structure is that it helps to explain the features observed in the myosin filament images determined by 2D filtration by Kensler (2005a) using the same images that were used in our current single particle analysis. Fig. 5(I–L) shows projected views of our 3D map which reproduce the features seen in the filament images of Kensler (2005a). In particular what was described as the “comma-shaped” heads on level 2 (boxed density in Fig. 3(B) in Kensler, 2005a) is now seen as a projecting density which is curved upwards in the 3D map (Fig. 3(A–D)). The comma-shape feature is also boxed in Fig. 5(J and L)). The projected views of the 3D map are also consistent with the “saw-tooth” pattern of myosin head densities described by Kensler (2005a) (labelled by red lines in Fig. 4(B) in Kensler, 2005a) and clearly demonstrated in the 2D re-projection of the current 3D map (Fig. 5(I–L)). It is now evident that this “saw-tooth” pattern observed in both fish skeletal and rabbit cardiac myosin filaments is due to both the 3-stranded arrangement of the heads as well as the azimuthal perturbation inherent in both filaments.4.2Axial and azimuthal crown perturbationsComparison of our new map for rabbit cardiac muscle myosin filaments with that for fish skeletal muscle myosin filaments (AL-Khayat et al. 2006; Fig. 4) shows that the angular appearance of the single particle 3D reconstructions and the clear axial and azimuthal perturbations are common to both filament types. In both cases levels 1 and 3 have heads pointing on average in roughly the same azimuthal directions compared with level 2 which in both cases is rotated by roughly 60° compare Fig. 3(F) with Fig. 3(G; 110 Å). The radial perturbation is minor in both the fish and rabbit myosin filaments. Comparison between the angles and axial spacings/separations between the three levels in both fish skeletal and the current rabbit heart myosin filaments are shown in Table 1. It can be seen that the angular perturbation in rabbit is slightly larger than in fish. It can also be seen that the angular separations are closer to 60° in rabbit compared to the corresponding ones in fish. This could be due to the current C-zone analysis allowing better resolution of the angles than those in fish. Moreover, the axial spacing between levels 2 and 1 is lower in rabbit (−134.4 Å) compared to that in fish (−149.0 Å) whereas the spacing is larger between levels 2 and 3 in rabbit (154.0 Å) compared to fish (135.0 Å).The axial perturbations measured for rabbit cardiac myosin filaments by Kensler (2005a) between what we define as levels 2 and 1, and 2 and 3 were −132.9 Å and 153.0 Å, respectively, as measured from six filtered images. [Note that Fig. 4(E) in Kensler (2005a) referred to our levels 1, 2 and 3 as 2, 3 and 1, respectively. His level 0 in Fig. 4(E) (Kensler, 2005a) is our level 2.] Our new values based on the average of data plotted in Fig. 4 are very similar at −134.4 Å and 154.0 Å, respectively (Table 1).4.3Location of non-myosin protein densitiesThere is evidence from the earlier work on fish skeletal myosin filaments that the main C- (or X-) protein density is located at level 1. A similar conclusion can be reached for the rabbit cardiac myosin filaments (Fig. 3(E)) where the strongest projected density is also at level 1. A further important observation in AL-Khayat et al. (2006) was that the azimuthal perturbation between the three crowns probably adjusts the head locations to fit perfectly within the hexagonal filament lattice common to all vertebrate striated muscles so that the myosin heads do not clash with neighbouring actin filaments. The same conclusion can be reached for the cardiac myosin filaments studied here. That the myosin filaments from different species and from skeletal and cardiac muscles show similar perturbations and a similar location for C- (X-) protein argues that there may be a single basic three-dimensional plan common to all vertebrate striated muscle myosin filaments, presumably with minor variations.A feature seen in the fish myosin filament reconstruction is that there is axial density linking levels 3 and 1 (see Fig. 3(F)). Similar density is absent between levels 1 and 2 and 2 and 3. However, this linking density between levels 3 and 1 is rather more marked in the cardiac reconstruction (arrowed in Fig. 3(B), Fig. 3(G; 110 Å), arrowed in Fig. 5(J)) which could be due to the selection of particles in this case from only the C-zone. We therefore tentatively attribute this “extra” density to C-protein in the cardiac thick filament and infer that similar connecting C-protein density also occurs in fish skeletal muscle myosin filaments. In the X-ray diffraction modelling (see Fig. 5(b) and (d) in AL-Khayat and Squire, 2006) a similar region of longitudinal connecting density between levels 3 and 1 was also present. Since C-protein and other accessory proteins were not included in the analysis, the best fit model had the myosin heads on level 1 dipping down and almost touching the heads on level 3 to account for this connecting density. The current 3D-EM reconstruction suggests the alternative possibility that this axial connecting density may be due to the presence of C-protein at this location.Skeletal muscle C-protein is known to be a modular structure composed of ten fibronectin 3-like (Fn-3) and immunoglobulin I-like (Ig-I) domains (termed C1–C10), with an N-terminal Pro-Ala-rich domain; (Pro-Ala)-C1(IgI)-C2(IgI)-C3(IgI)-C4(IgI)-C5(IgI)-C6(Fn3)-C7(Fn3)-C8(IgI)-C9(Fn3)-C10(Ig) (Offer et al., 1973; Bennett et al., 1986; Vaughan et al., 1993). In the case of cardiac C-protein there is an additional N-terminal domain C0(Ig) before the Pro-Ala domain and there is an insertion between the first two Ig domains (C1 and C2) (Yasuda et al., 1995). The connectivity of mass in our reconstruction running between the projected myosin head masses of levels 1 and 3 (arrowed in Figs. 3(B) and 5(J)) could be titin or C-protein. If the density corresponds to C-protein, this may be consistent with the C-terminal part of the C-protein (domains C6 to C10) running axially along the axis of the myosin filament as proposed by Squire et al. (2003). It could be this longitudinal connecting density that causes the myosin head masses on levels 3 and 1 to be kept azimuthally aligned. The C0 to C5 part of C-protein is thought to be able to project out from the myosin filament and to bind to actin (Squire et al., 2003).A weak feature of uncertain origin in our new map and seen in the circumferential sections of the current 3D reconstruction at low radius (e.g. dashed red lines in Fig. 3(G; 90 and 110 Å radius), is a continuous circumferential density halfway between levels 1 and 2. Although this feature was less continuous in the 3D map of fish skeletal muscle (Fig. 3(F)), a similar circumferential band of density at a radius of 50 Å was present in the earlier 3D map of the isolated frog skeletal muscle thick filament (Stewart and Kensler, 1986; Fig. (11a); note that the numbering of the crossbridge levels in the frog myosin filament map is different from the current rabbit myosin filament map). It could be argued that this continuous density in the circumferential section of the rabbit cardiac myosin supports the model for the structure of C-protein proposed by Moolman-Smook et al. (2002) with a trimeric collar arrangement around the myosin filament. At present it is unclear whether the difference in the density between the rabbit cardiac and fish skeletal myosin maps is due to a genuine difference between the filaments or is a difference due to the present more selective analysis where only the C-zone area was included.We have further investigated the locations of non-myosin densities by investigating the circumferential sections at a radius closer to the backbone of the filament where the myosin heads do not contribute as much and therefore any densities seen may be attributed to myosin rods, C-protein or titin. This is shown in Fig. 3(I) which is a circumferential section at a radius of 75 Å from the filament axis in the current rabbit EM reconstruction. The red box in Fig. 3(I) highlights the circumferential density above crown 1, discussed earlier (Fig. 3(G; 90 and 110 Å dashed red lines)) that may conceivably be attributed to a C-protein collar. In addition to this there is a weak meshwork of density, possibly attributable to titin or other features of the filament backbone, in the regions between what are taken to be the myosin head masses on crowns 1 and 3 and 2 and 3 as shown arrowed in Fig. 3(I). These densities were not seen in the reconstruction of the fish skeletal myosin filaments of AL-Khayat et al. (2006) and may correlate with the strong meridional reflections on the 10th and 11th layer lines in the transforms of the rabbit cardiac myosin filaments (Kensler, 2005a, 2005b). The reflection on the 11th layer line corresponds to the 39 Å spacing expected for the domains of titin. This weak meshwork of density may correspond to the pattern of beaded striations at this spacing on the backbone reported by Kensler (2005b).4.4The arrangement of myosin heads in cardiac muscle myosin filamentsWe have investigated the possible organisation of the myosin head pairs in our new 3D map by asking whether the observed densities are consistent with the 2D crystal structure of vertebrate smooth muscle two-headed myosin determined by Wendt et al. (2001). In order to obtain the best fit and due to the fact that the crowns were not identical, the two-headed Wendt structure was modelled in different orientations in each crown (Fig. 5(A–H). The fitting was done for each crown individually within PyMOL by orienting the Wendt structure by hand within one of the three projecting mass densities on that crown. The other two 3-fold-symmetry-related Wendt structures on each crown were then positioned using the 3-fold transformation command in the CCP4 software package (http://www.ccp4.ac.uk/). It appears that crowns 1 and 3 can both accommodate the Wendt model quite well, but that crown 2 does not fit. This agrees with the conclusions for fish skeletal myosin filament (AL-Khayat et al., 2006) in that the Wendt structure fitted reasonably well into two crowns (levels 1 and 3) but not the third (level 2). We conclude that the Wendt et al. (2001) structure is not a totally universal myosin head arrangement even though it fits well to the helically averaged crowns in tarantula invertebrate striated muscle myosin filaments (Woodhead et al., 2005). Instead it appears to occur on two of the three levels of fish skeletal and rabbit cardiac muscle myosin filaments, but not on the third.Moreover, the Wendt model fits much better within crown 1 in the current rabbit cardiac myosin filament map than it did on crown 1 in the fish skeletal myosin. In addition, in the current reconstruction, the Wendt model fits somewhat less well within crown 3 than it did on the same crown in fish skeletal myosin. The projected myosin head density on level 2 in the current 3D map is more curved upwards (Fig. 3(A–D), Fig. 5(A–D)) than observed on the same level in the fish skeletal myosin filament (AL-Khayat et al., 2006). The difference in the fitting between the current map and the previous fish skeletal could be due to the fact that myosin filaments are different in cardiac and skeletal muscles. Obtaining a map at higher resolution would be beneficial in order to assess this difference in more detail.There could be several possible reasons for the difference observed in the fit of Wendt model to the third level of fish skeletal and rabbit cardiac myosin filament compared to the very good fit in tarantula myosin filaments (Woodhead et al., 2005). One of these possibilities could be that these four species have different regulatory systems in that that both vertebrate smooth (Wendt model) and tarantula striated muscles are myosin-regulated by phosphorylation of their regulatory light chains (Aksoy et al., 1976; Sobieszek, 1977; Craig et al., 1987), whereas both fish skeletal and rabbit cardiac muscles are regulated by calcium binding to their thin filaments (Kendrick-Jones et al., 1970; Lehman and Szent-Gyorgyi, 1975).Recent work by Jung and Craig (2007) reported preliminary results that show that myosin molecules isolated from tarantula and Limulus striated muscles (both phosphorylation-regulated) can have similar head–head interactions to those seen in vertebrate smooth and scallop striated myosin. Unlike these myosin-regulated muscles, head–head interaction was not observed in vertebrate skeletal and cardiac muscle myosins, in agreement with previous shadowing observations. However, their preliminary data showed that these unregulated myosin molecules can also exhibit head–head interactions if blebbistatin is bound to the heads (slowing phosphate release by promoting the ‘closed’ conformation of switch 2 in the motor domain of the head). Therefore the Jung and Craig (2007) results support the idea that myosin head–head interaction may be quite a common motif in the inhibited state of myosin II molecules. Nevertheless, vertebrate and invertebrate systems are evolutionarily widely separated. In addition to the fact that the various species have different regulation systems in that smooth, tarantula, Limulus and scallop myosins are all myosin-regulated whereas both skeletal and cardiac are thin filament-regulated, the tarantula, Limulus and scallop myosin filaments are all perfect/true helical systems. This true helical symmetry requires that there are the same orientations, configurations and interactions between the two-heads on each crown. On the other hand, vertebrate myosin filaments (skeletal and cardiac) are not true helical systems. The perturbation from helical symmetry could be reflecting differences in the configurations, interactions and functions of the two-headed myosin molecules on different crowns. It is therefore probably not surprising, as we have shown in the current rabbit cardiac myosin and the previous fish skeletal myosin (AL-Khayat et al., 2006), that the pairs of heads at the three crowns do not all share exactly the same arrangement (the Wendt model). The third possibility, along with the differences in regulation and helicity, is that tarantula, Limulus and scallop myosin filaments all have different compositions of accessory proteins (paramyosin core, etc.) compared to fish skeletal and rabbit cardiac (which have surface-located titin and C-protein). It could be that the extra surface proteins in vertebrate myosin filaments disrupt the fitting of the Wendt structure in one of the three levels, although it still fits well into the other two crowns. Fourthly, the differences in lattice arrangements between these species and the need to fit the underlying 9-fold structure of vertebrate myosin filaments into the 6-fold filament array, presumably requires there to be systematic differences in the head arrangements on different crowns.We therefore conclude that, even if the vertebrate myosin heads do sometimes tend to take up the Wendt configuration in isolation, as per the work of Jung and Craig (2007) on isolated myosin molecules, the vertebrate myosin filaments with their perturbed helical myosin head arrangement, the presence of surface accessory proteins, the special lattice organisation in vertebrates, and their different regulatory system may all reflect the need for heads to depart from this arrangement on one crown.4.5The origin of the perturbations in vertebrate striated muscle myosin filamentsA method to test the effect of C-protein on the myosin filament structure is to study a system which does not have C-protein. Harris et al. (2002) reported the development of a knockout mouse for cardiac C-protein which shows hypertrophy and significant contractile defects, but retains normal sarcomeric structure. In a recent study, Kensler and Harris (2008) examined the structure of myosin filaments isolated from the C-protein knockout mouse heart muscle. They reported that myosin filaments isolated from the knockout mouse hearts appeared similar in length and diameter to the wildtype filaments, but the “forbidden” meridional reflections, thought to derive from a perturbation from helical symmetry in the wildtype filament, were weaker or absent in the Fourier transforms of the cMyBP-C−/− myocardial myosin filaments. In addition, the crossbridge array in the absence of cMyBP-C appeared to be more easily disordered and sensitive to the negative staining conditions. These studies provide support for the idea that C-protein interactions with myosin may affect the myosin head arrangement as proposed by AL-Khayat et al. (2006). As noted by Kensler and Harris (2008), 2D studies of the filaments do not allow a determination of whether an azimuthal perturbation is present in the filaments. The large azimuthal perturbation observed for cardiac myosin filaments in this study would not give rise to the observed forbidden meridional reflections which arise from an axial crossbridge perturbation and/or the presence of additional mass on a 430 Å repeat. It is only 3D analysis, as in the present study, that allows a detailed and separate study of the axial and azimuthal components. This highlights the advantage of the single particle method in showing the azimuthal perturbation which is not evident from the observed forbidden meridionals in X-ray diffraction patterns and FFTs from 2D electron micrograph images.The 3D reconstructions of the vertebrate thick filament (both the current rabbit cardiac and the previously published fish skeletal myosin filaments (AL-Khayat et al., 2006)) have clearly shown the presence of a perturbation in the crossbridge arrangement, but until now have not addressed the question of whether the perturbation in the crossbridges extends to the packing of the myosin rods in the backbone. We have shown here that the rabbit cardiac 3D map contained perturbations that were pronounced at large radius and reducing at smaller radius. We also note that our current rabbit cardiac myosin filament map complements the reported abstract of the wildtype mouse cardiac myosin filament of Perez-Zoghbi et al. (2007) who have stated that the three crowns in their map were not identical. However, our evidence suggesting the greater observed helical perturbation of the heads at higher radius compared to at lower radius from the filament axis is distinct to our current study. Future higher resolution 3D analysis work in both wildtype and C-protein knockout mouse heart muscle myosin filaments, as well as in our current studies of the isolated rabbit cardiac myosin filament, should help to select between the alternatives: EITHER that the azimuthal and axial perturbations within the myosin filament are an inherent feature of the packing of the myosin rods in a non-helical way within the filament backbone, OR that the myosin head origins lie on a perfect helix and that the conformational perturbations are confined to the myosin heads only and are due to the presence of titin, which does not have a 143 Å repeat, and the effect of titin in locating C-protein on every third crown level. Our evidence so far from the present analysis is that the second of these options is more likely.\n\nREFERENCES:\n1. AksoyM.O.WilliamsD.SharkeyE.M.HartshorneD.J.A relationship between Ca2+ sensitivity and phosphorylation of gizzard actomyosinBiochem. Biophys. Res. Commun.6919763541130907\n2. AL-KhayatH.A.HudsonL.ReedyM.K.IrvingT.C.SquireJ.M.Myosin head configuration in relaxed insect flight muscle: X-ray modelled resting cross-bridges in a pre-powerstroke state are poised for actin bindingBiophys. J.85220031063107912885653\n3. AL-KhayatH.A.MorrisE.P.SquireJ.M.Single particle analysis: a new approach to solving the 3D structure of myosin filaments. ReviewJ. Muscle Res. Cell Motil.258200463564415750848\n4. AL-KhayatH.A.SquireJ.M.Refined structure of bony fish muscle myosin filaments from low-angle X-ray diffraction DataJ. Struct. Biol.155200621822916884926\n5. AL-KhayatH.A.MorrisE.P.KenslerR.W.SquireJ.M.3D structure of fish muscle myosin filaments by single particle analysisJ. Struct. Biol.155200620221716731006\n8. BennettP.CraigR.StarrR.OfferG.The ultrastructural location of C-protein, X-protein and H-protein in rabbit muscleJ. Muscle Res. Cell Motil.719865505673543050\n9. CantinoM.SquireJ.M.Resting myosin crossbridge organisation in frog muscle thick filamentsJ. Cell Biol.10219866106183484742\n10. CraigR.PadronR.Kendrick-JonesJ.Structural changes accompanying phosphorylation of tarantula muscle myosin filamentsJ. Cell Biol.1051987131913722958483\n11. CraigR.AlamoL.PadronR.Structure of the myosin filaments of relaxed and rigor vertebrate striated muscle studied by rapid freezing electron microscopyJ. Mol. Biol.22819924744871453458\n12. CrowtherR.A.PadronR.CraigR.Arrangement of the heads of myosin in relaxed thick filaments from tarantula muscleJ. Mol. Biol.1841985430439\n13. CrowtherR.A.HendersonR.SmithJ.M.MRC image processing programsJ. Struct. Biol.11619969168742717\n14. DeLanoW.L.The PyMOL User’s Manual2002DeLano ScientificSan Carlos, CA, USA\n15. EakinsF.AL-KhayatH.A.MorrisE.P.KenslerR.W.SquireJ.M.3D structure of fish muscle myosin filamentsJ. Struct. Biol.137200215416312064942\n16. HarfordJ.J.SquireJ.M.“Crystalline” myosin cross-bridge array in relaxed bony fish muscle: low-angle X-ray diffraction from plaice fin muscle and its interpretationBiophys. J.5019861451553730499\n17. HarrisS.P.BartleyC.R.HackerT.A.McDonaldK.S.DouglasP.S.GreaserM.L.PowersP.A.MossR.L.Hypertrophic cardiomyopathy in cardiac myosin binding protein-C knockout miceCirc. Res.90200259460111909824\n18. HudsonL.HarfordJ.J.DennyR.C.SquireJ.M.Myosin head configuration in relaxed fish muscle: resting state myosin heads must swing axially by up to 150 Å or turn upside down to reach rigorJ. Mol. Biol.27319974404559344751\n19. HuxleyH.E.Electron microscope studies on the structure of natural and synthetic protein filaments from striated muscleJ. Mol. Biol.16196328130814064165\n20. HuxleyH.E.BrownW.The low-angle X-ray diagram of vertebrate striated muscle and its behaviour during contraction and rigorJ. Mol. Biol.3019673834345586931\n22. Kendrick-JonesJ.LehmanW.Szent-GyorgyiA.G.Regulation in molluscan musclesJ. Mol. Biol.5419703133264250215\n23. KenslerR.W.Mammalian cardiac muscle thick filaments: their periodicity and interactions with actinBiophys. J.82320021497150811867464\n24. KenslerR.W.The mammalian cardiac muscle thick filament: crossbridge arrangementJ. Struct. Biol.1493200530331215721584\n25. KenslerR.W.The mammalian cardiac muscle thick filament: backbone contributions to meridional reflectionsJ. Struct. Biol.1493200531332415721585\n26. KenslerR.W.HarrisS.P.The structure of isolated cardiac myosin thick filaments from cardiac myosin binding protein-C knockout miceBiophys. J.9420081707171817993479\n27. KenslerR.W.StewartM.Frog skeletal muscle thick filaments are three-strandedJ. Cell Biol.961983179718026602135\n28. KenslerR.W.StewartM.An ultrastructural study of the cross-bridge arrangement in the fish skeletal muscle thick filamentJ. Cell Sci.9419893914012632577\n29. KenslerR.W.StewartM.The relaxed crossbridge pattern in isolated rabbit psoas muscle thick filamentsJ. Cell Sci.10519938418487691850\n30. KenslerR.W.WoodheadJ.L.The chicken muscle thick filament: temperature and the relaxed crossbridge arrangementJ. Muscle Res. Cell Motil.16199579907751407\n31. LabeitS.KolmererB.Titins: giant proteins in charge of muscle ultrastructure and elasticityScience27019952932967569978\n32. LehmanW.Szent-GyorgyiA.G.Regulation of muscular contraction: distribution of actin-control and myosin-control in the animal kingdomJ. Gen. Physiol.661975130125778\n33. LudtkeS.J.BaldwinP.R.ChiuW.EMAN: Semiautomated software for high-resolution single-particle reconstructionsJ. Struct. Biol.1281999829710600563\n34. Moolman-SmookJ.FlashmanE.De LangeW.LiZ.CorfieldV.RedwoodC.WatkinsH.Identification of novel interactions between domains of myosin binding protein-C that are modulated by hypertrophic cardiomyopathy missense mutationsCirc. Res.91200270471112386147\n35. MorrisE.P.SquireJ.M.FullerG.W.The 4-stranded helical arrangement of myosin heads on insect (Lethocerus) flight muscle thick filamentsJ. Struct. Biol.1071991237249\n36. OfferG.MoosC.StarrR.A new protein of the thick filaments of vertebrate skeletal myofibrils: extraction, purification and characterisationJ. Mol. Biol.7419736536764269687\n37. OfferG.KnightP.J.BurgessS.A.AlamoL.PadronR.A new model for the surface arrangement of myosin molecules in tarantula thick filamentsJ. Mol. Biol.298200023926010764594\n38. PatwardhanA.PaulD.AL-KhayatH.A.MorrisE.P.A measure for the angle between projections based on the extent of correlation between corresponding central sectionsJ. Mol. Biol.344200470772415533440\n39. PaulD.PatwardhanA.SquireJ.M.MorrisE.P.Single particle analysis of filamentous and highly elongated macromolecular assembliesJ. Struct. Biol.148200423625015477103\n41. SeidmanJ.G.SeidmanC.E.The genetic basis for cardiomyopathy: from mutation identification to mechanistic paradigmsCell104200155756711239412\n42. SjostromM.SquireJ.M.Cryo-ultramicrotomy and myofibrillar fine structure: a reviewJ. Micros.1111977239278\n43. SjostromM.SquireJ.M.Fine structure of the A-band in cryo-sections I: the structure of the A-band of human skeletal muscle from ultrathin cryo-sections negatively stainedJ. Mol. Biol.10919774968839534\n44. SobieszekA.Ca-linked phosphorylation of a light chain of vertebrate smooth-muscle myosinEur. J. Biochem.731977477483139309\n45. SquireJ.M.General model of myosin filament structure II: myosin filaments and cross-bridge interactions in vertebrate striated and insect flight musclesJ. Mol. Biol.7219721251384567398\n46. SquireJ.M.General model of myosin filament structure III: molecular packing arrangements in myosin filamentsJ. Mol. Biol.7719732913234765362\n47. SquireJ.M.Symmetry and three-dimensional arrangement of filaments in vertebrate striated muscleJ. Mol. Biol.9019741531604453009\n48. SquireJ.M.HarfordJ.J.EdmanA.C.SjostromM.Fine structure of the A-band in cryo-sections III: cross-bridge distribution and the axial structure of the human C-zoneJ. Mol. Biol.15519824674947086899\n49. SquireJ.M.LutherP.K.KnuppC.Structural evidence for the interaction of C-protein (MyBP-C) with actin and sequence identification of a possible actin-binding domainJ. Mol. Biol.3313200371372412899839\n50. StewartM.KenslerR.W.Arrangement of myosin heads in relaxed thick filaments from frog skeletal muscleJ. Mol. Biol.19219868318513495665\n51. StewartM.KenslerR.W.LevineR.J.Three-dimensional reconstruction of thick filaments from Limulus and scorpion muscleJ. Cell. Biol.10119854024112410430\n52. TajsharghiH.ThornellL.E.LindbergC.LindvallB.HenrikssonK.G.OldforsA.Myosin storage myopathy associated with a heterozygous missense mutation in MYH7Ann. Neurol.54200349450014520662\n53. TrinickJ.Cytoskeleton—titin as a scaffold and a springCurr. Biol.619962582608805236\n54. van HeelM.SchatzM.Fourier shell correlation threshold criteriaJ. Struct. Biol.151200525026216125414\n55. van HeelM.GowenB.MatadeenR.OrlovaE.V.FinnR.PapeT.CohenD.StarkH.SchmidtR.SchatzM.PatwardhanA.Single-particle electron cryo-microscopy: towards atomic resolutionQuart. Rev. Biophys.332000307369\n56. VaughanK.T.WeberF.E.EinheberS.FischmanD.A.Molecular cloning of chicken myosin-binding protein (MyBP) H (86-kDa protein) reveals extensive homology with MyBP-C (C-protein) with conserved immunoglobulin C2 and fibronectin type III motifsJ. Biol. Chem.2681993367036767679114\n57. VibertP.Helical reconstruction of frozen-hydrated scallop myosin filamentsJ. Mol. Biol.22319926616711542113\n58. WatkinsH.RosenzweigA.HwangD.-S.LeviT.McKennaW.J.SeidmanC.E.SeidmanJ.G.Characteristics and prognostic implications of myosin missense mutations in familial hypertrophic cardiomyopathyN. Engl. J. Med.3261992110811141552912\n59. WatkinsH.ConnerD.ThierfelderL.JarchoJ.A.MacRaeC.McKennaW.J.MaronB.J.SeidmanJ.G.SeidmanC.E.Mutations in the cardiac myosin binding protein-C gene on chromosome 11 cause familial hypertrophic cardiomyopathyNat. Genet.1119954344377493025\n60. WendtT.TaylorD.TrybusK.M.TaylorK.Three-dimensional image reconstruction of dephosphorylated smooth muscle heavy meromyosin reveals asymmetry in the interaction between myosin heads and placement of subfragment 2Proc. Natl. Acad. Sci. USA9820014361436611287639\n61. WoodheadJ.L.ZhaoF.-Q.CraigR.EgelmanE.H.AlamoL.PadronR.Atomic model of a myosin filament in the relaxed stateNature43620051195119916121187\n62. YasudaM.KoshidaS.SatoN.ObinataT.Complete primary structure of chicken cardiac C-protein (MyBP-C) and its expression in developing striated musclesJ. Mol. Cell Cardiol.271995227522868576942"
+}

batch_13/PMC2533003.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2533003",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2533003\nAUTHORS: Tristan RA Lane, Wilby J Williamson, Joshua M Brostoff\n\nABSTRACT:\nWe present the case of a 70 year-old man with carbon monoxide poisoning following a house fire. A significant smoking history and likely underlying chronic lung pathology complicated treatment, as due to symptomatic retention of carbon dioxide we were unable to use high-flow oxygen to facilitate the elimination of carbon monoxide. We suggest that patients with risk factors for obstructive lung disease be monitored extremely carefully during treatment for carbon monoxide toxicity.\n\nBODY:\nCase reportA 70 year-old retired man presented to the Emergency Department following a house fire. The flat below his had caught fire during the night, awakening the patient and his family. Smoke rose through the floors and windows, and the patient and his wife were exposed to significant smoke inhalation. Following rescue by the fire brigade and initial high-flow oxygen therapy by the ambulance service, the man arrived at hospital drowsy and less responsive than he had been at the scene.Past medical history included hypertension, type 2 diabetes, L4/5 disc prolapse. Medications were bendrofluazide, amlodipine, metformin, atorvastatin, and a multivitamin. He denied any allergies. There was no family history of note, and the patient was a non-drinker but had a significant smoking history of at least 20 cigarettes per day for over 50 years, and was a current smoker.Examination revealed no evidence of external burns, facial burns, singed nostril hair, hoarseness, stridor, or overt evidence of airway obstruction. However the patient was expectorating carbonaceous sputum. Initial blood pressure was 139/84, pulse 74 beats per minute, and pulse oximeter saturations of 100% on high-flow oxygen. The patient was oriented but drowsy with a GCS of 14/15, losing a point for eye-opening.Oxygen was removed, and a subsequent arterial blood gas analysis showed pH 7.4 (NR 7.35–7.45), pO2 10.1 kPa (NR > 10.6), and pCO2 of 5.46 kPa (NR 4.6–6.0). High-flow oxygen at 15 litres/minute with a reservoir bag was recommenced as the patient's pulse oximeter saturations fell to 92% on air. Over the next hour the patient became increasingly difficult to rouse and a repeat arterial blood gas analysis showed a pH of 7.32 and a raised pCO2 of 7.80 kPa. With the substitution of controlled flow oxygen at 0.24 FiO2 the patient rapidly became more alert, and despite persistent oxygen saturations of 94%, suffered no subjective or objective dyspnoea. A later arterial blood gas analysis showed normalisation of the pCO2 level.Blood results were unremarkable apart from a COHb level of 9% (NR < 2%), compared with his wife who had no symptomatology or clinical signs, but an initial COHb level of 11%.Repeat measurements four hours later showed that the patient's COHb level had dropped to 5%, while his non-smoking wife had a level of only 1%. Both patients were observed overnight and were discharged the following morning with no short-term ill effects of their exposure.DiscussionCarbon monoxide (CO) is a tasteless, odourless gas resulting from incomplete combustion. While almost certainly under-diagnosed, there are nearly 200 documented serious poisonings every year in the UK, occurring most commonly from house fires, faulty gas heaters and car exhausts [1]. CO has approximately 240 times the affinity of oxygen for binding to haemoglobin, and forms the COHb complex which impairs tissue oxygen delivery, inhibits mitochondrial oxidative phosphorylation, and inactivates cytochrome oxidase [2].There is a spectrum of clinical features from headache, nausea, and flu-like symptoms through to coma with hyperventilation, convulsions, pulmonary oedema, myocardial ischaemia and cherry-red skin colouring. Levels of COHb correlate poorly with clinical features, but it is generally accepted that an initial level of greater than 15% suggests significant toxicity.Management depends on the severity of the poisoning. The initial aims are removal from the source, and administration of high-flow oxygen. CO elimination has a dependent relationship with the FiO2: in room air the half-life of COHb is up to 5 hours; with high-flow oxygen and a reservoir mask this reduces to approximately 70 minutes; and hyperbaric oxygen reduces this further to about 25 minutes [3,4].Our difficulty in the above case was that the patient almost certainly had undiagnosed COPD, and hence when high-flow oxygen was given – the preferred initial management strategy – the patient retained CO2 and became increasingly obtunded, necessitating a reduction in the FiO2, and hence prolonging the clearance of carbon monoxide.The gold standard treatment for severe CO poisoning is hyperbaric oxygen therapy. This markedly raises the arterial oxygen level, and in COPD patients prone to CO2 retention would clearly cause significant elevation of pCO2. Furthermore emphysematous bullae may rupture under an elevated pressure, hence COPD is a relative contra-indication to hyperbaric oxygen therapy.A search of Pubmed found nothing published on the management of carbon monoxide poisoning in patients with chronic obstructive lung disease, and clearly a careful balance needs to be found between the level of administered oxygen, the patient's pCO2, and the required rate of clearance of CO.We suggest that patients with carbon monoxide poisoning and a significant smoking history – even if not formally diagnosed with COPD – have regular ABG analysis during treatment to ensure that they are not developing a dangerous respiratory acidosis. Carbon dioxide retention in such patients limits the use of uncontrolled high-flow oxygen, and thus in certain circumstances early intubation may need to be considered. The use of hyperbaric oxygen therapy in such patients should be considered only with extreme caution.AbbreviationsABG: Arterial blood gas; CO: Carbon monoxide; CO2: Carbon dioxide; COHb: Carboxy-haemoglobin; COPD: Chronic obstructive pulmonary disease; FiO2: Fraction of inspired oxygen; GCS: Glasgow coma score; NR: Normal range.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsCase report and discussion written jointly by JMB, WJW and TRAL. All authors have read and approved the final manuscript.ConsentWritten informed consent was obtained from the patient for publication of this case report. A copy of the written consent is available for review by the Editor-in-Chief of this journal.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2533103.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2533103",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2533103\nAUTHORS: Konstantinos D. Pantazis, Ioannis S. Elefsiniotis, Hero Brokalaki\n\nABSTRACT:\nThere is an obvious, significant, and diachronic reduction of the\nprevalence of HBV infection in Greece, concerning the general population as\nwell as some traditionally high-risk groups, mainly as a result of constant\ninforming and the widespread initiation of preventive and prophylactic\nmeasures, as well as the improvement of health care services. Nevertheless,\nthere are special groups and populations (economical refugees, religious\nminorities, HIV-positive patients, abroad pregnant women, prostitutes, etc.) who\nrepresent sacs of high HBV endemicity and need epidemiological supervision and intervention, in order to limit the spread of the infection and to further\nimprove the existing epidemiological data.\n\nBODY:\nHepatitis B virus (HBV) infection is a major public\nhealth problem worldwide, as more than 2 billion people have been infected,\nwhereas more than 350 million present chronic HBV infection. It is estimated\nthat a significant proportion (15–40%) of chronic HBV infected patients develop\nliver cirrhosis, liver failure, and primary hepatocellular carcinoma (HCC),\nmaking chronic hepatitis B one of the 10 major causes of death worldwide [1].The prevalence of the infection varies\namong different areas of the world and among special populations within the\nsame area. About 45% of the world population lives in countries in which HBV\ninfection is hyperendemic (South-Eastern Asia and sub-Saharan Africa),\nwhere the percentage of hepatitis B surface antigen (HBsAg) seropositivity outranges 8% of the total\npopulation and the percentage of viral exposure varies between 70% and 90%. In these countries, HBV infection\nis vertically transmitted usually during the perinatal period and horizontally\nduring the infancy and early childhood age, leading to extremely high rates of\nchronic HBV infection and preserving the reservoir of chronic hepatitis B\npatients worldwide [2, 3]. Southern countries of\nCentral and Eastern Europe as well as those of the Mediterranean basin, the Amazon’s sink, Middle East, and Northern\nAfrica are regions of medium HBV infection\nseroprevalence (HBsAg seropositivity between 2–7%), whereas\ncountries of North-Western Europe and North America are considered to be countries\nwith low HBV endemicity (HBsAg seropositivity <2%). In\nthese countries, the infection is basically transmitted horizontally during the\nadolescent and early adult period, mainly by high-risk sexual activity [4].Moreover, the annual incidence of HCC is extremely\nhigh in hyperendemic regions of HBV infection (10–150 cases/100.000 people per year) compared to regions of lower\nprevalence (1–3 cases of HCC/100.000 people per year), a finding that\nsupports the causative relationship of HBV and HCC. The primary HCC is one of\nthe most common tumours (5th in frequency\nworldwide) and is considered responsible for 300 000–500 000\ndeaths annually, representing the 3rd most common cause of death in the\nhyperendemic countries of HBV infection [5].Patients with chronic hepatitis B have a threefold\nprobability to die within a decade compared to general population, mainly by\ncauses directly related to chronic liver disease (HCC, hepatocellular failure,\nvariceal bleeding, etc.) [6]. Regarding\nthe cost of patients’ hospitalization, it seems that patients with chronic HBV\ninfection spend more time in the hospital and need more specialized and\nexpensive in-hospital treatment compared to the general population. An\nepidemiological and economotechnical study from a set area of Scotland showed a\nsignificant difference in the duration of hospitalization and the mean cost of\ntreatment, as well as in the number of admissions and readmissions of patients\nwith chronic HBV infection, compared to the control group [6]. The annual cost of\ntreating patients with chronic HBV infection has been retrospectively estimated\nand seems to significantly differentiate between patients in procirrhotic or\ncompensated cirrhotic stages of the disease under treatment and those who\nexhibit decompensated cirrhosis, especially when they undergo liver\ntransplantation procedure. Even the palliative treatment of end-stage HCC costs\nhigher than long-term treatment of early stages of the disease [7].A significant decline of the incidence of acute hepatitis\nB cases (67%) has been observed in the USA\nafter 1990, especially among\nhigh-risk groups (such as intravenous drug users, homosexual men, and health\ncare professionals), according to Center of Diseases Control (CDC). Widespread\nvaccination programs, predelivery evaluation of pregnant women and\npassive-active immunoprophylaxis of infants born from chronic HBV infected\nmothers, blood donors control, and immunization of the majority of health care\nprofessionals, children and teenagers have resulted in that decline.\nConsidering the lack of those preventive and precaution measures in many\nhyperendemic countries of the world and the massive immigration observed in the\nlast fifteen years, mainly from countries of median or high prevalence of the\ninfection to countries of low HBV prevalence, the epidemiological data of HBV\ninfection seem to be\nsignificantly modified nowadays.In Greece,\nthe existing published epidemiological data are difficult to be representative\nto the general population data, mainly because they are referred to specific\npopulations (blood donors, religious-ethnic minorities, economic immigrants,\nsoldiers, children, pregnant women) or high-risk groups (prisoners, drug users,\nhaemodialysis patients, HIV, prostitutes, etc.) over- or underestimating the\nproblem. The National Institute of Employment and the National Statistical\nService reported that the number of recorded foreign immigrants in Greece at the inventory of 2001 was\napproximately 800 000\n(7.2% of recorded Greek population) and the majority of them (65%) came from Albania, a country of high\nendemicity of HBV infection.In general, the comparison of studies from the early\n70s with others from the late 80s shows a significant decline of HBV prevalence\nin Greece\nof about 50–80%. Studies on military recruits exhibit a significant decline of\nHBsAg prevalence, from 4% in 1973 to 0.95% in 1999, mainly due to the\nmodification of socioeconomical and medical practice parameters and to the\nsignificant increase of successfully vaccinated people [8, 9]. General population published data\nare limited in Greece\nwhereas the best-studied group is that of blood donors. In the larger cohort of\nblood donors studied for 6 consecutive years in Athens, the mean prevalence of HBsAg was\n0.84% [10]. A study\nof a large population of blood donors in Crete island for 5 consecutive years revealed that the mean prevalence of HBsAg\n(0.4%) was significantly lower than the\none reported from other compartments of our country [11]. Equally, low\npercentages of HBsAg prevalence were observed in children of school age and\nteenagers of a rural population of Crete (0.33%) and in high-risk hospitalized\npatients (2.66%), reflecting partially the epidemiological data of the general\npopulation of the island [12, 13]. A significant decline of HBsAg prevalence by 2.15% and of the\npercentage of HBV exposure by 22.6% are reported in the first published study\nconcerning the general population, in the region of South-Western Greece,\nwhereas in the region of Epirus in a 3-year prospective study of blood donors,\nthe HBsAg seropositivity was of 0.85%, relatively higher than the\none reported from other compartments of Greece [14, 15].The immigration of populations from countries of high\nendemicity of the infection seems to contribute significantly to the rapid\nmodification of epidemiological data, especially in areas bearing the high\nburden of immigrants. In a serological control of 1020 refugees from South Albania who came and work in Ioannina region, the percentages of HBsAg\nseropositivity and HBV exposure were 22.2% and 70.6%, respectively. Moreover, a\nsignificant proportion of chronic HBV infected Albanian patients were also hepatitis\nB e antigen-positive [HBeAg(+)] (21.1%) and 12.7% of them were chronic hepatitis\nD virus (HDV) infected, representing a\npopulation of high infectivity regarding viral hepatitis [16]. Data regarding\nimmigrants living in Athens\nshowed that the prevalence of HBsAg was significantly high (15.4%) especially among Albanian and Asian\nrefugees [17].\nLikewise, significant differences of the prevalence of HBV infection are\nreported in the religious minority of Thrace moslems comparing to Greeks from formal Soviet Union who live in the same area\nand the native habitants of Thrace\n(9.3% versus 4.3% versus 3.4%, respectively), revealing a population of high\nendemicity of HBV infection within our country [18].Perinatal (vertical) transmission is one of the most\ncommon ways of HBV infection dispersion, mainly in hyperendemic areas of the\nworld and seropositive pregnant women and their neonates represent the basic\ntarget groups for the elimination of vertical transmission, which preserves the\nmain reservoir of chronic HBV infected patients. In the larger group of women\nat reproductive age studied in Greece, the mean prevalence of HBsAg was 1.15%,\nbut was significantly higher among women of Albanian (5.1%) and Asian origin\n(4.2%) compared to Greek women (0.29%). The vast majority (71.34%) of\nHBsAg-positive women were of Albanian origin and about a third of them\nexhibited significant viral replication at perinatal period [19].Intravenous drug users and prisoners are also\nhigh-risk groups for viral hepatitis who were studied epidemiologically in our country. In a study of\n544 prisoners who were intravenous drug users concomitantly, 6.5% of them\nexhibited HBsAg-positivity whereas among prisoners for sexual\noffences the percentage of HBsAg-positivity was higher enough (13%) [20, 21]. The widespread of HBV infection in the\nGreek drug-users community observed in the past decade, according to published\nstudies [22] seems\nto be changed nowadays. In an our recently published study, we investigated the\npresence of serological markers of HBV infection in intravenous drug users\n(IVDU) with chronic hepatitis C virus (HCV) infection and the results were\ncorrelated to the time of drug usage initiation. We found that drug use\ninitiation before 1992 was significantly related to HBV exposure while the vast\nmajority of relatively newer drug users (drug initiation after 1992) were HBV\nseronegatives [23].In Greek oncologic patients with solid tumors, the\npercentages of HBsAg-seropositivity were 5.3% and those of HBV\nexposure 44%, while a proportion of them (14%) had clinical and/or biochemical\nexacerbation of liver disease during the chemotherapy schedule; this percentage\nis expected to be much higher in patients with haematological malignancies,\nwith or without steroid administration according to the international\nliterature [24].\nFurthermore, significantly higher prevalence of HBV infection is observed in\nchronic alcoholic patients—with or without\nestablished liver disease—compared to\nhealthy blood donors or nonalcoholic hospitalized patients, which is ascribed\nto their specific and some times unexpected behavior [25].Chronic liver disease due to coexisting chronic viral\nhepatitis B and/or C as well as a consequence of hepatotoxicity of the\nantiretroviral therapy is one of the most important factors of morbidity and\nmortality of human immunodeficiency virus- (HIV-) infected patients, since the\nclassic opportunistic infections and the complications of severe\nimmunodeficiency have been significantly diminished. In HIV-seropositive\npatients the probability of chronic HBV infection and chronic liver disease\nafter exposure to HBV is extremely high and is characterized by very high\nlevels of viraemia (serum HBV-DNA) and extremely low percentages of spontaneous\nloss and/or seroconversion of HBeAg [26]. About 13% of a group of HIV-positive patients, mainly\nhomosexual men (68.5%), were HBsAg-positive and the percentage of their\nexposure to HBV infection was 67.4%, representing a high-risk group for viral\nhepatitis [27]. In\na cohort study in Greece, regarding 737 HIV-positive patients, the percentage\nof HBsAg-seropositivity was 12.1% while the majority of\nHIV/HBV coinfected patients (60.9%) were also HBeAg-positive and they exhibited\nextremely high levels of HBV viral load, suggesting the direct uptake of\npreventive measures for this high-risk group, in order to protect its health\nand the public health in general [28].An obvious trend of reduction of the HBV prevalence is\nobserved in other traditionally high-risk groups, such as prostitutes, since\nthe percentage of HBsAg-seropositivity (11%) at the early 80s has been\nreduced to 3.3% in the late 90s; these data need to be reevaluated because of\nthe high percentage of people trafficking for sexual reasons, mainly from\ncountries of median-high HBV prevalence to our country, who are not reported or\nsanitary controlled and are thought to represent the most significant cause of\nHBV transmission in our days [29, 30]. Data from a\nstudy of the Department of Public Order were presented in daily press and\naccording to conservative calculations there are more than 14 000\npeople who are illegally sexually prostituted in our country.Finally, the professional exposure in the health care\nworker field has been studied since the 70s, where the percentages of HBsAg\nseroprevalence were 2.4% for medical students and 4.6% for nursing staff. In a\nrecent study from Hippokration Hospital of Athens including 400 nursing staff,\nthe percentage of HBsAg-seropositivity was 1.25% and the HBV exposure\nrate was 16%, but only 61% of that high-risk group were efficiently vaccinated,\nso there is a significant percent of population unprotected for various but \nnonaccepted reasons [31].In conclusion, although there is an obvious,\nsignificant, and diachronic reduction of the prevalence of HBV infection in Greece, concerning the\ngeneral population as well as some traditionally high-risk groups, mainly as a\nresult of constant informing and the widespread initiation of preventive and\nprophylactic measures as well as the improvement of health care services, there\nare special groups and populations (economical refugees, religious minorities,\nHIV-positive patients, abroad pregnant women, prostitutes, etc.) who represent\nsacs of high-HBV endemicity and need epidemiological supervision and\nintervention, in order to limit the spread of the infection and to further\nimprove the existing epidemiological data.\n\nREFERENCES:\n1. LokASChronic hepatitis BThe New England Journal of Medicine2002346221682168312037146\n2. GustIDEpidemiology of hepatitis B infection in the Western Pacific and South East AsiaGut199638, supplement 2S18S238786047\n3. ChenC-JWangL-YYuM-WEpidemiology of hepatitis B virus infection in the Asia-Pacific regionJournal of Gastroenterology and Hepatology200015, supplement 2E3E610921373\n4. WHOGlobal Tuberculosis ControlWHO Report 2000, Geneva, Switzerland\n5. LavanchyDHepatitis B virus epidemiology, disease burden, treatment, arid current and emerging prevention and control measuresJournal of Viral Hepatitis20041129710714996343\n6. SteinkeDTWestonTLMorrisADMacdonaldTMDillonJFEpidemiology and economic burden of viral hepatitis: an observational population based studyGut200250110010511772975\n7. ButlerJRPiankoSKordaRJThe direct cost of managing patients with chronic hepatitis B infection in AustraliaJournal of Clinical Gastroenterology20043810, supplement 3S187S19215602169\n8. PapaevangelouGHepatitis B immunization programme: lessons learnt in GreeceVaccine199816, supplement 1S45S479915034\n9. StamouliMGizarisVTotosGPapaevangelouGDecline of hepatitis B infection in GreeceEuropean Journal of Epidemiology199915544744910442470\n10. KyriakisKPFoudoulakiLEPapouliaEISofroniadouKESeroprevalence of hepatitis B surface antigen (HBsAg) among first-time and sporadic blood donors in Greece: 1991–1996Transfusion Medicine200010317518010972911\n11. KoulentakiMSpanoudakisSKantidakiEPrevalence of hepatitis B and C markers in volunteer blood donors in Crete. A 5-year studyJournal of Viral Hepatitis19996324324810607237\n12. LionisCFrangoulisEKoulentakisMBiziagosEKouroumalisEPrevalence of hepatitis A, B, and C markers in school children of a rural area of Crete, GreeceEuropean Journal of Epidemiology19971344174209258547\n13. KoulentakiMErgazakiMMoschandreaJPrevalence of hepatitis B and C markers in high-risk hospitalised patients in Crete: a five-year observational studyBMC Public Health20011, article 171811208260\n14. GogosCAFoukaKPNikiforidisGPrevalence of hepatitis B and C virus infection in the general population and selected groups in South-Western GreeceEuropean Journal of Epidemiology200318655155712908721\n15. ZervouEKDalekosGNBoumbaDSTsianosEVValue of anti-HBc screening of blood donors for prevention of HBV infection: results of a 3-year prospective study in Northwestern GreeceTransfusion200141565265811346702\n16. DalekosGNZervouEKarabiniFTsianosEVPrevalence of viral markers among refugees from southern Albania: increased incidence of infection with hepatitis A, B and D virusesEuropean Journal of Gastroenterology and Hepatology1995765535587552639\n17. RoussosAGoritsasCPappasTSpanakiMPapadakiPFertiAPrevalence of hepatitis B and C markers among refugees in AthensWorld Journal of Gastroenterology20039599399512717844\n18. SklirosEALionisCFoudoulakiLSotiropoulosAKouroumalisESpandidosDHepatitis B and C markers in a Kurdish refugee camp in GreeceJournal of Gastroenterology and Hepatology200116783984011446899\n19. ElefsiniotisISGlynouIPantazisKDFotosNVMagaziotouIKadaHPrevalence of chronic HBV infection among 13,581 women at reproductive age in Greece: a prospective single center studyJournal of Clinical Virology200532217918015653423\n20. MallioriMSypsaVPsichogiouMA survey of bloodborne viruses and associated risk behaviours in Greek prisonsAddiction19989322432519624725\n21. GiotakosOBourtsoukliPParaskeyopoulouTPrevalence and risk factors of HIV, hepatitis B and hepatitis C in a forensic population of rapists and child molestersEpidemiology and Infection2003130349750012825735\n22. Roumeliotou-KarayannisATassopoulosNKarpodiniETrichopoulouEKotsianopoulouMPapaevangelouGPrevalence of HBV, HDV and HIV infections among intravenous drug addicts in GreeceEuropean Journal of Epidemiology1987321431463475213\n23. ElefsiniotisISHeroBMariolisASerological profile of HBV infection and liver histopathology among injecting drug users with chronic HCV infection in GreeceEuropean Journal of Internal Medicine200516749650016275544\n24. AlexopoulosCGVaslamatzisMHatzidimitriouGPrevalence of hepatitis B virus marker positivity and evolution of hepatitis B virus profile, during chemotherapy, in patients with solid tumoursBritish Journal of Cancer1999811697410487614\n25. DalekosGNZervouEMerkouropoulosMHTsianosEVPrevalence of hepatitis B and C viruses infection in chronic alcoholics with or without liver disease in Ioannina, Greece: low incidence of HCV infectionEuropean Journal of Epidemiology199612121258817173\n26. SorianoVMiróJMGarcía-SamaniegoJConsensus conference on chronic viral hepatitis and HIV infection: updated Spanish recommendationsJournal of Viral Hepatitis200411121714738553\n27. DimitrakopoulosATakouAHaidaAMolangeliSGialerakiAKordossisTThe prevalence of hepatitis B and C in HIV-positive Greek patients: relationship to survival of deceased AIDS patientsJournal of Infection200040212713110841086\n28. ElefsiniotisISMpotsiCPantazisKDPaparizosVLong-term follow-up of HIV/HBV coinfected patients from a Greek centerJournal of Clinical Virology200430328028115135750\n29. KaklamaniEKyriakidouATrichopoulosDPapoutsakisGKoumandakiIKaralisDHepatitis B serology in Greek prostitutes: significance of the different serum markersJournal of Hygiene19808422572617358965\n30. TsakrisAKyriakisKPChryssouSPapoutsakisGSeroprevalence of hepatitis B markers among female and transsexual sex workers in Athens, GreeceGenitourinary Medicine1997733223224\n31. Antypa-TheothoropoulouAPouyioukaMKarabassiVPapadouliZLamprinoudiAKontou-KastellanouCSero-epidemiologic survey of hepatitis B markers in healthcare workers of a major Greek HospitalJournal of Chemotherapy 200517p. 129"
+}

batch_13/PMC2533316.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2533316",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2533316\nAUTHORS: Robert Ehehalt, Richard Sparla, Hasan Kulaksiz, Thomas Herrmann, Joachim Füllekrug, Wolfgang Stremmel\n\nABSTRACT:\nBackgroundMechanisms of long chain fatty acid uptake across the plasma membrane are important targets in treatment of many human diseases like obesity or hepatic steatosis. Long chain fatty acid translocation is achieved by a concert of co-existing mechanisms. These lipids can passively diffuse, but certain membrane proteins can also accelerate the transport. However, we now can provide further evidence that not only proteins but also lipid microdomains play an important part in the regulation of the facilitated uptake process.MethodsDynamic association of FAT/CD36 a candidate fatty acid transporter with lipid rafts was analysed by isolation of detergent resistant membranes (DRMs) and by clustering of lipid rafts with antibodies on living cells. Lipid raft integrity was modulated by cholesterol depletion using methyl-β-cyclodextrin and sphingolipid depletion using myriocin and sphingomyelinase. Functional analyses were performed using an [3H]-oleate uptake assay.ResultsOverexpression of FAT/CD36 and FATP4 increased long chain fatty acid uptake. The uptake of long chain fatty acids was cholesterol and sphingolipid dependent. Floating experiments showed that there are two pools of FAT/CD36, one found in DRMs and another outside of these domains. FAT/CD36 co-localized with the lipid raft marker PLAP in antibody-clustered domains at the plasma membrane and segregated away from the non-raft marker GFP-TMD. Antibody cross-linking increased DRM association of FAT/CD36 and accelerated the overall fatty acid uptake in a cholesterol dependent manner. Another candidate transporter, FATP4, was neither present in DRMs nor co-localized with FAT/CD36 at the plasma membrane.ConclusionOur observations suggest the existence of two pools of FAT/CD36 within cellular membranes. As increased raft association of FAT/CD36 leads to an increased fatty acid uptake, dynamic association of FAT/CD36 with lipid rafts might regulate the process. There is no direct interaction of FATP4 with lipid rafts or raft associated FAT/CD36. Thus, lipid rafts have to be considered as targets for the treatment of lipid disorders.\n\nBODY:\nBackgroundUptake of long chain fatty acids (LCFAs) is important for many cellular functions and the understanding of the uptake mechanisms is an important target for treatment of lipid disorders [1-4]. The molecular mechanisms of fatty acid transport across the plasma membrane are still a matter of debate and the predominating mechanism likely differs from cell to cell (for reviews see [5-8]). In general, two possible groups of mechanisms are discussed: simple diffusion and saturable transport processes. Whereas the uptake based on the passive diffusion process depends on the activity of intracellular metabolism creating a transmembrane downhill concentration gradient; the saturable process is regulated by expression of certain proteins and lipids at the plasma membrane. Much effort has been spent to identify candidate proteins that are directly involved in the facilitated fatty acid uptake mechanism. So far four candidates (FABPpm, FAT/CD36, FATP family proteins, caveolin-1) have been discussed [9-11]. Whereas FAT/CD36 and FABPpm are membrane associated fatty acid binding proteins that are thought to mediate their dissociation from albumin and accumulation at the outer leaflet of the plasma membrane [12], followed by flip-flop across the phospholipid bilayer to the cytosolic site; it has been suggested that FATPs are real transporters directly involved in the uptake process of LCFAs across the membrane bilayer [5]. However, recent data from our laboratory suggest that this group of proteins are rather enzymes that indirectly facilitate the translocation process by encompassing acyl-coA activity [13].The composition of membrane lipids also modulates fatty acid uptake. In particular, there is increasing evidence that cholesterol is of crucial importance. Cholesterol depletion in 3T3 adipocytes, HMEC, HEK293 or HepG2 cells decreased LCFA uptake and this effect was reversible after re-addition of cholesterol [14-17]. Caveolin-1 has been suggested to regulate the cholesterol content of the plasma membrane [8,18] and LCFA uptake. Caveolin-1 can bind LCFAs [10]. Caveolin-1 knock out mice showed a reduced mass of adipocytes and increased serum free fatty acids, indicating that LCFA uptake into adipocytes might be impaired [19]. LCFA uptake is increased by caveolin-1 overexpression and inhibited in caveolin-1 knockout mouse fibroblasts [20] or by expression of the dominant negative caveolin-1 mutant CAVDGV [16]. Interestingly, the inhibitory effects of CAVDGV can be reversed by replenishing the cell membranes with cholesterol and can be mimicked by methyl-β cyclodextrin treatment [21].All these studies point out that cholesterol is critically involved in LCFA uptake. However, little is known about the mechanisms by which cholesterol modulates this process. We have previously hypothesized that the association of FAT/CD36 with lipid rafts might determine this process [8]. Rafts are lateral assemblies of sphingolipids and cholesterol within cellular membranes involved in compartmentalization of membrane processes [22]. Biochemically, lipid raft constituents are characterized by their insolubility in low concentration of detergents such as Triton X-100 [23]. In this regard, it has been demonstrated that a fraction of FAT/CD36 associates with detergent-resistant membranes (DRMs) in a cholesterol-dependent manner [16]. The reduction of overall LCFAs uptake by cholesterol depletion was as effective as the specific inhibition of the FAT/CD36 function by sulfo-N-succinimidyl oleate (SSO). Simultaneous treatments had no additional effect, suggesting that both procedures target the same cellular compartment [16].The most straightforward interpretation of these findings is that there are two pools of FAT/CD36 at the plasma membrane, one associated with lipid rafts, in which LCFAs are transported, and another outside of rafts, where no transport occurs. That means, FAT/CD36 has to reside in lipid rafts to facilitate FA uptake. Cholesterol depletion would shift FAT/CD36 into the surrounding lipid bilayer and make this protein non-functional. If FAT/CD36 facilitated uptake of fatty acids requires that the receptor is associated with lipid rafts, then regulation of this association may represent a mechanism by which cellular uptake of fatty acids can be regulated. In this paper we provide evidence demonstrating that association of FAT/CD36 with lipid rafts is critical for the uptake process to occur.MethodsReagents and antibodiesMethyl-β-cyclodextrin, sphingomyelinase and myriocin were from Sigma. Antibodies used were mouse anti-placenta alkaline phosphatase (PLAP) (Dako Cytomation), rabbit anti-PLAP [24]; rabbit anti-FATP4 (C4 anti FATP4 described in [13]); mouse anti-FLAG (Sigma), Cy5 and Cy3 donkey anti-mouse/rabbit (Jackson Immuno Research); mouse monoclonal antibody anti GFP 3E6 (Invitrogen); rabbit anti GFP KG77 [25]; mouse anti-FAT/CD36 (Biosource).ConstructsGFP-FATP4 have been described previously [13]. CD36-FLAG was kindly provided by Douglas M. Lubin, Washington University School of Medicine, St. Louis, MO. GFP-TMD is a membrane anchored version of the green fluorescent protein. The signal sequence of human CD8 (MALPVTALLLPLALLLHAARP) is followed by an epitope tag (VSV-G; YTDIEMNRLGK). Next is EGFP, followed by a 16 amino acid glycosylation tag from human rhodopsin (NGTEGPNFYVPFSNAT) and the transmembrane domain of podocalyxin (EFEDRFSMPLIITIVCMASFLLLVAALYGCCHRK). This plasmid is identical to the construct GFP-tail described in [26] except that the cytoplasmic tail has been removed and behaves like a non-raft protein [26].Oleate uptakeCumulative uptake of oleate was based on Stremmel and Berk [27]. Adherent cells were incubated for 5 min at 37°C with [3H]-oleate solution 170 μM (0.68 μCi/ml [3H]-oleate (Amersham)) BSA fatty acid free (Sigma) in PBS; After stopping and washing with ice-cold 0.5% BSA in PBS, cells were lysed with 1 M NaOH and aliquots analyzed for protein concentration (Biorad) and radioactivity by scintillation counting as before [13].Cells, transient transfection, cholesterol/sphingolipid depletionCOSJ (ATCC CRL-1651) and Vero (CCL-81) cells were maintained under standard tissue culture conditions with the appropriate culture media (COSJ: D-MEM Invitrogen 4,5 g glucose/L 10% FBS 2 mM L-Glutamine and Vero: D-MEM Invitrogen 1 g glucose/L 5% FBS 2 mM L-Glutamine). Cells grown to near confluency (10 cm2) were transfected with 4 μg plasmid-DNA and 10 μl lipofectamine 2000 (Invitrogen). Analysis was performed 16–20 h hours after transfection. For cholesterol depletion the cells were treated for 30 min with 10 mM methyl-β-cyclodextrin (MβCD) in DMEM. Cholesterol determinations were done using the Amplex Red Cholesterol Assay kit (Molecular Probes). For inhibition of sphingolipid synthesis the cells were seeded at a density of 0.2 × 106 in 10 cm2 dishes and were grown for 3 d in complete medium in the presence of 5 μM myriocin. During this period the medium was changed once. Growing COS cells in the presence of myriocin for a longer time caused the cells to detach and undergo apoptosis. The extent of sphingolipid-depletion was estimated microscopically on cells incubated for 30 min with 25 μg/ml of rhodamine-conjugated cholera toxin subunit B (Rh-CTB) (Molecular Probes) at 4°C, which was followed by a wash and subsequent fixation in 4% paraformaldehyde in PBS (PFA). The expression of FATP4 and FAT/CD36 for the different conditions were in initial experiments by Western blotting and found not to differ significantly (data not shown).Immunofluorescence and antibody-induced patchingFor immunofluorescence microscopy, cells were fixed for 4 min with 4% PFA at 8°C followed by an incubation in methanol at -20°C for 4 min. Fixed cells were incubated for 1 h at room temperature with the appropriate dilution of antibodies in PBS/0.2% gelatine (see below). After three washes with PBS/0.2% gelatine they were incubated with the respective secondary antibodies in PBS/0.2% gelatine for 1 h at room temperature.To aggregate raft proteins the respective antibodies were diluted in CO2 independent medium (GIBCO) containing 2 mg/ml BSA. The polyclonal antibodies against PLAP were diluted 1:35; the monoclonal anti-GFP (3E6) and anti-FAT/CD36 1:50. The cells were incubated for 45 min with the respective combination of antibodies at 10°C, briefly washed and further incubated for 45 min at 10°C with mixed fluorescently labelled secondary antibodies. Cy3-labelled secondary antibodies were diluted 1:500, and the Cy5-labelled ones 1:100. The cells were fixed as described above. Fluorescent images were acquired on an Olympus microscope and arranged with Adobe Photoshop.Preparation of detergent resistant membranes (DRMs)Detergent extraction with Triton X-100 was performed as described before for N2a cells [25]. Cells were grown in 3.5 cm dishes, transfected and 10–12 h later washed once with PBS and scraped on ice into 1.5 ml homogenisation buffer (250 mM Sucrose, 10 mM Hepes, 2 mM EDTA) and after centrifugation (5 min 2000 rpm) cell pellets were homogenized in homogenisation buffer containing 20 μg/ml each of chymostatin, leupeptin, antipain and pepstatin A (Sigma) through a 26 G needle and centrifuged for 5 min at 3000 rpm. The postnuclear supernatant was subjected to extraction for 30 min at 4°C in 1% Triton X-100. The extracts were adjusted to 40% OptiPrep (Axis-Shield) and overlaid in a TLS 55 centrifugation tube with 30% OptiPrep/TNE, and TNE (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA old protocol/25 mM Tris-HCl, pH 10.8, 150 mM NaCl, 5 mM EDTA new protocol since 03.2007). The gradients were centrifuged at 400000 g in a Beckman SW41 rotor for 20 h at 4°C. Fractions were obtained and used for Western blotting as described [16] and protein levels quantified by densitometry.Statistical analysisAll values are reported as mean and standard error of the mean (SEM). The Kruskal-Wallis test was used to test for statistical significance. Probability values of p < 0.05 were set as threshold for statistical significance.ResultsCholesterol depletion inhibits long chain fatty acid uptake in COS cellsTo study the role of lipid rafts in long chain fatty acid uptake we first analyzed again the effect of cholesterol depletion also in COS cells. Cholesterol depletion was done by treatment with the complexing agent methyl-β-cyclodextrin (MβCD) that extracts plasma membrane cholesterol. Immediately prior to testing uptake of [3H]-oleic acid, COS cells were treated with 10 mM MβCD for 30 min. By this the cholesterol levels could be reduced to 50% of those of control cells. To easily monitor the role of FAT/CD36 in fatty acid uptake cells were transiently tranfected with cDNA of FAT/CD36. Using lipofectamine as a mediator up to 50% of cells could be transfected. Overexpression of FAT/CD36 resulted in an increased overall uptake of [3H]-oleic acid by >20% within 5 min (Figure 1). Depletion of cholesterol with MβCD decreased fatty acid uptake by >50% in both control cells and FAT/CD36 transfected cells (Figure 1). These results show again that cholesterol is critically involved in [3H]-oleate uptake.Figure 1Cholesterol depletion inhibits [3H]-oleate uptake. COS cells were transiently transfected with FAT/CD36 and treated with 10 mM MβCD leading to a ~50% decrease in total cellular cholesterol (not shown). Afterwards [3H]-oleate uptake within 5 min was analysed. FAT/CD36 increased [3H]-oleate uptake significantly (p < 0.05). Cholesterol depletion decreased overall fatty acid uptake to less then 50% of control cells (p < 0.05). Data are expressed as mean and SEM of at least n = 6 experiments. The ratio has been arbitrarily set to 100% in cells that were not cholesterol depleted.Inhibition of sphingolipid synthesis decreases LCFA uptakeRafts are cholesterol and sphingolipid-enriched microdomains. To find out whether sphingolipids are also important for LCFA uptake, we tested the effect of inhibition of sphingolipid biosynthesis.COS cells were cultured in the presence of 5 μM myriocin, which inhibits the first step of de novo ceramide synthesis by interacting with the serine-palmitoyl transferase or treated with 1 U of sphingomyelinase that is supposed to deplete plasma membrane sphingolipids. Cells were either treated for 3 d with myriocin before cotransfection of FAT/CD36 and FATP4 or treated 1 d after cotransfection for 15 min with 1 U of sphingomyelinase. [3H]-oleate uptake within 3 min was measured. Both treatments resulted in a reduction of [3H]-oleate uptake by 17 ± 3% (myriocin) and 15 ± 2% (sphingomyelinase), respectively (Figure 2A). The ganglioside GM1 localizes to lipid rafts and its plasma membrane level is altered by inhibition of ceramide synthesis [28]. We monitored the effect of myriocin and sphingomyelinase by staining of the ganglioside GM1 with rhodamine-conjugated cholera toxin subunit B (Rh-CTB). Myriocin treatment led to a decrease in staining, suggesting a decreased amount of total sphingolipids (Figure 2B). Sphingomyelinase resulted in similar results (data not shown). Thus, [3H]-oleate uptake can also be altered by depletion of sphingolipids.Figure 2Myriocin and sphingomyelinase treatment decreases [3H]-oleate uptake. COS cells were either grown before cotransfection of FAT/CD36 and FATP4 for 3 d in the presence or absence (control) of 5 μM myriocin or were treated after co-transfection for 15 min with 1 U sphingomyelinase. (A) Overexpressing FAT and FATP4 increased [3H]-oleate uptake significantly (p < 0.05). Myriocin and sphingomyelinase treatment inhibited [3H]-oleate uptake significantly (p < 0.05). Data are expressed as mean and SEM from n = 3 independent experiments. (B) Staining of COS cells with Rh-CTB after 3 d treatment with 5 μM myriocin. The amount of the ganglioside GM1 on the surface is reduced after myriocin treatment. Bar: 200 μm. Both images in panel B were recorded with the same exposure time.FAT/CD36 co-patches with placental alkaline phosphatase and segregates from non-raft marker GFP-TMDLipid rafts are small and highly dispersed at the plasma membrane. In fibroblasts using photonic force microscopy it has been found that individual rafts have a size of about 50 nm and contain approximately 3000 sphingolipid molecules and probably a subset of 10–30 protein molecules [29]. Therefore they are too small to be resolved by light microscopy. However, raft and non-raft markers can be cross-linked with antibodies into visible patches [25,30,31]. Raft proteins co-patch with each other and segregate away from non-raft markers. Thus, cross-linking with antibodies can be used as a tool to analyse raft association. We therefore tested whether antibody cross-linking induces co-patching of FAT/CD36 with a raft marker, the glycosyl phosphatidylinositol (GPI)-anchored protein placental alkaline phosphatase (PLAP). As a non-raft marker, we used GFP-TMD. COS cells were transfected and the respective plasma membrane antigens cross-linked with antibodies recognizing the extracellular part of the membrane proteins. FAT/CD36 clearly co-localized together with PLAP at the plasma membrane in the majority of cells and segregated from GFP-TMD (Figure 3). As expected [13] FATP4 showed an intracellular pattern completely different from FAT/CD36. As it could not be detected at the plasma membrane no co-patching with FAT/CD36 (Figure 4) could be seen. This supports again our previous data [13] that FATP4 is an enzyme at the ER level, indirectly involved in fatty acid uptake due to enzyme activity.Figure 3Co-patching of PLAP and GFP-TMD (Podo) with FAT/CD36. 20 h after transient transfection, Vero cells were incubated for 45 min at 10°C with respective primary antibodies, washed and incubated for 45 min with mixed Cy5 and Cy3 fluorescently-labelled secondary antibodies. PLAP and GFP-TMD were patched with polyclonal rabbit anti PLAP and polyclonal serum KG77 against GFP, respectively. Patching of FAT/CD36 was achieved with monoclonal antibody from Biosource international. FAT/CD36 co-localized with the raft marker protein PLAP and segregated from GFP-TMD. Bar: 10 μm.Figure 4FAT/CD36 does not co-localize with FATP4. 20 h after transient transfection with FAT/CD36 and FATP4-GFP, FAT/CD36 was clustered with anti-FAT/CD36 antibody from Biosource and Cy3 fluorescently labelled secondary antibody. (A) FAT/CD36 is mainly localized at the plasma membrane. (B) FATP4-GFP shows a reticular staining pattern as described before [13] representing ER membranes. (C) Patched FAT/CD36 does not show any co-localisation with FATP4-GFP.Antibody induced cross-linking increases the association of FAT/CD36 to detergent resistant membranesVarious proteins associate to lipid rafts with different kinetics and partition coefficients [32]. It has been shown that antibody-induced patching may stabilize association of raft proteins with detergent-resistant membranes (DRMs) [22,31]. We therefore analysed the association of FAT/CD36 to DRMs under cross-linking with antibodies.In initial experiments COS cells were transfected with FAT/CD36 and then extracted with Triton X-100 and subjected to OptiPrep™ step gradient centrifugation. Under those conditions in three independent experiments no significant effect of antibody cross-linking could be detected. FAT/CD36 was found in both Triton X-100 resistant and soluble fractions showing that there are two pools of FAT/CD36 in cellular membranes. However, when FATP4 was co-expressed with FAT/CD36 raft association of the FAT/CD36 was increased (Figure 5). Co-expression of the cytosolic green fluorescent protein (GFP) or the ER marker sec61-GFP [33] did not change DRM association of FAT/CD36 (data not shown). Performing cross-linking under co-expressing conditions revealed a clearly increased fraction of FAT/CD36 in the DRM fraction (Figure 5). Therefore, DRM association of FAT/CD36 can be increased by cross-linking with antibodies, which probably reflects increased raft affinity caused by oligomerization. Similar results have been demonstrated for other proteins, which by forming oligomers increase their raft association [34-39]. The fact that increased association of FAT/CD36 to DRMs occurs especially in FATP4 overexpressing cells indicates a promoting effect of FATP4 for FAT/CD36 raft association.Figure 5Effect of antibody cross-linking on association of FAT/CD36 to DRMs. 20 h after transient transfection of FAT/CD36, FATP4 or both COS cells were were lysed in 1% Triton X-100/TNE at 4°C. (A) After floatation in an OptiPrep step-gradient FAT/CD36 was found in two pools, in DRMs and in soluble membranes (lane 1–3). Co-expression of FATP4 resulted in an increased relative amount of FAT/CD36 found in DRMs (lane 6). Antibody cross-linking of FAT/CD36 using an mouse anti human FAT/CD36 antibody from Biosource shifted FAT/CD36 towards the DRM fractions (lane 6). No significant amount of FATP4 was found in DRMs, indicating that FAT/CD36 and FATP4 might be in distinct compartments within the cell. Flotillin-2, a typical raft protein was used as a control to estimate the quality of DRM isolation. The results are representative of three others experiments carried out independently. (B) Quantification; FATP4 expression and antibody-induced patching significantly increased the amount of FAT/CD36 in the top two fractions (DRM associated). The amount in the top two fractions was correlated to the total amount of protein in all fractions. Data are expressed as mean and SEM of n = 3 experiments. Asterisk indicates significant differences to cells transfected with FAT/CD36 only (p < 0.05).Cross-linking of FAT/CD36 with antibodies increases oleate uptakeIf long chain fatty acid uptake were to take place in lipid rafts, then antibody cross-linking should not only induce co-patching with raft markers at the surface of living cells and increase association between FAT/CD36 and DRMs. It should also increase LCFA uptake, provided that the antibody does not neutralize LCFA binding or transport To find this out, we analyzed the effect of antibody cross-linking on overall [3H]-oleate uptake. Cells were co-transfected with FATP4 and FAT/CD36 and [3H]-oleate uptake analyzed within 5 min in the presence of anti-FAT/CD36 from Biosource (Figure 6). Antibody enhanced cross-linking indeed increased [3H]-oleate uptake significantly. We next examined the effect of cholesterol depletion on antibody-induced [3H]-oleate uptake. Immediately prior to the uptake assay transfected cells were treated for 30 min with 10 mM methyl-β-cyclodextrin and again uptake within 5 min was analyzed. After cholesterol depletion with methyl-β-cyclodextrin, [3H]-oleate uptake was no longer enhanced by antibody cross-linking (Figure 6). Thus, under conditions in which rafts are supposed to be disrupted (cholesterol depletion) increased uptake of [3H]-oleate due to cross-linking with antibodies is not detectable. Thus, lipid raft integrity is important for FAT/CD36 function in fatty acid uptake.Figure 6Effects of antibody cross-linking (x-link) and cholesterol depletion on [3H]-oleate uptake. Cells were transiently transfected with FLAG-FAT/CD36 and FATP4 and then treated or not for 30 min with 10 mM methyl-β-cyclodextrin (MβCD). Afterwards overall [3H]-oleate uptake within the first 5 min in presence or absence of anti-FAT/CD36 from Biosource was analysed. Quantification of three independent experiments is shown. The ratio has been arbitrarily set to 100% in cells neither cross-linked nor cholesterol depleted. Cross-linking increased overall fatty acid uptake significantly (p < 0.05). However under conditions of cholesterol depletion, antibody enhanced fatty acid uptake was not apparent.DiscussionThe data presented here strengthen the evidence that FAT/CD36 partition into lipid rafts and that LCFA uptake mediated by FAT/CD36 depends on this association.First of all, we could show that LCFA uptake was critically dependent on the integrity of lipid rafts. Lipid rafts are cholesterol and sphingolipid enriched microdomains. Removal of raft lipids from cells leads to disruption of raft functions [22]. Consistent with previous results on other cells [14-17] we could demonstrate that by decreasing cellular levels of cholesterol, LCFA uptake in COS cells was also inhibited by > 50%. Additionally, we could show that by decreasing ceramide synthesis with myriocin or depleting membrane sphingolipids using external sphingomyelinase we were able to lower LCFA uptake. Sphingolipid depletion was less dramatic, probably because it is more difficult to lower cellular sphingolipid levels than cholesterol levels. The effects of sphingolipid synthesis inhibition may also be more pleiotropic because there are many different kinds of sphingolipids (sphingomyelins and glycosphingolipids). Nevertheless, we found that plasma membrane GM1 levels were reduced after myriocin treatment, demonstrating that not only cholesterol but also sphingolipids modulate LCFA uptake.Another important finding supporting raft association was that after cross-linking with antibodies FAT/CD36 co-patched with placental alkaline phosphatase (PLAP), a GPI-anchored raft protein, in living cells and segregated from patches formed by cross-linking of GFP-TMD that served as a non-raft marker. This assay has previously been used to monitor if proteins associate with lipid rafts at the cell [25,39]. Antibody cross-linking of surface proteins leads to the formation of plasma membrane clusters that can be easily observed during light microscopy. It has been shown that co-patching is dependent on plasma membrane cholesterol, as patching is inhibited by cholesterol removal [31].Antibody cross-linking increased DRM association of FAT/CD36 as was previously demonstrated for other raft proteins [25,31]. Because a considerable amount of FAT/CD36 was also found in the soluble fraction after detergent treatment, FAT/CD36 is probably found (in steady state) in two membrane pools, one raft associated and another localized in the surrounding bilayer. How partitioning between these two pools is regulated is an open question, but interestingly it has been shown that FAT/CD36 can dimerize [36]. Oligomerization of raft components is known to lead to increased raft affinity [37,38]. Therefore this mechanism might be important to regulate raft association of FAT/CD36.It has been previously hypothesized that the detergent soluble fraction might represent intracellular FAT/CD36 and that upon cholesterol depletion FAT/CD36 might be stuck in the secretory pathway to the plasma membrane [16]. It has been described, that FAT/CD36 can indeed translocate between the plasma membrane and intracellular compartments [40]. However, recent data from Covey et al [17] using immunofluorescence and flow cytometry convincingly show that in 3T3-L1 adipocytes cholesterol depletion inhibits uptake of LCFAs without affecting FAT/CD36 or caveolin-1 distribution within the cells, indicating that cholesterol levels regulate LFCA uptake via a pathway that does not involve altered surface localization of FAT/CD36. Our cross-linking data, which were obtained on living cells at 10°C, can explain these findings because under these conditions intracellular antigens are not accessible to antibodies. Therefore increased DRM association of FAT/CD36 (Figure 5) after cross-linking with antibodies likely represent a horizontal recruitment of plasma membrane non-raft associated FAT/CD36 into raft domains. The enrichment of DRM associated FAT/CD36 could only be detected in presence of co-expressed FATP4. Whether the FATP4 acitivity as acyl-CoA synthetase drives cellular fatty acid influx and, thus, the requirement for raft associated FAT/CD36 or whether FATP4 is actively involved in the raft constitution process remains to be elucidated. FATP4 has been originally described to by a major fatty acid transporter at the apical membrane of enterocytes [11] and it has been speculated that both proteins might cooperate at the plasma membrane in fatty acid uptake [5,41]. However recent data could not support this view [13]. FATP4 seems rather to be an enzyme for acyl-CoA activity at the ER level and therefore indirectly increases fatty acid uptake. Deletion of FATP4 resulted in a perinatal lethality with a phenotype reminiscent of lethal restrictive dermopathy [42]. By histology the skin was characterized by a hyperproliferative hyperkeratosis with a disturbed epidermal barrier. Lipid analysis of the skin revealed an increased proportion of ceramides and cholesterol. Within the ceramide fraction very long chain fatty acid substitutes were significantly reduced [42]. These finding suggests that FATP4 is involved in ceramide biosynthesis, possibly by providing CoA activity for very long chain fatty acids. It is intriguing to speculate that this protein might therefore have also impact on lipid raft integrity. Rafts are enriched in sphingolipids (that are made from ceramides) and their long chain fatty acid substitutes have been implicated to be important to connect the outer and the inner layer of the microdomain. FATP4 would therefore be indirectly involved in the fatty acid uptake process by providing the molecules for a functional lipid raft assembly. This could explain our results of an enhanced raft association of FAT/CD36 after overexpression of FATP4.In general clustering of lipid rafts can be achieved by different mechanisms [22]. Besides antibodies also other ligands (e.g. insulin), lectins or linker proteins have been discussed. Which would be a reasonable mechanism for FAT/CD36 clustering? It is conceivable that under physiological conditions the clustering of FAT/CD36 is mediated by binding of long chain fatty acids. It was indeed shown that the artificial, long chain fatty acid [3H]-sulfo-N-succinimidyl oleate (SSO) – when incubated with adipocytes – binds to FAT/CD36 with high affinity and is almost exclusively found in DRMs [16]. Thus the concentration of fatty acids presented to the plasma membrane may regulate the rate of fatty acid uptake by providing a sufficient functional platform in the form of FAT/CD36-raft complexes.ConclusionOur data support a crucial role for lipid rafts in LCFA uptake. Compartmentalization of FAT/CD36 at the cell surface by lipid rafts seems to be important in regulating its involvement in LCFA uptake. Taken all data together it seems possible that fatty acid binding increases raft affinity of FAT/CD36. The protein is shifted from the surrounding bilayer to these microdomains were LCFA uptake is likely to occur.AbbreviationsLCFA: long chain fatty acids; SSO: sulfo-N-succinimidyl oleate; DRM: detergent resistant membrane.Authors' contributionsRE precipitated in the design of the study and carried out the immunofluorescent experiments. He drafted the manuscript. RS carried out the uptake and the floatation experiments. JF, WS, HK and TH participitated in the design of the study and helped to draft the manuscript. All authors read and approved the final manuscript.AchnowledgementsWork was supported by Stiftung Nephrologie and the Dietmar-Hopp-Stiftung\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2535604.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2535604",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2535604\nAUTHORS: Tim Eisen, Athena Matakidou, Richard Houlston\n\nABSTRACT:\nBackgroundPart of the inherited risk to lung cancer is likely to include common, low risk alleles. The identification of this class of susceptibility is contingent on association-based analyses. We established GEnetic Lung CAncer Predisposition Study (GELCAPS) to collect DNA and clinico-pathological data from a large series of cases and a series of spouse/partner controls, thereby generating a key resource for the identification of low risk alleles.MethodsGELCAPS was one of the first genetic epidemiological trials in the UK to be adopted by the National Cancer Research Network (NCRN) onto its portfolio with the participation of over 100 oncology departments specialising in the management of lung cancer.ResultsSamples from over 5,000 independent lung cancer cases and 2,000 controls have so far been assembled through GELCAPS.ConclusionGELCAPS represents one of the largest datasets of its type in the world capable of informing on the contribution of low penetrance alleles to the development of lung cancer and the influence of genetic variation on outcome. In addition our experience in developing the GELCAPS serves to illustrate how large DNA biobanks for genetic analyses can be rapidly generated within the UK using the NCRN.\n\nBODY:\nBackgroundLung cancer is a major cause of cancer mortality worldwide [1]. In the United Kingdom, it accounts for more than 33,000 cancer deaths each year (Cancer Research UK). The disease is frequently cited as a malignancy solely attributable to environmental exposure, principally tobacco smoking. It has, however, long been postulated that individuals may differ in their susceptibility and there is strong evidence from epidemiological studies for a familial risk [reviewed in [2]]. Direct evidence for a genetic predisposition is provided by the increased risk of lung cancer associated with a number of rare Mendelian cancer syndromes, such as in carriers of germline TP53 [3]and RB [4,5] mutations, as well as in patients with Bloom's [6] and Werner's [7] syndromes.The two major types of lung cancer, non-small cell lung cancer (NSCLC), and small cell lung cancer (SCLC) account for 75% and 25% of cases respectively. Although the histological features are different between these (reflected in differences in patterns of gene expression), there are similarities in the spectrum of underlying somatic genetic alterations suggesting commonality in pathogenesis. Moreover, the observation that the familial risks are not subtype dependent [8-13] and that histological concordance between affected family members is poor [9] is consistent with the hypothesis of a \"generic\" inherited susceptibility to lung cancer.The genetic basis of inherited susceptibility to lung cancer outside the context of the rare Mendelian cancer predisposition syndromes is at present undefined, but a model in which major gene loci account for the excess familial risk seems unlikely. One hypothesis about the allelic architecture of susceptibility proposes that part of the genetic risk is caused by disease loci, which include common, low penetrance alleles. This \"common-disease common-variant\" hypothesis implies that conducting association analyses based on scans of Single Nucleotide Polymorphisms (SNPs) should be a powerful strategy for identifying low-penetrance variants [14,15].Previous studies aimed at identifying low penetrance alleles for lung cancer susceptibility have largely been based on a candidate gene approach formulated on preconceptions as to the role of specific genes in the development of the disease. Perhaps not surprisingly most studies have to date only evaluated a restricted number of polymorphisms, primarily in genes implicated in the metabolism of tobacco-associated carcinogens and the protection of DNA from carcinogen-induced damage. However, without a clear understanding of the biology of lung cancer predisposition the definition of suitable genes for the disease is inherently problematic making an unbiased approach to loci selection highly desirable.Despite much research, few definitive low penetrance susceptibility alleles for lung cancer have been to date unequivocally been identified through candidate-based association studies. As with many other diseases, positive associations have been reported for various polymorphisms of genes such as GSTT1 [16], GSTM1 [17], ERCC2 [18], CYP1A1 [19], and TP53 [20] from small studies, but few of the initial positive results have been replicated in subsequent studies. The inherent statistical uncertainty of case-control studies involving just a few hundred cases and controls seriously limits the power of such studies to reliably identify genetic variants conferring modest but potentially important risks.In addition to genetic variation affecting the risk of developing cancer it is increasingly being recognised that genetic variation, not necessarily in the same genes, may also affect clinical outcome. As with case-control association studies aiming to identify novel susceptibility alleles the same issues of study power pertain to the search for prognostic markers and such studies are again contingent on access to large case-series.Following the sequencing of the human genome, large-scale harvests of SNPs have been conducted and > 10 million documented. Patterns of linkage disequilibrium (LD) between SNPs have been characterised allowing subsets of SNPs (tagging SNPs) to be selected that capture a large proportion of the common sequence variation in the human genome. This coupled with the advent of highly efficient analytical platforms allow whole genome-wide studies (GWAS) for disease associations to be conducted cost effectively. The relationship between patients' genotype and risk of lung cancer is now open for exploration.The identification of genes associated with cancer predisposition and determination of their contribution to disease incidence are contingent on having DNA samples from large, systematic series of cancer patients. The resulting genetic epidemiological data provides the information on which to base the identification, counselling and management of at-risk individuals. The National Cancer Research Network (NCRN) was established to provide support for clinical cancer research in England and is one of the most substantial and constructive developments in the area of cancer research to be made in recent years in the United Kingdom. In England, serving a population of 50 million people, the NCRN is made up of 34 geographically distinct Networks covering the entire country. Within each Network there are clinical research support staff and infrastructure to promote accrual of patients into trials and studies, and the collection of high quality clinico-pathological data and appropriate biological samples. Hence the NCRN presents a major scientific initiative not only in the field of clinical trials but also in the field of genetic epidemiology.To create a resource for identifying low penetrance alleles for lung cancer we established GELCAPS (GEnetic Lung CAncer Predisposition Study) in March 1999 to collect DNA and clinico-pathological data from a large series of lung cancer cases. Within 5-years of setting up the initiative by linkage with the NCRN it has been possible to create a world-class resource of biological and clinico-pathological data from over 5,000 individuals with lung cancer.Methods/DesignEligibility criteriaAll patients diagnosed with lung cancer between March 1999 and July 2004 were eligible for the study. To ensure that data and samples were collected from bona fide lung cancer cases and avoid issues of bias from survivorship only incident cases with histologically or cytologically (only if not adenocarcinoma) confirmed primary disease were ascertained. Partners of recruited lung cancer patients with no personal history of cancer were recruited as controls.Procedural outlineA standardised questionnaire was used to collect basic demographic characteristics-sex, date of birth, ethnic group (White, Black-Caribbean, Black-African, Black-other, Indian, Pakistani, Chinese, Other), country of birth, current area of residence – in addition to details on active and past smoking history (including type of tobacco product, amount smoked, age at first cigarette and age at any major change of smoking habits), exposure to asbestos, occupational history, and personal past medical history. All questionnaires were self-administered and no surrogate responders were used. An open question was used to obtain information on family history of cancer involving first-degree relatives. A positive history of lung cancer was only assigned when detailed information was provided identifying the family member affected by lung cancer. The referring clinician using a standard registration form supplied clinico-pathological details (type of lung cancer, stage at presentation) of patients.Coupled with patient recruitment their spouses/partners who had no known past or current history of malignancy were invited to participate for the purposes of contributing to the generation of a control series. For these individuals details of sex, date of birth, ethnic group, place of birth, current area of residence and smoking history were collected through a self-administered questionnaire. 10–20 ml EDTA-venous blood samples were collected from all participants. Consent forms, questionnaires, registration forms and blood samples were returned to the Institute of Cancer Research (ICR) by mail. Blood samples collected were stored at -80°C prior to DNA extraction and quantification.It is our intention to collect outcome data on all cases entered into GELCAPS. In the first stage of this process subsets of participating centers were asked to provide the clinical details on the outcome of the recruited lung cancer patients. Records were requested based on their date of accrual, with those accrued at the beginning of the study being requested first. A standard proforma was used to collect information on diagnosis, stage, treatment and survival. Fully informed consent was obtained from all patients alive at the time of outcome data collection. Outcome forms were returned to the ICR by mail and details were stored electronically.Statistical considerationsThe primary aim of establishing GELCAPS was to generate a DNA resource of lung cancer patients sufficiently large to robustly identify low penetrance alleles by association studies of genetic polymorphisms. From the outset we envisaged that at some juncture such searches would be conducted on a genome-wide basis. It is well recognised that as such studies involve typing a vast number of markers, a large number of false positive associations will inevitably be generated and only a small number of markers will be truly associated with disease susceptibility. Hence associations need to attain a high level of statistical significance to be established beyond reasonable doubt and significance levels of ~10-7 have been proposed as being appropriate [14]. The original target of GELCAPS was to assemble a series to include ~2,000 cases. This figure had been arrived upon on the basis of upon contemporaneous views of the probable impact of common alleles on disease risk. During development of GELCAPS studies of other common diseases indicated that common disease alleles are likely to be associated with risks typically in the range of 1.1–1.5. To identify alleles conferring such risks is contingent on sample sets twice that of our original target and we therefore revised our target accordingly in order to have ~80% power to identify an association between SNP genotype and risk.Ethical considerationsIn generating DNA registries such as GELCAPS ethical considerations are central to study design. One of the particular strengths of studies such as GELCAPS is that once constructed the DNA database can be probed repeatedly for different existing and newly identified candidate risk factor genes. It is not feasible to contact all study entrants to seek further written consent for specific test therefore, the information sheet and study discussion was centred on the general concept of 'genetic analyses'. As these investigations were to be solely for research to find new gene(s) predisposing to cancer it was implicit that no individual results will be conveyed to persons. In publications of findings no study entrant would be identifiable. As with all studies of this nature we clearly stated that if a study entrant wished to withdraw their DNA sample and all information held on them would be destroyed. To ensure confidentiality data is held under secure conditions at the ICR Institute of Cancer Research and information held on study entrants will not be divulged to any person or agency without the prior written agreement of the study entrant.All clinical information and biological samples were obtained only after fully informed consent was obtained from participating individuals, and in accordance with the tenets of the Declaration of Helsinki. Ethical approval for the study was obtained from the London Multi-Centre Research Ethics Committee (MREC/98/2/67) and local ethical committees. Personal information was stored in accordance with the Data Protection Act (1998).Extraction of DNA, storage and quality assuranceDNA was extracted from EDTA blood samples using either a standard salt extraction procedure or using the Chemagen system (Chemagen Biopolymer-Technologie AG, Arnold-Somerrfield-Ring 2, 5499 Baeswelder, Germany, Picogreen quantified (Quant-it, Invitrogen, Paisely, UK) and normalised to 100 ug/ul in TE buffer. DNAs stocks are being stored in Eppendorf tubes (Barkhausenwe 1 22339 Hamburg, Germany) at -80°C. To avoid subjecting stock DNAs being to repeated thawing and freezing we have generated a series of \"master\" 96 deep well plates of samples from which DNAs can be readily robotically abstracted for genotyping studies. Fidelity of DNA is being constantly evaluated by monitoring performance in the different genotyping platforms.ResultsAfter securing ethical permissions at a national level through the Multi Research Ethics Committee we started recruitment to GELCAPS in March 1999. Ascertainment of cases was restricted to 28 centres and accrual was maximally 10–20 patients per month. After GELCAPS was incorporated into the NCRN (National Cancer Research Network) portfolio in March 2002 it was subsequently rolled out across England after individual centers had obtained local ethical permissions. Adoption by the NCRN was associated with a significant increase in patient and control accrual (Figure 1). Eventually 140 oncology centers (Figure 2) became active participants in GELCAPS with patient ascertainment averaging ~100 cases per month. The remit and operational procedure by which patients are accrued to NCRN adopted studies does not allow collection of compliance data within each participating center. However, we estimate based on our intimate knowledge of the clinical activities of three centers that patient accrual to GELCAPS is ~70% of those invited to participate.Figure 1Accrual of cases and controls to GELCAPS.Figure 2Centres in the UK recruiting to GELCAPS after NCRN adoption.The original target of GELCAPS was to assemble a series of 2,000 lung cancer cases. Given the efficiency by which samples were being accrued following adoption of GELCAPS by the NCRN a new target of at least 4,000 cases was deemed to be eminently feasible within the time frame for which funding had been secured.We terminated accrual to GELCAPS in July 2004 by which time samples from 5,269 cases with primary lung cancer and 2,094 controls had been recruited. The majority of cases were male (64%) reflecting the sex preponderance of disease. Whilst the mean age of controls was comparable to cases (62.9 years, SD = 10.6) not surprisingly 69% were female (Table 1). Similarly, the prevalence of smoking was significantly higher amongst cases compared to controls. A high proportion of the cases ascertained had been diagnosed with lung cancer at a young age (Table 1); specifically 1,617 (~31%) of the cases were aged less than 60 years old at diagnosis, compared with < 10% in the general population. The frequency of the various forms of lung cancer was, however, in keeping with that observed in UK general population – ~23% being affected with SCLC and ~73% with NSCLC (Table 1).Table 1Characteristics of lung cancer patients recruited to GELCAPSCasesControlsTotal5,2692,094Gender (Male: Female)3,382 (64.2%)645 (30.8%)Age at diagnosis (years)887 (35.8%)446 (69.1%) <4053 (1.0%)64 (3.1%) 40–49302 (5.7%)166 (7.9%) 50–591,146 (21.7%)514 (24.5%) 60–691,868 (35.4%)756 (36.1%) 70–791,590 (30.1%)510 (24.4%) 80+310 (5.9%)84 (4.0%) Mean (SD)65.1 (10.0)62.9 (10.6)Ethnicity Arabic3 (0.05%)1 (0.05%) Asian7 (0.13%)6 (0.29%) Bangladeshi1 (0.01%)0 Black-African4 (0.08%)1 (0.05%) Black-Caribbean31 (0.59%)8 (0.38%) Black-Other1 (0.01%)0 Indian16 (0.30%)4 (0.19%) Jewish-Ashkenazi12 (0.22%)6 (0.29%) Pakistani8 (0.15%)0 White5,065 (96.1%)1,947 (93.0%) Other/not specified82 (1.6%)121 (5.8%)Reported asbestos exposure807 (15.2%)125 (6.0%)Family history of lung cancer750 (14.2%)212 (10.1%)Smoking habits Never-smokers307 (5.8%)718 (34%) All smokers Age first started smoking (SD)16.5 (4.0)17.9 (4.8%) Pack years in Smokers (SD)47.2 (30.6)30.4 (22.1%)Histology of cancerSmall cell (SCLC)1,193 (22.6%)Non-small cell (NSCLC)3,815 (72.4%) Squamous1,905 (49.9%) Adenocarcinoma (including variants)1,110 (29.1%) Large cell10 (0.3%) Brochoalveolar44 (1.2%) Adenosquamous11 (0.3%) Neuroendocrine20 (0.5%) NSCLC unspecified715 (18.7%)Sarcoma5 (0.1%)Unclassified primary256 (4.9%)Tumour stage at presentation, by histology SCLC  Limited168 (50.4%)  Extensive165 (59.6%) NSCLC  I151 (14.1%)  II140 (13.1%)  III457 (42.7%)  IV323 (30.2%)To date we have acquired follow up data on 1,187 patients; specifically, information on the staging, management and clinical outcome permitting comparison patients randomly drawn from the general population. Stage at presentation for each of the different subtypes of lung cancer was similar to that observed in the general population; specifically, for patients with SCLC, somewhat less than half (43%) presented with limited disease and of the patients with NSCLC, 13% had stage I, 15% had stage II, 43% had stage III, and 29% had stage IV disease. The majority of patients with limited stage SCLC had been treated with a combination of radical radiotherapy and chemotherapy, whilst all patients received chemotherapy. The main treatment modality for SCLC patients with extensive disease was chemotherapy. Patients with early stage NSCLC (stage I and II disease) were mainly treated with surgical resection of the primary tumor whilst about one third received chemotherapy and radical radiotherapy. The mainstay treatment modality of patients with stage III and IV NSCLC was chemotherapy. Overall the median survival time (MST) for the subset of 1,187 GELCAPS patients was 18.6 months. Prognosis was significantly correlated with stage at presentation, with those presenting early have a far better survival (Figure 3). Patients with SCLC had a MST of 26.0 and 10.5 months if diagnosed with limited and extensive disease respectively. For those with NSCLC, MSTs ranged from 12.1 months for stage IV patients to 32.3 months for stage I disease.Figure 3Survival from lung cancer in patients according to stage at presentation: A) Patients with SCLC, B) Patients with NSCLC. In both SCLC and NSLC survival was significantly better (P < 0.0001) in patients presenting with early stage disease compared to those presenting with late stage disease in both Log rank tests of the difference in distribution of survival curves and in Cox-proportional hazard test, adjusting for age, sex, year of presentation and treatment with platinum-based chemotherapy. Statistical analyses performed using STATA version 8.0 (College Station, Tx, USA).DiscussionRecent data from GWASs of breast [21,22], prostate [23-27] and colorectal (CRC) cancer [28-31] provides strong evidence for the involvement of common disease-causing alleles and suggests that a relatively large number of genes influence the aetiology in most cancers in the patient population as a whole.To exploit the advances brought about by the human genome projects, future work in cancer genetics will be dependent upon the acquisition of large well-characterised cohorts of cancer cases. Here we have demonstrated that the centralisation of cancer services in the UK offers an opportunity to establish large, well-characterised cohorts by targeting collection to the largest centres. Moreover mobilising NCRN networks provides a means of delivering consistently the data and sample collection to complete genetic epidemiology studies, relating to the detection of main effects on the required scale.Because ascertainment of cases through GELCAPS has been based on clinical centres specialising in the treatment of lung cancer a high proportion of cases have been diagnosed young. While this means cases are not fully representative of disease in the general population the distribution of age at diagnosis serves to empower GELCAPS for identifying disease-causing alleles by virtue of genetic enrichment.Given that constitutional genotypes may well influence patient prognosis it is highly desirable that survivorship is not confounding influence on sample collection. As survival rates in patients recruited to GELCAPS were not significantly different to those documented in previously published audits of lung cancer in the UK there is no evidence that \"healthy study participant\" selection will have genetically biased ascertainment. For all participants, sex, ethnicity and age at sampling have been documented. The geographical area of birth and area of residence within the UK is known for all of the individuals and this information can be used to allow analyses stratified by region of residence, reducing any effects of population stratification. The possibility of population stratification leading to false inference of disease-genotype association can readily be addressed by adjusting for known region/ethnicity or by using information on unlinked genetic markers.We acknowledge the potential problem of differential bias in genotyping samples accrued from different sources. Although the samples collected through GELCAPs have been ascertained from many clinical centres we have no evidence that this has affected sample quality as we have previously documented call rates of 99.8% in samples genotyped for 1,500 SNPs [32-35] and Quantile-Quantile plots of test association statistics provide no evidence for differential biasThe NCRN research networks are established within cancer care networks where access to partners is readily available and direct. They are not designed to collect samples from the general population so our choice of collecting samples from partners was a pragmatic one appropriate for the NCRN. Inevitably in studies such as GELCAPS a smaller number of samples from controls will be collected than from cases since in addition to lack of compliance many patients do not have a current partner. The sex of controls ascertained through initiatives such as GELCAPS will usually be of the opposite gender to cases, and controls are potentially over-matched with respect to many lifestyle risk factors. Theses limitations can be offset to a large degree by using samples collected from the healthy spouses/partners of one cancer as a source of controls for a different cancer. This is something we are currently pursuing with respect to a similar NRCN sponsored initiative the National Study of Colorectal Cancer Genetics (NSCCG)Because of the difficulty of obtaining sufficiently detailed data on environmental exposure in studies such as GELCAPS, and because there are issues to do with comparability of exposure data from controls assembled from different studies, it is acknowledged that studies of environmental risk factors including gene-environment interaction will be limited in resources such as GELCAPS. The main value of collections such as GELCAPS will be in studies of genetic risk factors and gene-gene interactions; hypotheses regarding gene-environment interaction require alternative datasets, such as the European Prospective Investigation into Cancer and nutrition (EPIC) study [36], which are centred around population based-cohorts. Accepting such limitations our experience in developing GELCAPS serves to illustrate how large DNA databases for genetic analyses can rapidly be developed in the UK. At present we have only collected outcome data on around 20% of cases recruited to GELCAPS. By completing the collection of follow up data on all cases we shall be able to assemble a unique series for examining the influence of constitutional genotype on clinical outcome in the population setting.ConclusionFinally, it is noteworthy that the value of GELCAPS has been demonstrated in a recent GWAS of lung cancer we have conducted in which we have been able to robustly identify a susceptibility variant for the disease mapping to 15q [37].AbbreviationsGELCAPS: Genetic Lung Cancer Predisposition StudyCompteting interestsThe authors declare that they have no competing interests.Authors' contributionsTE and RSH were the principal investigators for the GELCAPS, devised the study. AM helped in study development and was responsible for database design, and management of study coordinators. All authors contributed to the paper.Availability & requirementsCancer Research UK: GELCAPS: National Cancer Research Network: NHS Cancer Plan: National Study of Colorectal Cancer Genetics: European Prospective Investigation into Cancer and Nutrition: Pre-publication historyThe pre-publication history for this paper can be accessed here:\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2536668.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2536668",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2536668\nAUTHORS: Thomas Borchert, Joerg Krueger, Annette Hohe\n\nABSTRACT:\nBackgroundVariety protection is of high relevance for the horticultural community and juridical cases have become more frequent in a globalized economy due to essential derivation of varieties. This applies equally to Calluna vulgaris, a vegetatively propagated species from the Ericaceae family that belongs to the top-selling pot plants in Europe. We therefore analyzed the genetic diversity of 74 selected varieties and genotypes of C. vulgaris and 3 of Erica spp. by means of RAPD and iSSR fingerprinting using 168 mono- and polymorphisms. The same data set was utilized to generate a system to reliably identify Essentially Derived Varieties (EDVs) in C. vulgaris, which was adapted from a method suggested for lettuce and barley. This system was developed, validated and used for selected tests of interest in C. vulgaris.ResultsAs expected following personal communications with breeders, a very small genetic diversity became evident within C. vulgaris when investigated using our molecular methods. Thus, a dendrogram-based assay to detect Essentially Derived Varieties in this species is not suitable, although varieties are propagated vegetatively. In contrast, the system applied in lettuce, which itself applies pairwise comparisons using appropriate reference sets, proved functional with this species.ConclusionThe narrow gene pool detected in C. vulgaris may be the genetic basis for juridical conflicts between breeders. We successfully tested a methodology for identification of Essentially Derived Varieties in highly identical C. vulgaris genotypes and recommend this for future proof of essential derivation in C. vulgaris and other vegetatively propagated crops.\n\nBODY:\nBackgroundCalluna vulgaris L. (Hull.), an exclusive species within the genus Calluna, has increased its economic weight, and not only within the German horticultural industry over the last few decades: In 2005 almost 100 million plants were produced in Germany, of which about 30% were exported to other European countries [1] where the demand is also still increasing. Although merely a handful of breeders are commercially active in breeding C. vulgaris, more than 300 varieties now exist, which are or have been protected at the Bundessortenamt, Hannover (BSA) [2] and/or the Community Plant Variety Office (CPVO), Angers, France [3]. More than 50% of applications for variety protection at the CPVO date from 2003 or later, which supports the argument of the increasing importance of C. vulgaris.Breeding efforts in C. vulgaris primarily aim at a special type of its inflorescence, the so-called bud flowers (Fig. 1). Flowers of these plants do not open during the entire reproduction phase from August to December which makes them appear visually attractive for a long period of time when not many other flowering ornamental outdoor plants are available in the northern hemisphere. This phenotype is closely linked with and possibly caused by a lack of anthers. This connection, in turn, has a severe impact on breeding methods because interesting bud-flowering genotypes are only applicable as the female parent in crossings. In addition, there is only sparse information and hypotheses available concerning the inheritance of this trait. Therefore – and since C. vulgaris is a vegetatively propagated crop – breeding in C. vulgaris over the past few decades was to a large extent performed by selection of spontaneous mutations, rather than by systematic crossings (personal communications with breeders). The actual variety composition in Europe offers a mixture of normal flowering and bud flowering types (state: 01/2008) with main focus on the latter (~85%). Some special forms (e.g. 'Radnor' with filled flowers or 'Peace' as a multi-bracteate type) are present as well. However, due to the problems described above, the actual gene pool used in breeding of C. vulgaris is presumably quite narrow.Figure 1Shoots of two C. vulgaris genotypes representing the main inflorescence types. left: normal ('White Mite'), right: bud ('Anneliese').Therefore, in this study the genetic diversity within the species C. vulgaris was examined with molecular DNA techniques, comprising a selection of 64 economic important and partially still-protected varieties from Germany, including varieties from other European countries and the USA, 5 genotypes resulting from crossings, as well as a selection of 5 wild plants of different origin. Moreover, 3 different genotypes of Erica spp. were included in this study as an anticipated outgroup [see Additional file 1].In the case of C. vulgaris, variety protection assessments as executed by the CPVO and as described in the Protocol for Distinctness, Uniformity and Stability Tests for Calluna L. (Hull.), LING, Scots Heather (CPVO-TP/94/Final of 06/11/2003), comprises 22 phenotypic traits in total but only 18 traits for bud-flowering varieties, which are tested in 2 flowering seasons with 30 plants (replications). Herein, problems arise from continually increasing applications for protection of bud-flowering genotypes, from their partial identicalness in many of these traits and from the subjectiveness that is inherent in the measurement of phenotypic traits. Moreover, breeding of bud flowering types requires backcrossing, which is – in contrast to mutant selection or 'cosmetic breeding' – 'true breeding', but which also contributes to the narrow gene pool. Previously, these drawbacks led to some juridical disputes in the field of variety derivation in C. vulgaris in Germany.The problem of variety derivation and the need for an appropriate protection system was already identified decades ago and is especially pressing in the context of global marketing. In Europe, the Act of Convention from 1991 followed on from a Convention on the Protection of Plant Varieties [4] and first introduced the term 'Plant Breeder's Rights'. Today it is acknowledged by 65 member states (, state: 01/2008). Variety protection in these member states is based on DUS-tests (distinctness, stability, uniformity: see above). Despite increasing testing efforts, these tests remain sketchy since the investigated traits may be influenced by several factors, e.g. environmental changes, and are evaluated by subjective ratings so that molecular markers have become a desirable tool [5]. But also the so-called 'fingerprinting techniques' – although widely recommended as a supplement to phenotypic tests e.g. by [6] – entail a number of problems, since only a random sample i.e. a subset of the genome, can be examined. Therefore, any statistical method applied to this problem has to be able to maintain a delicate balance in order to avoid excessive identification of false positives on the one hand, as well as false negatives on the other [6-8].Several case studies have been recently published in the context of ED-conflicts, however, for the most part these do not concern vegetatively propagated species, since it is generally assumed that variation in such varieties does not occur, which would allow clear-cut molecular genotyping. [9] tested essential derivation in various vegetatively propagated ornamentals (Rhododendron, Rosa, Phalaenopsis) by AFLP-genotyping and constructing UPGMA-dendrograms. However, these investigations relied on the assumption of total genetic stability within vegetatively propagated varieties and therefore dispensed with any statistical analysis. From our point of view, this is not appropriate for all vegetatively propagated species, because – for example with Calluna – phenotypic variations (sports) are well-known and are based on genotypic variation. From these experiences we support EDV-identifying systems with respect to statistical validation as with the one introduced by [10] for lettuce and barley [11], which is based on the definition of a minimum distance (threshold) for distinctness. Such procedures are necessary since proving identity is more difficult than proving distinction with molecular markers [12]. Lettuce is a self-fertilizer and consequently genetic variation within varieties can be expected to be very low. Moreover total variation between today's cultivars should be somewhat reduced due to an intensive breeding history. For this reason, [11] suggested that ED-conflicts should not be analyzed through the construction of a dendrogram visualizing hypothetical kinship relations, but instead by the examination of all pairwise genetic distances within an appropriate reference population, and then comparing these results to the distance between actual varieties in question.As a result, another aim of our study, drawing on the publication by [11], was to implement a comparable concept of identification of EDVs in C. vulgaris based on molecular data resulting from RAPD and iSSR techniques. Our system proposal is critically evaluated with regard to essential premises e.g. variation and stability [13], its success in C. vulgaris, and its practicability in the future.The results presented here were obtained during a BMWi-(Federal Ministry of Economics and Technology) funded cooperation between the IGZ and a German breeding company (Heidepflanzen Peter de Winkel, ). Thus, variety denotation is ciphered in cases where the breeder's interests may be affected.ResultsEstimation of genetic diversity and kinship relations within C. vulgarisUsing RAPD- and iSSR-techniques, we achieved a total of 129 (RAPD) and 39 (iSSR) distinguishable and reproducible bands. This corresponds to 9.9 bands/RAPD primer and 7.8 bands/iSSR primer. The combined results of RAPD and iSSR studies are shown in the dendrogram in Fig. 2. While the three Erica genera do cluster as an outgroup, all tested genotypes from the Calluna species cluster to the right of one node. Interestingly, the wild-types from Thuringia (Ruhla) and from the Italian Alps (San Remo) cluster as an additional outgroup within the Calluna species while the other wild-types available (Löhnstein, Niederohe, Tiefenthal, all from the Lüneburger Heide in Germany) are grouped within the rest of the Calluna genotypes.Figure 2Dendrogram consisting of 74 C. vulgaris and 3 Erica spp. genotypes. Constructed from 168 mono- and polymorphisms amplified from 13 RAPD and 5 iSSR-primers and based on the Dice/Nei and Li coefficient with subsequent UPGMA-clustering. Nodes with strong support (> 85%) by bootstrapping (n = 10.000, PHYLIP) are marked with **, moderately supported groups (50% – 85%) are marked with *, varieties of interest for the involved company are ciphered by CV# where # is replaced by increasing numbers. Variety encryption is known to the authors and the company, respectively. For purposes of clarity and according to their regional provenance, genotypes have been classified by symbols as indicated.The statistical significance of our data was investigated with the resampling method of bootstrapping as initially described by [14] using the software Winboot and n = 10,000 replications. Those few nodes with moderate support (50% < p < 85%), as well as strong support (p ≥ 85%), which appeared both in the NTSYSpc-constructed tree as well as in the majority-rule consensus tree of Winboot, are marked with * and **, respectively, in Fig. 2. The linked genotypes to the right of these nodes may be considered to be linked in real kinship. Despite the high number of analyzed bands, all other linkages are statistically unconfirmable within the present data set.Identifying EDVs in C. vulgarisDue to former juridical conflicts concerning property rights of varieties in the genus Calluna we endeavored to develop a reliable statistical system for identifying EDVs in this species based on the results from the first part of our study. Since the dendrogram analysis did not support statistically significant decisions on kinship relations and probably would not do so even after analysis with a clearly expanded data set, we decided to implement a method based on a procedure published by [11] for similar analyses in lettuce. We therefore created appropriate Reference Sets of 25 varieties (Table 1) for each pair of tested genotypes (Test Set) in question and then computed primer-wise and pair-wise similarity values within each set. The Test Sets were chosen to represent non-ambiguous EDV or clear non-EDV cases for proof of concept, as well as several cases of interest in Calluna (Table 2). This non-ambiguousness was derived from personal communications with the involved breeding company in case of the EDV-pair. The test of a BC1 against the parents as a clear non-EDV case was performed with our own crossings.Table 1Identifying essential derivation in C. vulgaris. Reference SetsSet ASet B1Niederohe'Sandy'2San Remo'Annegret'3'Adrie'4'Allegro'5'Boskoop'6'Carmen'7'C. W. Nix'8'Dark Beauty'9'Findling'10'Glenmorangie'11'Johnson's Variety'12'Long White'13'Mariella'14'Marlies'15'McDonalds of Glencoe'16'Minima Smith's Variety'17'Mrs. Pinxteren'18'Multicolor'19'Orange Queen'20'Peace'21'Radnor'22'Sandhammeren'23'Silver Knight'24'Underwoodii'25'Wickwar Flame'Identifying essential derivation in C. vulgaris: Reference-Sets A and B chosen for the Test-Sets. Set B differs from A only in the exchange of two wild-type genotypes (Niederohe, San Remo) versus 2 varieties ('Sandy', 'Annegret') as indicated by summarizing both columns of Set A and B for the residual 23 varieties.Table 2Identifying essential derivation in C. vulgaris. #Test-Setsselection criteriaReference SetHypothesisResult1'Maria'Maria HellMaria Hell = known sport of 'Maria' according to information from a breederAyesyes2'Maria'BC1-individualprogeny testingAnono3'Roter Oktober'BC1-individualprogeny testingAnono4'Melanie'Anette''Anette' = sport of 'Melanie' according to information given by BSA docAyesyes5'Melanie''Sandy''Sandy' = sport of 'Melanie' according to information given by BSA docAyesyes6'Annegret''Anneliese''Anneliese' = sport of 'Annegret' according to information given by BSA docAYesyes7'Fritz Kircher'CV7re-testing results from former investigationsAyesno8'Karla''Venetia'similar cultivars from different breedersAnono9'Minka''Miranda'similar cultivars from different breedersAnono10SanRemoRuhlawild-type testingBnono11NiederoheLöhnsteinwild-type testingBnono12NiederoheSanRemowild-type testingBnonoTest Sets and their criteria for selection, the applied Reference Set and our initial hypothesis with regard to whether or not ED was to be expected. The first three tests were used as proof of concept, meaning consistency of hypothesis and validation of the eligibility of the method; tests 4–12 are true testings.After extensive testing we selected a threshold provided by the highest Dice value of the 98% lowest values of all pairwise comparisons within the reference set (Fig. 3). This threshold was chosen in order to prevent the BC1 individual from being categorized as essentially derived from the backcross parent which constitutes an essential prerequisite for validation of our test since backcrossing is the normal breeding system in bud-flowering Calluna. The 98% thresholds in both Reference Sets differ due to the necessary adjustment of the reference set according to the test in question (exchange of wild-type genotypes against varieties from the upper cluster of the dendrogram): 98%-Set A: 0.865 Dice similarity value, 98%-Set B: 0.893 Dice similarity value.Figure 3Identifying essential derivation in C. vulgaris I. Validation of the method using three sets for proof-of-concept: One set of a known essentially derived variety pair and two sets of genotypes involved in backcrossing, marked by black symbols. Additionally six pairs of varieties of interest have been tested against the chosen threshold of 0.865 Dice similarity value, which was derived from Reference Set A.For proof of concept we tested, on the one hand, one pair of individuals ('Maria' and Maria Hell), from which it was known that the latter was derived from the first one. On the other hand, an individual from a backcross progeny was tested against both parents, which should result in the categorization of being non-derived. As expected, the first result was positive and the second one negative, using the threshold as given above (Fig. 3). Moreover, similarity between the BC1 individual and the backcross parent was clearly higher than between the BC1 individual and the second parent. The Dice value of the comparison with the backcross parent was actually slightly above the threshold; however, overlapping error bars indicated that the similarity was nevertheless not sufficiently high for these two genotypes to be categorized as essentially derived.Regarding the 'true tests', the results were negative for several pairs of morphologically similar cultivars from different breeders (Fig. 3), as well as for wild genotypes of different origin (Fig. 4). In contrast, when testing the cultivars 'Melanie' and 'Anette', their genetic similarity was found to exceed the threshold, thus confirming the public data supplied by the BSA according to which 'Anette' is a sport of 'Melanie'. This was also confirmed for 'Melanie' vs. 'Sandy' and 'Annegret' vs. 'Anneliese' (Fig. 3).Figure 4Identifying essential derivation in C. vulgaris II. Test of three pairs of wild types of different origin using Reference Set B (0.893 Dice similarity value).The last test concerned a pair of cultivars ('Fritz Kircher' vs. CV7) which have in a former, non-public study been characterized as being essentially derived from one another using dendrogram analysis. In our investigation, however, their genetic similarity is lower than the threshold, thus clearly indicating an absence of essential derivation.DiscussionUntil now, molecular data on genetic diversity within the species C. vulgaris was only available for regionally restricted wild-type populations [15-17], not for varieties used in commercial breeding. An actual and urgent necessity for a comprehensive study in C. vulgaris can be deduced from several points: the number of applications for variety protection is currently increasing considerably, whereas the information given in the registration schedules is at least occasionally unreliable or equivocal (e.g. a bud flowering variety is said to be the result of selfing of another bud-flowering genotype, which is biologically impossible due to the total loss of anthers in bud-flowering genotypes). This leads to an ambiguous situation with regard to variety derivation. Additionally, molecular data are needed for concerted breeding works and the elimination of coincidence in this process.Since it is technically simple to accomplish and requires no a priori sequence information, iSSR- and RAPD-PCR [18-20] are widely used techniques in different species; but RAPDs in particular may be 'considered the practice of PCR without a clue' [21]. All the same, both techniques provide a uniformly distributed amplification of DNA fragments throughout the genome of eukaryotic organisms due to the nature of their origin, and were shown to be an adequate molecular tool for studying DNA polymorphisms (e.g. [22,23]). The same was true for our investigations as we observed very robust inner-laboratory reproducibility: here, a value as low as 0.46% of missing data within the 77 × 168 similarity matrix was achieved.The dendrogram resulting from the combined computation of both RAPD and iSSR banding patterns showed a low genetic variability within the species C. vulgaris: almost all tested varieties and genotypes are grouped at a Dice/Nei & Li similarity value of 0.80, or even higher. This confirmed our hypothesis of a narrow gene pool, which was expected by the breeding experiences and methods applied of the participating company (personal communications) and its competitors. Moreover, one has to bear in mind that C. vulgaris is the only species within the genus Calluna and that crossing with other genera of the Ericaceae is thus impossible, thereby assisting in the conservation in nature, too, of a slender genetic diversity. We esteem the clear discrimination between Erica and Calluna as one argument of reassurance for our methodological approach and consider the dendrogram to be unbiased in the sense of an essential prerequisite for picturing genetic data [11]. The fact that wild type genotypes from the Lüneburger Heide are grouped this near to economically important varieties is another piece of evidence for our line of argument in respect of a significantly narrow gene pool in C. vulgaris. In addition, to our knowledge, breeding in C. vulgaris began in exactly this area of Germany by collecting incidentally originated bud-flowering genotypes. Our results might thus confirm this hypothesis, especially since the wild types from Thuringia and the Italian Alps do not cluster within this group.Another interesting feature of the resulting dendrogram is that the data were insufficient to support more than the few marked nodes (marked with * or ** in Fig. 2) as statistically significant. However, we do not consider the amount of bands i.e. mono-/polymorphisms from our data as generally too sparse, since [24] showed that an estimation of diversity within one population using approx. 200 dominant (i.e. AFLP) markers is as efficient as using 50 codominant (i.e. microsatellite) markers. Therefore, it is our suspicion, that the dendrogram method is not suitable for EDV identification in species with narrow gene pools.ED issues arise for varieties that successfully passed DUS testing. An EDV is (i) predominantly derived from an initial variety, (ii) clearly distinguishable from it and except for these differences (iii) conforms to the initial variety in the expression of essential characteristics [5]. We consider it to be of paramount importance to apply a well adjusted system for identification of these EDVs for each species, and in our case for C. vulgaris, since the range of similarities presented in Fig. 2 proved the hypothesis of some breeders that the economically important varieties (and the genus Calluna in general) are closely related and thereby may readily lead to ED disputes, as has already been the case in the past.As explained above, construction of a dendrogram proved to be no satisfactory tool for EDV identification in C. vulgaris – contrary to the results obtained for other vegetatively propagated species presented by [9] for Phalaenopsis, Rosa and Rhododendron. Another example is given by [25]. Using AFLPs, they proved that Rosa × hybrida original varieties are not more closely linked than 0.80 Jaccard's index. In contrast, the genetic similarities in so-called mutant groups were always higher than 0.96 (but not 1.0). Their dendrogram assay is therefore correctly rated as a suitable method to unambiguously distinguish rose EDVs from their initial variety. In addition, the detection of polymorphisms between sports and the original variety may be considered somewhat coincidental since molecular markers only cover a small portion of the target organism's genome. [26] demonstrated, that in cut roses RAPD-polymorphisms between a variety and its sports did occur in two varieties, but were not reproducible. Using AFLPs the authors were even able to amplify stable polymorphisms in sports of another variety. However, they were still able to distinguish vegetatively and sexually propagated progenies, since amplification in seedlings constantly resulted in a higher number of polymorphisms.We ascribe our differing results to the coincidence of two phenomena in C. vulgaris. First, stable genetic conditions – which could be reasonably anticipated for vegetatively propagated species – are worthy of discussion in the context of Calluna, since the phenomenon of sport/reversion (a type of somatic mutation) is well-known by breeders.Moreover, the very narrow gene pool in C. vulgaris gives rise to high genetic similarities, even if a new variety was obtained through crossing, due to the fact that even quite different individual plants, e.g. a wild type from the Lüneburger Heide and a bud flowering variety, show a considerable proportion of monomorphic bands in RAPD and iSSR analyses. Such lack of genetic diversity is our main reason for focusing on a system for EDV-identification involving a reference-set, as this is the important difference to e.g. the rose cases mentioned above: even in Rosa × hybrida more than 10,000 varieties exist, resulting from some 150 years of breeding efforts [25], and they are still clearly distinguishable. The opposite situation is, in fact, the result of the differing breeding methods applied in C. vulgaris: breeding for a common phenotype (bud-flowering) and repeated back-crossing are generally accepted reasons that promote the development of narrow gene pools [25].By working with a system similar to that described for lettuce and barley by [11], we were successful in both, identifying well-known essentially derived genotypes as well as discriminating between a genotype resulting from backcrossing and its parents (Fig. 3). We considered these results as a proof of concept for our method and additionally analyzed other test-sets whose information of origin we regarded to be unreliable, questionable or simply of interest. Here, information on variety derivation was primarily confirmed by our method as outlined in table 2. Moreover, the system discriminated phenotypically similar varieties from different breeders as well as wild genotypes of different origin, thus also confirming the hypotheses.ConclusionAs a result of these findings, we would like to suggest the outlined method as an appropriate system for EDV-testing in C. vulgaris. Applicability to other vegetatively propagated crops should be tested, as well as the combined use of 'fixed' and 'random/unmapped markers' as suggested by [11]. Moreover, we recommend the inclusion of at least three independent gDNA isolations of different individuals per genotype, since inner-varietal identity cannot be presumed and is hard to verify, even in vegetatively propagated crops.MethodsDNA techniques: isolation of genomic DNA (gDNA)gDNA of C. vulgaris genotypes was isolated according to [27]. About 200 mg young leaf tissue (stored over night and frozen in liquid nitrogen) was homogenized in 2 ml tubes in a mixer mill (MM301, Retsch) using 2 stainless steal balls (Ø = 5 mm). The tissue was resuspended in buffer A (50 mM Tris-HCl pH 8.0, 5 mM EDTA, 350 mM sorbitol, 1% β-mercaptoethanole, 10% PEG-6000) and centrifuged for 1 min at 4°C and 8,000 rpm (Sigma 3K30, rotor-no. 12148). The resulting pellet was again resuspended in buffer B (50 mM Tris-HCl pH 8.0, 5 mM EDTA, 350 mM sorbitol, 1% β-mercaptoethanole, 1% sodiumsarcosyle, 0.1% CTAB, 710 mM NaCl) and incubated for 30 min at 60°C. After adding 0.8 volumes chloroform-isoamyl alcohol 24:1, the samples were centrifuged for 15 min at 4°C at 15,300 rpm. The supernatant was transferred to a new 2 ml reaction tube and incubated at -20°C for 30 min after adding 0.75 volumes isopropanole. After centrifugation (5 min at 4°C at 5,000 rpm) the pellet was washed with 70% ethanol, air-dried and resuspended in 500 μl TE buffer. To each sample, 1 ng RNAse (Carl Roth GmbH) was added, followed by incubation for 15 min at 37°C. Subsequently, phenol-chloroform extraction step was performed twice and the resulting supernatant containing purified gDNA was pelleted at -20°C for 60 min after addition of 0.1 volumes 3 M sodium acetate and 0.75 volumes isopropanole. The precipitated gDNA was washed twice with 70% ethanol, air-dried and resuspended in 100 μl TE. Long-time storage was achieved at -20°C.DNA techniques: PCR amplification and electrophoresisAmplifications of RAPD-fragments generated from random decamer primers (Carl Roth GmbH) were performed in a Primus 96 advanced thermocycler (peqlab GmbH) using the following protocol: 5 min at 95°C, [1 min at 95°C, 1 min at 35°C, 1 min at 72°C]35×, 10 min at 72°C. The reaction mixture for a total volume of 25 μl contained 1× reaction buffer, 2.5 mM MgCl2, 1 U Taq.-DNA Polymerase (recombinant, Invitrogen), 0.2 mM of each dNTP (Invitrogen), 0.5 μM primer (Carl Roth GmbH, MWG Biotech AG), 10 ng gDNA and the adequate amount of sterile deionized H2O.Amplification of iSSR-Fragments was performed following the same protocol as described for RAPDs, with the altered annealing temperatures according to primer length. Table 3 provides an overview of primers used in this study; these were chosen after screening 60 decamer primers for reproducibility. Decamer primers were obtained as random primer kits from Carl Roth GmbH; iSSR primers were synthesized by MWG Biotech AG. iSSR primers given in Table 3 were chosen by referring to the common di- and trinucloetide motifs in plants (AC/TG)n and (AAG/TTC)n (e.g. [28]).Table 3List of iSSR- and RAPD-primers. typedenominationsequence (5' → 3')sourceiSSR17898B(CA)6-gTaccording to [21]17898C(CA)6-AC17899(CA)6-gg17901B(gT)6-TTP02(AAg)6-Cgaccording to [28]RAPDRX13ACgggAgCAArandom primer kitsRX14ACAggTgCTgRY01gTggCATCTCRY13gggTCTCggTRY15AgTCgCCCTTRY16gggCCAATgTRY17gACgTggTgARY18gTggAgTCAgRZ04AggCTgTgCTRZ05TCCCATgCTgRZ07CCAggAggACRZ12TCAACgggACRZ17CCTTCCCACTList of primers used for PCR-amplification of mono- and polymorphic fragments within gDNA of C. vulgaris, their sequence, and the source according to which the sequences were selected.Electrophoretic separation of the amplification products was performed in 23 × 25 cm 1.5% agarose gels by applying 7 V/cm for 2.5 hours. The gel contained ethidium bromide for visualization of fragments at 254 nm. Documentation was carried out with a digital imaging system (Biostep GmbH).Reactions were repeated at least twice before fragments were used for distance calculations.Statistics: gel analysis and phylogenetic calculationsGel analysis (band detection, noise reduction, size calibration, fragment matching) was performed with the Phoretix 1D Advanced software (Nonlinear Dynamics). The selection of bands derived from each primer was performed by objective criteria (e.g. thresholds for routine band detection and matching and recommendations to ensure reproducibility, e.g. the exclusion of fragments of very high and very low size). The banding data were transformed to a computable 0/1 matrix in the common [OTU × band] layout.Phylogenetic as well as dendrogram calculations were conducted with the NTSYSpc 2.20 L software (©1986–2006, Applied Biostatistics Inc.). Qualitative banding values were computed using the SimQual module with the similarity coefficient of Dice [29] and Nei and Li [30], respectively. Subsequent UPGMA clustering was conducted within the Sahn module, while the module Treeview was used to visualize the data set as a dendrogram.For Bootstrapping using the Dice coefficient, Winboot [31] was used (replications given in the text) which finally constructs a majority-rule consensus tree based on the Consense module of the PHYLIP software.EDV-testing by application of the tail principleA system first published by [11] for EDV identification in lettuce and barley was adapted for C. vulgaris as follows. Since a priori pedigree information is unavailable for C. vulgaris and the application of the pedigree principle – a threshold selection based upon inclusion of 'identity by descent' probabilities – was not possible, we selected the tail principle for the identification of a threshold from a distribution of pair-wise similarities from a reference-set. The configuration of the Reference Set (25 varieties) for each Test Set (2 varieties) was adapted by using the information gained from our phylogenetic results which matches the integration of the calibration principle. In the context of [11], we decided not to include known EDVs in this set, e.g. mutation-derived varieties or sports, since these would only represent extremely high values within the reference-set and would complicate data interpretation. Both polymorphic and monomorphic markers were analyzed. Detached primer-wise computation of similarity values becomes applicable by assuming a genetic independence between primers and a uniform distribution of primer binding sequences throughout the genome. Arithmetic means, standard deviations, standard errors as well as its medians were calculated according to [10] and [11] in analyzing the inner- and inter-set-similarities and also to define a threshold for identifying EDVs in Calluna. This threshold was positioned in such a way that varieties known to be ED exceed the threshold and non-EDVs do not.As carried out before for RAPD and iSSR values, these data were computed for their similarity values with the coefficient of Dice (Nei & Li) by NTSYSpc 2.20 L. Only in case of Dice similarity values above the threshold and non-overlapping error bars, a pair of genotypes is categorized as being essentially derived. In a case where similarity values are below the threshold or where they are above but with overlapping error bars, a pair of genotypes is categorized as not essentially derived.AbbreviationsAFLP: Amplified Fragment Length Polymorphism; ASSINSEL: 'International Association of Plant Breeders for the Protection of Plant Varieties'; EDV: Essentially Derived Variety/Varieties; BSA: Bundessortenamt; CPVO: Community Plant Variety Office; gDNA: genomic DNA; ISF: International Seed Federation; iSSR: inter Simple Sequence Repeats; RAPD: Randomly Amplified Polymorphic DNA; UPOV: International Union for the Protection of new Varieties of Plants; UPGMA: Unweighted Pair Group Method with Arithmetic MeanAuthors' contributionsAfter methodological setup, TB carried out the complete RAPD section from laboratory work to analysis and drafted the manuscript. JK performed the complete iSSR part from laboratory work to analysis. AH designed the study and participated in drafting the manuscript.All authors read and approved the final manuscript.Supplementary MaterialAdditional file 1Complete list of included varieties and genotypes, their country of origin and pedigree information where known (source is given in the last column). Column 3 defines the flower type either as normal, bud, multi-bracteate (multi) or filled. Sources of information are either the Bundessortenamt (BSA:doc), the appropriate website (web) of the 'The International Register of Heather Names'  or personal communications (personal contact).Click here for file\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2546440.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2546440",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2546440\nAUTHORS: Klaus Dittmann, Claus Mayer, Rainer Kehlbach, H Peter Rodemann\n\nABSTRACT:\nBackgroundTo elucidate the role of src kinase in caveolin-1 driven internalization and nuclear transport of EGFR linked to regulation of DNA-repair in irradiated cells.ResultsIonizing radiation resulted in src kinase stabilization, activation and subsequent src mediated caveolin-1 Y14- and EGFR Y845-phosphorylations. Both phosphorylations were radiation specific and could not be observed after treatment with EGF. Inhibition of EGFR by the antibody Erbitux resulted in a strong accumulation of caveolin/EGFR complexes within the cytoplasm, which could not be further increased by irradiation. Radiation-induced caveolin-1- and EGFR-phosphorylations were associated with nuclear EGFR transport and activation of DNA-PK, as detected by phosphorylation at T2609. Blockage of src activity by the specific inhibitor PP2, decreased nuclear transport of EGFR and inhibited caveolin-1- and DNA-PK-phosphorylation. Knockdown of src by specific siRNA blocked EGFR phosphorylation at Y845, phosphorylation of caveolin-1 at Y14 and abolished EGFR transport into the nucleus and phosphorylation of DNA-PK. Consequently, both knockdown of src by specific siRNA and also inhibition of src activity by PP2 resulted in an enhanced residual DNA-damage as quantified 24 h after irradiation and increased radiosensitivity.ConclusionSrc kinase activation following irradiation triggered caveolin-1 dependent EGFR internalization into caveolae. Subsequently EGFR shuttled into the nucleus. As a consequence, inhibition of internalization and nuclear transport of EGFR blocked radiation-induced phosphorylation of DNA-PK and hampered repair of radiation-induced double strand breaks.\n\nBODY:\nBackgroundMany human tumor cells are characterized by over-expression of epidermal growth factor receptor (EGFR), a protein that promotes growth and aggressiveness and resistance of cancer cells to chemo- and radiotherapy [1-5]. EGFR can be phosphorylated in response to binding of its specific ligands (EGF, TGF alpha and Amphiregulin) [6,7] and after exposure to unspecific stimuli like ionizing radiation [8], UV-radiation [9], hypoxia [10], hyperthermia [11], oxidative stress [12] and trans-activation by G-protein coupled receptors [13,14]. Ligand-dependent as well as ligand-independent phosphorylation of EGFR results in receptor internalization [15] and intracellular signaling [4,5,16-18]. Up to date internalization is assumed to be essential for receptor silencing and inactivation. Indeed, EGF treatment results in internalization of EGFR into coated pits followed by receptor degradation [19]. As reported by Khan [12], exposure to oxidative stress can lead to internalization of EGFR by caveolae and this process is associated with peri-nuclear accumulation of EGFR.A characteristic constituent of caveolae is caveolin. In vertebrates the caveolin gene family has three members: CAV1, CAV2, and CAV3, coding for the proteins caveolin-1, caveolin-2 and caveolin-3, respectively. Caveolins form oligomers and associate with cholesterol and sphingolipids in certain areas of the cell membrane, leading to the formation of caveolae. Caveolae are involved in receptor independent endocytosis [20]. Furthermore Caveolin-1 is an integral transmembrane protein and an essential component in interactions of integrin receptors with cytoskeleton-associated and signaling molecules [21]. Compartmentation into caveolae prevents EGFR degradation and simultaneously enables intracellular EGFR signaling [12]. These findings suggest a new function of EGFR – depending on its intracellular localization -, which supplements its functions described so far. The idea of additional EGFR functions is further supported by the observation, that peri-nuclear EGFR can be transported into cell nucleus in response to irradiation [5]. As we and others have reported earlier [4,22-24], nuclear EGFR is linked with activation of DNA-PK and regulation of non-homologous end-joining DNA-repair resulting in increased radioresistance [5]. As reported recently [1], nuclear EGFR detection in tumors biopsies correlated strongly with treatment resistance and bad prognosis.In the present study, we focused on the radiation-induced nuclear translocation process of EGFR via caveolae. Evidence is provided that inhibition of src activity blocks the caveolin-dependent EGFR internalization and nuclear EGFR transport, which results in impaired DNA-repair.Materials and methodsCell culture, transfection, irradiation and colony formation assayExperiments were performed with the human bronchial carcinoma cell line, A549 (ATCC) and the human squamous carcinoma cell line FaDu (ATCC, origin head and neck cancer). Cells were irradiated with 200-kV photons (Gulmay RS 225, dose rate 1 Gy/min) at 37°C. The EGFR-inhibitory antibody Erbitux was purchased from Merck KG aA, Germany and was administered to the cells at a concentration of 30 nM 1 h before irradiation. PP2 (4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo [3,4-d]pyrimidine) was received from Sigma and cells were treated at a concentration of 100 nM PP2 dissolved in DMSO for 1 h. For silencing of src cells were treated with specific siRNA for 72 hours before irradiation. Control non-silencing siRNA (sense UUCUCCGAACGUGUCACGUtt; antisense ACGUGACACGU-UCGGAGAAtt) and siRNA targeting src (sense ACUCGCCUUCUUAGAGUUUtt; antisense AAACUCUAAGAAGGCGAGUtt) probes were purchased from MWG-Biotech AG. Both were transfected at a concentration of 30 nM using Lipofectamine 2000 transfection reagent according to manufacture's protocol (Invitrogen). For colony formation assay cells were grown to confluence, treated as indicated and irradiated. After 6 hours cells were typsinized and seeded at a density of 500 cells in 78 cm2 plates. After 10 days colonies were fixed, stained and counted. Radiation survival curves were plotted after normalizing for the cytotoxicity induced by siRNA treatment or vehicle alone. Clonogenic survival curves were constructed from at least three independent experiments.Subcellular fractionationCytoplasmic and nuclear extracts were prepared according to the instructions of the NE-PER® nuclear and cytoplasmic extraction kit (Pierce, Rockford, IL, USA).Western blot analysis and immune-precipitationAfter irradiation, as described above, cells were lysed and proteins were resolved by SDS-PAGE. Western blotting was performed according to standard procedures [25]. The primary antibodies were diluted as follows: anti-EGFR (BD Transduction Laboratories, clone 13) 1:1000; anti EGFR pY845 (nanotools, clone 12A3) 1:1000; anti-EGFR pY992 (abcam, polyclonal) 1:500); anti-EGFR pY1173 (Cell signaling, clone 53A5) 1:1000; anti-phosphotyrosin (Santa Cruz, clone PY20) 1:500); anti-src (Santa Cruz, clone H-12) 1:1000; anti-src Y416 (cell signaling, polyclonal), 1:1000; anti-caveolin-1 (BD Transduction Laboratories, clone 2297) 1:1000; anti-caveolin pY14 (BD Transduction Laboratories, clone 56) 1:1000; anti-DNA-PK (PharMingen, clone 4F10C5) 1:500; anti-DNA-PK pT2609 (Rockland) 1:1000; anti-lamin B1 (Biozol, clone ZL-5) 1:1000. Quantification of binding was achieved by incubation with a secondary peroxidase-conjugated antibody with the ECL system (Amersham).EGFR was immune-precipitated from cytoplasmic and nuclear protein fractions prepared from 20 × 106 cells with EGFR antibody clone 13 (BD Transduction Laboratories). Immune-precipitation was performed as described [26].Quantification of γH2AX-foci formationCells cultured on CultureSlides (Becton Dickinson) were incubated with PP2 or src-siRNA, irradiated and fixed with 70% ice-cold ethanol 24 h after irradiation. For immune-fluorescence analysis cells were incubated with γH2AX antibody (Upstate, clone JBW301)(1:500) at room temperature for 2 h. Positive foci were visualized by incubation with a 1:500 dilution of Alexa488-labelled goat anti-mouse serum (Molecular Probes) for 30 min. Coverslips were mounted in Vectashield/DAPI (Vector Laboratories). For each data point 300 to 500 nuclei were evaluated.ResultsCaveolin-1 associated EGFR internalization following irradiation was triggered by src kinaseIonizing radiation induced protein stabilization of cytoplasmic src in A549 cells within 10 to 20 min (Fig. 1A). Stabilization of src was associated with phosphorylation at residue Y416, which indicates src activation [27]. Radiation-induced stabilization and activation of src led to complex formation with caveolin-1 and EGFR as shown by means of a src-specific immune-precipitation. Caveolin-1 associated EGFR was phosphorylated at residue Y845, which is a known src kinase specific phosphorylation site [28]. Moreover, increased src activity was associated with increased caveolin-1 and EGFR binding and phosphorylation at Y14 and respectively Y845 (Fig. 1A). EGF treatment also stabilized src protein and triggered a weaker Y416 phosphorylation. However, caveolin-1 and EGFR were not increased over the control in the src associated complex following EGF treatment (Fig. 1B). To elucidate the role of src-driven EGFR phosphorylation at residue Y845 following irradiation, we determined the phosphorylation status of additional tyrosine residues of EGFR after irradiation with 4 Gy (Fig. 1C). In general, tyrosine phoshorylation of EGFR was markedly increased. Especially phosphorylation at residues Y845 and Y1173 were strongly induced by irradiation, whereas the induction of phosphorylation at residue Y992 was weaker.Figure 1Radiation induced caveolin-1 linked EGFR internalization is triggered by src kinase. Confluent A549 cells were irradiated with 4 Gy (A). At the time points given, src protein was immune-precipitated from cytoplasmic protein fraction. Proteins were separated by SDS-PAGE and after blotting the protein amounts of src, caveolin1 and EGFR were quantified. The same procedure was applied after cell stimulation with EGF (B). Phosphorylation of EGFR at the tyrosine No. 845, 992 and 1173 was quantified with help of specific antibodies following a standard western procedure. Equal protein amounts for immune-precipitation were documented by showing actin expression within aliquots of input proteins. (C). Protein expression was quantified by densitometry and shown as fold-induction relative to untreated control. The mean protein expression derived from three experiments was given below each band.The antibody Erbitux stabilized the cytoplasmic caveolin-1/EGFR complexIn agreement with Fig. 1, immune-precipitation targeting caveolin-1 indicated also an increasing complex formation with EGFR following radiation exposure in A549 cells (Fig. 2). Pre-incubation (1 h) with the EGFR inhibitory antibody Erbitux, resulted in a pronounced stabilization of the caveolin-1/EGFR complex in non-irradiated cells. This complex formation could not be further increased by radiation exposure.Figure 2Stabilization of caveolin1/EGFR complex in cytoplasm after incubation with EGFR inhibitory antibody Erbitux. Confluent A549 cells were treated with human IgG or EGFR inhibitory antibody Erbitux at a concentration of 30 nM for 1 h. Subsequently cells were irradiated with 4 Gy and cytoplasmic and nuclear proteins were isolated at time points given. Proteins were separated by SDS-PAGE and after blotting EGFR protein was quantified with help of a specific antibody. Experiments were performed three times; shown are representative results. Equal protein input for immune-precipitation was documented by showing actin expression within aliquots of input proteins. Expression of specific proteins was quantified by densitometry and shown as fold-induction relative to untreated control.Src kinase inhibitor PP2 prevented radiation-induced EGFR transport into the nucleus and hampered radiation-induced activation of DNA-PKAs already shown earlier [5], ionizing radiation triggered EGFR transport into the nucleus and complex formation with DNA-PK (Fig. 3) [5]. Treatment with the src kinase inhibitor PP2 (100 nM for 1 h) was sufficient to block radiation-induced nuclear translocation of EGFR in A549 cells (Fig. 3). Interestingly, basal amount of nuclear EGFR protein was not affected by PP2 treatment. However, total nuclear caveolin-1 protein and its phosphorylated form was decreased by PP2 treatment. In addition, phosphorylation of DNA-PK at residue T2609, which is essential for non-homologous end-joining DNA-repair process [29], was inhibited in response to PP2 incubation, whereas the amount of total DNA-PK protein was unchanged.Figure 3Inhibition of nuclear EGFR transport and phosphorylation of DNA-PK by src inhibitor PP2. Confluent A549 cells were treated with src inhibitor PP2 at a concentration of 100 nM for 1 h. Subsequently cells were irradiated with 4 Gy and nuclear proteins were isolated at the time points given. Proteins were separated by SDS-PAGE and after blotting protein amounts were quantified by help of specific antibodies. Expression of specific proteins was quantified by densitometry, normalized to lamin B1 and shown as fold-induction relative to untreated control. Experiments were performed three times; shown are representative results.Src siRNA decreased phosphorylation of cytoplasmic EGFR at Y845, reduced EGFR transport into nucleus and impaired phosphorylation of DNA-PK at T2609 after irradiationSrc protein expression was effectively knocked down in A549 and FaDu cells 72 h after transfection with src specific siRNA (Fig. 4A/C). Repression of src-protein markedly inhibited radiation-induced phosphorylation of cytoplasmic EGFR at Y845 and caveolin Y14 (Fig. 4A/C) after irradiation. In agreement with the data of Fig. 3, src knockdown also reduced radiation-induced EGFR shuttling into nucleus (Fig. 4B/D). Moreover, like PP2 inhibitor (see Fig. 3) src siRNA treatment led to reduced phosphorylation of DNA-PK at residue T2609 (Fig. 4B). It is noteworthy, that basal amount of nuclear EGFR protein was increased already by src siRNA treatment alone. However, this increase did not correspond to an activation/phosphorylation of DNA-PK at T2609.Figure 4Inhibition of EGFR transport into the cytoplasm/nucleus by src siRNA. Exponentially growing A549 cells were incubated either with control- or src-specific siRNA for 72 h at a concentration of 30 nM. Subsequently confluent cells were irradiated with 4 Gy and cytoplasmic (A) and nuclear proteins (B) were isolated at the time points given. Corresponding results were presented for FaDu cells (C and D). Expression of specific proteins was quantified by densitometry, normalized either to actin or lamin B1 and shown as fold-induction relative to untreated control. Experiments were performed three times; shown are representative results.Blocking of src signaling increased level of residual DNA-damage following irradiationIrradiation of confluent A549 cells treated with control siRNA resulted in a radiation- dose dependent increase in residual γH2AX positive repair foci, which represent un-repaired DNA double strand breaks (Fig. 5A). Pre-treatment with src specific siRNA increased the number of radiation-induced residual γH2AX foci 24 h after irradiation, by a factor of 1.5 – 2 (Fig. 5A). Likewise pre-treatment of cells with the src kinase inhibitor PP2 for 1 h, resulted also in a significant increase in residual γH2AX foci (Fig. 5B). In both cases inhibition of DNA-damage repair was correlated with increased radiosensitivity, as determined by means of colony formation assay (Fig. 5C/D).Figure 5Inhibition of DNA-repair and clonogenic survival by src siRNA or src inhibitor PP2. A549 cells were either treated with src siRNA (72 h) (A/C) or with src inhibitor PP2 (1 h) (B/D). Subsequently cells were irradiated with 2, 4 and 6 Gy and after 24 h cells were fixed. Residual damage was visualized by incubation with γH2AX antibody (A/B). Each bar represents the mean ± SE of residual repair foci positive for γH2AX per cell nucleus. For each data point 300 nuclei were evaluated. Asterisks indicate significant differences (Student's t-test * p < 0.05). For colony formation assay cells pretreated either with src siRNA (C) or PP2 (D) were irradiated and seeded at a density of 500 cells per 78 cm2 dish. After 10 days colonies were fixed and stained. Surviving fractions were calculated on the basis of colony counts and plating efficiency. Each value represents the mean of three independent experiments. Differences were considered as statistically significant for p < 0.05 (t-test) and were marked with an asterisk *.DiscussionIt is generally accepted, that the epidermal growth factor receptor is localized within the cell membrane and will be internalized following activation and dimerization [30]. Indeed, such a scenario can be observed following EGF stimulation [30], which initiates proliferation associated signaling. However, the EGFR can be also activated by oxidative stress [12], radiation, [5,8] and G-coupled receptors [13]. The molecular mechanisms of this ligand independent activation of EGFR are not fully understood. However, ligand independent stimulation of EGFR, e.g. by ionizing radiation [8], is clearly characterized by receptor internalization also. The data presented herein, give new insights into the mechanism of EGFR internalization process and the intra-nuclear function of EGFR following exposure to ionizing radiation.Several pathways enable endocytic transport of cargo molecules from the surface of eukaryotic cells into cytoplasm [31]. The two best understood pathways, relevant for EGFR internalization, are the clathrin-coated pit [31] and the caveolin [32] driven internalization mechanisms. As shown by Khan et al. [12], the clathrin-coated pit associated EGFR internalization can be observed following treatment with EGF and results in a fast degradation and silencing of receptor function. In contrast, treatment with H2O2 leads to EGFR internalization into caveolae, which sort internalized EGFR into a per-nuclear localization associated with an ongoing receptor signaling [12]. In agreement with these data, we could show, that exposure to ionizing radiation induced a caveolin-1 associated EGFR internalization, whereas EGF treatment failed to trigger complex formation between src, EGFR and caveolin-1. Sorting into different compartments in response to different stimuli may explain signal discrimination at the level of activated EGFR. Like for H2O2 treatment [12], exposure to ionizing radiation also mediates the src driven phosphorylations of EGFR at Y845 and of caveolin-1 at Y14, which is needed for internalization of EGFR into caveolae [12]. In response to radiation not only EGFR phosphorylation at Y845 – which is Src dependent – was observed, but also phosphorylation at Y992 and Y1173 could be observed. Both are described as autophosphorylation sites [33]. This implicates that ionizing radiation activates not only src kinase, but also EGFR kinase and both kinases contribute to altered phosphorylation pattern of EGFR following radiation exposure. Caveolin-1 phosphorylation seems to be critical for caveolae formation [20]. On the contrary, Y845 phosphorylation of EGFR probably is rather essential in regulation of EGFR-kinase activity than in formation of coated pits or caveolae [15]. However, as shown by us and by Khan [12] src activity is crucial for radiation- and H2O2-induced formation of caveolae. Nevertheless, the molecular mechanism responsible for activation of src has to be resolved. From our data it appears, that radiation leads to a fast activation of src, which is documented by phosphorylation of src at residue Y416. This phosphorylation is described as an autophosphorylation [27]. As activating molecular switch several mechanisms are discussed: (i) oxidation associated structural modifications result in activation of src kinase [34], (ii) inhibition of a phosphatase leads to auto-activation of kinase [35], (iii) G-coupled receptor signaling mediates src activation [14]. Which of these potential mechanisms is relevant for radiation-induced src kinase activity is currently unclear and is subject of ongoing investigations.As shown herein, treatment with Erbitux, which binds to the extracellular domain of EGFR, results in receptor internalization and formation of an intracellular complex of EGFR, caveolin-1 and Erbitux. Internalized EGFR however can not be activated by EGF and this observation may explain growth inhibitory effects of Erbitux.Khan et al. observed a peri-nuclear EGFR accumulation due to caveolin-1 driven internalization after exposing cells to H2O2 [12]. We could also detect a peri-nuclear localization of the EGFR [5] in irradiated cells, which is accompanied by a nuclear EGFR shuttling [5]. Based on these results we hypothesized, that peri-nuclear EGFR serves as a pool for nuclear EGFR transport following irradiation. This hypothesis is supported by the observation that inhibition of src either by its specific inhibitor PP2 or by specific siRNA, prevents nuclear translocation of EGFR by blocking caveolin-1 driven EGFR internalization. It is noteworthy, that caveolin-1 driven EGFR internalization occurs predominantly following treatment of cells with genotoxic agents. This observation is in favor with the idea, that EGFR internalization and nuclear transport of EGFR are linked with DNA-repair processes [23,36]. This assumption is supported by the observation, that caveolin-1 driven EGFR internalization is not observed after EGF treatment. As shown for irradiated cells nuclear EGFR is found in complex with DNA-PK, which is an essential compound of non-homologous end-joining DNA-repair [5]. As reported earlier [37], inhibition of EGFR nuclear transport by Erbitux, markedly impaired radiation associated activation of DNA-PK and increased cellular radiosensitivity [37]. In agreement with that, inhibition of src, which blocks EGFR internalization and subsequently nuclear transport after irradiation, abolished activation of DNA-PK, inhibited DNA-repair and increased radiosensitvity. Based on the data presented, it can be concluded, that the radiation-induced activation and nuclear translocation of EGFR is mediated through src kinase activity in a caveolin-1 dependent process. As blocking of these processes markedly effects repair of DNA-double strand breaks, this EGFR-coupled radiation response mechanism offers new interventional molecular targets for cancer therapy, especially by radiation therapy.ConclusionEGFR internalization by caveolin-1 is a stress specific cellular reaction, which is src kinase activity dependent. Linked with EGFR internalization nuclear transport can be observed following irradiation. Nuclear EGFR transport can be hampered by inhibition of src. Consequently, src inhibition is associated with inhibition of EGFR triggered activation of DNA-PK, which leads to an inhibition of DNA-repair and cell survival.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsKD, CM and RK performed experiments and interpreted data; the authors contribution to this research are reflected in the order shown. HPR supervised all aspects of this research. KD and HPR prepared the manuscript. All authors read and approved the final manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547148.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547148",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547148\nAUTHORS: Peter B. Stockwell\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547308.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547308",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547308\nmeasurement of the alcoholic strength\nof potable spirits\n\nAUTHORS: R. G. Lidzey, B. M. Stockton, M. E. B. Brown\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547558.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547558",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547558\nAUTHORS: C. Ferré, M. J. Castiñeiras, M. J. Alsina, R. Galimany, J. Riera\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547572.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547572",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547572\nAUTHORS: No authors listed\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547708.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547708",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547708\nAUTHORS: D. Migneault, R. K. Forcé\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547709.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547709",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547709\naccess chemistry analyser\n\nAUTHORS: Susan B. Schotters, James H. McBride, Denis O. Rodgerson, Margaret H. McGinley, Marilyn Pisa\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2547828.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2547828",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2547828\nAUTHORS: R. Puchades, A. Maquieira, J. Atienza, M. A. Herrero\n\nABSTRACT:\nFlow injection analysis (FIA) has emerged as an increasingly used laboratory tool in chemical analysis. Employment of the technique for on-line sample treatment and on-line measurement in chemical process control is a growing trend. This article reviews the recent applications of FlA. Most papers refer to on-line sample treatment. Although FIA is very well suited to continuous on-line process monitoring, few examples have been found in this areamost of them have been applied to water treatment or fermentation processes.\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2548001.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2548001",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2548001\nAUTHORS: Kevin J. Halloran\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2548084.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2548084",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2548084\nvoltammograms by target transformation\nfactor analysis\n\nAUTHORS: Ling Gao, Shouxin Ren\n\nABSTRACT:\nNo Abstract\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2548397.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2548397",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2548397\nAUTHORS: Radosław Bandomir, Mariusz Krawczyk, Jacek Namieśnik\n\nABSTRACT:\nWe present the results of a first stage of development work on a new type of analyzer for hydrogen and C1–C3 hydrocarbons concentration measurements in the lower explosive limit range, based on single pellistor sensor with artificial neural network data postprocessing.\n\nBODY:\nNo Body Content\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2553246.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2553246",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2553246\nAUTHORS: Titus Schleyer, Heiko Spallek, Brian S Butler, Sushmita Subramanian, Daniel Weiss, M Louisa Poythress, Phijarana Rattanathikun, Gregory Mueller\n\nABSTRACT:\nBackground As biomedical research projects become increasingly interdisciplinary and complex, collaboration with appropriate individuals, teams, and institutions becomes ever more crucial to project success. While social networks are extremely important in determining how scientific collaborations are formed, social networking technologies have not yet been studied as a tool to help form scientific collaborations. Many currently emerging expertise locating systems include social networking technologies, but it is unclear whether they make the process of finding collaborators more efficient and effective.Objective This study was conducted to answer the following questions: (1) Which requirements should systems for finding collaborators in biomedical science fulfill? and (2) Which information technology services can address these requirements?Methods The background research phase encompassed a thorough review of the literature, affinity diagramming, contextual inquiry, and semistructured interviews. This phase yielded five themes suggestive of requirements for systems to support the formation of collaborations. In the next phase, the generative phase, we brainstormed and selected design ideas for formal concept validation with end users. Then, three related, well-validated ideas were selected for implementation and evaluation in a prototype.Results Five main themes of systems requirements emerged: (1) beyond expertise, successful collaborations require compatibility with respect to personality, work style, productivity, and many other factors (compatibility); (2) finding appropriate collaborators requires the ability to effectively search in domains other than your own using information that is comprehensive and descriptive (communication); (3) social networks are important for finding potential collaborators, assessing their suitability and compatibility, and establishing contact with them (intermediation); (4) information profiles must be complete, correct, up-to-date, and comprehensive and allow fine-grained control over access to information by different audiences (information quality and access); (5) keeping online profiles up-to-date should require little or no effort and be integrated into the scientist’s existing workflow (motivation). Based on the requirements, 16 design ideas underwent formal validation with end users. Of those, three were chosen to be implemented and evaluated in a system prototype, “Digital|Vita”: maintaining, formatting, and semi-automated updating of biographical information; searching for experts; and building and maintaining the social network and managing document flow.Conclusions In addition to quantitative and factual information about potential collaborators, social connectedness, personal and professional compatibility, and power differentials also influence whether collaborations are formed. Current systems only partially model these requirements. Services in Digital|Vita combine an existing workflow, maintaining and formatting biographical information, with collaboration-searching functions in a novel way. Several barriers to the adoption of systems such as Digital|Vita exist, such as potential adoption asymmetries between junior and senior researchers and the tension between public and private information. Developers and researchers may consider one or more of the services described in this paper for implementation in their own expertise locating systems.\n\nBODY:\nIntroductionSocial networking technologies have become one of the latest “killer applications” on the Internet, with some sites such as MySpace and Facebook amassing large numbers of users in a very short period of time [1]. While those sites initially focused on younger demographics such as teenagers and college students, they are now encompassing rapidly growing segments of adult and/or professional users. Professionals are beginning to employ such systems for, among other things, extending their professional networks (ie, by learning about colleagues of colleagues), locating experts to solve specific problems, and finding collaborators.Social networking approaches have the potential to help scientists find appropriate collaborators more quickly and efficiently than is currently the case. Over the past several decades, science has become significantly more collaborative, both generally [2,3], as well as in biomedicine [4]. The increasing frequency with which the terms interdisciplinarity and multidisciplinarity appear in the literature [5] illustrates this strong trend toward collaboration. As a result, collaboration with the right individuals, teams, and institutions is becoming ever more crucial to project success. New programmatic initiatives such as the Roadmap [6,7] and the Clinical and Translational Science Award (CTSA) [8] programs of the National Institutes of Health (NIH) and the Janelia Farm Research Campus of Howard Hughes Medical Institute [9] in the United States demonstrate that funding agencies and research organizations are not just passively observing this trend but are actively encouraging it.Currently, most researchers use one of two primary methods to find new collaborators [10]. One approach is to turn to colleagues in their existing social network [10-12]. Colleagues, especially senior ones or those “in the know,” are often able to quickly identify promising candidates for collaboration, to provide input on their potential compatibility and credibility, and to make an introduction. The second method is to search for potential collaborators through published works [10], done most commonly in online databases such as PubMed and Google Scholar. Information from these databases helps the collaboration seeker gauge the potential collaborator’s competence, credibility, and interest, but it provides no support for gaining access. Soliciting collaboration may begin with “cold calling” if no connection through a third person is possible.A third method for finding collaborators is to use databases of researchers partially or exclusively designed for the purpose. Knowledge management systems of this type, which include “expertise locating systems” [13], “knowledge communities” [14,15], and “communities of practice” [16,17], all provide, to varying degrees, support for finding experts and, by extension, potential collaborators. In the literature, the functions and definitions of these types of systems are not cleanly separated. It appears that expertise locating systems (also called expertise locator systems) as their core function most directly focus on the ability to find individuals knowledgeable in a particular problem/domain.The computer-supported cooperative work (CSCW) literature contains numerous examples of systems designed to connect people with each other to solve specific problems [13,18-23]. The Expertise Recommender [13] is a recommendation system to help company workers locate persons best qualified to assist with a specific problem. The Zephyr Help Instance [18] and ReachOut [19] are examples of simple lightweight collaborative systems to tap the expertise within a company. Email is sometimes used to exploit weak and latent ties within a professional community [24]. Most of these systems serve to help a person solve a specific problem at a particular point in time. Consequently, one of their most important functions is to help identify the person who is best equipped to assist with solving the problem in a specified time frame.In this study, we are focusing on the much bigger challenge of establishing the long-term collaborations typical in biomedical science. In this case, not only are researchers looking for the most qualified expert, but they also will most likely enter into a long-term relationship. Evaluating an individual’s promise for such a long-term relationship requires information, engagement, and effort much beyond what is needed for finding an expert for singular (or even episodic) problem solving. A thorough literature search located only one report of a system [21] specifically designed to help scientists meet this challenge.In contrast to the dearth of reports in the literature, electronic systems purporting to make it easier to help scientists find collaborators abound. Similar to social networking sites such as Facebook, such systems endeavor to help individuals make connections to others that are not likely to be made in an off-line context [25,26]. Among the more established systems is the Community of Science (COS), which provides a “database of detailed, first-person profiles of more than 480,000 R&D professionals and scholars” [27]. Another system, the Faculty Research Interest Project (FRIP) [21], is in use at the University of Pittsburgh and currently indexes 1926 research faculty of the six schools comprising the Health Sciences Center. Another system, ExpertFinder [20], has been designed to help employees of The MITRE Corporation locate experts within the company. LinkedIn, Innocentive, Index Copernicus Scientists, Research Crossroads, and BiomedExperts are some of the more recent commercial offerings that advertise large directories of professionals/scientists. A thorough search for literature evaluating how well these systems facilitate the initiation of collaborations yielded no results. While these systems provide significant value to individuals looking for someone with specific expertise, anecdotal evidence suggests that they currently do not play a significant role in helping researchers establish collaborations.However, there are good reasons to suspect that expertise locating systems could help scientists find the most appropriate collaborator(s) more quickly and efficiently than is currently the case [26]. General trends in scientific research are compelling scientists to become more collaborative than they already are. As academic/research institutions extend the scale and scope of their research portfolio and, in the process, the numbers of their research faculty, more individuals are available for collaboration, either locally or remotely. At the same time, online databases, such as Google and PubMed, make locating collaborators easier. The number of potential collaborators is also increased by modern communication and collaborative technologies—many remote collaborations that would have once been considered impractical have now become feasible. The result is an “embarrassment of riches” for scientists seeking collaborators. Unfortunately, with this ever-expanding pool of potential collaborators, the task of selecting optimal collaborators is becoming more onerous and requires more effort from researchers, simply because there may be many more good options to choose from than previously possible. Studies have shown that when faced with this type of social overload, individuals are more likely to adopt competitive or withdrawal strategies and thus tend to be less cooperative [28]. Systems that help scientists “quality filter” the realm of possibilities for the most promising potential collaborators could help alleviate this social overload [29] and achieve more appropriate collaboration decisions at lower cost to the collaboration seeker.The confluence of the trends of increasing scientific collaboration, the emergence of social networking as a powerful mediator of social interaction, and the growing availability of information about scientists and their work presents a significant opportunity to investigate whether expertise locating systems can make the process of finding collaborators more effective and efficient. Current systems are relatively new and have an uncertain track record. One immediate question that occurs is whether those systems are responsive to the requirements of scientists seeking collaborators. In answer to this question, the main goal of this study was to develop preliminary, generalizable requirements for expertise locating systems for biomedical scientists. Its second goal was to design a set of services responsive to these requirements, implement them in a prototype system, and formatively evaluate them with representative end users.The main focus of this paper is to describe services and functions useful for expertise locating systems in general, not their implementation in a specific system. This study has been conducted as part of the University of Pittsburgh’s Clinical and Translational Science Institute in response to the core challenge to accelerate scientific discovery and the application of its results. As the other 23 current CTSA awardees in the United States are pursuing the same goal, our results are highly significant in that context. In addition, we hope that scientists and developers of expertise locating systems consider our results in the context of their own projects, potentially adopt/implement them, and conceptualize and design additional services as necessary.MethodsThis project proceeded in two phases: the background research phase and the generative phase. While the background research phase of the project emphasized discovering as much as possible about the relevant problem domain, the generative phase was intended to develop as many viable solutions as possible and then to choose one or more approaches to implement in a prototype. The project team included two faculty from the Center for Dental Informatics (TS and HS), one faculty from the Katz Graduate School of Business (BB) at the University of Pittsburgh, and two faculty (Susan Fussell and Brad Myers) and five senior masters students (SS, DW, LP, PR, and GM) from the Human-Computer Interaction Institute at Carnegie Mellon University. The project took place from January to August 2007. We describe the two main project phases briefly below.Background Research PhaseWe began the background research phase with a systematic literature review on relevant topics from the computer-mediated communication, social network theory, and computer-supported cooperative work literature. Keywords included “expertise locating systems,” “expertise management systems,” “knowledge communities,” “knowledge management,” “knowledge management systems,” “communities of practice,” and “virtual communities.” We searched Medline, the ISI Web of Science, the ACM Portal, and the IEEE Digital Library (all available years). From this material, we generated an affinity diagram [30] of issues and questions involved in the initiation of collaboration. We then performed contextual inquiries (CI) [31] with 10 researchers at Carnegie Mellon University and the University of Pittsburgh from a range of disciplines and levels of seniority. Since we could not directly observe researchers forming collaborations, the contextual inquiry was based on retrospective accounts. We also used a technique called directed storytelling in which we presented hypothetical situations to the interviewees and had them walk us through what they would do in each given situation. For each CI session, we generated workflow, sequence, and cultural models [31].\n                    Table 1Researcher affiliation, gender, seniority, collaborator count, and perceived collaborative loadSchoolGenderSeniorityNumber of CollaboratorsPerceived Collaborative LoadMedicineMJunior3-4too fewMedicineFJunior4too fewMedicineMSenior4too fewMedicineFJunior7too fewDentalMJunior7too fewRehabilitationMSenior9too fewPublic HealthFJunior10too fewPharmacyFJunior25too fewPharmacyMJunior6just rightMedicineMSenior6-8just rightMedicineFJunior8just rightNursingFJunior8just rightRehabilitationMSenior8just rightPharmacyMJunior9just rightPharmacyMSenior9just rightMedicineMJunior10just rightMedicineMSenior10just rightDentalMSenior15just rightNursingFJunior20just rightMedicineFJunior20just rightPublic HealthMSenior30just rightRehabilitationMJunior30-50just rightPublic HealthMSenior16-20too manyMedicineMSenior24too manyPublic HealthFSenior40too manyPublic HealthMSenior7n/aDentalMSenior15n/a\n                In a parallel study, we conducted semistructured interviews with 27 scientists at the University of Pittsburgh (see Table 1). The interviews contained 10 main questions and focused on current and previous collaborations, finding collaborators, solving problems in research, and information needs and information resource use of participants. The interview study was conceived as a pilot study since few formal investigations of these topics have been reported in the literature [10]. The interviewers conducted the interviews individually and transcribed their notes shortly thereafter.We analyzed the semistructured interviews using grounded theory [32], an approach in which the interviewer and one other researcher annotated each transcript independently. Annotations were formulated as themes from which the annotators induced initial hypotheses about the attitudes, motivation, and behavior of the interviewees. A third researcher summarized all annotations and themes, as well as whether they supported or refuted the particular hypothesis or hypotheses they related to.We modeled three of the semistructured interviews in accordance with the CI method described above and added the resulting workflow, sequence, and cultural models to the 10 sets of models developed during the CI phase. We did this in order to increase the variety of observations and add insights that may have been articulated during the interviews but not during the CI sessions. Subsequently, we consolidated the data into single flow, sequence, and cultural models. The flow model provided a good view of actors and their roles and the flow of information among them. The cultural model identified the cultural aspects that have a strong influence on whether and how collaborations are formed.We then derived a detailed set of requirements from the consolidated models and the results of the interviews and categorized them into five main themes: compatibility, communication, intermediation, information quality and access, and motivation. These themes served as the basis for developing the design ideas during the generative phase, which we describe next.Generative PhaseThe generative phase began with brainstorming design ideas for systems to help facilitate the establishment of collaborations in light of the system requirements we had formulated. Two initial brainstorming sessions resulted in a total of over 40 ideas. The ideas included semiautomatic updating of online profiles; locating collaborators through colleagues or matching research interests in published papers; utilizing online journal clubs, online video presentations, and live question-and-answer sessions; social tagging of research papers; facilitating directed social contact through methods such as ride sharing and hobby groups; and creating systems to support matchmaking through “social hubs,” such as department chairs. Several of the ideas drew on functions available in the Web 2.0 and ubiquitous/mobile computing technology spaces.Sixteen of the 40 ideas generated during the brainstorming phase were selected for formal concept validation. During this phase, we evaluated the design concepts with nine researchers at the University of Pittsburgh Health Science Campus. The participants represented scientists at the junior, senior, and executive levels with varying research foci (basic, clinical, and translational) at several schools. We presented each design idea as a real-life scenario to the participant and solicited feedback on its functionality and usefulness. Thus, we used the viewpoint of the end user as a central guiding principle for shaping our designs, an approach crucial to the development of user-centered applications [31]. The scenarios employed “personae,” which are archetypal representations of individuals that represent either the participant or individuals they would encounter when interacting with the system. For instance, “Carlos” was characterized as an inexperienced junior researcher at the School of Pharmacy in the early stages of his career. He had few contacts and was willing to be less selective about collaborative projects in order to gain experience and expand his network. “Bernice,” on the other hand, was a well-known biomedical researcher who demanded a rigorous work style and could afford to discontinue collaborations she felt were fruitless. The personae brought life to the scenarios and allowed participants to act and react naturally with regard to the proposed ideas. A facilitator presented the scenarios and guided user feedback through scenario-specific questions. At least one other observer was present to record notes. The sessions were audio-recorded as a reference for analysis.For each design idea, the individual ratings of the researchers were combined into a summary score that ranged from 1 (not needed) to 4 (very much needed). At the same time, the project team rated the feasibility of implementing each idea on a scale from 1 (low) to 3 (high). The feasibility rating integrated judgments about how difficult it would be to implement each idea based on technical, environmental, and cultural considerations.Based on the feedback from the concept validation sessions, we selected three related ideas for implementation and evaluation in a prototype. We implemented the design first as a wire frame, then as a high fidelity prototype. We performed think-aloud evaluations with four scientists using three use cases. The use cases described common scenarios that we asked participants to complete using the Digital|Vita prototype. Two observers kept notes on the interaction of each participant with the system, focusing on functions that were found to be either problematic or useful. The development team then brainstormed system improvements and implemented them to the highest degree possible. The high-fidelity prototype was used to produce a video about the system, which served as a way to solicit input from senior decision makers and external reviewers.The studies conducted as part of the background research phase and generative phase were approved by the University of Pittsburgh Institutional Review Board (IRB approval numbers: 0612065 and PRO07050299).ResultsThe background literature review, affinity diagramming, contextual inquiry, and semistructured interviews yielded five themes of requirements for systems to support the formation of collaborations. The themes are compatibility, communication, intermediation, information quality and access, and motivation. We briefly describe the themes below.Themes for System RequirementsCompatibilityBeyond expertise, successful collaborations require compatibility with respect to personality, work style, productivity, and many other factors [10,33]. Although exceptions exist, the majority of researchers interviewed saw compatibility of personality and work style as a prerequisite to collaboration. Therefore, more than a simple overlap of interests is needed to create a successful collaboration. The researchers we interviewed indicated that they would not trust an impersonal recommendation or suggestion made by a system about potential compatibility, putting them somewhat at odds with what users of dating sites and Facebook are apparently willing to do [34,35]. If researchers cannot assess compatibility with potential collaborators personally, they primarily appear to trust personal recommendations from colleagues. For this reason, expertise locating systems should show social connections between the collaboration seeker and potential collaborators.CommunicationFinding appropriate collaborators requires the ability to effectively search in domains other than your own using information that is comprehensive and descriptive [29,36]. However, researchers are often unlikely to be very familiar with the terminology they need in order to find a specific area of expertise in another domain [29]. One way that researchers currently solve this problem is by asking boundary-spanning colleagues and friends familiar with both realms about whom they should contact for help with solving a particular research problem. While the system should provide researchers with the ability to search directly for expertise, it should also make explicit who in their own professional network may be able to guide them effectively to other experts for resolving questions in different disciplines, organizational units, or research groups. A second requirement for communication is to broaden the ability to search for experts using more information than just publication databases such as Medline or Google Scholar. Those databases typically describe the knowledge and expertise of a researcher in less detail and less comprehensively than a complete curriculum vitae (CV). The CV can support a richer form of evaluation because it provides a more complete picture of the individual’s research-related activities, such as grants, grant reviews, patents, editorships, and positions in associations. Expertise locating systems should therefore not only allow the user to search a potential collaborator’s publications, but also their research interests, grant submissions, and biographical information.IntermediationSocial networks are very important for finding potential collaborators, assessing their suitability and compatibility, and establishing contact with them. Established researchers often use existing connections with colleagues as their primary resource for locating new collaborators. However, junior researchers with few or no contacts within the desired field may have significant difficulty initiating collaborations [26]. Researchers are more likely to contact a friend or colleague who they think will know an expert than to cold-call the author of a relevant research paper [24]. Advantages of personal contact include a higher likelihood of compatibility between parties, increased chances of a timely response (which is an issue when there is a status differential), and a less intimidating (and potentially face-saving) method of contacting a new party. Websites such as Facebook and LinkedIn circumvent cold-calling by integrating recommendation services and allowing users to see friends of friends. In this way, users’ networks are actually expanded to include their friends’ networks in addition to their own. The tendency to use friends/colleagues as intermediaries strongly supports the power and influence of existing social networks and suggests that a successful collaboration-networking site will need to leverage this construct for both identification and access [11,12]. Eysenbach [Eysenbach Medicine 2.0 Editorial, this issue] suggests the new scholarly term “apomediation” for the function of the intermediary, emphasizing the positive guidance toward high-quality resources (in this case, collaborators).Information Quality and AccessInformation profiles must be complete, correct, up-to-date, and comprehensive and allow fine-grained control over access to information by different audiences. Missing, incorrect, and out-of-date information and poor indexing (for instance, through the use of nonstandard vocabularies) of information profiles [15,20,29,37] make it difficult for a collaboration seeker to obtain the information necessary to assess the suitability and appropriateness of a prospective collaborator. Several commercial services, such as the Community of Science, rely on the user to keep their profile up-to-date and correct at all times. Others generate a “preliminary” profile for scientists from public sources, such as PubMed, Computer Retrieval of Information on Scientific Projects (CRISP), and the USPTO database, which the individual must correct and complete in order to generate a comprehensive, up-to-date profile. Several researchers we interviewed indicated that they had public online profiles but that they did not spend much time keeping them current. Given the many factors that collaboration seekers take into account when evaluating potential collaborations, information profiles should be as complete, correct, and up-to-date as possible [26]. In addition, some researchers indicated that given a choice, they would be selective about what information they would consider making public about themselves. For instance, interviewees seemed much more willing to disclose current ideas to collaborators from within their institutions than to those from competing institutions.MotivationKeeping online profiles up-to-date should require little or no effort and be integrated into the scientist’s existing workflow. In order to provide relevant and up-to-date information to colleagues, researchers must have an incentive to supply the information and keep it current. For example, our study participants regularly invested time updating information in their biographical and professional documents, such as their CV, biosketches for grants, and faculty evaluation forms. They were highly motivated to do so because these documents must be up-to-date in order to obtain grant funding, provide background information when invited to lecture or consult, and participate in university evaluations such as promotion and/or tenure decisions. There is no such motivation to update online profiles. In addition, our contextual inquiries showed that a major problem with the existing workflow is that researchers need to reformat and update the same information in multiple documents. We recognized this as an opportunity to draw researchers into using an online system. If it were possible to streamline the process by synchronizing information in multiple documents, the system would provide an incentive to keep information updated. Therefore, before a database of personal researcher profiles can be used as a tool to initiate collaboration, it must to be adopted as a repository of biographical and research-related information by a wide range of researchers. This “critical mass” problem is one of the classic challenges to the adoption of CSCW systems [38].\n                    Figure 1Results of the concept validation phase for 16 design ideas using nine scientist participants; need rating: 1 (low) to 4 (high); feasibility rating: 1 (low) to 3 (high)\n                Services for Optimizing How Scientific Collaborations Are EstablishedAs described in the Methods section, the research team generated a large number of ideas for one or more systems to support researchers in locating collaborators. Figure 1 shows the results of the concept validation phase for the 16 design ideas. It is important to note that there is not a 1:1 correspondence between the requirement themes and application ideas. Rather, the different application ideas are responsive to one or more requirement themes to different degrees. In selecting the ideas to be implemented, we aimed to respond to the requirements as best as possible within the context of a software application. The combination of three highly validated ideas, Profile Updating, CV/Profile Formatting, and Connections through Colleagues, appeared to satisfy our constraints most closely and were chosen to be implemented in a prototype system which we dubbed “Digital|Vita.” The three main sets of services implemented in Digital|Vita are the following:1. Maintaining, formatting, and semiautomated updating of biographical information: This set of services allows users to maintain biographical information and output it to several standard formats.2. Searching for experts: These services provide capabilities for searching for potential collaborators using a range of search criteria and allow searchers to exploit the social network represented in Digital|Vita in the process.3. Building and maintaining the social network and managing document flow: These services allow users to build a network of social connections, group colleagues into teams, and manage the flow of biographical documents within their teams.The following sections briefly describe these sets of services.Maintaining, Formatting, and Semiautomated Updating of Biographical InformationThis service is provided by the My Information (see Figure 2) data management function in DigitalVita, which stores biographical information about a user in a comprehensive and detailed manner. Information typically found in CVs, such as education, academic appointments, grants, and publications, can be entered and edited by the user. The items making up each collection, such as single publications, are stored as separate records and logically divided into fields in the database, enabling fine-grained information extraction and display.\n                    Figure 2The My Information component in Digital|Vita allows the user to enter and update biographical information through manual or semiautomated processes\n                Three services in My Information allow the user to enter and update biographical information:Importing information from existing sources: The primary method for populating biographical information is extraction from existing sources such as the National Library of Medicine’s Medline and the National Institutes of Health’s CRISP databases. A similar approach to retrieving and aggregating data from existing sources is being used in many other systems [13,20-23]. In Digital|Vita, records from these sources are pre-matched (for instance, through a name search) to the user, and the user simply confirms which records pertain to them. (This approach is used by the Faculty Research Interests Project (FRIP) [21] system currently in use at the University of Pittsburgh. When Digital|Vita is implemented at the University of Pittsburgh, publications will be imported from the existing FRIP database.)Propagating information through social networks: A second mechanism for acquiring biographical information is the semiautomated synchronization of updates made by colleagues in Digital|Vita. The process is semiautomated because all affected scientists are automatically notified about updated information, but each of them has to manually approve the update for inclusion in their own information. For instance, when Digital|Vita users manually enter a paper, they have the option of selecting coauthors from within the Digital|Vita system. When an entry is saved, Digital|Vita automatically propagates this update to the coauthors and displays it on each user’s Digital|Vita home page (see Figure 2). The coauthors can then confirm or reject the update for their own personal profile.Manually entering and updating information: The third mechanism for entering and updating information is manual entry. This is appropriate for data for which existing sources are neither available nor accessible. Examples of such data include professional appointments, degrees, and publications indexed in services that do not allow automated retrieval.\n                    Figure 3The My Documents component provides functions to output biographical information to several standard formats, customize information content, archive old versions, and include updates to biographical information selectively\n                While My Information allows the user to input and manage their biographical information, the My Documents function helps the user produce and archive several forms of output from that data. My Documents includes three services:Output to several standard formats: The current design of Digital|Vita provides for several standard output formats for biographical information such as a university-specific CV, a brief CV, and NIH and NSF biosketches. Users can choose the desired output format, generate the new document, and edit it according to their preferences.Customization and versioning: The ability to easily customize document content was deemed essential for the researchers we interviewed because they typically adapt biographical documents for specific grant applications, even if the format required for each is the same. With this service, users can customize documents with a simple checkbox approach—if an item is checked, it is included in the specific document. My Documents also supports versioning so that older versions of a specific document are available on demand.Selective updating: The system makes it explicit when the existing version of a document does not include recently updated information (see Figure 3) in order to allow the user to make an informed choice about including or excluding such updates. When the user customizes documents with information that is not contained in the My Information database, Digital|Vita allows the user to back-propagate the information to My Information. Thus, users do not have to interrupt their current workflow in order to make updates to My Information.As the user edits a specific document, the system displays the length of the document in pages in order to allow the user to observe page limits. In addition, the user can preview the printed version of the document; send it to colleagues in their professional network and recipients through email; and save the document in predetermined file formats.Searching for ExpertsOur background research indicated that researchers consider a variety of factors when choosing potential collaborators. For many, searching Medline and Google Scholar is only the first step in acquiring several types of information about their colleagues. The purpose of the My Information section in Digital|Vita is to store rich and comprehensive profiles of researchers in the database and make them available for flexible and powerful searching by others.Simple and advanced search of profiles: The first step in finding an expert within Digital|Vita is to allow users to query profiles flexibly. While the simple search in Digital|Vita only offers the capability to query profiles using keywords, the advanced search adds institution, department, location (for institutions with multiple campuses), publication activity, and relevance. (Relevance is a score indicating the level of expertise of the “hit” regarding the desired research topic.) Search results return key information about each hit (see Figure 4). They include academic affiliation, research interests, publications, and number of citations. Users can sort the search results and compare the appropriateness of potential collaborators. A potential trade-off of this design results from the fact that status, seniority, and relative experience of a person are now explicitly communicated. This could affect the decisions collaboration seekers make because a well-published and experienced researcher is now clearly identifiable as compared to a less published, less experienced researcher. Making these distinctions highly visible may potentially reduce the opportunities junior researchers are offered. On the other hand, it may allow the searcher to target a collaborator’s level of experience and expertise more directly. When users have identified one or more promising candidates for collaboration, they can access detailed profiles. Researchers’ profile pages contain information they have approved for inclusion by managing the My Profile section of their Digital|Vita. Thus, researchers have relatively granular control over which information is published about them. Typically, the profile page displays detailed information about their background, research interests, and publications (with links to PubMed for abstracts and, in some cases, full-text articles).Exploiting the social network to search: The search results page also displays connections through colleagues (identified by an icon symbolizing a social network, see icon next to \"Wendy Roberts\" in Figure 4). The icon shows two nodes if the individual is a colleague who belongs to the user’s social network and three nodes if the individual is a colleague of a colleague. This design paradigm resembles the functionality of sites such as Facebook and LinkedIn. Users may elect to search only in their extended network (ie, among individuals who are in the social network of their colleagues). Junior researchers may find this feature helpful to avoid having to cold-call potential collaborators. The Digital|Vita design currently does not provide a mechanism for asking a colleague for an introduction electronically, as other systems do.\n                    Figure 4A sample search results screen in Digital|Vita shows brief profiles of potential collaborators\n                Building and Maintaining the Social Network and Managing Document FlowThis service is managed in the My Colleagues section of Digital|Vita. It is intended for researchers to keep track of their collaborators, colleagues within their department, and general professional network of colleagues within Digital|Vita. It is the area in the system where users build the social network that they are able to exploit when they search for collaborators (see above) and where they manage document flow between themselves and their research teams.Creating links to colleagues: The value of social networks for recommending collaborators has been discussed earlier [39-41]. An obvious hurdle to establishing a social network is that there are few information sources from which data can be drawn to populate it directly. Nonetheless, in order to reduce the work for users, Digital|Vita generates suggestions for individuals to be included in a user’s social network by matching coauthors on papers and collaborators on grants with existing researchers in Digital|Vita. Users can then decide individually whether to include the suggested individuals in their social network. However, coauthorship and collaboration on grants are typically no more than partial indicators of collaboration [3]. Therefore, Digital|Vita users can ask anyone in the system to become their colleague. In this case, the system sends an electronic invitation, which the recipient either can accept or reject. In case of rejection, the recipient can opt to provide a reason. The requester is then notified about the recipient’s decision. One design alternative that was considered was not to notify the requester about the recipient’s decision. This alternative would provide a way to save face for both parties [42]. We decided against this design, however, in order to prevent users from thinking that the system was not working when receiving no response to their request.Assembling research teams: My Colleagues also provides a simple mechanism to label and organize groups of colleagues. The primary use of these groups is to manage the flow of biographical documents among them. The rationale for this feature is that many collaborations in academia arise within the context of pursuing a particular funding opportunity. Since one key activity in preparing grant applications is collecting biographical information from each team member, it was logical to add functions to Digital|Vita to support this effort. Users can create any number of named groups drawing from their list of colleagues on record in the system. As in real life, individual colleagues can belong to more than one group. Groups can be annotated with relevant information, for instance with the identifier of the funding opportunity the group is working on.Managing biographical document flow: Once a Digital|Vita user has created a named group, he or she can issue an electronic request to the group specifying the type of document requested (eg, NIH biosketch), the purpose for the request, and the date the information is needed by. Team members respond to requests through the system, which gives each person the opportunity to customize the requested document before it is sent. Digital|Vita issues automatic reminders to team members who have not responded by the due date. (Requesting documents in this manner is also possible between individuals.) The status of requests to and from other colleagues, as well as responses, is tracked in a Document Inbox. The Document Inbox allows users to send or request a document, as well as view and manage their recent document requests. Historical requests are accessible through a link to an archive. Before sending a document, users can preview it to ensure they are sending the correct document and that it contains the desired information. If new additions to the user’s biographical information have been made, the document can be edited directly before sending. Users can also decline a document request. Requests are archived automatically after the due date of the document has passed or when the user has sent the requested document.In summary, maintaining, formatting, and semiautomated updating of biographical information; searching for experts; and building and maintaining the social network and managing document flow are three sets of services designed to make the process of finding collaborators more efficient and effective and so facilitate the establishment of collaborations. We have focused on describing the Digital|Vita functionality as separate services in order to allow other researchers and developers to implement them selectively or all together in other systems.Additional Information About Digital|VitaThe preceding section presents a relatively abbreviated description of the functionality of the Digital|Vita system. A video illustrating a prototype of the system and its use is available in the Multimedia Appendix. In addition, the final report (dated July 2007) about the Digital|Vita prototype project, which includes a comprehensive description of the problem space, research, and development methods and the Digital|Vita design and functionality, including the design rationale, is available online. At present, the Digital|Vita development team is writing detailed system specifications for the development of a production application.DiscussionThe problem of connecting scientists with each other is not new. However, doing so efficiently and effectively has taken on particular relevance and urgency in an age when much of science is migrating to a multidisciplinary, collaborative, and team-oriented model. At the same time, while electronic systems to help connect scientists have existed for some time, to this point they appear to have played only a minor role in helping scientists form collaborations.Systematic approaches to designing systems to help researchers find collaborators are only in their infancy. We began this study with two basic research questions: (1) What requirements should systems for finding collaborators in biomedical science fulfill? and (2) Which information technology services can address these requirements? We believe that we have made an important contribution to the design of expertise locating systems with regard to both questions. The five main themes we have identified as requirements for such systems (compatibility, communication, intermediation, information quality and access, and motivation) show that collaboration seeking is a complex activity that does not depend simply on the ability to retrieve factual information about potential collaborators. It is clear from our exploration of these themes that social connectedness, personal and professional compatibility, and power differentials influence the formation of collaborations. This means that systems that do not model and leverage the social context are at a clear disadvantage in satisfying the social requirements for establishing collaborations.On the other hand, a rich informational representation of potential collaborators also appears to be important. Checking PubMed and Google for publications of a potential collaborator was only a starting point for many of our interviewees. Detailed investigation included other information resources, such as the NIH’s CRISP, as well as patent and other databases. Because of the fragmentation of information about potential candidates, a thorough background search on potential collaborators is time and effort intensive. The cost of a search, therefore, appeared to be a barrier to finding the most appropriate and qualified collaborators. It was therefore logical for our design to focus on the most comprehensive and up-to-date, but customarily also least accessible, information profile available: the CV of the individual scientist.Systems such as the Community of Science have long made biographical information a centerpiece of researcher profiles. Why do we think Digital|Vita may succeed where others have failed? Digital|Vita is centered on one component of a workflow that scientists almost universally perform on an ongoing basis—maintaining and updating the CV—and adds functions to support the establishment of collaborations. This design mirrors Payton’s [37] approach to use trails through an information space to identify individuals with common interests. In both cases, information useful for expertise location is a by-product of activities that are already being performed. In addition, CV maintenance in Digital|Vita remains in its local context. Moreover, institutions typically have idiosyncratic formats for CVs and evaluations, and thus systems designed to manage biographical information must be able to format it according to local requirements. To our knowledge, while DigitalVita includes this function, none of the major commercial expertise locating systems, such as the CoS, Collexis, and Research Crossroads, provide this functionality, which is a major barrier to their adoption.Managing biographical information within Digital|Vita not only requires no extra effort from a scientist compared to the traditional approach, it actually reduces effort because the raw biographical information is converted automatically to several frequently used standard formats. Making this workflow a central feature of Digital|Vita may prompt researchers to at least explore the collaboration-seeking functionality of the system.However, the simple availability of features to search for collaborators does not mean that they will be used. Encouraging researchers to seek collaborators through Digital|Vita as opposed to traditional methods faces significant obstacles. For instance, established researchers often are so well-informed and well-connected that they, on average, will outperform any electronic system. We therefore anticipate that Digital|Vita may be primarily attractive to younger scientists (who may be using social networking tools in their life outside of work) and scientists who are new to the University of Pittsburgh or who are planning to collaborate with individuals in disciplines that they are not very familiar with. Digital|Vita also faces a complex challenge in keeping information about a researcher private while at the same time marketing that researcher to maximum effect. We believe that the granular control Digital|Vita provides in determining what information is public and what is not will help individuals adjust their public profile to their preferences. Other potential barriers to adoption include establishing an initial critical mass of profiles adequate for finding and choosing collaborators and integrating the systems and its capabilities with the regular work practice of the institution and individual researchers.Future work on the Digital|Vita system will take two major directions. After development and implementation of the production system, we plan to design additional functions intended to improve the matching process among potential collaborators. Most likely, this research strand will focus on the development of algorithms to help pinpoint the most promising collaborators and bring new potential collaborative opportunities to a researcher’s attention. A second direction for the Digital|Vita effort will be to identify other information technology services to help scientists find and access resources that are useful for their work. For instance, we are currently working on a directory of computational resources at the University of Pittsburgh to support scientific problem solving.It is clear that electronic systems in support of research, and specifically those supporting the establishment of collaborations, will become increasingly important in the future. As more and more science goes “digital,” both in its execution as well as in its documentation, systems such as Digital|Vita will become essential to the everyday life and activities of scientists.\n\nREFERENCES:\n1. WeaverACMorrisonBBSocial NetworkingComputer20084129710010.1109/MC.2008.61\n2. AdamsJDBlackGCClemmonsRStephanPEPatterns of research collaborations in U.S. universities, 1981-1999National Bureau of Economic Research Productivity Program Meeting; December 6, 2002; Cambridge, MA\n3. KatzJSMartinBRWhat is research collaboration?Res Policy1997326111810.1016/S0048-7333(96)00917-1\n4. ArzbergerPFinholtTAData and collaboratories in the biomedical community. Report No. CREW-02-012002Ann Arbor, MICollaboratory for Research on Electronic Work, School of Information, University of Michigan\n5. BraunTSchubertAA quantitative view on the coming of age of interdisciplinarity in the sciences 1980-1999Scientometrics2003958118318910.1023/A:1025439910278\n6. ZerhouniEliasMedicine. The NIH RoadmapScience2003103302564263721452606610.1126/science.1091867302/5642/6314526066\n7. ZerhouniElias AUS biomedical research: basic, translational, and clinical sciencesJAMA2005921294111352816174693294/11/135210.1001/jama.294.11.135216174693\n8. ZerhouniElias AClinical research at a crossroads: the NIH roadmapJ Investig Med2006554417131715285510.2310/6650.2006.X0016\n9. CechThomas RRubinGerald MNurturing interdisciplinary researchNat Struct Mol Biol20041211121166915578058nsmb1204-116610.1038/nsmb1204-116615578058\n10. KrautREGalegherJEgidoCRelationships and tasks in scientific research collaborationHum-Comput Interact1987-198831315810.1207/s15327051hci0301_3\n11. ContractorNSZinkDChanMIKNOW: a tool to assist and study the creation, maintenance, and dissolution of knowledge networksIshida T, editor. Community Computing and Support Systems: Social Interaction in Networked Communities1998New YorkSpringer201217\n12. ContractorNSMongePRManaging knowledge networksManage Commun Q200211116224925810.1177/089331802237238\n13. McDonaldDWAckermanMSExpertise recommender: a flexible recommendation system and architectureProceedings of the 2000 ACM Conference on Computer Supported Cooperative Work (CSCW ’00)2000New YorkACM23124010.1145/358916.358994\n14. EricksonTKelloggWAKnowledge communities: online environments for supporting knowledge management and its social contextAckerman MS, Pipek V, Wulf V, editors. Sharing Expertise: Beyond Knowledge Management2003Cambridge, MAMIT Press299325\n15. de VriesSKommersPOnline knowledge communities: future trends and research issuesInt J Web Based Commun20041111512310.1504/IJWBC.2004.004804\n16. JohnsonCMA survey of current research on online communities of practiceInternet Higher Educ200141456010.1016/S1096-7516(01)00047-1\n17. MillenDRFontaineMAMullerMJUnderstanding the benefit and costs of communities of practiceCommun ACM20024454697310.1145/505248.505276\n18. AckermanMSPalenLThe Zephyr Help Instance: promoting ongoing activity in a CSCW systemTauber MJ, editor. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems: Common Ground (CHI ‘96)1996New YorkACM26827510.1145/238386.238528\n19. JacoviMSorokaVUrSWhy do we ReachOut?: functions of a semi-persistent peer support toolProceedings of the 2003 International ACM SIGGROUP Conference on Supporting Group Work (GROUP ‘03)2003New YorkACM16116910.1145/958160.958186\n20. MattoxDMayburyMTMoreyDEnterprise expert and knowledge discoveryBullinger H, Ziegler J, editors. Proceedings of the HCI International '99 (the 8th International Conference on Human-Computer Interaction) on Human-Computer Interaction: Communication, Cooperation, and Application Design. Volume 21999Mahwah, NJLawrence Erlbaum303307\n21. FriedmanP WWinnickB LFriedmanC PMickelsonP CDevelopment of a MeSH-based index of faculty research interestsProc AMIA Symp2000265911079886D20013611079886\n22. StreeterLALochbaumKEWho knows: a system based on automatic representation of semantic structureRIAO 88 Program: Conference with Presentation of Prototypes and Operational Demonstrations, User-Oriented Content-Based Text and Image Handling1988Paris, FranceCID380388\n23. MockusAHerbslebJDExpertise browser: a quantitative approach to identifying expertiseProceedings of the 24th International Conference on Software Engineering (ICSE '02)2002New YorkACM50351210.1145/581339.581401\n24. FlynnDASeeking peer assistance: use of e-mail to consult weak and latent tiesLibr Inform Sci Res2005271739610.1016/j.lisr.2004.09.005\n25. LampeCEllisonNSteinfieldCA face(book) in the crowd: social searching vs. social browsingProceedings of the 2006 20th Anniversary Conference on Computer Supported Cooperative Work (CSCW '06)2006New YorkACM Press16717010.1145/1180875.1180901\n26. EhrlichKLocating expertise: design issues for an expertise locator systemAckerman MS, Pipek V, Wulf V, editors. Sharing Expertise: Beyond Knowledge Management2003Cambridge, MAMIT Press137158\n27. Community of Science: resources for research, worldwideWho we areCommunity of Science2008http://www.cos.com/about/\n28. BaumACalesnickLEDavisGEGatchelRJIndividual differences in coping with crowding: stimulus screening and social overloadJ Pers Soc Psychol19821043482183010.1037/0022-3514.43.4.821\n29. McDonaldDWEvaluating expertise recommendationsEllis C, Zigurs I, editors. Proceedings of the 2001 International ACM SIGGROUP Conference on Supporting Group Work (GROUP '01)2001New YorkACM21422310.1145/500286.500319\n30. BeyerHHoltzblattKContextual designInteractions199961324210.1145/291224.291229\n31. BeyerHHoltzblattKContextual Design: Defining Customer-Centered Systems1998San Francisco, CAMorgan Kaufmann\n32. GlaserBGBasics of Grounded Theory Analysis: Emergence vs. Forcing1992Mill Valley, CASociology Press\n33. MattessichPWMonseyBRCollaboration: What Makes It Work. A Review of Research Literature on Factors Influencing Successful Collaboration. St1992Paul, MNAmherst H. Wilder Foundation\n34. FioreATDonathJSOnline personals: an overviewCHI '04 Extended Abstracts on Human Factors in Computing Systems2004New YorkACM1395139810.1145/985921.986073\n35. MerkleERRichardsonRADigital dating and virtual relating: conceptualizing computer mediated romantic relationshipsFam Relat20004949218719210.1111/j.1741-3729.2000.00187.x\n36. LuttersWGAckermanMSBosterJMcDonaldDWMapping knowledge networks in organizations: creating a knowledge mapping instrumentChung HM, editor. Proceedings of the Sixth Americas Conference on Information Systems (AMCIS '00) [CD-ROM]2000Atlanta, GAAIS Press20142018\n37. PaytonDWDiscovering collaborators by analyzing trails through an information spaceJensen D, Goldberg H, editors. Proceedings of the AAAI Fall Symposium on Artificial Intelligence and Link Analysis1998Menlo Park, CAAAAI Press8487\n38. GrudinJWhy CSCW applications fail: problems in the design and evaluation of organization of organizational interfacesProceedings of the 1988 ACM Conference on Computer-Supported Cooperative Work (CSCW '88)1988New YorkACM859310.1145/62266.62273\n39. YangSJHChenIYLA social network-based system for supporting interactive collaboration in knowledge sharing over peer-to-peer networkInt J Hum-Comput St20081661365010.1016/j.ijhcs.2007.08.005\n40. McDonaldDWRecommending collaboration with social networks: a comparative evaluationProceedings of the SIGCHI Conference on Human Factors in computing Systems (CHI '03)2003New YorkACM59360010.1145/642611.642714\n41. TerveenLMcDonaldDWSocial matching: a framework and research agendaACM Trans Comput-Hum Interact200512340443410.1145/1096737.1096740\n42. AokiPMWoodruffAMaking space for stories: ambiguity in the design of personal communication systemsProceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '05)2008New YorkACM18119010.1145/1054972.1054998"
+}

batch_13/PMC2553845.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2553845",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2553845\nAUTHORS: Raffi Tonikian, Yingnan Zhang, Stephen L Sazinsky, Bridget Currell, Jung-Hua Yeh, Boris Reva, Heike A Held, Brent A Appleton, Marie Evangelista, Yan Wu, Xiaofeng Xin, Andrew C Chan, Somasekar Seshagiri, Laurence A Lasky, Chris Sander, Charles Boone, Gary D Bader, Sachdev S Sidhu\n\nABSTRACT:\nPDZ domains are protein–protein interaction modules that recognize specific C-terminal sequences to assemble protein complexes in multicellular organisms. By scanning billions of random peptides, we accurately map binding specificity for approximately half of the over 330 PDZ domains in the human and Caenorhabditis elegans proteomes. The domains recognize features of the last seven ligand positions, and we find 16 distinct specificity classes conserved from worm to human, significantly extending the canonical two-class system based on position −2. Thus, most PDZ domains are not promiscuous, but rather are fine-tuned for specific interactions. Specificity profiling of 91 point mutants of a model PDZ domain reveals that the binding site is highly robust, as all mutants were able to recognize C-terminal peptides. However, many mutations altered specificity for ligand positions both close and far from the mutated position, suggesting that binding specificity can evolve rapidly under mutational pressure. Our specificity map enables the prediction and prioritization of natural protein interactions, which can be used to guide PDZ domain cell biology experiments. Using this approach, we predicted and validated several viral ligands for the PDZ domains of the SCRIB polarity protein. These findings indicate that many viruses produce PDZ ligands that disrupt host protein complexes for their own benefit, and that highly pathogenic strains target PDZ domains involved in cell polarity and growth.\n\nBODY:\nIntroductionModular protein–protein recognition domains are involved in the assembly of numerous intracellular complexes that mediate diverse cellular functions. Thousands of recognition domains are contained within the human proteome, and these have been classified into over 70 distinct families [1]. The PDZ (PSD-95/Discs-large/ZO-1) domain family is particularly interesting because it plays a key role in the development of multicellular organisms, in which PDZ domains are often found as components of multidomain scaffolding proteins involved in cell polarity and intercellular interactions [2,3]. PDZ domains are often embedded in proteins that assemble specialized subcellular sites, such as epithelial junctions [4], neuronal postsynaptic densities [5], and immunological synapses of T cells [6]. The biological importance of PDZ domains is further underscored by the identification of various PDZ-containing proteins as human disease and pathogen effector targets [4,7–15].Although the human genome encodes over 250 PDZ domains in over 100 proteins, most studies to date have focused on individual family members or a handful of domains [16–20]. Nevertheless, these studies uncovered general features of PDZ domain structure and function. Aside from unusual cases in which phospholipids [21,22] or internal motifs [23,24] are recognized, PDZ domains assemble intracellular complexes principally by recognition of C-terminal sequences in which specificity is mediated by interactions between ligand side chains and the PDZ domain binding surface [2]. Early studies grouped PDZ domains into two main specificity classes based on two ligand positions: class 1 (X[T/S]XφCOOH) and class 2 (XφXφCOOH), where X is any residue and φ is a hydrophobe [16,25]. Less common classes of PDZ domains, such as class 3 recognizing the motif X[ED]XφCOOH, were also identified [17]. However, subsequent studies have shown that the PDZ binding cleft can interact specifically with up to seven C-terminal ligand residues, enabling differentiation between biologically diverse ligands [20]. A recent large-scale analysis of mouse PDZ domains confirmed the highly specific nature of PDZ–ligand interactions but did not address the issue of PDZ domain classification [26]. To better understand how PDZ domains mediate cellular function and how hundreds of family members may compete for hundreds of potential ligands, we conducted a large-scale analysis using phage-displayed random peptide libraries. This establishes a specificity map and comprehensive classification system for the PDZ domain family, which provides insight into domain function and evolution and can be used to make novel discoveries about PDZ domain signaling systems.ResultsPDZ Domain Specificity PotentialWe used C-terminal peptide-phage display [27,28] to conduct a large-scale analysis of PDZ domain specificity, focusing on human and, for comparison, the simple metazoan Caenorhabditis elegans. Previous studies have shown that a phage-displayed combinatorial peptide library approach is a powerful tool to elucidate PDZ domain specificity and may be used to identify biologically relevant targets [20,29,30]. We cloned 72 out of 82 worm and 96 out of 254 human PDZ domains detected by three domain detection tools, BLAST [31], PFAM [32], and SMART [33]. The domain boundaries were defined as the union of all predicted domain regions plus ten amino acids on each side.The 168 cloned domains were expressed as GST fusion proteins, and 145 of these (57 worm and 88 human) could be purified in a stable, soluble form. These 145 proteins were used in binding selections with a C-terminal phage-displayed library containing greater than 10 billion random peptides, and we were successful in obtaining binding peptides against 82 domains (28 worm and 54 human). Failure to find binding peptides for the remaining domains may be due to instability of isolated domains or a requirement for larger, structured ligands not represented in our random peptide library [2]. Nonetheless, by sequencing approximately 10,000 binding clones, we were able to isolate approximately 3,100 unique peptide ligands for 82 PDZ domains (Tables S1 and S2). The domains used in this analysis and all associated peptides are available in a computer-readable format at http://baderlab.org/Data/PDZ.Consistent with the canonical preference of PDZ domains for hydrophobic C termini, the vast majority of the selected sequences (>97%) terminate with a hydrophobic residue, and for each domain, we therefore aligned the sequences on the basis of the C-terminal anchor position. The small number of peptides that are not canonical C-terminal binders were not considered in our analysis, although they are made available in our peptide files for others to analyze. To statistically model the binding specificity of each PDZ domain to enable computational analysis, each aligned peptide ligand set was used to create a position weight matrix (PWM). Each matrix column captures the amino acid binding preference of a PDZ domain at a ligand position as a probability distribution. From this PWM, the specificity of each ligand position is visualized as a sequence logo [34] and summarized using a specificity potential (SP) score ranging from least specific (any amino acid is recognized, SP score = zero) to most specific (only a single amino acid is recognized, SP score = one) (Figure 1A). For 72 PDZ domains, we had sufficient peptide data (n > 10) to calculate reliable SP scores. Our analysis reveals that essentially all PDZ domains recognize the last three ligand positions (0, −1, and −2), the majority recognize positions −3 and −4, and some recognize positions −5 and −6 (Figure 1B, SP > 0.2; Figure 1C, mean SP). The total specificity score per domain (SP t), calculated by summing the SP scores across the last nine ligand positions (Figure 1D), shows that most PDZ domain binding sites achieve high specificity through recognition of multiple features of the last seven residues of C-terminal peptide ligands (Table S3). These patterns are conserved across worm and human. Furthermore, there is no significant difference between the SP t values for worm (mean = 3.2 ± 0.9) and human domains (mean = 3.1 ± 1.0), indicating that increased human genome complexity has not been accompanied by a corresponding increase in overall PDZ domain specificity.Figure 1PDZ Domains Are Highly Specific across Multiple Ligand PositionsA total of 72 PDZ domains (each with greater than ten peptides) corresponding to 2,998 ligands were analyzed to assess specificity for each ligand position. Specificity was measured using the SP score, which ranges from zero (least specific) at a given ligand position to one (most specific). Bars are colored as follows: all PDZ domains (black), human (grey), or worm (white).(A) Specificity profile for a representative PDZ domain (C34F11.9a-1) with SP scores shown above each ligand position.(B) Fraction of PDZ domains exhibiting significant specificity (SP > 0.2) at each ligand position.(C) The mean SP value at each ligand position.(D) The distribution of total SP (SP t) summed over all ligand positions.We created a specificity map organizing all 82 successfully mapped PDZ domains. Hierarchical clustering was used to automatically place similar PDZ binding profiles (described by PWMs) in close proximity (Figure 2). This map reveals that approximately 90% of the domains fit into 16 distinct specificity classes, and the remainder represent unique specificities. By considering all recognized ligand positions, our comprehensive specificity map significantly expands the canonical PDZ domain classification system, which assigned only two main classes on the basis of specificity for ligand position −2 [16].Figure 2Specificity Map Classifies the PDZ Domain FamilyAll 82 PDZ domains studied were clustered to create a specificity map, which was used as a guide to manually define PDZ specificity classes. Of the 82 domains, 73 are assigned to one of 16 classes, labeled to the right of each domain name. For consistency with the established PDZ domain classification system [16], each class is denoted by a numeral based on the specificity for position −2, followed by a letter to account for specificity across the rest of the binding site. C. elegans domains are highlighted in yellow. Sets marked with identical Roman numerals in parentheses are homologous PDZ domains in human/worm orthologs. Domains that exhibit unique specificities not part of any class are denoted by asterisks (*). The 16 classes are defined by the following C-terminal motifs: 1a (φ[K/R]XSDV); 1b (Ω[R/K]ET[S/T/R/K]φ); 1c (φφETXL); 1d (ETXV); 1e (TWΨ); 1f (ΩΩTWΨ); 1g (φφφ[T/S][T/S]ΩΨ); 1h (φφ[D/E][T/S]WΨ); 2a (FDΩΩC); 2b (WXΩFDV); 2c (WΩφDΨ); 2d (φφX[E/D]φφφ); 2e (φφφφ); 2f ([D/E]φΩφ); 3a (WΩ[S/T]DWΨ); 4a (ΩφGWF); φ, hydrophobic (V, I, L, F, W, Y, M); Ω, aromatic (F, W, and Y); Ψ, aliphatic (V, I, L, and M); and X, nonspecific.PDZ Domains Are Versatile and RobustConsidering each ligand position independently, there are a striking number of distinct specificities for the last six positions (Figure 3). For instance, all domains prefer hydrophobic C termini, but there are eight distinct specificities for position 0. This suggests a vast potential for the PDZ domain family to bind different sequences. To assess the contributions of domain binding site positions to ligand binding and the specificity capacity of the PDZ domain family, we mutated ten Erbin PDZ domain (ERBB2IP-1) core binding-site positions and determined specificity profiles for each of the 91 single-residue mutants. The core binding positions are those that make contact (closer than 4.5 Å) with the peptide ligand in all of nine different PDZ domain structures (Protein Data Bank [PDB] entries 1N7T, 2H2B, 2H2C, 1I92, 2HE2, 1BE9, 2GZV, 1IHJ, and 1N7F). At each of the ten positions, mutations were made to amino acids that are abundant in the 82 natural human and worm PDZ domains for which we collected phage-derived specificity profiles (Figure 2).Figure 3Distinct Specificities of PDZ Domain Binding SitesThe specificity profiles of 72 PDZ domains reveal eight, seven, eight, seven, five, and three distinct specificities for ligand positions 0, −1, −2, −3, −4, and −5, respectively. At each position, distinct specificities are shown (magenta) with either the single-letter amino acid code or symbols, as follows: +, positive charge; −, negative charge; φ, hydrophobic (V, I, L, F, W, Y, and M); Ψ, aliphatic (V, I, L, and M); and Ω, aromatic (F, W, and Y).To minimize potential destabilization caused by structurally deleterious mutations, selections were performed at 4 °C, and under these low stringency conditions, the wild-type specificity profile (Figure 4) was somewhat less specific than that at room temperature (wild-type ERBB2IP-1 logo in Figure 2, class 1h). Phage selections were successful in all cases, and a total of approximately 3,400 unique ligands were sequenced. We compared the specificity profile of each mutant (Figure S1) to that of the wild type and visualized the differences as a heat map (Figure 4).Figure 4Sequence Determinants of PDZ Domain SpecificityHeat map summary of the effects of mutations on the specificity of ERBB2IP-1. Each row represents one mutant, ordered by PDZ domain binding-site position (labeled to the right of each set of rows), and each column represents one ligand position. Mutations were chosen to represent the diversity of amino acids in 82 natural PDZ domains for which we have phage data. To minimize potential destabilization caused by structurally deleterious mutations, selections were performed at 4 °C, and under these low stringency conditions, the wild-type specificity profile, shown at top left, was somewhat less specific than that at room temperature (Figure 2). The mutation listed to the left of each row, at the PDZ domain position listed at the right according to a structure-based nomenclature [55], causes a change in specificity, shown in each row. The blue-to-red gradient indicates increasing difference relative to wild type, normalized per column with significant differences highlighted in green (greater than one standard deviation away from the mean difference over the column). Selected mutant profiles are highlighted (depicted as sequence logos to the left of the corresponding row), with significant specificity changes in the logo boxed in red. Structures of ERBB2IP-1 with a bound peptide ligand [36] are shown with mutated positions depicted as spheres. Red side chains indicate ligand positions for which specificity is altered by mutations at PDZ positions shown as red spheres.Positions 0 and −2 define the most commonly used PDZ classification system [16]. The only significant change in specificity for position 0 was due to a substitution at position β1:β2–7, which lines the hydrophobic pocket that accepts the C-terminal ligand side chain. Changes in specificity for position −2 were caused by mutations in the α2 helix positions α2–1 and α2–5, which are close to the ligand side chain at this position [35]. The four substitutions at position α2–1 that do not alter specificity significantly (Y, N, Q, and K) are all capable of forming hydrogen bonds, and thus, can substitute functionally for the wild-type H, which hydrogen bonds with T at position −2 of an optimal ligand for ERBB2IP-1 [36]. The remaining 14 mutations result in class 2 specificity profiles with preference for hydrophobes at position −2. These results indicate that specificity for ligand positions −2 and 0 depends mainly on direct amino acid residue side chain interactions.Wild-type ERBB2IP-1 prefers W at ligand position −1, and although no mutation completely alters this preference, six mutations expand specificity to include other residue types. These mutations occur at four positions spread throughout the PDZ domain, and only the effects at one position (β3:α1–1) can be explained by changes in direct residue contacts. Mutations at two positions located far from position −1 (α2–1 and β2–4) expand specificity to include D, and it is likely that these mutations alter the ligand orientation and allow a D side chain to interact with the R side chain at position β3–5, which sits between ligand positions −1 and −3 [36]. The influence of indirect effects is demonstrated dramatically by the introduction of a flexible G residue at β3–4, which produces a PDZ domain with two distinct specificity profiles, one wild type and the other completely altered at positions −1 and −3. Thus, specificity for ligand position −1 can be influenced by direct and indirect interactions at positions throughout the PDZ binding site.Mutations at seven PDZ positions affect specificity for position −3, but only three of these positions (β2–2, β3–4, and β3–5) are in direct contact with this ligand position. Interestingly, many mutations at four other positions (β1:β2–7, β3:α1–1, α2–1, and α2–5) accentuate, rather than alter, the wild-type preference for negatively charged residues at position −3. A similar situation appears to exist for the upstream −4 and −5 positions, as the slight preference of the wild-type domain for hydrophobic residues at these positions (Figure 2, class 1h) is accentuated by mutations at five positions (β1:β2–7, β2–2, β2–4, β3–5, and α2–5), and most of these effects cannot be explained by changes in direct contacts. These effects may be caused by ligand orientation changes, which may allow for more favorable interactions between the PDZ domain and ligand residues upstream of position −2. Additionally, some mutations may weaken the energetic contributions from interactions with the three C-terminal ligand positions, and thus, ligand binding may become more dependent on favorable interactions involving upstream ligand positions. Thus, weaker interactions with ligand residues at positions −3, −4, and −5 can be affected in multiple indirect ways by mutations at numerous positions in the binding site.Our mutational and specificity prediction analyses provide general insights into PDZ domain specificity and have implications for prediction of binding specificity from domain sequence. Because specificity for ligand position −2 is mediated by local contacts with the PDZ α2 helix, point mutations at α2–1 and α2–5 are sufficient to substantially alter this specificity. Furthermore, there is a strong correlation between specificity for ligand position −2 and the sequence at the α2–1 position, as 44 of our 82 mapped domains (Figure 2) contain an H at the α2–1 position, and 37 (84%) of these prefer ligands containing T/S at position −2. We were not able to detect other strong correlations between individual PDZ domain positions and ligand specificity. In contrast, specificity for −1, −3, and positions further upstream depends on positions scattered throughout the PDZ domain and likely involves indirect conformational effects that subtly alter specificity at these positions without changing the specificity class. Thus, specificity in a PDZ domain is determined by multiple structural and chemical mechanisms involving both direct contacts and cooperative, long-range effects (Figure 5). Consequently, the binding site must be considered as a whole to accurately predict specificity from primary sequence.Figure 5Mutations Affecting PDZ Domain SpecificityERBB2IP-1 (grey) is shown with a bound peptide ligand (WETWVCOOH; cyan) (PDB entry 1N7T) [36]. PDZ domain binding-site positions that were subjected to mutagenesis are shown as spheres. In each panel, PDZ domain positions at which mutations affected specificity for the indicated ligand position are colored red and other mutagenized positions are colored green. PDZ domain positions are labeled in black according to a structure-based nomenclature [55].The clear selection of ligands by all PDZ mutants shows that the domain can function under high mutational pressure. Furthermore, 35 of 91 mutations analyzed caused a significant change in specificity for at least one ligand position (Figure 4). Taken together, these results show that PDZ domains are versatile and robust, as mutations frequently cause a change, rather than a loss of function.Conserved Specificity and Domain ExpansionBecause expansion of PDZ and other modular domains is correlated with increased organism complexity [37], we asked what role PDZ versatility plays in the evolution of complexity. Almost all PDZ specificity classes we define contain human and worm representatives, indicating that most of the human specificities are also present in the worm. Our dataset contains six worm/human ortholog gene pairs with mapped PDZ domain binding specificity in both species (Figure 6). Four of these pairs have nearly identical specificity profiles, and two are very similar. This level of conservation across more than one billion years of evolution separating worm from human [38] suggests that these specificity profiles are important for biological function. The limited number of conserved specificity classes used across two distant species suggests that most PDZ domain specificity classes arose early in evolution, and evolutionary constraints prevented new classes from forming following the divergence of worm and human. Consequently, additional complexity in the human PDZ domain family compared to that of the worm apparently arose through domain expansion and shuffling, rather than from the evolution of radical new specificities.Figure 6Specificity Profiles of Orthologous Domains Are Highly ConservedAll worm and human ortholog pairs with mapped PDZ domains in our dataset are shown. The domain architecture, as defined by SMART [33], is shown for each worm (top) and human (bottom) protein in an ortholog pair. The specificity profiles defined by peptide phage display are shown below or above the worm or human PDZ domains, respectively. The name and length of each protein is indicated on the left or right, respectively. The orthologous protein pairs are drawn to scale. The following protein pairs could be unambiguously identified as orthologs on the basis of common domain architecture and high sequence identity: (A) C34F11.9a/DVL2, (B) F54E7.3/PARD3, (C) Y54G11A.10/LIN7A, (D) C25F6.2a/DLG1, (E) W03F11.6/MLLT4, and (F) F17E5.1a/CASK.Specificity Predicted from Primary SequenceGiven the limited number of natural specificity classes and the mutant Erbin PDZ domain–ligand correlations observed, we asked whether we could use primary sequence to classify binding specificity of wild-type PDZ domains. Our extensive dataset based on phage-displayed random peptide libraries shows a clear correlation between binding-site identity and specificity. Domain pairs with binding-site sequence identities greater than 70% have specificity profiles with equivalent similarity to those within a specificity map class (Figure 7). This is the first time such a correlation has been shown for PDZ domains. We find that 69 of the remaining 254 unmapped worm and human PDZ domains have greater than 70% binding-site identity to mapped domains, and thus are predicted to have near identical binding profiles (Figure 8 and Table S4). An analogous analysis using full-length PDZ domain sequences reveals that domains with greater than 50% overall identity also exhibit highly similar specificity profiles (Figure S2). Thus, by combining experimentally mapped and predicted PDZ domain binding specificities, we roughly double the size of our PDZ domain specificity map and achieve 45% coverage of 336 predicted worm and human PDZ domains. As these novel rules require only in silico analysis of primary sequence and perform well across worm and human, it should be possible to predict the specificity of a given PDZ domain sequence in any organism.Figure 7PDZ Domain Sequence Identity Accurately Predicts Binding Specificity(A) ERBB2IP-1 structure (grey) is shown with a bound peptide ligand (WETWVCOOH; colored) [36]. PDZ domain binding site positions are shown as spheres, and positions that were analyzed by mutagenesis are colored green. PDZ positions are labeled in black according to a structure-based nomenclature [55], and peptide positions are labeled in red. We defined the PDZ binding site as 17 residues that make contact with the ligand (closer than 4.5 Å) in at least one of nine different structures (PDB entries 1N7T, 2H2B, 2H2C, 1I92, 2HE2, 1BE9, 2GZV, 1IHJ, and 1N7F).(B) The relationship between binding-site sequence identity and specificity profile similarity. Each point represents a pair of PDZ domains from our mapped set. Red circles represent pairs assigned to the same class, as defined in our specificity map, and blue squares represent all other pairs. The lower-right quadrant, absent of data points, contains an example for one pair of PDZ domains (ERBB2IP-1 and LRRC7–1), which exhibit a specificity profile similarity of 0.95 and a binding-site sequence identity of 0.88 (sequence mismatches are shown in red).Figure 8Prediction of PDZ Domain SpecificityA network view of predicted PDZ domain specificities. Worm and human PDZ domains are shown as blue or pink nodes, respectively. Diamonds denote domains with experimentally phage-mapped specificity profiles, and circles denote domains with predicted specificity profiles. Lines connect domains with greater than 70% sequence identity in the binding site, and line width is proportional to sequence identity. Connected domains are predicted to have high specificity profile similarity scores (>0.83). Network was created using Cytoscape 2.5 [54].Endogenous PDZ InteractionsOne major application of our PDZ domain specificity map is protein interaction prediction. As previously observed for numerous PDZ domains, phage display selects high-affinity peptide ligands through an iterative panning process, some of which are physiologically relevant [20,29,30,39,40]. These studies have also demonstrated a strong correlation between phage-derived PWM scores and affinities determined for synthetic peptides. However, the in vivo ligand interactions for any given PDZ domain depend on its intrinsic peptide specificity, the concentration and context of the protein in which it is located, and the range and concentration of accessible ligands. Also, some ligands may interact with suboptimal affinities to regulate specific biological processes. Thus, endogenous C termini closely matching our mapped specificities are likely to bind the given PDZ domain in vitro, but determination of in vivo binding requires additional experimental support.To significantly reduce the human PDZ interactome search space and prioritize interactions for future experimental testing, we detected the best matched C-terminal sequences in the human proteome for individual domains using a PWM-based scoring algorithm and a score threshold that stringently allows only the top few hits (Table S5). The network of potential human PDZ domain mediated protein interactions obtained in this manner contains 322 interactions between our 54 experimentally mapped PDZ domains and 228 human proteins. These high-scoring ligands are significantly enriched in known PDZ interactors (27 interactions are known, p = 8.6 × 10−18) (Table S5) and in gene function annotation consistent with known PDZ ligand-associated functions (Figure S3) [2–4]. Thus, our prioritized list is likely enriched in novel bona fide human PDZ protein interactions.Our prioritization approach is useful because potential ligands can be considered for experimental follow-up in order of similarity to the phage-mapped specificity profile. For instance, when studying a particular protein of interest, it may be useful to expand the list of potential ligands to include additional lower-scoring ligands that may nonetheless be physiologically relevant. To illustrate the utility of this approach, we focused on DLG1, one of the first and best characterized PDZ-containing proteins (Table S6) [2,5,41], and extended the potential ligand list for the three DLG1 PDZ domains by choosing a less stringent score threshold. Our predicted interactions capture eight of the 11 known ligands for DLG1 (Table S7) and identify many additional potential ligands with scores comparable to those of the known ligands. The list includes many known ligands for the closely related DLG homologs (DLG2, −3, and −4) (Table S6), and is enriched in gene function annotations consistent with known functions of the DLG homologs (Figure S4), which are involved in establishing and maintaining cell polarity, and interact with ion channels, guanyl-nucleotide exchange factors, and other signal transduction proteins [2,5,41].Pathogenic PDZ InteractionsPathogenic viruses and bacteria use short linear peptide motifs that target PDZ domains and other peptide-binding modules to perturb host signaling networks [8,10–13]. To study the extent of this pathogenic subversion of host cellular processes, we computationally identified 89 viral proteins with C termini matching mapped PDZ domain specificities better than the potential endogenous interactors defined above. These cover all PDZ domain specificity classes (Table S5). Our results suggest that many viruses specifically target distinct PDZ domain classes with high-affinity ligands that compete with endogenous interactors and interfere with normal physiology.To further explore viral targeting of PDZ domain proteins, we focused on SCRIB, a protein known to be targeted by human papilloma virus (HPV) [10]. SCRIB contains multiple PDZ domains and is involved in establishing and maintaining membrane polarity in epithelia, neurons, and T cells [4–6,40]. We identified numerous potential SCRIB viral ligands in our initial network and in an additional network derived from a recent database of avian influenza genomes [13]. We used a less stringent score cutoff because we desired a more focused and sensitive search designed to be experimentally validated. We measured in vitro affinities of ten potential SCRIB viral ligands using synthesized peptides (peptides 1–10, Figure 9A). Each peptide interacts with at least one SCRIB PDZ domain, but not, in general, with the first PDZ domain of ZO-1 (TJP1–1), which has an overlapping, but different, specificity [20]. As a further test, we found that a Herpes virus ligand (peptide 12) matching the specificity profiles of SCRIB and ZO-1 PDZ domains bound to both, while another ligand (peptide 13) matching only the ZO-1 PDZ domain specificity profile interacted only with this domain. These experiments show that our specificity map is useful to guide experiments and that viral proteins contain C-terminal motifs that are capable of specifically targeting distinct sets of PDZ domains.Figure 9Viral Proteins Interfere with Host Cellular Function by Targeting the PDZ Domains of SCRIB(A) Many viral proteins bind SCRIB PDZ domains. Affinities were determined as IC50 values for peptides representing viral C termini binding to SCRIB PDZ domains and the first PDZ domain of ZO-1 (TJP1–1) [20]. Ligand sequence positions that match the specificity profiles for SCRIB, TJP1–1, or both, are colored green, blue, or red, respectively. Orange and yellow indicate high-affinity (IC50 < 10 μM) or moderate-affinity (IC50 > 10 μM) interactions, respectively. Asterisks (*) indicate no detectable interaction (IC50 > 500 μM). Double asterisks (**) indicate influenza A strain designations [13], rather than RefSeq accession numbers.(B) Loss of the late phase of T cell polarization induced by our designed synthetic peptide that targets SCRIB PDZ domains 1, 2, and 3. The receptor Crtam interacts with the PDZ domains of SCRIB to control cell growth and maintain polarity of T cells [40]. These effects are reversed by the addition of our designed peptide (Π-RSWFETWV, peptide 14) that binds with high affinity to the SCRIB PDZ domains, but not by a designed nonbinding peptide with mutations at the 0 and −2 positions (Π–RSWFEAWA, peptide 15). The symbol Π denotes the internalization sequence from the Antennapedia protein (RQIKIWFQNRRMKWKK), which has been shown to be internalized into cells [56]. Naive Crtam−/− CD4 T cells were electroporated with plasmid DNA expressing Crtam or a mock DNA control. Cells were treated with synthetic peptides (1.0 μM) and stained for Talin, a marker for the leading edge of polarized T cells [40].(C) Our designed SCRIB PDZ-binding peptide (peptide 14) triggers T cell proliferation. Cells were treated with plasmid DNA and peptides, as described in (B), and cellular proliferation was measured by the incorporation of [3H]-thymidine. Data are representative of three independent experiments. Error bars indicate the standard deviation (SD). The p-value was determined by statistical analysis performed with a control using the Dunnett method.Some of our predicted viral ligands are known, whereas many are novel. The HPV E6 protein (peptide 1) is known to disrupt SCRIB function and causes loss of epithelial cell polarization and concomitant hyperproliferation [10], and the PDZ-binding motif is only present in high-risk, oncogenic HPV strains. PDZ-binding motifs are also present in the Tax proteins of human T-lymphotrophic (HTLV) type 1 viruses (HTLV-1) (peptide 2) that cause adult T-cell leukemia/lymphoma (ATLL), but are absent from HTLV type 2 viruses that do not cause lymphoproliferative disorders [12]. The HTLV-1 Tax protein has been shown to interact with other PDZ domains [12], but we now show previously unreported interactions with SCRIB. In another example, the NS1 proteins of influenza A from avian and highly pathogenic human strains typically contain negatively charged residues at position −3 and can interact with many PDZ domains, whereas those from less pathogenic human strains typically contain positively charged residues at this position and show low reactivity with PDZ domains [13]. This is consistent with our specificity map (Figure 2) and with our binding data showing that the C termini of NS1 proteins from two avian influenza strains (peptides 9 and 10) interact with the SCRIB PDZ domains, but one from a low-pathogenicity human strain (peptide 11) does not. These confirmations, and new viral–SCRIB interactions we find involving herpes, vaccinia, myxoma, and fibroma viruses (peptides 3–8), suggest that many pathogenic viruses utilize a common mechanism to advantageously target SCRIB-mediated complexes involved in cell polarity and growth control.To further explore this hypothesis, we investigated whether the PDZ-binding motif alone is sufficient to induce effects associated with pathogenic viral infections. Based on our PDZ specificity map, we designed a synthetic peptide that binds to the SCRIB PDZ domains with high affinity (peptide 14). We recently described an interaction between the C terminus (ESIVCOOH) of the cell-surface receptor Crtam and SCRIB in T cells, which is critical to attenuate proliferation and maintain a late phase of T cell polarity [40]. Our designed SCRIB PDZ-binding peptide abrogates these functions and causes loss of late-phase T cell polarization and reverses the hypoproliferative effects of Crtam. In contrast, a nonbinding control peptide (peptide 15) has no effect (Figure 9). These effects are remarkably similar to the hyperproliferative phenotypes previously observed to be induced by the Tax protein of HTLV-1 in lymphocytes [12] and the E6 protein of high-risk HPV in epithelial cells [10]. The induction of hyperproliferation is likely to be advantageous for viral replication in general, and thus, it appears that the ability to disrupt polarity by interference with SCRIB PDZ domain complexes is a crucial factor in conferring high pathogenicity to many viruses, including HPV, HTLV, and influenza A.DiscussionWe have presented the first large-scale specificity map of a domain family across species, based on approximately 3,100 peptide ligands, covering approximately one half of the combined set of 336 predicted PDZ domains encoded by the human and C. elegans genomes. We find that the PDZ domain family is surprisingly complex and diverse, recognizing up to seven C-terminal ligand residues and forming at least 16 unique specificity classes across human and worm. Further interpretation of our map reveals that PDZ domains are versatile, capable of binding diverse ligand sequence motifs, and are robust under high mutational load. Highly mutagenized WW and SH3 domain sequences also retain the ability to recognize proline-rich sequences [42,43], suggesting that functional robustness under high mutational pressure may be a general feature of peptide-binding modules. Although we find that the PDZ domain family likely evolved by domain expansion rather than from evolution of radically new specificities, the robustness of the PDZ domain may be ideal for supporting rapid evolution of interaction networks through testing of many functional variants under evolutionary pressures that select for novel ligands.For the first time to our knowledge, we find a predictive correlation between PDZ domain sequence and binding specificity. This correlation bridges a gap in our ability to predict protein interactions and signaling networks from a genome. Because the correlation holds across worm and human, it can likely be used for accurate in silico predictions of PDZ domain specificity in other species. The predictive correlation will likely improve by considering additional features of the domain and ligand, including additional physicochemical and structural properties, class-specific binding sites, and cooperative and indirect effects of residues across the entire domain [44].One major application of the PDZ domain specificity map is the prediction of interaction networks that provide insights into functions of PDZ domains in the cell. We have described a preliminary use of this map to prioritize human PDZ-mediated protein interactions, which directly led to novel insights into normal [40] and pathogen biology. Ideally, computational predictions would be more robust and less dependent on experimental support. A major impediment to domain-based protein interaction prediction is assessing predicted interaction validity, due to a dearth of bona fide PDZ ligands in the literature. This is further complicated because many interactions are known only for full-length proteins, often containing multiple PDZ domains. Thus, there are not enough known domain–peptide interactions to robustly learn optimal prediction parameters and accurately measure sensitivity and specificity. For instance, there is only one known worm interaction for our mapped PDZ domains, involving the lin-7 PDZ-containing protein that interacts with the C terminus of let-23 (the worm homolog of the epidermal growth factor receptor) [45], but this single interaction does agree with our data. Notwithstanding this limitation, higher-scoring potential interactors are more likely to be relevant, as supported by comparison to known PDZ domain interactions, analysis of gene function annotation, and comparison to well-studied examples [10,20,29,40]. This is not surprising, as similar computational methods with the same goal were successful in previously published PDZ domain studies [18,20,40]. Our list of prioritized interactions is a useful resource for biologists interested in further studying potential interactions involving PDZ domains. Ideally, this will lead to discovery of additional PDZ interactions that can be used to improve performance of computational protein interaction prediction methods.The PDZ specificity map is useful for a number of applications. Our analysis reveals numerous viral proteins that may interact with PDZ domains to hijack host cellular networks for pathogen benefit. Based on our data, we were able to design synthetic viral-like peptides that target a specific biological system in human T cells. Analogously, therapeutics could be designed to alter PDZ-based cell systems for medical benefit [9], following approaches similar to those used for the development of peptidomimetics targeting other peptide-binding modules [46]. Our specificity map will prove invaluable for guiding peptidomimetic design, as it offers starting points for numerous PDZ domain specificity classes, provides optimal PDZ-binding peptides useful for target validation in cellular assays, and helps identify related domains and potential in vivo interaction partners that must be considered for cross-reactivity. Further, the versatility and robustness of the PDZ domain make it ideal for use in engineering synthetic biological systems [47].Correct use and interpretation of our specificity map requires understanding of its physiological relevance. The map was constructed using optimal binding ligands detected by phage display. In the cell, however, natural ligands often bind suboptimally to enable regulation in signaling systems, and may have noncanonical binding modes. Multiple cellular factors must be considered to determine physiologically relevant binding using our data, including affinity, concentration, localization, and competition between similar PDZ domains for the same ligands.Nevertheless, our C-terminal PDZ ligand dataset defines the diverse specificities of the PDZ domain family that have enabled the evolution of complex cellular architecture and provides a strong foundation for further work investigating physiologically relevant interactions. Further, the experimental and computational methods we describe are readily applicable to dozens of families of peptide recognition domains, covering a significant fraction of cell signaling proteins in eukaryotic genomes. We anticipate that derivation of specificity maps for all peptide recognition domains will enable the accurate prediction of physiologically relevant wiring diagrams directly from sequenced genomes.Materials and MethodsPDZ domain identification.For cloning, the domain boundaries were defined as the union of all domains found by a combination of the National Center for Biotechnology Information (NCBI) BLAST [31], PFAM [32], and SMART [33] with an additional ten amino acids on each side, as described previously [28]. For computational analysis, PDZ domain sequences were defined using hmmpfam precomputed by Ensembl [48] and downloaded from Ensembl 36 (homo_sapiens_core_36_35i and caenorhabditis_elegans_core_36_140c) using the Ensembl Perl API. Other domain resources were used to manually supplement this list when PDZ domains of interest were not found in Ensembl. Percent sequence identity was determined using a multiple sequence alignment of all human and worm PDZ domain sequences generated using MUSCLE 3.6 [49] with default parameters. Sequence identity was calculated as: number of matched positions divided by (aligned sequence length minus gap positions).PDZ domain cloning and expression.DNA fragments encoding PDZ domains of interest (Tables S1 and S2) were amplified from cDNA using the polymerase chain reaction and were cloned into vectors designed for the expression and purification of PDZ domains fused to glutathione S-transferase, as described [20,28]. All expression vectors were verified by DNA sequencing.Selection of PDZ domain ligands.C-terminal phage-displayed peptide libraries (>1010 unique members) were used to isolate ligands for PDZ domains using a series of iterative panning steps, as described [28]. Specific binding clones were individually tested for positive interactions with cognate PDZ domains by phage ELISA. Specific binding clones derived from sibling phages (identical in DNA sequence) were only counted once. This resulted in the isolation of 3,100 unique peptides from over 10,000 sequenced peptides for 82 PDZ domains. Data for six domains (ERBB2IP-1, SCRIB-1, SCRIB-2, SCRIB-3, TJP1–1, and TJP2–2) were from a previous study [20]. For other natural PDZ domains, a random decapeptide library was used in binding selections performed at room temperature. For the analysis of ERBB2IP-1 point mutants, a random heptapeptide library was used at 4 °C. A final manual inspection of all sequences removed a small number (92) that did not conform to the canonical C-terminal binding mode or did not agree with the major specificity profile (these sequences are available in our online files at http://baderlab.org/Data/PDZ).Specificity potential.For each PDZ domain, the set of peptide ligands was used to create a binding profile statistical model as a PWM. The SP for a given column (position) of a PWM was calculated as is done for the letter height in a sequence logo [34], except normalized to range from 0 to 1 instead of 0 to 4.32 (log 20). A SP value of one means the given PDZ domain is completely specific for a single amino acid at that position, and a value of zero means that there is no preferred amino acid at that position. As no domains exhibited specificity at position −9, the SP value for each position was corrected for bias in the peptide library by subtracting the specificity score at position −9 for the entire set of 3,066 unique PDZ domain ligands (peptides found to bind to the 82 natural PDZ domains in the specificity map).Specificity map construction.A specificity map (Figure 2) was constructed by clustering all 82 natural human and worm PDZ domain specificity profiles. We used hierarchical agglomerative clustering with average linkage with a custom specificity profile (PWM) distance metric, defined below. Each set of binding peptides was aligned, anchored by the C terminus, and used to create a specificity profile statistical model as a position weight matrix (PWM). Since the peptide library was constructed using a 32-codon set defined by the NNK nucleotide ambiguity codes, it is expected that some amino acids occur more frequently than others. To correct for this bias, the PWM was normalized by dividing amino acid frequencies by their expected frequency in the NNK codon set, following established practice [36]. To consider physicochemical similarities among amino acids and enable a more biologically relevant PWM similarity calculation, PWMs were recalculated to use a reduced amino acid alphabet of five groups constructed as follows: STQN (polar), KRH (basic), DE (acidic), FLAMPWIVCY (hydrophobic), and G. Distance between PWM pairs a and b, D(a,b), was then calculated using the following distance metric:where a = PWM A; b = PWM B; w = number of columns in the PWM (i.e., ten–amino acids peptide length); ∑ = alphabet used in PWM = the number of rows in the PWM (i.e., five groups for the reduced alphabet defined above); and L = a letter from the reduced alphabet.This distance metric is normalized such that zero represents perfectly similar PWMs and one represents perfectly dissimilar PWMs. Similarity is calculated as 1 − distance. The clustering results were visualized in standard fashion (as a tree with branch lengths corresponding to PWM pair distances). Leaf ordering for graphical tree display was optimized using the algorithm of Bar-Joseph et al. [50]. Clustered PWMs were graphically represented using sequence logos [34] and displayed as leaves on the cluster tree. Summary PWMs were constructed for all tree nodes as averages of all leaf node PWMs connected to that node and displayed on each tree node with a size proportional to their horizontal position in the tree. Software for creating this tree is available from http://baderlab.org/Software/LOLA. The tree was manually annotated by an expert (SSS) to define specificity classes.Erbin mutant heat map construction.Each row of the heat map depicted in Figure 4 represents one of 91 Erbin mutants, and each column represents one of seven positions in the mutant specificity profile. To quantify the difference between mutant and wild-type profiles, both profiles were statistically modeled as PWMs and compared using the distance metric, D, described above, on a per PWM position basis (one position per heat map table cell). PWMs used 20 amino acids, instead of the reduced set described above, to provide a more fine-grained measure of PWM distance (which requires that the ∑ parameter in Equation 1 be the set of 20 amino acids). Resulting differences were then normalized across all 91 mutants per position (that is, over an entire column). The linear color gradient represents the difference of the mutant specificity profile compared to wild type, from minimum (blue) to maximum (red) distance. Significant differences are greater than one standard deviation away from the mean and highlighted in green. The map was then manually annotated with relevant sequence logos and structures.Prioritization of endogenous PDZ domain ligands.The PWM representing the specificity profile for each of the 82 mapped PDZ domains was used to search for C-terminal matches in the RefSeq human proteome (∼33,700 proteins) and viral proteome (∼54,600 proteins) sets, downloaded May 21, 2007 [51]. SCRIB viral ligands were chosen in a more focused, but earlier, search of the RefSeq viral database from July 15, 2006, containing approximately 48,000 viral proteins and an additional set of influenza virus proteomes [13]. One pseudocount was added to each cell of the PWM to allow a low level of matching for amino acids that are not seen by phage display, but nevertheless, may be involved in a natural interaction. Matching potential ligands not having a hydrophobic C terminus were removed, since it is known from structural evidence that the PDZ domain is highly specific for hydrophobes at this position. A small number (∼10 ligands) were eliminated using this filter. PWM scores are calculated as the negative base 10 logarithm of the normalized probability of the PWM sequence match, such that low, positive scores represent better PWM matches. To enable comparisons across PWMs, scores were normalized to the range defined by the maximum and minimum possible scores that could be produced by the given PWM. Human proteome PWM score thresholds were calculated automatically for each PDZ domain by progressively testing increasing score thresholds and choosing a cutoff score when the number of new hits at a given threshold was higher than the cumulative number of hits of all previously tested thresholds (not including the last score tested). Viral PWM score cutoffs were defined to be better than the best human scoring match to ensure that only viral interactions with closer PWM matches than any human protein were predicted as viral PDZ ligands. Viral proteome matches in proteins that contained the string “phage” in their descriptions were removed, as these viruses likely target bacteria, not eukaryotic cells.Additional ligands can be found by choosing a more liberal score threshold. We chose not to optimize the score threshold to maximize overlap of predicted interactions to a benchmark, since we could find no suitable benchmark. Available interactions in which the PDZ domain involved in the interaction was known were too few, and remaining interactions involved full-length proteins without domain-level resolution. We did not complete this conservative prioritization for worm due to the absence of almost any known interactions for worm PDZ domain containing proteins and poor Gene Ontology (GO) annotation coverage for predicted ligands.The overlap statistic was computed based on all protein interactions involving our mapped PDZ domains in the UniHI database [52]. Approximately 8% (27) of prioritized interactions are known, which corresponds to a p-value of 8.6 × 10−18. The p-value was computed by calculating the overlap of 1,000 random shufflings of our prioritized interactions with the UniHI benchmark. Our randomly shuffled prioritized interactions overlapped approximately 6.7 interactions in the benchmark on average, with a standard deviation of approximately 3.4 and with a normal distribution. The normal distribution was used to calculate the p-value.Gene Ontology.For predicted endogenous PDZ domain ligands, GO term enrichments were computed against all available GO annotation packaged with BiNGO on January 17, 2007, using the BiNGO Cytoscape plugin [53,54] with HUGO gene identifiers, the hypergeometric statistical test of significance, and Benjamini and Hochberg False Discovery Rate (FDR) correction with a significance level of 0.05.Software.All computational analyses were performed using custom Java software built with BioJava 1.4 and JFreeChart. Free, open source Java software for visualizing and clustering specificity profiles is available from http://baderlab.org/Software/LOLA and for searching sequence databases using a specificity profile to find potential protein interactions via a Cytoscape plugin from http://baderlab.org/Software/BRAIN.Data.All peptide sequences are available from http://baderlab.org/Data/PDZ and have been submitted to the DOMINO and PDZBase (accession codes cpe_3 to cpe_176) databases.Affinity assays.Peptides were synthesized with acetylated C termini. The binding affinities of peptides for PDZ domains were determined as 50% inhibition concentration (IC50) values using competition ELISAs, as described [20]. The IC50 value was defined as the concentration of peptide that blocked 50% of PDZ domain binding to immobilized peptide.T cell assays.Naïve Crtam−/− CD4 T cells were purified and activated with plate-bound anti-CD3 and anti-CD28, as described [40]. On day 4, T cells were electroporated with 4 μg of pIRES_GFP or pIRES_GFP/Crtam plasmid DNA by Amaxa Nucleofector (program X-01). Synthetic peptides (1.0 μM) were added into the cultures, and after 6 h, transfected cells were restimulated at 1 × 106 cells/ml with plate-bound antibodies and were fixed 14 h later for Talin staining. After 42 h, [3H]-thymidine (1 μCi/well) was added, and the plates were harvested 8 h later.Supporting InformationFigure S1Specificity Profiles for Point Mutants of ERBB2IP-1Each column heading shows the wild-type sequence at each position, which is labeled according to a structure-based nomenclature shown in Figure 4 [36]. Each column shows the specificity profiles for the point mutants analyzed at that position, and the identity of each mutation is indicated to the left of each profile. The wild-type profile as observed at 4 °C is shown for comparison in the box at bottom left. The specificity profiles were derived from approximately 3,400 binding peptide sequences.(3.85 MB PDF)Click here for additional data file.Figure S2PDZ Domain Sequence Identity Accurately Predicts Binding SpecificityThe relationship between overall PDZ domain sequence identity and specificity profile similarity. Each point represents a pair of PDZ domains from our mapped set. Red circles represent pairs assigned to the same class, as defined in our specificity map, and blue squares represent all other pairs.(267 KB PDF)Click here for additional data file.Figure S3Gene Ontology Terms Associated with Endogenous Prioritized Human PDZ LigandsOverrepresented terms for the human proteins in Table S5 were calculated using the BiNGO plugin for Cytoscape and shown as circles [53,54]. Arrows connect less specific to more specific terms, as defined in GO. The area of a given node is proportional to the number of genes annotated in the corresponding GO category in our set of prioritized ligands. The node color scale is proportional to the p-value of the overrepresentation of the GO term in the set relative to the number of genes in the genome. White nodes are not significantly overrepresented, however they are included in order to illustrate the GO structure within the three different categories.(A) GO biological process.(B) GO molecular function.(C) GO cellular localization.(850 KB PDF)Click here for additional data file.Figure S4Gene Ontology Terms Associated with Endogenous Predicted Ligands for the PDZ Domains of DLG1Overrepresented GO biological process terms for the proteins in Table S6 were calculated using the BiNGO plugin for Cytoscape and shown as circles [53,54]. The analysis was performed as in Figure S3.(270 KB PDF)Click here for additional data file.Table S1Summary of Analyzed C. elegans PDZ DomainsThe domains are colored as follows: green, purified and peptide-phage selections were successful; blue, purified but peptide-phage selections were unsuccessful; grey, not cloned or could not be purified in a soluble form from Escherichia coli. The listed amino acid ranges indicate the length of the constructs used in the analysis and not necessarily the PDZ domain boundaries defined by computational domain identification.(43 KB PDF)Click here for additional data file.Table S2Summary of Analyzed Human PDZ DomainsThe domains are colored as follows: green, purified and peptide-phage selections were successful; blue, purified but peptide-phage selections were unsuccessful; grey, not cloned or could not be purified in a soluble form from E. coli. The listed amino acid ranges indicate the length of the constructs used in the analysis and not necessarily the PDZ domain boundaries defined by computational domain identification.(47 KB PDF)Click here for additional data file.Table S3\nSP Values for Human and C. elegans PDZ DomainsValues were only determined for 72 domains that had ten or more selected peptides. C. elegans domains are highlighted in yellow, and values greater than or equal to 0.2 are highlighted in green.(41 KB PDF)Click here for additional data file.Table S4PDZ Domain Specificity PredictionSpecificity of unmapped domains is predicted to be highly similar (>0.83 profile similarity) to the mapped domains with greater than 70% sequence identity in the binding site. The species of origin is shown to the right of each domain.(9.07 MB PDF)Click here for additional data file.Table S5Click here for additional data file.Prioritized Endogenous and Viral Ligands for Human PDZ DomainsPDZ domains are listed in alphabetical order and prioritized ligands are listed in ascending order by interaction score. Lower interaction scores are better. For each domain, only those viral ligands with better scores than the best endogenous ligand are shown. Viral or known endogenous ligands are highlighted in magenta or yellow, respectively.(74 KB PDF)Table S6Prioritized Endogenous Ligands for the PDZ Domains of DLG1PDZ domains are listed in numerical order, and for each, the 50 ligands with the best prediction score are listed in ascending order by interaction score. Lower interaction scores are better. Known ligands reported in PDZBase for DLG1 or the other DLG homologs (DLG2, −3, and −4) are highlighted in yellow or blue, respectively.(41 KB PDF)Click here for additional data file.Table S7Known Ligands for the PDZ Domains of DLG1Known ligands reported in PDZBase are ordered in ascending order of interaction score for the highest scoring PDZ domain (Table S6). Lower interaction scores are better. NA denotes ligands that were not predicted by the scoring algorithm.(24 KB PDF)Click here for additional data file.\n\nREFERENCES:\n1. PawsonTNashP\n2003\nAssembly of cell regulatory systems through protein interaction domains\nScience\n300\n445\n452\n12702867\n2. HarrisBZLimWA\n2001\nMechanism and role of PDZ domains in signaling complex assembly\nJ Cell Sci\n114\n3219\n3231\n11591811\n3. ShengMSalaC\n2001\nPDZ domains and the organization of supramolecular complexes\nAnnu Rev Neurosci\n24\n1\n29\n11283303\n4. HumbertPRussellSRichardsonH\n2003\nDlg, Scribble and Lgl in cell polarity, cell proliferation and cancer\nBioessays\n25\n542\n553\n12766944\n5. KimEShengM\n2004\nPDZ domain proteins of synapses\nNat Rev Neurosci\n5\n771\n781\n15378037\n6. Ludford-MentingMJOliaroJSacirbegovicFCheahETPedersenN\n2005\nA network of PDZ-containing proteins regulates T cell polarity and morphology during migration and immunological synapse formation\nImmunity\n22\n737\n748\n15963788\n7. AartsMLiuYLiuLBesshohSArundineM\n2002\nTreatment of ischemic brain damage by perturbing NMDA receptor- PSD-95 protein interactions\nScience\n298\n846\n850\n12399596\n8. AltoNMShaoFLazarCSBrostRLChuaG\n2006\nIdentification of a bacterial type III effector family with G protein mimicry functions\nCell\n124\n133\n145\n16413487\n9. DevKK\n2004\nMaking protein interactions druggable: targeting PDZ domains\nNat Rev Drug Discov\n3\n1047\n1056\n15573103\n10. DoorbarJ\n2006\nMolecular biology of human papillomavirus infection and cervical cancer\nClin Sci\n110\n525\n541\n16597322\n11. FreseKKLeeSSThomasDLLatorreIJWeissRS\n2003\nSelective PDZ protein-dependent stimulation of phosphatidylinositol 3-kinase by the adenovirus E4-ORF1 oncoprotein\nOncogene\n22\n710\n721\n12569363\n12. HallWWFujiiM\n2005\nDeregulation of cell-signaling pathways in HTLV-1 infection\nOncogene\n24\n5965\n5975\n16155603\n13. ObenauerJCDensonJMehtaPKSuXMukatiraS\n2006\nLarge-scale sequence analysis of avian influenza isolates\nScience\n311\n1576\n1580\n16439620\n14. RogeljBMitchellJCMillerCCMcLoughlinDM\n2006\nThe X11/Mint family of adaptor proteins\nBrain Res Rev\n52\n305\n315\n16764936\n15. UematsuKKanazawaSYouLHeBXuZ\n2003\nWnt pathway activation in mesothelioma: evidence of Dishevelled overexpression and transcriptional activity of beta-catenin\nCancer Res\n63\n4547\n4551\n12907630\n16. SongyangZFanningASFuCXuJMarfatiaSM\n1997\nRecognition of unique carboxyl-terminal motifs by distinct PDZ domains\nScience\n275\n73\n77\n8974395\n17. StrickerNLChristophersonKSYiBASchatzPJRaabRW\n1997\nPDZ domain of neuronal nitric oxide synthase recognizes novel C-terminal peptide sequences\nNat Biotechnol\n15\n336\n342\n9094134\n18. VaccaroPBrannettiBMontecchi-PalazziLPhilippSHelmer CitterichM\n2001\nDistinct binding specificity of the multiple PDZ domains of INADL, a human protein with homology to INAD from Drosophila melanogaster\nJ Biol Chem\n276\n42122\n42130\n11509564\n19. van HuizenRMillerKChenDMLiYLaiZC\n1998\nTwo distantly positioned PDZ domains mediate multivalent INAD-phospholipase C interactions essential for G protein-coupled signaling\nEMBO J\n17\n2285\n2297\n9545241\n20. ZhangYYehSAppletonBAHeldHAKausalyaPJ\n2006\nConvergent and divergent ligand specificity among PDZ domains of the LAP and zonula occludens (ZO) families\nJ Biol Chem\n281\n22299\n22311\n16737968\n21. ZimmermannPMeerschaertKReekmansGLeenaertsISmallJV\n2002\nPIP(2)-PDZ domain binding controls the association of syntenin with the plasma membrane\nMol Cell\n9\n1215\n1225\n12086619\n22. WuHFengWChenJChanLNHuangS\n2007\nPDZ domains of Par-3 as potential phosphoinositide signaling integrators\nMol Cell\n28\n886\n898\n18082612\n23. HillierBJChristophersonKSPrehodaKEBredtDSLimWA\n1999\nUnexpected modes of PDZ domain scaffolding revealed by structure of nNOS-syntrophin complex\nScience\n284\n812\n815\n10221915\n24. PenkertRRDiVittorioHMPrehodaKE\n2004\nInternal recognition through PDZ domain plasticity in the Par-6-Pals1 complex\nNat Struct Mol Biol\n11\n1122\n1127\n15475968\n25. NourryCGrantSGBorgJP\n2003\nPDZ domain proteins: plug and play\nSci STKE\n2003\nRE7\n12709532\n26. StifflerMAChenJRGrantcharovaVPLeiYFuchsD\n2007\nPDZ domain binding selectivity is optimized across the mouse proteome\nScience\n317\n364\n369\n17641200\n27. FuhGPisabarroMTLiYQuanCLaskyLA\n2000\nAnalysis of PDZ domain-ligand interactions using carboxyl-terminal phage display\nJ Biol Chem\n275\n21486\n21491\n10887205\n28. TonikianRZhangYBooneCSidhuSS\n2007\nIdentifying specificity profiles for peptide recognition modules from phage-displayed peptide libraries\nNat Protoc\n2\n1368\n1386\n17545975\n29. LauraRPWittASHeldHAGerstnerRDeshayesK\n2002\nThe Erbin PDZ domain binds with high affinity and specificity to the carboxyl termini of delta-catenin and ARVCF\nJ Biol Chem\n277\n12906\n12914\n11821434\n30. ZhangYAppletonBAWuPWiesmannCSidhuSS\n2007\nStructural and functional analysis of the ligand specificity of the HtrA2/Omi PDZ domain\nProtein Sci\n16\n1738\n1750\n17656586\n31. AltschulSFGishWMillerWMyersEWLipmanDJ\n1990\nBasic local alignment search tool\nJ Mol Biol\n215\n403\n410\n2231712\n32. FinnRDMistryJSchuster-BocklerBGriffiths-JonesSHollichV\n2006\nPfam: clans, web tools and services\nNucleic Acids Res\n34\nD247\n251\n16381856\n33. SchultzJCopleyRRDoerksTPontingCPBorkP\n2000\nSMART: a web-based tool for the study of genetically mobile domains\nNucleic Acids Res\n28\n231\n234\n10592234\n34. SchneiderTDStephensRM\n1990\nSequence logos: a new way to display consensus sequences\nNucleic Acids Res\n18\n6097\n6100\n2172928\n35. DoyleDALeeALewisJKimEShengM\n1996\nCrystal structures of a complexed and peptide-free membrane protein-binding domain: molecular basis of peptide recognition by PDZ\nCell\n85\n1067\n1076\n8674113\n36. SkeltonNJKoehlerMFTZobelKWongWLYehS\n2003\nOrigins of PDZ domain ligand specificity: structure determination and mutagenesis of the Erbin PDZ domain\nJ Biol Chem\n278\n7645\n7654\n12446668\n37. VogelCChothiaC\n2006\nProtein family expansions and biological complexity\nPLoS Comput Biol\n2\ne48\ndoi:10.1371/journal.pcbi.0020048\n16733546\n38. WangDY-CKumarSHedgesSB\n1999\nDivergence time estimates for the early history of animal phyla and the origin of plants, animals and fungi\nProc R Soc Lond B\n266\n163\n171\n39. RunyonSTZhangYAppletonBASazinskySLWuP\n2007\nStructural and functional analysis of the PDZ domains of human HtrA1 and HtrA3\nProtein Sci\n16\n2454\n2471\n17962403\n40. YehJHSidhuSSChanAC\n2008\nRegulation of a late phase of T cell polarity and effector functions by Crtam\nCell\n132\n846\n859\n18329370\n41. FunkeLDakojiSBredtDS\n2005\nMembrane-associated guanylate kinases regulate adhesion and plasticity at cell junctions\nAnnu Rev Biochem\n74\n219\n245\n15952887\n42. PanniSDenteLCesareniG\n2002\nIn vitro evolution of recognition specificity mediated by SH3 domains reveals target recognition rules\nJ Biol Chem\n277\n21666\n21674\n11929862\n43. RussWPLoweryDMMishraPYaffeMBRanganathanR\n2005\nNatural-like function in artificial WW domains\nNature\n437\n579\n583\n16177795\n44. LocklessSWRanganathanR\n1999\nEvolutionarily conserved pathways of energetic connectivity in protein families\nScience\n286\n295\n299\n10514373\n45. SimskeJSKaechSMHarpSAKimSK\n1996\nLET-23 receptor localization by the cell junction protein LIN-7 during C. elegans vulval induction\nCell\n85\n195\n204\n8612272\n46. ZobelKWangLVarfolomeevEFranklinMCElliottLO\n2006\nDesign, synthesis, and biological activity of a potent Smac mimetic that sensitizes cancer cells to apoptosis by antagonizing IAPs\nACS Chem Biol\n1\n525\n533\n17168540\n47. DueberJEYehBJChakKLimWA\n2003\nReprogramming control of an allosteric signaling switch through modular recombination\nScience\n301\n1904\n1908\n14512628\n48. HubbardTJAkenBLBealKBallesterBCaccamoM\n2007\nEnsembl 2007\nNucleic Acids Res\n35\nD610\nD617\n17148474\n49. EdgarRC\n2004\nMUSCLE: a multiple sequence alignment method with reduced time and space complexity\nBMC Bioinformatics\n5\n113\n15318951\n50. Bar-JosephZGiffordDKJaakkolaTS\n2001\nFast optimal leaf ordering for hierarchical clustering\nBioinformatics\n17\nSuppl 1\nS22\nS29\n11472989\n51. DurbinREddySRKroghAMitchisonG\n1998\nBiological sequence analysis: probabalistic models of proteins and nucleic acids\nCambridge\nCambridge University Press\n356\n52. ChaurasiaGIqbalYHanigCHerzelHWankerEE\n2007\nUniHI: an entry gate to the human protein interactome\nNucleic Acids Res\n35\nD590\n594\n17158159\n53. MaereSHeymansKKuiperM\n2005\nBiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks\nBioinformatics\n21\n3448\n3449\n15972284\n54. ShannonPMarkielAOzierOBaligaNSWantJT\n2003\nCytoscape: a software environment for integrated models of biomolecular interaction networks\nGenome Res\n13\n2498\n2504\n14597658\n55. AppletonBAZhangYWuPYinJPHunzikerW\n2006\nComparative structural analysis of the Erbin PDZ domain and the first PDZ domain of ZO-1. Insights into determinants of PDZ domain specificity\nJ Biol Chem\n281\n22312\n22320\n16737969\n56. VucicDDeshayesKAckerlyHPisabarroMTKadkhodayanS\n2002\nSMAC negatively regulates the anti-apoptotic activity of melanoma inhibitor of apoptosis (ML-IAP)\nJ Biol Chem\n277\n12275\n12279\n11801603"
+}

batch_13/PMC2556326.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2556326",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2556326\nAUTHORS: Eric J Hill, Ann B Vernallis\n\nABSTRACT:\nBackgroundThe direction of cytokine secretion from polarized cells determines the cytokine's cellular targets. Leukemia inhibitory factor (LIF) belongs to the interleukin-6 (IL-6) family of cytokines and signals through LIFR/gp130. Three factors which may regulate the direction of LIF secretion were studied: the site of stimulation, signal peptides, and expression levels. Stimulation with IL-1β is known to promote IL-6 secretion from the stimulated membrane (apical or basolateral) in the human intestinal epithelial cell line Caco-2. Since LIF is related to IL-6, LIF secretion was also tested in Caco-2 following IL-1β stimulation. Signal peptides may influence the trafficking of LIF. Two isoforms of murine LIF, LIF-M and LIF-D, encode different signal peptides which have been associated with different locations of the mature protein in fibroblasts. To determine the effect of the signal peptides on LIF secretion, secretion levels were compared in Madin-Darby canine kidney (MDCK) clones which expressed murine LIF-M or LIF-D or human LIF under the control of an inducible promoter. Low and high levels of LIF expression were also compared since saturation of the apical or basolateral route would reveal specific transporters for LIF.ResultsWhen Caco-2 was grown on permeable supports, LIF was secreted constitutively with around 40% secreted into the apical chamber. Stimulation with IL-1β increased LIF production. After treating the apical surface with IL-1β, the percentage secreted apically remained similar to the untreated, whereas, when the cells were stimulated at the basolateral surface only 20% was secreted apically. In MDCK cells, an endogenous LIF-like protein was detected entirely in the apical compartment. The two mLIF isoforms showed no difference in their secretion patterns in MDCK. Interestingly, about 70% of murine and human LIF was secreted apically from MDCK over a 400-fold range of expression levels within clones and a 200,000-fold range across clones.ConclusionThe site of stimulation affected the polarity of LIF secretion, while, signal peptides and expression levels did not. Exogenous LIF is transported in MDCK without readily saturated steps.\n\nBODY:\nBackgroundMost cells secrete a variety of cytokines, either continuously or in response to stimulation. The spatial regulation of cytokine secretion is important since it determines which neighbouring cells respond, but little is known about how cytokine secretion in the constitutive pathway is regulated. With their prominent apical and basolateral domains, epithelial cells have been studied extensively for their ability to sort membrane proteins to specific domains. Sorting depends on hierarchical signals and occurs in multiple stages of the secretory pathway [as reviewed in [1-3]]. Lipid raft-associated and independent pathways have been identified for transport from the trans-Golgi to the apical membrane [4]. Less is known about the sorting of secreted proteins. N-glycans and O-glycans as well as proteinaceous patches have been proposed as sorting signals. Lectins may cooperate with lipid rafts in the transport routes [5].LIF is a secreted, glycosylated cytokine, of the IL-6 family, that exhibits pleiotropic activities in a wide range of tissues and cell types. LIF is produced by epithelial cells during development and during infection and inflammation, including in the intestine [6], the uterus [7] the lung [8], and the kidney [9].When cultured on permeable supports, Caco-2 cells display the features of differentiated small intestinal enterocytes, forming polarized monolayers with tight junctions. We chose to characterise LIF secretion by Caco-2, since Moon et al. [10] made an intriguing observation about IL-6 secretion in these cells. They reported that IL-6 secretion was higher from the membrane domain (apical or basolateral) that was stimulated with IL-1β than from the non-stimulated domain, suggesting that secretion could be locally facilitated by a pro-inflammatory stimulus. Whether or not LIF is secreted in a similar way is interesting since LIF and IL-6 are related cytokines and may be co-regulated.We also studied LIF secretion in MDCK, a renal cell line which displays characteristics of distal tubule cells and collecting duct cells [11]. MDCK were selected since the cell line is well suited to exogenous expression. Murine LIF (mLIF) and human LIF (hLIF) are easily distinguished from endogenous canine LIF in ELISAs. Inducible expression systems based on tetracycline-regulated promoters are available for MDCK which allow the effects of different expression levels on secretion to be measured. Few studies have systematically examined the effects of varying expression levels. Marmorstein et al. [12] conducted a study of vascular endothelial growth factor (VEGF165) and transforming growth factor β1(TGF-β1) secretion at different expression levels in a retinal pigment epithelial cell line (RPE-J) by varying the amount of adenovirus used to transfect the cells. Their results indicated that these cytokines were secreted in apical pathways that were readily saturated, suggesting specific transporters. For example, VEGF dropped dramatically from 100% apical secretion to 30% apical with increased expression levels. We tested whether this result could be generalized to mLIF and hLIF in MDCK.Exogenous expression also facilitated study of the contribution of LIF signal peptides to LIF trafficking. Two secreted isoforms of LIF are derived from the murine LIF gene. The isoforms, corresponding to transcripts LIF-D and LIF-M, encode different signal peptides, which in at least some fibroblasts appeared to affect the localization of LIF protein. After cleavage of the signal peptides, the mature murine proteins had the same polypeptide sequence but the LIF-D protein was secreted and freely diffusible; whereas the LIF-M was secreted but remained associated with the extracellular matrix [ECM;  [13]]. LIF-D and LIF-M transcripts are also produced from the hLIF gene [14,15]; however, differences in localization of the products have not been reported. If, as in fibroblasts, LIF-M was associated with the ECM in epithelial cells, LIF-M secretion might be preferentially directed toward the basolateral membrane, where the ECM is present. To test this hypothesis, LIF-M and LIF-D were exogenously expressed in MDCK.Our results suggest that exogenous LIF is predominantly secreted in an apical direction in MDCK, regardless of signal peptide or expression level, indicating a lack of specific transporters for LIF.ResultsCaco-2 secreted LIFThe polarized secretion of a number of cytokines including IL-6 has been described in Caco-2 cells [10]. Secretion of LIF by Caco-2 has not been previously reported. Caco-2 cells were grown as a confluent monolayer in tissue culture flasks for two weeks to allow them to fully differentiate. Conditioned media was then collected over a 48 hour period and assayed for LIF-like activity using a BA/F3-mLIFR-mgp130 assay (hLIF will activate mLIFR/mgp130). The conditioned media displayed a LIF-like activity, which was detectable over background in up to a 1:16 dilution and did not support BA/F3 parental cells (lacking mLIFR/mgp130). A polyclonal antibody raised against eukaryotic human LIF inhibited the proliferative responses by 75% (data not shown). The neutralization by anti-LIF antibodies demonstrates that a large proportion if not all of the activity in the conditioned media is due to hLIF itself and not to other LIFR-dependent ligands.An hLIF ELISA, which uses antibodies raised against glycosylated hLIF, detected 32 pg/ml hLIF in the Caco-2 conditioned media. This concentration is similar to what was predicted from the BA/F3-mLIFR/mgp130 assays. The ELISA result demonstrates that Caco-2 constitutively secrete small amounts of LIF.Caco-2 preferentially secreted LIF from the basolateral membraneTo determine the polarity of LIF secretion, Caco-2 cells were grown on permeable filters. The formation of a sealed monolayer after 14 days was confirmed by measurements of the transepithelial electrical resistance (TEER), which, in unstimulated cells reached approximately, 1,800 Ω.cm2. Following conditioning for 24 hours, the media from the apical and basal chambers was analyzed by ELISA. As the volumes recovered from the apical and basolateral chambers differed (0.9 ml vs 1.4 ml), amounts rather than concentrations are reported. Unstimulated Caco-2 cells produced on average 46 pg total LIF protein of which 41% was secreted into the apical chamber (Fig. 1a and 1b). This differential accumulation of LIF is likely to represent genuine differences in secretion since transport of LIF across the monolayer was very limited (data not shown).Figure 1Caco-2 regulated the direction of secretion of LIF and IL-6 in response to IL-1β. Caco-2 were grown on permeable supports. a, c) The total amount of LIF and IL-6 secreted into both apical and basal chambers was determined by ELISA 24 hours after addition of IL-1β (1 ng/ml) into either the apical (A), basal (B) or both chambers (A+B). b, d) The percentage of LIF and IL-6 secreted into the apical chamber following treatment. Values represent the mean +/- SEM (n = 3). ND (not detected). *p < 0.05 vs. no stimulation; #p < 0.05 apical vs. basolateral stimulation.Secretion of LIF and IL-6 by Caco-2 was influenced by the site of stimulation with IL-1βApical or basal treatment alone with IL-1β resulted in an increase in the total amount of LIF secreted, but the increase was not statistically significant. Simultaneous treatment of both sides (dual stimulation) resulted in a significant 7-fold increase relative to unstimulated cells (Fig. 1a). Differences between apical and basolateral stimulation were apparent when considering the polarity of secretion. With apical stimulation, 42% of the LIF was secreted apically, whereas with basolateral stimulation only 20% of the LIF was secreted apically, demonstrating that the site of stimulation is important in determining the direction of secretion (Fig 1b). Dual stimulation yielded an intermediate percentage secreted apically. The effect of IL-1β on LIF trafficking is not due to a disruption of the monolayer integrity since stimulation with IL-1β did not have a significant effect on monolayer permeability (data not shown).The samples collected for LIF secretion were also assayed for IL-6 using an IL-6 ELISA. In contrast to LIF, IL-6 secretion was not detected in untreated cells (Fig. 1c). Apical, basal or dual stimulation with IL-1β all increased IL-6 secretion, but only the basal and dual increases were significant, reaching an average of 284 pg with dual stimulation. When stimulated at the apical membrane, 57% of the IL-6 was secreted into the apical chamber (Fig. 1d). In contrast, basolateral treatment with IL-1β resulted in only 33% secreted apically. With stimulation of both sides of the monolayer, an intermediate percentage was secreted apically. As with LIF, therefore, the polarity of IL-6 secretion was differentially regulated by IL-1β at the apical and basolateral membranes. The two cytokines show a closely related but not identical pattern of secretion.MDCK secreted a LIF-like factor from the apical membranesThe availability of tetracycline-responsive systems (Tet off) for MDCK makes it an attractive cell line for manipulating secretion. Demonstrating that MDCK secreted endogenous LIF in a polarized manner would suggest that the cells actively sort LIF. Studying the secretion of endogenous LIF in MDCK, however, is problematic since antibodies to canine LIF are not available. In lieu of an ELISA, conditioned media from unstimulated MDCK was assayed on BA/F3-mLIFR-mgp130 cells. The media demonstrated a LIF-like activity that was dependent upon LIF-R/gp130 (data not shown). Since the efficiency with which this assay detects the canine-LIF-like activity is unknown, the assay can only measure relative amounts of LIF-like activity. When monolayers of MDCK were grown on permeable filters, and conditioned media was collected after 48 hours, all of the LIF-like activity was detected in the apical compartment (Fig. 2). The apical activity could be diluted up to at least 8-fold and still stay within the linear range of the assay. Taking account of the greater dilution of proteins secreted into the basal compartment, if 20% or more of the LIF-like activity was secreted into the basal compartment, it should have been detected in the bioassay. Stimulating the MDCK cells with human IL-1β did not increase the secretion of the LIF-like activity (data not shown) perhaps because the human cytokine is a weak agonist for the canine receptor [16].Figure 2MDCK apically secreted an endogenous LIF-like factor. MDCK cells were grown on permeable supports for 24 hours and the conditioned media was collected from the apical and basal chambers. The collected media was then analyzed for LIF-like activity using the BA/F3 mLIR-mgp130 cell assay. Results are expressed as the A570 value of cells assayed for proliferation by 3-(4,5-dimethylthiazolyl-2)-2, 5-diphenyl tetrazolium bromide (MTT). Values represent the average activity present in a dilution series of media collected from an individual transwell assayed in triplicate. Two Transwells® are shown with one denoted by a square and one by a triangle. A single representative experiment is shown of n = 3.MDCK mLIF clones secrete LIF primarily from the apical membraneSince the level of expression has been shown to influence the direction of secretion of growth factors, several stable MDCK Tet-off LIF cell lines were generated for study using the PBI-L expression vector. PBI-L contains a bi-directional promoter, allowing clones to be selected according to their expression of luciferase and/or LIF. To test the polarity of secretion at the minimum and maximum levels of LIF expression, clones were grown on permeable supports and analyzed both in the fully switched on state (no doxycycline; No Dox) or fully switched off (1 μg/ml Dox). Both mLIF and hLIF-expressing clones were generated since differences in their surface-exposed amino acids might result in differences in protein sorting.As MDCK cells are responsive to stimulation with hLIF [17], the effects of mLIF and hLIF on monolayer integrity were first tested. Eukaryotic mLIF or hLIF (1 ng/ml) was added to both sides of confluent monolayers and the effects on TEER were tested after 24 hours. MDCK cells produced TEERs in the region of 700–850 Ω.cm2 after five days growth on filters. No significant difference in TEER was observed following stimulation with either mLIF or hLIF in comparison to untreated cells. At most small amounts of LIF were transported across the monolayer. After adding LIF to the apical or basal side, less than 5% of the recovered LIF was detected in the opposite chamber.Four MDCK-Tet-off mLIF-D clones were analyzed by growing the clones on permeable supports and then assaying the mLIF which accumulated in the apical or basal compartment after 24 hours. Clones 9 and 19 secreted on average 61 and 7 nanograms of mLIF respectively with around 67% of the mLIF secreted apically. They did not differ significantly in their preference for apical secretion at lower expression levels (< 1 ng) observed with 1 μg/ml Dox (63% apical; Fig. 3a and 3b). mLIF-D clone 11 did not produce detectable amounts of mLIF in the presence of 1 μg/ml Dox but with lower amounts of Dox (0.5 ng/ml and 1 ng/ml), mLIF was detected (21 pg and 57 pg respectively). The difference in the percentage secreted apically between No Dox and 0.5 ng/ml Dox is statistically significant, but the difference is small (65% vs. 71%). mLIF-D clone 4, without Dox, secreted 62% apically and with Dox, secreted 48% apically. This 14% drop in the percentage secreted apically was significant but was small compared to the dramatic changes observed for VEGF in RPE-J cells (see discussion). Comparing across expression levels from picograms to nanograms of LIF, there is no compelling evidence for trafficking machinery which can be saturated by mLIF-D.Figure 3MDCK Tet-off cells secreted exogenous mLIF-D and mLIF-M predominantly in the apical direction across a range of expression levels. MDCK cells grown on permeable filters were incubated without doxycycline (No Dox) or varying concentrations of doxycycline (Dox) to regulate the expression level of mLIF-D (a, b) or mLIF-M (c-f). After 24 hours, conditioned media was collected from the apical and basal chambers and analyzed for mLIF by ELISA. a, c, e) Total amount of mLIF secreted. b, d, f) Percentage of mLIF secreted into the apical chamber. Values are expressed as mean +/-SEM (n = 3). ND (not detected at 1 μg/ml Dox). *P < 0.05; vs 1 μg/ml Dox. #P < 0.05; vs. 0.5 ng/ml Dox; ~P < 0.05; vs. 1 ng/ml Dox.In order to identify any differences in the pattern of secretion between the LIF-D and LIF-M signal peptides, two MDCK Tet-off-mLIF-M clones were also examined over a range of expression levels (Fig 3c–f). Without Dox, mLIF-M clone F produced 11 ng of mLIF (78% in the apical chamber), whereas with Dox this clone produced significantly less mLIF (less than 1 ng) and secreted less of it apically (71%). Without Dox, mLIF-M clone S produced 8 ng of mLIF (66% in the apical chamber), but with 1 μg/ml Dox it did not produce detectable amounts of mLIF. The experiments were repeated with less Dox (Fig. 3e and 3f). In these experiments, clone S secreted without Dox, 31 ng of mLIF (70% in the apical chamber). With 1 ng/ml Dox, clone S secreted significantly less mLIF (253 pg), but secreted a similar percentage apically (74%). Thus of the two clones one showed a difference in the proportion secreted apically at different levels of expression and one did not. Both clones secreted the majority of the mLIF in the apical direction, which is similar to the pattern seen with clones expressing mLIF-D. There is no evidence for a difference between the two signal peptides in directing the polarity of mLIF secretion.MDCK hLIF clones also preferentially secrete LIF from the apical membraneFour MDCK Tet-off hLIF-D clones were also tested for their pattern of secretion (Fig 4). Without Dox, hLIF-D clone 3 produced over 4 μg of hLIF of which 70% of the hLIF was secreted apically. With Dox this clone produced significantly less hLIF (10 ng) but secreted a similar percentage apically (69%). hLIF-D clones 30 and 16 behaved in a similar fashion. Clone 29 produced the least amount of LIF and showed little if any response to Dox. Without Dox, it produced 211 pg of hLIF of which 70% was in the apical chamber, whereas with Dox it produced 220 pg of which 73% was secreted apically. Of the four clones, only clone 29 showed a significant difference in apical secretion but this very small difference (3%) occurred in the absence of a significant change in expression levels of hLIF. Considering the clones together, over a range of concentrations from picograms to micrograms of LIF, the pattern of exogenous hLIF secretion is similar to that of exogenous mLIF in MDCK cells, with about 70% of the LIF secreted apically. There is no evidence for saturation of a sorting step in LIF transport.Figure 4MDCK Tet-off cells secreted exogenous hLIF-D predominantly in the apical direction across a range of expression levels. MDCK cells grown on permeable filters were incubated without doxycycline (No Dox) or with 1 μg/ml of doxycycline (Dox) to regulate the expression of hLIF-D. After 24 hours, conditioned media was collected from the apical and basal chambers and assayed for hLIF by ELISA. a) Total amount of hLIF secreted. b) Percentage of hLIF secreted into the apical chamber. Values are expressed as mean +/- SEM (n = 3). *P < 0.05 vs.1 μg/ml Dox.DiscussionThe expression of pro-inflammatory cytokines and their receptors by normal (unstimulated) intestinal epithelial cells points towards a role for these cytokines in the normal growth and maintenance of the gut epithelium. The secretion of LIF by unstimulated Caco-2 cells is consistent with the detection of LIF from human colonic epithelial cells [6] and the presence of LIF in mRNA obtained from murine intestinal tissue [18]. Guimbaud et al [19] reported that Caco-2 cells did not make LIF perhaps because they used a less sensitive ELISA. Whilst the human colonic epithelium has been reported to lack LIFR expression, pericryptal fibroblasts express the receptor, suggesting an interaction of LIF produced by colonic epithelial cells with pericryptal fibroblasts [6]. LIF levels are increased in inflammation, as demonstrated by high levels of LIF in patients with ulcerative colitis [19]. In contrast to LIF, we did not observe constitutive secretion of IL-6 from Caco-2. This may reflect different physiological roles for LIF and IL-6 in the normal intestinal epithelium.It is striking that the site of LIF and IL-6 secretion is influenced by the site of stimulation. Such preferential secretion displays a variety of features. It is observed with some cytokines and acute phase proteins but not all secreted proteins [10,20], with only some inducers [21], and in a variety of polarized cell types [21-24]. The mechanisms which underlie preferential secretion at the site of stimulation have not been defined and merit future study. Enhanced secretion at the site of stimulation could result from a global response such as an increase in the amount of membrane incorporated into the stimulated membrane or a change in membrane composition. If the mechanisms are global, however, there must remain transport routes which are not affected by these changes. For example, Moon et al. demonstrated in Caco-2 that although IL-6 displayed a marked preference for secretion through the membrane stimulated by IL-1β, the secretion of complement component C3 displayed strong basolateral secretion regardless of the membrane stimulated [10].Conditioned media from MDCK cells demonstrated LIF-like activity in BA/F3 mLIFR-mgp130. LIF expression by MDCK would not be surprising since collecting duct cells are known in rat to express LIFR and LIF. Following ischemia-reperfusion injury LIF is expressed more widely by the epithelial cells of the kidney, where it may participate in regeneration by stimulating the proliferation of epithelial cells [25,26]. LIFR is expressed at both the apical and basolateral membranes of MDCK [17]. The apparent selectivity for apical secretion of a LIF-like activity observed here may reflect paracrine signalling between epithelial cells and is broadly consistent with the apical secretion observed in the exogenous LIF-expressing clones. A sensitive ELISA for canine LIF is required for a more definitive comparison between endogenous and exogenous expression.We generated cell clones from MDCK Tet-off cells in which LIF expression was regulated by a tetracycline-inducible promoter. The clones demonstrated a wide range of LIF expression levels when the promoters were induced or repressed. Exogenous mLIF was mostly secreted in an apical direction irrespective of whether the signal peptide was from LIF-D or LIF-M. This suggests that the LIF-M signal peptide does not direct the secretion of LIF basolaterally toward the ECM in MDCK. We have not measured the binding capacity of the ECM for LIF in MDCK cells. If the binding capacity is similar to that observed in the osteosarcoma cell line UMR-106 [27], however, it would be easily saturated when LIF is highly expressed in the absence of doxycycline and is unlikely to have affected the measurement of polarity. Cell-type differences may be important in the recognition of sorting signals. When proteins that are polarized in epithelial cells are expressed in fibroblasts, they display specific secretion routes [28,29], however, it is not known if routes used by endogenous proteins in fibroblasts are present in epithelial cells. LIF may also require a specific \"sorting escort\" as has been observed for components of secretory granules [30].Interestingly, the level of LIF expression did not influence the proportion of LIF secreted apically and basolaterally. The high levels of LIF produced were comparable to those observed with VEGF when its expression level was manipulated in RPE-J cells [12], but there was no evidence in MDCK of a shift in LIF polarity with increasing expression comparable to the shift observed for VEGF (100% to 30% apical). This failure to observe saturation may reflect a lack of specific sorting receptors for LIF.The polarity observed with exogenous mLIF and hLIF (70% apical) in MDCK may simply reflect the total volume of vesicles delivered to the apical and basolateral surfaces. This hypothesis is attractive since passive incorporation of LIF into vesicles would account for the lack of saturation. To our knowledge total vesicular traffic has not been measured in MDCK so it is not known what percentage secreted apically would be consistent with passive transport. Comparisons with other four-helical cytokines, however, suggest that intrinsic features of LIF are likely to play a role even in passive transport. Rat growth hormone which is unglycosylated displayed only 35–40% secreted apically when expressed exogenously in MDCK [31,32], indicating differential sorting of LIF and growth hormone. LIF more closely resembles erythropoietin in its secretion pattern. Wild-type erythropoietin displayed predominantly apical secretion when exogenously expressed in MDCK. The percentage apical secretion dropped to about 50% after the removal of one of the N-glycosylation sites [33], suggesting that either the N-glycosylation serves as a sorting signal or it is required for the conformation of a proteinaceous signal. Further experiments could reveal if the glycosylation of LIF was similarly required for apical secretion. One way in which the glycosylation/conformation of LIF could be important is if it mediates a preferential affinity for the lipids that compose apical vesicles or tubules. Constitutive secretory proteins have been shown to be associated with the lumenal face of membranes in the secretory pathway [34]. If LIF can directly interact with lipid domains, it may explain why saturation was not detected for LIF. The high expression levels achieved with exogenous LIF will make experiments with confocal microscopy and detergent solubilization feasible.The apical secretion of exogenous LIF and an endogenous LIF-like activity in MDCK contrasts with the basolateral preference seen in Caco-2. LIF is not unusual in this regard. Differences in secretion patterns of transmembrane proteins between these two cell lines have been reported and attributed to differences in the recognition of basolateral trafficking signals [35]. We haven't determined whether the preferential secretion of LIF at the site of stimulation is unique to Caco-2 or shared with MDCK. Similarly we haven't determined whether the lack of a role for LIF signal peptides and the absence of saturation are unique to MDCK. For each of these findings, further experiments are required to test whether they extend to both epithelial cell lines.ConclusionThe results presented provide evidence for the secretion of LIF by intestinal epithelial cells and a LIF-like activity from kidney epithelial cells. Interesting questions have emerged as to how the site of stimulation is able to influence the pattern of protein secretion and how the secretion of exogenous LIF is directed to the apical and basolateral poles in the apparent absence of specific protein sorters. More mechanistic approaches are now required to pursue these questions.MethodsCell cultureCaco-2 cells were maintained in DMEM, high glucose, with L-glutamine (Gibco), 20% FCS (Labtech), 1% MEM (Gibco), 100 U/ml penicillin, and 100 μg/ml streptomycin (Gibco). BA/F3 mLIR-mgp130 cells were maintained in RPMI 1640 (Gibco), supplemented with 10% FCS, 50 U/ml penicillin, 50 μg/ml streptomycin, and recombinant human LIF (20 ng/ml). Parental BA/F3 cells were maintained in RPMI 1640 (Gibco), supplemented with 10% FCS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 100 pg/ml IL-3 (R & D systems). MDCK Tet-off cells (BD Biosciences) were maintained in DMEM high glucose containing 50 U/ml penicillin, and 50 μg/ml streptomycin and 1 μg/ml puromycin (Sigma). Polarized secretion experiments were performed with cells grown on 4.7 cm2, 0.4 μm pore size polycarbonate membrane Transwell® inserts (Costar). Caco-2 cells were seeded into Transwell® inserts at a density of 4 × 105 cells/well. Media in each chamber was replaced every two-three days for fourteen days, after which time monolayer integrity was demonstrated by measuring transepithelial resistance (TEER; described below). For stimulation experiments, monolayers were stimulated with 1 ng/ml IL-1β (R & D systems) for 24 hours. MDCK cells were seeded into Transwell® inserts at a density of 106 cells/well. Media in each chamber was replaced every 2–3 days for five days, after which time monolayer integrity was verified by measuring TEER.Epithelial monolayer integrityThe TEER of each monolayer was measured with an epithelial voltmeter using 'chopstick electrodes' (EVOM World Precision Instruments, USA). The mean TEER was calculated from three different positions. The intrinsic resistance of the insert (permeable support alone) was subtracted from the total resistance (cell monolayer and permeable support) to calculate the resistance of the monolayer. The resistance was corrected for surface area of the permeable support (4.7 cm2) and the TEER was expressed as Ohms cm2 (Ω.cm2). TEER was determined before the start and at the end of each period of collecting conditioned media.Proliferation assayActivity of conditioned media was assessed by a cell proliferation assay. BA/F3 mLIR-mgp130 is a pro-B-cell line that has been stably transfected with cDNA, encoding both components of the mLIF receptor: mLIFR and mgp130. The assay was performed as described previously [36]. Proliferation was measured using 1-(4,5-Dimethylthiazol-2-yl)-3,5-diphenylformazan (MTT; [37]). Neutralization of LIF-like activity was achieved by adding 10 μg/ml of rabbit anti-hLIF (Chemicon; AB1886) to the assay.ELISAsLIF and IL-6 concentrations in cell supernatants were determined by specific ELISA: hLIF, hIL-6 (Bender MedSystems), and mLIF (Quantikine ELISA; R & D Systems). In our experiments, the detection limits for eukaryotic hLIF and hIL-6 were 10 pg/ml. The Quantikine mLIF kit detected bacterial mLIF down to 20 pg/ml. The kit also detected purified glycosylated mLIF that we prepared in eukaryotic cells.Construction of LIF expression plasmidsmLIF cDNA encoding the LIF-D signal peptide was isolated from the plasmid pXMT2 mLIF-D [13] using the enzyme EcoR1. The fragment was subcloned into the plasmid PcDNA3.1- (Invitrogen) to create the plasmid PcDNA3-mLIF-D. Murine LIF cDNA encoding the LIF-M signal peptide was isolated from the plasmid pXMT2 mLIF-M [13] using EcoR1 and Xho1. The fragment was subcloned into the plasmid, PcDNA3.1 (-), to create the plasmid PcDNA3.1-mLIF-M. To ensure a comparable 5' end to the mLIF clones, hLIF-D cDNA was generated from a pre-existing PcDNA3.1-hLIF plasmid by PCR. The 5' oligonucleotide was directed to the 5' signal peptide and included a BamH1 site and the 3' oligonucleotide was directed to the 3' end of the LIF coding sequence and included an EcoR1 restriction site. The resulting fragment was subsequently ligated into PcDNA3.1 to create the plasmid PcDNA3.1-hLIF-D. cDNAs encoding the mLIF-D, mLIF-M and hLIF-D signal peptides were isolated from their respective PCDNA3.1 plasmids using the restriction enzymes Nhe1 and HindIII and were subsequently ligated into the plasmid pBI-L (BD Biosciences). All constructs generated by PCR were verified by automated DNA sequencing (Functional Genomics, University of Birmingham).Stable expression in MDCK Tet-OFF cellsMDCK Tet-Off cells (BD Biosciences) were transfected with PBI-L-mLIF-D, PBI-L-mLIF-M, and PBI-L hLIF-D using calcium phosphate. Stably transfected cells were selected in media containing 100 ng/ml hygromycin and 1 μg/ml doxycycline and cloned using cloning cylinders (Sigma). Cells expressing exogenous LIF were identified by luciferase activity transcribed from the bi-directional promoter of PBI-L and by LIF ELISA.Data AnalysisResults are presented as mean ± standard error of the mean (SEM) of three separate experiments in which duplicate wells were analyzed except where otherwise specified. For comparing total secretion in the presence and absence of doxycycline, unpaired t tests were performed. For comparing proportions, the data was first normalized by taking the arcsine of the square root of the proportion. When more than two results were compared, an analysis of variance (ANOVA) followed by the Tukeys test was performed.Authors' contributionsEJH and ABV both contributed to the experimental work and the drafting of the manuscript. Both authors read and approved the final manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2556395.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2556395",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2556395\nAUTHORS: Hege Langli Ersdal, Douwe A. A. Verkuyl, Kenneth Björklund, Staffan Bergström\n\nABSTRACT:\nBackgroundObstructed labour remains one of the leading causes of maternal and foetal death and morbidity in poorly resourced areas of the world, where the 24 hours availability of a caesarean section (CS) cannot be guaranteed, and the CS related mortality rate is still high. In these settings, reinstatement of symphysiotomy has been advocated. The objectives were, in1994; to study perioperative and long-term complications of symphysiotomy and compare them to those related to CS for similar indications, in 1996; to measure the symphyseal width after symphysiotomy and compare it to that after normal vaginal delivery, and, in 1998; to assess knowledge, attitudes and practice related to symphysiotomy among doctors and midwives in Zimbabwe.Methods and FindingsThirty-four women who had undergone symphysiotomy and 29 women who had undergone a CS for obstructed labour were interviewed. The symphyseal widths of 19 women with a previous symphysiotomy were compared to that of 92 women with previous normal vaginal deliveries, using ultrasound technique.Forty-one doctors and 39 midwives, in three central hospitals and seven district hospitals in Zimbabwe, were interviewed about symphysiotomy. None of the 34 women reported serious soft tissue injuries or infections post symphysiotomy. Long-term complications after symphysiotomy do not differ notably from those after CS for similar indications. The intra-articular width of the symphysis pubis is increased after a symphysiotomy. Seventy-nine of the 80 interviewed health care workers knew about symphysiotomy. One obstetrician had performed symphysiotomies. Two-thirds of the participants considered symphysiotomy an obsolete and second-class operation, but lifesaving and appropriate in remote areas of Zimbabwe. Ten of 13 midwives in remote areas wanted to carry out symphysiotomies themselves.ConclusionsNo severe complications due to symphysiotomy were revealed in this study. The results suggest that a modest permanent enlargement of the pelvis post symphysiotomy (together with the absence of a scarred uterus) may facilitate subsequent vaginal delivery. Doctors and midwives working in district hospitals have a more positive attitude to symphysiotomies than the colleagues in central hospitals. Obstetricians (who would have to do the teaching), working in the large urban hospitals almost exclude symphysiotomy as an alternative management in Zimbabwe.\n\nBODY:\nIntroductionSymphysiotomy was developed in Europe late in the 18th century as a means to save the lives of mother and child in obstructed labour. In the early days the separation of the bones was performed after exposing the symphysis. With the introduction of aseptic techniques in the mid 19th century and the subcutaneous symphysiotomy by Frank [1], using local anaesthetics, the operation gained acceptance and became part of the obstetric arsenal in many parts of Europe [2]–[5]. The technique was further refined by Zárate in Argentina [6], and symphysiotomy was extensively practised in Latin America and Ireland up to the mid 20th century [4], [7], [8]. Seedat and Crichton in Durban, South Africa, further developed the technique [9] and symphysiotomy came to be used in Africa for the ensuing decades [9]–[12]. Currently, outside Africa, symphysiotomies are still performed in Papua New Guinea [13].Mechanical obstruction in the second stage complicates 1–2% of labours [14]. In developing countries this percentage might be higher because women often do not attain their (pelvic) growth potential due to stunting. The facilities and staffing level of non-private hospitals in developing countries make it often inevitable that CSs are associated with an appreciable mortality rate. In addition, there is no guarantee that a woman with a scar in the uterus can, or even would like to, reach in time a (still adequately staffed) hospital for the next delivery. The World Health Organization (WHO) estimates annual global obstructed labour related maternal mortality at 50,000 [15]. On top of that, prolonged labours frequently cause life long handicaps to mother and child. According to the United Nations Population Fund (UNFPA) there are 2 million women living with a vesico-vaginal fistula caused by obstructed labour (http://www.endfistula.com/fast_facts.htm). The woman in obstructed labour should not be abandoned in her plight.Symphysiotomy is a potential solution in selected cases of obstructed labour with mild to moderate cephalopelvic disproportion [12], [16]–[18]. A review of 5,000 symphysiotomies performed during the twentieth century concluded that: “If valid conclusions can be drawn from one hundred years of retrospective studies, there is considerable evidence to support a reinstatement of symphysiotomy in the obstetric arsenal, for the benefit of women in obstructed labour and their offspring” [16]. A theoretical model shows that under basic circumstances only a few symphysiotomies are needed to prevent one maternal death [19]. Still, symphysiotomies are rarely performed systematically, with notable exceptions [17], [20]. In contrast, Onah demonstrated a dislike for CS among pregnant Nigerian women [21], and a preference for symphysiotomy above CS, if given the choice [22].In view of the ongoing controversy we found reasons to contribute to the discussion with a rare long term case control follow up study. At the same time we evaluate for the first time symphyseal width after symphysiotomy and after normal vaginal delivery with ultrasound technique, and knowledge, attitudes and practice related to symphysiotomy among doctors and midwives in Zimbabwe.MethodsThe present paper builds on two follow-up studies post symphysiotomy performed in Bulawayo, Zimbabwe in1994 and 1996 and one study in 1998 on knowledge, attitudes and practice among doctors and midwives at three levels of the health system in Zimbabwe; the academic hospital in Harare (Harare Central Hospital), two central hospitals in Bulawayo (United Bulawayo Hospital and Mpilo Central Hospital) and a convenience sample of seven district hospitals (Beitbridge, Lupane, Esigodini, Gwanda, Hwange, Victoria Falls and Kariba).Follow-up post symphysiotomy, studies in 1994 and 1996Permission to conduct the interviews and ultrasound examinations, including ethical clearance, was obtained from the medical superintendents at United Bulawayo Hospitals (UBH) and Mpilo Hospital. All women were informed about the research project (in English or a local language) when seen by the first author (HLE) together with a local interpreter or an ultrasound technician. Since these follow-up studies were conducted among a partly illiterate population the women gave explicit verbal consent to involvement. Nobody refused. The interpreter or the ultrasound technician bore witness to the oral consent process.In 1994, 18 women who had undergone a symphysiotomy previously (S group) and 29 women with previous CS (CS group) were interviewed.The second author (DAAV) had made a serious effort to record basic data and the addresses of all the women who had a symphysiotomy in the UBH and the referring Matabeleland South Province district hospitals during 1990–1994. Also women who had a symphysiotomy performed even earlier and who were seen incidentally at the antenatal clinic or in labour were added to the list. Altogether 61 women were registered. Symphysiotomies had mostly been performed by consultants in UBH, and by experienced district doctors who considered symphysiotomy to be the best option in each of the cases.The CS group consisted of women having undergone caesarean delivery for a probable pelvic outlet obstruction. They were recorded in the register as having had a CS after a failed vacuum extraction. It was not possible to find enough time-matched failed vacuum extractions followed by CS in the registers of the UBH, therefore we included the delivery records of the other tertiary hospital in Bulawayo, Mpilo Hospital, which caters to similar women. Seventy women were selected in this way and we tried to match the time since the index delivery as much as possible.Serious efforts were made to visit women in both groups. Fourteen women in the S group and 24 women in the CS group were found at home. Letters written in local languages and English, promising a check up and expressing commitment to pay travelling costs and compensation (for themselves and necessary accompanying person), were sent to the remainder. After this, four and five women, respectively, appeared.The first author (HLE) prepared the interview in cooperation with DAAV. HLE conducted all the interviews and took notes, assisted by a single interpreter speaking fluent English, and the local languages.The interview was divided in four sections. Firstly, the baseline data (age, parity before and after index intervention, and years since index intervention) were confirmed and registered. Secondly, the woman was asked about the child's condition when born and subjectively experienced perioperative complications (pain, fistulae, laceration, hemorrhage or infection). Thirdly, she was questioned about specific current complaints (pain when walking, dancing, jumping or carrying, painful scar, dyspareunia, infertility, and incontinence), and finally, the women were asked about any subsequent pregnancies and deliveries.In 1996, 19 women who had undergone a symphysiotomy previously (S group) and 92 women with only previous normal vaginal deliveries (NVD group) were interviewed and the symphyseal joints were examined by ultrasound at UBH. Ethical clearance for this study was also given in advance by The National Committees for Research Ethics in Norway.The register of women who had undergone symphysiotomy in the area around Bulawayo had in the mean time expanded to 70. Letters were written, inviting them to UBH for a checkup. Nineteen women appeared representing the S group, three of whom had also been interviewed in 1994. Two of these 19 women were pregnant (both 8 months). Eleven women lived in and around Bulawayo, four lived 20–100 kilometres outside the city, and four lived over 100 kilometres away.At the Department of Obstetrics and Gynaecology in UBH, 92 women with previous normal vaginal deliveries, referred for ultrasonography for different reasons, were during one month consecutively invited to have their symphyseal width recorded. Further inclusion criteria were parity 1–8, having delivered exclusively vaginally, age between 15 and 45 years, and no orthopaedic disorders. All the selected women were informed about the study, and all of them accepted to participate. Thirty-four of the women included were pregnant.Baseline data for both groups were registered and women in the S group were interviewed by HLE using the same semi-structured interview as in 1994. There was always an ultrasound technician present who spoke the local languages and English. Afterwards, women in both groups were sonographically examined by HLE with a Philips Sterling 1993 model, utilizing 5 mega Hertz convex sector probe to measure the internal symphyseal gap. Björklund et al had concluded in 1996 [23] that estimating the width of the symphysis pubis by ultrasonography offers at least the same precision as radiography. The image of the symphysis is seen as two curved lines representing the anterior surfaces of the pubic bones, turning parallel into the symphyseal gap, where the width may be measured. Measuring the width may sometimes be difficult due to the shape of the gap. The opposing edges of the gap might be highly irregular and there is a widening mostly upwards, but also occasionally downwards. The measuring point was defined as the narrowest part of the symphyseal joint gap in the upper half of the symphysis. The women were in a supine position with extended hips while the measurements were taken.For statistic calculation Student's t-test was used.Knowledge, attitudes and practice among doctors and midwives, study in 1998The interviews were not announced beforehand. All relevant doctors and midwives present at the respective maternity units at the time of the visit were invited to participate, and everyone accepted. Permission to conduct the interviews, including ethical clearance, was obtained from the medical superintendents at each location.The first and second authors (HLE and DAAV) prepared the questions in cooperation with a local general practitioner in Bulawayo. The interview consisted of two parts. Firstly, the participant was asked about her/his practice and if he/she could describe the procedure. Secondly, the interview included a series of statements on symphysiotomy. The person interviewed was asked to react affirmative or negative to closed questions. HLE conducted all the interviews.In the data analysis the interview locations were divided into urban and rural, comparing the responses of specialists in obstetrics and gynaecology, junior doctors and district doctors and different categories of midwives.ResultsFollow-up post symphysiotomy, studies in 1994 and 1996Baseline dataThe baseline data from the three groups (S, CS, and NVD) are presented in Table 1. Twelve women in the S group had delivered before their symphysiotomy, five with a CS. Women in the NVD group had no abdominal deliveries, and their average parity when seen was 2.6.10.1371/journal.pone.0003317.t001Table 1Baseline data from the three groups; CS, S, and NVD.CS 1994S 1994+1996NVD 1996n = 29n = 34n = 92AverageRangeAverageRangeAverageRangeMaternal age26,815–4026,118–4231,119–45Parity before index operation10–50,80–6Interval since operation, years2,90–1540–15Total number of deliveries at follow up2,31–62,51–82,61–8Reported perioperative complications after symphysiotomyThere was one neonatal death during the hospital admission associated with a symphysiotomy among the 70 deliveries on record.None of the 34 interviewed women in the S group reported serious soft tissue injuries in the birth canal, e.g. laceration, fistulae, and hemorrhage, or post-operative infection. One woman delivered a baby in a district hospital, the neonatal death, with an undiagnosed hydrocephalus and experienced a too wide separation of the symphysis resulting in damage to the tissues supporting the urethra/bladder neck. She suffered from stress incontinence as soon as the urinary catheter had been removed. Thirteen (38%) of the women did not remember their symphysiotomy delivery as being particularly painful.Current complaints\nTable 2 shows a comparison of current complaints in the S and CS groups.10.1371/journal.pone.0003317.t002Table 2Comparison of current complaints after CS and after S.CSSPn = 29n = 34%%Pain when walking*\n13,823,50,33Pain when dancing6,92,90,52Pain when jumping6,98,80,78Pain when carrying6,98,80,84Painful scar51,72,9<0,01Dyspareunia17,229,4**\n0,41Infertility6,900,12Incontinence6,92,90,47*CS 4/4 pain when walking any distance.S 1/8 pain when walking any distance, 7/8 pain when walking long distance.**10/10 some pain over symphysis pubis when abduction of the legs.The outcomes are very similar for the two interventions, except for scar pain (p<0.01).Eight women (23.5%) in the S group reported that they felt pain when walking; seven of them had tenderness or pain over the symphysis pubis only after walking 10–20 km. The woman who delivered a baby with hydrocephalus reported pain in the symphysis pubis and a feeling of instability in the pelvis when walking any distance. She was the only woman in the S group with urinary stress incontinence at follow up. She could presently cope with her incontinence if she remembered to void frequently. She managed her daily duties, but was not able to walk far from the village. None of the other women in the S group experienced instability in the symphysis pubis or the pelvic region.Half of the women in the CS group experienced pain in the scar. They also mentioned itching, pins and needles. Five women in the CS group reported pain deep in vagina during coitus, none of them having had this problem before the CS. Ten women in the S group reported some pain over the symphysis pubis with abduction of the legs during coitus, but none experienced pain in vagina. The women were not asked about possible psychosocial complications arising from this.Width of symphysis pubis and subsequent deliveries\nTable 3 shows the width of the symphysis pubis in women post symphysiotomy and in women with only previous normal vaginal deliveries. The average symphyseal width after symphysiotomy was 9.2 mm. Among women with previous normal vaginal deliveries the average distance was 4.7 mm (p<0.01). The outcome of subsequent deliveries after CS and after symphysiotomy is presented in table 4.10.1371/journal.pone.0003317.t003Table 3The width of symphysis pubis after S and after NVD.SNVDPnAverageRangenAverageRangemmmmmmmmNot pregnant179,22,0–20,4584,71,9–7,8<0,01Pregnant213,912,8–15,0345,82,9–9,7<0,0110.1371/journal.pone.0003317.t004Table 4Outcome of subsequent deliveries after CS and S.CSSn%n%Total number of women2934No of women with subsequent delivery827,61647,1Total number of deliveries1022Normal vaginal deliveries6601986,4Caesarean section440313,6Knowledge, attitudes and practice among doctors and midwives, study in 1998Eighty doctors and midwives were invited to participate in this study, and all accepted. The interviews in the urban settings comprised 34 doctors and 26 midwives, and in the rural settings 7 doctors and 13 midwives. Thus a total of 41 doctors and 39 midwives were interviewed. Of the doctors in the urban settings, 10 were obstetricians and 24 junior doctors. In the district hospitals, all the 7 doctors interviewed were non-specialist. Of the midwives in the urban settings, 13 were clinical instructors, 10 qualified midwives and 3 student midwives. In the rural settings, 4 were clinical instructors, 7 were qualified midwives and 2 were student midwives.Seventy-nine of the 80 participants knew about symphysiotomy, and 76 could describe the technique, including 16 of the 17 midwife instructors. One junior doctor was not aware of the intervention. One of the ten obstetricians had occasionally performed a symphysiotomy the other nine did not practice the intervention, but indicated that they would be able to carry it out. Among the 24 junior doctors in the urban settings, six wanted to learn the skill. One of the seven district doctors thought he could do a symphysiotomy if need be, another two wanted to learn the procedure.\nTable 5 shows the responses of the doctors to a series of statements on symphysiotomy, comparing the opinions of obstetricians and junior doctors working in the large cities with those of district doctors. The latter have more positive attitudes to symphysiotomy. Table 6 compares the responses of doctors and midwives in the urban and rural settings, respectively, to the same series of statements as those of Table 5. All the rural midwives (n = 13) regarded symphysiotomy as a lifesaving operation appropriate for remote areas, and 23 of the 39 midwives (59%) thought that the procedure should be taught to midwives. One third of the doctors and two thirds of the midwives thought that symphysiotomy should be performed, if only for teaching purposes, in central hospitals.10.1371/journal.pone.0003317.t005Table 5Affirmative answers on statements about symphysiotomy among obstetricians and junior doctors in three hospitals in the two largest cities of Zimbabwe; Harare and Bulawayo, and among district doctors at seven district hospitals.UrbanRuralObstetriciansJunior doctorsDistrict doctorsn = 10 (%)n = 24 (%)n = 7 (%)Symphysiotomyis a harmful operation6 (60)16 (67)2 (28)is obsolete and second-class8 (80)16 (67)4 (57)can be lifesaving3 (30)12 (50)5 (71)is appropriate in very* remote areas5 (50)9 (37)5 (71)Symphysiotomy hasnegligible maternal mortality8 (80)18 (75)6 (86)high maternal morbidity8 (80)15 (63)2 (28)uneventful subsequent deliveries7 (70)5 (21)2 (28)Symphysiotomy should beperformed in central hospitals2 (20)9 (37)3 (43)taught to district doctors2 (20)9 (37)4 (57)taught to midwives2 (20)5 (21)2 (28)*More remote areas than existing in Zimbabwe1998.10.1371/journal.pone.0003317.t006Table 6Affirmative answers on statements about symphysiotomy among doctors and midwives in three hospitals in the two largest cities of Zimbabwe (Harare and Bulawayo), and among doctors and midwives at seven district hospitals.UrbanRuralDoctorsMidwivesDoctorsMidwivesn = 34 (%)n = 26 (%)n = 7 (%)n = 13 (%)Symphysiotomyis a harmful operation22 (65)11 (42)2 (29)4 (31)is obsolete and second-class24 (71)17 (65)4 (57)9 (69)can be lifesaving14 (41)20 (77)5 (71)13 (100)is appropriate in remote areas14 (41)20 (77)5 (71)13 (100)Symphysiotomy is associated withnegligible maternal mortality26 (76)10 (38)6 (86)4 (31)high maternal morbidity23 (68)11 (42)2 (29)4 (31)uneventful subsequent deliveries12 (35)12 (46)2 (29)10 (77)Symphysiotomy should beperformed in central hospitals11 (32)15 (58)3 (43)9 (69)taught to district doctors11 (32)20 (77)4 (57)11 (85)taught to midwives7 (21)13 (50)2 (29)10 (77)Eight of the ten obstetricians stated that their opinions on symphysiotomy were based on what they were taught to believe in medical school or during specialist training, and that they had not been influenced by papers on symphysiotomy they had read later in their careers. Three of the ten obstetricians thought that randomised controlled trials of symphysiotomies were needed, whereas the other seven saw no need for further research into a procedure they regarded as outdated. Ninety percent of the junior doctors indicated that their opinion on symphysiotomy was based on what they were taught in medical school. One of the 39 midwives was aware of scientific studies on symphysiotomy, the rest had never heard of any such studies. Fifty-one percent of the doctors spontaneously commented that symphysiotomy had been abandoned in Western countries, and considered this fact an argument against the use of the intervention in Zimbabwe.DiscussionThere are no reports of maternal mortality directly related to symphysiotomy from the antibiotic era [16]. There are still large regional differences in CS-related mortality. A study from the adjacent Matabeleland-South province of 1,128 CSs performed in 7 district hospitals 1998–2000 (covering a period when the health/transport infrastructure was still good) revealed a CS-related mortality of 1.6% [24]. In circumstances where CSs are not very dangerous and where women have few children (hence few deliveries after a previous CS) the advantages of symphysiotomies are small and furthermore doctors cannot obtain or maintain the expertise needed [19], but symphysiotomies are still sometimes the best option even in this situation [25]–[29]. Verkuyl calculated that in breech presentations in district hospitals in sub-Saharan Africa a policy of replacing elective CS with trial of labour combined with a symphysiotomy if the foetal head became stuck, would, in nullipara, prevent one maternal death for two symphysiotomies performed [19]. Replacing emergency CS for failed vacuum extraction with a symphysiotomy was claimed in this paper to have a similar beneficial effect on maternal mortality [19].Hill et al estimated 535,900 maternal deaths worldwide in 2005 [30], of which 50% were concentrated in sub-Saharan Africa and there was no significant reduction in maternal mortality ratios between 1990 and 2005 in this area. For all other countries with data there was a decrease in maternal mortality ratios of 2.5% per year in the same time period [30]. Symphysiotomy is an alternative management in cases of mild to moderate cephalopelvic disproportion, live foetus and longitudinal lie. The British Journal of Obstetrics and Gynaecology proclaimed, in an editorial (2002), the rebirth of symphysiotomy to reduce maternal and foetal mortality and morbidity [31].According to the literature symphysiotomy is associated with immediate post-operative pain and discomfort [4], [10], [12], [13], [16]–[18], [32]–[38], and it is a painful procedure when insufficient local anaesthesia is given. All women will experience pain at the moment of delivery, and separating the different origins of pain in labour might be difficult, also because an episiotomy and mostly a vacuum extraction is part of the symphysiotomy procedure.The results from the present study and previous follow-up studies indicate that symphysiotomy confers an acceptable level of complaints in the long run [10], [16], [32], [36], [38]. Lasbrey assessed pain in the symphysis pubis, groin, hip, thigh, sacro-iliac joint and stress incontinence, mostly associated with physical activity, and found that these problems were small and did not interfere with day-to-day activities [36]. The prevalence of these symptoms at some time during the follow-up period or in a subsequent pregnancy was 58% in the symphysiotomy group and 60% in parous women with unassisted vaginal deliveries. Hartfield compared the outcomes of symphysiotomy and CS (for cephalopelvic disproportion) including a long-term follow-up study [10]. Prevalence figures of reported sub fertility (7%), stress incontinence (3%), and backache (25%) in the symphysiotomy group were similar to corresponding figures in the CS group. Infertility after symphysiotomy was not addressed in the present study, and the woman who reported incontinence was apparently subject to an incorrect judgement of the degree of cephalopelvic disproportion. Pain or tenderness over the symphysis pubis when walking long distances was observed in this study, but not reported nor mentioned as a parameter in previous follow-up studies [4], [10], [13], [32]–[38]. Dyspareunia was registered in 17,2% (CS group) and 29,4% (S group) of the women in this study, unfortunately women with normal vaginal deliveries were not asked about this. Overall, prevalence of maternal complaints after symphysiotomy and CS do not differ much, though they are somewhat different in nature.The study performed in 1996 indicates that the width of symphysis pubis is permanently increased after a symphysiotomy, potentially facilitating future vaginal deliveries. Over three-quarters of the women with previous symphysiotomy that had a subsequent delivery had given birth vaginally without complications, in agreement with previous observations [3]–[5], [20], [25], [27]. This is an important advantage for a young woman with a borderline pelvis and several pregnancies ahead, living in a rural area with a tradition to deliver at home far from a well equipped hospital.The studied S group, including almost 50% of the women recorded after a symphysiotomy, might be a biased sample. Maybe women with severe complications were kept away, or the other way around, women without complaints had little to gain from leaving home for 1–2 days. Health services were free and well organized at that time, so it is likely to believe that the network of clinics would have picked up a woman with severe handicap and referred her. All the women in the record were operated in Bulawayo or an adjacent mission hospital, and the registered addresses were in the same region. Both postal service and bus transport system were reliable in 1994 and 1996.Doctors and midwives, working in Zimbabwe in 1998, knew in theory what a symphysiotomy is, and the majority of the participants in this study considered the operation lifesaving and appropriate in remote areas. Knowledge, attitudes and practice may have changed with time, but the results from 1998 indicate that district doctors and midwives have a more positive attitude to symphysiotomy than colleagues working in the cities. Most positive are the midwives working in rural areas. They have to manage every day with poor medical facilities and limited resources, including transport. These midwives often see mothers and their babies in bad condition due to prolonged obstructed labour, and frequently have to handle the situation alone. Ten of 13 district midwives wanted to learn how to perform a symphysiotomy themselves, in agreement with an editorial in the British Journal of Obstetrics and Gynaecology of 2002: “Symphysiotomy is simple. ….It requires only local anaesthesia, a catheter, a scalpel and a pair of gloves. It can be performed safely and successfully by individuals who are not obstetricians” [31]. The technique utilized does not require special surgical skills, and experience indicates that symphysiotomy may be carried out by a well-trained midwife if training is provided by an experienced obstetrician [24], [32].On the other hand, the obstetricians are the most reluctant subgroup, 90% had never done a symphysiotomy, not even on a model like their colleagues in the United Kingdom have to do as part of their specialist training, to be able to handle emergency situations. The obstetricians are responsible for the education of new doctors and they will also influence the teaching of student midwives. Their lectures are of great importance because only a few of the participants in this study had access or paid attention to the scientific literature on symphysiotomy.The present situation in Zimbabwe (like in several other countries) demonstrates another important factor related to symphysiotomy. In the wake of the economical collapse, a previously well organised health service struggles to maintain standards. Midwives and doctors emigrate. Transferring patients is often not possible because fuel is unobtainable. Women with a scar of previous CS(s) are in mortal danger when pregnant. Some of them could have been delivered in the past by a symphysiotomy, and that would have prevented the scar and made her pelvis somewhat larger.ConclusionThe frequency and severity of long term complications post symphysiotomy do not differ from those after CS. The permanent increase of the width of the symphysis pubis, measured in this study, seems to facilitate subsequent vaginal deliveries. Therefore, the benefit of symphysiotomy in cases of mild to moderate cephalopelvic disproportion is positively correlated to the local risks of a CS (nearly always the alternative course of action), the number of deliveries expected in the future, and the chance that subsequent deliveries would not happen in a well equipped and staffed hospital. This study indicates that almost 50% of doctors and 85% of midwives working in a developing country agree with this statement.\n\nREFERENCES:\n1. FrankF\n1910\nÜber den subkutanen Symphysenschnitt und die suprasymphysäre Entbindung.\nMonatschr Geburtshilfe Gynäk\n22\n680\n693\n2. KupferbergH\n1927\nZur Therapie Beim Engen Becken.\nMonatsschr Gynäk Geburtshilfe\n76\n176\n183\n3. PeytavinC\n(Paris 1931) De la symphysiotomie partielle sous-cutanée.\nÉtude critique de 45 observations de symphysiotomie suivant le procédé de Zarate [dissertation]\n4. BarryAP\n1952\nSymphysiotomy or pubiotomy. Why? When? And how?\nIr Med J\n6\n49\n73\n5. DexeusSSalarichN\n1954\nDe la sinfisiotomía complementaria de recurso o de emergencia en el curso de la extracción fetal por las vías naturals.\nRev Esp Obstet Ginecol\n13\n358\n366\n14372370\n6. ZárateE\n1931\nDie Záratesch Technik der partiellen Symphyseotomie. Ihre Indikationen und ihr therapeutischer Wert bei patologischen Geburten nach Beobachtungen von 100 Fällen.\nArch Gynäkol\n147\n759\n767\n7. Ramírez OlivellaJOrtiz-PerezJ\n1935\nDoce años de sinfisiotomia subcutanea.\nArch Med Int\n1\n254\n273\n8. Peralta RamosA\n1948\nLa sinfisiotomía en la era antibiotica.\nObstet Ginec Lat-Amer\n6\n3\n12\n18865870\n9. SeedatEKCrichtonD\n1962\nSymphysiotomy: technique, indications and limitations.\nLancet\ni\n554\n10. HartfieldVJ\n1973\nA comparison of the early and late effects of subcutaneous symphysiotomy and of lower segment Caesarean section.\nJ Obstet Gynaecol Br Commonw\n80\n505\n514\n11. GebbieDAM\n1982\nSymphysiotomy.\nClin Obstet Gynecol\n9\n663\n683\n12. van RoosmalenJ\n1987\nSymphysiotomy as an alternative to cesarean section.\nInt J Gynaecol Obstet\n25\n451\n458\n2892703\n13. MolaGDL\n1995\nSymphysiotomy or cesarean section after failed trail of assisted delivery.\nPNG Med J\n38\n172\n177\n14. AbouzahrC\n1998\nProlonged and obstructed labour.\nMurrayCJLLopezAD\nHealth dimensions of Sex and reproduction\nGeneva\nWHO\n243\n266\n15. WHO\n1999\nReduction of maternal mortality. A joint WHO/UNFPA/UNICEF/World Bank Statement\nGeneva\nWHO\n16. BjörklundK\n2002\nMinimally invasive surgery for obstructed labour: a review of symphysiotomy during the twentieth century (including 5000 cases].\nBJOG\n109\n236\n248\n11950177\n17. Sunday-AdeoyeIMOkontaPTwomeyD\n2004\nSymphysiotomy at the Mater Misericordiae Hospital Afikpo, Ebonyi State of Nigeria (1982–1999]: A review of 1013 cases.\nJ Obst Gyn\n24\n525\n529\n18. van RoosmalenJ\n1990\nSafe motherhood: Caesarean section or symphysiotomy?\nAm J Obstet Gynecol\n163\n1\n4\n2375330\n19. VerkuylDAA\n2007\nThink globally act locally: the case for symphysiotomy.\nPlos Med\n4(3)\ne71\ndoi:10.1371/journal.pmed.0040071 www.plosmedicine.com\n17388656\n20. EzegwuiHUOlomuOOTwomeyED\n2004\nSymphysiotomy in a Nigerian hospital.\nInt J Gynaecol Obstet\n85\n272\n273\n15145265\n21. OnahHE\n2002\nFormal education does not improve the acceptance of cesarean section among pregnant Nigerian women.\nInt J Gynecol Obstet\n76\n321\n323\n22. OnahHEUgonaMC\n2004\nPreferences for cesarean section or symphysiotomy for obstructed labor among Nigerian women.\nInt J Gynecol Obstet\n84\n79\n81\n23. BjörklundKBergströmSLindgrenPGUlmstenU\n1996\nUltrasonographic measurment of the Symphysis pubis: A potential method of studying symphyseolysis in pregnancy.\nGynecol Obstet Invest\n42\n151\n153\n8938463\n24. van EygenLRutgersRAK\n2008\nCaesarean section as preferred mode of delivery in term breech presentations is not a realistic option in rural Zimbabwe.\nTrop Doc\n37\n1\n4\n25. WykesCBJohnstonTAPaterson-BrownSJohansonRB\n2003\nSymphysiotomy: a lifesaving procedure.\nBJOG\n110\n219\n221\n12618172\n26. GoodwinTMBanksEMillarLPhelanJ\n1997\nCatastrophic shoulder dystocia and emergency symphysiotomy.\nAm J Obstet Gynecol\n177\n463\n464\n9290471\n27. GlasterEVialYHohlfeldP\n1999\nCurrent aspects of symphysiotomy. Apropos of 1 case and review of the literature.\nJ Gynecol Obstet Biol Reprod (Paris)\n28\n519\n28\n10598344\n28. SpencerJAD\n1987\nSymphysiotomy for vaginal breech delivery: two case reports.\nBr J Obstet Gynaecol\n94\n716\n718\n3620418\n29. MenticoglouSM\n1990\nSymphysiotomy for the trapped aftercoming parts of the breech: a review of the literature and a plea for its use.\nAust NZJ Obstet Gynaecol\n30\n1\n9\n30. HillKThomasKAbouZahrCWalkerNSayLInoueMSuzukiE\n2007\nEstimates of maternal mortality worldwide between 1990 and 2005: an assessment of available data.\nLancet\n370\n1311\n1319\n17933645\n31. Editorial\n2002\nThe rebirth of symphysiotomy.\nBJOG\n109\n32. MolaMLamangMMcGoldrickIA\n1981\nA retrospective study of matched symphysiotomies and Caesarean sections at Port Moresby General Hospital.\nP.N.G. Med J\n24\n103\n112\n6950596\n33. PapeGL\n1999\n27 symphysiotomies.\nTrop Doct\n29\n248\n249\n10578647\n34. GreisenG\n1999\nThree-year follow-up of 8 patients delivered by symphysiotomy.\nInt J Gynaecol Obstet\n23\n203\n205\n35. HartfieldVJ\n1975\nLate effects of symphysiotomy.\nTrop Doct\n5\n76\n78\n124103\n36. LasbreyAH\n1963\nThe symptomatic sequelae of symphysiotomy. A follow up study of 100 patients subjected to symphysiotomy.\nS Afr Med J\n37\n231\n234\n37. BirdGCBalJS\n1967\nThe vacuum extractor in association with symphysiotomy at a provincial hospital in Kenya.\nE Afr Med J\n43\n51\n56\n38. BergströmSLublinHMolinA\n1994\nValue of symphysiotomy in obstructed labour management and follow up of 31 cases.\nGynecol Obstet Invest\n38\n31\n35\n7959323"
+}

batch_13/PMC2556659.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2556659",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2556659\nAUTHORS: Sulayman D Dib-Hajj, Mark Estacion, Brian W Jarecki, Lynda Tyrrell, Tanya Z Fischer, Mark Lawden, Theodore R Cummins, Stephen G Waxman\n\nABSTRACT:\nBackgroundParoxysmal extreme pain disorder (PEPD) is an autosomal dominant painful neuropathy with many, but not all, cases linked to gain-of-function mutations in SCN9A which encodes voltage-gated sodium channel Nav1.7. Severe pain episodes and skin flushing start in infancy and are induced by perianal probing or bowl movement, and pain progresses to ocular and mandibular areas with age. Carbamazepine has been effective in relieving symptoms, while other drugs including other anti-epileptics are less effective.ResultsSequencing of SCN9A coding exons from an English patient, diagnosed with PEPD, has identified a methionine 1627 to lysine (M1627K) substitution in the linker joining segments S4 and S5 in domain IV. We confirm that M1627K depolarizes the voltage-dependence of fast-inactivation without substantially altering activation or slow-inactivation, and inactivates from the open state with slower kinetics. We show here that M1627K does not alter development of closed-state inactivation, and that M1627K channels recover from fast-inactivation faster than wild type channels, and produce larger currents in response to a slow ramp stimulus. Using current-clamp recordings, we also show that the M1627K mutant channel reduces the threshold for single action potentials in DRG neurons and increases the number of action potentials in response to graded stimuli.ConclusionM1627K mutation was previously identified in a sporadic case of PEPD from France, and we now report it in an English family. We confirm the initial characterization of mutant M1627K effect on fast-inactivation of Nav1.7 and extend the analysis to other gating properties of the channel. We also show that M1627K mutant channels render DRG neurons hyperexcitable. Our new data provide a link between altered channel biophysics and pain in PEPD patients.\n\nBODY:\nBackgroundRecent genetic studies have identified sodium channel Nav1.7 as a major contributor to pain [1,2]. Nav1.7 is predominantly expressed in dorsal root ganglion (DRG) and sympathetic ganglion neurons [3-6], and specifically in most functionally-identified DRG nociceptive neurons [7]. Global knock-out of Nav1.7 in mice is neonatal lethal while nociceptor-specific knock-out results in elevated mechanical and thermal pain thresholds and an attenuated pain response [8]. Surprisingly, congenital loss of Nav1.7 in humans is not associated with cognitive or motor deficits, but causes complete indifference to pain [9-11]. In contrast, one set of gain-of-function mutations of Nav1.7 in inherited erythromelalgia (IEM) leads to severe episodes of pain, mostly in the feet and hands [12-18], and a different set of mutations results in paroxysmal extreme pain disorder (PEPD) [19]. Because some cases of PEPD do not carry mutations in Nav1.7 [19], it is important to genetically profile new PEPD cases to determine the molecular basis of the disease which may influence possible therapy options.Severe pain in PEPD patients accompanied by redness in the lower body can start in infancy, is induced by bowel movement or probing of perianal areas, and can be accompanied by tonic nonepileptic seizures, syncopes, bradycardia and occasionally asystole [20]. The cause for the seizures and cardiac symptoms is not well understood. Pain progresses with age to affect ocular and maxillary/mandibular areas and is triggered by cold, eating or emotional state [20]. Pain episodes can last seconds to minutes (and hours in extreme cases), and gradually subside.Whole-cell voltage-clamp studies have shown that all of the IEM mutations lower the voltage threshold for Nav1.7 activation, with most mutations increasing the ramp response of the channel [13,16,17,21-25]. In contrast, PEPD mutations in Nav1.7 impair fast-inactivation and can result in a persistent current [19]. The M1627K mutation, which is located in domain IV S4–S5 linker (DIV/S4–S5) and first identified in a sporadic case of PEPD from France, has been shown to cause a significant shift in the voltage-dependence of steady-state fast-inactivation of mutant channels but does not produce a persistent current [19]. The effects of M1627K on other properties that influence channel function, for example slow-inactivation, recovery from fast-inactivation, and responses to ramp stimuli, have not been reported. Also, while it is predicted that impaired fast-inactivation could be linked to repetitive firing in hyperexcitable DRG neurons [26], the effects of the PEPD mutation on DRG neuron excitability have not been experimentally demonstrated.We report here the identification of M1627K mutation in a new family with PEPD. We confirm that the M1627K mutation causes a large depolarized shift in fast-inactivation without altering channel activation, and show that mutant channels inactivate from the open state with slower kinetics, but without altering the rate for development of closed-state inactivation. We demonstrate that M1627K channels recover from fast-inactivation faster than wild type (WT) channels, and produce larger currents in response to a ramp stimulus. We also show, for the first time, that a PEPD mutation reduces the threshold for single action potentials and increases the number of action potentials in DRG neurons in response to graded stimuli, providing a link between altered channel biophysics and the pain that occurs in PEPD.ResultsClinical phenotypeAn English family with a history of PEPD in two generations was evaluated for a linkage with mutations in SCN9A. The proband is a 36 year old Caucasian female who presented to her physician with a lifelong history of episodic erythema and painful burning sensations from the waist downward. The episodes of erythema and extreme pain last for approximately 20 minutes. Her symptoms began in infancy and occurred several times a year, with pain triggered by bowel movement, passage of flatus or any painful stimulus in the lower half of her body. The patient experienced two attacks during each of her two deliveries, each of which occurred by Caesarian section under general anesthesia. Treatment with carbamazepine was initiated in infancy and currently has been successful in reducing her attacks to approximately one per year. As a teenager, the patient tried to reduce the daily dose; however, at 400 mg of carbamazepine per day she experienced an increase in the frequency of her attacks. Her symptoms dramatically improved upon returning to the original dose of 600 mg per day. Based on the clinical evaluation, the patient was diagnosed with PEPD.The proband's father and sister have a similar presentation of the disease since infancy. The father is currently not treated with any medication and feels that the frequency of attacks has increased with age. The proband's sister has been treated effectively with carbamazepine since childhood, but requires a lower dose of 200 mg/day. Prior to starting carbamazepine at one year of age, she had frequent episodes of pain and erythema from the waist downward. She self-reports that her symptoms have improved with age, and currently she has suffered just two attacks in the last five years. Unlike the proband, she can recall three attacks in her arms triggered by minor hand injuries. She also had two attacks during Caesarian section under general anesthesia.Identification of M1627K mutationGenomic DNA from the proband and her family (Figure 1A) was used to amplify all known exons of SCN9A, and their sequences were compared to the reference Nav1.7 cDNA [27]. Proband and control templates produced similar amplicons which were purified and sequenced. Sequence analysis identified a T to A in exon 26 (Figure 1B), corresponding to position 4879 of the reference sequence [27]. This mutation substitutes methionine (M) by lysine (K) at position 1627 of the polypeptide, which is located in the middle of the DIV/S4–5 linker. The missense mutation segregates with the disease in this family, with the affected father and a sister heterozygotes for the mutation and an unaffected brother carrying only the wild-type alleles (Figure 1A). M1627 is conserved in all human sodium channels (Figure 1C), and was previously identified independently in a sporadic case of PEPD from France [19].Figure 1Inheritance pattern of the M1627K mutation in DIV/S4–5 linker in Nav1.7 in familial PEPD. A, Inheritance of PEPD in two generations of a family with PEPD from the UK. Circles denote females; squares denote males, and symbols with diagonal line denote deceased individuals. The proband is indicated by an arrow. Blackened symbols indicate subjects affected with PEPD. B, Sequence traces of a segment of exon 26 in the region which encodes M1627. Left trace shows homozygous T4879 from exon 26 of the unaffected brother of the proband. Right trace shows heterozygous T4879A in exon 26 from the proband. This missense mutation leads to M1627K substitution. In (A), a (+) symbol denotes subjects heterozygous for the T4879A mutation in exon 26, and a (-) symbol denotes subjects without the mutation. Inheritance of PEPD segregates with the T4879A heterozygosity. C, Sequence alignment of DIV/S4–S5 linker from human sodium channels. The M1627K substitution in this family with PEPD is noted in the sequence from Nav1.7. The M1476I from a family with cold-induced myotonia occurs at the corresponding site as M1627 in Nav1.4. The M1652R from a family with LQT-3 occurs at the adjacent methionine residue in Nav1.5 (M1628 in Nav1.7).Voltage-clamp electrophysiology: activation and deactivationSodium currents from WT hNav1.7R, and the mutant channel M1627K were recorded from stably transfected HEK 293 cells in order to study the effects of this mutation on gating properties of the channel. Figure 2 shows representative WT (Figure 2A) and M1627K (Figure 2B) whole-cell currents elicited with a series of depolarizing test pulses from a holding potential of -100 mV. Although the peak current amplitude was smaller for M1627K channels (1.3 ± 0.2 nA, n = 15) than for WT channels (4.2 ± 0.7 nA, n = 12), the peak current-voltage relationship was similar for WT and M1627K channels (Figure 3A). The midpoint of activation (estimated by fitting the data with a Boltzman function) was -25.4 ± 1.0 mV (n = 14) for M1627K currents and -22.5 ± 1.9 mV (n = 11) for WT currents (p > 0.05). The kinetics of deactivation, which reflect the transition from the open to the closed state, of WT and M1627K channels were also examined by eliciting tail currents at a range of potentials after briefly activating the channels (at -20 mV for 0.5 ms). The time constants for deactivation were not altered at potentials ranging from -100 mV to -40 mV for the M1627K channel (Figure 3B). Together these data confirm that the M1627K mutation does not alter Nav1.7 activation properties.Figure 2M1627K currents decay slower than WT Nav1.7 currents. Representative WT (A) and M1627K (B) Nav1.7 currents are shown. Cells were held at -100 mV and currents were elicited with 50 ms test pulses to potentials ranging from -80 to 40 mV. For better comparison, WT and M1627K currents elicited with -30 mV (C) and +25 mV (D) depolarizations are shown superimposed. Although the rate of activation is not apparently altered, the decay phase is substantially faster for WT currents.Figure 3The M1627K mutation alters inactivation properties of Nav1.7. A, Normalized peak current-voltage relationship for WT (filled squares, n = 11) and M1627K (open circles, n = 15) channels. Cells were held at -100 mV and currents were elicited with 50 ms test pulses to potentials ranging from -80 to 40 mV. B, Time constants for tail current deactivation at repolarization potentials ranging from -40 to -100 mV for WT (filled squares, n = 8) and M1627K (open circles, n = 9) Nav1.7 channels. Time constants were obtained with single exponential fits to the deactivation phase of the currents. C, Comparison of steady-state fast-inactivation for WT (filled squares, n = 11) and M1627K (open circles, n = 15) Nav1.7 channels. Currents were elicited with test pulses to 0 mV following 500 ms inactivating prepulses. D, Comparison of slow-inactivation for WT (filled squares, n = 3) and M1627K (open circles, n = 4) Nav1.7 channels. Slow inactivation was induced with 10 s prepulses, followed by 100 ms pulses to -120 mV to allow recovery from fast-inactivation. A test pulse to 0 mV for 20 ms was used to determine the fraction of current available.Voltage-clamp electrophysiology: inactivationWe next examined the effects of the M1627K mutation on inactivation properties. The voltage-dependence of steady-state fast-inactivation was dramatically shifted in the depolarizing direction by the M1627K mutation (Figure 3C). The midpoint of fast-inactivation (V1/2, measured with 500 ms pre-pulses) was -75.3 ± 1.7 mV for WT (n = 11) and -56.0 ± 1.4 mV for M1627K (n = 15) channels. The slope of the steady-state inactivation relationship was 6.2 ± 0.1 mV/e-fold for WT channels and 9.3 ± 0.6 mV/e-fold for M1627K channels. Differences in the V1/2 and slope of fast-inactivation fit were significant (p < 0.001). However, the fraction of current remaining available after the -10 mV inactivation prepulse was not significantly different (p > 0.05) between WT (1.3 ± 0.3%) and M1627K (1.9 ± 0.5%) channels.By contrast, the voltage-dependence of slow-inactivation of hNav1.7R currents was only slightly altered by the M1627K mutation (Figure 3D). Ten second pre-pulses, followed by 100 ms recovery pulses to -120 mV to allow recovery from fast-inactivation, preceded the test pulse (to 0 mV for 20 ms) to determine the fraction of current available. The M1627K mutation reduced the fraction of slow-inactivated mutant channels that occurred at -80, -70 and -60 mV, which might contribute to increased channel availability at normal resting membrane potentials.It should be noted that in this study we used 500 ms inactivating prepulses to determine the voltage-dependence of steady-state fast-inactivation. The time-constant for development of fast-inactivation at -70 mV is ~140 ms for WT channels [28] and therefore a 500 ms prepulse, which allows ~97% of the channels to inactivate at -70 mV, is ideal for measuring the voltage-dependence of steady-state fast-inactivation. One concern is that 500 ms could allow channels to also enter a state of slow-inactivation. Our data does not support this conclusion because 1) the time constant for development of slow-inactivation of WT channels at -50 mV was greater than 1000 ms and 2) accumulation of channels in a slow-inactivated state after 500 ms at -50 mV was less than 10% [29]. To further investigate this issue, we compared the voltage-dependence of WT and M1672K channel fast-inactivation using 100 ms prepulses (Figure 4). Although the inactivation curve for WT channels was more depolarized with the 100 ms prepulse than the 500 ms prepulse (due to incomplete fast-inactivation at negative potentials) the inactivation curves for M1672K channels were nearly identical (Figure 4). These data support the conclusion that the M1672K mutation causes a depolarizing shift in the voltage-dependence of fast-inactivation.Figure 4Effect of prepulse duration on voltage-dependence of inactivation. A, Comparison of steady-state fast-inactivation for WT (squares, n = 3) and M1627K (circles, n = 4) Nav1.7 channels. Currents were elicited with test pulses to 0 mV following either 100 ms (filled symbols) or 500 ms (unfilled symbols) inactivating prepulses. The apparent voltage-dependence of inactivation of WT Nav1.7 channels is altered by the prepulse duration. In contrast the apparent voltage-dependence of inactivation of M1627K mutant channels is not significantly altered by changing the prepulse duration from 100 to 500 ms. Data are presented as mean ± S.E.Mutant channels exhibited slower decay kinetics (Figure 2) compared to WT channels, and we observed significant differences in the time constants for current inactivation between WT and M1627K channels. The rate of open-state inactivation was quantified by fitting the decay phase of the macroscopic current with a single exponential function. The time constants estimated from these fits are plotted as a function of the test potentials (Figure 5A). The time constants were much larger for M1627K currents than for WT currents over voltages that range from -30 to +40 mV. At -30 mV, for example, WT currents inactivated with a time constant of 5.8 ± 0.8 ms (n = 6) and M1627K currents inactivated with a time constant of 9.9 ± 1.3 ms (n = 6). At +25 mV, WT currents inactivated with a time constant of 0.44 ± 0.01 ms (n = 6) and M1627K currents inactivated with a time constant ten times larger (4.9 ± 0.4 ms, n = 6). These differences were statistically significant (p < 0.05). We also examined the development of closed-state inactivation at voltages ranging from -90 to -50 mV (Figure 5A, triangular symbols). Cells were stepped to the inactivation potential (from a holding potential of -100 mV) for increasing durations, and then stepped to the test potential (0 mV) to measure the fraction of the remaining available channels. The data from these cells were fitted with a single exponential function. Interestingly, the time course for the development of inactivation from the closed state was not significantly altered by the M1627K mutation, indicating that the M1627K mutation has a greater effect on open-state inactivation than closed-state inactivation.Figure 5The M1627K mutation alters rate of open-state inactivation and recovery from inactivation properties of Nav1.7. A, Time constants for development of fast-inactivation as a function of voltage for WT (filled symbols, n = 6) and M1627K (open symbols, n = 6) Nav1.7 channels are shown. Open-channel inactivation time constants were measured at voltages ranging from -45 to +40 mV by fitting the decay phase of currents elicited with depolarizing pulses with single exponentials. The time constants for development of closed-state inactivation were estimated from single exponential fits to time courses measured at inactivation potentials ranging from -90 to -50 mV for WT (filled triangles, n = 6) and with M1627K (open triangles, n = 6) Nav1.7 channels. B, Recovery from inactivation kinetics are faster for M1627K mutant channels (open circles, n = 7) than for WT Nav1.7 channels (filled squares, n = 7). Time constants were estimated from single exponential fits to time courses measured at recovery potentials ranging from -140 to -60 mV. The recovery from inactivation voltage protocol involved prepulsing the cell to -20 mV for 20 ms to inactivate all of the current, then stepping the membrane potential back to the recovery potential for increasing recovery durations prior to the test pulse to 0 mV. The maximum pulse rate was 0.5 Hz.Voltage-clamp electrophysiology: recovery from fast-inactivationThe time course for recovery from fast-inactivation (repriming) of WT and M1627K channels was measured at recovery voltages ranging from -140 to -60 mV. Fast-inactivation was induced with 20 ms inactivating prepulses to -20 mV. The time course for recovery from inactivation for both WT and M1627K currents could be fitted with single exponential functions. Recovery from inactivation was significantly faster for M1627K channels than for WT channels (Figure 5B). For example, the time constant for recovery of WT channels (τ = 92 ± 11 ms, n = 7) was almost 4-fold larger at -70 mV than the corresponding time constant for M1627K channels (τ = 26 ± 3 ms, n = 7).Voltage-clamp electrophysiology: response to a ramp stimulusIEM mutations are known to produce enhanced responses to small, slow depolarizations (ramp stimuli) [1], but the ramp response has not been studied in PEPD mutations [19]. Figure 6A shows the ramp current recorded from a M1627K cell producing a current with peak amplitude of 2.8 nA, compared to the ramp current recorded from a WT cell producing a comparable current with a peak amplitude of 3.4 nA. The ramp currents elicited with slow ramp (0.2 mV/ms) depolarizations from -100 to +20 mV were expressed as a percentage of peak current (Figure 6B) and the relative amplitude of the ramp currents were significantly larger for M1627K channels (7.0 ± 1.0%; n = 9) than for WT channels (1.0 ± 0.2%; n = 7). We compared the relative amplitude of the ramp currents with the extent of overlap between the voltage-dependence of activation and steady-state fast-inactivation (Figure 6C). As shown, the relative amplitude of the WT and M1627K ramp currents, and the voltage-dependence of these currents, correlates reasonably well with the respective overlap between the voltage-dependence of activation and steady-state fast-inactivation.Figure 6The M1627K mutation increases the amplitude of currents elicited by slow ramp depolarizations. A, Representative ramp currents elicited by 600 ms long ramp depolarizations from -100 to +20 mV are shown recorded from a HEK293 cell expressing WT channels (black trace) and from one expressing M1627K channels (red trace). The peak transient current amplitude elicited in the WT cell was larger (3.4 nA) than that of the M1627K cell (2.8 nA). B, The average relative ramp current (ramp current divided by peak transient current amplitude) is larger for M1627K cells (red trace, n = 9) than for WT cells (black trace, n = 7). C, The properties of the averaged ramp currents are compared to the overlap between the voltage-dependence of activation (derived from the current-voltage relationship) and steady-state fast inactivation. The inverted amplitude of the WT ramp current was scaled so that it corresponded to the peak overlap between WT activation and steady-state fast inactivation (black shaded area). As can be seen, voltage-dependence of the WT ramp current corresponds to the region of overlap. Furthermore, the relative amplitude and voltage-dependence of the M1627K ramp currents corresponds to the overlap between M1627K activation and steady-state fast inactivation (red shaded area).Current Clamp electrophysiologyPrevious studies have not examined the effect of PEPD mutations on DRG neuron excitability [19]. We assessed the effect of the M1627K mutation on DRG neuron excitability by recording in current-clamp mode from DRG neurons transfected with GFP and either WT or the M1627K mutant construct. The GFP-positive cells that were selected for recording were similar for the two groups in cell size, as measured either by apparent cell body diameter (24.1 ± 0.6 μm for M1627K vs. 25.5 ± 0.5 μm for WT) or by measured whole-cell capacitance (23.4 ± 1.5 pF for M1627K vs. 23.6 ± 1.4 pF for WT). The expression of the M1627K mutant channels caused a significant reduction in the threshold to generate action potentials. DRG neurons (n = 28) that express WT channels required on average about 300 pA to reach threshold (292 ± 37 pA), whereas DRG neurons that express the M1627K mutant channel (n = 31) only required, on average, about half the amount of current to reach threshold (154 ± 20 pA; p < 0.005). One possibility for the reduction in threshold is an increase in input resistance that would give a larger depolarization to a fixed stimulus current. However, the difference of the input resistance between the two groups of transfected cells did not reach statistical significance (M1627K: 795 ± 105 MΩ; WT: 662 ± 67 MΩ; p = 0.3). A similar small, but non-significant, difference had been reported for the A863P IEM mutant [17].No significant difference in resting membrane potential (RMP) was observed between the two groups of cells (M1627K: -60.1 ± 1.2 mV; WT: -60.1 ± 1.6 mV; p = 0.97). Additionally, the amplitude of action potential elicited at threshold showed no significant differences between M1627K transfected DRG neurons (n = 31) compared to WT transfected DRG neurons (n = 28) for peak (M1627K: 60.3 ± 2.2 mV, WT: 54.8 ± 2.9 mV, Mann-Whitney rank sum test, p = 0.24) or for after-hyperpolarization (M1627K: -64.2 ± 1.3 mV, WT: -59.2 ± 1.8 mV, t-test, p = 0.07).DRG neurons that express the M1627K mutant channels fired more action potentials in response to suprathreshold graded stimuli compared to neurons that express WT channels. When the number of spikes elicited by a series of 1-second long depolarizing current injections was quantified, it is clear that on average the M1627K transfected cells responded with more spikes over the entire current injection range compared to WT transfected cells (Figure 7). The difference in firing was significantly different (p < 0.05) for the current injections between 200 pA to 750 pA. Since the threshold value for the M1627K mutant was significantly lower than for WT channel, one possible explanation would be that we are not comparing cells at the equivalent rheobase. Therefore, we carried out further analysis to compare these two groups as a function of fold-threshold. This analysis clearly demonstrated that the M1627K mutant causes increased excitability, reaching significance for 3× threshold as well for the maximum response that could be evoked (Figure 8).Figure 7M1627K mutant channels increase frequency of firing of DRG neurons compared to WT Nav1.7 channels. The mean total number of action potentials (defined as spikes in membrane potential greater than 0 mV) for a current injection pulse of 1 second duration is plotted as a function of stimulus current intensity. The responses of neurons expressing M1627K mutant channels (n = 31) was significantly elevated over the responses of neurons expressing WT Nav1.7r channels (n = 28) for stimulus current injections exceeding 200 pA (p < 0.05).Figure 8Firing frequency is increased in current clamped DRG neurons expressing M1627K mutant channels when stimulus intensities are compared at fold-threshold. For each cell the firing frequency at 2× and 3× threshold was determined and then averaged. Since threshold is lower for M1627K expressing neurons (154 ± 19 pA) compared to hNav1.7r-WT expressing neurons (292 ± 37 pA), the mean stimulus intensity at 2× and 3× was similarly scaled (319 ± 40 pA for M1627K versus 617 ± 68 for WT at 2× threshold; 483 ± 59 pA for M1627K versus 926 ± 101 pA for WT at 3× threshold). The firing frequency was significantly different between M1627K and WT for 3× threshold and peak firing frequency (p < 0.05 and p < 0.01) The lowest stimulus intensity that elicited the peak firing frequency was not significantly different (545 ± 50 pA for M1627K versus 564 ± 60 pA for WT).Higher rates of firing appeared to occur in M1627K cells with the lowest threshold values, which often did not reach their maximum firing rate until current injections of >10× threshold. In contrast, most of the WT transfected cells only produced a maximum of 1 or 2 action potentials even for current injections exceeding 1 nA (but still less than 4× threshold). Examples of these firing phenotypes are shown in Figure 9. The responses from a cell transfected with the M1627K mutant sodium channel in this case had a threshold of only 25 pA and reached a maximum response of 18 impulses/second to a current injection of only 350 pA. In contrast, this example of a WT transfected cell had threshold of 450 pA and fired a maximum of 2 spikes in response to a current injection of 1200 pA.Figure 9Representative traces illustrating the differences in action potential firing in DRG neurons expressing either WT or M1627K channels. The upper panels are the responses of a DRG neuron expressing M1627K channels in which the threshold for action potential generation was 25 pA. Panel A: stimulus intensity was 50 pA (2× threshold); panel B: stimulus intensity was 75 pA (3× threshold); panel C: stimulus was 350 pA (14× threshold) at which the peak rate was recorded. The lower panels are the responses of a DRG neuron expressing WT Nav1.7 channels in which the threshold for action potential generation was 450 pA. Panel D: stimulus intensity was 900 pA (2× threshold); panel E: stimulus was 1400 pA (3× threshold); panel F: shows that the peak response for this cell occurred at 1200 pA.DiscussionIn this study, we identified a mutation, M1627K, in Nav1.7 from a previously unreported family with PEPD. Life-long symptoms in two members of the family are controlled by the sodium channel blocker carbamazepine. We investigated the functional effect of the M1627K substitution on Nav1.7, and confirm that it causes a large depolarizing shift in the voltage-dependence of steady-state fast-inactivation, with no effect on channel activation. Using current clamp, we show that M1627K mutant channels lower threshold for single action potentials and increase the number of action potentials in response to graded suprathreshold stimuli in small DRG neurons, most of which are nociceptors. Thus we show, for the first time, that a PEPD mutation produces nociceptor hyperexcitability.Missense mutations in Nav1.7 have been linked to two types of inherited painful neuropathies, early-onset IEM [12-18], and PEPD [19]. While missense mutations in Nav1.7 have been found in most families with IEM and PEPD in whom the gene was sequenced, both IEM and PEPD may be genetically heterogenous because mutations in the coding exons of SCN9A were not found in 5 cases of PEPD [19], or in cases of familial early-onset [30] and adult-onset IEM [31], suggesting that other target genes or mutations in non-coding regions of SCN9A might underlie these cases. Mutations in the coding exons of sensory neuron-specific sodium channels Nav1.8 and Nav1.9 have been ruled out as causative in these cases of inherited erythromelalgia [30]. We present in this study a previously unreported familial case of PEPD from the UK with the mutation M1627K, in the DIV/S4–S5 linker, which is also present in a sporadic case of a male patient from France [19,20].The patients in our study have responded favorably to treatment with carbamazepine, similar to those reported previously [20]. Carbamazepine targets voltage-gated sodium channels [32] and potently inhibits TTX-S channels in DRG neurons [33], including those specifically produced by wild-type Nav1.7 channels [34]. Although we did not examine the sensitivity of M1672K channels to carbamazepine in this study, carbamazepine has been shown to block persistent currents generated by I1461T and T1464I PEPD Nav1.7 mutant channels [19], and it is likely that inhibition of M1672K channel activity contributes to the therapeutic action of carbamazepine in the two patients examined in our study.Our observations of a 19 mV depolarizing shift in the voltage-dependence of fast-inactivation for M1627K confirm the linkage of impaired fast-inactivation of Nav1.7 and PEPD reported by Fertleman et al [19]. We report here that recovery from fast-inactivation was accelerated in the M1627K, consistent with a destabilized inactivation state of the channel. We have also observed that M1627K display an increased ramp current, similar to IEM mutations [13,16,17,21-25]. M1627K displayed a trend toward a small (< 3 mV) hyperpolarizing shift in activation, but this was not statistically significant. Depolarizing shifts in fast-inactivation have also been observed in some [13,16,17] but not all [21-25] mutations that cause IEM. However, the depolarizing shifts in fast-inactivation associated to date with IEM (= 10 mV) are smaller than the shifts in PEPD mutations ([19] and this study). It is intriguing that the IEM mutation A863P with a +10 mV depolarizing shift in fast-inactivation [13,16,17] does not yield symptoms of PEPD, suggesting that other factors, perhaps genetic makeup or bigger shifts in the voltage-dependence of fast-inactivation may contribute to clinical manifestations of the disease.Gating properties of Nav1.7, for example slow recovery from fast-inactivation and an ability to respond to ramp stimuli [28,35], suggest that, normally, it may act as a \"threshold\" channel which boosts subthreshold stimuli, and thus sets the gain in nociceptors [36,37]. Therefore, it is not surprising that mutations lowering the voltage-threshold for channel activation as in IEM lead to DRG neuron firing in response to a weaker stimulus that may normally be innocuous. Mutations that impair fast-inactivation as in PEPD allow more current to pass through the mutant channel, and thus induce stronger depolarization that brings the DRG neuron closer to the voltage-threshold for all-or-none action potential firing. By analogy to mutations in the cardiac channel Nav1.5 which cause arrhythmias and in neuronal channel Nav1.1 which causes epilepsy [26], the M1627K PEPD mutation which impairs fast-inactivation of Nav1.7 would be expected to increase repetitive firing, leading to hyperexcitablity of DRG neurons.Indeed, we now show in this study that a PEPD mutation in human Nav1.7 channels in DRG neurons renders these cells hyperexcitable. Current clamp recordings showed a lower threshold for single action potentials, and an increased firing rate in response to suprathreshold stimuli, but did not show a change in resting membrane potential for DRG neurons expressing M1627K channels. However, Harty et al [13,16,17] have shown that depolarization of RMP produced by an IEM mutation (A863P) contributes to, but is not solely responsible for, the increase in DRG neuron hyperexcitability produced by that mutation. Thus, impaired fast-inactivation, accelerated repriming, and the enhanced response to slow depolarizations may all have contributed to the hyperexcitability of DRG neurons expressing M1627K.The DIV/S4–5 linker is highly conserved in length and sequence (Figure 1C) among all sodium channels described to date, suggesting an important role in the normal functioning of the channel. Increasing the length of DIV/S4–S5 linker in Nav1.4 channels renders the mutant channels non-functional [38]. Importantly, mutations in this linker in several channels underlie pathological conditions [19,38-44]. Interestingly, substitution of the first methionine (Ma) in this linker (Figure 1C) with a positively charged residue (M1627K) in Nav1.7 causes PEPD ([19] and this study), while substitution with a hydrophobic residue isoleucine (M1476I) in Nav1.4 causes cold-induced myotonia [42]. Subsitution of the second methionine (Mb) with a positively charged residue (M1652R) in Nav1.5 causes LQT-3 syndrome [44]. All three disorders are linked to hyperexcitability of the cell in which they are expressed, irrespective if it is a neuron or a myocyte. The similarity of the outcome suggests a common mechanism of action, consistent with a conserved function of this linker in channel gating.Site-directed mutagenesis studies have suggested that the DIV/S4–5 linker contributes to the receptor for the fast-inactivation tripeptide IFM in loop 3 (L3) which links DIII and DIV [45-47]. Structural studies have shown that this linker can acquire an α-helical structure [48] with several residues including the MaMb (Figure 1C) forming a hydrophobic cluster that is important for inactivation, but indicate that these residues do not interact directly with the IFM motif [47,49,50]. Taken together, these studies suggest a model of two antiparallel α-helices [48]; this structure positions MaMb to interact with Y1470Y1471 (numbers according to Nav1.7) in L3. Substitution of the residues that correspond to Ma ([19] and this study) or Mb [44] or Y1470Y1471 residues destabilizes the inactivated state of the channel and yields similar gating changes in several channels. Interestingly, phosphorylation by Fyn kinase of Y1495 which is predicted to interact with Ma in Nav1.5 (equivalent to Y1471 in Nav1.7), produces a significant depolarizing shift in the voltage-dependence but no effect on the rate of steady-state fast-inactivation [51]. Thus the introduction of a charged residue at either of these two sites destabilizes this interaction and leads to impaired binding of the inactivation gate with its receptor.In summary, our results show that a PEPD mutation produces hyperexcitability in DRG neurons. Our findings also confirm the impairment of fast-inactivation previously associated with PEPD mutations, but show that, in addition, a PEPD mutation can enhance the response of the Nav1.7 channel to small, slow depolarizations and accelerate repriming. These data contribute to a better understanding of the pathophysiology of pain in patients with PEPD and provide additional support for efforts to develop Nav1.7-specific therapeutics for treatment of neuropathic pain.Materials and methodsPatientsThe proband (Figure 1) is a 36 year old female with a history of erythema and burning pain in the lower parts of the body. Family consent was obtained according to an approved institutional review board protocol and blood samples were then withdrawn and analyzed for mutations in SCN9A.Exon screeningGenomic DNA (gDNA) was purified from venous blood. Human Caucasian variation panel DNA (25 males, 25 females; The Coriell Institute, Camden, NJ) was used as a normal population control. Coding exons and flanking intronic sequences were amplified and sequenced as described previously [13]. Briefly, PCR amplification was carried out using 150 ng gDNA, 1 μM primers and Expand Long Template polymerase (Roche, Indianapolis, IN) in a 50 μl reaction volume for 35 cycles (95°C × 30 s, 55°C × 30 s and 72°C × 1 min.). Short exons were amplified using two primers, whereas exon 26 required four sets of primers to cover its entirety. Genomic sequences were compared to the reference Nav1.7 cDNA [27] to identify sequence variation. Sequencing was performed at the Howard Hughes Medical Institute/Keck Biotechnology Center at Yale University. Sequence analysis used BLAST (National Library of Medicine) and Lasergene (DNAStar, Madison, WI).Voltage-clamp electrophysiologyThe plasmid carrying the TTX-resistant (TTX-R) version of human Nav1.7 cDNA (hNav1.7R) was previously described [35]. The M1627K mutation was introduced into hNav1.7R using QuickChange XL II site-directed mutagenesis (Stratagene, La Jolla, CA). Transfected HEK 293 cells, grown under standard culture conditions (5% CO2, 37°C) in Dulbeccos's Modified Eagle's Medium supplemented with 10% fetal bovine serum, were treated with G418 for several weeks, and stable cell lines that express the mutant channel were selected.Whole-cell patch-clamp recordings were conducted at room temperature (~21°C) using an EPC-10 amplifier and the Pulse program (v 8.5, HEKA Electronic, Germany). Fire-polished electrodes (0.8–1.5 MΩ) were fabricated from 1.7-mm VWR capillary glass using a Sutter P-97 puller (Novato, CA). Average access resistance was 2.1 ± 0.1 MΩ (mean ± SE, n = 27). Voltage errors were minimized using ~40–75% series resistance compensation to achieve identical (1.5 ± 0.2 mV) voltage error after series resistance compensation for the two groups. Capacitance artifacts were canceled using computer-controlled circuitry of the patch clamp amplifier and linear leak subtraction was used for all voltage clamp recordings. Recordings were always started 3.5–4 minutes after establishing the whole-cell configuration. Membrane currents were filtered at 5 kHz and sampled at 20 kHz. The pipette solution contained (in mM): 140 CsF, 1 EGTA, 10 NaCl and 10 HEPES (pH 7.3). The standard bathing solution was (in mM) 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, and 10 HEPES (pH 7.3). Data were analyzed using Pulsefit (HEKA Electronic, Germany) and Origin (Microcal Software, Northampton, MA) software. Unless otherwise noted, statistical significance was determined (p < 0.05) using an unpaired t-test. Results are presented as mean ± SEM and error bars in the figures represent standard errors (SE).Transfection of DRG neuronsThe protocol for the care and sacrifice of rats used in the study was approved by the Veterans Administration CT Healthcare system IACUC. DRG tissue from 1- to 5-day old Sprague Dawley rats were harvested and dissociated using a protocol that was adapted from Rizzo et al [52]. Briefly, dissected ganglia were placed in ice cold oxygenated complete saline solution (CSS), which contained (in mM) 137 NaCl, 5.3 KCl, 1 MgCl2, 25 sorbitol, 3 CaCl2, 10 N-2-hydroxyethylpiperazine-N' -2-ethanesulfonic acid (HEPES); pH 7.2. They were then transferred to an oxygenated, 37°C CSS solution containing 1.5 mg/ml Collagenase A (Roche Applied Science, Indianapolis, IN) and 0.6 mM EDTA and incubated with gentle agitation at 37°C for 20 min. This solution was then exchanged with an oxygenated, 37°C CSS solution containing 1.5 mg/ml Collagenase D (Roche Applied Science, Indianapolis, IN), 0.6 mM EDTA and 30 U/ml papain (Worthington Biochemical, Lakewood, NJ) and incubated with gentle agitation at 37°C for 20 min. The solution was then aspirated and the ganglia triturated in DRG media [(DMEM/Fl2 (1:1) with 100 U/ml penicillin, 0.1 mg/ml streptomycin (Invitrogen, Carlsbad, CA) and 10% fetal calf serum (Hyclone, Logan, UT)], which contained 1.5 mg/ml bovine serum albumin (Sigma-Aldrich, St. Louis, MO) and 1.5 mg/ml trypsin inhibitor (Roche Applied Science, Indianapolis, IN).Either WT hNav1.7R or M1627K mutant channels were transiently transfected into the DRG neurons, along with enhanced-GFP, by electroporation with a Nucleofector II (Amaxa, Gaithersburg, MD) using Rat Neuron Nucleofector Solution and program G-013. The ratio of sodium channel to GFP constructs was 5:1. Immediately after transfection, the cells were allowed to recover for 5 minutes in Ca2+- and Mg2+-free culture medium (DMEM + 10% FBS). The cell suspension was then diluted with DRG media containing 1.5 mg/ml bovine serum albumin and 1.5 mg/ml trypsin inhibitor, 80 μl was plated on 12 mm circular poly-D-lysine/laminin pre-coated coverslips (BD Biosciences, Bedford, MA) and the cells incubated at 37°C in 5% CO2 for 30 min. DRG media (1 ml/well), supplemented with 50 ng/ml each of mNGF (Alomone Labs, Jerusalem, Israel) and GDNF (Peprotec, Rocky Hill, NJ), was then added and the cells maintained at 37°C in a 5% CO2 incubator for 18–48 hr before recording.Current-clamp electrophysiologySmall (20–30 μm diameter) GFP-labeled DRG neurons were used for current-clamp recording. Neurons with round cell body morphology and clear processes were selected for analysis and recordings were performed between 20- and 50-hours post-transfection. Pipette resistance was 1–3 MΩ when filled with the pipette solution which contained (in mM): 140 KCl, 0.5 EGTA, 5 HEPES, and 3 Mg-ATP, pH 7.3 with KOH (adjusted to 315 mOsm with dextrose). The extracellular solution contained the following (in mM): 140 NaCl, 3 KCl, 2 MgCl2, 2 CaCl2, 10 HEPES, pH 7.3 with NaOH (adjusted to 320 mOsm with dextrose). Formation of a GΩ seal and the transition to whole-cell configuration was performed in voltage-clamp mode before proceeding to the current-clamp recording mode. Recordings were obtained using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA) connected to a Digidata 1422 interface controlled by Clampex software (Molecular Devices). Cells with resting membrane potentials (RMP) more negative than -40 mV that were stable (<10% variation during the first 5 minutes) were included for data analysis. Input resistance was determined by fitting the slope of a line fit to hyperpolarizing voltage responses to current steps of -5 pA to -25 pA in 5 pA increments. Threshold was determined by the first action potential elicited by a series of depolarizing current injections that increased in 5 pA increments. Action potentials were counted by detecting membrane potential transients that exceeded the threshold value of 0 mV during 1-sec long depolarizing current injections. Data are expressed as means ± SEM. Student's t-test was used to assess the significance (p < 0.05) of differences between parameters measured from DRG neurons transfected with WT or M1627K mutant channels. Where indicated, the Mann-Whitney rank sum test was applied because of non-normal data distributions. Statistical analysis was performed using Sigmaplot software (Systat Software, San Jose, CA).Competing interestsThe authors declare that they have no competing interests.Authors' contributionsSDD-H and TRC participated in the experimental design and interpretation of the data. ME and BJ collected, analyzed and interpreted electrophysiological data. LT identified the M1627K mutation, made the mutant M1627K construct and established the stable cell lines. TZF and ML collected and interpreted clinical data and confirmed a diagnosis of PEPD. SGW conceived the project, participated in the experimental design and interpretation. All authors participated in writing of the manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2556766.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2556766",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2556766\nAUTHORS: Haiyan Gao, David A Harrison, Gareth J Parry, Kathleen Daly, Christian P Subbe, Kathy Rowan\n\nABSTRACT:\nIntroductionCritical care outreach services (CCOS) have been widely introduced in England with little rigorous evaluation. We undertook a multicentre interrupted time-series analysis of the impact of CCOS, as characterised by the case mix, outcome and activity of admissions to adult, general critical care units in England.MethodsData from the Case Mix Programme Database (CMPD) were linked with the results of a survey on the evolution of CCOS in England. Over 350,000 admissions to 172 units between 1996 and 2004 were extracted from the CMPD. The start date of CCOS, activities performed, coverage and staffing were identified from survey data and other sources. Individual patient-level data in the CMPD were collapsed into a monthly time series for each unit (panel data). Population-averaged panel-data models were fitted using a generalised estimating equation approach. Various potential outcomes reflecting possible objectives of the CCOS were investigated in three subgroups of admissions: all admissions to the unit, admissions from the ward, and unit survivors discharged to the ward. The primary comparison was between periods when a formal CCOS was and was not present. Secondary analyses considered specific CCOS activities, coverage and staffing.ResultsIn all, 108 units were included in the analysis, of which 79 had formal CCOS starting between 1996 and 2004. For admissions from the ward, CCOS were associated with significant decreases in the proportion of admissions receiving cardiopulmonary resuscitation before admission (odds ratio 0.84, 95% confidence interval 0.73 to 0.96), admission out of hours (odds ratio 0.91, 0.84 to 0.97) and mean Intensive Care National Audit & Research Centre physiology score (decrease in mean 1.22, 0.31 to 2.12). There was no significant change in unit mortality (odds ratio 0.97, 0.87 to 1.08) and no significant, sustained effects on outcomes for unit survivors discharged alive to the ward.ConclusionThe observational nature of the study limits its ability to infer causality. Although associations were observed with characteristics of patients admitted to critical care units, there was no clear evidence that CCOS have a big impact on the outcomes of these patients, or for characteristics of what should form the optimal CCOS.\n\nBODY:\nIntroductionCritical care outreach services (CCOS) were introduced widely into the National Health Service (NHS) in England in 2000 as an important component of the vision for the future of critical care services [1]. The three main objectives of CCOS were to avert admissions or ensure timely admission to critical care, to enable discharges from critical care, and to share skills with ward staff. There was no prescribed model for CCOS; Critical Care Networks and NHS Trust Critical Care Delivery Groups were encouraged to develop their own locally customised service. Despite little evidence for their benefit, CCOS were introduced without any formal prospective evaluation.A wide range of services falling under the umbrella of CCOS have been developed, introduced, incrementally implemented and improved over time [2]. These services vary in terms of their objectives (such as meeting one or more of the three main objectives or other additional objectives), activities (such as direct bedside support, follow-up of patients discharged from critical care to the ward, or education and training), staffing (such as doctor-led or nurse-led, or size of team), hours of work (such as round the clock or office hours) and coverage of wards (such as selected wards only or complete coverage) [3]. A systematic review on the effectiveness of CCOS [4] indicated that published research on the impact of CCOS is limited, there is insufficient evidence to confirm their effectiveness, and more comprehensive research is needed. As a result of the wide variation in the models of service delivery adopted and potentially wide variation in the stage of implementation and use, CCOS cannot now be evaluated using the gold-standard research design, a multicentre, randomised controlled trial.The aim of this study was to undertake a multicentre, interrupted time-series analysis of the impact of CCOS at the critical care unit level, as characterised by the case mix, outcome and activity of admissions to adult, general critical care units participating in the Case Mix Programme, which is the national comparative audit of critical care in England, Wales and Northern Ireland.Materials and methodsThe analysis sought to examine trends in pre-specified outcomes over time in those critical care units participating in the Case Mix Programme for which CCOS data were available from a previously completed survey.Data sourcesCase Mix Programme DatabaseThe Case Mix Programme Database (CMPD) is a high-quality clinical database of case mix, outcome and activity data on consecutive admissions to adult, general critical care units in England, Wales and Northern Ireland [5]. Data are collected by trained data collectors according to precise rules and definitions, and are validated both locally and centrally before being pooled into the CMPD. A total of 393,205 validated admissions to 172 critical care units between January 1996 and December 2004 were extracted from the CMPD.The Intensive Care National Audit & Research Centre (ICNARC) physiology score is an illness severity score calculated from the ICNARC risk prediction model [6], based on physiological measurements from the 24 hours after admission to critical care. Admissions were classified as either medical, elective surgical, or emergency surgical, on the basis of the source of admission to the unit and the National Confidential Enquiry into Perioperative Death classification of surgery, as described previously [5].Survey data and other sourcesThe results of a national survey of the evolution of CCOS in England [3] were used to identify units with formal CCOS, to characterise the CCOS in terms of the activities undertaken, coverage and staffing, and to identify important time-dependent confounders.A total of 191 acute NHS hospitals in England completed the survey. The survey data were validated extensively by a software data entry check, random-sample double data entry, and data cleaning.The following time-dependent variables were identified from the survey and, where necessary, other sources.The primary comparison was between periods when a formal CCOS was and was not present in the hospital housing the critical care unit, defined as at least one member of staff with funded time dedicated to the CCOS. Hospitals that were represented both in the CMPD and in survey data were contacted for details of the date on which the CCOS formally started, because this was not included in the survey.Secondary comparisons were performed by using the following variables to characterise each CCOS:1. Aspects of outreach activity, eight binary variables: (a) ward follow-up, (b) outpatient follow-up, (c) telephone advice, (d) direct bedside clinical support, (e) informal bedside teaching, (f) formal educational courses, (g) use of physiological track and trigger warning systems, and (h) audit and evaluation of outreach activity.2. Coverage of CCOS, two categorical variables: (a) temporal (24 hours and 7 days a week; 12 to 23 hours and 7 days per week; less than 12 hours and 7 days per week; selected days), and (b) locational (all wards/selected wards only).3. Staffing of CCOS, two categorical variables: (a) no medical involvement or some medical involvement (medical staff with dedicated funded sessions allocated to the CCOS), and (b) small team (fewer than three whole-time equivalent staff per ten level 3 or flexible level 2/3 beds) or large team (three or more whole-time equivalent staff per ten level 3 or flexible level 2/3 beds).All analyses were adjusted for the following confounding variables: number of level 3 beds (general and specialist); number of level 2 beds (general and specialist); number of flexible level 2/3 beds (general and specialist); presence of a standalone general high-dependency unit; teaching status; Foundation Trust status; tertiary referral centre; presence of a 'hospital at night' service; presence of an acute pain team; presence of a nutrition team; availability of non-invasive ventilation on general wards; presence of an overnight ventilation facility in theatre/recovery; use of the Acute Life-threatening Events Recognition and Treatment (ALERT) course, or similar, for ward staff; presence of a formal resuscitation policy.The timings of the opening of standalone general high-dependency units, granting of Foundation Trust status and initiation of 'hospital at night' services were not included in the survey. These were sought from individual hospitals or from the Department of Health or Modernisation Agency websites.Outcome measuresA variety of potential outcomes that might reflect the CCOS objectives of averting admissions, ensuring timely admission and enabling discharge were investigated in the following three subgroups of admissions.1. All admissions to the unit: proportion of admissions direct from the ward.2. Admissions from a ward in the same hospital: (a) proportion of admissions receiving cardiopulmonary resuscitation (CPR) during 24 hours before admission, (b) proportion of admissions out of hours (22:00 to 06:59), (c) mean and SD of the ICNARC physiology score, (d) proportion of admissions having all active treatment withdrawn, and (e) mortality in the unit.3. Unit survivors discharged to the ward: (a) proportion of discharges occurring out of hours (22:00 to 06:59), (b) proportion of discharges designated as an 'early discharge due to shortage of beds', (c) hospital mortality, and (d) proportion of patients readmitted to the unit within 48 hours of discharge.Statistical analysesThe interrupted time-series analysis included all admissions in the CMPD from critical care units located in hospitals for which a completed survey form was received. Units were excluded if we were unable to ascertain the formal start date for the CCOS. Missing data in the time-dependent variables identified from the survey were replaced with the last value carried forward unless all values from 1996 to 2004 were missing, in which case that unit was excluded.Time series consist of sets of values for the same variables collected at regular or irregular intervals. Data in the CMPD are collected on an individual patient basis; however, collapsing the data into a time series of monthly average values for each critical care unit enabled us to use statistical techniques to model trends and cycles over time. Population-averaged panel-data models were fitted by using a generalised estimating equation approach, with robust (Huber–White) variance–covariance estimates to account for clustering at the unit level [7], and an autoregressive correlation structure of order 1 within units over time.The primary analysis was on the presence of a formal CCOS. Lagged effects over two months were included in the model because the effects of introducing a new service are not likely to be evident immediately after the introduction. Secondary analyses were on CCOS activities, coverage and staffing, as defined above.All analyses were adjusted for a linear time trend, seasonality (11 dummy variables for the months February to December), and the 14 time-dependent confounding variables. In addition, analyses of admissions out of hours were adjusted for unit occupancy, and analyses of unit survivors discharged to the ward were adjusted for age, ICNARC physiology score and surgical status.Interactions between the categorical variables representing CCOS coverage and staffing were tested in the corresponding models.A sensitivity analysis was conducted for the outcome of CPR before admission by including only those patients in hospital for at least 24 hours before admission, to exclude CPR occurring out of hospital. A sensitivity analysis was also conducted for admissions having all active treatment withdrawn, restricting to active treatment withdrawal occurring within 48 hours of admission, because these may represent futile admissions that are more likely to be averted by a CCOS.Statistical analyses were performed with Stata 9.2 (StataCorp LP, College Station, TX, USA).ResultsIn all, 130 units were identified both in the CMPD and in survey data. Of these, 111 indicated the presence of CCOS and were contacted to acquire the formal start date; 107 (96%) responded. The four units that did not respond, for which no date for the start of formal outreach services could be identified, were dropped from the analyses. A further 18 units were dropped from the analyses because of missing values in the time-dependent survey data.Of the original 130 units, 108 (83%) were included in the analyses, of which 79 (73%) had a formal CCOS starting between 1996 and 2004. There was a median of 36.5 (quartiles 24 to 47) months' data after the introduction of CCOS in these units. The 29 units with no formal CCOS or with CCOS starting after 2004 were included as non-intervention sites to improve the modelling of time trends and confounders.The characteristics of patients in the three subgroups of admissions are described in Table 1.Table 1Descriptive statistics for all admissions, admissions from the ward and discharges to the wardStatistics and outcomesAll admissionsAdmissions from the wardDischarges to the wardAdmissions, n (percentage)240,884 (100)56,082 (23.3)138,160 (57.4)Age (years) Mean (SD)59.3 (19.4)60.1 (19.0)58.6 (19.7) Median (quartiles)64 (48–74)65 (50–74)63 (46–74)Males, n (percentage)139,176 (57.8)30,437 (54.3)78,986 (57.2)ICNARC physiology score Mean (SD)18.2 (10.2)21.5 (10.8)14.5 (7.7) Median (quartiles)17 (10–24)20 (14–28)13 (9–19)Admission type, n (percentage) Non-surgical139,376 (57.9)56,082 (100)66,214 (47.9) Elective surgical53,563 (22.2)NA43,099 (31.2) Emergency surgical47,945 (19.9)NA28,847 (20.8) Hospital mortality235,551 (32.6)25,847 (46.9)16,184 (11.7)Outcomes for admissions from the ward, n (percentage) CPR 24 hours prior to admission5,349 (9.6) Admission out of hours (22:00–06:59)16,312 (29.1) Active treatment withdrawn8,670 (15.4) Unit mortality18,040 (32.2)Outcomes for discharges to the ward, n (percentage) Discharge out of hours (22:00–06:59)8,870 (6.4) Early discharge due to shortage of beds5,440 (3.9) Readmission within 48 hours1,919 (1.4)CPR, cardiopulmonary resuscitation; ICNARC, Intensive Care National Audit & Research Centre; NA, not applicable.The effects of the presence of a formal CCOS and its lag over two months on the predefined outcomes for the three subgroups of admissions are shown in Figures 1 to 3. The figures provide a graphical illustration of the effect estimates for the first, second, and third and subsequent months after the introduction of CCOS. The estimates for the first and second months represent the progression from no CCOS to having a CCOS; the estimate for the third and subsequent months represents the sustained effect of CCOS, presumed to remain constant for the life of the CCOS. Full details of the effect estimates are given in Additional file 1. There was no significant change in the proportion of all admissions coming from the ward (Figure 1). For admissions from the ward (Figure 2), the presence of a formal CCOS was associated with a significant decrease in CPR during 24 hours before admission, admission out of hours and the mean ICNARC physiology score. From the third month after the formal start date onwards, the effect estimate (95% confidence interval) and P value for these three outcomes were as follows: odds ratio 0.84 (0.73 to 0.96), P = 0.012; odds ratio 0.91 (0.84 to 0.97), P = 0.012; and decrease in mean 1.22 (0.31 to 2.12), P = 0.008, respectively. There was no significant change in the SD of the ICNARC physiology score, the proportion of admissions having all active treatment withdrawn, or unit mortality. For unit survivors discharged to the ward (Figure 3), there was an apparent increase in out-of-hours discharges (and an associated increase in hospital mortality) in the first month after the introduction of CCOS. This effect disappeared in the second and subsequent months.Figure 1The effect of critical care outreach services (CCOS) for all admissions to the unit. Effect estimate (odds ratio) and 95% confidence interval are shown for the first, second, and third and subsequent months after the introduction of CCOS.Figure 2The effect of critical care outreach services (CCOS) for admissions from the ward. Effect estimates and 95% confidence intervals are shown for the first, second, and third and subsequent months after the introduction of CCOS. CPR, cardiopulmonary resuscitation; ICNARC, Intensive Care National Audit & Research Centre; ICU, intensive care unit.Figure 3The effect of critical care outreach services (CCOS) for unit survivors discharged to the ward. Effect estimates and 95% confidence intervals are shown for the first, second, and third and subsequent months after the introduction of CCOS.The sensitivity analyses showed similar results on CPR before admission and active treatment withdrawal in the restricted subgroups.Full results of the secondary analyses on CCOS activities, coverage and staffing can be found in Additional file 1. We have the following observations.With regard to CCOS activities, the use of physiological track and trigger warning systems was associated with lower rates of CPR before admission (odds ratio 0.84, 95% confidence interval 0.72 to 0.98, P = 0.049) and the SD of the ICNARC physiology score (decrease in SD 0.06 (0.01 to 0.10), P = 0.010). Certain other activities were associated with statistically significant changes in outcomes, but with no plausible rationale for causality. For example, the presence of an outpatient follow-up service was associated with characteristics of admissions from the ward. It is likely that these represent spurious findings because of the number of tests performed.With regard to CCOS coverage, there were some statistically significant differences between coverage categories, but these were not consistent and did not show any expected 'dose-response' pattern.With regard to CCOS staffing, medical teams were associated with a lower proportion of ward admissions out of hours (odds ratio 0.92 (0.84 to 1.00), P = 0.046) and reductions in active treatment withdrawal (odds ratio 0.76 (0.59 to 0.97), P = 0.026) in comparison with teams with no medical involvement. Larger teams were associated with a higher proportion of all admissions coming from the ward (odds ratio 1.18 (1.02 to 1.35), P = 0.025), increased active treatment withdrawal in admissions from the ward (odds ratio 1.29 (1.02 to 1.64), P = 0.033) and higher hospital mortality for patients discharged to the ward (odds ratio 1.11 (1.02 to 1.21), P = 0.020) in comparison with smaller teams. The direction of causality in these associations is unclear.There were no significant interactions between the variables representing CCOS coverage and staffing.DiscussionThis study found that the presence of a formal CCOS was associated with a significant decrease in CPR rates during 24 hours prior to admission, out-of-hours admission (22:00 to 06:59) and mean ICNARC physiology score for admissions from the ward. There was no evidence for an association between the presence of a formal CCOS and the other outcomes investigated in this study. In particular, there was no effect on unit mortality for patients admitted to the critical care unit from the ward, and no sustained effect was seen on mortality or readmission rates for patients discharged alive from the critical care unit.Cardiopulmonary arrest is a clinically important adverse event that carries a high mortality. Such an event is often preceded by signs of physiological deterioration [8,9]. The findings in the present study suggest that the use of physiological track and trigger warning systems is an important part of CCOS activity. The use of such a system may lead to earlier intervention when a patient shows signs of deteriorating and may therefore reduce the CPR rate. A wide variety of track and trigger warning systems are in use, with little evidence of reliability, validity or utility [10]. In most previous studies it has been impossible to distinguish any effects of using a track and trigger system from other components of CCOS activity. Only one single centre study has evaluated the effect of introducing a track and trigger system in the absence of a specific CCOS or similar service providing the response [11]. The finding of reduced CPR is consistent with some previous studies in non-randomised before/after comparisons of CCOS or similar services [12-15]. However, other studies, including the MERIT cluster-randomised trial, have reported no significant effects on CPR rates [16-18]. CPR rates in patients admitted to critical care units may be reduced because arrest rates are reduced, but there are also other plausible explanations. It may be that the arrest rate remains the same but resuscitation is attempted less frequently through the more appropriate use of 'do not attempt resuscitation' decisions. Alternatively, it may be that the same number of arrests and resuscitation attempts are still taking place, but fewer of these patients are being admitted to critical care units because the CCOS determine admission to be futile. It is most likely that some combination of all these effects is taking place.Reductions in out-of-hours admissions to the intensive care unit (ICU) may result from a number of different processes. It may be that patients requiring critical care are being identified early and admitted appropriately during the working day, averting the need to admit the patient as an emergency in the middle of the night. Alternatively, it is possible that in hospitals with a CCOS that does not operate 24 hours per day, at-risk patients identified overnight are being left until the CCOS begins work in the morning rather than being referred directly to the ICU.The fact that acute severity of illness, as measured by the ICNARC physiology score, was reduced without an associated reduction in mortality may reflect lead-time bias – a reduction in the apparent severity of illness as a result of stabilisation before admission, rather than a true reduction in the underlying severity of illness [19]. However, true severity of illness may be affected by at least three processes if CCOS achieve the stated aim of averting admissions or ensuring timely admission. Averting ICU admissions that can be managed safely on the ward with the assistance of the CCOS would remove some of the least sick patients, resulting in an increase in the average severity of illness. Conversely, averting futile admissions that would not benefit from critical care by the increased used of decisions on treatment limitation would remove some of the sickest patients, resulting in a decrease in the average severity of illness. Finally, ensuring the timely admission of patients requiring critical care may enable them to be admitted at an earlier stage in the disease process, with lower severity of illness.The fact that other expected changes resulting from CCOS were not evident may be due to a genuine lack of benefit of CCOS or to the variability in the way in which these services were designed and implemented, and the funding available to them, leading to similar variability in their impact. There may be other factors not captured in the survey that could have had an impact on the effects of a CCOS, for example organisational or management and leadership styles or culture. There is some modest impact in places, but we must wait to see whether this will be sustained in the future.Overall, this study showed a very mixed picture. There is no clear evidence that CCOS have a big impact on patient outcomes. In addition, there do not seem to be any clear characteristics of what should form the optimal CCOS.The three major strengths of our study are the size, high-quality data and rigorous methodology. We performed a multicentre study on a national scale: data from 108 critical care units were included in the analyses, representing about half of all adult general critical care units in England. The CMPD has been independently evaluated in accordance with criteria for a high-quality database and scored highly [5]. The approach of interrupted time-series analysis has advantages over a simple before/after comparison because it controls for long-term trends and seasonality in the data, but it may be influenced by other events occurring at about the same time as the event of interest (historical bias) [20,21]. In the CMPD, we have time-series data for many critical care units (namely panel data or cross-sectional time-series data) [22]. The introduction of outreach services at different times in different locations produces a natural experiment by which we can reduce the effects of historical bias. Population-averaged panel-data models estimate the consistent (average) effect of CCOS across hospitals. This effect estimate is of most relevance for policy and planning decisions. All major potential confounding factors were identified and included in the study.There were several limitations to our study. First, the variations in the way in which CCOS have been implemented decrease our ability to analyse and understand their impact. However, because CCOS are widespread in England [3], a randomised controlled trial of their effectiveness is now infeasible. Well-controlled, multicentre observational studies are therefore likely to be the best way to gain additional insight into this topic. Second, the delivery of CCOS may have changed over the course of the study period. We were limited to information on the set-up of CCOS obtained from a survey conducted at a single point in time; however, had more detailed data been available, it is doubtful whether it would have been possible to fit such a complex model. Third, we observed associations with the introduction of CCOS but are unable to attribute causality. For example, we cannot determine from the data whether the decrease in CPR before admission to ICU was due to the prevention of arrests by earlier referral or to an increase in decisions on treatment limitation. Bradford Hill [23] has identified nine 'considerations for causality': strength of the association; consistency across observers, places, circumstances and times; specificity (that is, that the same association is not observed in other settings); temporal relationship; biological gradient (that is, dose-response); plausibility; coherence with what is already known in the area; experiment (which provides the strongest argument, when available); and analogy with similar situations. The multicentre interrupted time-series approach helps to establish consistency, specificity and temporal relationships. However, none of the associations could be considered to be overwhelmingly strong, and certain results, particularly among the secondary analyses, failed on consideration of plausibility or biological gradient. Fourth, although the population-averaged effect is the most relevant for policy decisions, it does not measure the expected benefit for an individual patient, because the population includes individuals with no potential to gain from the presence of CCOS. For this reason, we concentrated the analyses on subpopulations with the most potential to benefit. Finally, length of stay in critical care and in hospital may be important performance indicators and are strongly associated with costs, but these were not investigated because they are highly skewed variables, making it difficult to identify significant population-averaged effects.Further large, multicentre, prospective studies are required to identify which aspects of CCOS are truly effective. We propose to evaluate the impact of outreach services, at the patient level, by prospectively identifying admissions in the CMPD receiving outreach before and/or after their critical care episode.ConclusionAlthough some effects of CCOS were found, there is no clear evidence that CCOS have a big impact on outcomes of patients admitted to critical care. No clear characteristics of what should form the optimal CCOS could be identified, except that the use of physiological track and trigger warning systems seems potentially beneficial. There is some modest impact in places, but we must wait to see whether this will be sustained in the future and whether this is associated with improvements in important patient outcomes. Further large, multicentre prospective studies are required.Key messages• CCOS have been widely introduced in England with the aims of averting admissions to critical care, ensuring timely admission, enabling discharge and educating the ward staff.• Our interrupted time-series analysis demonstrates reductions in the proportion of admissions receiving CPR before admission, admission out of hours, and severity of illness for patients admitted to the ICU from the ward, but no effect on unit mortality.• There were no sustained effects on outcomes for unit survivors discharged to the ward.• Analysis of specific CCOS activities suggested that changes in admission characteristics may be attributable in part to the use of physiological track and trigger warning systems.AbbreviationsCCOS = critical care outreach services; CMPD = Case Mix Programme Database; CPR = cardiopulmonary resuscitation; ICNARC = Intensive Care National Audit & Research Centre; ICU = intensive care unit; NHS = National Health Service.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsHG and DAH led the design and analysis of the study and drafted the manuscript. GJP, KD, CPS and KR contributed to the design of the study, interpretation of results, and critical revision of the manuscript. All authors read and approved the final manuscript.Supplementary MaterialAdditional file 1A PDF file containing five tables listing detailed results of all primary and secondary analyses and sensitivity analyses.Click here for file\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2557141.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2557141",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2557141\nAUTHORS: Ching Ouyang, David D. Smith, Theodore G. Krontiris\n\nABSTRACT:\nVariation in gene expression may give rise to a significant fraction of inter-individual phenotypic variation. Studies searching for the underlying genetic controls for such variation have been conducted in model organisms and humans in recent years. In our previous effort of assessing conserved underlying haplotype patterns across ethnic populations, we constructed common haplotypes using SNPs having conserved linkage disequilibrium (LD) across ethnic populations. These common haplotypes cluster into a simple evolutionary structure based on their frequencies, defining only up to three conserved clusters termed ‘haplotype frameworks’. One intriguing preliminary finding was that a significant portion of reported variants strongly associated with cis-regulation tags these globally conserved haplotype frameworks. Here we expand the investigation by collecting genes showing stringently determined cis-association between genotypes and expression phenotypes from major studies. We conducted phylogenetic analysis of current major haplotypes along with the corresponding haplotypes derived from chimpanzee reference sequences. Our analysis reveals that, for the vast majority of such cis-regulatory genes, the tagging SNPs showing the strongest association also tag the haplotype lineages directly separated from ancestry, inferred from either chimpanzee reference sequences or the allele frequency-derived haplotype frameworks, suggesting that the differentially expressed phenotypes were evolved relatively early in human history. Such evolutionary signatures provide keys for a more effective identification of globally-conserved candidate regulatory haplotypes across human genes in future epidemiologic and pharmacogenetic studies.\n\nBODY:\nIntroductionVariation in allelic expression is very commonly observed in the human genome [1], [2] and, rather than alteration in protein products, may account for a significant fraction of inter-individual variation [3]. Therefore, identification of such variation is a major step toward understanding the differential predisposition to common diseases and variation in drug responses among individuals and ethnic populations. For example, slight changes in allelic expression of the tumor suppressor gene, APC, can affect predisposition to tumorigenesis [4]. Also, as recently illustrated, VKORC1 gene expression influences the warfarin maintenance dose [5]. In recent years, studies searching for association between genetic markers and quantitative gene expression profiling, referred to as genetical genomics [6], have been conducted in model organisms and humans (reviewed in [7], [8]). Loci associated with the variation of gene expression, described as expression quantitative trait loci (eQTL), have been identified both in cis and in trans for many genes.Following an assessment of common underlying haplotype patterns across ethnic populations, we previously reported the observation that pairwise linkage disequilibrium (LD), based on the commonly used correlation coefficient, r2, between single nucleotide polymorphisms (SNPs) selected from populations having African ancestry shows strong conservation across other non-African populations, but not vice versa [9]. This observation is likely the consequence of a major population bottleneck out of Africa. Using these LD-selected SNPs, we demonstrated a defined SNP haplotype structure that is highly conserved across all ethnic populations. Hence, a set of globally-applicable tagging SNPs could be feasibly defined. Two recent studies investigating haplotype/LD variation and the transferability of tagging SNPs across global populations have provided strong support for our observation [10], [11]. The conserved common haplotypes we defined clustered into a simple evolutionary structure of up to three “haplotype frameworks”. SNPs tagging such haplotype frameworks (fmSNPs) could generally be identified within defined LD blocks as the ones having the highest allele frequencies in African-ancestry populations. These allele-frequency-derived, ethnically-conserved frameworks were likely the ancestral haplotype backgrounds upon which more recent mutations have been superimposed. Interestingly, our preliminary analysis suggested that a significant portion of reported variants strongly associated with cis-regulation tagged these globally-conserved haplotype frameworks [9]. A conceptual illustration of ancestry-based haplotype clusters and the association with expression phenotypes is presented in Figure 1.10.1371/journal.pone.0003362.g001Figure 1Ancestry-based haplotype clusters and the association with expression phenotypes.(A) In this hypothetical example, five extant haplotypes are observed (1–5) within a chromosome segment showing strong LD (low recombination rate). These haplotypes are derived through five mutation steps (resulting in five SNPs in current populations) from the inferred ancestral sequence (boxed in black) and can be grouped into two major haplotype clusters (boxed in green and red). Separating the ancestry-based haplotype clusters are earlier mutation steps (G3 → A; C2 → T). Alleles of these SNPs can be applied for “tagging” the clusters (typed in green and red). Currently, ancestry is commonly inferred by either the allele frequencies of SNPs or the corresponding nucleotides in non-human primate species. When the frequency of SNP alleles is applied (preferably using those of African populations), the haplotype clusters are referred to as “haplotype frameworks” [9]. The SNPs tagging the frameworks are termed “framework SNPs” or “fmSNPs”. (B) Tree structure of the five extant haplotypes and the expression phenotype clusters. Given a simple hypothesis that an historical mutation creates a variant altering the expression phenotype (resulting either enhancing or suppressing expression), two alternative schemes of resulting phenotype clusters associated with the variant are illustrated. The left panel exemplifies an evolutionarily earlier expression alteration caused by a mutation tagging the ancestry-based haplotype clusters, and the right panel demonstrates the alteration caused by a more recent mutation (with the mutations boxed and the resulting expression phenotype clusters circled).In this report, we expanded the investigation of the relationship between cis-regulatory expression phenotypes and the SNPs tagging local haplotype frameworks (fmSNPs) by collecting and examining genes from major studies showing strong cis-regulatory association. We first delineated haplotype frameworks based on the high-density HapMap Phase II genotype data as described previously [9], followed by phylogenetic analysis among current major haplotypes and the corresponding haplotypes derived from chimpanzee reference sequences. We then measured the association between LD-derived tagging SNPs with expression phenotypes. As a consequence of this analysis, we observed significant correlation between SNPs showing the strongest association and SNPs tagging the major lineages directly separated from ancestry, inferred either from the frequency-derived haplotype frameworks (fmSNPs) or the chimpanzee reference sequences. We discuss the evolutionary implications of these findings for the origin and maintenance of expression variants in human populations, as well as for further genetic epidemiologic and pharmacogenetic studies.ResultsTo investigate the relationship between cis-regulatory expression phenotypes and the SNPs tagging local haplotype frameworks (fmSNPs), we analyzed a total of 26 genes (Table 1) showing stringently determined cis-association with expression phenotypes from five major studies (Morley et al. [12], Cheung et al. [13], Pastinen et al. [14], Deutsch et al. [15], Stranger et al. [16], reviewed by Pastinen et al. [17]). These studies were all conducted with lymphoblastoid cell lines of CEPH or HapMap CEU samples (Utah residents with ancestry from northern and western Europe), but using an earlier release (phase I) of HapMap genotype data having lower SNP density and employing different expression platforms. We first downloaded HapMap genotype data (Phase II; release 21) encompassing the gene pre-mRNA transcript and at least 10 kb upstream and downstream from initiation and termination sites, where predicted cis-regulatory modules (clusters of transcription factor binding sites) are most enriched [18]. We then constructed the local haplotype framework structure as previously described [9]. We used the originally-reported peak SNPs (SNPs showing strongest association) described in the above studies to serve as the lower-density screens (HapMap Phase I data) and then measured the association between gene expression and all tagging SNPs – this time taking advantage of the high-density HapMap Phase II data – within the block containing the reported peak SNPs. We applied public HapMap expression data across three major populations (GSE2552 [13] and GSE5859 [19], based on the Affymetrix platform, and GSE6536 [20], based on the Illumina platform).10.1371/journal.pone.0003362.t001Table 1Correlation between SNPs showing strongest association with cis-regulation and SNPs tagging haplotype frameworks.GeneGenomic region investigated1\nTagging SNP showing strongest association with cis-regulationReference numberAllele frequency2 in CEU/ YRI/ CHB+JPTRelative position to gene3\nTagging frequency-derived haplotype frameworksTagging lineages derived from chimpanzee referenceNominal p-value significant across multiple populations4\n\nClass I: SNPs showing strongest association with cis-regulation tag frequency-derived haplotype frameworks\n1HSD17B125\n,\n6\nchr11:43648880..43844743; 195.9 kbrs1083816226%/41%/20%intragenic\nYes\n\nYes\n✓2IRF55\n,\n6\nchr7:128161944..128194036; 32.1 kbrs228071442%/44%/47%dn 4.6 kb10\n\nYes\n\nYes\n✓3CD1517\nchr11:812985..838833; 25.9 kbrs407528926%/37%/5%up 2.3 kb10\n\nYes\n\nYes\n✓4CCT88\n,\n9\nchr21:29340517..29377880; 37.4 kbrs96595114%/29%/12%intragenic\nYes\n\nYes\n✓5PPAT5\n,\n6\nchr4:57090458..57152772; 62.3 kbrs968367929%/36%/29%intragenic\nYes\n\nYes\n✓6LOC3887966\nchr20:36472655..36507865; 35.2 kbrs37522789%/48%/13%intragenic\nYes\n\nYes\n✓7TMEM89\nchr16:351860..381907; 30.1 kbrs374388839%/29%/62%intragenic\nYes\n\nYes\n✓8CTBP15\nchr4:1185057..1242737; 57.7 kbrs375592046%/34%/70%up 0.7 kb10\n\nYes\n\nYes\n✓9ATF57\nchr19:55114271..55139002; 24.73 kbrs382677736%/18%/44%up 1.1 kb10\n\nYes\n\nYes\n✓10ARTS-17\nchr5:96112276..96179397; 67.1 kbrs3018730%/39%/45%intragenic\nYes\n\nYes\n11IL165\nchr15:79252254..79402155; 149.9 kbrs1163844425%/43%/2%intragenic\nYes\n\nYes\n12CTSH5\n,\n6\nchr15:76991161..77034474; 43.3 kbrs103693829%/86%/87%intragenic\nYes\nNo✓13CHI3L25\n,\n6\nchr1:111472322..111508101; 35.8 kbrs1204890037%/27%/13%up 4.2 kb10\n\nYes\nNo✓14VAMP85\nchr2:85706374..85730810; 24.4 kbrs373182838%/42%/33%intragenic\nYes\nNo✓\nClass II: SNPs showing strongest association with cis-regulation only tag lineages derived from chimp reference\n1BTN3A27\nchr6:26453120..26496524; 43.4 kbrs939371313%/3%/11%intragenicNo\nYes\n✓2SERPINB109\nchr18:59723724..59763455; 39.7 kbrs808549021%/81%/42%intragenicNo\nYes\n✓3LRAP5\n,\n6\nchr5:96231023..96319053; 88.0 kbrs224765049%/59%/56%intragenicNo\nYes\n✓4CAV29\nchr7:115669574..115742544; 73.0 kbrs1713876710%/1%/20%up 1.8 kb11\nNo\nYes\n✓5PAX87\nchr2:113679805..113762727; 82.9 kbrs1112317039%/28%/32%intragenicNo\nYes\n✓6CAT7\nchr11:34407053..34460178; 53.1 kbrs1083624413%/13%/54%intragenicNo\nYes\n✓7OAS17\nchr12:111797458..111830430; 33 kbrs185933640%/0%/27%dn 9.6 kb10\nNo\nYes\n\nOther:\n1RPS265\n,\n6\nchr12:54711952..54783960; 72.0 kbrs1117173938%/84%/27%dn 32.6 kb10\nNo13\nNo✓2CPNE15\n,\n6\n,\n9\nchr20:33667381..33726261; 58.9 kbrs1248040810%/8%/7%intragenicNo13\nNo✓3CSTB5\n,\n6\n,\n9\nchr21:44008259..44068882; 60.6 kbrs88098718%/1%/48%up 28.2 kb12\nNo13\nNo✓4RAB7L17\nchr1:202397009..202475780; 78.8 kbrs95136642%/16%/40%dn 52.3 kb10\nNo13\nNo✓5SFRS69\nchr20:41509931..41558894; 49.0 kbrs812481333%/1%/16%dn 13.2 kb10\nNo13\nNo1Including 10 kb upstream/downstream sequences of initiation/termination sites or an extended area to cover local LD block. Based on HapMap Phase II data release #21 in July 2006 and NCBI B35 assembly.2Frequency of rare allele derived in HapMap CEU Population versus the frequency of the same allele in YRI or CHB+JPT. CEU: CEPH (Utah residents with ancestry from northern and western Europe); YRI: Yoruba in Ibadan, Nigeria; CHB: Han Chinese in Beijing, China; JPT: Japanese in Tokyo, Japan.3Relative position (upstream; up / downstream; dn) to initiation/termination sites.4Based on public data from GSE 6536 (Illumina platform) or GSE 2552 / GSE 5859 (Affymetrix platform).5Reported in Morley et al., Nature 430, 743–7 (2004).6Reported in Cheung et al., Nature 437, 1365–9 (2005).7Reported in Pastinen et al., Hum. Mol. Genet. 14, 3963–71 (2005). Association data provided online from authors' website.8Reported in Deutsch et al., Hum. Mol. Genet. 14, 3741–9 (2005).9Reported in Stranger et al., PLoS Genet. 1, e78 (2005). Association data provided by the authors.10The LD block extends into intragenic region.11The LD block extends to upstream 0.8 kb.12The LD block extends to upstream 2.3 kb.13Haplotype framework-tagging SNP shows significant association in at least one population measured by either platform.A typical example, HSD17B1, is depicted in Figure 2, in which the intragenic SNP, rs4755741, was reported as the peak SNP [13]. To delineate the local haplotype framework structure from the YRI population and compare it to that of other major populations, we downloaded genotypes of HapMap SNPs encompassing a total of 195.9 kb, including 10 kb upstream of the transcription initiation site and 10 kb downstream of the termination site. We then selected SNPs in strong LD (r2>0.8) against at least one other SNP conserved across populations and inferred major haplotypes (>5%; see Panel B) within the block containing the peak SNP (the labeled triangular area in the LD plot in Panel A). To simplify the presentation in Figure 2, we show only SNPs with rare allele frequencies greater than 20% in either population (additional SNPs do not alter the primary result we obtained). These major haplotypes clustered into two frameworks, A and B, tagged by a set of fmSNPs having the highest allele frequency within the block (common and rare alleles are colored in green and red, respectively). Major haplotypes within each framework can be further tagged by other SNPs having lower allele frequencies (rare alleles colored in purple). For this gene, as well as many others (genes 1 to 14 in Table 1; also see Supporting Information Figures S1 for detailed analyses), the fmSNPs showed the strongest association with the expression phenotype. We designated these genes (14 of 26) as class I in Table 1.10.1371/journal.pone.0003362.g002Figure 2Delineation of underlying haplotype framework structure encompassing the HSD17B12 gene (Class I cis-regulatory gene).(A) Diagram depicting the HSD17B12 gene and its chromosomal position (reproduced from the HapMap graphical browser), aligned with the local LD structure determined in YRI (output from the Haploview program) using LD-selected SNPs. For simplifying this presentation, we focused on common SNPs (frequency >0.2) in either the HapMap YRI or CEU populations. Pairwise calculation of standardized LD, r2, was first determined using YRI data. SNPs in strong LD (r2>0.8) with at least one other SNP and also exhibiting conserved LD in CEU and CHB/JPT were selected for the LD plot and haplotype analyses. The original SNP reported to show the strongest association with expression (peak SNP) is marked with a solid black triangle at its physical position and mapped to its corresponding position in the LD plot. The LD block containing the peak SNP is surrounded with black lines. (B) Haplotype frameworks within the block containing the peak SNP. The major haplotypes (>5% in either population) and their population frequencies were inferred using the Haploview program. Five major haplotypes in the YRI population clustered into two haplotype frameworks (A and B) that can be tagged by a set of SNPs (fmSNPs) in strong LD and having the highest allele frequency within the block. The common alleles of fmSNPs are colored green, and the rare alleles red. The rare alleles of other lower-frequency SNPs are colored purple. Comparison of major haplotypes delineated in CEU and CHB/JPT using the same sets of SNPs showed an identical haplotype structure with a different frequency distribution as shown to the right. (Four SNPs having no genotype information in CEU were left blank.) All SNP reference (rs) numbers are shown above, with the original reported peak SNP, rs4755741, outlined in black. The chimpanzee nucleotides corresponding to each SNP are shown below. The colors of SNP alleles used in CEU, CHB/JPT, and chimpanzee follow the convention defined in YRI. The stars below chimpanzee nucleotides indicate polymorphisms located at (C/T)pG positions on either strand.In addition to the peak SNP-fmSNP correlation, we observed that a few SNPs showing the strongest association, despite having no correlation with fmSNPs, exhibited a unique characteristic: namely, that of being in strong LD against a relatively large number (the vast majority) of other SNPs within the LD block. As shown in Figure 3, at BTN3A2, using the same SNP selection criterion of pairwise LD (r2>0.8), the major haplotypes within the block were delineated (Panel B). The heritable, unidirectional allelic imbalance and the regulatory haplotype of this gene were also discussed in Pastinen et al. [21]. Although the reported peak SNP, rs9379851, was not in strong LD against frequency-derived fmSNPs (r2 = 0.04, 0.21, 0.40 in YRI, CEU, and CHB/JPT populations, respectively), it was highly correlated with many other SNPs tagging the haplotype B4 within the 24 kb LD block (3%, 12%, 9% frequency in YRI, CEU, and CHB/JPT, respectively).10.1371/journal.pone.0003362.g003Figure 3Delineation of underlying haplotype framework structure encompassing the BTN3A2 gene (Class II cis-regulatory gene).(A) Diagram depicting the BTN3A2 gene, its chromosomal position, and the local LD structure. This panel follows the convention in Figure 2 except that, for simplifying the presentation, we focused on common SNPs with frequency >0.1 in HapMap populations. (B) Haplotype frameworks within the block containing the peak SNP. This panel also follows the convention in Figure 2. The pairwise LD measure, r2, between the peak SNP and fmSNP is shown in all three populations. Sets of SNPs in strong LD, determined using YRI genotypes and based on the criterion of r2>0.8, are depicted at the bottom. The number of SNPs in each bin is shown to the left. The SNP set marked in red, containing an extraordinarily large number of SNPs relative to other bins (tagging haplotype B4 within the block), shows the strongest association with expression phenotypes.To date, most genetical genomics studies are solely based on association tests between individual SNPs and expression phenotypes. One advantage of our haplotype-based approach is its capability of incorporating evolutionary analysis. Currently, there are two general approaches for inferring ancestry — one is based on the frequency of SNP alleles and the other on the comparison of corresponding nucleotides in species closely related to human beings, e.g., chimpanzees. Independent studies have reported that there was a general agreement between the two approaches [22], [23]. The more common human allele generally matches the corresponding nucleotide in the chimpanzee genome (76% concordance as reported in Hacia et al. [22]). Given the conservation of the haplotype frameworks defined by fmSNPs across other out-of-Africa populations [9], these frameworks are likely haplotype backgrounds upon which more recent mutations having lower allele frequencies have been superimposed. Since all genes selected for this study are reported cis-regulatory genes, we considered whether our observed correlation between fmSNPs and SNPs showing the strongest association with expression differences was a consequence of selection in earlier human history and whether genes behaving like BTN3A2 were under more recent selection in the African population, resulting in population-specific frequency distortion.We subsequently conducted evolutionary analysis of the common haplotypes across all 26 genes (Supporting Information Figures S1). For all the SNPs employed in our haplotype construction, we mapped the corresponding chimpanzee nucleotides using chimpanzee reference sequences, followed by median-joining (MJ) network analysis to derive phylogenetic relationships among all major haplotypes. As shown in Figure 4A, at HSD17B12, the two frequency-derived haplotype frameworks (A and B) were separated directly from the ancestral haplotype. In addition, we performed coalescent-based likelihood analysis to draw the maximum likelihood genealogical relationships among the common haplotypes. The result also supported the hypothesis that underlying these haplotype frameworks were older mutations closer to the root of the gene tree. Thus, the differential expression pattern of these haplotypes was likely to have appeared early in human evolutionary history.10.1371/journal.pone.0003362.g004Figure 4Phylogenetic relationships among current major haplotypes and the association to expression phenotypes.Median-joining (MJ) network analysis was conducted using the Network program for HSD17B12 (Class I; shown in panel A) and BTN3A2 (Class II; shown in panel B). The major haplotypes in HapMap populations shown in Figure 2 and 3 were entered, using their population haplotype frequencies, along with the corresponding chimpanzee haplotype. The SNPs located at (C/T)pG positions on either strand (marked with stars in panel B of the previous figures) were generally excluded from this analysis because of their potentially high mutation rates. Each haplotype is represented by a circle. The area of each circle, except for the chimpanzee reference (colored yellow), reflects the observed frequency of each haplotype in the total dataset (YRI, CEU, CHB/JPT). The portion of YRI, CEU, and CHB/JPT chromosomes in each circle is denoted with green/red, white, and grey colors, respectively. The colors of haplotype frameworks A and B follow the green and red convention in previous figures. The length of lines between any two haplotype nodes is proportional to the number of mutation steps. The rs numbers of SNPs are labeled along the lines. The median vector (mv) is shown as a small circle and can be interpreted as possibly extant unsampled haplotypes. For each set of SNPs in strong LD (generally marking the branches in the lineage), a SNP with the most complete genotypes (underlined) was chosen for testing association. The nominal p-values for these SNPs in each major population, based on expression data sets GSE6536 (Illumina platform) and GSE2552/GSE5859 (Affymetrix platform), are shown in order. The coalescent-based maximum likelihood tree structure and the regression of expression phenotypes are plotted at the bottom of each panel. The set of SNPs in LD showing the strongest association with expression phenotypes is typed in red in the phylogenetic network and shown as red dots in the genealogical tree.For genes showing the features exemplified by BTN3A2, we observed a weaker correlation between the more common allele and the chimpanzee nucleotide, i.e., a more significant proportion of the rare alleles matched the ancestral nucleotides. Phylogenetic analysis, typified by the results of Figure 4B, suggested that the ancestral haplotype was between the haplotype, B4, carrying a relatively large number of tagging SNPs in strong LD, including the peak SNP, and the rest of the B haplotypes. Similarly, the coalescent-based maximum likelihood tree structure also suggested that the accumulation of such a long stretch of B4-tagging SNPs likely occurred early in the tree. Hence, for such genes, the differential expression pattern also evolved early, except that the frequency of the cis-regulatory haplotype showing differential expression was often lower in current populations having African ancestry, presumably a consequence of more recent population-specific selection. We designated these genes (following only the chimpanzee-inferred ancestry) as Class II (7 genes out of 26) in Table I. Five of them showed the characteristic of carrying a relatively large number of tagging SNPs in strong LD (BTN3A2, SERPINB10, LRAP, CAV2, OAS1).Since the majority of our sampled genes exhibited the same evolutionarily conserved feature, we next asked whether their cis-regulatory phenotypes were, as expected, also conserved across populations. Based on one set of expression data in YRI (GSE6536, Illumina platform) and two sets data in CHB/JPT (GSE5859, Affymetrix platform; GSE6536, Illumina platform), we tested the association for all tagging SNPs (Figure 4 and Supporting Information Figures S1). Only three genes in Class I and II did not show significant association in at least one other population (Table 1). In addition, for all genes showing significant cis-association across multiple ethnic populations, the direction of allelic effect on expression was always consistent, strongly supporting the hypothesis that the cis-regulation was derived early and still being maintained in current populations.Overall, among the 26 cis-regulatory loci we analyzed, frequency-derived fmSNPs showed strongest association with expression phenotypes for 14 genes (Class I), 11 of which also demonstrated the chimpanzee-inferred ancestry. Seven genes (Class II) followed only the chimpanzee-inferred ancestry, but not the frequency-derived haplotype frameworks. A total of 80% (21/26) of genes followed either frequency- or chimpanzee- inferred ancestry. To determine if this distribution of cis-regulatory loci could be the result of chance, we performed simulations using LD-selected common SNPs within the analyzed LD blocks of the same 26 genes. Assuming that every SNP had the same probability to be the cis-regulatory variant, our simulations, under a completely random-occurrence scenario, resulted in an average of only 13 genes showing an association of a cis-regulatory variant with the SNPs tagging the major branches separated from ancestry. When compared to the total of 21 genes actually observed, our simulation resulted in a significant deviation (p = 10−6). Therefore, we rejected the null hypothesis of randomness. We concluded that, in the 26 genes, there was a higher probability of SNPs tagging the major lineages separated from ancestry to be cis-regulatory variants. Also of note, for the five genes showing no correlation with either ancestry inference, the frequency-derived fmSNPs of these five genes all showed individually significant association in at least one population, measured by at least one platform.DiscussionEarlier linkage studies have shown that quantitative gene expression levels are significantly heritable [12], [24]. Although both cis- and trans-linkages have been detected, one interesting observation has been the enrichment of cis-linkages among the strongest signals, a phenomenon also observed in mice and rats [25]–[27]. Recent genetical genomics studies based on whole genome association tests have also revealed that a majority of signals for differential expression are cis-acting [13], [16]. Overall, current data suggest that cis-regulatory effects are more consistent and larger. In contrast, trans-acting signals are more modestly significant and often are not replicated (reviewed in [7], [17]).Since our current knowledge of trans-acting regulation may still be insufficient for comprehensive association studies [17], an adequate approach at this stage would be to focus on the identification of cis-regulatory genes that are heritable as a monogenic trait. Currently, most genetical genomics studies searching for cis-regulatory genes are based on association tests between individual SNPs and expression phenotypes. However, while the SNP density employed in the commercially available, high-throughput platforms keeps growing, the major trade-off is true associations failing to pass the stringent statistical correction for multiple testing. Our analysis indicates that, since the vast majority of true cis-regulatory genes carry evolutionarily common signatures, the use of such signatures (fmSNPs for class I genes and subhaplotypes with many SNPs in high LD for class II genes) should provide more effective identification of true positives. Also, since recent major studies have only focused on a limited number of expressed genes in lymphoblastoid cell lines, learning the common genetic characteristics of identified cis-regulatory genes from these studies should help future identification of other globally-conserved cis-regulatory genes across different tissues.Genetical genomics studies, often based on different platforms with different experimental designs, have in the past shown poor correlation between studies [17]. Examples are given at LRAP and SFRS6 (Supporting Information Figures S1), where an apparent discrepancy across the two major commercial platforms, Illumina and Affymetrix, is shown. In the case of SFRS6, different 50mer probes used in the Illumina platform also produced a discrepancy, probably a consequence of different probes recognizing alternative transcripts. Other questions regarding statistical analysis and cell line variability have also arisen, leading to warnings to interpret results with caution [7], [28], [29]. We would like to note, however, that our observations were based on a collection of cis-regulatory genes from independent studies, conducted in different laboratories using different approaches, but confirmed using an independent dataset with a larger sample size. Although the number of genes we collected is limited in this study, we nonetheless observed common genetic features of these cis-regulatory genes that could be applied to a significant fraction of genes analyzed (21/26 in which the reported associations could be replicated). While it is possible that our observation was only a result of enrichment of a specific profile of cis-regulation using the top association hits from different studies, other examples fitting our observation have independently appeared in recent literature, for example, the clustering of VKORC1 and NPY haplotypes based on their expression phenotypes and their correlation to drug and stress response [5], [30]. This suggests that these features may be a general and powerful means of discovering evolutionarily-conserved variants of gene expression. Since the variants were generally common in current populations, they will likely prove useful for validating expression differences across multiple tissues, populations and enhancing our understanding of the differential predisposition to common diseases and variation in drug responses in different ethnic groups.Recent surveys have shown that many gene coding regions in the human genome do not show an excess of low-frequency alleles, suggesting that balancing selection might be more common than previously thought (reviewed by Bamshad et al. [31]). Our analyses also revealed that haplotypes in current populations carrying high- and low-expression phenotypes were nearly exclusively evolved early in human evolutionary history (Figure 4), likely as a consequence of balancing selection. Therefore, disease gene variation taking the form of cis-acting eQTLs may have a narrower allelic spectrum toward high population frequencies, as predicted by the common diseases/common variant (CDCV) model that genetic risk of common diseases is often conferred by alleles having relatively high frequencies [32].Methods\ncis-regulatory genes included in this studyAs shown in Supporting Information Table 1, a total of 44 genes showing stringently determined cis-association with expression phenotypes were initially collected from five major studies [12]–[16]. These studies were all conducted with lymphoblastoid cell lines of CEPH or HapMap CEU samples (Utah residents with ancestry from northern and western Europe) using an earlier release (phase I) of HapMap genotype data, but employing different expression platforms. Although the majority of these genes demonstrated prior positive results for linkage or allelic imbalance (AI) assays, we added a further validation step by confirming the cis-association using an independent dataset having a relatively large sample size [33] (GSE8052; Affymetrix platform; 400 UK samples). Thirty of the 44 genes passed the genome-wide significance threshold (a LOD score of 6.076, corresponding to a false discovery rate of 0.05, as listed in supplementary table 1 in Dixon et al. [33]). Of the thirty genes, four genes were excluded from our analysis (GSTM1 and GSTM2: known region of structure variation [deletion]; PSPHL: probe 205048_s_at mapped to a region having no annotated gene in the b35 assembly; POMZP3: HapMap SNP density too low for our haplotype analysis). Overall, we included 26 genes for our haplotype and cis-association analysis listed in Table 1.SNP selection for delineating haplotype framework structureSNP data from HapMap release 21/Phase II in July, 2006, based on NCBI b35 assembly and dbSNP b125, were downloaded using the graphical browser provided by the International HapMap Project (http://www.hapmap.org/). For regions encompassing at least 10 kb upstream and downstream from initiation and termination sites of the pre-mRNA transcript, genotypes (forward strand) of 60 YRI (Yorubans of Ibadan, Nigeria), 60 CEU (Utah residents with ancestry from northern and western Europe) individuals (parents of family trios) and 90 CHB/JPT (Han Chinese in Beijing, China, and Japanese in Tokyo, Japan) were employed for LD-based SNP selection and local haplotype framework analyses. For delineating major haplotypes with frequencies greater than 5% in current populations, SNPs having rare allele frequencies greater than 5% in any population were screened first (unless otherwise noted in the figures), followed by selection using the pairwise LD measure [34], r2, for those in strong LD against at least one other SNP (based on the criterion of r2>0.8). SNPs showing conserved LD behavior across populations were employed in haplotype construction, as described in our previous publication [9].Delineation of LD and haplotype framework structureFor each analyzed region, the LD plotting, haplotype block partitioning, and the delineation and population frequency estimation of major haplotypes were performed by the HAPLOVIEW program, version 3.32 (http://www.broad.mit.edu/mpg/haploview) [35]. The haplotype block partitioning was generally determined with YRI data using one of the three methods (confidence intervals, four gamete rule, solid spine of LD) incorporated into the HAPLOVIEW program, depending on which covered the most extensive area containing the peak SNP. For some genes, we covered more extensive regions to increase informativeness. The haplotype frameworks were clustered based on YRI allele frequencies, as described in our previous publication [9].Genealogical analysisPhylogenetic relationships among major haplotypes were analyzed by the Median Joining (MJ) network algorithm packaged in the NETWORK program, version 4.201 (http://www.fluxus-engineering.com/sharenet.htm) [36]. The major haplotypes in either population, along with chimpanzee haplotypes, were entered with their population haplotype frequencies. The chimpanzee haplotypes were derived using corresponding nucleotides in the chimpanzee reference sequences, retrieved using the UCSC genome browser (http://genome.ucsc.edu/) [37]. SNPs located at (C/T)pG positions on either strand, because of their higher mutation rate, were generally excluded from this analysis.Coalescent-based genealogical analysis was performed by the GENETREE program version 9.0 (http://www.stats.ox.ac.uk/griff/software.html) [38]. It applies the Markov chain simulation to perform likelihood estimates of tree probabilities under the infinite site model. The major haplotypes in the three populations (denoted as subpopulations) at HSD17B12 and BTN3A2 (shown in Figure 4) were entered using their population haplotype frequencies. The chimpanzee corresponding alleles of the polymorphic sites were designated as ancestral alleles.Association analysisLD-derived SNP bins were defined from SNPs within the block containing the originally reported peak SNPs using the TAGGER program (“tagger pairwise” option) incorporated into the HapMap graphical browser [39]. For each bin (generally marking the branches in the genealogical analysis), a tagging SNP was selected based on the completeness of genotypes across the three populations for testing association. Association analysis between each tagging SNP and two sets of HapMap expression data, based on two (Affymetrix and Illumina) platforms and across three HapMap populations, (GEO accession number GSE2552 [13], GSE5859 [19], and GSE6536 [20]), was conducted by following the regression methods described in Cheung et al. [13] (discussed in [40]). The nominal p-value of each tagging SNP was used for the determination of SNPs showing strongest association. The value of 0.05 was used as our cutoff for statistical significance.Monte-Carlo-based simulationWe tested the hypothesis whether cis-regulatory SNPs were randomly distributed along the genealogical tree versus an alternative that there was an enrichment or selection effect among the 26 genes in Table 1. We performed a Monte-Carlo-based simulation under the assumption that every common SNP has the same probability to be the cis-regulatory variant. For each gene, we randomly selected an LD-selected common SNP (as shown in all figures) under a binomial distribution. The number of trials was the count of common SNPs and the probability of a cis-regulatory variant was the number of SNPs tagging the major lineages separated from ancestry divided by the total number of LD-selected common SNPs across the investigated LD blocks of all genes. Our test statistic compared the average number of genes found in a series of 1,000,000 simulations, versus the observed 21 genes.Supporting InformationSupporting Information Figures S1(7.15 MB PDF)Click here for additional data file.\n\nREFERENCES:\n1. LoHSWangZHuYYangHHGereS\n2003\nAllelic variation in gene expression is common in the human genome.\nGenome Res\n13\n1855\n1862\n12902379\n2. PantPVTaoHBeilharzEJBallingerDGCoxDR\n2006\nAnalysis of allelic differential expression in human white blood cells.\nGenome Res\n16\n331\n339\n16467561\n3. PastinenTHudsonTJ\n2004\nCis-acting regulatory variation in the human genome.\nScience\n306\n647\n650\n15499010\n4. YanHDobbieZGruberSBMarkowitzSRomansK\n2002\nSmall changes in expression affect predisposition to tumorigenesis.\nNat Genet\n30\n25\n26\n11743581\n5. RiederMJReinerAPGageBFNickersonDAEbyCS\n2005\nEffect of VKORC1 haplotypes on transcriptional regulation and warfarin dose.\nN Engl J Med\n352\n2285\n2293\n15930419\n6. JansenRCNapJP\n2001\nGenetical genomics: the added value from segregation.\nTrends Genet\n17\n388\n391\n11418218\n7. de KoningDJHaleyCS\n2005\nGenetical genomics in humans and model organisms.\nTrends Genet\n21\n377\n381\n15908034\n8. GibsonGWeirB\n2005\nThe quantitative genetics of transcription.\nTrends Genet\n21\n616\n623\n16154229\n9. OuyangCKrontirisTG\n2006\nIdentification and functional significance of SNPs underlying conserved haplotype frameworks across ethnic populations.\nPharmacogenet Genomics\n16\n667\n682\n16906021\n10. ConradDFJakobssonMCoopGWenXWallJD\n2006\nA worldwide survey of haplotype variation and linkage disequilibrium in the human genome.\nNat Genet\n38\n1251\n1260\n17057719\n11. de BakkerPIBurttNPGrahamRRGuiducciCYelenskyR\n2006\nTransferability of tag SNPs in genetic association studies in multiple populations.\nNat Genet\n38\n1298\n1303\n17057720\n12. MorleyMMolonyCMWeberTMDevlinJLEwensKG\n2004\nGenetic analysis of genome-wide variation in human gene expression.\nNature\n430\n743\n747\n15269782\n13. CheungVGSpielmanRSEwensKGWeberTMMorleyM\n2005\nMapping determinants of human gene expression by regional and genome-wide association.\nNature\n437\n1365\n1369\n16251966\n14. PastinenTGeBGurdSGaudinTDoreC\n2005\nMapping common regulatory variants to human haplotypes.\nHum Mol Genet\n14\n3963\n3971\n16301213\n15. DeutschSLyleRDermitzakisETAttarHSubrahmanyanL\n2005\nGene expression variation and expression quantitative trait mapping of human chromosome 21 genes.\nHum Mol Genet\n14\n3741\n3749\n16251198\n16. StrangerBEForrestMSClarkAGMinichielloMJDeutschS\n2005\nGenome-wide associations of gene expression variation in humans.\nPLoS Genet\n1\ne78\n16362079\n17. PastinenTGeBHudsonTJ\n2006\nInfluence of human genome polymorphism on gene expression.\nHum Mol Genet\n15 Spec No 1\nR9\n16\n16651375\n18. BlanchetteMBatailleARChenXPoitrasCLaganiereJ\n2006\nGenome-wide computational prediction of transcriptional regulatory modules reveals new insights into human gene expression.\nGenome Res\n16\n656\n668\n16606704\n19. SpielmanRSBastoneLABurdickJTMorleyMEwensWJ\n2007\nCommon genetic variants account for differences in gene expression among ethnic groups.\nNat Genet\n39\n226\n231\n17206142\n20. StrangerBENicaACForrestMSDimasABirdCP\n2007\nPopulation genomics of human gene expression.\nNat Genet\n39\n1217\n1224\n17873874\n21. PastinenTSladekRGurdSSammakAGeB\n2004\nA survey of genetic and epigenetic variation affecting human gene expression.\nPhysiol Genomics\n16\n184\n193\n14583597\n22. HaciaJGFanJBRyderOJinLEdgemonK\n1999\nDetermination of ancestral alleles for human single-nucleotide polymorphisms using high-density oligonucleotide arrays.\nNat Genet\n22\n164\n167\n10369258\n23. StephensJCSchneiderJATanguayDAChoiJAcharyaT\n2001\nHaplotype variation and linkage disequilibrium in 313 human genes.\nScience\n293\n489\n493\n11452081\n24. MonksSALeonardsonAZhuHCundiffPPietrusiakP\n2004\nGenetic inheritance of gene expression in human cell lines.\nAm J Hum Genet\n75\n1094\n1105\n15514893\n25. SchadtEEMonksSADrakeTALusisAJCheN\n2003\nGenetics of gene expression surveyed in maize, mouse and man.\nNature\n422\n297\n302\n12646919\n26. HubnerNWallaceCAZimdahlHPetrettoESchulzH\n2005\nIntegrated transcriptional profiling and linkage analysis for identification of genes underlying disease.\nNat Genet\n37\n243\n253\n15711544\n27. DossSSchadtEEDrakeTALusisAJ\n2005\nCis-acting expression quantitative trait loci in mice.\nGenome Res\n15\n681\n691\n15837804\n28. WilliamsRBCotsapasCJCowleyMJChanENottDJ\n2006\nNormalization procedures and detection of linkage signal in genetical-genomics experiments.\nNat Genet\n38\n855\n856\n16874319\n29. AkeyJMBiswasSLeekJTStoreyJD\n2007\nOn the design and analysis of gene expression studies in human populations.\nNat Genet\n39\n807\n808\n17597765\n30. ZhouZZhuGHaririAREnochMAScottD\n2008\nGenetic variation in human NPY expression affects stress response and emotion.\nNature\n452\n997\n1001\n18385673\n31. BamshadMWoodingSP\n2003\nSignatures of natural selection in the human genome.\nNat Rev Genet\n4\n99\n111\n12560807\n32. LanderES\n1996\nThe new genomics: global views of biology.\nScience\n274\n536\n539\n8928008\n33. DixonALLiangLMoffattMFChenWHeathS\n2007\nA genome-wide association study of global gene expression.\nNat Genet\n39\n1202\n1207\n17873877\n34. DevlinBRischN\n1995\nA comparison of linkage disequilibrium measures for fine-scale mapping.\nGenomics\n29\n311\n322\n8666377\n35. BarrettJCFryBMallerJDalyMJ\n2005\nHaploview: analysis and visualization of LD and haplotype maps.\nBioinformatics\n21\n263\n265\n15297300\n36. BandeltHJForsterPRohlA\n1999\nMedian-joining networks for inferring intraspecific phylogenies.\nMol Biol Evol\n16\n37\n48\n10331250\n37. KentWJSugnetCWFureyTSRoskinKMPringleTH\n2002\nThe human genome browser at UCSC.\nGenome Res\n12\n996\n1006\n12045153\n38. GriffithsRCTavareS\n1994\nAncestral Inference in Population Genetics.\nStat Sci\n9\n307\n319\n39. de BakkerPIYelenskyRPe'erIGabrielSBDalyMJ\n2005\nEfficiency and power in genetic association studies.\nNat Genet\n37\n1217\n1223\n16244653\n40. SpielmanRSCheungVG\n2007\nReply to “On the design and analysis of gene expression studies in human populations”.\nNat Genet\n39\n808\n809"
+}

batch_13/PMC2557512.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2557512",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2557512\nAUTHORS: B W Hancock\n\nABSTRACT:\nNo Abstract\n\nBODY:\nClinical presentation and diagnosisRenal cell carcinoma is by far the commonest cancer in this group. Presentation may be with local symptoms. Thus haematuria, usually painless though sometimes with colicky pain secondary to blood clots, is common. Less often, a swelling or aching in the loin may be noticed by the patient. Often and increasingly the tumour is picked up coincidentally during an imaging investigation (e.g., ultrasound scan of the abdomen). Metastatic disease may be evident, in up to a third of patients, as bone, lung or brain secondaries or by systemic features such as weight loss, fever, malaise and anaemia. Such features may also be present with organ-confined disease.TreatmentSurgery is the cornerstone of treatment. Nephrectomy is potentially curative for tumours confined to the kidney and/or regional nodes and is of value in reducing tumour bulk in palliating advanced disease. With metatastic disease radiotherapy may help in palliating local symptoms. Conventional chemotherapy has little value and although immunotherapy with interferon-α or interleukin-2 does marginally improve progression-free and overall survival, it is at the cost of marked toxicity. However international efforts continue to develop ways of identifying the small minority of patients who will benefit greatly from immunotherapy. More recently knowledge of molecular pathways involved, particularly in clear cell renal carcinoma, has provided a rationale for targeted treatment. Agents have been identified, which target gene products, such as vascular endothelial growth factor (VEGF) and platelet derived growth factor (PDGF), implicated in tumourigenesis. Such agents have demonstrated significant activity in clinical trials and represent a major advance in treatment (Brugarolas, 2007).Incidence and survival patternsOnly a small number of kidney cancers have an inherited basis, as for example von Hippel–Lindau (VHL) disease where the cancer risk for a variety of tumours including renal carcinoma is transmitted in an autosomal dominant manner. It is now known that the VHL suppressor gene is mutated or silenced in most clear cell renal carcinomas though the initiator(s) of this event are still unknown. Information on risk factors for renal cell carcinoma has come from a large number of international case control and cohort studies (McLaughlin et al, 2006). Cigarette smoking and obesity are the most consistently established causal factors, smoking having a greater effect in men, obesity possibly in women. It is important to take these factors into account when interpreting incidence trends. Across the community obesity is increasing and smoking decreasing. From the data presented by Westlake et al (2008), there has been an improvement in relative survival over the period 1986–1999. Most likely this reflects general improvements in health care, and possibly earlier diagnosis, rather than in specific therapies – the 1980s and 1990s were the decades of biological therapy where early promise of improvements in survival never came to significant fruition. Deprivation still seems to be an important factor in prognosis in men, but less so in women; the reasons for this are unclear. Certainly these trends cannot be related to known changes in smoking and obesity but could be linked to changes over time in health care and earlier diagnosis across the socioeconomic spectrum.Future clinical implicationsIt is disappointing that survival figures for English patients (up to mid 1990s), were among the lowest in Europe but reassuring that improvements have particularly been seen during the late 1990s. All patients now have the benefit of having their case discussed by a multidisciplinary team including specialist urology and oncology clinicians, with access to optimal treatments and also information on clinical trials. For example, a new randomised study (SORCE) assessing the function of adjuvant sorafenib (one of the new generation of VEGF inhibitors) in high-risk resected renal cancer has recently been launched. For metastatic disease sunitinib and sorafenib (VEGF inhibitors) have now been licensed in the United Kingdom on the basis of evidence from large randomised controlled clinical trials (Motzer et al, 2007; Escudier et al, 2007). These agents and others are not yet generally available in the National Health Service and are currently being appraised by the National Institute for Clinical Excellence. There is every hope, therefore, that the improvements in survival reported in the accompanying article will continue into and beyond the first decade of the 21st century. In addition there is considerable interest in prognostic models (based on clinical and laboratory data) that have been constructed to provide meaningful risk stratification for clinical trials (Shuch et al, 2006), the results of which can then be extrapolated to tumour registry survival data and help in our understanding of the factors that influence trends and socioeconomic inequalities in survival.\n\nREFERENCES:\n1. Brugarolas J (2007) Renal cell carcinoma – molecular pathways and therapies. N Engl J Med\n356: 185–18717215538\n2. Escudier B, Eisen T, Stadler WM, Szczylik C, Oudard S, Siebels M, Negrier S, Chevreau C, Solska E, Desai AA, Rolland F, Demkow T, Hutson TE, Gore M, Freeman S, Schwartz B, Shan M, Simantov R, Bukowski RM, TARGET Study Group (2007) Sorafenib in advanced clear-cell renal-cell carcinoma. N Engl J Med\n356: 125–13417215530\n3. McLaughlin JK, Lipworth L, Tarone RE (2006) Epidemiologic aspects of renal cell carcinoma. Semin Oncol\n33: 527–53317045081\n4. Motzer RJ, Hutson TE, Tomczak P, Michaelson MD, Bukowski RM, Rixe O, Oudard S, Negrier S, Szczylik C, Kim ST, Chen I, Bycott PW, Baum CM, Figlin RA (2007) Sunitinib versus interferon alfa in metastatic renal-cell carcinoma. N Eng J Med\n356: 115–124\n5. Shuch BM, Lam JS, Belldegrun AS, Figlin RA (2006) Prognostic factors in renal cell carcinoma. Semin Oncol\n33: 563–57517045085\n6. Westlake S, Cooper N, Rachet B, Coleman MP (2008) Survival from cancers of the kidney and ureter in England and Wales up to 2001. Br J Cancer\n99(Suppl 1): S93–S9518813274"
+}

batch_13/PMC2559883.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2559883",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2559883\nAUTHORS: Chieh-Chun Chen, Sheng Zhong\n\nABSTRACT:\nBackgroundTo date, the reconstruction of gene regulatory networks from gene expression data has primarily relied on the correlation between the expression of transcription regulators and that of target genes.ResultsWe developed a network reconstruction method based on quantities that are closely related to the biophysical properties of TF-TF interaction, TF-DNA binding and transcriptional activation and repression. The Network-Identifier method utilized a thermodynamic model for gene regulation to infer regulatory relationships from multiple time course gene expression datasets. Applied to five datasets of differentiating embryonic stem cells, Network-Identifier identified a gene regulatory network among 87 transcription regulator genes. This network suggests that Oct4, Sox2 and Klf4 indirectly repress lineage specific differentiation genes by activating transcriptional repressors of Ctbp2, Rest and Mtf2.\n\nBODY:\nBackgroundTranscriptional control is a key regulatory mechanism for cells to direct their destinies. A large number of transcription factors (TFs) could simultaneously bind to a regulatory sequence. With the constellation of TFs bound, the expression level of a target gene is usually determined by the combinatorial control of a number of TFs. The interactions among regulatory proteins and their regulatory sequences collectively form a regulatory network. A major challenge in the study of gene regulation is to identify the interaction relationships within a regulatory network.A number of analytical methods have been proposed to reconstruct gene regulatory networks from gene expression and protein-DNA binding data. Association rule mining [1], Boolean Network [2], temporal models [3,4], ARACNE [5] and Bayesian networks [6-8] are among the most popular routes. For example, the Module Networks approach built a probabilistic model for the gene expression correlations between regulators and target genes and iteratively searched for the most compatible partition of targets genes to their respective regulators [9]. The correlation of gene expression patterns of regulators and the target genes is often the essential piece of information utilized by the current procedures. It is widely recognized that the statistical correlation of the regulators and the targets is often an inaccurate representation of the regulator-target relationship [10,11]. This is because the quantity of a TF's mRNA does not necessarily correlate to its active protein concentration, and even the active protein concentration does not necessarily correlate to its transcriptional efficiency on every target gene. Using correlation, or some transformed version of correlation measure as the basis for reconstructing regulatory networks is an approximation made for convenience of modeling and analysis, with a sacrifice of making spurious findings (see examples in [9]). A network reconstruction method based on quantities that closely represent the biophysical properties of TF-DNA binding, transcription activation and repression is still missing.Thermodynamic models of TF-DNA and TF-RNA polymerase (RNAP) interactions were pioneered by Buchler et al. [12] in prokaryotes. These models brought the stochastic interactions of TFs, regulatory sequences and RNAP into a statistical mechanics framework, and enabled a quantitative model for the transcription rate. Recently, Segal et al. [13] and our team [14] attempted to employ thermodynamic models in the study of eukaryotic gene regulation. Under a fix time point in drosophila development, Segal et al. demonstrated a thermodynamic model could predict the spatial expression patterns of segmentation genes in Drosophila [11]. In differentiating embryonic stem cells (ESCs), we showed that the interaction types of the TFs could be predicted from the temporal response of the target gene [14]. These successes made it tempting to experiment novel methods for reconstructing regulatory networks based on more biophysically appropriate metrics than correlation.We describe here a computational framework, called Network-Identifier, for inferring regulatory networks from time course gene expression data. The gene expression values at each time point are supposed to be at an equilibrium state, which is a general setting for most of time course data available. Applying to the analysis of five datasets of differentiation of murine ESCs, we identified a transcription network composed of 34 TF-TF interactions and 185 TF-target relationships. Data from RNAi [15] and chromatin immunoprecipitation coupled with microarray (ChIP-chip) data [16,17] independently validated a statistically highly significant fraction of these regulatory relationships.ResultsGene regulatory network in mouse ESCsESCs are derived from early mammalian embryos. ESCs possess two important characteristics that distinguish their importance in scientific and medical fields. First, they are capable of self-renewal through apparently unlimited, undifferentiated proliferation in cultured cell lines [18-20]; second, they have remarkable pluripotency potentials [21] to give rise to many different cell types in the body, which may contribute to the study of body development and regenerative medicine.We employed five time series microarray datasets of mouse ESCs in this study, including a dataset for retinoid acid induced differentiation [15] and four datasets for spontaneous differentiation of four ESC lines (three lines from [22]; one unpublished, S.Z. and W.H.W, manuscript in preparation). We restricted the analysis to the regulatory relationships among 747 genes that are annotated by Gene Ontology term Transcription Regulator Activity, and are present on the Affymetrix U72av2 array. We designated six known TFs, Oct4, Sox2, Nanog, Klf4, Esrrb and Tcl1 as regulators of this system, due to their previously characterized role in ESCs. Interaction-Identifier [14] was applied to each time course microarray dataset. A list of common TF Interaction forms across datasets was then generated by Evidence merger. Genes were then grouped by their predicted regulators as well as their roles of regulation, i.e. activator and repressor. Twelve gene groups were formed. ChIP-chip data are available for Oct4, Sox2, Nanog and Klf4. Five out of eight regulatory-target relationships involving these four regulators were significantly enriched with ChIP-chip verified relationships (Table 1). RNA knock-out experiments were performed for all the six regulators [15,17]. Nine out of twelve target gene groups involving these six regulators were enriched with RNAi verified regulatory relationships (Table 2). Note that when using RNAi data for testing the predicted regulatory role of a TF, we only counted the target genes whose changes of expression were in the consistent direction to the predicted role of its TF, but not counting all targets genes with any changes to both directions. These tests demonstrated that the predicted regulatory relationships were in general consistent to those derived from independent experiments.Table 1Validation by ChIP-chip dataRoleTF# of target genes# of genes verified in ChIPChi-SquareP-valueActivationNanog391210.469860.00121Sox21212112.4370.00042Oct46782.4361130.11857Klf4491813.907870.00019RepressionNanog47114.1901520.04066Sox2132195.7787380.01622Oct4103112.1212880.145264Klf462141.3351510.247891Table 2Validation by RNA interference dataRoleTF# of target genes# of genes verified in RNAiChi-SquareP-valueActivationNanog3999.7106040.00183Sox21212016.220835.6E-05Oct4671325.266045E-07Esrrb9562.9662060.085021Tcl12124.6504290.03105Klf4491625.212625.1E-07RepressionNanog4720.0187130.891192Sox2132129.0359170.00265Oct410373.9093970.04802Esrrb7365.4077210.02005Tcl12724.6635940.03081Klf462100.5373940.463515Finally, Network-Identifier identified the regulatory relationships that were predicted by expression data and had consistent evidence from either RNAi or ChIP-chip data. We used Cytoscape [23] to display the final reported regulatory relationships (Figure 1).Figure 1The gene regulatory network identified by Network-Identifier. Yellow nodes represent regulators. Green nodes represent genes promoting self-renewal and pluripotency. Red nodes represent genes used for differentiation. Sharp and blunt arrows represent activation and repression effects, respectively. Red and green lines represent activation and repression activities with RNAi evidence, respectively. Blue and black lines denote regulatory relationships with ChIP-chip evidence.87 regulators and target genes were reported in the ESC transcription network (Figure 1). In particular, the mutual regulation of Klf2 and Klf4 were recently shown to be an important module for maintaining the undifferentiated state of ESCs [17]. Utf1 and Myc are known to be key ESC transcription factors. The result that they are under the control of Oct4 and Klf4 underscores the importance of Klf4 in promoting self-renewal. Mtf2 has only recently been implied to inhibit differentiation by recruiting the polycomb group of transcription repressors [24]. This analysis indicates that Klf4 and Sox2 could synergistically activate Mtf2 in ESCs. The regulatory relationships for a number of genes involved in lineage specific differentiation were also identified. These include Gata6, Gata3, Sox17 and FoxA2. Inhibiting these lineage specific differentiation genes in ESCs is critical to maintain an undifferentiated state. Among the predicted network, there were a number of transcription repressors, including Ctpb2 and Rest. Ctpb2 was predicted to be activated by Oct4. Rest was predicted to be jointly regulated by Oct4 and Sox2. These results suggest that Oct4 and Sox2 could indirectly inhibit differentiation genes by activating transcription repressors such as Ctpb2 and Rest.DiscussionNetwork-Identifier is proposed to reconstruct transcription network based on biophysical models of transcription regulation. Multiple temporal gene expression datasets are used as inputs to Network-Identifier. ChIP-chip and RNAi data can also be utilized by Network-Identifier as independent validation datasets to further improve the predicted networks. Moreover, Network-Identifier has great flexibility in incorporating independent datasets other than ChIP-chip or RNAi data to reinforce the strength of validation.It should be recognized that there are still a number of simplifications made in the modeling of the biophysical properties of gene regulation. A number of molecular events are not included in the model. These include: 1) the interactions of more than two TFs, 2) long range interaction of enhancer binding TFs and RNAP, 3) DNA methylation and 4) chromatin structure and state. Future work that takes these molecular features and events into account will potentially improve the accuracy of network reconstruction.MethodsRevisiting the Interaction-Identifier methodWe previously described the Interaction-Identifier method for identification of the candidate form of interaction among TFs and RNAP on the promoter of a target gene [14]. Interaction-Identifier models how a given TF interaction form affects the transcript concentrations of a target gene at steady states. Searching the space of TF interaction forms, it identifies the form that minimizes the difference between model-derived target concentrations and the observed expression data. The method is composed of three components: (1) a thermodynamic model for translating a TF interaction form and TF concentrations into the equilibrium probability of RNAP binding to the promoter of the target gene; (2) a kinetic model for derivation of steady state transcript concentrations of the target gene; and (3) matching gene expression data to model-derived steady state concentrations and identifying the underlining TF interaction form (Figure 2).Figure 2Flowchart of the Interaction-Identifier method.Thermodynamic models for RNAP bindingThermodynamic models are based on the assumption that the level of gene expression is proportional to the equilibrium probability that RNAP binds to the promoter of interest; and these probabilities can be computed in a statistical mechanics framework. The Interaction-Identifier method follows recent efforts [12,25,26] to translate TF concentrations into the equilibrium probability of RNAP binding using thermodynamic models.Considering the transcription of the target gene is regulated by only one TF on its promoter, a promoter can then take one of the four possible states: (1) both the TF and the RNAP bind with the promoter; (2) Only the RNAP binds to the promoter; (3) Only the TF binds to the promoter; (4) neither the TF nor the RNAP binds with the promoter, where let the weight of promoter with no RNAP or TF be 1 and the weights qp, qTF and wTFpqpqTF denote the ratios between the probabilities of states 2, 3, 4, respectively. The probability of the promoter of the target gene being bound with an RNAP is:\n                        \n                           \n                              RNAP is\n                              :\n                              \n                                 \n                                    \n                                       q\n                                       p\n                                    \n                                    +\n                                    \n                                       w\n                                       \n                                          T\n                                          F\n                                          p\n                                       \n                                    \n                                    \n                                       q\n                                       \n                                          T\n                                          F\n                                       \n                                    \n                                    \n                                       q\n                                       p\n                                    \n                                 \n                                 \n                                    1\n                                    +\n                                    \n                                       q\n                                       p\n                                    \n                                    +\n                                    \n                                       q\n                                       \n                                          T\n                                          F\n                                       \n                                    \n                                    +\n                                    \n                                       w\n                                       \n                                          T\n                                          F\n                                          p\n                                       \n                                    \n                                    \n                                       q\n                                       \n                                          T\n                                          F\n                                       \n                                    \n                                    \n                                       q\n                                       p\n                                    \n                                 \n                              \n                              ,\n                              where \n                              \n                                 w\n                                 \n                                    T\n                                    F\n                                    p\n                                 \n                              \n                              =\n                              \n                                 {\n                                 \n                                    \n                                       \n                                          \n                                             1\n                                          \n                                          \n                                             \n                                                No interaction\n                                             \n                                          \n                                       \n                                       \n                                          \n                                             \n                                                10\n                                                ~\n                                                100\n                                             \n                                          \n                                          \n                                             \n                                                Activation\n                                             \n                                          \n                                       \n                                       \n                                          \n                                             0\n                                          \n                                          \n                                             \n                                                Repression\n                                             \n                                          \n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     A TF can serve as either an activator or a repressor, or simply it does not interact with the RNAP, represented by different wTFp. These effects can be simulated by choosing appropriate wTFp. If w is set to 1, it represents that RNAP and the TF independently bind to the promoter. If w is set to 10~100, it represents that the TF helps to recruit RNAP to the promoter. The larger w is the higher the synergism is. If w is set to 0 or close to 0, it represents that the TF blocks the RNAP binding, and thus the TF is a repressor.Under the statistical mechanics framework, similar expressions can be derived for genes with two regulatory TFs capable of binding to a promoter together with RNAP. By adjusting the interaction factors w, we can obtain an analytical form for the probability of RNAP binding under different forms of interactions among RNAP and the two TFs.A kinetic model for equilibrium transcript concentrationThe equilibrium concentration of the transcripts of a target gene is governed by its synthesis and degradation rates. Empirical data show that mRNA synthesis rate is proportional to the equilibrium probability of RNAP binding to promoters [12,25,26]. Interaction-Identifier further assumes that mRNA degradation rate is proportional to its concentration, which seems to be a reasonable assumption based on recent studies [27,28]. However the authors can always relax this assumption into more general forms at the model evaluation and model improvement stages. Empirical data on eukaryotic mRNA synthesis and degradation are available to estimate these rates [27-29]. Ordinary equations are used to implement this kinetic model.Inferring gene regulatory networkA computational framework for inferring gene regulatory networks (Network-Identifier) was developed based on thermodynamic modeling of transcription regulation. Network-Identifier requires more than one time course microarray experiments for the same biological process as input datasets. The method has three components: 1) Interaction-Identifier [14], 2) Evidence merger and 3) Verification component (Figure 3). For each time course dataset, Network-Identifier enumerates all possible regulatory forms on each target gene. These interaction forms include the activation or repression by a single TF, and the five interaction forms between any two TFs. Network-Identifier evaluates the fitness of each interaction form with Interaction-Identifier and ranks them according to their fitness. The 10 most likely interaction forms of TFs on a target gene are recorded as the Top-10 List. A built-in cutoff (default = 0.8) for Interaction-Identifier eliminates any interaction that is not well supported by data. It is therefore possible for a target gene to have less than 10 candidate TF interaction forms in its Top-10 List. The Top-10 Lists from every dataset are passed onto Evidence merger, which searches for the most frequently appeared interaction form in the Top-10 Lists of a target gene. This most frequently identified interaction form is passed onto the verification component. The verification component groups target genes according to their TF interaction forms. For each regulator-target relationship, for example TF-1 represses gene a, the target genes grouped into this relationship are subject to statistical tests. Chi-square tests are used to test whether the identified TF-target relationships are enriched with regulatory relationships identified from independent experimental data, such as ChIP-chip and RNAi data. Finally, if the tests are all insignificant, Network-Identifier will fail to report any regulatory network. If some of these tests are significant, suggesting there is consistency between the expression-derived regulatory relationships and those found by independent methods, Network-Identifier will invoke a compromise algorithm to report the regulatory relationships that are confirmed by at least two independent data sources. Currently the implemented compromise algorithm is to require the regulatory relationship identified by expression data to be reproduced in at least one of the two other experiments: ChIP-chip and RNAi. It is easy to substitute this algorithm with more sophisticated algorithms [30] or when some of the independent data are not available.Figure 3Flowchart of inferring a gene regulatory network. Input microarray datasets are shown in yellow and independent experimental data for validation are marked in orange. Intermediate results are shown in blue. Computational components are shown in green.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsCCC implemented the algorithm and performed data analysis. SZ initiated and supervised the project. CCC and SZ wrote the manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2559898.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2559898",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2559898\nAUTHORS: Jack Y Yang, Mary Qu Yang\n\nABSTRACT:\nBackgroundMany protein regions and some entire proteins have no definite tertiary structure, presenting instead as dynamic, disorder ensembles under different physiochemical circumstances. These proteins and regions are known as Intrinsically Unstructured Proteins (IUP). IUP have been associated with a wide range of protein functions, along with roles in diseases characterized by protein misfolding and aggregation.ResultsIdentifying IUP is important task in structural and functional genomics. We exact useful features from sequences and develop machine learning algorithms for the above task. We compare our IUP predictor with PONDRs (mainly neural-network-based predictors), disEMBL (also based on neural networks) and Globplot (based on disorder propensity).ConclusionWe find that augmenting features derived from physiochemical properties of amino acids (such as hydrophobicity, complexity etc.) and using ensemble method proved beneficial. The IUP predictor is a viable alternative software tool for identifying IUP protein regions and proteins.\n\nBODY:\nBackgroundProteins are composed of one or more chains of amino acids, and exhibit several levels of structure. The primary structure is defined by the sequence of amino acids comprising each chain, while the secondary structure is defined by local, repetitive spatial arrangements, which falls into three basic categories: helix, strand, and coil. The tertiary structure is defined by how the chain folds into a three-dimensional configuration, while the quaternary structure is concerned with how different chains combine into multisubunit or oligomeric, protein (protein complexes). Most proteins function only when folded into a particular configuration. Recently, a class of proteins has been discovered that do not fold into any particular configuration – instead of folding into specific 3-D structures, they exist as dynamic ensembles in their native state. These proteins have been variously called natively unfolded, natively disordered or Intrinsically Unstructured regions and Proteins (IUP) [1-7]. Unlike regular proteins, which unfold and lose their ability to function when subjected to environmental challenges such as detergents, urea, or heat, IUP may continue to function under such conditions, as they do not have to be folded into a particular configuration in order to carry out their function. An IUP protein usual does not involve with catalysis process that functioning as an enzyme, because catalysis requires tightest-binding to transition state, which binding specificity most likely requires structured active site, an IUP functions in signaling and have been associated with a wide range of protein functions such molecular recognition, molecular assembly and disassembly as well as protein modification. IUP regions also play a central role in diseases characterized by protein misfolding and aggregation [1-5]. Furthermore, the identification of such regions can aid both structure determination and sequence alignment, and may also aid in drug design. The identification of IUP regions from the primary structure of a protein is thus an important but difficult problem [1-5]. IUP can be identified through protein tertiary structure. Traditionally, the tertiary structure of proteins is determined using experimental methods such as X-ray crystallography, Overhauser Effect Enhanced Nuclear Magnetic Resonance spectroscopy (NMR), and Circular Dichroism Spectroscopy (CDR). However, these experimental methods are usually time consuming and often have their own limitations and problems. Since Dunker et. al. developed the first IUP predictor – PONDR [3], consequential development of IUP predicators includes disEMBL [6], and Globplot [7]. We had developed a number of IUP and membrane protein predictors [8-10] that use amino acid sequences as inputs and that give IUP and structured protein assignments as outputs. This paper is a continuation of the earlier IUP predictors we developed before [8-10]. Our predictors use protein primary structure information only and contain three parts: feature generation, classification and ensemble methods.MethodsFeature generation from primary structureThe first step in constructing a classifier is to choose the features that the classier uses. Performance of a given classifier depends on the set of features that are used. Our feature extraction is based on the physiochemical analysis of protein sequences.Features from compositions of peptide sequencesWe studied the characteristic of amino acid residues in sequences. We found that different properties of amino acid in sequence tend to encode structural information. It appears different protein folding classes can be identified by the differences in their amino acid compositions. Detailed analysis of patterns of sequences leads us some important discoveries. Analyzing the peptide patterns in sequences have convinced us that different protein folding classes can be identified by the differences in their amino acid compositions. We conclude that distinguishing peptide patterns in sequences provides useful information to detect different protein classes and/or discover new classes of proteins.The first set of 20 features is derived from first order statistics, regarding amino acid compositions in primary sequence. Let us define amino acid set as:A˜ = {A, C, D, E, F, G, H, I, K, L, M, N, P, Q, R, S, T, V, W, Y}.These twenty letters in the set A˜ represent the one-letter symbols for 20 amino acids. We use a window with length L centered at each amino acid residue to extract features. Let x(j) ∈ A~, j is the jth position of the amino sequence of protein. Hence, x(j) represents the amino acid at position j of the protein sequence. Let M denote the length of protein sequence. Since the window size shrinks [6] at the N-terminal and C-terminal of protein sequences, we utilized k0 and k as parameters to make the window size adjustable at the terminals. k0 and k are defined as follows:(1)k = max {1, i - (L-1)/2}and(2)k0 = min {M, i + (L-1)/2}Let Pi (a) represents the probability of an amino acid residue, which is denoted by \"a\"(a ∈ A~), inside the window centered at position j, then(3)\n                        \n                           \n                              \n                                 P\n                                 i\n                              \n                              (\n                              a\n                              )\n                              =\n                              \n                                 1\n                                 \n                                    k\n                                    −\n                                    \n                                       k\n                                       0\n                                    \n                                    +\n                                    1\n                                 \n                              \n                              \n                                 \n                                    ∑\n                                    \n                                       j\n                                       =\n                                       \n                                          k\n                                          0\n                                       \n                                    \n                                    k\n                                 \n                                 \n                                    \n                                       δ\n                                       \n                                          x\n                                          (\n                                          j\n                                          )\n                                          a\n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     where k0 and k are defined in Equation (1) and (2), a ∈ A~:We then constructed more features by second order statistics regarding the pattern of one amino acid followed by another amino acid in the window. Let's define Ψaman represent the pattern of a pair of amino acids, that is, amino acid am followed by amino acid an. Both am and an belong to amino acid set A~. Second order statistic features are calculated by the following equation:(4)\n                        \n                           \n                              \n                                 P\n                                 i\n                              \n                              (\n                              Ψ\n                              \n                                 a\n                                 m\n                              \n                              \n                                 a\n                                 n\n                              \n                              )\n                              =\n                              \n                                 1\n                                 \n                                    k\n                                    −\n                                    \n                                       k\n                                       0\n                                    \n                                    +\n                                    1\n                                 \n                              \n                              \n                                 \n                                    ∑\n                                    \n                                       j\n                                       =\n                                       \n                                          k\n                                          0\n                                       \n                                    \n                                    \n                                       j\n                                       =\n                                       k\n                                       −\n                                       1\n                                    \n                                 \n                                 \n                                    \n                                       δ\n                                       \n                                          x\n                                          (\n                                          j\n                                          )\n                                          \n                                             a\n                                             m\n                                          \n                                       \n                                    \n                                    \n                                       δ\n                                       \n                                          x\n                                          (\n                                          j\n                                          +\n                                          1\n                                          )\n                                          \n                                             a\n                                             n\n                                          \n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     where k0 and k are defined in (1), (2), am ∈ A~, an ∈ A~\n                        \n                           \n                              \n                                 δ\n                                 \n                                    x\n                                    (\n                                    j\n                                    )\n                                    ,\n                                    \n                                       a\n                                       m\n                                    \n                                 \n                              \n                              \n                                 δ\n                                 \n                                    x\n                                    (\n                                    j\n                                    +\n                                    1\n                                    )\n                                    ,\n                                    \n                                       a\n                                       n\n                                    \n                                 \n                              \n                              =\n                              {\n                              \n                                 \n                                    \n                                       1\n                                    \n                                    \n                                       \n                                          if amino acid x\n                                          (\n                                          j\n                                          )\n                                          =\n                                          \n                                             a\n                                             m\n                                          \n                                           and x\n                                          (\n                                          j\n                                          +\n                                          1\n                                          )\n                                          =\n                                          \n                                             a\n                                             n\n                                          \n                                       \n                                    \n                                 \n                                 \n                                    \n                                       0\n                                    \n                                    \n                                       \n                                          otherwise\n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     Next we introduce a 9-gram encoding scheme based on the physiochemical properties of amino acids. The 20 amino acid residues can be clustered into 9 groups as shown in Table 1. Then, we defined the 2-tuple code set as following:Table 1Encode scheme based on physiochemical properties of amino acidsGroupResiduesDescriptiong1CHighly conservedg2MHydrophobicg3N, QAmides, polarAcids, positive,g4D, Epolarg5S, TAlcoholsg6P, A, GAliphatic, smallg7I, V, LAliphaticg8F, Y, WAromaticg9H, K, RBases, chargedΩ = {g1, g2, g3, g4, g5, g6, g7, g8, g9}, and g ∈ ΩThere are two advantages to this 9-gram encoding scheme. First, for small window sizes, there may not be sufficient data to accurately represent the first and second order statistics of 20 amino acids; this will be less of a problem with the 9-gram encoding because there are fewer first and second order statistics to estimate than when the 20 amino acid encoding is used. Second, the 9-gram scheme can reduce computational complexity if it is used as an alternative.Modify equations (3) and (4) by the 9-gram encode scheme, we then calculate on the first order statistics and second order statistics of the 9-gram encoding scheme for each amino acid residue in a sequence using the same window accordingly.We also generated features by using protein family profiles or position-specific scoring matrices from PSI-BLAST. For a sequence of length N, an N*20 family profile is constructed based on the multiple alignment of homologues found during the PSI-BLAST search. Feature Pj(a) is the averaged log-odds of amino acid \"a\" in the neighbourhood of sequence position j as calculated below:(5)\n                        \n                           \n                              \n                                 p\n                                 j\n                              \n                              (\n                              a\n                              )\n                              =\n                              \n                                 1\n                                 \n                                    k\n                                    −\n                                    \n                                       k\n                                       0\n                                    \n                                    +\n                                    1\n                                 \n                              \n                              \n                                 \n                                    ∑\n                                    \n                                       j\n                                       =\n                                       \n                                          k\n                                          0\n                                       \n                                    \n                                    k\n                                 \n                                 \n                                    \n                                       S\n                                       \n                                          x\n                                          (\n                                          j\n                                          )\n                                       \n                                    \n                                    (\n                                    a\n                                    )\n                                 \n                              \n                           \n                           \n                        \n                     Sequence complexityThe complexity of the sequence is used as a features. According to the past research [1-5], Most likely IUP have low complexity patterns. K2-entropy is used to measure the local complexity of the amino acid sequence. The complex of each amino acid residue in the sequence is calculated in the same window as previous defined, which is given by:(6)\n                        \n                           \n                              \n                                 C\n                                 i\n                              \n                              =\n                              −\n                              \n                                 \n                                    ∑\n                                    \n                                       n\n                                       =\n                                       1\n                                    \n                                    \n                                       20\n                                    \n                                 \n                                 \n                                    \n                                       P\n                                       i\n                                    \n                                    (\n                                    \n                                       a\n                                       n\n                                    \n                                    )\n                                    \n                                       \n                                          log\n                                          ⁡\n                                       \n                                       2\n                                    \n                                    \n                                       P\n                                       i\n                                    \n                                    (\n                                    \n                                       a\n                                       n\n                                    \n                                    )\n                                 \n                              \n                           \n                           \n                        \n                     where an ∈ A~, and Pi is calculated by Equation (3).HydrophobicitySince hydropathy is an important determinant of protein-chain fold, calculation of hydropathy could be useful for IUP prediction. We therefore use relative hydrophobicity of each amino acid, called hydropathy in the feature space. The feature for hydropathy H(i) at position i is the average of hydropathy in the feature window for given hydropathy scale is calculated as following:(7)\n                        \n                           \n                              H\n                              (\n                              i\n                              )\n                              =\n                              \n                                 1\n                                 \n                                    k\n                                    −\n                                    \n                                       k\n                                       0\n                                    \n                                    +\n                                    1\n                                 \n                              \n                              \n                                 \n                                    ∑\n                                    \n                                       j\n                                       =\n                                       \n                                          k\n                                          0\n                                       \n                                    \n                                    k\n                                 \n                                 \n                                    H\n                                    y\n                                    d\n                                    r\n                                    o\n                                    p\n                                    a\n                                    t\n                                    h\n                                    y\n                                    (\n                                    j\n                                    )\n                                 \n                              \n                           \n                           \n                        \n                     There are several different hydropathy scales generated by different methods. Equation (7) is used for generating features based on four different hydropathy scales. However, by comparing joint probability distributions of intrinsically unstructured and structured proteins for all different hydropathy scales, we found that Kytes-Doolittle's scale [11] is best in distinguishing IUP from structured proteins. This best feature can be also selected from principle component analysis as showing in the feature selection session.IUP propensitiesFinally, we also use amino acid IUP propensities [7], which is a scale to measure how likely an amino acid is to be unfolded. Comparisons on two sets of propensities are shown in Table 2. The average of IUP propensity associated with jth residue in the sequence given in a window with length Lis calculated as below:Table 2Russell/Linding and Deleage/Roux: Disorder PropensitiesResidueRussell/LindingDeleage/RouxP0.552321.117G0.433230.6675N0.229890.479D0.227630.4645S0.142880.2965T0.008870.145H-0.001210.135C-0.00151-0.1255K-0.10001-0.0495R-0.17659-0.179Q-0.18768-0.055E-0.20469-0.2745Y-0.207510.0825F-0.22557-0.497M-0.2259-0.4765W-0.24338-0.257A-0.26154-0.275L-0.33793-0.4385V-0.38618-0.7055I-0.42224-0.515(8)\n                        \n                           \n                              R\n                              (\n                              j\n                              )\n                              =\n                              \n                                 1\n                                 \n                                    k\n                                    −\n                                    \n                                       k\n                                       0\n                                    \n                                    +\n                                    1\n                                 \n                              \n                              \n                                 \n                                    ∑\n                                    \n                                       j\n                                       =\n                                       \n                                          k\n                                          0\n                                       \n                                    \n                                    k\n                                 \n                                 \n                                    I\n                                    U\n                                    P\n                                    p\n                                    r\n                                    o\n                                    p\n                                    e\n                                    n\n                                     \n                                    (\n                                    j\n                                    )\n                                 \n                              \n                           \n                           \n                        \n                     The k and ko are defined in (1), (2).The selection on a better propensity feature is illustrated in the feature selection session using a distance gauge.Feature selection for high-dimensional dataThe above feature extraction step can generate 537 features for each amino acid residue in a sequence as summarized below:• 20 from first order statistics of 20 amino acids;• 400 from second order statistical of 20 amino acids;• 9 from first order statistics of 2-tuple code;• 81 from second order statistics of 2-tuple code;• 20 Features based on PSI-BLAST Profiles;• 1 from complexity;• 4 from different hydropathy scales;• 2 from different unstructured propensity scales.High-dimensional data requires a feature selection step to address the curse-of-dimensionality. In addition, a large number of features often make the learning algorithm scale poorly [1]. Generally, feature selection strategies are either wrapper-based, or embedded, or filter-based.Wrapper-based feature selectionWrapper algorithms use the interactions between feature selection and the learning algorithm by involving the learning algorithm in the feature selection step. If they are not over-fitting and are not so expensive computationally, wrappers would be the best feature selection algorithms, since they also depend on the inductive principles of the learning algorithms.Embedded feature selection algorithmsDecision Trees and CART (classification and regression trees) exemplify embedded feature selections; the process of selecting a feature to split at each node of the tree is implicitly a feature-selection step.Filter-based feature selectionFilter-based feature selections select features before the data is passed to a learning algorithm. They are used as pre-processing steps to model selections and learning. Since these algorithms are independent of any learning algorithms and they are used in the preliminary steps in learning, their computational complexities are usually not high and are much faster than wrapper-based algorithms.We studied several feature selection algorithms include information gain, T-test, the Chi-square goodness-of-fit test, the bi-normal separation (BNS), Fisher permutation test, and distance measurement.We are especially in favour of distance measures because they can reduce computational complexity efficiently. Let's consider a two-class problem. Given two features: X and Y, D (X) and D (Y) measure the separation of two classes subject to feature X and feature Y, respectively. The distance measure D (X) for feature X is defined as follows:(9)\n                        \n                           \n                              D\n                              (\n                              X\n                              )\n                              =\n                              \n                                 \n                                    \n                                       m\n                                       1\n                                    \n                                    −\n                                    \n                                       m\n                                       2\n                                    \n                                 \n                                 \n                                    \n                                       \n                                          \n                                             σ\n                                             1\n                                             2\n                                          \n                                          +\n                                          \n                                             σ\n                                             1\n                                             2\n                                          \n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     where m1 and m2 are the mean values of feature X for the class 1 and class 2, σ12 and σ22 are variance of feature X for the class 1 and class 2, respectively. The mean and variance for a given feature are calculated as follows:(10)\n                        \n                           \n                              \n                                 \n                                    \n                                       \n                                          m\n                                          =\n                                          \n                                             1\n                                             N\n                                          \n                                          \n                                             \n                                                ∑\n                                                \n                                                   i\n                                                   =\n                                                   1\n                                                \n                                                N\n                                             \n                                             \n                                                \n                                                   x\n                                                   i\n                                                \n                                             \n                                          \n                                       \n                                    \n                                 \n                                 \n                                    \n                                       \n                                          \n                                             σ\n                                             2\n                                          \n                                          =\n                                          \n                                             1\n                                             \n                                                N\n                                                −\n                                                1\n                                             \n                                          \n                                          \n                                             \n                                                ∑\n                                                \n                                                   i\n                                                   =\n                                                   1\n                                                \n                                                N\n                                             \n                                             \n                                                \n                                                   \n                                                      (\n                                                      \n                                                         x\n                                                         i\n                                                      \n                                                      −\n                                                      m\n                                                      )\n                                                   \n                                                   2\n                                                \n                                             \n                                          \n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     where xi is the value of feature X for ith instance in one class, and N is the number of instances in that class. Modifying equations (10–11) gives us the calculation of D(Y) corresponding to feature. Now we obtain both D(X) and D(Y). If D(X) > D(Y), feature X is selected; otherwise feature Y is selected.Distance measure can be in pair each time for every feature. After sorting these distances, we can select most important features for separating two classes. For example, for the feature on IUP propensities as showing in Tables 1, we choose Russell/Linding scale by the above decision rule. This also justify why they can build a single-feature simplest predictor.From the above analysis, clearly, distance-measure based feature selection method works in a pair-wise manner. We are also in favour of another feature selection method called principle component analysis (PCA) to reduce the feature dimension.Compare to previous approach, this particular feature selection scheme can handle multiple features simultaneously. PCA is also called Karhunen-Loeve (K-L) transformation. K-L transformation is an orthonormal transformation of a vector X→ to same dimensional vector Y→. In the transformation domain, the first principle component is the normalized linear combination with maximum variance; the second component has the next largest variance and so forth. Based on such ranking, only those with largest variance are preserved and the others are neglected. In fact the principle components are ranked by their ability to distinguish among classes. The implement [10] procedure is as follows. Assume there are N instances in the training set and M features, let X→ represent a population of N-dimensional vectors, mean value of each feature mx has been calculated. The mean value of each feature and KL transformation have been performed, the resulting covariance matrix of Y→ has been analyzed. According to the nature of those features, we divide them into four categories [10]: characterize of local compositions of amino acids, characterize patterns of local regions, characterize propensities of IUP, characterize physiochemical properties of local regions.Evidently, there are four different hydropathy scales which result in four features regarding amino acid hydrophobicity property. These features are correlated with each other. The correlation factor ρxy of two features X and Y can be obtained by the following equation:(11)\n                        \n                           \n                              \n                                 ρ\n                                 \n                                    x\n                                    y\n                                 \n                              \n                              =\n                              \n                                 \n                                    E\n                                    [\n                                    X\n                                    Y\n                                    ]\n                                    −\n                                    E\n                                    [\n                                    X\n                                    ]\n                                    E\n                                    [\n                                    Y\n                                    ]\n                                 \n                                 \n                                    \n                                       \n                                          (\n                                          E\n                                          [\n                                          \n                                             X\n                                             2\n                                          \n                                          ]\n                                          −\n                                          E\n                                          \n                                             \n                                                [\n                                                X\n                                                ]\n                                             \n                                             2\n                                          \n                                          )\n                                       \n                                    \n                                    \n                                       \n                                          (\n                                          E\n                                          [\n                                          \n                                             Y\n                                             2\n                                          \n                                          ]\n                                          −\n                                          E\n                                          \n                                             \n                                                [\n                                                Y\n                                                ]\n                                             \n                                             2\n                                          \n                                          )\n                                       \n                                    \n                                 \n                              \n                           \n                           \n                        \n                     Among those four features, we choose the best one by using decision rule based on PCA. It turns out the Kyte-Doottle scale generate the best feature compared to the other three scales on IUP prediction [10].After our feature selection procedure, we decide to use 59 selected features as input to our classifier. These features describe the amino acid compositions, hydropathy, complexity and IUP propensities from sequences. The experimental results show that these 59 features turn out to produce the best predicting accuracy for IUP classifier. The selected features are also cross checked with AAIndex  so that they are biologically significant.IUP predictorWe use the Recursive Maximum Contrast Tree Classifier (RMCT) we developed before [8-10]. This classifier utilizes K-Nearest Neighbour Classifier [12], where the nearest neighbours are defined by the tree structure [8-10,12]. After constructing the RMCT using the training data, the resulting tree can be used as a 2-class classifier. We define IUP residues as class 1, and structured protein peptide residues as class 0. K majority voting principle guides classification with RMCT on tree nodes. We calculate the distances between a test instance and instances in the training dataset by a distance measure D(n). We sort the distances in ascending order and get the first K minimum distances as our K decision-making instances in a tree node. Then the class label of the test instance is assigned according to the majority voting of class labels from these K decision-making distances.Detecting IUP is measured by the value P(L = 1) that is given by K majority voting rules based on the training instances and therefore is a real number between zero and one. AS shown in Figure 1 shows the IUP predictor interface, the y-axis on low-left window represents P(L = 1), while the x-axis represents the residue location (starting from number 1 to M – the length of sequence). P (L = 1) actually reflects a \"probability\" that a test instance is classified as class 1. If P(L = 1) is great than 1/2 by the K majority voting rule as described above, then this test instance should be classified as class 1, and if it is less than 1/2, it should be classified as class 0 (Default threshold value is set as 1/2). However such threshold can be adjustable in IUP classifier. The general rule for classification is that if P(L = 1) is greater than the threshold value (between 0 and 1), then the test instance is classified as class 1, otherwise it is classified as class 0. If we decrease the threshold value from 1/2, then more class 1 test instances can be detected. The trade off is that statistically, more class 0 instances are wrongly classified as class 1. Increasing the threshold has the effect of increasing the false negative rate and reducing the false positive rate. Varying threshold values is an essential technique in obtaining Receiver Operating Characteristic (ROC) curves. Once we set up a threshold value as the boundary, a test instance can be assigned a class label at each of the root, child and grandchild node (when N = 3) in RMCT classifier [8-10]. Finally classification decision (for N > 1) is made by the majority voting from the class labels of those nodes after each node has already made its own decision based on majority voting from the decision making instances in each node. Therefore, adjustable parameters in the IUP predictor are the window length L (used in generating features), the number of decision-making instances K in each node on RMCT and number of tree nodes participating in final classification decision (N).Figure 1Example of IUP.Ensemble methodsThe performance of IUP predictor can be improved using ensemble methods; a diverse class of methods that seek to combine the decisions of several classifiers in order to improve classification accrues [12-15]. We exploited the implementation of combing boosting with bagging to improve prediction accuracy.Combing Boosting with BaggingWe developed the Boosting with Bagging algorithm. Boosting is an algorithm that can be used to improve the performance of a classifier. While the original Boosting algorithm is due to Schapire [13], later Freund and Schapire introduced an improved algorithm called Adaboost, which was designed to handle 2-class classifiers. There were several extensions to the multi-class case, including Adaboost.M1 [14]. As we are interested in incorporating useful confidence information into IUP classifier, we combine bagging with a generalization of Adaboost.M1 called the CBoost algorithm [15] that allows confidence information to be incorporated. Our combined CBoost with bagging algorithm emphasizes on weaker learner for each boosting run.Assuming we have N training instances, then we construct classify f (x→i). Class label yi is either 0 or 1. The square error of classify f (x→i) is given by each boosting round, The training set is denoted as T = {(xi, yi), i = 1,... N} where xi is a feature vectors and yi corresponds to a class label. The boosting approach constructs a sequence of functions (also known as classifiers or hypotheses) h(x→), indexed by the parameter t; given by an instance (xi, y). The kth component of h(x→) reflects the confidence of the classifier that label k corresponds to the true class label y. The components of h(x→) are normalized, so that h(x→) specifies a distribution over labels. Associated to each instance(xi, yi) is a weight wit, which, when normalized, yields a probability distribution {Pit, i = 1,... N} over the training data. This distribution is supplied to the \"subroutine\" that actually constructs the classifier associated with boosting round t. The \"subroutine\" then uses these weights to construct the classifier h(x→). The error of the classifier constructed on round t are identified, and h(x→) relative to the distribution supplied on round t, and {Pit} are calculated. The coefficient of the classifier h(x→) in the combined hypothesis can then be calculated. The weights are then updated using the update rules. Those weights corresponding to instances x for which h(x→) assigns a large probability mass (near 1) to the correct class are decreased substantially more than those for which h(x→) assigns a small probability mass. It follows that the distribution Pit becomes more concentrated on those instances that have a high rate of misclassification. We then classify an attribute vector x by computing the linear combinations of h(x→). This procedure is combined with bagging.Bagging with CBoost can reduce variance error but not affect bias error [10]; this algorithm averages the predictions of several classifiers and then assigns the class label that is closest to the average.ConclusionWe applied our IUP predictor to the problem of identifying both IUP and structured regions in proteins. We have developed a new ensemble method called bagging with CBoost; that have improved the overall performance of our IUP predictor. We find that both feature selection and ensemble methods improve the performance. Also we find that extracting features based on physiochemical proprieties proved beneficial. Augmenting features derived from physiochemical proprieties of amino acids (such as hydrophobicity, complexity, etc.) followed by and feature selection step, and developing bagging with CBoost algorithm significantly improve the predicting accuracies on both IUP and structured proteins. Those are the key innovations of this version of IUP predictor. Because performance of IUP is on par with PONDR, IUP is a viable alternative predictor that can be useful in structural genomics.ResultsDataBoth IUP and structured protein training data are selected from the Protein Data Bank (PDB) based on X-ray crystallography or other reliable experimental results. IUP residues are verified in PDB as missing coordinates in X-ray crystallography data. The selected structured protein sequences consist of 290 non-redundant sequences and are completely folded proteins with stable 3D structures, and resolutions better than 2A. Their pair-wise sequence identities are less than 25% (PDB SELECT25).Experimental resultsStructural information on IUP and structured proteins for training are cross verified using PDB to ensure the reliability of training data. The test dataset are out of sample data.It appears that the performance of the IUP predictor depends on several parameters, including the window length L for feature generation, the number of decision making instances K inside a RMCT tree node and the number of tree nodes N that are used for decisions. Experimental results indicate that N is least sensitive and L is most sensitive, therefore, we focused on a series of experiments using a variety of different window L lengths L and decision-making instances K.We performed the experiments to test how the number of decision-making instances affects the performances of IUP prediction. As shown in Table 3, The classifier achieved the best performance in detecting IUP regions when the number of decision-making instances K equals to 9. However, the performance of identifying structured residues increases as K increases (monotonic increase function). To balance the predicting accuracies between IUP and structured residue simultaneously, we set K = 21 as default value in our predictor, as this value appears to offer best overall classification accuracy.Table 3The effect of decision-making instances (K) on the performance of IUPKTP (IUP)TN (Order)Average accuracy30.80310.71830.760750.80140.73360.767570.8040.73490.769490.81060.74140.776110.80360.74530.7744130.80150.7490.7752150.80230.75140.7768170.80250.75360.778190.80240.75750.7799210.80050.75970.7801230.79640.76240.7794Window length L significantly affects the performance as showing in Figure 2. The performances on IUP prediction is marked in blue, while pink line is the performance on predicting structured residues. Yellow line represents the overall accuracy. In Figure 2, we set our default window length L = 10, as this value appears to give a best overall result. Default value for N is set at N = 3.Figure 2The effect of window length for feature extraction on the performance of IUP.We made several observations, namely:▪ When the window length for extracting feature is less than 4, the predictor almost cannot identify any IUP region. This implicitly suggests existence of local force on each residue, and this force determines the fold of protein. It also indicates that a structural or functional motif contains several or more amino acid residues; window length L should not be too short. It should not be too long as a long sequence may contain several different structural or functional domains or motifs.▪ When window length L is approximately 10, the true positive rate reaches a maximum; beyond that, the true positive rate decreases as the window length increases.▪ The average accuracy reaches a maximum when window length is approximately 12; beyond that, the average accuracy keeps decreasing.▪ The true negative rate increases when window length L increases; after L is approximately increased to 28, the rate drops.▪ The three rate curves (true positive, true negative, overall) are crossed approximately at window length 14 (See Figure 2).We compared IUP performance against other predictors such as PONDR, GlobPlot and DisEMBL. Both GlobPlot and DisEMBL are developed by European Molecular Biology Laboratory (EMBL). While DisEMBL is based on the neural networks, GlobPlot is a single-feature (IUP propensity) classifier for IUP prediction. PONDR is predictor mainly based on neural network algorithm. In order to reliably test the performance of our IUP predictor, the test data was completely out-of-sample \"blind\" data and consist of 255 both IUP and ordered proteins. The prediction and comparison results are illustrated in Figure 3. The test result indicated the IUP predictor reached the same performance level of the most popular predictor PONDR and outperformed disEMB, and Globplot. For many test sequences, we found the IUP predictor performed best, while for many others, PONDR performed best. This suggested that our IUP predictor is a viable alternative to the most popular predictor – PONDR.Figure 3Comparison of our predictor (IUP) to DisEMBl, GlobPlot and PONDR VLXT.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsBoth authors contributed equally to the paper.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2567037.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2567037",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2567037\nAUTHORS: Jin Fan, Kevin G. Guise, Xun Liu, Hongbin Wang\n\nABSTRACT:\nVoluntary control of information processing is crucial to allocate resources and prioritize the processes that are most important under a given situation; the algorithms underlying such control, however, are often not clear. We investigated possible algorithms of control for the performance of the majority function, in which participants searched for and identified one of two alternative categories (left or right pointing arrows) as composing the majority in each stimulus set. We manipulated the amount (set size of 1, 3, and 5) and content (ratio of left and right pointing arrows within a set) of the inputs to test competing hypotheses regarding mental operations for information processing. Using a novel measure based on computational load, we found that reaction time was best predicted by a grouping search algorithm as compared to alternative algorithms (i.e., exhaustive or self-terminating search). The grouping search algorithm involves sampling and resampling of the inputs before a decision is reached. These findings highlight the importance of investigating the implications of voluntary control via algorithms of mental operations.\n\nBODY:\nIntroductionThe human body transmits 11 million bits of information per second (bps) to the brain, but our conscious mind can only process up to 50 bps (Information Theory, Britannica Online) [1]. For example, visual attention can select only 30 to 60 bits of information for processing with each glimpse [2]. Voluntary control [3] of information processing is therefore crucial to allocate resources and prioritize the processes so that those most relevant under a given situation can reach the level of focused consciousness. Although much of the reduction/selection (e.g., perceptual grouping [4]) has already occurred elsewhere and may be hard coded/wired, input information outside the current focus of attention cannot and should not be fully excluded because they may have survival value. Therefore, there is a need for dynamic and flexible control. Such control is also most needed when a great deal of computation is required prior to response generation, for example, during information processing in the presence of salient task-irrelevant distracters.The mechanisms of voluntary control, however, are not well understood. Although studies on the neural correlates of voluntary control routinely employ tasks that manipulate control in a qualitative manner, e.g., the Erickson flanker task [5] or the color-word Stroop task [6], they have gathered important findings. Brain structures involved in selective sensory processing of relevant visual targets have also been studied using a cued spatial-attention task [7]. Further advance in our understanding of the specific roles of these structures, however, will come from a more quantitative investigation of the relationship between behavioral/neural responses and voluntary control. This requires parametrically examining the algorithms that instantiate the mental operations of voluntary control.Searching for and identifying majority constituents of a group (e.g., if there are five children on a playground, three girls and two boys, then girls comprise the majority) is an important and common function of our daily lives. However, a hardwired circuit for the majority gate-based logic is inefficient to implement [8]. This may also apply to the human brain. Therefore, dynamic algorithms have to be employed for more flexible computations. In this study, we designed a majority function task to systematically manipulate the amount and content of input to examine the algorithms for the interplay of voluntary control with input. In this task, a set of 1, 3, or 5 horizontal arrows were presented simultaneously at 8 possible locations arranged as an octagon centered on a fixation cross. The ratio of left and right pointing arrows within a set was also manipulated. Participants were asked to determine the direction (left or right) in which the majority of arrows pointed, and to indicate their response via button press.One way to quantify information is to measure its entropy. According to Shannon's information theory [9], the information uncertainty in bits (entropy) depends on the amount/content of the input and the efficiency of encoding. Therefore, we define computational load as entropy, which is determined by the information amount/content of the input and the algorithms of mental operations used to encode and process the input. Examination of changes in reaction time (RT) with respect to computational load allowed us to test competing hypotheses of the algorithms of mental operations that may be adopted by participants to control information processing to reach a majority decision.MethodsParticipantsThirty adult volunteers participated in this study. After excluding six participants with a accuracy lower than 75% under the most difficult task condition, the final sample size included in this report is 24 (13 females and 11 males; mean age, 25.9 years; range, 22–38 years). Written informed consent was obtained from each participant following the procedure approved by the institutional review board of the Mount Sinai School of Medicine. All participants had normal or corrected-to-normal vision.Apparatus and proceduresThe task was compiled and run on a PC with a 17 inch LCD monitor, using E-Prime™ software (Psychology Software Tools, Pittsburgh, PA). The task was first explained using a paperboard illustrating each condition. Participants then performed a practice session on a PC with 6 blocks of trials with 12 trials in each block and 72 trials in total, using the same timing parameters as the actual test. The practice was continued until participants demonstrated at least 90% accuracy overall. Participants then performed the actual test.The majority function task (MFT)In this task, groups of arrows with set sizes of 1, 3, and 5 are randomly presented at 8 possible locations arranged as an octagon centered on a fixation cross. The arrows point either left or right, and are presented simultaneously (see \nFigure 1\n). The configuration of the 8 positions is similar to that used in previous studies on covert attention [10], [11]. On a computer screen, the length of the arrow is 6 mm, the radius from the fixation cross to the center of any arrow is 2 cm, and the viewing distance is 50 cm. The radius from the fixation cross to the center of an arrow subtends approximately 2.3° of visual angle. Participants' task is to indicate the direction in which the majority of the arrows point. To encourage speed without sacrificing accuracy, participants are instructed to make responses as rapidly as possible while maintaining a low error rate. In each trial, an arrow set is presented for 2500 ms, followed by a variable fixation period of 2000 to 3000 ms. Each trial lasts 5 s on average. Responses within the 2500 ms window terminate the display of the stimulus. There are three runs in this task. In each run, there are two blocks for each set size, six blocks in total. Each block has 12 trials. Within a block for a certain set size, arrows under different stimulus conditions are displayed in a random order, with each stimulus condition appearing an equal number of times. The order of the blocks is counterbalanced by a Latin square with reversed repetition within each run. The order is 135-531, 513-315, and 351-153 for the first, second, and third run, respectively. Here the number represents the set size. The total number of trials in each run is 72. Before and after each block there is a 5 s fixation period. There are also five 5 s fixation periods between blocks in each run. Each run lasts 395 s. The total trial number in this task is 216 and the task takes about 20 minutes.10.1371/journal.pone.0003522.g001Figure 1Illustration of representative stimulus configurations of the majority function task.In this task, arrows with set sizes of 1, 3, or 5 are randomly presented at 8 possible locations arranged as an octagon centered on a fixation cross. The arrows point either left or right, and are presented simultaneously. Participants' task is to indicate the direction in which the majority of arrows point. For example, if three arrows are presented, and two point to the left and one to the right (see the “2∶1” panel in the “Set size 3” column), the correct response should be “left”. The eight circles are for illustration of the locations and are not displayed during the experiment. The label for each condition is the ratio of the numbers in each category.Although the amount and content of input to be processed is varied in the majority function task, the response is only one bit because there are only two alternatives. Therefore, the variable related to the stages of response selection and execution, after stimulus preprocessing and categorization [12], is constant across all set sizes and stimulus conditions. In addition, the pattern presented in a current trial is independent of its preceding trial, in contrast to the serial-choice RT tasks.ResultsBehavioral results\n\nTable 1\n shows the experimental results including mean RT and accuracy under each condition. Although it is possible to infer the computational load of each condition from the accuracy of the responses [13], here we used RT as the main dependent variable. The mean RTs (520 ms, 884 ms, 1200 ms) for the three set sizes were significantly different, F(1, 23) = 792.18, p<0.01 (linear), and F(1, 23) = 2.69, ns (quadratic). In set size 3, the RTs under the two conditions were significantly different, F(1, 23) = 608.19, p<0.01. In set size 5, the RTs under the three conditions were significantly different, F(1, 23) = 813.95, p<0.01 (linear), and F(1, 23) = 25.09, p<0.01 (quadratic). The mean accuracy (99.5%, 98.7%, 94.6%) of the three set sizes were significantly different, F (1, 23) = 96.22, p<0.01 (linear), and F(1, 23) = 21.39, p<0.01 (quadratic). In set size 3, the accuracy under the two conditions were significantly different, F(1, 23) = 10.12, p<0.01. In set size 5, the accuracy under the three conditions were significantly different, F(1, 23) = 109.07, p<0.01 (linear), and F(1, 23) = 69.23, p<0.01 (quadratic). The positive change in RT, mean standard deviation, and error rate across conditions may represent the differences in the computational load.10.1371/journal.pone.0003522.t001Table 1RT (ms) and accuracy (%) under all stimulus conditions (n = 24).Set sizeStimulus conditionRatioRTAccuracyMeanSDMean SDa\nMeanSD10,11∶05207710799.50.93000, 1113∶0647110142100.00.0001, 0112∶1112115330997.53.9500000, 111115∶072413017499.80.900001, 011114∶1126119234998.62.400011, 001113∶2161520339285.26.7Note: a Mean of SDs across participants.Analysis of algorithms of mental operationsThe behavioral results suggest a relationship between RT and amount/content of inputs that goes beyond a simple linear or loglinear function, suggesting an interaction between uncertainty of inputs and mental operations in overall performance. We calculated computational load as a function of input information to be processed and the algorithms adopted by the human brain. Therefore, potential algorithms have to be compared and contrasted in order to find the most plausible one. Methods used in the analysis of short-term memory scan (e.g., [14]) can be used to analyze the algorithm of mental operations. Here we analyzed and compared the results using three plausible algorithms: exhaustive search, self-terminating search, and grouping search.Exhaustive searchIf we follow the equation to find the majority (the majority function), which takes all inputs and then returns the value which is most common among them, we would expect that, for all stimulus conditions within the same set size, computational load and RT would not be affected by the number of arrows pointing in a common direction. That is, the processing time for this algorithm is only affected by the amount, and not by content, of input. For example, for a set size of 5 arrows, RT would be the same for conditions in which 5, 4, and 3 arrows are pointing in the same direction. The data indicated that this was not the case.Self-terminating searchGiven that arrows are presented in random patterns and locations, assuming that human participants scan the arrows sequentially and terminate the scan when the majority of the arrows can be determined, we can compute the computational load in terms of bits under different input conditions. Let 0 and 1 represent left and right pointing directions, respectively. For set size 1, there are only two possible outcomes: 0 or 1. For set size 3, there are four combinations: 000, 001, 011, 111 (disregarding the order of the digits in each combination). For the set size 5, there are six combinations: 00000, 00001, 00011, 00111, 01111, and 11111. For set size 1, only one arrow with two alternatives has to be scanned. Therefore the computational load in bits is 1. For set size 3, if three arrows point in the same direction (000, or 111), only two arrows need to be scanned. Therefore, the computational load in bits is 2. However, if only two arrows point in the same direction, corresponding to the three patterns of 001, 010, and 100 (considering the order), there will be 2, 3, or 3 arrows that need to be scanned sequentially starting from the left and moving to the right. The same number of bits applies to the combinations of stimulus condition 011. On average 2 2/3 arrows need to be scanned. Therefore, the average computational load in bits is 2 2/3, with the best-case of 2 and worst-case of 3. Similarly, for set size 5, having 5, 4, or 3 arrows pointing in the same direction, the average the computational load in bits is 3, 3 3/5, and 4 1/2, with the best-case being 3 for all three conditions, and worst-case being 3, 4, and 5 for each of the three conditions, respectively (see \nTable 2\n).10.1371/journal.pone.0003522.t002Table 2Experimental conditions and estimates of input information and computational load under self-terminating and grouping search algorithms.Set sizeStimulus conditionRatioInput digitsSelf-terminating searchGrouping searchBestWorstAverageGroup a\nGroup sizeScan digitslog2\n10,11∶011111110.003000, 1113∶032221221.00001, 0112∶13232 2/33262.58500000, 111115∶053331331.5800001, 011114∶15343 3/52.537.52.9100011, 001113∶25354 1/2103304.91Note: a Number of grouping attempts on average required to obtain a congruent sample.\n\nFigure 2A\n shows a plot of the mean RT as a function of computational load in terms of bits assuming the self-terminating search algorithm was used. The regression analysis with RT as the dependent variable and computational load in bits as the independent variable was conducted. For the average case, RT = 110+312 · bits, R2 = 0.82, F(1, 4) = 17.73, p<0.05, indicating a good fit. Linear mixed-effects model analysis with computational load as the fixed effect and subject as the random effect showed that Akaike's information criterion (AIC) was 1945.06; the intercept was significant, F(1, 131) = 4.88, p<0.05; and the computational load was significant, F(1, 119) = 423.33, p<0.01. However, the RT of the incongruent condition of set size 3 (2∶1 condition, with a computational load of 2 2/3 bits) was significantly longer than the congruent condition of set size 5 (5∶0 condition, with a computational load of 3 bits), 1121 vs. 724, t(23) = 6.08, p<0.001. Given that the self-terminating algorithm predicts less computational load in the former condition than the latter, this evidence is against the self-terminating search and suggests that participants may have adopted strategies beyond the self-terminating search to perform the task.10.1371/journal.pone.0003522.g002Figure 2Reaction time (RT) as a function of computational load which is determined by processed information in bits, on average, assuming that the self-terminating search algorithm was adopted (A), and that the grouping search algorithm was adopted (B).The grouping search algorithm better predicts the linear relationship between the RTs and computational load relative to the self-terminating search algorithm.Grouping searchThe algorithm implemented in the human brain may not be as simple as the above-mentioned self-terminating search. When human participants analyze patterns in order to make a response, they may adopt a strategy in which they attempt to group and sample arrows with a majority size (over half of the total set size) based on their directions in order to achieve a high efficiency. For example, for a set size of 3, intuitively, a Boolean circuit of (a1 XNOR a2) OR (a1 XNOR a3) OR (a2 XNOR a3) makes sense. Here the exclusive nor (XNOR) returns “true” if input a1 and a2 are identical, and returns “false” if they are different. If the grouping of (a1 XNOR a2) returns true, the search can stop. However, this grouping (or sampling) process may have to be performed several times independently or recursively before a solution is reached based on a congruent sample. Therefore, more computation would be required under near-tie high uncertainty conditions within a certain set size than what would have been predicted based only on the self-terminating search algorithm. It is noteworthy to indicate that the maximum grouping size of 3 arrows should be within the capacity limit of the locations that can be selected at once [15].Let us assume that participants adopt such a grouping (sampling) strategy and search for a congruent sample with a majority grouping size. For set size 1, only 1 arrow needs to be scanned. For set size 3, for the condition in which all 3 arrows point in the same direction, only 1 grouping attempt needs to be made with 2 arrows being scanned; and for the condition in which 2 arrows point in the same direction, there will be 1 successful grouping out of an average of 3 attempts. Therefore, 6 arrows, the product of 3 grouping attempts and group size of 2 arrows, need to be scanned. Similarly, for set size 5, for the conditions in which 5, 4, or 3 arrows point in the same direction, 1, 2.5, 10 grouping attempts on average need to be made and 3, 7.5, and 30 arrows need to be scanned, respectively (see \nTable 2\n).If we use the majority group size (1, 2, and 3 for set sizes of 1, 3, and 5, respectively) as the information unit, assuming that each sampled group is equivalent to one unit of information, the information to be processed (i.e., the computational load) is logg (s), where the base g represents the group size and s is the number of arrows to be scanned. To convert this measure to bits (i.e., from base g to base 2), it is multiplied by log2 (g) [9]. Therefore, the computational load is log2 (g) • logg (s), which is equivalent to log2 (s). Here we converted the information to be processed in each condition to bits based on the average number of arrows that need to be sampled. It is worth noting that the calculation of 0 bits for the set size 1 condition does not mean that 0 bits of information need to be processed. The decision making step of the majority direction needs 1 bit. We can add 1 bit to all conditions, but this should not affect the general predictions of the grouping search algorithm.\n\nFigure 2B\n depicts RT as a function of computational load assuming the grouping search algorithm was adopted. The regression analysis with RT as a function of computational load of the grouping algorithm was also conducted on the group data. With the log2 (scan arrows) (i.e., computational load in bits) as the independent variable, RT = 458+242 · bits, R2 = 0.96, F(1, 4) = 105.48, p<0.001. Linear mixed-effects model analysis with computational load as the fixed effect and subject as the random effect showed that AIC was 1823.29; the intercept was significant, F(1, 38) = 246.36, p<0.01; and the computational load was significant, F(1, 119) = 1398.48, p<0.01. Compared to the self-terminating search model, the grouping search model fits the data better because the AIC value of the linear mixed-effects model for the grouping search was lower than for the self-terminating search.Self-report of the strategy adopted by participantsParticipants were queried regarding the strategy that they each employed during the task at the end of the study. Of 24 subjects, 17 reported that they scanned the stimulus display until they found either 2 arrows pointing in the same direction (for set size 3) or found 3 arrows pointing in the same direction (for set size 5). Eight of these subjects spontaneously mentioned that they found the task to be the easiest when 2 or 3 arrows pointing in the same direction were grouped together, and the remainder agreed that the task was the easiest when this occurred when prompted by the experimenter. Six participants spontaneously described use of a grouping strategy. For example, one participant reported, for a set size of 5 arrows, first identifying a group of three arrows. If all arrows in the group pointed in the same direction, the participant made the appropriate response. If only two of the three did, the participant scanned the rest of the display for a third arrow pointing in the same direction. If grouping was not possible, e.g. the arrows were evenly distributed about the crosshair, then a serial scanning strategy was adopted.DiscussionOther than exhaustive search, self-terminating search, which incorporates an additional stopping rule in which the participant scans the arrows one by one until the majority threshold is reached (e.g., 2 arrows pointing in the same direction within a set size of 3 arrows), is clearly the second most straightforward algorithm. Consistent with this algorithm, Figure 2A reveals that RT increases with the computational load in two cases: (a) across the three congruent conditions, in which all arrow(s) in each set point in the same direction, as indicated by the dashed line; and (b) within the two conditions of set size 3 and within the three conditions of set size 5, as indicated by the solid lines. However, the opposite prediction for the 2∶1 and 5∶0 conditions based on self-terminating search stands as evidence against the possibility that subjects adopted this algorithm.The fact that human visual attention can be directed towards more than one item simultaneously allows for the possibility of a grouping search algorithm, in which participants first select a sample of arrows with a size equal to the majority threshold and then process the sample. This is similar to perceptual grouping [4]. If all arrows in the sample happen to point in the same direction (congruent), then a response can be quickly generated. If not, a re-sampling kicks in until a congruent sample is found and a response is generated. We estimate the computational load for the grouping search as a logarithmic function of the product of the grouping size and the expected number of groups that need to be sampled in order to obtain one that is congruent. It is clear that RT increases monotonically as a function of the computational load and that this relationship is well approximated by a linear function (Figure 2B). The linear relationship between RTs and the computational load based on the grouping search strategy may suggest a tree-like structure representing the arrows to be sampled and a dichotomizing test.These results support the idea that RT is determined not only by the amount and content of the input but also by the algorithms of mental operations that people adopt in the face of information uncertainty. Situations in the real world are often more complex than laboratory choice-RT tasks and require more voluntary control. The majority function task is interesting in that it requires greater voluntary control of computation than tasks used for testing the conflict effect (e.g., [5]), although it also uses conflicting information to manipulate information uncertainty. With this task we highlight the role of voluntary control during information processing and provide a more general framework to account for the conflict effect. For example, in a variation of the flanker task [16] in which people were asked to detect the direction of the target arrow and ignore the distracters, we observed a typical conflict effect–the RT difference between the incongruent and congruent conditions–of about 50–150 ms. This can be accounted for by the computational load framework because the computational load for the congruent condition is 1 bit for the two alternative responses, whereas it is less than or equal to 2 bit for the incongruent condition because of incongruent flankers.In addition, performance in the majority function task cannot be fully predicted by a conflict effect account. For example, comparing two conflicting conditions in which the distribution of arrows are 2∶1 and 4∶1 in set sizes 3 and 5 respectively, the RT of the latter condition is significantly longer (1121 vs. 1261, t(23) = 4.55, p<0.001) , which is opposite of what the conflict effect account predicts, since the non-target to target ratio is larger in the former. Similar to other categorization tasks [12], the goal of computation is to identify the majority based on the input. Because any arrow could potentially belong to the majority subset if the set size is equal to or greater than 3, more than one arrow needs to be processed, either scanned one-by-one or randomly sampled and grouped as we tested above. However, the degree of uncertainty caused by conflicting information predicts RT within a given set size if the grouping search is adopted. For example, for the set size 5, the most uncertain condition with the distribution of arrows of 3∶2 requires 10 grouping attempts to be made on average before a solution is reached based on the grouping search algorithm. This may explain why its RT is much longer (1615 ms) compared to another less uncertain but also conflicting condition with the distribution of arrows of 4∶1 (1261 ms), which requires only 2.5 attempts on average to obtain a congruent sample.The majority function task reported in this paper has features of conflict, grouping, and input variation that are often elements of many separate tasks in the literature. The methods to compute the computational load in this task may be used to account for discrepancy between findings of previous studies on conflict effect using different tasks. This majority function task is similar to the visual motion task used in studies of perceptual decision making (e.g., [17]), but here we examine and model decision making on a system level by considering the algorithms potentially adopted by the brain to process discrete information. We cannot exclude other possible factors that might contribute to the current results such as the Gestalt effect based on the holistic perception of all congruent arrows, information reduction [18], or perceptual grouping [4]. This may account for overall faster responses and the relatively flat slope for the congruent conditions. Although certain common mechanisms might be involved in voluntary control, the underlying algorithms will vary and be task specific in different situations depending on different computational goals [19]. For example, under the high input information condition, which is beyond the grouping capacity limit (e.g., more than 5 arrows), other algorithms, such as those suggested for perceptual decision making regarding motion coherence, might be adopted by humans to find the majority.We argue that voluntary control is implemented by algorithms of mental operations, which are in turn implemented by brain networks. This study demonstrates that it is important and plausible to analyze the underlying algorithms for voluntary control by examining the relationship between the amount and content of input and RT. RT is a basic and central measure of mental operations in almost all cognitive tasks [20]. Early studies based on information theory [9] have found that choice RT is determined by the amount of information in bits that has to be processed to generate a correct response (e.g., [18], [21], [22]), though the causality in this relationship has been challenged [23], [24]. Some elegant models for the central mechanisms of choice RT have been proposed, and changes in RT as a function of information processing have been studied in the context of perceptual decision making (e.g., the sequential-sampling models, for a review, see ref. [25]), mental addition (subtraction) [26], visual search [27], and categorization [28]. In this study, we explicitly considered the underlying algorithms for voluntary control of information processing.\n\nREFERENCES:\n1. Information theory: Applications of information theory: Physiology. Britannica Online\n2. VerghesePPelliDG\n1992\nThe information capacity of visual attention.\nVision Res\n32\n983\n995\n1604866\n3. PosnerMIRothbartMK\n1998\nAttention, self-regulation and consciousness.\nPhilos Trans R Soc Lond B Biol Sci\n353\n1915\n1927\n9854264\n4. TreismanA\n1982\nPerceptual grouping and attention in visual search for features and for objects.\nJ Exp Psychol Hum Percept Perform\n8\n194\n214\n6461717\n5. EriksenBAEriksenCW\n1974\nEffects of noise letters upon the identification of a target letter in a nonsearch task.\nPerception and Psychophysics\n16\n143\n149\n6. StroopJR\n1935\nStudies of interference in serial verbal reactions.\nJournal of Experimental Psychology\n18\n643\n662\n7. HopfingerJBBuonocoreMHMangunGR\n2000\nThe neural mechanisms of top-down attentional control.\nNat Neurosci\n3\n284\n291\n10700262\n8. WalusKSchulhofGJullienGAZhangRWangW\n2004\nCircuit design based on majority gates for applications with quantum-dot cellular automata. Thirty-eighth Asilomar Conference on Signals, Systems & Computers.\nPacific Grove, , California\nIEEE, Piscataway NJ, ETATS-UNIS\n9. ShannonCEWeaverW\n1949\nThe Mathematical Theory of Communication.\nUrbana\nUniversity of Illinois Press\n10. CarrascoMMcElreeB\n2001\nCovert attention accelerates the rate of visual information processing.\nProc Natl Acad Sci U S A\n98\n5363\n5367\n11309485\n11. CarrascoMMcElreeBDenisovaKGiordanoAM\n2003\nSpeed of visual processing increases with eccentricity.\nNat Neurosci\n6\n699\n670\n12819786\n12. SmithEE\n1968\nChoice reaction time: an analysis of the major theoretical positions.\nPsychol Bull\n69\n77\n110\n4867597\n13. PosnerMIRossmanE\n1965\nEffect of size and location of informational transforms upon short-term retention.\nJ Exp Psychol\n70\n496\n505\n5833671\n14. SternbergS\n1966\nHigh-speed scanning in human memory.\nScience\n153\n652\n654\n5939936\n15. FranconeriSLAlvarezGAEnnsJT\n2007\nHow many locations can be selected at once?\nJ Exp Psychol Hum Percept Perform\n33\n1003\n1012\n17924803\n16. FanJMcCandlissBDSommerTRazAPosnerMI\n2002\nTesting the efficiency and independence of attentional networks.\nJ Cogn Neurosci\n14\n340\n347\n11970796\n17. NiwaMDitterichJ\n2008\nPerceptual decisions between multiple directions of visual motion.\nJ Neurosci\n28\n4435\n4445\n18434522\n18. PosnerMI\n1964\nInformation reduction in the analysis of sequential tasks.\nPsychol Rev\n71\n491\n504\n14216895\n19. MarrD\n1982\nVision.\nNew York\nFreeman\n20. PosnerMI\n1986\nChronometric Explorations of Mind.\nNew York\nOxford University Press\n21. HickWE\n1952\nOn the rate of gain of information.\nQ J Exp Psychol\n4\n11\n26\n22. HymanR\n1953\nStimulus information as a determinant of reaction time.\nJ Exp Psychol\n45\n188\n196\n13052851\n23. FittsPM\n1964\nPerceptual-motor skill learning.\nMeltonAW\nCategories of Human Learning\nNew York\nAcademic Press\n24. KornblumS\n1968\nSerial-choice reaction time: inadequacies of the information hypothesis.\nScience\n159\n432\n434\n5634661\n25. SmithPLRatcliffR\n2004\nPsychology and neurobiology of simple decisions.\nTrends Neurosci\n27\n161\n168\n15036882\n26. DehaeneSSpelkeEPinelPStanescuRTsivkinS\n1999\nSources of mathematical thinking: behavioral and brain-imaging evidence.\nScience\n284\n970\n974\n10320379\n27. TreismanAMGeladeG\n1980\nA feature-integration theory of attention.\nCognit Psychol\n12\n97\n136\n7351125\n28. MervisCBRoschE\n1981\nCategorization of natural objects.\nAnn Rev Psychol\n32\n89\n115"
+}

batch_13/PMC2569064.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2569064",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2569064\nAUTHORS: Vincent Fauveau, Della R Sherratt, Luc de Bernis\n\nABSTRACT:\nA crucial question in the aim to attain MDG5 is whether it can be achieved faster with the scaling up of multi-purpose health workers operating in the community or with the scaling up of professional skilled birth attendants working in health facilities. Most advisers concerned with maternal mortality reduction concur to promote births in facilities with professional attendants as the ultimate strategy. The evidence, however, is scarce on what it takes to progress in this path, and on the 'interim solutions' for situations where the majority of women still deliver at home. These questions are particularly relevant as we have reached the twentieth anniversary of the safe motherhood initiative without much progress made.In this paper we review the current situation of human resources for maternal health as well as the problems that they face. We propose seven key areas of work that must be addressed when planning for scaling up human resources for maternal health in light of MDG5, and finally we indicate some advances recently made in selected countries and the lessons learned from these experiences. Whilst the focus of this paper is on maternal health, it is acknowledged that the interventions to reduce maternal mortality will also contribute to significantly reducing newborn mortality.Addressing each of the seven key areas of work – recommended by the first International Forum on 'Midwifery in the Community', Tunis, December 2006 – is essential for the success of any MDG5 programme.We hypothesize that a great deal of the stagnation of maternal health programmes has been the result of confusion and careless choices in scaling up between a limited number of truly skilled birth attendants and large quantities of multi-purpose workers with short training, fewer skills, limited authority and no career pathways. We conclude from the lessons learnt that no significant progress in maternal mortality reduction can be achieved without a strong political decision to empower midwives and others with midwifery skills, and a substantial strengthening of health systems with a focus on quality of care rather than on numbers, to give them the means to respond to the challenge.\n\nBODY:\nBackgroundAs the international public health community marks the twentieth anniversary of the Safe Motherhood Initiative [1], more than 530 000 women still die each year from complications of pregnancy and childbirth, over 90% of them in South Asia and sub-Saharan Africa. Additionally, 10 to 20 million women annually suffer severe health problems as a result of pregnancy and childbirth, such as obstetric fistula or chronic infection. Seventy percent of maternal deaths are due to five major complications, the majority of which occur during labour, delivery and the post partum period. Approximately 15% of women will experience a complication during pregnancy, childbirth or the immediate postpartum period – most of which cannot be predicted, but almost all of which can be managed. Most maternal death and disability could be averted if:• all pregnancies were wanted,• all births were attended by skilled health professionals and• all complications were managed in quality referral facilities offering emergency obstetric care [2].While the focus of this paper is on the second of these conditions, it must not be forgotten that a large part of maternal deaths could be avoided if all women had access to family planning and reproductive health services. It must also be acknowledged that the interventions to reduce maternal death also significantly contribute to reducing newborn mortality.Saving mothers' lives is widely recognized as an imperative for social and economic development, as well as a human rights imperative, although until recently there has been limited evidence mapping such links[3]. It is the basic right of every woman and baby to have the best available care to enable them to survive pregnancy and childbirth in good health. Yet, while the techniques and strategies to address maternal health are well known and widely accepted, and the need for access to specialist emergency obstetric care services has a high level of evidence [4], the factor most neglected in the last decade was human resources required to implement these interventions. Although there is a general consensus that maternal mortality and morbidity cannot be reduced without midwives and others with midwifery skills, the numbers of these skilled providers have not significantly increased over the last two decades. Moreover, the actual numbers of skilled midwifery providers has started to decrease in some countries, as the result of migration, losses from HIV/AIDS and dissatisfaction with remuneration and working conditions. At the same time issues of quality of care remain crucial, particularly where health systems do not play their supportive role, as in many countries that have embarked in scaling up the number of community-based providers without giving sufficient attention to their skills. The World Bank estimates that maternal deaths would decrease by 73% if coverage of key interventions rose to 99% [5]. Access to essential maternal health care services, however, is riddled with inequities. The lower a woman's economic status, the less likely she is to have skilled assistance at delivery and lifesaving emergency obstetric care [2]. Geographical location, ethnicity and age are also related to disparities in access.WHO initiated a decade of special attention to the health workforce with the World Health Report 2006, 'Working together for Health'[6]. UNFPA, working jointly with the International Confederation of Midwives (ICM), plans to contribute to this global initiative on the health workforce by initiating in collaboration with their partners a global campaign to promote and rapidly scale-up the coverage of midwifery care. Midwives and others with midwifery skills are the representation of UNFPA's mandate within the health workforce, not only for their role in providing skilled delivery care, but also for their ability to deliver the essential sexual and reproductive health package in relation to maternal health. In addition, efforts to strengthen midwifery are also in line with UNFPA's mandate to promote gender equality, as midwives are key female members of the health workforce. However, for many reasons, some having to do with the fact that most midwives are women, there has been gross underinvestment, and sometimes no investment at all, in building or maintaining a cadre of professional midwives. In addition, midwives very often have low status within their community and receive little recognition. The vast majority of midwives thus suffer from the same gender-related inequalities as other women. The result has been insufficient investment in midwifery training, deployment and supervision, coupled with inadequate regulation and policies to support and protect midwives in their practice. Yet, without expert midwives to teach midwifery skills and supervise others, ensuring quality of care will not be possible and efforts to reduce maternal and newborn deaths will fail. A number of countries or states – particularly Sri Lanka, Malaysia, Tunisia, Thailand, Kerala, Tamil Nadu – have, however, successfully undertaken specific measures to make midwifery a respectful and attractive profession. Policy, advocacy and revision of regulatory systems were instrumental in order to professionalize midwifery and remove discriminatory legislation.The Millennium Development Goal 5 highlights the crucial role of midwives and others with midwifery skills on the path to improved maternal health by including as its second indicator the proportion of births attended by skilled health providers. Although the percentages are not specified, it is assumed that the target for 2015, \"universal access to a skilled birth attendant\", translates into between 90% and 100% coverage. Currently it is estimated that no more than 40% of births in low-income countries are assisted by properly skilled attendants – highlighting the large effort needed to reach the target of 90% coverage by 2015 [7]. According to WHO [2], an additional 334 000 midwives are required to fill this gap, not counting the number of doctors and other nurse providers. It can be argued that at least twice as many are required to achieve universal access to a full package of sexual and reproductive health care.In the past few years, the international public health community has made two significant advances. One by incorporating in to the new global health partnerships the health care professional organizations such as the International Confederation of Midwives (ICM) and the International Federation of Gynecology and Obstetrics (FIGO). The other by highlighting the key role of human resources for health (HRH) in the failure of health systems and the need to address HRH in priority in health system strengthening initiatives (GAVI-HSS, GFATM, Global Business Plan, Global Campaign for Health MDGs, International Health Partnership, etc).This paper aims at contributing to generating a massive effort to increase not only the coverage of all births by skilled attendants, but also the quality of this attendance by promoting the role of midwives and others with midwifery skills in improving maternal, reproductive and newborn health. The question, however, is whether countries should give priority to producing a relatively high number of multipurpose community-based providers to cover all villages or to produce a lower number of specialized, facility-based, professional and skilled maternal health providers [8].Situation and challengesEnsuring equitable access to a continuum of skilled care before, during and after childbirth, is recognized as a universal human right, and is critical for saving the lives of mothers and for their newborns [2,9-11]. However, skilled care requires skilled providers – a scarce commodity in most low-income countries. Much of the efforts in the lead up to the 20 year marking of the Safe Motherhood Initiative (SMI), have focused on the barriers to skilled care are at birth, among which the lack of qualified human resources appears the most challenging.The lack of skilled providers linked to a facility offering quality emergency obstetric and neonatal care (EmONC), is neither a new phenomena, nor is it only a problem of low-income countries. The need to invest in training of the midwifery workforce and ensuring that midwifery providers have appropriate life-saving skills have been topics of debate for many decades [12,13]. Yet, as estimates for the proportion of births attended by a skilled provider shows, the majority of women in developing countries still give birth without such assistance and the data reveals huge disparities and inequity, with women in low income families having little options or opportunities to access such healthcare [2,7]. However, the lack of access to health services occurs for a variety of reasons and not just because of lack of healthcare providers [14].A 'skilled birth attendant' (SBA) has been defined by the WHO in collaboration with the ICM and FIGO and has been endorsed by UNFPA, the World Bank and the International Council of Nurses in 2004 [15]. The definition builds on and seeks to add clarity to the initial definition in the 1999 Joint statement on Maternal Mortality [16] and the one developed by the Interagency Group for Safe Motherhood in 2000 [17], and sets better the minimal requirement for a skilled birth attendant.The 2004 definition states that a skilled birth attendant is: \"an accredited health professional – such as a midwife, doctor or nurse – who has been educated and trained to proficiency in the skills needed to manage normal (uncomplicated) pregnancies, childbirth and the immediate postnatal period, and in the identification, management and referral of complications in women and newborns.\" [15]As the above definition clearly shows SBAs are not a single cadre or professional group. SBAs are providers with specific midwifery competencies; they perform these competencies as professional midwives or, if trained in these competencies as general practitioners with midwifery competencies, or as nurses. Furthermore, not only must they have received proper training to carry out their tasks, but they must have developed the competencies to a level of proficiency. The total list of competencies for each type of skilled attendant will vary between the different professional groups, according to the scope of practice for each group. The list may even vary for cadres with same professional title in different countries, depending on the legislation and regulations and training curricula for each cadre. The common denominator, however, is the basic skills required to assist a woman during pregnancy, childbirth and after birth, including essential care to newborns – known internationally as 'midwifery skills' and defined as \"core competencies\". In addition, experts agree that the education of nurses and midwives must include development of problem-solving competencies, because the arrival of a woman at a referral facility is often the end of a long and complex decision-making process, influenced by the interpersonal relationships between the woman, her family members and the health providers. [18]Moreover it is known that to be effective, healthcare providers must work in a supportive enabling environment – which must include basic equipment and drugs as well as good communication and transportation systems – to ensure timely referrals when needed and have effective and supportive supervision. Yet, too often, the enabling and supportive environment is also lacking.Midwifery skillsThe 'core competencies' required of any skilled birth attendant outlined in the 2004 WHO ICM FIGO statement were intended to apply to any health worker providing midwifery care at any level of the healthcare system, including the primary care level. Included within the core competencies are the basic EmOC skills to which essential neonatal care has been added, as well as essential maternal and neonatal healthcare for preventive and promotional care and care of women and newborn with no complications. The list of 'additional competencies' was added in the 2004 statement to apply to those skilled birth attendants working in peripheral and or isolated settings, where referral to a district hospital is difficult. Whereas the 'advanced skills' are the surgical competencies required for comprehensive care (EmONC).Contention however remains as to which maternal health providers should have these core competencies. Is it all maternal health care providers? And, who should have just the core and who should have advanced or additional competencies? Moreover, the discussion on which maternal health workers can be trained or 'up-skilled', to ensure they have the required competencies to a level of proficiency, is causing concern in many countries.Even if there was a consensus on the above questions, there remains the issue of the maintenance of these competencies. And the issue of whether the legal and regulatory framework properly protects the rights of the healthcare provider to perform the life-saving interventions for maternal and newborn survival. Often they are seen as the prerogative only of physicians. Therefore, becoming competent, or scaling up the competencies of the maternity workforce, is only part of the overall issue to be addressed. To develop and implement a plan for the adequate production of their maternity workforce, the countries need to know how many of which type are needed, where they should be deployed, and how to retain them at their post, especially those working in rural areas.Why have the critical midwifery competencies been so neglected?One of the major reasons explaining why so many countries still have inadequate numbers of skilled midwifery providers is because those grappling with human resources have not paid attention to the need for 'proficiency' in the various competencies required to assist women and newborns. For too long it has been accepted that as long as the health worker received some (often too little) training in midwifery, this was sufficient. Too often there has been a lack of understanding and appreciation of the difference between competence – the ability to carry out a task to the required standard – and competencies, the discreet knowledge, skills, attitudes and experience required for individuals to perform their jobs correctly and proficiently [19].Additional reasons for the current shortfall in midwifery skills in many low-income countries include the lack of understanding and appreciation of what the professional midwife can offer, as well as an historical prioritisation on medical training of physicians over other healthcare providers. As argued in the World Health Report 2005, many countries facing current shortages of midwifery providers have been at the mercy of misguided, albeit well intentioned, advice from external donors recommending policy changes to create a multipurpose worker [2,20] or seeing midwifery care as a voluntary occupation that can be performed by a traditional healer or traditional birth attendant.Investing in a specialist midwifery provider is challenging in many countries because midwifery, as a predominantly female profession, does what is predominantly considered 'women's work' [21]. The double burden of being a woman, herself subject to gender inequalities, as well as being a female worker, puts tremendous pressure on midwives who do a very emotional and stressful job that can lead to high levels of occupational 'burn-out' [22-24]. Having responsibilities for their own home and child care, etc., and working with women in what some perceive as a female area – pregnancy and birth – is made even more difficult in those situations where women's status is low and where assisting childbirth is seen as low status or culturally unclean. On a positive note however, where midwives are respected they can, by working in the community, in close proximity to families, have the potential for offering career aspirations to girls and young woman and in so doing, may contribute to efforts to address gender inequity. [21]The failure of governments to provide competent, skilled midwifery health workers has been seen by some as a blatant case of gender inequality or lack of gender sensitive health policy [25]. Failure of governments to provide basic healthcare for the most vulnerable of its citizens at the most vulnerable time of life can be viewed in the light of the Committee on Economic, Social and Cultural Rights' General Comment 14 as a failure of governance [26].Why invest in midwives and others with midwifery skills?Investing in a specialist cadre of midwifery provider-professional midwives or others with midwifery skills – has been shown to make a difference in reducing maternal mortality in many countries. Indeed, historical evidence tells us that the countries that have succeeded in reducing their maternal mortality and morbidity have done so by ensuring skilled care at ALL births [8,27-29]. In particular, they have achieved this by ensuring that all home births were undertaken by 'trained and supervised midwives or, as in the case in Sweden and the UK, by making sure midwives not only referred all complicated cases – having first rendered first aid and offered first line management – but also reported all births and maternal deaths to the local public health physician or district health authority. [30]. Reviewing case studies from countries that have in recent years succeeded in reducing their maternal mortality ratio, Koblinsky suggested that, \"assistance at birth by a skilled birth attendant in the home or any health facility, supported by a functioning referral system, can reduce the MMR down to around 50 or below\" [28]. The recent Lancet series on maternal survival also point to the value of midwives working as a team in health centres [31]. Indeed, home delivery is not a good use of the time of scarce professionals, who should be concentrated in health centres.For skilled attendants to effectively contribute to achieving the MDGs however, they must be accessible, offer affordable women-centred care, and must be seen as a member of the health system and to be credible. For this they must be technically competent. Being seen by the community as a specialist in midwifery care contributes to credibility. The outstanding evolutionary feature of maternity-related health services in Sri Lanka and Malaysia is the pivotal role of trained and government employed midwives. They have been relatively inexpensive to both countries, yet they have been the cornerstones for the expansion of an extensive health system to rural communities. They have provided accessible maternity services in hospitals and communities, gained sustained respect from the communities they serve, and are described with affection and admiration by managers and policymakers in each country' [32]. As found in a study on access to emergency obstetric care and human resources in Tanzania, there is a positive correlation between having a professional qualification and clients' willingness to use health services [33].Professional midwives or others who meet the international definition of a midwife [34] (regardless of their title) and practice according to ICM's evidence-based essential midwifery competencies [35] do have all the essential basic midwifery competencies required for the provision of high quality skilled midwifery care, and more. Where they work in partnership with women and are acceptable by women and their communities, professional midwives (or those functioning with legal protection as a professional midwife) offer countries potential for meeting the broader reproductive health needs of communities [21,36], as well as contributing to universal primary health care for all [37]. As history has shown, midwives can be most useful in helping to ensure that health services reach those in greatest need, the poor and hard to reach communities [38,39].Quality or quantity?While there is a need to build the capacity of the maternity workforce in terms of quantity in order to reach out to all communities, it is even more important to consider quality. The debate on whether to prioritise quality or just have more numbers is at the heart of current discussions on skilled attendants, and strategic decisions are likely to have a strong impact on maternal mortality. Whilst everyone agrees it is not effective to look at human resources for health for a specific health issue in isolation [40], we argue that MNH services do have several unique characteristics that require specific attention when making decisions about the size, shape and production of the midwifery workforce. Specifically the need exists for:• High levels of technical competence in a number of very specific areas, both curative and promotive in nature. Maternal mortality reduction shows the greatest sensitivity to the presence of skilled maternal health providers [41].• Appropriate curricula that ensure sufficient time for hands-on practical training to become competent to the level of proficiency in all the requisite areas, as complications can arise quickly and without warning. What is required is repeated reflexive and intelligent practice [42,43]. Clinical instruction and mentorship are also paramount. Trainers must themselves be proficient in these competencies, although unfortunately in many low-income countries they are not [44].• Gender sensitivity. Although this can apply to all health service access [45], lack of a female provider is perceived as one the major barriers to why women do not use maternal health care [46,47].• Excellent inter-personal communication and cultural competencies, because of the high cultural sensitivity of pregnancy and birth. Nowhere else are interpersonal skills, linguistic skills and cultural appreciation more crucial to help the families with decision making in all aspects of reproductive health [18,46,47].• Motivation for the job – has been shown to be vital for providing quality care [48-50]. Midwifery providers must be available at all hours of the day and night – whenever birth takes place. Among the criteria that should be considered are demonstrating professionalism and positive attitude to patient, avoiding impersonal routine response, and resisting to corruption [51].For all the above reasons it is essential that curricula and training programmes prioritise midwifery skills – but sadly many current training programmes do not. Far too often, midwifery skills are seen as accessory, or add-on skills, and are afforded little time, typically at end of a programme, where there is little time for repeated hands-on practice.In terms of numbers, the largest barrier to overcome is the need for sufficient teachers and trainers who are competent in education and in midwifery theory and in clinical practice. Deciding on numbers depends on a complex set of criteria: number of training institutions and teachers, caseload, overall education standards, reservoir of suitable entrants, but also recruitment policies, fiscal space and budget. Historically, a population base ratio has been used to estimate the number of midwives needed in a given country. The most widely used ratio of one midwife to 5000 population developed by WHO in 1993 [12], assumes that one community midwife would be able to care for 200 pregnant women a year, including assisting at their births and giving postnatal follow up care. The ratio however does not take account of the skill-mix needed to care for obstetric emergencies, nor the different geographical circumstances, differences in fertility rate nor other personal or professional work demands on the midwife. UNFPA has recently called for using a new \"births by midwife\" indicator i.e. the number of births expected to be attended in all security by a qualified midwife [36] (see Figure 1).Figure 1Expected births per midwife ratio in selected countries.To achieve the right balance between numbers and quality, adequate funds and a cost-effectiveness analysis are necessary, in turn dependant upon having policies and strategies in place. To avoid repetition of past mistakes and the selection of misguided strategies, technical competence is critical to guide the decision.Towards solutions: key areasTime to scale up is limited. However, as countries like Indonesia have experienced, rapid scale up in numbers without ensuring full competencies of midwifery providers can be costly in terms of in-service training needs [52]. It is also possible to improve access to skilled care by better utilization of existing staff, and training mid-level providers in tasks that are usually undertaken by physicians [53,54]. Each country will need to take a considered approach, allowing fast scale-up while at the same time maintaining, or improving, quality. While there is a need to address the deficiencies in specific obstetric skills, especially surgical skills and specialist neonatal skills, it is the midwife who will ensure access to all. Graham et al estimate that on average there should be a minimum of five midwives for 1 obstetrician (or physician with obstetric skills) [55]. Midwives are also required to develop community capacity in order for communities to take their place in monitoring and evaluating maternity services and contributing to overall quality improvements [47].Midwives and other midwifery providers perform best within a multi-professional team of health workers – including peers – but also support workers who can conduct some of the non-specialist midwifery tasks under their supervision. Physicians with obstetric skills or mid-level providers with obstetric competencies (such as in selective surgical procedures) are best targeted at referral centres where surgery is possible. This partnership should be based on mutual respect and appreciation for each other's contribution, rather then on an outdated historical hierarchical model, which sees the midwife or other mid-level worker as subservient to the physician.In addition to training, capacity building and capacity-development require attention to structure, systems, roles, support, supervision, as well as logistics [56]. Above all, any new initiative must have inbuilt from the beginning a robust monitoring and evaluation systems, not only to demonstrate when progress is being made, but also to monitor quality improvement and future decision making that is at the heart of any capacity-development initiative [57].During the 1st International Forum on midwifery in the community held by UNFPA, ICM and WHO in 2006 [[58] and Additional file 1], a framework was proposed for rapid scale-up of midwifery providers, based on a capacity development model. The framework identifies seven interconnected areas of work (Figure 2):Figure 2Framework for addressing issues of scaling-up midwifery for the community level.1. Policy, legal and regulatory frameworks2. Ensuring equity to reach all3. Recruitment and education (pre- and in-service), accreditation,4. Empowerment, supervision and support5. Enabling environment, systems, community aspects6. Tracking progress, monitoring and evaluation, numbers and quality7. Stewardship, resource mobilization1. Policy, legal and regulatory frameworksAll the above areas of work are interrelated, but political and legislative action must be in the forefront. The protection to which mothers and children are entitled under the right to health framework cannot be regarded as 'charity'. It is an obligation of governments, irrespective of adverse conditions such as severe shortage of economic resources [2,9,21,22,59-61]. While governments cannot be held responsible for the actual care or omissions of care given by individual practitioners, they are responsible for ensuring that adequate mechanisms are in place for regulation, delegation of authority and training of the providers and that appropriate policies are implemented. Legal and regulatory frameworks are also needed to protect midwifery and medical providers.Action: create a coalition of interested stakeholders, including professional associations, to promote and influence policy changes. Such partnerships should be built on mutual respect and include community participation, for example civil society groups, from the start.2. Ensuring equity in reaching the poorIn all countries poverty is strongly associated with less access and use of healthcare, including skilled midwifery care at birth [62,63]. Evidence shows that even in relatively low-income groups, women with higher levels of autonomy find it easier to access maternal health services [64]. Furthermore, evidence shows that introduction of formal user fees and demands for payment 'under the table' have a negative influence on utilization of maternal health care services, particularly during childbirth [46,48].Action: making equity a national cause, in collaboration with and involving from the beginning he wider stakeholder group, such as the other ministries, and civil society, NGOs, faith-based and private healthcare providers, media and parliamentarians.3. Recruitment and education (pre- and in-service), accreditationRecruiting from and providing education within the local area can help ensure that service provision is culturally appropriate. Both pre-service and in-service education and training programmes should be based on a competency model, with those who teach midwifery in clinical or classroom settings being themselves competent in midwifery and having undertaken adequate preparation for their role. More work is needed to ensure that pre-service midwifery programmes have a better client-centered basis [51]. Improving quality of care depends on the new graduates' ability to practice their newly acquire skills in the real situation. There is a need to develop or strengthen accreditation systems, including ensuring periodic updating and professional continuing education programmes linked to re-registration or re-licensing.Action: promote national evidence-based standards for education programmes and institutions, ensuring that they are as important as evidence-based clinical standards and protocols. Incentive schemes may be needed in some situations, to encourage and support recruitment from local communities and/or recruitment from linguistically and culturally diverse communities.4. Empowerment, supervision and supportThe problems associated with getting staff to change their performance based on evidence are widely recognized [65]. Because the majority of women will not encounter a problem during pregnancy, childbirth or after birth, few providers may have hands-on practice of managing complications. Indeed, many midwives working at the community level may never have experienced in their initial training some of the problems and complication that they may meet during their professional career. Providing midwives with supportive supervision which helps build their capacity is essential, more so for those working in isolated practice or small teams in the community. For supervision to build capacity it must go further than assessing records and reviewing case registers. It needs to be supportive, undertaken by clinically competent midwives, allow free and open discussion of clinical practices, and give an opportunity for providers to acknowledge their weaknesses [66]. Supervision should empower midwives, should not focus on just filling in a checklist, and should be performed by provincial or national health offices.Action: Organize supervision as a separate function from the management of the midwifery service, although linked to it and indeed in some areas supervisors may have responsibility for both. Ensure that supervisors are competent in midwifery and receive in-service and updating training in supervising midwifery practice.5. Enabling environment, strengthening systems, community aspectsToo often this enabling environment is missing – often due to failures in health system management. For example, frequently the essential drugs for EmONC are not included in the national drugs list. It is now well known that health care practitioners cannot carry out all their tasks and function effectively if they have concern for their own safety or that of their family, or if they are anxious about their own health or the health of their family [6]. Caring for woman and newborns in an environment lacking essential drugs and equipment to save lives if a complication occurs is particularly stressful and de-motivating. Support from the local community and community leaders, and the active participation of men, are also vital to creating an enabling environment, despite the barriers to male participation [67].Actions: total quality care improvements, quality circles, as well as needs assessments, clinical audits, community surveys, confidential enquiries into maternal deaths, investigations of near-miss cases: all can be used as means of improving quality of care. A continuous supply of essential drugs down to the community level must be assured.6. Tracking progress, monitoring and evaluation for numbers and qualityUntil recently little attention has been paid to the need for permanent monitoring and periodic evaluation of large midwifery programmes. Very few current programmes have built-in evaluation, and there is consequent uncertainty about their health outcomes, and thus their effectiveness. Most safe motherhood programmes rely on fairly standard process indicators such as the UN indicators [68-71] that are most often used for measuring the availability and use obstetric services, but do not take into account quality, which is the product of technical capacity and culturally appropriate response.In addition, lack of a universal benchmark to define a skilled birth attendant has not only caused confusion and lack of validity around this indicator, but has led to great variations and thus an inability to make comparative judgments on programmes [6]. There are currently few reliable and tested tools to measure the midwifery competencies of healthcare providers, or to compare the performance and utilization of non-specialized midwifery providers against specialist provider [72-74].Actions: Establish regular monitoring based on routine data collection with an emphasis on quality. Monitoring and evaluation should involve midwives and midwifery providers at the community level, so that midwives and the community members can use the findings. This is particularly important for evaluating training initiatives, where – for pragmatic reasons – descriptive, non-experimental designs calling for before-and-after studies are the only option for assessing effectiveness.7. Stewardship, resource mobilizationWhile it is acknowledged that most countries need to take incremental steps towards implementing comprehensive health policies to respond to the needs of all citizens, very few have a well designed systematic plan to achieve this [75]. Forty African countries are currently engaged in developing and implementing their national Road Map for maternal and newborn care. Ensuring equitable midwifery care requires intensified actions and substantial investments, calling for increased funds, and better costing and budgeting [76]. In many countries parliamentarians and senior policy makers are not fully aware of the issues around access to midwifery care at the community level and often fail to understand the complexities involved. Furthermore, studies show that decentralization efforts too often focus on financial and structural reforms and do not take sufficient account of the human resource dimension [77,78].Actions: Governments must provide sufficient expenditure and proportionate investment of public resources in the maternal health sector, and focus expenditure on rectifying existing imbalances in the provision of health facilities, health workers and health services. This includes ensuring that the privatization of the health sector does not create a threat to the availability, accessibility, non-discrimination, acceptability and quality of maternal and newborn health services. Policy makers must also recognize that, even where safe motherhood programmes are built on increasing access to institutional birth, women and newborns need access to community-based midwifery care ante and post-natally, as women are more likely to seek skilled care for birth if they have access to such care ante-natally [79].Lessons learned in countriesThe issue of requiring a dedicated skilled provider for maternal and newborn health is gaining momentum in many parts of the world – despite pressures for a generic multipurpose healthcare provider. A survey conducted by WHO in the Africa region showed that among the 31 African countries who responded to the survey (out of 46), only 14 had a HRH policy and plan, an HRH situation analysis and an HRH operational plan [80]. For example, WHO-AFRO is about to publish a set of Midwifery Competencies for Africa, recommended by the Regional Committee in 2005 and developed through a series of consultations with countries. It is hoped that countries will use these competencies as benchmarking for agreeing who meets the definition of a skilled attendant. There are also positive signs to show that the various country Road Maps for maternal and, newborn health are offering important opportunities to integrate human resources issues in the national health plans and national sexual and reproductive health policies. Similarly in other regions there is a renewed interest in developing and supporting the specialist cadre of midwifery provider.Creating/promoting a specialist midwifery cadreThere are more examples of countries investing in increasing the numbers of multi-purpose maternal health providers, but some countries are also taking steps to strengthen and skill up their current midwifery providers, and/or creating a specialist cadre in an attempt to upgrade quality of obstetric care. For example, action has begun to re-establish midwifery in the south of Sudan, an area of huge deprivation following years of civil unrest which has left that part of the country with almost no health system. One of the first priorities undertaken with the assistance of the international donors following the signing of the Peace Accord has been to develop and initiate a programme to train midwives for the community. Elsewhere in Africa, new programmes for direct entry into midwifery training have just started, such as in Zambia.In Bolivia, with UNFPA support, plans have been agreed and work commenced to introduce a pre-service midwifery programme, at provincial university level, so that the midwives from this programme will be educated to a level equivalent of other healthcare providers such as nurses. The reason behind the decision to start such a programme is that, despite excellent results of the national insurance scheme, many women are still reluctant to be attended to by a professional provider until a problem arises, often too late. This is because in the rural areas, where the majority of families still live, people feel that healthcare providers at the facility do not respect the cultural requirements surrounding pregnancy and childbirth. This new programme for professional midwives will have a large component on social and cultural issues, as well as on technical midwifery care. The work is being undertaken with technical support from Chile, which is one of the countries with the longest history of professional midwives in Latin America [81]. Haiti is also in the process of re-opening the national school of nursing and midwifery, after many years of deterioration of their health system due to internal conflict.In many parts of Asia the same positive signs can also be observed. In 2006, Pakistan took the decision to mount a large initiative to train more than 58 000 community midwives. The first intake of students commenced in the summer of 2007. The competencies for this programme and the training of the midwife teachers were done in collaboration with and support from the ICM. The programme for introducing this new cadre has not taken a traditional vertical approach, but has started with strengthening the regulatory and accreditation system, through fortifying the Pakistan Nursing Council, establishing a new Midwifery Association (affiliated with the ICM), and working with the State Examinations Boards. The MOH supported by partners has also strengthened the training infrastructure, including upgrading and refurbishing training schools, as well as updating the staff working in the facilities where students will also undertake part of their training and where it is hoped they will refer clients after their graduation when needed. Afghanistan has recently re-opened their schools of midwives, after having started with launching a competency-based pre-service training curriculum. This successful programme allowed 1300 young midwives to graduate and make a dramatic impact on women's access to maternity care.'Skilling up' and increasing retention of the current maternity workersWith the exceptions cited above, very few countries have embarked on a scheme for introducing a new cadre of professional midwife. Most countries in all regions have mainly focused on scaling up and skilling-up those who are already functioning as midwives, or supporting and retaining the midwifery providers working in isolated places. Mauritania for example is expanding an obstetric risk insurance mechanism aimed at sharing costs related to obstetric complications among all pregnant women on a voluntary basis. The budget includes a number of incentives (30%) and duty allowances (13%) to compensate facilities and staff for increased workload and is aimed at suppression of informal payments by clients. A mechanism of special incentives to ensure better retention of health professionals in remote areas has also been established, while noting that this initiative is not without its challenges, given the increasing competition from an uncontrolled and rapidly developing private sector [82].Senegal on the other hand is focusing on strengthening management systems and capacity, especially at the district level. Professional staff are now receiving incentives and midwives, who are seen as the most cost-effective health professionals, are involved in maternal death reviews and focus groups to assist them in improving quality of care [83].Rwanda is also undertaking a management approach to ensuring that skilled midwifery providers are available and accessible, free of charge and offering quality care. A recent survey has shown improvements in health centre performance and higher productivity of health staff through output-based performance contracts [84].Mozambique, Malawi, Senegal, Tanzania and a few other African countries have for some years successfully trained mid-level cadres (health officers and midwives), as well as general practitioners to provide comprehensive emergency obstetric care including surgery. Their initial skills were deficient in terms of maternal and newborn health, and therefore as generalists they were unable to meet the needs of mothers and newborns. These trained health professionals are highly cost-effective as their training (and other related costs) is less costly in regard to the comparable performance of obstetric specialists. Furthermore, there is evidence showing their high level of retention [85,86].Expanding numbers of professional midwives to take services to the communityIn Malawi, as in many African countries, professional midwives mainly conduct institutional births, yet the majority of births still take place at home. Also, like many neighbouring countries, Malawi suffers from a huge deficit of all human resources for health, including physicians, with a ratio of 1.6/100 000 (health workers/population). Addressing HRH challenges is very difficult, but action is being taken to expand training institutions to accommodate more students; increase enrolment of nurse/midwives and other healthcare providers; and to skill up competencies to gain community midwifery clinical experience. Moreover, a community-oriented curriculum has been developed to train District Health Officers, as a specific response to the huge numbers lost through migration. The programme includes a minimum of community health (25%), plus surgical and medical specialties, including midwifery skills. A post-graduate programme has now also been added [87].Zimbabwe where maternal mortality increased between 1994 and 1999 from 283/100 000 to 695/100 000, is facing major challenges in relation to midwifery services, including high attrition rate (brain drain), inadequate midwife tutors, midwifery not seen as a lucrative post graduation career, and no recognition for the profession of midwifery. The curriculum has been revised, student midwives now have practical attachments (hands on experience), a new diploma in midwifery has been started, in-service training and on-the-job support (mentorship by a skilled midwife) are now standardized. Efforts to increase the capacities of training of teachers have resulted in development of a Masters with a major in Maternal and Child Health. WHO has also recently announced support for working in collaboration with the Royal College of Midwives (UK) to encourage some of the diaspora community who are in the UK working as midwife teachers to return for short stays to offer their services in Malawi.In some countries the low rate of skilled attendance is not because there are insufficient providers, but because of insufficient posts in the public sector to employ all available healthcare providers, even if they are known to have the necessary skills. In Kenya, an initiative began in 2004 to explore if it were possible to empower retired midwives and to support them to return to work as semi-private practitioners, still linked to and supervised by the local facility, under authority of the District Management Team. This pilot project has proved to be highly effective in increasing the numbers of skilled midwifery providers working in the country – particularly at the community level, where almost half the births still take place. As in some other countries in the region, the age of retirement from public service is low in Kenya, currently 55 years of age. Many professional women, who have delayed their own pregnancies and childrearing until after they have completed their studies and have had little time to work, find this age too early. Many are still supporting children through higher education and out of necessity are required to keep earning an income. Although a formal evaluation of the initiative is not yet available, all stakeholders are enthusiastic with the preliminary results, and the MoH has now asked all donors to support this initiative. A decision has been made in the new national RH strategy to roll-out the programme across the country. One of the keys of the success of this initiative according to UNFPA and MoH has been the involvement of the community in selecting which retired midwives to support. Those that have been selected are valued in their respective communities and are being well used by the local families. Preliminary results show that referrals for complications have increased significantly particularly referrals from midwives who have been able to identify problems or potential problems early.Conclusion: Scaling up and skilling upWe hope to have conveyed the message that for the sake of mothers and newborns both 'scaling up' coverage and 'skilling up' quality of care are necessary. In the event of scarce resources, however, we support the option of giving priority to quality of care over coverage, offering an adequate number of skilled professionals strongly supported by a well performing system, rather than the option of a high number of multi-purpose workers based in villages without adequate capacity, authority and support. We do not believe, and the experience of Bangladesh and Indonesia seem to confirm, that a high number of community-based, multi-purpose workers can be properly supported and funded to achieve the desired objective. Also, our message is that even though specialist skilled professionals are preferable, they cannot, and should not, work alone. We introduce a fundamental contrast between 'community-midwives', who we consider unable to fulfill the core life-saving functions and 'midwives in the community', who are midwives first, with all the skills attached to the definition.Overloading skilled professionals, particularly with tasks that can be done by others, is not cost-effective and can lead to burn out and poor quality. While 'multi-purpose community workers' can deliver other complementary services such as family planning and other primary health care services, it is not cost effective to produce multi-purpose workers with 'some' midwifery skills. Properly trained specialist skilled attendants such as professional midwives may take 3 to 4 years to train, they can have additional skills and deliver a broad range of primary healthcare, provided doing so does not interfere with the provision and maintenance of the competencies required to be a skilled birth attendant.Developing the needed workforce to ensure that women and newborns have access to a competent midwifery provider requires a comprehensive plan, tailored to the specific situation in each country. We believe that the framework developed by the participants at the 1st International Forum on Scaling up Midwifery for the Community can help countries to develop such a plan, while keeping a focus on quality [Additional file 1].While countries should keep in mind from the beginning the 'long-term strategy' consisting of most births taking place in health centres (even small facilities operated by teams of midwives) attended by skilled professionals operating in multidisciplinary teams, and backed up by accessible functioning referral hospitals, their health planners also need to be pragmatic and to consider possible 'interim strategies'. An example of one such strategy is professional midwives leading multi-purpose teams and supervising home births attended by other health workers. However, there must be time limits set for these interim strategies otherwise they might become permanent strategies, as was the case in too many settings over the past 20 years.Our final message is that monitoring and evaluation must be built into all plans from the very beginning, including for interim strategies, in an effort to produce evidence on how best to develop a competent midwifery workforce in low-resource settings. There must be a greater focus on continuous monitoring and periodic evaluations. Furthermore, monitoring and evaluation must focus on qualitative as well as quantitative data and look at the performance of providers – measuring how they are performing and identifying the system barriers that prevent quality performance.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsVF conceived the paper, drafted the outline, the research question and the conclusion, reviewed and edited the whole manuscript. DRS drafted the analysis of the current situation and the challenges, as well as the key areas of work to scale up midwifery. LdB drafted the chapter on lessons learned from countries. All three authors contributed to the reference search, read and approved the final manuscript.Supplementary MaterialAdditional file 1The Hammammet Call to Action, 15 December 2006. The full text of the joint inter-agency declaration and call to action for governments, regulatory bodies, professional health care organizations, educators and communities. Resulting from the first International Forum on Midwifery in the Community in Hammammet, TunisiaClick here for file\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2570921.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2570921",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2570921\nAUTHORS: Maria João Vieira, Eliane Silva, Costantina Desario, Nicola Decaro, Júlio Carvalheira, Canio Buonavoglia, Gertrude Thompson\n\nABSTRACT:\nNo Abstract\n\nBODY:\nTo the Editor: Canine parvovirus (CPV) emerged in the late 1970s, presumably by mutations in feline panleukopenia virus, and became a major viral pathogen of dogs worldwide (1). Between 1979 and 1981, the original type 2 virus (CPV-2) was replaced by a new genetic and antigenic variant, type 2a (CPV-2a). Between 1983 and 1984, CPV-2a was replaced by type 2b (CPV-2b), which differs from type 2a by only 1 epitope located at residue 426 of the VP2 capsid protein (2). CPV-2 does not replicate in cats, but the new variants replicate in dogs and cats (3). Recently, an antigenic change has been observed in a new strain, CPV-2c, isolated from domestic dogs in Italy (4). This variant was also detected in Vietnam (5), other countries in Europe (6), and the United States (7). CPV-2c was recently detected in cats (8) and is characterized by a replacement of aspartic acid with glutamic acid at residue 426 of the VP2 capsid protein.To identify CPV types 2, 2a, and 2b, PCR methods were developed (9). However, these methods could not distinguish type 2c from type 2b (4). Consequently, we used a PCR–restriction fragment-length polymorphism (RFLP) assay with endonuclease MboII. This enzyme can distinguish type 2c from other CPVs (4). Recently, a real-time PCR assay based on minor groove binder (MGB) probe technology was developed for rapid identification and characterization of the antigenic variants. This assay is based on 1 nucleotide polymorphism in the VP2 gene (10).In June 2006, a 10-week-old female dog (PT-32/07) was brought to the veterinary clinic in Figueira da Foz, Portugal, with clinical signs of parvovirus infection, after an episode of gastrointestinal disease in her littermates. Three littermates, also brought to the clinic, showed no signs of infection. None of the dogs were vaccinated against CPV. Clinical signs in dog PT-32/07 were lethargy, anorexia, vomiting, diarrhea, and a temperature of 39.3°C. Identical signs were observed in 1 littermate 3 days later; the 2 other dogs did not show any signs other than lethargy and loose stools.Rectal swab samples from all dogs were screened for CPV by using an immunomigration rapid test (Synbiotics Corporation, Lyon, France). Two of the dogs showed negative results, and 2 showed positive results. Feces, serum, and lingual swab samples were positive for parvovirus DNA. DNA was quantified by using a real-time PCR with TaqMan technology performed in an i-Cycler iQ (Bio-Rad Laboratories, Milan, Italy).CPV variants were characterized by using MGB probe technology. This technology uses type-specific probes labeled with different fluorophores (FAM and VIC) that can detect single nucleotide polymorphisms between CPV types 2a/2b and 2b/2c (10). MGB probes specific for type 2b were labeled with FAM in both type 2a/2b and 2b/2c assays, and MGB probes specific for type 2a (type 2a/2b assay) and type 2c (type 2b/2c assay) were labeled with VIC.All specimens from 1 dog (PT-32/07) were positive for the 2 variants of CPV type 2 (CPV 2b and CPV 2c). Conversely, of the 3 littermates, 2 were positive for CPV type 2b and 1 was positive for CPV type 2c in all samples (Table).TableDetection by minor groove binder probe assay of CPV antigenic variants in different specimens from dogs from the same litter (10 weeks old), Portugal, 2006*DogVaccinesRapid test result for CPVDays in clinicClinical courseSamplesTaqMan probeCPVFAM a/bFAM b/cVIC a/bVIC b/cPT-15/07None–7RecoveredFeces++––2bLingual swab++––2bSerum++––2bPT-16/07None+7RecoveredFeces++––2bLingual swab++––2bSerum++––2bPT-17/07None+7RecoveredFeces–––+2cLingual swab–––+2cSerum–––+2cPT-32/07None–7RecoveredFeces++–+2b/2cLingual swab++–+2b/2cSerum++–+2b/2c*CPV, canine parvovirus.A conventional PCR and RFLP analyses were performed by using the method of Buonavoglia et al. (4) with known positive CPV-2b and CPV-2c samples as controls to confirm our findings. The 583-bp PCR product obtained from the coinfected dog by using primer pair 555for/555rev was digested with MboII. Digestion generated 2 fragments (≈500 and 80 bp) in all dog samples. The CPV-2c control sample showed 2 fragments (≈500 and 80 bp), and CPV-2b control sample was not digested with MboII.We report CPV-2b and CPV-2c variants in samples from a dog with littermates that were positive for CPV-2b or CPV-2c during an episode of gastrointestinal disease. Coinfection with multiple CPV variants that showed high genetic diversity in the VP2 gene has recently been reported in a domestic cat (8). Continuous and rapid evolution of CPV may cause serious problems in diagnostic testing and vaccine efficacy. Antigenic variation may negatively affect vaccine efficacy if changes occur at major antigenic sites. Thus, continuous monitoring for novel genetic and antigenic virus types is needed. Additional studies are in progress to characterize nucleotide sequences of all CPV isolates from this case.\n\nREFERENCES:\n1. Truyen\nU. Evolution of canine parvovirus: a need for new vaccines?\nVet Microbiol. 2006;117:9–13. 10.1016/j.vetmic.2006.04.00316765539\n2. Parrish\nCR, Aquandro\nCF, Strassheim\nML, Evermann\nJF, Sgro\nJY, Mohammed\nHO. Rapid antigenic-type replacement and DNA and sequence evolution of canine parvovirus.\nJ Virol. 1991;65:6544–52.1942246\n3. Truyen\nU, Evermann\nJF, Vieler\nE, Parrish\nCR. Evolution of canine parvovirus involved loss and gain of feline host range.\nVirology. 1996;215:186–9. 10.1006/viro.1996.00218560765\n4. Buonavoglia\nC, Martella\nV, Pratelli\nA, Tempesta\nM, Cavalli\nA, Buonavoglia\nD, \nEvidence for evolution of canine parvovirus type 2 in Italy.\nJ Gen Virol. 2001;82:3021–5.11714979\n5. Nakamura\nM, Tohya\nY, Miyazawa\nT, Mochizuki\nM, Phung\nHT, Nguyen\nNM, \nA novel antigenic variant of canine parvovirus from a Vietnamese dog.\nArch Virol. 2004;149:2261–9. 10.1007/s00705-004-0367-y15503211\n6. Decaro\nN, Desario\nC, Addie\nD, Martella\nV, Vieira\nMJ, Elia\nG, \nMolecular epidemiology of canine parvovirus, Europe.\nEmerg Infect Dis. 2007;13:1222–4.17953097\n7. Hong\nC, Decaro\nN, Desario\nC, Tanner\nP, Pardo\nMC, Sanchez\nS, \nOccurrence of canine parvovirus type 2c in the United States.\nJ Vet Diagn Invest. 2007;19:535–9.17823398\n8. Battilani\nM, Scagliarini\nA, Ciulli\nS, Morganti\nL, Prosperi\nS. High genetic diversity of the VP2 gene of a canine parvovirus strain detected in a domestic cat.\nVirology. 2006;352:22–6. 10.1016/j.virol.2006.06.00216822535\n9. Senda\nM, Parrish\nCR, Harasawa\nR, Gamoh\nK, Muramatsu\nM, Hirayama\nN, \nDetection by PCR of wild-type canine parvovirus which contaminates dog vaccines.\nJ Clin Microbiol. 1995;33:110–3.7699026\n10. Decaro\nN, Elia\nG, Martella\nV, Campolo\nM, Desario\nC, Camero\nM, \nCharacterisation of the canine parvovirus type 2 variants using minor groove binder probe technology.\nJ Virol Methods. 2006;133:92–9. 10.1016/j.jviromet.2005.10.02616313976"
+}

batch_13/PMC2572733.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2572733",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2572733\nAUTHORS: Mathias Pessiglione, Predrag Petrovic, Jean Daunizeau, Stefano Palminteri, Raymond J. Dolan, Chris D. Frith\n\nABSTRACT:\nSummaryHow the brain uses success and failure to optimize future decisions is a long-standing question in neuroscience. One computational solution involves updating the values of context-action associations in proportion to a reward prediction error. Previous evidence suggests that such computations are expressed in the striatum and, as they are cognitively impenetrable, represent an unconscious learning mechanism. Here, we formally test this by studying instrumental conditioning in a situation where we masked contextual cues, such that they were not consciously perceived. Behavioral data showed that subjects nonetheless developed a significant propensity to choose cues associated with monetary rewards relative to punishments. Functional neuroimaging revealed that during conditioning cue values and prediction errors, generated from a computational model, both correlated with activity in ventral striatum. We conclude that, even without conscious processing of contextual cues, our brain can learn their reward value and use them to provide a bias on decision making.\n\nBODY:\nIntroductionHumans frequently invoke an argument that their intuition can result in a better decision than conscious reasoning. Such assertions may rely on subconscious associative learning between subliminal signals present in a given situation and choice outcomes. For instance, clinicians may improve their therapeutic decisions through learned associations between treatment outcomes and subliminal signs presented by their patients. Likewise, poker players can improve their gambles through a learned association between monetary outcomes and subliminal behavioral manifestations of their opponents (the so-called “gamblers' tell”).The idea that such instrumental conditioning can occur subconsciously has been around for almost a century (Thorndike, 1911). This assumption originally rested on observations that rewards and punishments shape behavioral responses in species allegedly lacking conscious awareness. However, subliminal conditioning studies in humans have so far been restricted to Pavlovian paradigms such as fear conditioning (Clark and Squire, 1998; Knight et al., 2003; Morris et al., 1998; Olsson and Phelps, 2004), where subliminal stimuli (like unseen faces) are paired with unpleasant events (like white noise) to increase automatic responses (like skin conductance). To our knowledge, subliminal instrumental conditioning, where decision making would be biased by unperceived cues predicting rewards or punishments, has never been previously demonstrated.Our subliminal conditioning task was adapted from a published supraliminal task, wherein subjects selected between visual cues so as to learn choices that maximized monetary outcomes (Pessiglione et al., 2006). In our previous study, we modeled subjects' behavior by optimizing the free parameters of a standard machine learning algorithm (termed Q learning), to get maximal likelihoods for the observed decisions. When we regressed key output variables of the optimized model against simultaneously acquired functional neuroimaging data, we showed that prediction errors were expressed in the striatum. Postexperimental debriefing indicated that some subjects managed to understand the statistical structure of the task, while others appeared to rely on what they referred to as their intuition. These latter reports suggest that subjects can improve their decisions without consciously following the incremental steps of the Q-learning procedure.The motivating assumption of the current experiment was that processes associated with striatal learning are not consciously accessible but, nonetheless, influence choice decision making. Indeed, if contextual cues reach awareness, other brain systems are likely to play a role, as expressed in conscious reasoning or strategic control, which allows one to develop explicit knowledge of statistical contingencies. However, if the cues remain unseen, learning would solely depend on a subconscious processing that involves the striatum, with an algorithmic structure similar to a Q learning, which does not embody explicit information about statistical contingencies. Under these assumptions, we predicted that, if in our task visual cues were masked, both striatal activity and behavioral choices would still reflect Q-learning outputs.ResultsA prerequisite for the present study was to demonstrate efficient masking of the visual cues. These cues were novel abstract symbols, which were scrambled and mixed to create mask images. To assess visual awareness, we successively displayed two masked cues on a computer screen and asked subjects whether they perceived a difference or not. We reasoned that if subjects are unable to correctly perceive any difference between the masked cues, then they are also unable to build conscious representations of cue-outcome associations. The procedure has the advantage of not showing the cues unmasked, so that, by the end of the experiment, subjects had no idea what the cues look like.The perceptual discrimination task was performed outside the scanner at the beginning of the experiment, in order to adapt duration of cue display to each individual, and in the scanner at the end of the experiment, to check for any effect of learning or change in visual conditions. For all subjects, cue duration was set at either 33 or 50 ms and was kept fixed through the entire experiment. In every individual, correct guessing on the final assessment did not differ from chance (p > 0.05, chi-square test). At group level, average percentage of correct responses for the 20 subjects was 48% ± 3%, which again was not different from chance (p > 0.5, two-tailed paired t test). Average d′ was 0.08 ± 0.20, showing that, even when correcting for response bias, signal detection was not different from zero (p > 0.5, two-tailed paired t test). Thus, subjects remained unable to discriminate between the different masked cues, from the beginning to the end of conditioning sessions.We employed the same masking procedure in the subliminal conditioning task, in which cues were paired with monetary outcomes (Figure 1). From these outcomes (−£1, £0, +£1), subjects had to learn when to take the risky response (either “Go” or “NoGo,” depending on subjects). Subjects were also told that, for the risky response, the outcome would depend on the cue hidden behind the masking image (see instructions in Supplemental Data available online). As they would not see the cues, we encouraged them to follow their intuition, taking the risky response if they had a feeling they were in a winning trial and choosing a safe response if they felt it was a losing trial. Note that if subjects always made the same response, or if they performed at chance, their final payoff would be zero.As a dependent variable to assess for conditioning effects, we used monetary payoff, which corresponds to the area below the reward and above the punishment learning curves (Figure 2A). Overall subjects won money in this task, on average £7.5 ± £1.8 (p < 0.001, one-tailed paired t test), indicating that the risky response was more frequently chosen following reward predictive relative to punishment predictive cues. Both reward and punishment conditions also differed significantly from the neutral condition (p < 0.05, one-tailed paired t test). There was no significant difference (p > 0.5, two-tailed paired t test) between the reward and punishment condition: on average subjects won £24.3 ± £1.9 and avoid losing £23.2 ± £2.1. Learning curves showed that responses improved symmetrically for rewards and punishments, ending with 63% ± 5% of correct responses on average. Surprisingly, this plateau was reached at around the halfway point of the learning session. The effects of subliminal conditioning were subsequently assessed with a preference judgment task, in which cues were uncovered and rated by the subjects from the most to least liked (Figure 2B). Ratings were significantly higher for reward compared to punishment cues (p < 0.01, one-tailed paired t test), consistent with subjects having learned the affective values of subliminal cues, such that these values were able to bias their preferences. When uncovering the cues, subjects were also asked to signal any cue that they may have seen during conditioning sessions; none was reported as previously seen.To model instrumental conditioning, we implemented a standard Q-learning algorithm (Pessiglione et al., 2006), with inputs from individual histories of cues, choices, and outcomes. On every trial, the model estimates the likelihood of the risky response from the value of the displayed cue. If the risky response was actually taken, the model then updates the value of the displayed cue in proportion to the prediction error. The parameters of the model were optimized such that likelihoods of risky responses provided the best fit of subjects' actual responses across conditioning sessions (Figure 2A). The Q values and prediction errors generated by this optimized algorithm were then used as regressors for analysis of brain imaging data (see Figure S1).We recorded brain activity while subjects performed the subliminal conditioning task, using functional magnetic resonance imaging (fMRI). We first examined brain regions reflecting Q value at the time of cue onset, increasing their response to reward-predicting cues and decreasing their response to punishment-predicting cues, across learning sessions. After correction for multiple comparisons (family-wise error, p < 0.05), we noted significant correlated activity in ventral striatum bilaterally (Figures 3A and 3B, left). The same region was also significantly activated at the time of outcome in keeping with prediction errors being expressed at this time point (Figure 3C, left). In a second analysis, we computed regression coefficients for the different conditions at the time of cue and outcome onsets, separately for the first and second half of conditioning sessions. Contrasts with the neutral condition were then averaged over all ventral striatal voxels showing significant activation at the most conservative threshold in the first analysis. This confirmed that from the first to the second half of conditioning sessions, ventral striatal responses increased for reward cues and decreased for punishment cues (Figure 4A, left). At the time of outcome onset, the same ventral striatal region reflected positive prediction errors in the reward condition and negative prediction errors in the punishment condition. In keeping with the Q-learning model, both positive and negative prediction errors decreased from the first to the second half of conditioning sessions. Thus, across subliminal conditioning, the ventral striatal response was consistent with the expression of Q values (for unseen cues) and prediction errors (based on visible outcomes).We further examined variability in individual performance to explain why some subjects won more money than others. More precisely, we searched for brain regions where coefficients of Q-value regressors correlated with individual payoffs. These regions were confined to extrastriate visual cortex (Figure 3A, right) at the most conservative threshold (familywise error, p < 0.05), spreading into the ventral occipitotemporal stream (Figure 3B, right) with a more liberal threshold (uncorrected, p < 0.001). Contrast estimates confirmed that extrastriate voxels progressively differentiated the reward and punishment cues from the first to the second half of conditioning sessions (Figure 4A, right). At the time of outcome onset, these extrastriate regions responded positively for both rewards and punishments, showing no evidence for encoding of prediction errors. Thus, during the subliminal conditioning task, the extrastriate visual cortex learned to discriminate between unseen cues according to their reward value but did not express outcome-related prediction errors (Figure 3C, right).To further assess whether the ventral striatum and visual cortex were able to discriminate between the subliminal cues, we extracted time courses of BOLD response. These time courses were averaged over trials, sessions, and subjects, separately for the reward and punishment conditions (Figure 4B). We found that BOLD responses to reward and punishment cues significantly differed after two acquisition volumes (3.9 s) in the ventral striatum (one-tailed paired t test, p < 0.01) and after three (5.85 s) in the visual cortex (one-tailed paired t test, p < 0.01).Finally, we ascertained whether neuroimaging and behavioral effects of subliminal conditioning were driven by subjects scoring at the high end in perceptual discrimination performance. We tested for correlations between correct guessing assessed in the final perceptual discrimination test and coefficients of Q-value regressors in both the ventral and extrastriate cortex. None was significant; Pearson's correlation coefficients were respectively −0.25 and −0.18. We also tested correlation of correct guessing with monetary payoffs from conditioning sessions and differential ratings in the preference judgment task. Again, none was significant; Pearson's correlation coefficients were respectively 0.26 and 0.29.DiscussionWe provide evidence that instrumental learning can occur without conscious processing of contextual cues. This finding might relate to previous evidence for implicit or procedural learning, where behavioral responses can be adapted to the statistical structure of stimuli that fails to be reported explicitly (Bayley et al., 2005; Berns et al., 1997; Destrebecqz and Cleeremans, 2001; Seitz and Watanabe, 2003). Interestingly, implicit/procedural learning has been suggested to involve the basal ganglia, in contrast with explicit/declarative memory which would involve the medial temporal lobe (Cohen et al., 1985; Milner et al., 1998; Poldrack and Packard, 2003; Squire, 1992). In implicit learning tasks, such as serial reaction time or probabilistic classification, authors have claimed that subjects can achieve good acquisition without explicit knowledge of the task structure. However, methods for assessing awareness of statistical contingencies have been criticized, principally on the issue that questions were too demanding in terms of memory (Lagnado et al., 2006; Lovibond and Shanks, 2002; Wilkinson and Shanks, 2004). Thus, to formally test whether instrumental conditioning can occur without awareness, we took a more stringent approach: masking the cues, so that they remained unperceived.We believe our methodology avoids most previous problems related to assessing awareness, by demonstrating that subjects were not able to discriminate between masked cues (without the help of rewards and punishments), rather than retrospectively assessing awareness of contingencies. Moreover, postconditioning recognition tests would not be sufficient in our case, since subjects would not need to identify cues for associative learning to be conscious. Indeed, they could learn associations between risky response outcomes and tiny fragments of the visual dynamic pattern formed by the mask/cue/mask flashing. However, postconditioning debriefing questions might be informative in explaining why subjects could not discriminate between masked cues. Thus, when we explicitly asked subjects to state what the cues looked like, they reported in majority of cases that they had no idea. When the subjects were presented with the cues, now unmasked, they reported surprise at seeing the symbols while asserting that they had never seen them before. This suggests that during conditioning, subjects had no a priori representational knowledge to guide a visual search for cues hidden behind the masks. We believe that absence of an a priori representation is a crucial feature of our design, which, in addition to visual masking, prevented subjects from consciously seeing the cues.Using this methodology, we show that pairing rewards and punishments can guide behavioral responses and even condition preferences for abstract cues that subjects could not consciously see. Note that if cues were visible, learning curves would have been optimized in one trial or two; hence we are not claiming that conscious awareness is unhelpful in supraliminal instrumental conditioning. However, in our subliminal conditioning task, conscious strategies (such as win-stay/lose-switch) might have been detrimental, which would explain why learning curves were limited well below the optimum.We also identified brain circuitry associated with subliminal instrumental conditioning. The ventral striatum responded to subliminal cues and to visible outcomes in a manner that closely approximates Q-learning algorithm, expressing reward expected values and prediction errors, just as was reported in supraliminal instrumental conditioning studies (O'Doherty et al., 2004; Pagnoni et al., 2002; Pessiglione et al., 2006; Yacubian et al., 2006). Interestingly, there is no need for representing the statistical structure of the task in Q learning, which is an incremental procedure updating the expected values of chosen actions according to the subject's history of reward and punishment outcomes. This accords well with views that the striatum is a major player in implicit/procedural learning (Graybiel, 2005; Hikosaka et al., 1999; Packard and Knowlton, 2002) and with evidence that ventral striatum encodes reward-related information (Delgado, 2007; Knutson and Cooper, 2005; Pecina et al., 2006).For the sake of simplicity, we have described ventral striatum activity as directly reflecting key outputs of Q-learning algorithm: Q value at the time of cue onset and prediction error at the time of outcome. There are nonetheless other variables in machine learning literature that would also correlate with ventral striatum activity and which could provide an alternative interpretational framework for our study. In particular, it is important to note that average Q values (over the reward, neutral, and punishment conditions) remain around zero during our conditioning paradigm. Hence, Q value is approximately equal to Q value minus average Q value, which can be seen as equivalent to a cue prediction error (actual Q value minus predicted Q value). Our data are therefore equally compatible with the notion that the ventral striatum encodes prediction errors at the time of both cue and outcome onsets. However, because prediction errors represent a function of Q values, the brain has to learn about Q values in order to signal prediction errors. Thus, whether we consider the ventral striatum as encoding a Q value or a prediction error does not alter our central conclusion: namely, the human brain can learn the reward value of subliminal cues, so as to later influence behavioral choices.It is of interest that extrastriate visual cortex also reflected the reward value of subliminal cues, but not outcome-related prediction errors. Modulation of visual cortex activity by monetary incentives has already been reported in neuroimaging studies of supraliminal processes, such as visuomotor transformation, attentional control, and working memory (Krawczyk et al., 2007; Ramnani and Miall, 2003; Small et al., 2005). In our case, the modulation suggests that conditioning involves an interaction between the extrastriate cortex (which would discriminate cues according to their visual properties) and the ventral striatum (which would tag cues with affective values depending on reward prediction errors). However, we acknowledge that we do not as yet have a complete account of how the brain produces behavioral effects of subliminal conditioning. Notably, we failed to identify the brain regions mapping affective values onto motor commands, which would complete the circuit from visual cues to behavioral responses. Further experiments will be necessary to fill in these explanatory gaps.More generally, our approach, combining perceptual masking and computational modeling, can be extended over the field of functional neuroimaging. Computational reinforcement learning theory has proven useful to model both brain activity and behavioral choices in human and nonhuman primates (Daw and Doya, 2006; McClure et al., 2004; O'Doherty et al., 2007). Brain activity reflecting sophisticated computations are unlikely to be accessed by the conscious mind, which takes minutes to solve even simple equations. This brain activity would therefore represent unconscious processes, which we formally demonstrated here in the case of instrumental conditioning. Combining masking and modeling can, in principle, make more tractable the identification of basic neuronal mechanisms shared within other species, eliminating the use of reportable knowledge that might be unique to humans. It might also help assess the integrity of these same basic mechanisms in patients with neurological or psychiatric conditions, avoiding confounds generated by conscious strategic compensations.Experimental ProceduresSubjectsThe study was approved by the National Hospital for Neurology and Neurosurgery and the Institute of Neurology joint Ethics Committee. Subjects were recruited via Gumtree website and screened for exclusion criteria: left handedness, age below 18 or above 39, regular taking of drug or medication, history of psychiatric or neurological illnesses and contra-indications to MRI scanning (pregnancy, claustrophobia, metallic implants). All subjects gave informed consent prior to taking part. We scanned 20 subjects: 11 males (mean age 26.8 ± 6.3 years) and 9 females (mean age 23.8 ± 3.3 years). Two more subjects were scanned but discarded from the analysis, because they eventually could describe parts of the visual cues, were above chance level in the perception task, and won unusually large amounts of money in the conditioning task. Subjects were told that they would play for real money, but at the end of the experiment their winnings were rounded up to a fixed amount.Behavioral Task and AnalysisSubjects first read the instructions (see Supplemental Data) about the different tasks, which were later explained again step by step. Before scanning, subjects were trained to perform the conditioning task and the perception task on practice versions. In the scanner, they had to perform three sessions of the conditioning task, each containing 120 trials and lasting 13 min, and one session of the perception task, containing 120 trials and lasting about 7 min. The abstract cues were letters taken from the Agathodaimon font. The 12 cues shown in the scanner were randomly assigned to the four task sessions, each session hence employing 3 new cues. The same two masking patterns (see Figure 1), one displayed before and the other after the cue, were used in all task sessions. The sequence of display and the cue-outcome associations were also randomized for every subject.The perceptual discrimination task was used to select the appropriate duration for cue display, which was then kept to either 33 or 50 ms for the entire experiment. In this task, subjects were flashed two masked cues, 3 s apart, displayed on the center of a computer screen, each following a fixation cross. They had to report whether or not they perceived any difference between the two visual stimulations. The response was given manually, by pressing one of two buttons assigned to “same” and “different” choices. The perceptual discrimination task was then employed as a control for awareness at the end of conditioning sessions. We checked with a chi-square test that in all included subjects performance was not significantly different from chance level (50% of correct responses). We also calculated d′ measure, which is the difference between normalized rates of hits (correct “different” response) and false alarms (incorrect “different” responses). We ensured that this measure was not significantly different from zero, at group level, using one-tailed paired t test.The instrumental conditioning task involved choosing between pressing or not pressing a button, in response to masked cues. After showing the fixation cross and the masked cue, the response interval was indicated on the computer screen by a question mark. The interval was fixed to 3 s and the response was taken at the end: “Go” if the button was being pressed, “No” if the button was released. The response was written on the screen as soon as the delay had elapsed. Subjects were told that one response was safe (you do not win or lose anything) while the other was risky (you can win £1, lose £1, or get nothing). The risky response was assigned to Go for half of the subjects, and to NoGo for the other half, such that motor aspects were counterbalanced between reward and punishment conditions. Subjects were also told that the outcome of the risky response would depend on the cue that was displayed between the mask images. In fact, three cues were used, one was rewarding (+£1), one was punishing (−£1), and the last was neutral (£0). Because subjects were not informed about the associations, they could only learn them by observing the outcome, which was displayed at the end of the trial. This was either a circled coin image (meaning +£1), a barred coin image (meaning −£1), or a gray square (meaning £0).Subjects were then debriefed about their visual perceptions and their response strategies. They reported responding either at chance, following their intuition, or following logical rules. None of them had the slightest idea of what the cues looked like. For the preference judgment task, the cues were then shown unmasked on a computer screen. The three cues used for a given session were displayed side by side, the position being randomized. Subjects were asked to rate them in order of preferences: 3 for the most liked, 2 for the intermediate, and 1 for the least one.To assess for instrumental conditioning, we used one-tailed paired t tests comparing individual earnings with chance level (which is £0). Similarly, to assess for preference conditioning, we used one-tailed paired t tests comparing differential rating of winning and losing cues with chance level (which is 0).Computational ModelWe used a standard Q-learning algorithm (Sutton and Barto, 1998), which has been shown previously to offer a good account of instrumental choice in both humans and monkeys (Daw and Doya, 2006; McClure et al., 2004; O'Doherty et al., 2007). For each cue, the model estimates the expected value of the risky response, on the basis of individual sequences of choices and outcomes. This value, termed a Q value, is essentially the amount of reward expected from choosing the risky response given the contextual cue. These Q values were set at 0.1 before learning, and after every risky response the value of the cue was updated according to the Rescorla-Wagner rule: Q(t + 1) = Q(t) + α∗δ(t). Following this rule, values are increased if the outcome is better than expected, and decreased in the opposite case. The prediction error was δ(t) = R(t) − Q(t), with R(t) defined as the reinforcement obtained from choosing the risky response at trial t. In other words, the prediction error δ(t) is the difference between the expected outcome, i.e., Q(t), and the actual outcome, i.e., R(t). The reinforcement magnitude R was +1 and −1 for winning and losing £1, and 0 for neutral outcomes. Given the Q value, the associated probability of choosing the risky response was estimated by implementing the softmax rule: P(t) = 1/(1 + exp(−Q(t)/β)). This rule ensures that likelihood will be superior to 0.5 for positive values and inferior to 0.5 for negative values. The learning rate α concerns the amplitude of value changes from one trial to the next. The temperature β concerns the randomness of decision making. These two free parameters, α and β, were adjusted to maximize the probability (or likelihood) of the actual choices under the model. With the constraint that the parameters should be identical for reward and punishment cues we found: α = 0.1 and β = 0.9. The model was then used to create statistical regressors corresponding to the Q values and prediction errors, for analysis of brain images.Images Acquisition and AnalysisT2∗-weighted echo planar images (EPI) were acquired with blood oxygen-level dependent (BOLD) contrast on a 3.0 Tesla magnetic resonance scanner. We employed a tilted plane acquisition sequence designed to optimize functional sensitivity in the orbitofrontal cortex and medial temporal lobes (Deichmann et al., 2003). To cover the whole brain with a short TR (1.95 s), we used the following parameters: 30 slices, 2 mm slice thickness, 2 mm interslice gap. T1-weighted structural images were also acquired, coregistered with the mean EPI, normalized to a standard T1 template, and averaged across subjects to allow group level anatomical localization. EPI images were analyzed in an event-related manner, within a general linear model, using the statistical parametric mapping software SPM5 (Wellcome Trust center for NeuroImaging, London, UK). The first 5 volumes of each session were discarded to allow for T1 equilibration effects. Preprocessing consisted of spatial realignment, normalization using the same transformation as structural images, and spatial smoothing using a Gaussian kernel with a full-width at half-maximum of 6 mm.We used two different statistical linear regression models for our analyses. In both every trial was modeled as having two time points, corresponding to cue and outcome onsets. In the first model, two separate regressors were created for cues and outcomes, respectively modulated by the Q values and prediction errors computed by our optimized algorithm. In the second model, 12 separate regressors were created corresponding to the two time points (cues and outcomes) times the two conditioning phases (first and second half of each session) times the three conditions (reward, neutral, and punishment). In all cases, the regressors of interest were convolved with a canonical hemodynamic response function (HRF). To correct for motion artifact, subject-specific realignment parameters were modeled as covariates of no interest. Linear contrasts of regression coefficients were computed at the individual subject level and then taken to a group level random-effects analysis. At group level, we performed two statistical analyses: first a one-sample t test to find brain regions where regression coefficients were significant across subjects, and second a correlation with individual payoffs to find brain regions where regression coefficients increased with higher conditioning effect. A threshold of p < 0.05 after familywise error (FWE) correction for multiple comparisons was applied to avoid any a priori on brain localization. A more liberal threshold (p < 0.001, uncorrected) was also used to observe the extension of significant activations. To further illustrate activations, time courses were estimated by fitting a flexible basis set of finite impulse responses (FIRs), separated from the next by one scan (1.95 s). Both regression coefficients and time courses were then averaged across subjects, pooling together the voxels that passed the conservative threshold in statistical parametric maps (SPMs).\n\nREFERENCES:\n1. BayleyP.J.FrascinoJ.C.SquireL.R.Robust habit learning in the absence of awareness and independent of the medial temporal lobeNature436200555055316049487\n2. BernsG.S.CohenJ.D.MintunM.A.Brain regions responsive to novelty in the absence of awarenessScience2761997127212759157889\n3. ClarkR.E.SquireL.R.Classical conditioning and brain systems: the role of awarenessScience280199877819525860\n4. CohenN.J.EichenbaumH.DeacedoB.S.CorkinS.Different memory systems underlying acquisition of procedural and declarative knowledgeAnn. N Y Acad. Sci.444198554713860122\n5. DawN.D.DoyaK.The computational neurobiology of learning and rewardCurr. Opin. Neurobiol.16200619920416563737\n6. DeichmannR.GottfriedJ.A.HuttonC.TurnerR.Optimized EPI for fMRI studies of the orbitofrontal cortexNeuroimage19200343044112814592\n7. DelgadoM.R.Reward-related responses in the human striatumAnn. N Y Acad. Sci.11042007708817344522\n8. DestrebecqzA.CleeremansA.Can sequence learning be implicit? New evidence with the process dissociation procedurePsychon. Bull. Rev.8200134335011495124\n9. GraybielA.M.The basal ganglia: learning new tricks and loving itCurr. Opin. Neurobiol.15200563864416271465\n10. HikosakaO.NakaharaH.RandM.K.SakaiK.LuX.NakamuraK.MiyachiS.DoyaK.Parallel neural networks for learning sequential proceduresTrends Neurosci.22199946447110481194\n11. KnightD.C.NguyenH.T.BandettiniP.A.Expression of conditional fear with and without awarenessProc. Natl. Acad. Sci. USA1002003152801528314657356\n12. KnutsonB.CooperJ.C.Functional magnetic resonance imaging of reward predictionCurr. Opin. Neurol.18200541141716003117\n13. KrawczykD.C.GazzaleyA.D'EspositoM.Reward modulation of prefrontal and visual association cortex during an incentive working memory taskBrain Res.1141200716817717320835\n14. LagnadoD.A.NewellB.R.KahanS.ShanksD.R.Insight and strategy in multiple-cue learningJ. Exp. Psychol. Gen.135200616218316719649\n15. LovibondP.F.ShanksD.R.The role of awareness in Pavlovian conditioning: empirical evidence and theoretical implicationsJ. Exp. Psychol. Anim. Behav. Process.28200232611868231\n16. McClureS.M.YorkM.K.MontagueP.R.The neural substrates of reward processing in humans: the modern role of FMRINeuroscientist10200426026815155064\n17. MilnerB.SquireL.R.KandelE.R.Cognitive neuroscience and the study of memoryNeuron2019984454689539121\n18. MorrisJ.S.ÖhmanA.DolanR.J.Conscious and unconscious emotional learning in the human amygdalaNature39319984674709624001\n19. O'DohertyJ.DayanP.SchultzJ.DeichmannR.FristonK.DolanR.J.Dissociable roles of ventral and dorsal striatum in instrumental conditioningScience304200445245415087550\n20. O'DohertyJ.P.HamptonA.KimH.Model-based fMRI and its application to reward learning and decision makingAnn. N Y Acad. Sci.11042007355317416921\n21. OlssonA.PhelpsE.A.Learned fear of “unseen” faces after Pavlovian, observational, and instructed fearPsychol. Sci.15200482282815563327\n22. PackardM.G.KnowltonB.J.Learning and memory functions of the Basal GangliaAnnu. Rev. Neurosci.25200256359312052921\n23. PagnoniG.ZinkC.F.MontagueP.R.BernsG.S.Activity in human ventral striatum locked to errors of reward predictionNat. Neurosci.52002979811802175\n24. PecinaS.SmithK.S.BerridgeK.C.Hedonic hot spots in the brainNeuroscientist12200650051117079516\n25. PessiglioneM.SeymourB.FlandinG.DolanR.J.FrithC.D.Dopamine-dependent prediction errors underpin reward-seeking behaviour in humansNature44220061042104516929307\n26. PoldrackR.A.PackardM.G.Competition among multiple memory systems: converging evidence from animal and human brain studiesNeuropsychologia41200324525112457750\n27. RamnaniN.MiallR.C.Instructed delay activity in the human prefrontal cortex is modulated by monetary reward expectationCereb. Cortex13200331832712571121\n28. SeitzA.R.WatanabeT.Psychophysics: Is subliminal learning really passive?Nature42220033612621425\n29. SmallD.M.GitelmanD.SimmonsK.BloiseS.M.ParrishT.MesulamM.M.Monetary incentives enhance processing in brain regions mediating top-down control of attentionCereb. Cortex1520051855186515746002\n30. SquireL.R.Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humansPsychol. Rev.9919921952311594723\n31. SuttonR.S.BartoA.G.Reinforcement Learning1998MIT PressCambridge, MA\n32. ThorndikeE.L.Animal Intelligence: Experimental Studies1911MacmillanNew York\n33. WilkinsonL.ShanksD.R.Intentional control and implicit sequence learningJ. Exp. Psychol. Learn. Mem. Cogn.30200435436914979810\n34. YacubianJ.GlascherJ.SchroederK.SommerT.BrausD.F.BuchelC.Dissociable systems for gain- and loss-related value predictions and errors of prediction in the human brainJ. Neurosci.2620069530953716971537"
+}

batch_13/PMC2572835.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2572835",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2572835\nAUTHORS: Xinjiang Cai, David E. Clapham\n\nABSTRACT:\nThe mammalian CatSper ion channel family consists of four sperm-specific voltage-gated Ca2+ channels that are crucial for sperm hyperactivation and male fertility. All four CatSper subunits are believed to assemble into a heteromultimeric channel complex, together with an auxiliary subunit, CatSperβ. Here, we report a comprehensive comparative genomics study and evolutionary analysis of CatSpers and CatSperβ, with important correlation to physiological significance of molecular evolution of the CatSper channel complex. The development of the CatSper channel complex with four CatSpers and CatSperβ originated as early as primitive metazoans such as the Cnidarian Nematostella vectensis. Comparative genomics revealed extensive lineage-specific gene loss of all four CatSpers and CatSperβ through metazoan evolution, especially in vertebrates. The CatSper channel complex underwent rapid evolution and functional divergence, while distinct evolutionary constraints appear to have acted on different domains and specific sites of the four CatSper genes. These results reveal unique evolutionary characteristics of sperm-specific Ca2+ channels and their adaptation to sperm biology through metazoan evolution.\n\nBODY:\nIntroductionIn spermatozoa, Ca2+ influxes through plasma membrane Ca2+ channels play a key role in mediating sperm maturation, motility, and the acrosome reaction [1]–[3]. Sperm motility is driven by flagellar ATP-dependent dynein motor proteins. Basic flagellar architecture is conserved across species; flagellar movement does not require calcium, or even a plasma membrane. In general, sperm acquire motility after they exit the male reproductive system.Both initiation of motility and modulation of motility vary in a species-specific fashion. Initiation of sperm motility display many features distinct from modulation of motility. In many mammals, motility is characterized by symmetrical movement of the tail and progressive movement (normal swimming), but the initiating signal or signals are poorly understood. Hyperactivated motility in mammals occurs later, as the sperms encounter progressively more alkaline environments as they ascend the reproductive tract. The primary characteristic of sperm hyperactivated motility in well-characterized mammalian species is a large bend angle between head and tail [4], [5]. This larger sweep of the tail results in substantially more force than the force of sperm cells swimming before hyperactivation [6]. This force has been proposed to free sperm from surfaces or trapping spaces in the uterus and oviduct, and/or penetrate the cumulus and the thick protective wall of the zona pellucida [7]. In sea urchin, the chemoattractant peptide, Resact, initiates an increase in intracellular [Ca2+] and change in the bend angle, resulting in larger swimming arcs [8]. In contrast, marine fish spermatozoa motility [9] is initiated by the large change in osmolality as they exit the fish and encounter seawater, but this initiation of motility is not Ca2+-dependent [10]. These marine teleosts have up to 10 times the beat frequency of mammalian sperm cells, but swim only for short distances to eggs deposited in seawater [9]. A change in flagellar bend angle, equivalent to hyperactivation in mammals, has not been reported, and indeed may not be required for sperm docking to the marine fish egg pyle. In birds and reptiles, sperm are often stored in specialized tubules connected to the oviduct. In chickens, initiation, or re-initiation of motility may be related to temperature sensitivity of a plasma membrane Ca2+ ATPase [11]. Thus, although all spermatozoa have the basic ATP-dynein motor required for normal swimming, they appear to differ in modulation of motility in a species-specific manner.Several types of Ca2+-permeant channel proteins, such as voltage-gated Ca2+ (CaV), CatSper, cyclic nucleotide-gated (CNG), and Transient Receptor Potential (TRP) channels, have been detected in spermatocytes or spermatozoa [1]–[3], [12], but until recently their function could only be assayed in spermatocytes. Among the Ca2+-permeant channels, CatSper channels mediate the dominant Ca2+-carrying current in mouse epididymal spermatozoa and induce sperm hyperactivation during sperm capacitation [13]. Intracellular alkalinization triggers CatSper activation and subsequent hyperactivation of motility in mice. In addition, in the absence of CatSper current, sperm cell motility endurance declines [12]–[15].The mammalian CatSper family of ion channels is composed of four members, CatSper1-4 [14], [16]–[19]. All four CatSper proteins are expressed in sperm cells and functionally localized to the principal piece of the sperm tail [12]. Like most of the voltage-gated ion channels [20], the six-transmembrane-spanning (6-TMS) CatSpers are believed to form a tetrameric structure. All four CatSpers seem to be required to mediate functional alkalinization-activated Ca2+-selective sperm currents [12]. Targeted deletion of any of the four CatSper genes results in mouse male infertility and an identical sperm cell phenotype including loss of sperm hyperactivation [12], [14], [21]–[23], while mice lacking other ion channel genes, if viable, are fertile [1]. Interestingly, a sperm-specific two-TMS protein, CatSperβ, is associated with the CatSper1 channel subunit in mouse testis [24]. CatSperβ displays a similar localization pattern with CatSpers, copurifies with CatSper proteins, and is absent in CatSper1\n−/− sperm. Thus, CatSperβ is an auxiliary subunit of the CatSper channel complex [24], but its role in channel activity is unknown.In the absence of suitable heterologous expression systems, our current understanding of the physiological role of CatSpers is primarily based on studies using mouse models [12]–[14], [21]–[23]. Reproductive proteins usually undergo rapid evolution and functional divergence because selective evolutionary forces such as sperm competition, sexual selection and sexual conflict may act on reproductive proteins to generate phenotypic diversity through speciation [25]–[27]. In addition, comparative genomics studies and evolutionary analyses can provide insights into molecular, biophysical, and biochemical analyses of functional and regulatory mechanisms of ion channels and transporters [28]–[33]. Therefore, to better understand the functional role of the CatSper channel complex, we set out to determine the evolutionary origins and patterns of CatSpers and CatSperβ using extensive database mining and rigorous phylogenetic analyses, and to analyze the evolutionary rates and functional divergence of CatSpers and CatSperβ.Materials and MethodsDatabase mining, sequence alignment and phylogenetic analysisTBlastN and BlastP searches [34] using protein sequences of Homo sapiens CatSper and CatSperβ protein sequences were performed on the genomic and protein databases of the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/blast/), Ensembl (http://www.ensembl.org/Multi/blastview), and the Joint Genome Institute (JGI) (http://www.jgi.doe.gov/). H. sapiens and Nematostella vectensis CatSper and CatSperβ sequences were also used for BlastN and TBlastN searches of three sponge databases - Amphimedon queenslandica genomic traces at NCBI, SpongeBase (http://spongebase.uni-mainz.de/), and Oscarella carmela EST database [35]. TBlastN searches were also conducted on the survey genomic database of the elephant shark, Callorhinchus milii (http://esharkgenome.imcb.a-star.edu.sg/). Protein sequences of the bacterial NaV channel superfamily were obtained from the bacterial protein cluster CLS1187052 at the NCBI protein cluster database and sequences characterized previously [36].Sequence alignments, manual editing, and phylogenetic analysis for the collected dataset were carried out essentially as previously described [29], [32], [37].Chromosome syntenyThe Evolutionarily Conserved Regions (ECR) browser (http://ecrbrowser.dcode.org/) was used for initial screening of syntenic chromosomal regions among H. sapiens, Mus musculus, Gallus gallus, Fugu rubripes, and Xenopus tropicalis genomes. However, only M. musculus and G. gallus genomes showed substantial synteny in H. sapiens chromosomal regions flanking the CatSpers and CatSperβ genes, and were used for further analysis. The flanking orthologous genes were identified on the ECR browser and/or the NCBI Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/).Estimation of non-synonymous (dN) to synonymous (dS) nucleotide substitution ratio\nH. sapiens and M. musculus mRNA sequences of CatSper and CatSperβ genes were retrieved from the NCBI database, and then converted to codon alignments by the PAL2NAL server [38], using corresponding protein sequence alignments. dN and dS values were calculated with the codeml program implemented in the PAML package [39].Results and DiscussionDevelopment of four distinct CatSper genes and CatSperβ in early EumetazoaMouse gene knockout studies indicate that all four CatSper subunits are required to mediate functional Ca2+-selective sperm currents necessary for sperm hyperactivation [12]–[14], [21]. However, previous reports have suggested the presence of fewer copies of CatSper homologues in early deuterostomes: three in Ciona intestinalis\n[40] and two in sea urchin testis [24]. Here, extensive genomic analysis of two sea squirts, C. intestinalis and Ciona savignyi, the sea urchin, Strongylocentrotus purpuratus, and the most basal extant chordate lineage, the amphioxus Branchiostoma floridae\n[41], [42], demonstrated the presence of four CatSper subunits and single copies of CatSperβ in these four species (Fig. 1A and Supplementary Table S1). Therefore, the CatSper channel complex containing four CatSper subunits and one CatSperβ had developed in early deuterostomes.10.1371/journal.pone.0003569.g001Figure 1Phylogenetic reconstruction of the evolutionary history of CatSpers and CatSperβ in metazoans.The phylogenetic trees of the CatSper protein family (A) and the CatSperβ protein family (B) were constructed using the maximum likelihood approach [37]. Two putative primitive Ca2+ channels (MbrCaVS5 and NveCaVS17) were used as the outgroup for the CatSper family. Bootstrap values of more than 60 are shown at corresponding branches. The CatSper1-4 groups are indicated by rectangular bars, with invertebrate CatSpers filled with white and vertebrate CatSpers with gray. Note that protein sequences that failed in the chi-square test in Tree-Puzzle [69] or contained more than 15% gaps in the refined alignments were not subjected to phylogenetic analysis (Table S1). Abbreviations used: Aca, A. carolinensis; Bfl, B. floridae; Bta, B. taurus; Cfa, C. familiaris; Cin, C. intestinalis; Cja, C. jacchus; Cpo, C. porcellus; Csa, C. savignyi; Eca, E. caballus; Hsa, H. sapiens; Mdo, M. domestica; Mmu, M. mulatta; Mus, M. musculus; Nve, N. vectensis; Oan, O. anatinus; Pan, P. anubis; Ptr, P. troglodytes; Rno, R. norvegicus; Spu, S. purpuratus.As one of the unicellular ancestors of Metazoa [43], choanoflagellates possess homologues of many components of animal Ca2+ signaling and amplification pathways (the Ca2+ signaling ‘toolkit’ [44]) but not CatSper channels [33]. To further determine the evolutionary origins of CatSper channels, we next searched the currently available databases of metazoans that branched off the metazoan stem before the radiation of bilaterians – the genome sequences of Cnidarian Nematostella vectensis\n[45] and Placozoa Trichoplax adhaerens at JGI, the EST database of Ctenophora Pleurobrachia pileus, and EST and genomic trace databases of ancestral metazoan Porifera sponges.Indeed, N. vectensis has four CatSper gene homologues, which can be classified into CatSper1-4 based on phylogenetic analysis, and a single homolog of CatSperβ (Fig. 1A and Table S1). Three EST sequences were also identified in the P. pileus EST database, which were further categorized as CatSper1 and CatSper2 (Table S1). No significant hits were found in the sponge EST and genomic trace databases and in the Placozoa genome. It should be noted that the negative results from currently available sponge databases should be viewed as provisional until the complete genome or sperm EST sequences from sponges have been examined.The transmembrane segments (TMS) and the putative pore region of human and mouse CatSpers are highly conserved [17]. Sequence alignment of CatSper sequences identified in this study showed similar observations across invertebrate and vertebrate species (Fig. 2). The key aspartate residue of CatSpers, presumably corresponding to the highly conserved glutamate residues (or occasionally aspartate residues) in the pore loop regions that underlie Ca2+ selectivity in many CaV channels [46], is absolutely conserved (Fig. 2).10.1371/journal.pone.0003569.g002Figure 2Conservation of the putative pore domain of CatSpers from invertebrate and vertebrate species.The putative pore regions of CatSpers from selected species, corresponding to amino acids 210–239 of human CatSper3 (GenBank Accession No., NP_821138.1), were aligned and manually edited to improve alignments. Invertebrate CatSpers are indicated with white rectangular bars and vertebrate CatSpers with gray rectangular bars. The key aspartate residue of CatSpers is overlined with an asterisk symbol. For abbreviations, please refer to the Fig. 1 legend.The identification of four CatSper homologues and CatSperβ in Cnidarians, but not in choanoflagellates, suggests that the origin of the CatSper complex could be dated back to early Eumetazoa, much earlier than previously thought [24]. Many other ion channels and transporters usually underwent further expansion in basal vertebrates after divergence from Urochordata [29], [31], [32], [47], [48]. However, the CatSper channel complex appears to be preserved from early metazoans to vertebrates without any further duplication events. The stringent requirement for four different CatSpers to form a putative heterotetrameric channel complex might render any further gene duplication of CatSpers excessive and the gene then be degenerated. For instance, a recent primate-specific duplication of CatSper2 resulted in a CatSper2 pseudogene (Table S4).Lineage-specific gene loss of CatSper and CatSperβ across the metazoan evolutionSperm cells express a variety of Ca2+ channels and transporters [2], the majority of which are also expressed in other tissues and are often highly conserved from early metazoans to mammals, such as TRP channels [49], [50], Na+/Ca2+ exchangers (NCX) [29], [48], and sarco/endoplasmic reticulum Ca2+ ATPases [51], [52]. In contrast, little is known about the detailed evolutionary pattern of sperm-specific ion channels.\nFig. 3 shows the comprehensive evolutionary genomics of the CatSper ion channel complex. We have identified novel CatSper sequences from non-mammal vertebrates, such as the Anole lizard Anolis carolinensis, and cartilaginous fishes Callorhinchus milii and Leucoraja erinacea (Table S1). The CatSper channel complex is completely absent from a diverse sampling of protostome genomes. Interestingly, even though they are present in basal Chordata (amphioxus and sea squirts), all four CatSpers and CatSperβ appear to be lacking in several vertebrate lineages such as Agnatha (jawless fishes), Teleostei (bony fishes), Amphibia (frogs), and Aves (birds). This distinctive lineage distribution pattern of the CatSper channel complex (Fig. 3) is in drastic contrast to that of most ion channels and transporters characterized previously, which are usually highly conserved in metazoans [29], [31], [32], [40], [47], [48].10.1371/journal.pone.0003569.g003Figure 3The metazoan phylogeny describing the presence or absence of CatSpers and CatSperβ in examined metazoan genomes.The phylogenetic branching patterns of metazoans (not proportional to the evolutionary rates) were extracted from the Tree of Life project (http://www.tolweb.org/tree/) as of July 30, 2008. The metazoan lineages known to contain the CatSper channel complex are indicated by boxes with a solid line and gray background, while those believed to lack the channel complex are marked by boxes with a broken line and white background. Two lineages (Porifera and cartilaginous fishes) are underlined since the results obtained from these lineages are based solely on survey genome or genomic traces rather than complete genome sequences. The availability of more metazoan genome sequences in the future will probably expand and/or refine the lineage branches shown here.Two main evolutionary mechanisms might account for the intermittent pattern of presence or absence of genes in the genomes: lineage-specific gene loss and horizontal gene transfer [53], [54]. A bacterial voltage-gated Na+ channel, NaVBP, is believed to have the closest relationship with CatSpers, before the mammalian CaV and NaV channel classes [19]. Similar to CatSpers, NaVBP is activated by alkalinization and is important for motility [55]. Horizontal gene transfer between prokaryotes and metazoans is not common, but possible [56]. Perhaps during the process of metazoan evolution, a bacterial channel protein like NaVBP became incorporated into sperm cells of some, but not all metazoan genomes. Such a bacterial channel would have undergone further functional divergence to enhance sperm motility by inducing Ca2+-dependent hyperactivation and subsequent species-dependent adaptation to reflect phylogeny. Alternatively, the absence of the CatSper complex in the specific metazoan lineages described above could be explained as a result of differential gene loss. Thus, in sperm cells of those metazoan lineages, Ca2+ influx across plasma membrane might not be important for sperm motility endurance or hyperactivated motility. Of course, other types of Ca2+ channels/sources might have substituted for the CatSper channel complex.If lineage-specific gene loss had occurred, it might first render the gene degenerate and non-functional. Subsequent neutral evolution could then mask a substantial portion of the gene sequence, but short fragments in the coding sequence of the degenerated gene might still be present, with flanking functional genes on the syntenic chromosomes conserved across closely related species. Therefore, we compared syntenic chromosomal regions containing CatSper and CatSperβ genes in mouse and human genomes and those in bird, amphibian, and bony fish genomes.By examining genomic regions with the ECR browser and the NCBI genomic database, we found highly conserved synteny between selected regions of the mouse and human genomes and those of the chicken genome for CatSper2-4 and CatSperβ (Fig. 4 and Table S2, S3, S4, S5, S6). As shown in Fig. 4A and Table S2, 10 genes flanking the 5′- and 3′-end of CatSperβ on human chromosome 14 and corresponding orthologous genes on chicken chromosome 5 are syntenic. By using three different gene-finding programs [37], subsequent examination of chicken genomic sequence located between SMEK1 and TC2N genes did not yield any obvious gene coding region, but a portion of the genomic sequence could be translated into a 66-aa fragment with high similarity to human CatSperβ (Fig. 4A) and other CatSperβ sequences (data not shown). Similar observations have also been observed for CatSper2 and CatSper3 (Tables S4 and S5). Thus, CatSper sequences might have been degenerated in chick genomes, with short fragments of coding sequences still present.10.1371/journal.pone.0003569.g004Figure 4Chromosomal synteny between human and chicken genomes and sequence alignment of degenerated CatSperβ and CatSper3 fragments from bird genomes.\nA, chromosomal regions harboring degenerate DNA fragment of putative CatSperβ on chicken chromosome 5 with synteny to human chromosome 14. Shown here are ten genes flanking H. sapiens CatSperβ on human chromosome 14 and syntenic genes on chicken chromosome 5 (not to the scale of base-pair length). CatSperβ is indicated with a black rectangle and other genes with gray rectangles. The location of the degenerate genomic fragment of putative chicken CatSperβ is specified with an open rectangle, and the translated sequence is aligned with HsaCatSperβ. B, H. sapiens CatSper3 is aligned with the short stretch of sequences translated from putative degenerate chicken and zebra finch CatSper3 fragments. Abbreviations for flanking genes can be found in Table S2.We next examined the preliminary genome sequence of the zebra finch Taeniopygia guttata. We also found a stretch of genomic sequence that could be translated into a short CatSper3 fragment sharing sequence similarity with a region in human CatSper3 and the putative chicken CatSper3 fragment (Fig. 4B). Such a short CatSper4 fragment was also made in the genome of the snail Lottia gigantean (data not shown). In addition, the EST libraries of a non-insect arthropod, the Arctic springtail Onychiurus arcticus, contain an EST clone (GenBank Accession No. EW749693.1) encoding a partial protein sequence in which the C-terminal (aa 68–218) displayed high similarity to TMS domains of CatSper4, with N. vectensis CatSper4 as the most significant hit (E-value, 3×e−27). The N-terminal 67 aa share no sequence homology to other CatSpers or known proteins.Taken together, we conclude that the intermittent presence/absence pattern of the CatSper channel complex in metazoan genomes was likely derived from lineage-specific gene loss. Importantly, CatSperβ is present in genomes in which CatSpers are identified, but is missing in genomes that lack CatSpers. Thus, functional association between CatSperβ and CatSpers is not only critical for CatSperβ stability at the cellular level, i.e., mouse sperm cells [24], but might also relate to the all-or-none presence of CatSpers and CatSperβ at the genomic level.Mouse spermatozoa have only two primary ion channel currents in normal saline solution, ICatSper and IKsper\n[15], [57]. The Ca2+ current is mediated by CatSpers while the K+ current is mediated presumably by mSlo3. Both currents are effectively gated by alkalinization and voltage changes. Activated K+ current hyperpolarizes the flagellar plasma membrane and thus maximizes the entry of Ca2+ via Catsper by increasing driving force on Ca2+. In sea urchin, direct whole-sperm voltage clamp recordings have not been possible, but indirect evidence suggests that the chemoattractant peptide Resact activates a cyclic-nucleotide-gated (CNG) K+-selective conductance [8] (presumably, Sp-tetraKCNG channel [58]) . This K+ current may act in place of IKSper to enhance Ca2+ entry. Mammalian spermatocytes may contain CNGB1 and perhaps CNGA3 channel proteins [59], but to date, currents consistent with CNG channels have not been observed in mouse sperm [57]. The simplest interpretation is that CatSpers are critical for the alteration of motility patterns for chemotaxis (sea urchin) and hyperactivation (mammals), but that the type of K channel activated (CNG or mSlo3) modulates CatSper's roles specific to the species. Thus, sea urchin eggs, with their thin walls, are easy for sperm to penetrate but sperm must be guided to the egg in the less restricted ocean environment by chemotactic peptide alteration of sea urchin sperm motility. In contrast, mouse spermatozoa require more force, and thus hyperactivated motility, to penetrate the thicker oocyte wall. Whether there is an additional chemotactic factor guiding spermatozoa in mammals is not established. Since birds and fish have thin oocyte cell walls, CatSper may have provided no evolutionary advantage and was degenerated in these species.Such an extensive lineage-specific gene loss of an entire ion channel complex through metazoan evolution, especially in vertebrates, has not been documented in other channels and transporters. In some cases, such as for Na+/Ca2+ exchangers, one member, NCX4, which arose in basal vertebrates, persevered in teleost, amphibian and reptilian species but was lost in mammals and birds. Most of NCX members, NCX1-3, however, were still retained in all vertebrates examined [48]. Interestingly, a sperm-specific and unusual putative Na+/H+ exchanger required for sperm motility and fertility [60] also shows similar extensive lineage-specific gene loss in metazoans (Cai, X, unpublished observation). Thus, further evolutionary genomics studies of sperm-specific channels and transporters will no doubt shed novel insights into physiological roles of ion transport in sperm biology.Our database search did not identify the putative primordial CatSper sequence. It remains possible that the introduction of the primordial, probably distantly related, CatSper channel could be made through horizontal gene transfer between prokaryotes and primitive metazoan species, such as single 6-TMS domain NaV channels NaVBP [19], NaChBac [61], and 11 bacterial NaChBac homologues [36]. All these bacterial NaV channel homologues share a glutamate residue as the key acidic residue in the putative channel pore region [36], in contrast to the key aspartate residue conserved in CatSpers (Fig. 2). We analyzed 40 more bacterial protein sequences (protein cluster CLS1187052), which display high sequence homology and structural similarity to NaChBac. None of these 40 bacterial homologues grouped with CatSper sequences (Fig. 5). However, two (GI No. 56963529 and 134099759) have the key aspartate residue in the putative pore region, and it would be interesting compare their Ca2+-selectivity, voltage-dependence and kinetics with the CatSper channel.10.1371/journal.pone.0003569.g005Figure 5Phylogenetic tree of the 6-TMS bacterial ion channel family.A bootstrapped maximum parsimony tree was constructed with a NaV channel homologue in choanoflagellates (MbrNaV-S3L) as an outgroup. Two CatSper sequences, NveCatSper2 and HsaCatSper2, and two putative CaV channel homologues, MbrCaVS5-1 and NveCaV-S17, are also included. The position of the key aspartate residue in the pore region of two bacterial proteins is marked by gray circles. Bootstrap values of >50 are shown at corresponding branches. Each branch of the tree is labeled with the GI numbers in the NCBI protein database for most organisms. NaChBac, NaVSP and NaVPZ channels were functionally characterized previously [36], [61]. Abbreviations: Hsa, H. sapiens; Mbr, M. brevicollis; Nve, N. vectensis.Rapid evolution of the CatSper channel complexReproductive system proteins are subjected to selective evolutionary forces and often undergo accelerated evolution and functional divergence [25]–[27]. Indeed, mammalian sperm-specific proteins, mostly cell surface binding proteins and enzymes, are rapidly evolving [26]. Sperm-specific channels and transporters have not been extensively studied, but positive selection on indel substitutions does occur in the first exon of the CatSper1 gene in mammals [27], [62].Non-synonymous (dN) and synonymous (dS) nucleotide substitution values as well as dN/dS ratios are often used for detecting evolutionary rates [63]. We calculated dN and dS values of human-mouse orthologous gene pairs of CatSper and CatSerβ genes, using the codeml program implemented in the PAML package [39]. The dN value of CatSper and CatSerβ genes (average = 0.268, Table 1) is 5.8- and 3.2-fold larger, respectively, than housekeeping genes and non-sperm tissue-specific genes [64], and 1.5 fold larger than other sperm-specific genes [26]. Thus, both CatSpers and CatSperβ appear to be subjected to high selective forces that promote amino acid diversity.10.1371/journal.pone.0003569.t001Table 1\ndN/dS ratios between human and mouse CatSper and CatSperβ genes.Full-length\ndN\n\ndS\n\ndN/dS\nTMS\ndN\n\ndS\n\ndN/dS\n\nCatSper1\n0.3531.3580.260\nCatSper1\n0.0961.5470.062\nCatSper2\n0.2260.6980.324\nCatSper2\n0.13560.4040.336\nCatSper3\n0.2331.2340.189\nCatSper3\n0.1791.2560.143\nCatSper4\n0.2021.0460.193\nCatSper4\n0.0981.2830.077\nCatSperβ\n0.3280.8010.409Average0.2681.0270.275Ref. [68]\nMeansMeansMeansRef. [27]\nMeansMeansMeansHousekeeping genes0.0460.4470.093Other tissue-specific genes0.0730.410.19Tissue-specific genes0.0830.492\n*0.172Sperm-specific genes0.180.450.50*Calculated from 7 groups of tissue specific genes.All data shown here are derived form comparison of human-mouse orthologs.Abbreviation: TMS, transmembrane segments.The dS values (Table 1) indicate that CatSper and CatSerβ genes contain high mutation rates, with an ∼2-fold increase over those of housekeeping genes, other tissue-specific genes and sperm-specific genes. Thus, CatSper and CatSerβ genes have an average dN/dS ratio that is smaller than other sperm-specific genes. Nevertheless, the average dN/dS ratio of CatSper and CatSerβ genes is ∼3- and 1.6-fold larger than housekeeping genes and other tissue-specific genes, respectively (Table 1). In addition, even though mammalian synonymous mutation rates (dS values) are generally considered selectively neutral, recent studies suggest that synonymous mutations might also be subjected to selection, possibly through their effects on splicing and/or mRNA stability [65]. Thus, CatSper and CatSerβ mRNA might also be processed, as shown in recent studies on human and mouse homologues of the cystic fibrosis transmembrane conductance regulator Cl− channel [66]. Interestingly, CatSperβ has the highest dN/dS ratio of the genes examined (0.409; Table 1). The exact number of CatSperβ subunits in the channel complex is not yet known. Presumably, based on the observations of other tetrameric 6-TMS voltage-gated ion channels such as K+ channels [67], multiple CatSperβ proteins may be in the CatSper1-4 channel complex.Rapid evolution of all four CatSper proteins appeared to have occurred at different domains in CatSpers. For instance, CatSper1 underwent strong positive selection in the N-terminal cytoplasmic domain [27], [62], but displayed evolutionary constraint in the 6-TMS domains (dN/dS, 0.062) (Table 1). CatSper 4 displays similar constraints on its TMS domains with a dN/dS value of 0.077. In contrast, the TMS domains of CatSper2 and CatSper3 were driven by higher selective forces (dN/dS, 0.336 and 0.143, respectively). The distinct evolutionary patterns in CatSper channel domains may promote channel diversity in modulating potential protein or second-messenger interactions [12]. CatSperβ, as perhaps the sole accessory protein, may have evolved even faster to accommodate concerted evolution at differential evolutionary sites of CatSper proteins, resulting in a high dN/dS ratio.Functional divergence of the CatSper protein familyThe identification of many CatSper sequences, especially from invertebrates, allows us to calculate the functional divergence between CatSper1-4 (as shown previously [32], [68]). The coefficients of functional divergence between pairs of CatSper1-4 groups were estimated to be significant, ranging from 0.336±0.092 (CatSper3 vs. CatSper4) to 0.536±0.116 (CatSper2 vs. CatSper4) (Table S7). Thus, each CatSper group appears to have experienced altered evolutionary constraints after divergence from a putative ancestral duplication. Consistently, functional distance analysis also shows similarly long functional branch lengths (Fig. S1), implying that all four CatSpers diverged at comparable overall evolutionary rates from the putative primordial CatSper protein. Nevertheless, despite the similar overall evolutionary rates, the evolutionary processes might act differentially on specific sites that are unique in each CatSper gene (data not shown). Our data suggest that CatSper1-4 have acquired evolutionary novelties through substantial altered functional constraints after possible ancestral replication. Finally, rapid evolution of CatSpers (Table 1) still plays an important role in modulating CatSper functions in mammalian lineages.In conclusion, we have carried out a comprehensive evolutionary genomics study of CatSper and CatSperβ proteins that constitutes the Ca2+ channel complex critical for sperm Ca2+ hyperactivation in mammals. With continued advances in genome biology, the evolutionary genomics approaches undertaken here will further illuminate the lineage-specific distribution of Ca2+ channels and transporters at the genomic scale and greatly facilitate deciphering the Ca2+ signaling codes in a species- and/or tissue-specific manner.Supporting InformationFigure S1Tree topology of functional distance analysis of the CatSper protein family. Type I functional branch length bF was calculated as described in Materials and methods in Table S7. bF is an estimation of evolutionary distance of each CatSper group (1–4) to the putative primordial CatSper protein before replication (center circle).(2.31 MB TIF)Click here for additional data file.Table S1List of CatSper Proteins Used for Analyses(0.31 MB PDF)Click here for additional data file.Table S2Genome Synteny - CatSperβ(0.07 MB PDF)Click here for additional data file.Table S3Genome Synteny - CatSper1(0.06 MB PDF)Click here for additional data file.Table S4Genome Synteny - CatSper2(0.07 MB PDF)Click here for additional data file.Table S5Genome Synteny - CatSper3(0.06 MB PDF)Click here for additional data file.Table S6Genome Synteny - CatSper4(0.06 MB PDF)Click here for additional data file.Table S7(0.03 MB PDF)Click here for additional data file.\n\nREFERENCES:\n1. DarszonAAcevedoJJGalindoBEHernandez-GonzalezEONishigakiT\n2006\nSperm channel diversity and functional multiplicity.\nReproduction\n131\n977\n988\n16735537\n2. Jimenez-GonzalezCMichelangeliFHarperCVBarrattCLPublicoverSJ\n2006\nCalcium signalling in human spermatozoa: a specialized ‘toolkit’ of channels, transporters and stores.\nHum Reprod Update\n12\n253\n267\n16338990\n3. WhitakerM\n2006\nCalcium at fertilization and in early development.\nPhysiol Rev\n86\n25\n88\n16371595\n4. IshijimaSBabaSAMohriHSuarezSS\n2002\nQuantitative analysis of flagellar movement in hyperactivated and acrosome-reacted golden hamster spermatozoa.\nMol Reprod Dev\n61\n376\n384\n11835583\n5. CarlsonAEWestenbroekREQuillTRenDClaphamDE\n2003\nCatSper1 required for evoked Ca2+ entry and control of flagellar function in sperm.\nProc Natl Acad Sci U S A\n100\n14864\n14868\n14657352\n6. FauciLJDillonR\n2006\nBiofluid mechanics of reproduction.\nAnnual Review of Fluid Mechanics\n38\n371\n394\n7. SuarezSS\n2008\nRegulation of sperm storage and movement in the mammalian oviduct.\nInt J Dev Biol\n52\n455\n462\n18649258\n8. StrunkerTWeyandIBonigkWVanQLoogenA\n2006\nA K+-selective cGMP-gated ion channel controls chemosensation of sperm.\nNat Cell Biol\n8\n1149\n1154\n16964244\n9. CossonJGroisonALSuquetMFauvelCDreannoC\n2008\nMarine fish spermatozoa: racing ephemeral swimmers.\nReproduction\n136\n277\n294\n18524881\n10. OdaSMorisawaM\n1993\nRises of intracellular Ca2+ and pH mediate the initiation of sperm motility by hyperosmolality in marine teleosts.\nCell Motil Cytoskeleton\n25\n171\n178\n8324831\n11. ThomsonMFWishartGJ\n1991\nTemperature-mediated regulation of calcium flux and motility in fowl spermatozoa.\nJ Reprod Fertil\n93\n385\n391\n1838568\n12. QiHMoranMMNavarroBChongJAKrapivinskyG\n2007\nAll four CatSper ion channel proteins are required for male fertility and sperm cell hyperactivated motility.\nProc Natl Acad Sci U S A\n104\n1219\n1223\n17227845\n13. KirichokYNavarroBClaphamDE\n2006\nWhole-cell patch-clamp measurements of spermatozoa reveal an alkaline-activated Ca2+ channel.\nNature\n439\n737\n740\n16467839\n14. RenDNavarroBPerezGJacksonACHsuS\n2001\nA sperm ion channel required for sperm motility and male fertility.\nNature\n413\n603\n609\n11595941\n15. NavarroBKirichokYChungJJClaphamDE\n2008\nIon channels that control fertility in mammalian spermatozoa.\nInt J Dev Biol\n52\n607\n613\n18649274\n16. QuillTARenDClaphamDEGarbersDL\n2001\nA voltage-gated ion channel expressed specifically in spermatozoa.\nProc Natl Acad Sci U S A\n98\n12527\n12531\n11675491\n17. LobleyAPierronVReynoldsLAllenLMichalovichD\n2003\nIdentification of human and mouse CatSper3 and CatSper4 genes: characterisation of a common interaction domain and evidence for expression in testis.\nReprod Biol Endocrinol\n1\n53\n12932298\n18. JinJLO'DohertyAMWangSZhengHSandersKM\n2005\nCatsper3 and catsper4 encode two cation channel-like proteins exclusively expressed in the testis.\nBiol Reprod\n73\n1235\n1242\n16107607\n19. ClaphamDEGarbersDL\n2005\nInternational Union of Pharmacology. L. Nomenclature and structure-function relationships of CatSper and two-pore channels.\nPharmacol Rev\n57\n451\n454\n16382101\n20. CaiX\n2008\nSubunit stoichiometry and channel pore structure of ion channels: all for one, or one for one?\nJ Physiol\n586\n925\n926\n18079155\n21. QuillTASugdenSARossiKLDoolittleLKHammerRE\n2003\nHyperactivated sperm motility driven by CatSper2 is required for fertilization.\nProc Natl Acad Sci U S A\n100\n14869\n14874\n14657366\n22. CarlsonAEQuillTAWestenbroekRESchuhSMHilleB\n2005\nIdentical phenotypes of CatSper1 and CatSper2 null sperm.\nJ Biol Chem\n280\n32238\n32244\n16036917\n23. JinJJinNZhengHRoSTafollaD\n2007\nCatsper3 and Catsper4 are essential for sperm hyperactivated motility and male fertility in the mouse.\nBiol Reprod\n77\n37\n44\n17344468\n24. LiuJXiaJChoKHClaphamDERenD\n2007\nCatSperbeta, a novel transmembrane protein in the CatSper channel complex.\nJ Biol Chem\n282\n18945\n18952\n17478420\n25. SwansonWJVacquierVD\n2002\nThe rapid evolution of reproductive proteins.\nNat Rev Genet\n3\n137\n144\n11836507\n26. TorgersonDGKulathinalRJSinghRS\n2002\nMammalian sperm proteins are rapidly evolving: evidence of positive selection in functionally diverse genes.\nMol Biol Evol\n19\n1973\n1980\n12411606\n27. PodlahaOWebbDMTuckerPKZhangJ\n2005\nPositive selection for indel substitutions in the rodent sperm protein catsper1.\nMol Biol Evol\n22\n1845\n1852\n15930155\n28. HilleB\n2001\nIonic Channels of Excitable Membranes\nSunderland, MA\nSinauer Associates Inc\n814\n29. CaiXLyttonJ\n2004\nThe cation/Ca2+ exchanger superfamily: phylogenetic analysis and structural implications.\nMol Biol Evol\n21\n1692\n1703\n15163769\n30. YuFHYarov-YarovoyVGutmanGACatterallWA\n2005\nOverview of molecular relationships in the voltage-gated ion channel superfamily.\nPharmacol Rev\n57\n387\n395\n16382097\n31. CaiX\n2007\nMolecular evolution and structural analysis of the Ca2+ release-activated Ca2+ channel subunit, orai.\nJ Mol Biol\n368\n1284\n1291\n17400243\n32. CaiX\n2007\nMolecular evolution and functional divergence of the Ca2+ sensor protein in store-operated Ca2+ entry: stromal interaction molecule.\nPLoS ONE\n2\ne609\n17622354\n33. CaiX\n2008\nUnicellular Ca2+ signaling ‘toolkit’ at the origin of Metazoa.\nMol Biol Evol\n25\n1357\n1361\n18385221\n34. AltschulSFMaddenTLSchafferAAZhangJZhangZ\n1997\nGapped BLAST and PSI-BLAST: a new generation of protein database search programs.\nNucleic Acids Res\n25\n3389\n3402\n9254694\n35. NicholsSADirksWPearseJSKingN\n2006\nEarly evolution of animal cell signaling and adhesion genes.\nProc Natl Acad Sci U S A\n103\n12451\n12456\n16891419\n36. KoishiRXuHRenDNavarroBSpillerBW\n2004\nA superfamily of voltage-gated sodium channels in bacteria.\nJ Biol Chem\n279\n9532\n9538\n14665618\n37. CaiXZhangY\n2006\nMolecular evolution of the ankyrin gene family.\nMol Biol Evol\n23\n550\n558\n16280541\n38. SuyamaMTorrentsDBorkP\n2006\nPAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments.\nNucleic Acids Res\n34\nW609\n612\n16845082\n39. YangZ\n2007\nPAML 4: phylogenetic analysis by maximum likelihood.\nMol Biol Evol\n24\n1586\n1591\n17483113\n40. OkamuraYNishinoAMurataYNakajoKIwasakiH\n2005\nComprehensive analysis of the ascidian genome reveals novel insights into the molecular evolution of ion channel genes.\nPhysiol Genomics\n22\n269\n282\n15914577\n41. DelsucFBrinkmannHChourroutDPhilippeH\n2006\nTunicates and not cephalochordates are the closest living relatives of vertebrates.\nNature\n439\n965\n968\n16495997\n42. PutnamNHButtsTFerrierDEFurlongRFHellstenU\n2008\nThe amphioxus genome and the evolution of the chordate karyotype.\nNature\n453\n1064\n1071\n18563158\n43. KingNWestbrookMJYoungSLKuoAAbedinM\n2008\nThe genome of the choanoflagellate Monosiga brevicollis and the origins of metazoan multicellularity.\nNature\n451\n783\n788\n18273011\n44. BerridgeMJBootmanMDRoderickHL\n2003\nCalcium signalling: dynamics, homeostasis and remodelling.\nNat Rev Mol Cell Biol\n4\n517\n529\n12838335\n45. PutnamNHSrivastavaMHellstenUDirksBChapmanJ\n2007\nSea anemone genome reveals ancestral eumetazoan gene repertoire and genomic organization.\nScience\n317\n86\n94\n17615350\n46. EllinorPTYangJSatherWAZhangJFTsienRW\n1995\nCa2+ channel selectivity at a single locus for high-affinity Ca2+ interactions.\nNeuron\n15\n1121\n1132\n7576655\n47. JacksonHAMarshallCRAcciliEA\n2007\nEvolution and structural diversification of hyperpolarization-activated cyclic nucleotide-gated channel genes.\nPhysiol Genomics\n29\n231\n245\n17227887\n48. OnCMarshallCRChenNMoyesCDTibbitsGF\n2008\nGene structure evolution of the Na+-Ca2+ exchanger (NCX) family.\nBMC Evol Biol\n8\n127\n18447948\n49. TrevinoCLSerranoCJBeltranCFelixRDarszonA\n2001\nIdentification of mouse trp homologs and lipid rafts from spermatogenic cells and sperm.\nFEBS Lett\n509\n119\n125\n11734218\n50. VenkatachalamKMontellC\n2007\nTRP channels.\nAnnu Rev Biochem\n76\n387\n417\n17579562\n51. LawsonCDorvalVGoupilSLeclercP\n2007\nIdentification and localisation of SERCA 2 isoforms in mammalian sperm.\nMol Hum Reprod\n13\n307\n316\n17376796\n52. Vazquez-MartinezOCanedo-MerinoRDiaz-MunozMRiesgo-EscovarJR\n2003\nBiochemical characterization, distribution and phylogenetic analysis of Drosophila melanogaster ryanodine and IP3 receptors, and thapsigargin-sensitive Ca2+ ATPase.\nJ Cell Sci\n116\n2483\n2494\n12766186\n53. KrylovDMWolfYIRogozinIBKooninEV\n2003\nGene loss, protein sequence divergence, gene dispensability, expression level, and interactivity are correlated in eukaryotic evolution.\nGenome Res\n13\n2229\n2235\n14525925\n54. SalzbergSLWhiteOPetersonJEisenJA\n2001\nMicrobial genes in the human genome: lateral transfer or gene loss?\nScience\n292\n1903\n1906\n11358996\n55. ItoMXuHGuffantiAAWeiYZviL\n2004\nThe voltage-gated Na+ channel NaVBP has a role in motility, chemotaxis, and pH homeostasis of an alkaliphilic Bacillus.\nProc Natl Acad Sci U S A\n101\n10566\n10571\n15243157\n56. AnderssonJODoolittleWFNesboCL\n2001\nGenomics. Are there bugs in our genome?\nScience\n292\n1848\n1850\n11358998\n57. NavarroBKirichokYClaphamDE\n2007\nKSper, a pH-sensitive K+ current that controls sperm membrane potential.\nProc Natl Acad Sci U S A\n104\n7688\n7692\n17460039\n58. GalindoBEde la Vega-BeltranJLLabarcaPVacquierVDDarszonA\n2007\nSp-tetraKCNG: A novel cyclic nucleotide gated K+ channel.\nBiochem Biophys Res Commun\n354\n668\n675\n17254550\n59. HofmannFBielMKauppUB\n2005\nInternational Union of Pharmacology. LI. Nomenclature and structure-function relationships of cyclic nucleotide-regulated channels.\nPharmacol Rev\n57\n455\n462\n16382102\n60. WangDKingSMQuillTADoolittleLKGarbersDL\n2003\nA new sperm-specific Na+/H+ exchanger required for sperm motility and fertility.\nNat Cell Biol\n5\n1117\n1122\n14634667\n61. RenDNavarroBXuHYueLShiQ\n2001\nA prokaryotic voltage-gated sodium channel.\nScience\n294\n2372\n2375\n11743207\n62. PodlahaOZhangJ\n2003\nPositive selection on protein-length in the evolution of a primate sperm ion channel.\nProc Natl Acad Sci U S A\n100\n12241\n12246\n14523237\n63. LiWH\n1997\nMolecular Evolution\nSunderland, Massachusetts\nSinauer Associates\n64. ZhangLLiWH\n2004\nMammalian housekeeping genes evolve more slowly than tissue-specific genes.\nMol Biol Evol\n21\n236\n239\n14595094\n65. ChamaryJVParmleyJLHurstLD\n2006\nHearing silence: non-neutral evolution at synonymous sites in mammals.\nNat Rev Genet\n7\n98\n108\n16418745\n66. PaganiFRaponiMBaralleFE\n2005\nSynonymous mutations in CFTR exon 12 affect splicing and are not neutral in evolution.\nProc Natl Acad Sci U S A\n102\n6368\n6372\n15840711\n67. TorresYPMoreraFJCarvachoILatorreR\n2007\nA marriage of convenience: beta-subunits and voltage-dependent K+ channels.\nJ Biol Chem\n282\n24485\n24489\n17606609\n68. WangYGuX\n2001\nFunctional divergence in the caspase gene family and altered functional constraints: statistical analysis and prediction.\nGenetics\n158\n1311\n1320\n11454777\n69. SchmidtHAStrimmerKVingronMvon HaeselerA\n2002\nTREE-PUZZLE: maximum likelihood phylegenetic analysis using quartet and parallel computing.\nBioinformatics\n18\n502\n504\n11934758"
+}

batch_13/PMC2576179.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2576179",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2576179\nAUTHORS: Fumiyuki Goto, Kimiko Nakai, Takanobu Kunihiro, Kaoru Ogawa\n\nABSTRACT:\nBackgroundPatients suffering from dizziness due to vertigo are commonly encountered in the department of otolaryngology. If various clinical examinations do not reveal any objective findings, then the patients are referred to the department of internal medicine or psychiatry. In many cases, the diagnosis is psychological dizziness. Phobic postural vertigo, which was first reported by Brandt T et al in 1994, is supposed to be a type of psychological dizziness. The diagnosis is based on 6 characteristics proposed by Brandt et al. Patients are usually treated with conventional medical therapy, but some cases may be refractory to such a therapy. Psychotherapy is recommended in some cases; however, psychotherapy including autogenic training, which can be used for general relaxation, is not widely accepted. This paper describes the successful administration of autogenic training in a patient suffering from phobic postural vertigo.Case presentationWe present a case of a patient who suffered from phobic postural vertigo. A 37-year-old female complained of dizziness. She had started experiencing dizziness almost 3 years She was intractable to many sort of conventional therapy. In the end, her symptom disappeared after introduction of autogenic training.ConclusionAutogenic training can be a viable and acceptable treatment option for phobic postural vertigo patients who fail to respond to other therapies. This case emphasizes the importance of autogenic training as a method to control symptom of phobic postural vertigo.\n\nBODY:\nBackgroundPatients suffering from dizziness are commonly encountered at the otologist in the department of otorhinolaryngology. In the absence of any organic abnormality, such patients are diagnosed with psychogenic dizziness and are often referred to the department of psychosomatic medicine or to a psychologist, because dizziness and vertigo are complaints common to various psychiatric conditions like major depression, somatoform disorder, and anxiety disorders. A representative etiology is phobic postural vertigo (PPV), which was first proposed by Brandt T et al in 1994 [1]. The diagnosis of PPV is based on 6 characteristics proposed by this group. The key features for correct diagnosis are spontaneous (sometimes stimulus-induced) postural vertigo and unsteadiness in maintaining an upright position and walking; the correct diagnosis is not based on anxiety but on subjective dizziness itself. The correct diagnosis of PPV is important for better prognosis. These patients often have an obsessive personality. Psychotherapy including autogenic training (AT) and cognitive behavior therapy (CBT), which can be used for general relaxation and to treat disturbed emotions, is a good treatment option. However, there are no reports on the application of AT to patients with PPV. The present paper describes the successful administration of AT to a patient suffering from PPV intractable to several conventional therapies. Written informed consent was obtained from the patient for the presentation of this case.Case presentationA 37-year-old female complained of dizziness. She had started experiencing dizziness 3 years ago, following the infertility treatment that she had received. She experienced dizziness following an injection of human menopausal gonadotropin administered by a gynecologist and a visit psychologist. In addition to her dizziness, she also suffered from insomnia, tinnitus, and anxiety. Therefore, she was referred to a psychologist. However, the treatment of tranquilizers such as benzodiazepines and antidepressants such as serotonin selective re-uptake inhibitors (SSRIs) failed to cure the dizziness and only slightly improved her insomnia. She was therefore referred to our department for further examination and treatment. She expressed her dizziness as an event wherein she experienced frequent paroxysmal earthquakes occurring within seconds. The frequency of such episodes had recently increased to once every 5 minutes. She felt stable while doing her household tasks and she had never fallen. Audio-vestibular examination, including pure tone audiometry, posturography, and head MRI, revealed no abnormal findings. Her blood examination findings were normal; there was no spontaneous or evoked nystagmus. However, the peripheral part of her hand and foot often became pale due to poor peripheral circulation, an observation similar to Raynaud's phenomenon. She also experienced chronic headache and insomnia, whereby she woke up every 2 hours during the night. The results of the psychological examination were as follows: Self-rating Depression Scale (SDS), 47; Japanese version of the Cornell Medical Index (CMI), III; Manifest Anxiety Scale (MAS), 27; and Maudsley Obsessional-Compulsive Inventory (MOCI [2]), 9. MAS indicated a high level of anxiety. We deduced that her dizziness was due to psychosomatic reasons together with poor peripheral circulation. We prescribed setiptiline maleate and an additional herbal medicine, which is known to improve peripheral circulation. Within 2 weeks her symptoms slightly improved and the level of dizziness reduced to less than one third. However she didn't want to keep taking these drugs, since she want to have a baby. No abnormality was reported in any physical examination, including posturography. We diagnosed the patient's condition as phobic postural vertigo. Although she often experienced palpitations, cardiological examination reported no abnormal findings. These results indicate the existence of autonomic dysfunction due to psychological stress, including anxiety. We decided to focus on treating the patient's anxiety and the supposed autonomic dysfunction. After 1 month following the patient's first visit, AT was introduced by a clinical psychologist so as to ease her mental stress. The psychotherapy consisted of one 45-minute session every 3 weeks. The first session began with a brief introduction to the general background information about the cognitive approach, after which the patient was instructed how to perform AT. Thereafter, the patient performed AT in a relaxed sitting position on a chair for 10 minutes 3 times a day. No self-monitoring was advised. The patient was instructed to carry out slow and deep abdominal breathing at the beginning of AT and regular breathing during AT. She diligently and regularly continued this AT routine 3 times a day at her home, according to a written timetable. She learned all 6 standard formulas of AT in 6 psychotherapy sessions. Astonishingly, after the introduction of AT, her mood stabilized and her dizziness, insomnia, and headache disappeared in a few weeks. The dose of clotiazepam was reduced to 5 mg once a day. No additional treatment was administered. At 6- and 9-month follow-ups, the patient was free from dizziness, insomnia, and headache.ConclusionThe diagnosis of PPV is based on 6 characteristics proposed by Brandt et al. The diagnostic criteria [1] are as follows: (1) dizziness and subjective disturbance of balance while standing or walking despite normal clinical balance tests; (2) fluctuating unsteadiness for seconds to minutes, or momentary perceptions of illusory body perturbations; (3) usually a perceptual stimulus or social situation as a provoking factor with a tendency for rapid conditioning, generalization, and avoidance behavior; (4) anxiety and vegetative symptoms during or after vertigo; (5) obsessive-compulsive personality type, labile affect, or mild depression; and (6) onset frequently after a period of emotional stress, serious illness, or a vestibular disorder. Patients sometimes exhibit anxiety reactions. The different possible treatments for PPV include pharmacological treatment and psychotherapy. In this case, we clearly showed that patients with PPV that is intractable to many types of drugs can be alleviated by AT, which is a type of psychotherapy.AT was developed by the German psychiatrist Johannes Schultz and can be achieved by daily self-training sessions of 10 to 15 minutes [3]. AT is a technique for influencing one's autonomic nervous system and it can be used to alleviate many stress-induced psychosomatic disorders. Schultz emphasized parallelism between AT, yoga and meditation. AT has been widely applied as a relaxation technique and has been viewed as a highly effective method for controlling pain and reducing drug dependence substantially [4].We used a psychological approach to treat this patient because her symptoms were closely related to her anxiety and autogenic dysfunction and were refractory to conventional therapy. We speculate a vicious circle of PPV and the effect of AT, as shown in Fig. 1. We now assume patients who have PPV with the following characteristics to be candidates for psychotherapy: (1) patients with dizziness, which is intractable to conventional therapy; (2) patients with insomnia; (3) patients with a high anxiety level; (4) patients that have a variety of complaints in addition to dizziness; and (5) patients in whom vertigo is believed to be triggered by stress. However, AT is not recommended for patients with low-level anxiety, those with little motivation, or those who lack the intellectual capacity to understand and perform AT [5,6]. AT can be a viable and acceptable treatment option for a patient with PPV refractory to other therapies.Figure 1The vicious circle of phobic postural vertigo (PPV) and the effect of autogenic training.Authors' contributionsAll authors read and approved the final manuscript. FG and KN participated in the treatment of the patient and drafted the manuscript. TK and KO provided instructions and advice on the treatment strategy.ConsentWritten informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2577183.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2577183",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2577183\nAUTHORS: Sophie E. L. Chamberlain, Jian Yang, Roland S. G. Jones\n\nABSTRACT:\nWe have previously shown that spontaneous release of glutamate in the entorhinal cortex (EC) is tonically facilitated via activation of presynaptic NMDA receptors\n(NMDAr) containing the NR2B subunit. Here we show that the same receptors mediate short-term plasticity manifested by frequency-dependent facilitation of evoked glutamate release at these synapses. Whole-cell patch-clamp recordings were made from layer V pyramidal neurones in rat EC slices. Evoked excitatory postsynaptic currents showed strong facilitation at relatively low frequencies (3 Hz) of activation. Facilitation was abolished by an NR2B-selective blocker (Ro 25-6981), but unaffected by NR2A-selective antagonists (Zn2+, NVP-AAM077). In contrast, postsynaptic NMDAr-mediated responses could be reduced by subunit-selective concentrations of all three antagonists. The data suggest that NMDAr involved in presynaptic plasticity in layer V are exclusively NR1/NR2B diheteromers, whilst postsynaptically they are probably a mixture of NR1/NR2A, NR1/NR2B diheteromers and NR1/NR2A/NR2B triheteromeric receptors.\n\nBODY:\n1. INTRODUCTIONA huge amount of research has been devoted to the study of the\nphysiology, pharmacology, function, and pathology of NMDA receptors (NMDAr). This has been extensively reviewed elsewhere (e.g., [1–6]). Native NMDAr\nare heteromeric structures, and consist of NR1 subunits, which are obligatory,\nin combination with one or more of four subtypes of NR2 subunit (NR2A-D).\nFunctional receptors are tetramers, comprising two NR1 subunits and two NR2\nsubunits, where the functional unit is probably an NR1/NR2 heterodimer. The\nfunctional properties of NMDAr, such as single channel conductance, the degree\nof voltage-dependent Mg2+ block, and deactivation kinetics depend on\nwhich of the four NR2 subunits is assembled in the receptor. For example, NR2A\nand NR2B-containing channels have a high single channel conductance (40–50 pS) whereas\nNR2C and NR2D are lower (15–35 pS). NR2A-containing\nreceptors display fast decay kinetics (around 100 milliseconds), whereas NR2B\nand C are much slower (250 milliseconds), and NR2D slower still (4 seconds) [5, 7]. In addition to functional differences, various subunit combinations display\npharmacological differences in susceptibility to antagonists and regulatory\nmechanisms (such as sensitivity to H+, Zn2+, polyamines).Synaptic transmission is a highly dynamic and plastic process,\nmodified on-demand by a myriad of instantaneous, short, intermediate, and\nlong-term regulatory mechanisms. Much attention has been devoted to the study\nof the role of NMDAr in synaptic plasticity, particularly in long-term\npotentiation (LTP) and depression (LTD). These studies have largely focussed on\nNMDAr at postsynaptic sites. However, dynamic regulation of synaptic strength\ncan also involve receptors on presynaptic terminals, which provide a powerful,\nsynapse-delimited control of transmitter release, and the existence of\npresynaptic NMDAr (preNMDAr) is now firmly established. Neurochemical [8–11] and\nimmunolocalization studies [12–15] provided\nearly indications for preNMDAr. We\nprovided the first clear functional demonstration of preNMDAr, showing that the\ncompetitive antagonist, 2-AP5, could reduce the frequency of spontaneous\nexcitatory postsynaptic currents (sEPSCs) at glutamate synapse in the rat\nentorhinal cortex (EC), indicating a tonic facilitatory effect of preNMDAr on\nglutamate release [16]. PreNMDAr are now known to modify both glutamate and\nGABA release in a wide variety of locations and tissues [17–33].Increasing attention is being paid to the role of preNMDAr as\nmediators of both long-term alterations in synaptic strength, and in moment-to-moment\nand short-term activity-dependent changes in transmitter release. For example,\na role of preNMDAr in LTD has been demonstrated in cerebellum [34], visual [22, 33], and somatosensory [17] cortex. Conversely, involvement of preNMDAr in LTP\nhas been demonstrated in amygdala [26, 32]. More intermediate forms of\npotentiation of glutamate [30] and\nGABA transmission [23], over a time scale of minutes, may\nalso involve preNMDAr. As noted above, we found that preNMDAr are tonically\nactivated by ambient glutamate [17, 35], providing instantaneous control over\nthe level of glutamate release at EC synapses. Similar results have been\nreported for other areas [22, 27, 28, 33]. In addition, we found that preNMDAr are activated after action potential-driven synaptic\nrelease of glutamate, increasing the probability of subsequent release and\nallowing them to mediate short-term, frequency-dependent facilitation of\nglutamate transmission [16, 35].We have also demonstrated that the tonic\nfacilitatory effect of preNMDAr on spontaneous glutamate release is likely to\nbe predominantly mediated by NR2B-containing NMDAr, since the increase induced\nby 2-AP5 was mimicked [35, 36] by relatively specific blockers of the NR2B\nsubunit, \nifenprodil [37], and Ro 25-6981 [38]. In addition, an\nantagonist with some specificity (albeit weak) for the NR2A subunits,\nNVP-AAM077 [39] had little effect. Others have also concluded that preNMDAr are\nlikely to be predominantly NR2B-containing [27, 33, 40]. Postsynaptically, both\nNR2A and NR2B contribute to glutamate transmission, although there is\ncontroversy over whether diheteromeric NR1/NR2A and NR1/NR2B coexist at the\npostsynaptic density, or are segregated between synaptic and extrasynaptic\nlocations, or even in a synapse-specific way [3]. The contribution of\ntriheteromeric NR1/NR2A/NR2B receptors is also still a matter of debate [3, 41].In the present study, we have extended our\nstudies in the EC to examine the contribution of NR2A and NR2B receptors to\nshort-term plasticity of glutamate transmission, by examining the effects of\nrelatively specific blockers on the preNMDAr mediated, frequency-dependent\nfacilitation of evoked glutamate release. In addition, we have used the same\nagents to determine whether postsynaptic NMDAr may differ from those on\npresynaptic terminals.2. METHODS2.1. Slice preparationExperiments were performed in accordance with the U.K. Animals\n(Scientific Procedures) Act 1986, European Communities Council Directive 1986\n(86/609/EEC), and the University of Bath ethical review document. Slices\ncontaining EC and hippocampus were prepared from male \nWistar rats\n(P28–35), which were\nanaesthetized with an intramuscular injection of ketamine (120 mg/kg) plus\nxylazine (8 mg/kg) and decapitated. The brain was rapidly removed and immersed\nin oxygenated artificial cerebrospinal fluid (aCSF) chilled to 4°C. Slices (350–400 μm) were cut\nusing a Vibroslice, and stored in aCSF bubbled with 95% O2/5% CO2,\nat room temperature. Following recovery for at least 1 hour, individual slices\nwere transferred to a recording chamber mounted on the stage of a Zeiss\nAxioskop FS or an Olympus BX50WI microscope. The chamber was perfused (2.0 ml/min) with oxygenated aCSF (pH 7.4) at 31–33°C. The aCSF\ncontained (in mM) NaCl (126), KCl (3), NaH2PO4 (1.4),\nNaHCO3 (19), MgSO4 (2), CaCl2 (2), and D-glucose (10). Neurones were visualized\nusing differential interference contrast optics and an infrared video camera.2.2. Electrophysiological recordingPatch pipettes were pulled from borosilicate\nglass on a Flaming/Brown microelectrode puller. For recording spontaneous\n(sEPSCs) or evoked (eEPSCs) excitatory postsynaptic currents, pipettes were\nfilled with a Cs-gluconate-based solution containing (in mM) D-Gluconate (100),\nHEPES (40), QX-314 (1), EGTA (0.6), NaCl (2), MgCl2 (5), TEA-Cl (1),\nphosphocreatinine (5); ATP-Na (4), GTP-Na (0.3), MK-801 (2). Solutions were\nadjusted to 290 mOsmol, and to pH 7.3 with CsOH. Whole-cell\nvoltage clamp recordings (holding potential −60 mV unless otherwise stated)\nwere made from neurones in layer V of the medial division of the EC, using an Axopatch 200B amplifier (Molecular Devices, Calif., USA). Series resistance compensation was not employed, but access resistance (10–30 MΩ) was\nmonitored at regular intervals throughout each recording and cells were\ndiscarded from analysis if it changed by more than ±10%. Liquid junction\npotential (12.3 mV) was estimated using the Junction Potential Calculator\nincluded in pClamp-8 software (Molecular Devices, Calif., USA),\nand compensated for in the holding potentials.eEPSCs were elicited by electrical stimulation (bipolar pulses, 10–50 V, 0.02 millisecond\nduration) via a bipolar tungsten electrode placed on the surface of the slice\nin layer V of the lateral EC. The stimulation intensity was adjusted to give\nsubmaximal (approx. 50–60% maximum\namplitude) responses.2.3. Monitoring presynaptic NMDAr activityIn all these experiments, MK-801 (2 mM) was included in the patch\npipette solution to block postsynaptic NMDAr. This allowed us to record\nAMPA-receptor mediated responses in isolation, and to monitor activity at\npreNMDAr uncontaminated by postsynaptic receptor effects. This approach was\ndeveloped by us [16, 35, 42], and has been used successfully\nby others to block postsynaptic NMDAr in the recorded neurone [17, 27, 28, 32, 33, 40]. When whole-cell access was gained, neurones were voltage clamped at 0 mV,\nand synaptic stimulation was delivered at 2 Hz for 30–40 seconds to\nallow blockade of postsynaptic NMDAr by MK-801 dialyzed into the cell via the\npatch pipette solution. Membrane potential was then clamped at −60 mV and\nsingle shock stimulation delivered at low frequency (0.05 Hz) to evoke AMPAr\nmediated EPSCs. At 2 or 3 minute intervals, the single shock was replaced with\nstimulation at 3 Hz for 10 seconds. Such stimulation results in a\nfrequency-dependent facilitation of the AMPAr-mediated EPSC, which we have\nshown previously to be dependent on activation of preNMDAr [35]. We used the\ndegree of frequency-dependent facilitation of AMPAr-mediated eEPSCs as a\nquantitative measure of preNMDAr activation.2.4. Monitoring postsynaptic NMDAr activityIn these experiments, MK-801 was omitted from\nthe patch pipette solution. When whole-cell access was gained, control eEPSCs\nwere recorded at a holding potential of −60 mV, before addition of the AMPAr antagonist, NBQX,\nand the GABAAr-antagonist, bicuculline to the bath perfusion. After\n10–12 minutes, the\nholding potential was changed to +40 mV to record isolated NMDAr-mediated EPSCs\nas positive going currents. These were evoked at low frequency (0.05 Hz) until\nstable amplitudes were recorded, before addition of antagonists to the bath.2.5. Data analysisData were recorded to computer hard disk using Axoscope software. Minianalysis (Synaptosoft, Decatur, Ga, USA) was used for\nanalysis of EPSCs offline. In the studies of preNMDAr, the average peak\namplitude of the 8 responses before each episode of 3 Hz stimulation was\ndetermined. During the period of 3 Hz stimulation, the amplitude of the 8\nlargest events was determined and normalized to the average amplitude of the\npreceding low-frequency events to obtain a quantitative measure of frequency-dependent\nfacilitation in the presence and absence of antagonists. In these studies, we\nalso analyzed AMPAr-mediated sEPSCs, by determining interevent interval (IEI),\namplitude, rise (10–90%), and decay\ntimes. sEPSCs were detected automatically using a\nthreshold-crossing algorithm. Threshold varied from neurone to neurone but was\nalways maintained at a constant level in any given recording. At least 200\nevents were sampled during a continuous recording period for each neurone under\neach condition. Cumulative probability distributions of IEI were compared using\nthe Kolmogorov-Smirnoff test. In\nexperiments on postsynaptic NMDAr, responses were quantified by measuring mean\npeak amplitudes of at least 5 NMDAr-mediated eEPSCs evoked at low frequency at\nintervals throughout the study. In these studies, the vast majority of sEPSCs\nwere blocked, as recordings were conducted in the presence of NBQX. Occasional\nslow sEPSCs mediated by NMDAr were recorded, their frequency was very low (2-3 per minute) and\nprecluded meaningful analysis.2.6. MaterialsSalts used in preparation of aCSF were “Analar” grade and purchased\nfrom Merck/BDH or Fisher Scientific (Dorset, UK). All drugs were applied by bath perfusion. MK-801,\nNMDA, NBQX, D-2-AP5, bicuculline methiodide, and Ro 25-6981 ((αR,βS)-α-(4-hydroxyphenyl)-β-methyl-4-(phenylmethyl)-1-piperidinepropanol\nhydrochloride) were obtained from Tocris (Bristol, UK). TPEN (N,N,N′, N′-Tetrakis-(2-pyridylmethyl)-Ethylenediamine) \nwas obtained from Sigma (UK). UBP302\n((S)-1-(2-amino-2-carboxyethyl)-3-(2-carboxybenzyl)\npyrimidine-2,4-dione) was a kind\ngift from Dr. Dave Jane, University of Bristol, and NVP-AAM077 ((R)-[(S)-1-(4-bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydroquinoxalin-5-yl)-methyl]-phosphonic\nacid) was a gift from Dr. Yve\nAuberson at Novartis (Basel, Switzerland).3. RESULTS3.1. Presynaptic NMDAar\nFigure 1(a) shows eEPSCs evoked in a layer V\nneurone at 3 Hz, with postsynaptic NMDAr blocked by internally dialyzed MK-801.\nThe first 6 responses evoked during a train of 30 at 3 Hz are shown and\ndemonstrate the facilitation seen at this relatively low frequency. As reported\npreviously [35], the facilitation of the AMPAr-mediated eEPSCs was entirely\ndependent on presynaptic NMDAar activation, since it could be abolished by\n2-AP5 (n = 5, Figure 1(b)). Likewise, the NMDAr channel blocker, MK-801, also\nabolished frequency facilitation (n = 10, Figure 1(b)). In some neurones,\nfacilitation was replaced by a weak frequency-dependent depression of eEPSCs in\nthe presence of the blockers. This can be seen as a reduction in mean amplitude\nof eEPSCs in the presence of the blockers (e.g., Figure 1(b)). In a further 5\nneurones, we confirmed the specificity of the effect by testing the effects of\nGluR5 subunit specific antagonist of kainate receptors (UBP 302, 20 μM), since\nwe have recently shown that these receptors mediate a similar short-term facilitation\nof glutamate transmission at 3–5 Hz in layer III\nof the EC (Chamberlain S.E.L and Jones R.S.G. unpublished). UBP 302 had no\neffect on facilitation in layer V (not shown) confirming its dependence on\nNMDAar. Interestingly, 2-AP5 had no effect on frequency facilitation in layer\nIII of the EC (not shown), so although similar short-term plasticity is seen in\nboth layers, its underlying mechanism is lamina-specific.Since neither 2-AP5 nor MK-801 has selectivity\nfor NR2A v NR2B subunits [5], the data do not indicate the subunit composition\nof NMDAr responsible for short-term frequency-facilitation. To determine the\nreceptor involved, we have examined the effect of more specific antagonists.\nFirst, we tested the effects of Ro 25-6981. This is an allosteric inhibitor of\nNMDA receptors, which binds to a site on the N-terminal domain of the NR2\nsubunit, with a high degree of selectivity (>3000 fold) for NR2B over NR2A\n[38]. Figure 2(a) shows that Ro 25-6981 at 500 nM abolished the frequency\nfacilitation of eEPSCs, again revealing a weak depression. A lower\nconcentration (200 nM, n = 3) of Ro 25-6981 resulted in a mean maximal reduction\nin frequency-facilitation of 69 ± 7%. At these concentrations, the drug should\nhave little or no effect on NR2A subunits [38], strongly suggesting that\nNR2B-containing receptors are primarily responsible for this form of short-term\nplasticity at layer V synapses. This would agree with previous studies that\nhave shown the tonic facilitatory effect on spontaneous release is likely to be\nNR2B-mediated [16, 35, 43]. Accordingly, Ro 25-6981 resulted in a substantial\nincrease in IEI of sEPSCs from 277 ± 82 milliseconds (5.5 ± 1.9 Hz) to 764 ± 261 milliseconds\n(2.1 ± 0.7 Hz) recorded in the same neurones (cf. [36, 43]). KS analysis of cumulative probability distributions\nconfirmed a highly significant change. There was no concurrent change in mean\namplitude, rise, or decay time (not shown).Next, we examined the effect of NVP-AAM077 in 5\nneurones. This is a competitive antagonist that shows some selectivity for\nreceptors containing the NR2A subtype. Initial reports indicated a greater than\n100 fold selectivity of the compound for NR2A over NR2B [39, 44]. However,\nrecently, it has been suggested that the selectivity is closer to 10 fold when\nthe affinity of the two subtypes for glutamate is accounted for ([41], see also\n[45, 46]). Thus, at the concentration employed here (400 nM), we might expect\nalmost complete blockade of NR2A receptors, but it is possible that substantial\ninhibition of NR2B would also occur [41]. Nevertheless, NVP-AAM077 did not significantly affect the frequency-dependent facilitation of eEPSCs (see Figure 2(b)). If anything, the facilitation was slightly (although not significantly) increased. These data suggest that NVP-AAM077 may have\nreasonable selectivity for the NR2A receptor in our preparation, but that these\nreceptors are not involved in presynaptic short-term plasticity at layer V\nsynapses. Further support for this was obtained from analysis of sEPSCs. The\nmean IEI in control was 443 ± 230 milliseconds (4.0 ± 0.9 Hz), and this decreased\nslightly to 377 ± 180 milliseconds (4.5 Hz) with the addition of NVP-AAM077.\nLikewise, there was no change in amplitude, rise, or decay times of sEPSCs (not\nshown).In view of the controversy over the selectivity\nof NVP-AAM077, we also tested (n = 5) the effects of Zn2+, which has\nbeen shown to discriminate between NR2A and NR2B receptors. Like Ro 25-6981\nat NR2B subunits, Zn2+ binds to the N-terminal domain of the NR2A\nsubunit to exert a voltage-independent inhibition with >100 fold selectivity\nover NR2B [47–49]. However, as\nwith NVP-AAM077, a relatively high concentration of Zn2+ (300 nM)\nfailed to alter frequency-dependent facilitation of eEPSCs (see Figure 2(c)).\nIn addition, it had little effect on the IEI (200 ± 150 v 298 ± 170 milliseconds, see Figure 2(d)), amplitude (17.7 ± 3.4 v 15.4 ± 2.2 pA), rise (1.9 ± 0.3 v 2.1 ± 0.4 milliseconds), or decay times\n(24.6 ± 1.6 v 27.3 ± 1.3 milliseconds) of\nsEPSCs (cf. [43]). Thus, the\ndata from both NVP-AAM077 and Zn2+ studies militate strongly against\na role for NR2A receptors in presynaptic frequency-dependent facilitation in\nlayer V of the EC. The ability of Ro 25-6981 to block facilitation strongly\nindicates that presynaptic plasticity at these synapses is dependent only on\nNR2B-containing receptors.A recent paper [50] suggested that activation\nof postsynaptic NR2B-containing receptors at a similar frequency (3.3 Hz) to\nthat employed by us to elicit frequency-dependent facilitation induced a\nlong-term depression of the NMDAr-mediated currents themselves (primarily by\ndecreasing fractional Ca2+ currents carried by the receptors). We\nwere interested to see if the repetitive activation of the presynaptic\nNR2B-containing receptors would induce any decrement in frequency facilitation\nat layer V synapses. In 5 neurones, we induced facilitation of eEPSCs and\nmonitored the degree of facilitation but without the addition of any blockers. Overall\nthere was an initial decrease in the degree of facilitation of AMPAr-mediated\neEPSCs from the first to second episode, but thereafter it was remarkably\nconsistent (see Figure 3(a)). However, when we looked at absolute amplitude of\neEPSCs, there was a small, but consistent, increase over the course of the\nstudies. This applied to events evoked at both low and high frequencies (see Figure\n3(b)). We also examined the time course of these changes in the neurones tested\nwith 2-AP5 (see Figure 4). The antagonist appeared to prevent the progressive\nincrease in amplitude of the low-frequency events at the same time as blocking\nthe frequency-dependent facilitation. This limited protocol may suggest the\nshort-term frequency-dependent facilitation could underlie a longer-term\nenhancement of glutamate transmission. As the postsynaptic NMDAr were already\nblocked (by internal MK-801), this is likely to involve the presynaptic,\nNR2B-containing receptors.3.2. Postsynaptic NMDArWe now wished to determine the contribution of\nNR2A/B subunits to NMDAr at postsynaptic sites in layer V of the EC, so we\ntested the same antagonists used in the presynaptic experiments for effects on\nisolated NMDAr-mediated eEPSCs. As expected, the nonspecific blockers 2-AP5\n(n = 5) and MK-801 (n = 9) both abolished the slow eEPSCs recorded at +40 mV in the\npresence of NBQX and bicuculline (not shown). Ro 25-6981 (n = 5) also elicited a\nconcentration dependent reduction in postsynaptic NMDAr responses at\nconcentrations that would be expected to retain selectivity for NR2B-containing\nreceptors (see Figure 5(a)). The slow eEPSCs were essentially abolished by\nRo 25-6981 at 500 nM. This suggests that NR1/NR2B receptors dominate at\npostsynaptic sites as they do presynaptically. However, when we tested\nNVP-AAM077 (n = 6), we again found a concentration-related reduction in\npostsynaptic responses with around 80% inhibition at 500 nM (see Figure 5(b)).\nComparison with the data of Neyton and Paoletti [41] suggests that the effect\nof NVP-AAM077 could be explained by blockade of both NR2B and NR2A receptors\nsince 500 nM was sufficient to abolish NR2A responses in oocytes, but also to\nexert around 60% block of NR2B. However, this is at odds with its failure to\nalter preNMDAr-dependent facilitation, which is clearly an NR2B-mediated\nresponse. Studies with Zn2+ (n = 6) failed to\nsubstantially clarify the situation. The divalent cation also elicited a\nconcentration-dependent reduction in slow eEPSCs (see Figure 5(c)). The\nconcentrations employed exert around an 80% voltage-independent block of NR2A\nreceptors expressed in oocytes, but retain a considerable degree of selectivity\nwith regard to block of NR2B receptors [47, 49]. These data do suggest a role\nfor NR2A receptors at postsynaptic sites, but it is puzzling that Ro 25-6981 essentially also abolished NMDAr EPSC, when\nit would be expected to have little effect on NR2A receptors.We performed two more sets of experiments to look at this question\nfurther. In 5 neurones, we first perfused a low concentration of Ro 25-6981 (200 nM), to partially block the\nNMDAr EPSC. We then added a low concentration of Zn2+ (100 nM). In\nthese neurones, Ro 25-6981 resulted in inhibition of around 45%, and with the addition of\nZn2+ there was a further reduction to around 90–100%, which\nclearly indicates a role of both NR2A and NR2B in mediating the postsynaptic\nresponse (see Figure 5(d)). Finally, there is evidence that under control\nconditions, NR2A-containing receptors may be substantially blocked by Zn2+,\npresent in the ACSF as a result of contamination of other salts used in its\npreparation [47]. Although addition of Zn2+ clearly reduced slow\neEPSCs in our experiments, we also examined whether there was significant\nblockade of the NR2A receptor in control\nrecordings by testing the effect of the Zn2+-chelator, TPEN (2 μM),\nin 3 neurones. This had no effect on the mean amplitude of NMDAr eEPSCs (125.3 ± 25.1 v 111.9 ± 26.1 pA) suggesting that our\nresults with antagonists were unlikely to be confounded by Zn2+-contamination.Finally, as noted above, relatively low\nfrequency, repetitive activation of NR2B receptors has been shown to induce a\ndepression of postsynaptic NMDA responses per se [50]. In 7 neurones, we determined the effects of a brief\nperiod of repetitive stimulation (3 Hz, 40 seconds) on postsynaptic NMDAr\neEPSCs in 5 neurones. Overall, during the repetitive stimulation there was a\nsmall (15%), progressive decrease in the first 10–15 seconds, and\nthereafter the amplitude reached a plateau (see Figure 6(a)). We then recorded\nNMDAr eEPSCs at low frequency (0.05 Hz) over the subsequent 30 minutes. There\nwas an initial period (5 minutes) where responses appeared to be slightly\ndepressed and thereafter a recovery followed by a slight increase before recovery\nto control levels (see Figure 6(b)). However, apart from a brief period around\n20 minutes there was no significant difference compared to control.4. DISCUSSIONWe originally demonstrated that the presynaptic\nNMDAar mediating facilitation of glutamate release in the EC was likely to be\npredominantly NR2B-containing, as the frequency of sEPSCs was decreased by the\nN2B antagonist, ifenprodil [35]. Other work supports the conclusion that\npreNMDAr that facilitate spontaneous glutamate release at cortical synapses are\nprimarily NR2B-containing. We found that Ro 25-6981 but not NVP-AAM077 or Zn2+ reduced sEPSC frequency ([36], present study), and similar results with Ro\n25-6981 and Zn2+ were reported for synapses in layer II/III of the\nvisual cortex [28]. Jourdain et al. [27] reported that presynaptic NR2B receptors\nwere responsible for the increase in mEPSC frequency in dentate granule neurones\nseen after stimulation of glutamate release from adjacent astrocytes, as it was\nblocked by ifenprodil. We now show that the same receptor is likely to mediate\nshort-term plasticity of evoked glutamate release in layer V of the EC. Thus,\nthe facilitation of eEPSCs at the relatively low frequency of 3 Hz was blocked by\nRo 25-6981. The lack of effect of NVP-AAM077 and Zn2+ suggests that\nNR2A receptors do not contribute to facilitation of either spontaneous or evoked glutamate\nrelease at EC synapses. We cannot rule out a role of NR2A receptors at higher\nfrequencies, although Sjöström et al. [33] have reported that frequency facilitation at\n30 Hz at layer V synapses in visual cortex is greatly reduced by ifenprodil,\nsuggesting that NR2B dominate at other presynaptic sites as well.It is somewhat surprising that only presynaptic\nNR2B receptors appear to modulate release. Postembedding immunolabeling studies\nhave shown the presence of NR1 subunits in presynaptic terminals in cortex and\nhippocampus [12–14, 51–53]. Whilst a\nhost of studies have demonstrated NR2B subunits at presynaptic locations [15, 51, 54–59], similar\nstudies have also indicated the presence of NR2A subunits [51, 52, 60–62] although, to\ndate, there are no similar studies specifically related to the EC.The presence of all three subunits suggests\nthat both NR1/NR2A and NR1/NR2B diheteromeric receptors and possibly also\nNR1/NR2A/NR2B triheteromers could be expressed in cortical presynaptic\nterminals, and this may well be the case. However, it is clear from the\npharmacological experiments presented here and elsewhere, that NR1/NR2B\nreceptors are predominantly responsible for short-term NMDAr-mediated facilitation\nof glutamate release (but see, [63]). The properties of NR2B subunits differ from\nNR2A, in a way that may make them more suited to the task of presynaptic\nfacilitation (see [6, 7, 64–66]). NR2B\nsubunits have a higher affinity for both glutamate and glycine, and show less\ndesensitization. The two subunits confer similar single channel conductance to\ndiheteromeric receptors (around 50 pS),\nbut they have very different deactivation kinetics, with NR1/NR2A receptors\nhaving decay time constants of 50–100 milliseconds,\nand NR1/NR2B receptors in the order of 200–400 milliseconds.\nBoth are Ca2+-permeable, but NR2B receptors exhibit a higher\nfractional Ca2+-current than NR2A (see [66, 67]). Both subunits also\ndisplay Ca2+-dependent inactivation, but this is more pronounced for\nNR2A. The presence of NR2B subunits results in prolonged EPSPs compared to\nthose seen when NR2A subunits dominate (see [3, 7, 66]). Thus, it seems likely\nthat activation of presynaptic NR2B-containing receptors would mediate a slowly\ndeactivating opening of the NMDAr channel and a greater Ca2+-influx\ninto the presynaptic terminals than any influx mediated by NR2A receptors. Ca2+-influx\nvia the NMDAr is responsible for instantaneous control of spontaneous glutamate\nrelease [35]. With a deactivation time of around 300 milliseconds, repetitive\nactivation of NR1/NR2B receptors would readily result in temporal summation of\npresynaptic Ca2+-entry leading to the short-term facilitation at\neven relatively low-frequency stimulations seen here and previously [35].It is interesting to speculate on a physiological or pathological\nrole for short-term plasticity\nmediated by preNMDAr. State-dependent rhythms and oscillatory activity at\nvarious frequencies occur in the networks of the EC including ripples and sharp\nwaves (>100 Hz), gamma (30–80 Hz), theta (4–8 Hz), and slow\nwaves (0.1–0.5 Hz) [68–71], and these\nmay be involved in mnemonic processing in temporal lobe structures. There is a\nconsensus that theta oscillations are intimately involved in declarative memory\nand spatial navigation (see [72–74]), and it is\npossible that information encoding involved in these processes is reliant on an\nincrease in entorhinal-hippocampal delta/theta coherence [73]. The facilitation\nof glutamate transmission mediated by preNMDAr that we describe is readily\nelicited at frequencies in the low theta range. Thus, we could speculate that\nthese receptors may be involved in the generation of theta activity in the EC,\nand the proposed role of this activity in short-term memory and coding of\nspatial information (e.g., [72, 74]).At a pathological level, it is noteworthy that, oscillations at\ndelta (1-2 Hz) and theta\nfrequency may be associated with epilepsy. In patients with temporal lobe\nepilepsy, there is a generalized increase in EEG activity in the delta/theta\nrange, and the most common pattern of discharges after the initiation of ictal\nevents is a rhythmic delta/theta activity (e.g., [75, 76]). Also, in rats made\nchronically epileptic following kainic acid injection, epileptiform events in\nsuperficial layers of the EC were sometimes followed by spontaneous theta\noscillations in layer V [77]. We recently showed that preNMDAr function\ndeclines in adulthood, but is markedly enhanced in age-matched, chronically\nepileptic rats [36] and there is evidence for a similar increased function in\nhuman temporal lobe epilepsy [78]. We could speculate that this increased\npreNMDAr function could result in enhanced generation of delta/theta activity in epileptic\nconditions. Of further interest in this regard is the observation that\nincreased delta/theta EEG activity (albeit in patients with generalized\nabsence/myoclonic seizures) is normalized by the anticonvulsant drugs, valproate, and lamotrigine\n[79–81]. We have also\nshown that at least one anticonvulsant drug (felbamate) can block the preNMDAr\n[42]. This raises the possibility that some anticonvulsants could alter delta/theta\noscillations by targeting preNMDAr.Whatever the function of short-term plasticity,\nand the involvement of preNMDAr in it, there is increasing evidence that these\nreceptors may also contribute to longer term forms of plasticity, apparently\nmediating both LTD [17, 22, 33, 34] and LTP [26, 32] at\na variety of synapses. In at least one case, LTD appears to be mediated by NR2B-containing\nreceptors [33], so both short- and long-term plasticity of glutamate\ntransmission could involve Ca2+-influx via presynaptic NR2B\nreceptors. We have also shown recently that preNMDAr are rapidly mobile \nand can diffuse between locations near release sites and more distal\nlocations in the terminal membrane [82]. Trafficking of receptors in the\npresynaptic membrane appears to be influenced by ongoing activity levels, and\nexerts an intermediate (over 10 seconds of minutes) form of plasticity. Thus,\npresynaptic NR2B receptors may be heavily involved in both plasticity and\nmetaplasticity at glutamate synapses in EC and other cortical synapses.In the present study, we also present evidence for differences in pre-\nand postsynaptic NMDAr at layer V synapses. Whilst preNMDAr-mediated effects\nare exclusively dependent on NR1/NR2B-containing diheteromers, both NR2B and\nNR2A appear to contribute to postsynaptic responses. However, the relative\ncontributions of the two subunits are not clear. The ability of low\nconcentrations of both Zn2+ and Ro 25-6981 to reduce postsynaptic\nNMDAr responses could suggest that they are dependent on a mix of NR1/NR2A and NR1/NR2B\ndiheteromeric receptors. However, concentrations of either blocker, that should\nlargely retain selectivity at the respective subtypes, were able to almost\nabolish postsynaptic responses. This could suggest that the postsynaptic\nreceptors could be largely triheteromeric NR1/NR2A/NR2B receptors. Although\ntriheteromeric receptors do exhibit high affinity for both NR2A and NR2B\nselective blockers, it seems likely that they exhibit a reduced maximal inhibitory\neffect to either, and that maximal blockade requires occupation of both sites\n[83]. This does not fit well with our finding that combined application of low\nconcentrations of Zn2+ and Ro 25-6981 could also abolish\npostsynaptic responses, which would better support a mediation by a mix of\nNR1/NR2A and NR1/NR2B diheteromeric receptors. It should also be noted that the\nability of NMDA antagonists to block the receptors is not just dependent on the\nNR2 subunit present, but is also modified by which splice variant of the NR1\nsubunit with which it combines [47, 49]. We do not know which NR1 subunit(s)\nmay be present in the EC. Thus, overall it is difficult to define exactly what\nthe postsynaptic receptor population, but the most likely scenario is a mix of\nNR1/NR2A, NR1/NR2B, and NR1/NR2A/NR2B receptors.A number of studies have suggested that NR1/NR2A, NR1/NR2B, and\nNR1/NR2A/NR2B receptors may contribute to postsynaptic responses at other\ncortical synapses [84–86]. There is\nsupport also for synapse-specific segregation of NR2A and NR2B-containing\nreceptors (e.g., [87, 88]) and spatial segregation between subsynaptic and\nextrasynaptic sites (e.g., [86]). The controversy over whether subunit\ncomposition and spatial location are linked, and the difficulties in defining\nthe role of triheteromeric receptors has been well reviewed recently [3]. We\ncannot make any firm conclusions regarding these aspects in the EC, but our\ndata do suggest that postsynaptic NR1/NR2A, NR1/NR2B, and NR1/NR2A/NR2B\nreceptors all contribute to postsynaptic responses at glutamate synapses in\nlayer V of the EC, in contrast to presynaptic sites where NR1/NR2B receptors\nmay have exclusive control. Increasing numbers of studies have documented LTP\nand LTD at synapses in the EC [89–95]. The EC is\nclearly a pivotal site in learning and memory functions resident in the\ntemporal lobe. We have shown that preNMDAr mediate short-term forms of\nplasticity in the EC. In experiments employing a limited protocol of repetitive\nactivation, we found that this short-term plasticity may lead to longer-term\nplasticity (either pre- or postsynaptically), and the aim now is to examine in\ndetail the relationship between short-term effects and long-term plasticity and\nmetaplasticity at these synapses.\n\nREFERENCES:\n1. Cull-CandySGBrickleySFarrantMNMDA receptor subunits: diversity, development and diseaseCurrent Opinion in Neurobiology200111332733511399431\n2. GardoniFDi LucaMNew targets for pharmacological intervention in the glutamatergic synapseEuropean Journal of Pharmacology2006545121016831414\n3. KöhrGNMDA receptor function: subunit composition versus spatial distributionCell and Tissue Research2006326243944616862427\n4. LlansolaMSanchez-PerezACauliOFelipoVModulation of NMDA receptors in the cerebellum. 1. Properties of the NMDA receptor that modulate its functionThe Cerebellum20054315416116147947\n5. PaolettiPNeytonJNMDA receptor subunits: function and pharmacologyCurrent Opinion in Pharmacology200771394717088105\n6. YamakuraTShimojiKSubunit- and site-specific pharmacology of the NMDA receptor channelProgress in Neurobiology199959327929810465381\n7. Cull-CandySGLeszkiewiczDNRole of distinct NMDA receptor subtypes at central synapsesScience's STKE20042004255re16 pages\n8. BustosGAbarcaJForrayMIGyslingKBradberryCWRothRHRegulation of excitatory amino acid release by N-methyl-D-aspartate receptors in rat striatum: in vivo microdialysis studiesBrain Research19925851-21051151355000\n9. FinkKBönischHGöthertMPresynaptic NMDA receptors stimulate noradrenaline release in the cerebral cortexEuropean Journal of Pharmacology199018511151172171957\n10. KrebsMODesceJMKemelMLGlutamatergic control of dopamine release in the rat striatum: evidence for presynaptic N-methyl-D-aspartate receptors on dopaminergic nerve terminalsJournal of Neurochemistry199156181851824785\n11. MartinDBustosGABoweMABraySDNadlerJVAutoreceptor regulation of glutamate and aspartate release from slices of the hippocampal CA1 areaJournal of Neurochemistry1991565164716551672884\n12. AokiCVenkatesanCGoC-GMongJADawsonTMCellular and subcellular localization of NMDA-R1 subunit immunoreactivity in the visual cortex of adult and neonatal ratsThe Journal of Neuroscience1994149520252228083731\n13. DeBiasiSMinelliAMeloneMContiFPresynaptic NMDA receptors in the neocortex are both auto- and heteroreceptorsNeuroReport1996715–17277327768981465\n14. PetraliaRSYokotaniNWentholdRJLight and electron microscope distribution of the NMDA receptor subunit NMDAR1 in the rat nervous system using a selective anti-peptide antibodyThe Journal of Neuroscience19941426676968301357\n15. PetraliaRSWangYXWentholdRJThe NMDA receptor subunits NR2A and NR2B show histological and ultrastructural localization patterns similar to those of NR1The Journal of Neuroscience19941410610261207931566\n16. BerrettaNJonesRSGTonic facilitation of glutamate release by presynaptic N-methyl-D-aspartate autoreceptors in the entorhinal cortexNeuroscience19967523393448931000\n17. BenderVABenderKJBrasierDJFeldmanDETwo coincidence detectors for spike timing-dependent plasticity in somatosensory cortexThe Journal of Neuroscience200626164166417716624937\n18. BreukelAIMBesselsenELopes da SilvaFHGhijsenWEJMA presynaptic N-methyl-D-aspartate autoreceptor in rat hippocampus modulating amino acid release from a cytoplasmic poolEuropean Journal of Neuroscience19981011061149753118\n19. CasadoMDieudonnéSAscherPPresynaptic N-methyl-D-aspartate receptors at the parallel fiber-Purkinje cell synapseProceedings of the National Academy of Sciences of the United States of America20009721115931159711016958\n20. ChenY-HWuM-LFuW-MRegulation of presynaptic NMDA responses by external and intracellular pH changes at developing neuromuscular synapsesThe Journal of Neuroscience1998188298229909526015\n21. CochillaAJAlfordSNMDA receptor-mediated control of presynaptic calcium and neurotransmitter releaseThe Journal of Neuroscience19991911932059870950\n22. CorlewRWangYGhermazienHErisirAPhilpotBDDevelopmental switch in the contribution of presynaptic and postsynaptic NMDA receptors to long-term depressionThe Journal of Neuroscience200727379835984517855598\n23. DuguidICSmartTGRetrograde activation of presynaptic NMDA receptors enhances GABA release at cerebellar interneuron-Purkinje cell synapsesNature Neuroscience20047552553315097992\n24. GlitschMMartyAPresynaptic effects of NMDA in cerebellar Purkinje cells and interneuronsThe Journal of Neuroscience19991925115199880571\n25. HuangHBordeyAGlial glutamate transporters limit spillover activation of presynaptic NMDA receptors and influence synaptic inhibition of Purkinje neuronsThe Journal of Neuroscience200424255659566915215288\n26. HumeauYShabanHBissièreSLüthiAPresynaptic induction of heterosynaptic associative plasticity in the mammalian brainNature2003426696884184514685239\n27. JourdainPBergersenLHBhaukaurallyKGlutamate exocytosis from astrocytes controls synaptic strengthNature Neuroscience200710333133917310248\n28. LiY-HHanT-ZGlycine binding sites of presynaptic NMDA receptors may tonically regulate glutamate release in the rat visual cortexJournal of Neurophysiology200797181782317093111\n29. LienC-CMuYVargas-CaballeroMPooMVisual stimuli-induced LTD of GABAergic synapses mediated by presynaptic NMDA receptorsNature Neuroscience20069337238016474391\n30. MameliMCartaMPartridgeLDValenzuelaCFNeurosteroid-induced plasticity of immature synapses via retrograde modulation of presynaptic NMDA receptorsThe Journal of Neuroscience20052592285229415745954\n31. RobertABlackJAWaxmanSGEndogenous NMDA-receptor activation regulates glutamate release in cultured spinal neuronsThe Journal of Neurophysiology19988011962089658041\n32. SamsonRDParéDActivity-dependent synaptic plasticity in the central nucleus of the amygdalaThe Journal of Neuroscience20052571847185515716421\n33. SjöströmPJTurrigianoGGNelsonSBNeocortical LTD via coincident activation of presynaptic NMDA and cannabinoid receptorsNeuron200339464165412925278\n34. CasadoMIsopePAscherPInvolvement of presynaptic N-methyl-D-aspartate receptors in cerebellar long-term depressionNeuron200233112313011779485\n35. WoodhallGEvansDICunninghamMOJonesRSGNR2B-containing NMDA autoreceptors at synapses on entorhinal cortical neuronsThe Journal of Neurophysiology20018641644165111600627\n36. YangJWoodhallGLJonesRSGTonic facilitation of glutamate release by presynaptic NR2B-containing NMDA receptors is increased in the entorhinal cortex of chronically epileptic ratsThe Journal of Neuroscience200626240641016407536\n37. WilliamsKIfenprodil discriminates subtypes of the N-methyl-D-aspartate receptor: selectivity and mechanisms at recombinant heteromeric receptorsMolecular Pharmacology19934448518597901753\n38. FischerGMutelVTrubeGRo 25-6981, a highly potent and selective blocker of N-methyl-D-aspartate receptors containing the NR2B subunit. Characterization in vitroThe Journal of Pharmacology and Experimental Therapeutics19972833128512929400004\n39. AubersonYPAllgeierHBischoffSLingenhoehlKMorettiRSchmutzM5-Phosphonomethylquinoxalinediones as competitive NMDA receptor antagonists with a preference for the human 1A/2A, rather than 1A/2B receptor compositionBioorganic & Medicinal Chemistry Letters20021271099110211909726\n40. BrasierDJFeldmanDESynapse-specific expression of functional presynaptic NMDA receptors in rat somatosensory cortexThe Journal of Neuroscience20082892199221118305253\n41. NeytonJPaolettiPRelating NMDA receptor function to receptor subunit composition: limitations of the pharmacological approachThe Journal of Neuroscience20062651331133316452656\n42. YangJWetterstrandCJonesRSGFelbamate but not phenytoin or gabapentin reduces glutamate release by blocking presynaptic NMDA receptors in the entorhinal cortexEpilepsy Research2007772-315716417980555\n43. LiY-HHanT-ZMengKTonic facilitation of glutamate release by glycine binding sites on presynaptic NR2B-containing NMDA autoreceptors in the rat visual cortexNeuroscience Letters2008432321221618248890\n44. LiuLWongTPPozzaMFRole of NMDA receptor subtypes in governing the direction of hippocampal synaptic plasticityScience200430456731021102415143284\n45. BerberichSPunnakkalPJensenVLack of NMDA receptor subtype selectivity for hippocampal long-term potentiationThe Journal of Neuroscience200525296907691016033900\n46. WeitlaufCHonseYAubersonYPMishinaMLovingerDMWinderDGActivation of NR2A-containing NMDA receptors is not obligatory for NMDA receptor-dependent long-term potentiationThe Journal of Neuroscience200525378386839016162920\n47. PaolettiPAscherPNeytonJHigh-affinity zinc inhibition of NMDA NR1-NR2A receptorsThe Journal of Neuroscience19971715571157259221770\n48. RachlineJPerin-DureauFLe GoffANeytonJPaolettiPThe micromolar zinc-binding domain on the NMDA receptor subunit NR2BThe Journal of Neuroscience200525230831715647474\n49. TraynelisSFBurgessMFZhengFLyuboslavskyPPowersJLControl of voltage-independent Zinc inhibition of NMDA receptors by the NR1 subunitThe Journal of Neuroscience19981816616361759698310\n50. SobczykASvobodaKActivity-dependent plasticity of the NMDA-receptor fractional Ca2+ currentNeuron2007531172417196527\n51. AdamsMMFinkSEJanssenWGMShahRAMorrisonJHEstrogen modulates synaptic N-methyl-D-aspartate receptor subunit distribution in the aged hippocampusThe Journal of Comparative Neurology2004474341942615174084\n52. ContiFBarbaresiPMeloneMDucatiANeuronal and glial localization of NR1 and NR2A/B subunits of the NMDA receptor in the human cerebral cortexCerebral Cortex19999211012010220224\n53. KharaziaVNWeinbergRJImmunogold localization of AMPA and NMDA receptors in somatic sensory cortex of albino ratThe Journal of Comparative Neurology1999412229230210441757\n54. ChartonJPHerkertMBeckerC-MSchröderHCellular and subcellular localization of the 2B-subunit of the NMDA receptor in the adult rat telencephalonBrain Research199981626096179878886\n55. FujisawaSAokiCIn vivo blockade of N-methyl-D-aspartate receptors induces rapid trafficking of NR2B subunits away from synapses and out of spines and terminals in adult cortexNeuroscience20031211516312946699\n56. KotakVCFujisawaSLeeFAKarthikeyanOAokiCSanesDHHearing loss raises excitability in the auditory cortexThe Journal of Neuroscience200525153908391815829643\n57. O'DonnellRMolon-NoblotSLaroquePRigbyMSmithDThe ultrastructural localisation of the N-methyl-D-aspartate NR2B receptor subunit in rat lumbar spinal cordNeuroscience Letters20043711242915500960\n58. RadleyJJFarbCRHeYDistribution of NMDA and AMPA receptor subunits at thalamo-amygdaloid dendritic spinesBrain Research200711341879417207780\n59. ValtschanoffJGBuretteAWentholdRJWeinbergRJExpression of NR2 receptor subunit in rat somatic sensory cortex: synaptic distribution and colocalization with NR1 and PSD-95The Journal of Comparative Neurology1999410459961110398051\n60. AokiCFujisawaSMahadomrongkulVShahPJNaderKErisirANMDA receptor blockade in intact adult cortex increases trafficking of NR2A subunits into spines, postsynaptic densities, and axon terminalsBrain Research20039631-213914912560119\n61. JanssenWGMVissavajjhalaPAndrewsGMoranTHofPRMorrisonJHCellular and synaptic distribution of NR2A and NR2B in macaque monkey and rat hippocampus as visualized with subunit-specific monoclonal antibodiesExperimental Neurology2005191supplement 1S28S4415629759\n62. PetraliaRSSansNWangY-XWentholdRJOntogeny of postsynaptic density proteins at glutamatergic synapsesMolecular and Cellular Neuroscience200529343645215894489\n63. LucciniEMusanteVNeriERaiteriMPittalugaAN-methyl-D-aspartate autoreceptors respond to low and high agonist concentrations by facilitating, respectively, exocytosis and carrier-mediated release of glutamate in rat hippocampusJournal of Neuroscience Research200785163657366517671992\n64. LaurieDJSeeburgPHLigand affinities at recombinant N-methyl-D-aspartate receptors depend on subunit compositionEuropean Journal of Pharmacology199426833353457528680\n65. MoriHMishinaMStructure and function of the NMDA receptor channelNeuropharmacology19953410121912378570021\n66. Cull-CandySGBrickleySGNMDA receptorsto appear in Encyclopedia of Life Sciences, http://www.els.net.\n67. SobczykAScheussVSvobodaKNMDA receptor subunit-dependent [Ca2+] signaling in individual hippocampal dendritic spinesThe Journal of Neuroscience200525266037604615987933\n68. ChrobakJJLörinczABuzsákiGPhysiological patterns in the hippocampo-entorhinal cortex systemHippocampus200010445746510985285\n69. ChrobakJJBuzsákiGSelective activation of deep layer (V-VI) retrohippocampal cortical neurons during hippocampal sharp waves in the behaving ratThe Journal of Neuroscience19941410616061707931570\n70. CunninghamMODaviesCHBuhlEHKopellNWhittingtonMAGamma oscillations induced by kainate receptor activation in the entorhinal cortex in vitroThe Journal of Neuroscience200323309761976914586003\n71. CunninghamMOPervouchineDDRaccaCNeuronal metabolism governs cortical network response stateProceedings of the National Academy of Sciences of the United States of America2006103145597560116565217\n72. BuzsákiGTheta rhythm of navigation: link between path integration and landmark navigation, episodic and semantic memoryHippocampus200515782784016149082\n73. FellJKlaverPElfadilHSchallerCElgerCEFernándezGRhinal-hippocampal theta coherence during declarative memory formation: interaction with gamma synchronization?European Journal of Neuroscience20031751082108812653984\n74. VertesRPHippocampal theta rhythm: a tag for short-term memoryHippocampus200515792393516149083\n75. AlperKRaghavanMIsenhartRLocalizing epileptogenic regions in partial epilepsy using three-dimensional statistical parametric maps of background EEG source spectraNeuroImage20083931257126518024085\n76. DericiogluNSaygiSIctal scalp EEG findings in patients with mesial temporal lobe epilepsyClinical EEG and Neuroscience2008391202718318415\n77. TolnerEAKloostermanFvan VlietEAWitterMPLopes da SilvaFHGorterJAPresubiculum stimulation in vivo evokes distinct oscillations in superficial and deep entorhinal cortex layers in chronic epileptic ratsThe Journal of Neuroscience200525388755876516177045\n78. SteffensMHuppertzH-JZentnerJChauzitEFeuersteinTJUnchanged glutamine synthetase activity and increased NMDA receptor density in epileptic human neocortex: implications for the pathophysiology of epilepsyNeurochemistry International200547637938416095760\n79. ClemensBValproate decreases EEG synchronization in a use-dependent manner in idiopathic generalized epilepsySeizure200817322423317697790\n80. BélaCMónikaBMártonTIstvánKValproate selectively reduces EEG activity in anterior parts of the cortex in patients with idiopathic generalized epilepsy. A low resolution electromagnetic tomography (LORETA) studyEpilepsy Research2007752-318619117624734\n81. ClemensBPirosPBessenyeiMHollódyKLamotrigine decreases EEG synchronization in a use-dependent manner in patients with idiopathic generalized epilepsyClinical Neurophysiology2007118491091717258504\n82. YangJChamberlainSELWoodhallGLJonesRSGMobility of NMDA autoreceptors but not postsynaptic receptors in the rat entorhinal cortexThe Journal of Physiology2008586204905492418718983\n83. HattonCJPaolettiPModulation of triheteromeric NMDA receptors by N-terminal domain ligandsNeuron200546226127415848804\n84. Al-HallaqRAConradsTPVeenstraTDWentholdRJNMDA di-heteromeric receptor populations and associated proteins in rat hippocampusThe Journal of Neuroscience200727318334834317670980\n85. LuoJWangYYasudaRPDunahAWWolfeBBThe majority of N-methyl-D-aspartate receptor complexes in adult rat cerebral cortex contain at least three different subunits (NR1/NR2A/NR2B)Molecular Pharmacology199751179869016349\n86. TovarKRWestbrookGLThe incorporation of NMDA receptors with a distinct subunit composition at nascent hippocampal synapses in vitroThe Journal of Neuroscience199919104180418810234045\n87. ItoIFutaiKKatagiriHSynapse-selective impairment of NMDA receptor functions in mice lacking NMDA receptor epsilon 1 or epsilon 2 subunitThe Journal of Physiology199750024014089147327\n88. KumarSSHuguenardJRPathway-specific differences in subunit composition of synaptic NMDA receptors on pyramidal neurons in neocortexThe Journal of Neuroscience20032331100741008314602822\n89. AlonsoAde CurtisMLlinásRPostsynaptic Hebbian and non-Hebbian long-term potentiation of synaptic efficacy in the entorhinal cortex in slices and in the isolated adult guinea pig brainProceedings of the National Academy of Sciences of the United States of America19908723928092841979169\n90. BourasRChapmanCALong-term synaptic depression in the adult entorhinal cortex in vivoHippocampus200313778079014620873\n91. ChapmanCARacineRJPiriform cortex efferents to the entorhinal cortex in vivo: kindling-induced potentiation and the enhancement of long-term potentiation by low-frequency piriform cortex or medial septal stimulationHippocampus1997732572709228524\n92. CheongMYYunSHMook-JungIKangYJungMWInduction of homosynaptic long-term depression in entorhinal cortexBrain Research2002954230831012414114\n93. ChenY-HWuM-LFuW-MRegulation of presynaptic NMDA responses by external and intracellular pH changes at developing neuromuscular synapsesThe Journal of Neuroscience1998188298229909526015\n94. CraigSComminsSPlastic and metaplastic changes in the CA1 and subicular projections to the entorhinal cortexBrain Research20071147112413917368431\n95. SolgerJWoznyCManahan-VaughanDBehrJDistinct mechanisms of bidirectional activity-dependent synaptic plasticity in superficial and deep layers of rat entorhinal cortexEuropean Journal of Neuroscience20041972003200715078576"
+}

batch_13/PMC2577928.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2577928",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2577928\nAUTHORS: Robert A Smith, Robert Curtain, Mick Ovcaric, Lotti Tajouri, John MacMillan, Lyn Griffiths\n\nABSTRACT:\nTo investigate the migraine locus around the C19p13 region through analysis of the NOTCH3 gene (C19p13.2-p13.1), previously shown to be a gene involved in CADASIL and the TNFSF7 gene (C19p13), homologous to the ligands of TNF-alpha and TNF-beta, genes that have previously been associated with migraine. The NOTCH3 gene was analysed by sequencing all exons with known CADASIL mutations in a typical (non-familial hemiplegic) migraine family (MF1) that has previously been shown to be linked to C19p13. The TNFSF7 gene was investigated through SNP association analysis using a matched case-control migraine population. NOTCH3 gene sequencing results for affected members of MF1 proved to be negative for all known sequence variants giving rise to mutations for CADASIL. TNFSF7 gene chi-square results showed non-significant P values across all populations tested against controls, except for the MO subgroup which displayed a possible association with the TNFSF7 SNP (genotype, allele analysis P = 0.036, P = 0.017 respectively). Our results suggest that common migraine is not caused by any known CADASIL mutations in the NOTCH3 gene of interest. However, the TNFSF7 gene displayed signs of involvement in a MO affected population and indicates that further independent studies of this marker are warranted.\n\nBODY:\nINTRODUCTIONTypical migraine is a common debilitating disorder affecting approximately 12% of the Western population [1]. A large US epidemiological study indicated that the disease affects around 18% of women and 12 % of men in the overall population [2]. The disease has two main subtypes, migraine without aura (MO) affecting ~ 73% and migraine with aura (MA) affecting about 33%. Migraine with aura is defined as the more severe type and both subtypes can occur within the same family. The number of genes involved in common migraine is unknown; although genes for a rare severe sub-type of migraine Familial Hemplegic Migraine (FHM) have been identified on chromosome 19 [3] and recently FHM type 2 has been identified on chromosome 1 [4].Mutations in the neuronal calcium channel gene (CACNA1A) on 19p13.1-13.2 have been shown to cause FHM [3]. Calcium channel genes may also be involved in typical migraine (MA and MO) as some studies have indicated that the FHM locus may contribute to migraine with and without aura [5]. We have previously reported linkage in one large typical migraine family [6] to the CACNA1A region on chr19. Also in a study by Terwindt et al. (1998) [7] the authors detected an FHM mutation in the CACNA1A gene in a typical migraine patient with MA, suggesting that FHM may be a rare and severe form of MA.Typical migraine has been extensively studied on chromosome 19 showing positive linkage to causal markers in the C19p13 area [5,6,8-10], but only in FHM has the chromosome 19 gene and mutations within been identified [3]. The NOTCH3 gene on C19p13.2-p13.1 has previously been shown to be a gene involved in CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) [11]. CADASIL is an inherited stroke syndrome that leads to dementia. The key features of the disorder are migraine, recurrent subcortical events, and dementia, in association with diffuse white matter abnormalities on neuroimaging [12]. The gene encodes a large transmembrane receptor [13], which has been shown to be expressed in vascular smooth muscle cells in adult human tissues [14]. In a study by Joutel et al. 1997 [13] this large gene, containing 33 exons, was found to contain a cluster of mutations in exons 3 and 4 affecting 64% (32/50) of tested patients with the CADASIL disorder. Of the other 36% of CADASIL sufferers, 26% (13/50) have mutations in exons 2,5,11,14,18,19,22 or 23 and the other 10% (5/50) had no detectable mutation in the NOTCH3 gene [13]. Mutations identified to date in NOTCH3have all been missense mutations involving a loss or gain of a cysteine amino acid residue [13].The NOTCH3 gene may also be implicated in migraine, as migraine shows some symptomatic overlap with CADASIL. This has been suggested from studies performed by Hutchinson et al. (1995) [15] and Verin et al. (1995) [16] where the authors have found some correlations between the two disorders. Hutchinson et al. (1995) [15] used MRI to study 15 members of an Irish family, 10 of whom had evidence of CADASIL. Five members of this family had hemiplegic migraine. They proposed that hemiplegic migraine may be an allelic disorder to CADASIL. Migraine with aura (MA) is also included as one of the clinical hallmarks of CADASIL, with many CADASIL patients showing an unusually high frequency of attacks of migraine with atypical aura [17]. In 2003, Oberstein et al. [18] found that among 6 individuals who carried a mutation in the NOTCH3 gene, there was an increase in white matter hyperintensities on brain MRI, compared to controls and migraine with aura was more common in these subjects than in their controls. In addition, a recent association study by Schwaag et al. identified a significant association of the synonymous rs1043994 polymorphism with migraine [19]. However, not all NOTCH3 mutations may be involved in migraine, as an association study by Borroni et al. [20] indicated that the that functional polymorphism T6746C of NOTCH3 did not have any association with migraine in the tested Italian population.The TNFSF7 gene (Tumor Necrosis Factor Ligand Superfamily, member 7) localized on C19p13 is a surface antigen found on activated, but not resting, T and B lymphocytes [21]. It is a 19 amino acid protein containing a 20-amino acid hydrophilic N-terminal domain that lacks a signal sequence; an 18-amino acid hydrophobic region that presumably functions as a transmembrane anchor; and a C-terminal domain that contains 2 potential N-linked glycosylation sites is extracellular classifying TNFSF7 as a type II transmembrane protein. TNFSF7 is homologous to the ligands of the TNF receptor family, including TNF-alpha, TNF-beta and the CD40 ligand, showing 19 to 24% amino acid sequence identity in the extracellular region. [21]. TNF-alpha and TNF-beta have both been previously associated with migraine. Rainero et al. 2004 [22], recently found an association between the tumor necrosis factor-alpha -308 G/A gene polymorphism and migraine. In a group of 299 migraine patients and 306 control subjects, the association of this polymorphism with the occurrence and clinical characteristics of migraine was tested. They found that homozygosity for the G allele was associated with an increased risk of migraine (odds ratio [OR] = 2.85, p <0.001) [22]. Also Empl et al. 2003 [23] suggested that TNF-alpha could contribute to migraine pain generation and in their study of TNF-alpha and its soluble receptor sTNF-RI, they found that migraine patients tended to have less concentration levels of sTNF-RI (794 +/- 158 pg/ml) than controls (945 +/- 137 pg/ml) [23]. The authors suggested that if TNF-alpha plays a role in migraine physiopathology, migraine patients may lack sufficient antagonistic sTNF-RI to neutralize hyperalgesic TNF-alpha during a migraine attack, causing pain [23]. The TNF-beta gene has been associated with migraine without aura. Trabace et al. 2002 [24] found the frequency of the TNFB*2 allele of the TNF-beta gene was significantly increased in patients with migraine without aura as compared with the control group (78.72% versus 61.4%, Pc =.004) [24]. They stated that carriage of the TNFB*2 allele confers a high risk for the development of migraine without aura. The data supported the hypothesis that TNF-beta could be a susceptibility gene in migraine without aura [24]. Since TNFSF7 is homologous to the ligands of the TNF receptor family i.e. TNF-alpha and TNF-beta, (and is localized to a migraine susceptibility area at C19p13) we decided to investigate this gene for association with migraine.This study investigated the migraine susceptibility locus C19p13 by testing two genes, from this genomic region, that may possibly be involved in the disorder. The NOTCH3 gene was analysed by sequencing all exons with known CADASIL mutations through a family previously linked to C19p13 [6]. The TNFSF7 gene was also investigated using SNP case-control association analysis in a population of migraineurs and matched controls.MATERIALS AND METHODOLOGYSubjectsThe subjects who participated in this study were all of Australian Caucasian origin and were diagnosed for migraine by a detailed questionnaire and a clinical neurologist in accordance to the International Headache Society guidelines [25]. The control population consisted of individuals who indicated on a questionnaire that they had never suffered from migraine or any similar condition and that none of their first or second degree relatives suffered from migraine or similar conditions. This research was approved by the Griffith University Ethics Committee and all subjects participating in the study gave consent. Blood samples for the pedigree population as well as the association population were collected through the Genomics Research Centre clinic. Genomic DNA, from these patients, was extracted utilizing DNA extraction methods described previously [26, 27].For mutation analysis, 12 affected DNA samples obtained from migraine family 1 (MF1) (Fig. 1), a pedigree showing linkage to 19p13 in a previous study [6], were screened for the known exon 3 and exon 4 CADASIL mutations (Table 1) in the NOTCH3 gene. The pedigree family members tested (indicated by an arrow in Fig. 1) were migraine sufferers of which eight were diagnosed MA and four MO affected, two were males (both MA affected) and ten females (six MA and four MO affected). The age of these migraine affected pedigree members range from 31 to 87 years.Secondly 3 of the 12 family members, from three separate branches of MF1 (Fig. 1), were further screened by sequencing the rest of all known exons containing CADASIL mutations in the NOTCH3 gene (Table 1). The association study included a case-control population of 220 migraineurs and 220 sex, age and ethnicity matched controls.Mutation AnalysisMutation screening was performed by direct sequencing of exons, exhibiting known CADASIL mutations, in the NOTCH3 gene utilizing genomic DNA samples from migraine affected pedigree members of MF1 (Fig. 1). Templates of PCR DNA fragments were generated in the forward and reverse direction using standard PCR (outlined below in association study) and sequencing conditions (ABI protocols). Table 2 shows a list of exon primer sequences used in sequencing for the known mutations. The resulting products were electrophoresed on an ABI377 Sequencer (Applied Biosystems) and analysed using Sequencher software (Gene Codes Corporation) against a control sequence (Accession number: NM 000435).Association AnalysisThe TNFSF7 SNP, for association analysis, was selected utilizing the software SNPbrowserTM (ABI) and Ensembl Genome Browser, NCBI SNP databases. This synonymous coding SNP (refSNP 1862511, C/T variation, codes for amino acid Cysteine) is contained within exon 3 of the TNFSF7 gene with a minor allele frequency of approximately 0.3 (T allele). Primer sequences for the SNP were designed by utilizing Primer ExpressTM v 2.0 software (ABI) and are displayed below.Forward primer: 5’- AGCACTGGGCCGCTCC – 3’Reverse primer: 5’- CAAAAGTGTCCCAGTGAGGTTG – 3’The PCR reaction utilized final concentrations of 1.75mM, 0.2µM, 200µM, 1X and 1U for MgCl2, Primers, dNTPs, 10x Buffer and Taq polymerase, respectively. Approximately 40ng of DNA per reaction was used in the PCR. Cycling conditions for PCR consisted of 94°C for 4 minutes, then 35 cycles of 94°C for 1 minute and 60°C for 1 minute. The final PCR extension consisted of 72°C for 2 minutes.SNP genotype analysis was performed using a restriction enzyme site that overlapped the SNP site (CTCCTC). Restriction enzyme digest conditions utilized 10ul of PCR product and a 10ul mix of 3U of restriction enzyme with 1X enzyme buffer and H2O. Digest products were loaded and electrophoresised on 3% standard agarose gels running at 110 volts for 1 hour. The genotyping allele set and corresponding restriction digest fragment size consisted of: C allele = 192,100bp, T allele = 292bp, C/T = 292,192,100bp.Genotype and allele frequencies for the SNP variant were calculated from observed genotype counts. The expected genotype proportions according to the Hardy-Weinberg law were calculated and compared to observed genotypes as a control for systematic genotyping error and population stratification. Genotype and allele frequencies were initially assessed for association with migraine, then MA and MO populations were investigated using conventional contingency table analyses incorporating the standard chi-squared test for independence.RESULTSMutation AnalysisThe DNA samples from MF1 that were examined for NOTCH3 – CADASIL mutations were from individuals affected with typical migraine in a pedigree previously shown to be linked to chromosome 19p13 [6]. Firstly 12 members from MF1 were screened for exon 3 and 4 mutations of the NOTCH3 gene using direct sequencing methods - a diagnostic screening process that is utilized first in detecting CADASIL mutations. The results revealed no variations or mutations detected in the sequence of the two exons within the twelve migraine family one DNA templates. Secondly, 3 members of MF1 were chosen from three separate pedigree branches to test for the rest of the exons with known CADASIL mutations in NOTCH3 (Table 1). These results also proved to be negative with no known sequence variants detected in the extended exon sequence analysis. The sequencing results displayed that of a normal coding sequence for the NOTCH3 gene (Accession number: NM 000435).Association AnalysisTotal distribution of the rs1862511 SNP genotype and allele frequencies of the TNFSF7 gene in Migraine (Total), MA, MO and Control Groups are displayed in Table 3.The results of chi-square analysis of the TNFSF7 SNP (rs1862511) are displayed in Table 4. All population data of observed genotypes fitted the expected genotype proportions according to the Hardy-Weinberg law. Chi-square results showed non-significant P values (P > 0.05) across all populations tested against controls. The only exception was for the MO subgroup which displayed a positive association for both alleles (P = 0.017) and genotypes (P = 0.036) for the TNFSF7 SNP (Table 4).DISCUSSIONThe procedure for genetic diagnostic testing for CADASIL mutations has recently been revised. We undertook a similar approach to this revision in testing for these mutations in common migraine. A study in Britain [28], found 15 different point mutations in the NOTCH3 gene in 48 families, 73% of which were in exon 4, 8% in exon 3, and 6% in each of exons 5 and 6. The authors suggested that on the basis of this spectrum the suggested protocol for genetic diagnostic testing for CADASIL would be to screen exon 4 and proceed to mutational screening of exons 3, 5, and 6 where indicated [28]. A similar approach can be undertaken in diagnostic testing for other diseases with known mutations in genes causing a disorder, including FHM whereby two genes and corresponding mutations have been identified in causing this rare subtype of migraine.The NOTCH3 gene has been localised to C19p13, a region showing linkage to MF1 a typical (not FHM) migraine pedigree. Our sequencing results for this gene, specifically testing exons with known CADASIL mutations proved to be negative. These results indicate that common migraine, at least in this pedigree, is not caused by mutations in the NOTCH3 gene.TNF-alpha and TNF-beta have both been previously associated with migraine [22, 24]. Since TNFSF7 is localized to the same migraine susceptibility area at C19p13 and shows homology to the ligands of the TNF-alpha and TNF-beta genes (both localized at C6p21.3), we decided to also investigate this gene for involvement in migraine. A suitable informative SNP (rs1862511) was selected for an association analysis study involving a matched case-control population. Overall chi-square results comparing migraine (total) and MA sub-populations with sex, age and ethnicity matched controls proved to be non-significant, with P values of P = 0.136, P = 0.589 respectively obtained (Table 4). These values were both above the threshold of P<0.05 for significance. However, the MO subgroup displayed signs of possible involvement of the TNFSF7 SNP with this sub-population compared to controls. P values for genotype and allele analysis were P = 0.036 and P = 0.017, respectively (Table 4). These results indicate that this genetic variant may play a role in migraine without aura and warrant further investigation.Overall the results presented here do not support a relationship between NOTCH3 and common migraine. Although CADASIL patients often display migraine symptoms there was no evidence that CADASIL mutations from the NOTCH3 gene are involved in typical migraine in our C19p13 linked [6] migraine family samples. However, this does not discount the possibility of a role for mutations in the as yet unscreened exons in NOTCH3 in migraine susceptibility. The TNFSF7 gene also localized to the C19p13 region did not show any association with typical migraine or the MA subtype, however a weak association was found with MO affected individuals. Since this gene shows homology to the TNF receptor family and both TNF-alpha and beta have previously shown association with typical migraine [22, 24], the TNFSF7 gene (localizedon C19p13, a migraine linked region, see [5,6,8-10]) was investigated in migraine. Our association study showed signs of possible involvement of this gene with the MO subtype only, but clearly more work needs to be done in increasing sample size numbers and also confirmation from other research groups to give an indication of a relationship of this gene to migraine without aura.CONCLUSIONIt is clear that typical migraine is a complex disorder that may involve several genes on various chromosomes. Since the C19p13 region has been implicated in a number of studies it is considered a ‘hotspot’ for the disorder. More extensive sequencing of the NOTCH 3 gene that may identify novel mutations that relate to migraine should be undertaken. Further studies need to be performed with TNFSF7 and MO affected individuals and also other possible TNF receptor family homologues on C19p13, such as TNFSF9 (C19p13.3) and TNFSF14 (C19p13.3) should be tested for migraine involvement. It may be useful to explore the TNF receptor-migraine interaction further by studying these genes localized to the C19p13 region.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2579486.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2579486",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2579486\nAUTHORS: Yuko Seko, Noriyuki Azuma, Yoriko Takahashi, Hatsune Makino, Toshiyuki Morito, Takeshi Muneta, Kenji Matsumoto, Hirohisa Saito, Ichiro Sekiya, Akihiro Umezawa\n\nABSTRACT:\nBackgroundThe sclera maintains and protects the eye ball, which receives visual inputs. Although the sclera does not contribute significantly to visual perception, scleral diseases such as refractory scleritis, scleral perforation and pathological myopia are considered incurable or difficult to cure. The aim of this study is to identify characteristics of the human sclera as one of the connective tissues derived from the neural crest and mesoderm.Methodology/Principal FindingsWe have demonstrated microarray data of cultured human infant scleral cells. Hierarchical clustering was performed to group scleral cells and other mesenchymal cells into subcategories. Hierarchical clustering analysis showed similarity between scleral cells and auricular cartilage-derived cells. Cultured micromasses of scleral cells exposed to TGF-βs and BMP2 produced an abundant matrix. The expression of cartilage-associated genes, such as Indian hedge hog, type X collagen, and MMP13, was up-regulated within 3 weeks in vitro. These results suggest that human ‘sclera’-derived cells can be considered chondrocytes when cultured ex vivo.Conclusions/SignificanceOur present study shows a chondrogenic potential of human sclera. Interestingly, the sclera of certain vertebrates, such as birds and fish, is composed of hyaline cartilage. Although the human sclera is not a cartilaginous tissue, the human sclera maintains chondrogenic potential throughout evolution. In addition, our findings directly explain an enigma that the sclera and the joint cartilage are common targets of inflammatory cells in rheumatic arthritis. The present global gene expression database will contribute to the clarification of the pathogenesis of developmental diseases such as high myopia.\n\nBODY:\nIntroductionThe eye receives information from the outside as the retinal image, converting it into electrical signals for the brain, leading to visual perception. The retinal image is stabilized by the balance of intraocular pressure and the curvatures of the scleral and corneal envelope. In order to keep this balance, the rigidity of the sclera and the cornea are essential, especially the sclera must be rigid enough for the eyeball to be rotated by powerful extraocular muscles adhering to the sclera. The sclera and the corneal stroma that are anatomically continuous have common characteristics such as mechanical rigidity, and share a common origin, i.e., the neural crest. However, the cornea and the sclera are different in transparency: the cornea is completely transparent to produce a sharp image on the retina; the sclera is opaque to avoid the internal light scattering affecting the retinal image. This corneal transparency has been attributed to significant changes in the structure, especially of collagen fibrils, in the latter stages of development [1]. Multipotent progenitor/precursor cells of corneal stroma are identified from the mouse eye [2]. On the other hand, existence of multipotent progenitor/precursor cells in the sclera remains unclarified. Although the sclera does not contribute significantly to visual perception, scleral diseases such as refractory scleritis, scleral perforation and pathological myopia are considered incurable or difficult to cure.Microarray analysis of murine scleral development [3] and global sequencing analysis from the human scleral cDNA library [4] have been reported. To clarify pathogenesis of developmental diseases such as high myopia, a database of genes expressed in the sclera of younger donors is important. We here demonstrate with a global expression database of human infant sclera that the sclera derived from the neural crest evolutionarily retains characteristics of cartilage.ResultsIsolation and cell culture of human scleral cellsScleral tissues were excised from surgical specimens collected during treatment for retinoblastoma. The scleral tissue was cut into smaller pieces and cultured in the growth medium. The scleral cells began growing out almost one week after the start of cultivation. Scleral cells exhibited a fibroblast-like spindle shape or polygonal shape in morphology when cultured in monolayer (Fig. 1A). The cells from PD 5 to PD 31 rapidly proliferated in culture, and propagated continuously (Fig. 1B). The cells stopped replicating and became broad and flat at PD 43 or 264 days, indicating that they had entered senescence. The morphological changes are PD-dependent.10.1371/journal.pone.0003709.g001Figure 1Proliferation of human ‘sclera’-derived cells.A. Photograph of primary cultured human ‘sclera’-derived cells by phase-contrast microscope. B. Growth curve of cultured human ‘sclera’-derived cells. Vertical axis indicates population doublings (PD) and horizontal axis indicates days after innoculation of human ���sclera’-derived cells.Global outlook by hierarchical clustering and PCATo clarify the specific gene expression profile of scleral cells, we compared the expression levels of 54,675 probes in the cultured scleral cells and other cultured cells (Table 1) using the Affymetrix GeneChip oligonucleotide arrays. We first performed hierarchical clustering and PCA on the expression pattern. PCA showed similarity between scleral cells and chondrocytes derived from elastic cartilage (Fig. 2A). Hierarchical clustering analysis based on all probes showed similarity between scleral cells and chondrocytes (Fig. 2B). This similarity led us to hypothesize that the scleral cells are chondrocytes when proliferated ex vivo, or have a chondrogenic potential. We then performed PCA from the expression data of cartilage-associated genes, including aggrecan, Sox9, and parathyroid hormone receptor (Table S1). These genes are categorized as “cartilage condensation” or “proteoglycan biosynthesis” according to Gene Ontology. PCA based on cartilage-associated genes demonstrated that scleral cells are grouped into the same category that includes chondrocytes, synovial cells, and synovial fluid-derived cells (Fig. 2C). The synovial cells and synovial fluid-derived cells used in this study have a strong chondrogenic potential [5]–[7]. Hierarchical clustering analysis based on the cartilage-associated genes also demonstrated that sclera, cartilage, synovium, and joint fluid are categorized into the same group (Fig. 2D, Fig. 2E, Fig. S1).10.1371/journal.pone.0003709.g002Figure 2Global gene expression analysis of cultured human cells.A. Three-dimensional representation of PCA of gene expression levels (Human Genome U133 Plus 2.0: 54,675 probes). The gene expression data from scleral cells following one passage from the primary cultured cells (equivalent to appropriately 4 PDs) were used for PCA. Sclera and cartilage are positioned closely adjacent (shown in circle). B. Hierarchical clustering analysis based on the expression of all genes (Human Genome U133 Plus 2.0: 54,675 probes, NIA Array Analysis) shows similarity between scleral cells and chondrocytes. C. PCA of the cartilage-associated gene expression (Table S1). Sclera, cartilage, synovium, and joint fluid are positioned closely adjacent (shown in circle). D. Hierarchical clustering analysis based on expression levels of the cartilage-associated genes (NIA Array Analysis). Sclera, cartilage, synovium, and joint fluid are categorized into the same group. E. Hierarchical clustering analysis (TIGR MeV, see the Materials & Methods) with the heat map, based on expression levels of the cartilage-associated genes. Each row represents a gene; each column represents a cell population. Sclera, cartilage, synovium, and joint fluid are categorized into the same group. Cells derived from cartilage, synovium, and joint fluid are capable of generating cartilage in vivo [7], [34].10.1371/journal.pone.0003709.t001Table 1Human cells analyzed in this study.TitleDescriptionBone marrowBone marrow-derived cell (P1)HepatocyteHepatocyte (P0)EndometriumEndometrial cellSynoviumSynovium-derived cell (P1)Joint fluidJoint fluid-derived cell (P1)MuscleMuscle-derived cell (P1)BoneCancellous bone-derived cell (P1)FatSubcutaneus fat-derived cell (P1)Amniotic epitheliumAmniotic epithelial cell (P4)Umbilical cord (1)Umbilical cord-derived cell (P0) (1)Umbilical cord (2)Umbilical cord-derived cell (P0) (2)CartilageAuricular cartilage-derived cell (P1)ScleraSclera-derived cell (P1)Cornea (stroma)Keratocyte (P1)PeriostiumPeriostium-derived cell (P1)DermisDermal fibroblast (P2)Cortical boneCortical bone-derived cell (P3)Gene chip analysis was performed using RNAs from the cells obtained from each tissue. The cells obtained from bone marrow, liver, synovium, joint fluid, muscle, bone, and fat were cultivated as previously described [31]–[33]. Amniotic epithelial cells and umbilical cord-derived cells were cultured after each tissue was manually separated from the placenta and minced by surgical knife and scissors. Auricular cartilage-derived cells, periostium-derived cells, dermal fibroblasts, and cortical bone-derived cells started to be cultured after each tissue was manually separated from surgical specimens from patients with polydactyly or microtia. Keratocytes and scleral cells were obtained from corneal stroma and sclera (also see the Materials and Methods section). “Endometrium” was obtained from the homogenized endometrial cells under liquid nitrogen. All cells were harvested under signed informed consent, with the approval of the Ethics Committee of the National Institute for Child and Health Development, Tokyo. Signed informed consent was obtained from donors and the surgical specimens were irreversibly de-identified. All experiments handling human cells and tissues were performed in line with the Tenets of the Declaration of Helsinki. Global gene expression profiles of those cells are uploaded to GEO accession #GSE10934 at http://www.ncbi.nlm.nih.gov/geo/index.cgi.P: passage. P0 and P1 represents primary cell culture and cell culture one passage after starting primary culture from tissues, respectively.Chondrogenesis of human scleral cellsAfter reaching 70–80% sub-confluence, we started the micromass culture of scleral cells. Four weeks after culture in a chondrogenic medium containg TGF-β1 and BMP2, a pellet of human scleral cells exhibited a spherical shape (Fig. 3A). This pellet showed an alcian blue positive extracellular matrix, indicating that cultured micromasses of scleral cells exposed to TGF-β1 and BMP2 produce an abundant matrix (Fig. 3B). RT-PCR analysis demonstrated that scleral cells at passage 0 expressed aggrecan, COL2A, SOX5, SOX6, SOX9, and PTHR1 mRNAs (Fig. 3C). These expressions were maintained in the cells after 10 population doublings. After in vitro chondrogenesis of scleral cells, COL10A, SOX5, IHH, and MMP13 mRNA expressions increased. After human scleral cells labeled with DiI were implanted into a rat cartilage defect, the cells expressed type II collagen (Fig. 3D). These results demonstrated that human scleral cells retained chondrogenic potential both in vitro and in vivo.10.1371/journal.pone.0003709.g003Figure 3Chondrogenesis of human ‘sclera’-derived cells.A. In vitro chondrogenesis. ‘Sclera’-derived cells were centrifuged to make a pellet and cultured in chondrogenic medium for 4 weeks. Macroscopic feature is shown. B. Histological section of a pellet by micromass culture in a chondrogenic medium stained with alcian blue. Bar: 100 µm. C. Reverse transcriptase-PCR for cartilage-associated genes. Total RNAs were prepared from scleral cells at passage 0, at 10 population doublings, after in vitro chondrogenic induction, and normal cartilage as a positive control. D. Histological sections 4 weeks after transplantation of human scleral cells into cartilage defect of the knee in a rat. (a) Toluidin blue staining. (b) Immunohistochemistry. Human scleral cells were labeled with DiI (red). Nuclei were stained with DAPI (blue). Type II collagen was shown as green.DiscussionTracing back of human scleral cells to chondrocytes through cultivationThis study was undertaken to investigate if human sclera has a chondrogenic nature like chicken sclera [8], [9]. Bioinformatics of human scleral cells suggest similarity between scleral cells and chondrocytes, and this similarity may be attributed to evolution of the sclera (Fig. 4), that is, animals such as elasmobranch, teleost fish, amphibians, reptiles and birds incorporate the development of a cup of hyaline cartilage in the sclera [10]. Scleral cartilage is hypothesized to counter against the traction force of the extraocular muscle and against the accommodative force to move or deform the lens by intraocular muscles. In this paper, we employ the global gene expression approach to human scleral cells. As a result, scleral cells and chondrocytes are found to share common chondrogenic characteristics.10.1371/journal.pone.0003709.g004Figure 4The distribution of scleral cartilage in vertebrates.The chondrogenic nature of the sclera is conserved across species. The figure is modified from Franz-Odendaal, TA, et al., 2006 [10]. Species that have cartilage in the sclera are underlined; species with either absence or presence of cartilage in the sclera, depending on family, are dot-underlined; species without cartilage in the sclera are non-underlined.Simulation of chondrogenic process during developmentThe phenotype of the differentiated chondrocyte is characterized by the synthesis, deposition, and maintenance of cartilage-specific extracellular matrix molecules, including type II collagen and aggrecan [11]–[13]. Three-dimensional culture is a prerequisite for exhibition of this chondrogenic phenotype in vitro since the phenotype of differentiated chondrocytes is unstable in culture and is rapidly lost during serial monolayer subculturing [14]–[16]. The expression pattern of cartilage-associated genes in sclera-derived cells after induction is consistent with that of chondrocytes during development (Fig. 3C, Fig. S2): a) Consistent expression of type II collagen and aggrecan, markers of early-phase chondrogenesis [17], [18] in sclera-derived cells, indicates that sclera-derived cells retain their chondrogenic nature as a default state; b) Induction of type X collagen and MMP13 genes after pellet formation of sclera-derived cells may simulate late-stage chondrogenesis. In addition, other chondrocyte-associated genes, such as sox5, IHH, and PTHR1 were also up-regulated. Sox5 functions as a transcription factor necessary for chondrogenesis [19], [20], IHH promotes chondrogenesis as a cytokine [21], and PTHR1 mediates parathyroid hormone signaling as a specific receptor [18]. These results suggest that ex vivo culture of sclera-derived cells simulates the developmental process of chondrogenesis. Despite the chondrogenic nature of sclera-derived cells, lack of cartilage in the sclera in humans may be attributed to cis- and trans-regulation of cartilage-associated gene(s), or an unclarified inhibitory mechanism that was altered during evolution (Fig. 4).Implication of chondrogenic nature of sclera in diseasesThe fact that the gene expression pattern of the human fibrous sclera is similar to that of cartilage is interesting not only as comparative anatomy but also from a patho-etiological view point. The sclera and the joint cartilage are common targets for inflammatory cells in rheumatic arthritis [22], [23] or polychondritis [24], implying common proteins between the sclera and the synovium. Although the target protein(s) remains unclarified, our findings directly explain an enigma that both the sclera and the joint cartilage are affected in rheumatic arthritis. Furthermore, mutations in genes for type II and type XI collagen are a cause of Stickler syndrome [25], [26]. Patients with Stickler syndrome have joint deformity and severe high myopia due to an abnormality of the sclera. These affected lesions may be attributed to the chondrogenic nature of human sclera. In conclusion, our present study shows a chondrogenic potential of human sclera and explains the etiology of scleral disorders, at least in part. In addition, we would like to emphasize that the first database of gene expression in the human infant sclera (uploaded to GEO accession #GSE10934 at http://www.ncbi.nlm.nih.gov/geo/index.cgi) may contribute to the elucidation of scleral diseases in the future.Materials and MethodsIsolation and cell culture of human scleral cellsScleral tissues were excised from surgical specimens as a therapy of retinoblastoma, under signed informed consent, with the approval (approval number, #156) of the Ethics Committee of the National Institute for Child and Health Development, Tokyo. Signed informed consent was obtained from donors, and the surgical specimens were irreversibly de-identified. All experiments handling human cells and tissues were performed in line with the Tenets of the Declaration of Helsinki. The scleral pieces were cut into smaller pieces and cultured in the growth medium (GM): Dulbecco's modified Eagle's medium (DMEM)/Nutrient mixture F12 (1:1) with high glucose supplemented with 10% fetal bovine serum, insulin-transferrin-selenium, and MEM-NEAA (GIBCO).Oligonucleotide microarrayTotal RNAs were isolated from cultured scleral cells in the growth medium without any induction of differentiation to perform the gene chip analysis. Total RNA was extracted from a total of 5×106 cultured human scleral cells and other mesenchymal cells (Table 1) using RNeasy Plus mini-kit® (Qiagen, Maryland, USA) according to the manufacturer's instructions. A comprehensive expression analysis was performed using 2 µg of total RNA from each sample and GeneChip® Human Genome U133 plus 2.0 probe arrays (Affymetrix, Santa Clara, CA) according to the manufacturer's instructions. To normalize the variations in staining intensity among chips, the ‘Signal’ values for all probes on a given chip were divided by the median value for expression of all genes on the chip. To consider genes containing only a background signal, probes were eliminated only if the ‘Signal’ value was less than 10, or the Detection call was ‘Absent’ in any sample using GeneSpring software version 7.2 (Agilent Technologies, Palo Alto). The gene chip analysis was carried out on 8 independent scleral cultures.Hierarchical clustering and principal component analysis (PCA)To analyze the gene expression data in an unsupervised manner by gene chip array, we used hierarchical clustering and principal component analysis (NIA Array; http://lgsun.grc.nia.nih.gov/ANOVA/\n[27], TIGR MeV; http://www.tm4.org/mer.html\n[28]). The hierarchical clustering techniques classify data by similarity and the results are represented by dendrogram. PCA is a multivariate analysis technique which finds major patterns in data variability. Hierarchical clustering and PCA were performed on the data of gene chip analysis (a single assay for each sample) to group scleral cells and other mesenchymal cells into subcategories (Table 1).In vitro chondrogenesisTwo hundred thousand scleral cells were placed in a 15-ml polypropylene tube (Becton Dickinson) and centrifuged for 10 minutes. The pellet was cultured in DF-C medium™ containing 0.1 µM dexamethasone, 1 mM sodium pyruvate, 0.17 mM ascorbic acid-2-phosphate, 0.35 mM proline, 6.25 µg/ml bovine insulin, 6.25 µg/ml transferrin, 6.25 µg/ml selenous acid, 5.33 µg/ml linoleic acid, 1.25 mg/ml BSA, 5 ng/ml TGF-β1, 5 ng/ml BMP2, and 3% fetal bovine serum (TOYOBO). The medium was replaced every 3 to 4 days for 28 days. For microscopy, the pellets were embedded in paraffin, cut into 5-µm sections, and stained with alcian blue [29], [30].In vivo chondrogenesisUnder anesthesia, full thickness cartilage defects were created in the trochlear groove of the femur in SD rats. The defects were filled with DiI-labeled human scleral cells. The rats were returned to their cages after the operation and allowed to move freely. Animals were sacrificed with an overdose of sodium pentobarbital at 4 weeks after the operation. Specimens were dissected and embedded in paraffin. The sections were stained with toluidine blue and immunohistochemically stained with anti-type II collagen antibodies (clone F-57, DAIICHI FINE CHEMICAL, Co. Ltd., Toyama, Japan). All animals received humane care in compliance with the “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the “Guide for the Care and Use of Laboratory Animals” prepared by the Institute of Laboratory Animal Resources and published by the US National Institutes of Health (NIH Publication No. 86-23, revised 1985). The operation protocols were accepted by the Laboratory Animal Care and Use Committee of the Research Institute for Child and Health Development (2003-002).Reverse transcriptase-PCRTotal RNA was isolated with an RNeasy Plus mini-kit. Cartilage pellets were digested with 3 mg/ml Collagenase D for 3 hours at 37°C.The following PCR primer sets were used for cartilage-associated genes: aggrecan, sense (5′-TACACTGGCGAGCACTGTAAC -3′) and antisense (5′-CAGTGGCCCTGGTACTTGTT-3′), product size, 71 bp; collagen, type II, alpha 1, sense (5′-TTCAGCTATGGAGATGACAATC -3′) and antisense (5′-AGAGTCCTAGAGTGACTGAG -3′), product size, 472 bp; collagen, typeX, alpha 1, sense (5′-CACCTTCTGCACTGCTCATC-3′) and antisense (5′-GGCAGCATATTCTCAGATGGA-3′), product sizem, 104 bp; SOX5, sense (5′-AGCCAGAGTTAGCACAATAGG -3′) and antisense (5′-CATGATTGCCTTGTATTC -3′), product size, 619 bp; SOX6, sense (5′-ACTGTGGCTGAAGCACGAGTC -3′) and antisense (5′-TCCGCCATCTGTCTTCATACC -3′), product size, 562 bp; SOX9, sense (5′-GTACCCGCACTTGCACAAC-3′) and antisense (5′-TCGCTCTCGTTCAGAAGTCTC-3′), product size 72 bp; Indian hedgehog homolog (IHH), sense (5′-TGCATTGCTCCGTCAAGTC-3′) and antisense (5′-CCACTCTCCAGGCGTACCT-3′), product size 88 bp; parathyroid hormone receptor 1(PTHR1), sense (5′-CCTGAGTCTGAGGAGGACAAG-3′) and antisense (5′-CACAGGATGTGGTCCCATT-3′), product size 86 bp; matrix metallopeptidase 13 (MMP13), sense (5′-CCAGTCTCCGAGGAGAAACA-3′) and antisense (5′-AAAAACAGCTCCGCATCAAC-3′),product size, 85 bp, and GAPDH, sense (5′- GCTCAGACACCATGGGGAAGGT-3′) and antisense (5′-GTGGTGCAGGAGGCATTGCTGA-3′), product size, 474 bp.Supporting InformationFigure S1Global gene expression analysis of cultured human cells. Hierarchical clustering analysis based on expression levels of the cartilage-associated genes (NIA Array Analysis). We performed gene chip analysis (a single assay for each analysis) for eight independent primary scleral cultures from five patients (donors). We started eight independent cultures from three different scleral sites of Donor 2 (e.g. the anterior site 1.5 mm apart from the limbs, the middle part, and the posterior part), 2 different scleral sites of Donor 5, and three scleral sites of Donor 1, 3, and 4. We performed hierarchical clustering analysis, using these independent cultures and obtained consistent results, that is, “sclera”-derived cells are categorized into one sub-group. Furthermore, the sclera, cartilage, synovium, and joint fluid are categorized into the same group.(0.07 MB PDF)Click here for additional data file.Figure S2Cartilage-associated gene expressions in cultured fibroblasts derived from the dermis and the sclera. Cartilage-associated gene expressions by RT-PCR in cultured fibroblasts derived from the dermis and the sclera. Aggrecan, COL2A, IHH and PTHR mRNA expressions were clearly stronger in the scleral fibroblasts compared to the dermal fibroblasts, indicating that chondrogenic nature could be specific for the sclera among collagenous tissues.(0.01 MB PDF)Click here for additional data file.Table S1Cartilage-associated genes(0.01 MB PDF)Click here for additional data file.\n\nREFERENCES:\n1. ConnonCJMeekKMKinoshitaSQuantockAJ\n2004\nSpatial and temporal alterations in the collagen fibrillar array during the onset of transparency in the avian cornea.\nExp Eye Res\n78\n909\n915\n15051472\n2. YoshidaSShimmuraSShimazakiJShinozakiNTsubotaK\n2005\nSerum-free spheroid culture of mouse corneal keratocytes.\nInvest Ophthalmol Vis Sci\n46\n1653\n1658\n15851565\n3. ZhouJRappaportEFTobiasJWYoungTL\n2006\nDifferential gene expression in mouse sclera during ocular development.\nInvest Ophthalmol Vis Sci\n47\n1794\n1802\n16638983\n4. YoungTLGuoXDKingRAJohnsonJMRadaJA\n2003\nIdentification of genes expressed in a human scleral cDNA library.\nMol Vis\n9\n508\n514\n14551531\n5. SakaguchiYSekiyaIYagishitaKMunetaT\n2005\nComparison of human stem cells derived from various mesenchymal tissues: superiority of synovium as a cell source.\nArthritis Rheum\n52\n2521\n2529\n16052568\n6. YoshimuraHMunetaTNimuraAYokoyamaAKogaH\n2007\nComparison of rat mesenchymal stem cells derived from bone marrow, synovium, periosteum, adipose tissue, and muscle.\nCell Tissue Res\n327\n449\n462\n17053900\n7. KogaHMunetaTJuYJNagaseTNimuraA\n2007\nSynovial stem cells are regionally specified according to local microenvironments after implantation for cartilage regeneration.\nStem Cells\n25\n689\n696\n17138960\n8. SekoYShimokawaHTokoroT\n1995\nExpression of bFGF and TGF-beta 2 in experimental myopia in chicks.\nInvest Ophthalmol Vis Sci\n36\n1183\n1187\n7730028\n9. SekoYTanakaYTokoroT\n1995\nInfluence of bFGF as a potent growth stimulator and TGF-beta as a growth regulator on scleral chondrocytes and scleral fibroblasts in vitro.\nOphthalmic Res\n27\n144\n152\n8538991\n10. Franz-OdendaalTAHallBK\n2006\nSkeletal elements within teleost eyes and a discussion of their homology.\nJ Morphol\n267\n1326\n1337\n17051547\n11. ArcherCWMcDowellJBaylissMTStephensMDBentleyG\n1990\nPhenotypic modulation in sub-populations of human articular chondrocytes in vitro.\nJ Cell Sci\n97 (Pt 2)\n361\n371\n2277097\n12. HauselmannHJFernandesRJMokSSSchmidTMBlockJA\n1994\nPhenotypic stability of bovine articular chondrocytes after long-term culture in alginate beads.\nJ Cell Sci\n107 (Pt 1)\n17\n27\n8175906\n13. ReginatoAMIozzoRVJimenezSA\n1994\nFormation of nodular structures resembling mature articular cartilage in long-term primary cultures of human fetal epiphyseal chondrocytes on a hydrogel substrate.\nArthritis Rheum\n37\n1338\n1349\n7945499\n14. BenyaPDShafferJD\n1982\nDedifferentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels.\nCell\n30\n215\n224\n7127471\n15. BonaventureJKadhomNCohen-SolalLNgKHBourguignonJ\n1994\nReexpression of cartilage-specific genes by dedifferentiated human articular chondrocytes cultured in alginate beads.\nExp Cell Res\n212\n97\n104\n8174647\n16. LefebvreVPeeters-JorisCVaesG\n1990\nProduction of collagens, collagenase and collagenase inhibitor during the dedifferentiation of articular chondrocytes by serial subcultures.\nBiochim Biophys Acta\n1051\n266\n275\n2155668\n17. PittengerMFMackayAMBeckSCJaiswalRKDouglasR\n1999\nMultilineage potential of adult human mesenchymal stem cells.\nScience\n284\n143\n147\n10102814\n18. ShukunamiCShigenoCAtsumiTIshizekiKSuzukiF\n1996\nChondrogenic differentiation of clonal mouse embryonic cell line ATDC5 in vitro: differentiation-dependent gene expression of parathyroid hormone (PTH)/PTH-related peptide receptor.\nJ Cell Biol\n133\n457\n468\n8609176\n19. LefebvreVLiPde CrombruggheB\n1998\nA new long form of Sox5 (L-Sox5), Sox6 and Sox9 are coexpressed in chondrogenesis and cooperatively activate the type II collagen gene.\nEmbo J\n17\n5718\n5733\n9755172\n20. SekiyaITsujiKKoopmanPWatanabeHYamadaY\n2000\nSOX9 enhances aggrecan gene promoter/enhancer activity and is up-regulated by retinoic acid in a cartilage-derived cell line, TC6.\nJ Biol Chem\n275\n10738\n10744\n10753864\n21. KobayashiTSoegiartoDWYangYLanskeBSchipaniE\n2005\nIndian hedgehog stimulates periarticular chondrocyte differentiation to regulate growth plate length independently of PTHrP.\nJ Clin Invest\n115\n1734\n1742\n15951842\n22. JaysonMIJonesDE\n1971\nScleritis and rheumatoid arthritis.\nAnn Rheum Dis\n30\n343\n347\n4326763\n23. BarrCCDavisHCulbertsonWW\n1981\nRheumatoid scleritis.\nOphthalmology\n88\n1269\n1273\n7322474\n24. IsaakBLLiesegangTJMichetCJJr\n1986\nOcular and systemic findings in relapsing polychondritis.\nOphthalmology\n93\n681\n689\n3523358\n25. AnnunenSKorkkoJCzarnyMWarmanMLBrunnerHG\n1999\nSplicing mutations of 54-bp exons in the COL11A1 gene cause Marshall syndrome, but other mutations cause overlapping Marshall/Stickler phenotypes.\nAm J Hum Genet\n65\n974\n983\n10486316\n26. Van CampGSnoeckxRLHilgertNvan den EndeJFukuokaH\n2006\nA new autosomal recessive form of Stickler syndrome is caused by a mutation in the COL9A1 gene.\nAm J Hum Genet\n79\n449\n457\n16909383\n27. SharovAADudekulaDBKoMS\n2005\nA web-based tool for principal component and significance analysis of microarray data.\nBioinformatics\n21\n2548\n2549\n15734774\n28. SaeedAISharovVWhiteJLiJLiangW\n2003\nTM4: a free, open-source system for microarray data management and analysis.\nBiotechniques\n34\n374\n378\n12613259\n29. SugikiTUyamaTToyodaMMoriokaHKumeS\n2007\nHyaline cartilage formation and enchondral ossification modeled with KUM5 and OP9 chondroblasts.\nJ Cell Biochem\n100\n1240\n1254\n17115412\n30. SekiyaILarsonBLVuoristoJTRegerRLProckopDJ\n2005\nComparison of effect of BMP-2, -4, and -6 on in vitro cartilage formation of human adult stem cells from bone marrow stroma.\nCell Tissue Res\n320\n269\n276\n15778851\n31. MoritoTMunetaTHaraKJuYJMochizukiT\n2008\nSynovial fluid-derived mesenchymal stem cells increase after intra-articular ligament injury in humans.\nRheumatology (Oxford)\n47\n1137\n1143\n18390894\n32. SegawaYMunetaTMakinoMNimuraAMochizukiT\nin press\nMesenchymal stem cells derived from synovium, meniscus, anterior cruciate ligament, and articular chondrocytes share similar gene expression profiles.\nJournal of Orthopaedic Research\n33. TsurugaYKiyonoTMatsushitaMTakahashiTKasaiN\n2008\nEffect of intrasplenic transplantation of immortalized human hepatocytes in the treatment of acetaminophen-induced acute liver failure SCID mice.\nTransplant Proc\n40\n617\n619\n18374145\n34. MochizukiTMunetaTSakaguchiYNimuraAYokoyamaA\n2006\nHigher chondrogenic potential of fibrous synovium- and adipose synovium-derived cells compared with subcutaneous fat-derived cells: distinguishing properties of mesenchymal stem cells in humans.\nArthritis Rheum\n54\n843\n853\n16508965"
+}

batch_13/PMC2579739.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2579739",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2579739\nAUTHORS: Gian Luigi Mariottini, Elisabetta Giacco, Luigi Pane\n\nABSTRACT:\nThe toxicity of Cnidaria is a subject of concern due to its influence on humans. In particular, jellyfish blooms can highly affect human economical activities, such as bathing, fishery, tourism, etc., as well as the public health. Stinging structures of Cnidaria (nematocysts) produce remarkable effects on human skin, such as erythema, swelling, burning and vesicles, and at times further severe dermonecrotic, cardio- and neurotoxic effects, which are particularly dangerous in sensitive subjects. In several zones the toxicity of jellyfish is a very important health problem, thus it has stimulated the research on these organisms; to date toxicological research on Cnidarian venoms in the Mediterranean region is not well developed due to the weak poisonousness of venoms of jellyfish and anemones living in this area. In spite of this, during last decades several problems were also caused in the Mediterranean by stinging consequent to Cnidarian blooms mainly caused by Pelagia noctiluca (Forsskål, 1775) which is known to be the most venomous Mediterranean jellyfish. This paper reviews the knowledge on this jellyfish species, particularly considering its occurrence and toxicity.\n\nBODY:\nPelagia noctiluca: Distribution and BloomsThe mauve stinger Pelagia noctiluca (Cnidaria: Scyphozoa) is a small pelagic jellyfish generally pink-, mauve- or light brown-coloured, with a phosphorescent bell measuring 3 to 12 cm in diameter in adult specimens, whose edge is provided with lappets and tentacles; in this species the nematocysts stud the tentacles, the oral arms, as well as the upper surface of the bell [1]. Pelagia noctiluca has direct development, so its cycle do not comprise the benthic scyphistoma stage.This species has a wide distribution; in general, it is indicated as typical of warm waters, but due to currents it can enter temperate and cold seas, thus it can be found in tropical zones as well as in colder areas, such as in the north Atlantic and in the north Pacific; it seems to not go over 42°S and is vertically distributed mostly between 150 m depth and the surface [2], but during the day it can also be found between 300 and 500 m, with a maximum of 1,400 m [3].Pelagia noctiluca was collected in intertidal zones of the Pacific Ocean along the Californian coast [4] and is widely distributed in the Mediterranean Sea and in some zones of the Atlantic Ocean. In Atlantic waters it is frequent along the Southern Atlantic French coast in summer [5]; it was observed repeatedly in swarms around the British Islands [6], and in the North Sea and also, although infrequently, off the Dutch coast [7, 8]. Strandings of Pelagia noctiluca were observed in 1988 on beaches of the western coast of Ireland [9]. The occurrence of Pelagia noctiluca in the Mid Atlantic Sea, West to Ireland, was recorded recently [10]. This jellyfish is not considered a normal resident of the North Sea and small records were made during ‘70s off the Scottish north and east coasts and in the area of Shetland and west of the Norwegian Deep; in spite of this, in 1982 Pelagia noctiluca were more abundant than usual, presumptively carried in by currents [11].In the Mediterranean Sea swarms of Pelagia noctiluca occur usually in pelagic waters from March to May; in summer months, from June and August, isolated big specimens can also appear [12]. In the Southwestern Mediterranean this Cnidarian was indicated as the most frequent jellyfish along Tunisian coasts, mostly in autumn and winter; in this area the presence of Pelagia noctiluca depends on local and particular winds and currents [13]. Anyhow, this species was signalized as uncommon along Tunisian coasts [14, 15] and in front of Algeria [16].In the Mediterranean region the studies on jellyfish had an increase during late 70’s and early 80’s of the past century, when all the Mediterranean basin was affected by an abnormal proliferation of such organisms [17, 18]. This proliferation was mainly due to Pelagia noctiluca, which from the irritant/toxic point of view, is also the most important Mediterranean jellyfish. In that years the proliferation of other species, such as Rhizostoma pulmo, Cotylorhiza tuberculata, Aurelia aurita and Chrysaora hysoscella, was recorded at a lesser extent and caused less health problems because of their scant poisonousness.The Pelagia noctiluca bloom started in the Eastern Basin and in the Adriatic Sea (the latter was overall the most interested by the phenomenon) and subsequently spread to the Western Basin, although with less intensity. The main period concerned was 1981–1984; then, as suddenly it arose, the phenomenon finished quickly, even though sporadic proliferations were recorded in subsequent years.In the Eastern Mediterranean the studies concerning the distribution of Pelagia noctiluca and of other jellyfish were carried out mainly during summer [19]; some scientists performed research on the occurrence of this jellyfish in Greek and Egyptian waters [20], along Turkey [21, 22] and Lebanese coasts [23], and lessepsian jellyfish species coming from the Red Sea were observed in Israeli waters [24, 25]. Along Maltese coasts the studies on the distribution of Pelagia noctiluca were carried out from 1980 to 1986 [26].In the Adriatic Sea Pelagia noctiluca caused several problems to bathers and sea-workers and its distribution was studied along Croat coasts during the summer of 1983 [27] and seasonally in 1985 [28], and along Italian coasts during the period 1976–1986, mainly during the summer months, when its occurrence and seasonal variations were related with the variations of environmental factors [20, 29–32]. The influence of water temperature on swimming behaviour of Pelagia noctiluca was emphasized [33]; this factor could account for coastal aggregations occurring in the Adriatic coasts in summer. After research performed from 1976 to 1983, according to [34] the highest densities of Pelagia noctiluca in the Adriatic Sea were recorded in pelagic waters with high salinity and low nutrients. It was asserted that the dense aggregations of this organism in coastal shallow waters were wind- current- and tide-caused [35].In the Tyrrhenian Sea and in the Ligurian Sea the consequences of the bloom were remarkably lower than in other Mediterranean zones; nevertheless, notable alterations in comparison with the normal situation were recorded everywhere from 1982 to 1988 [20, 36–42]. Furthermore, in some receptive coastal zones, such as in the Spotorno Bay (Western Ligurian Riviera), the proliferation of Pelagia noctiluca was evident and reached a peak during some periods (September 1984 – January 1985) with no apparent regard to the season [43]. Archived research containing data on the occurrence of Pelagia noctiluca during a 200 year-period showed that in the zone of Villefranche-sur-Mer (western Ligurian Sea) there were ‘years with Pelagia noctiluca’ and ‘years without Pelagia noctiluca’ every about 12 years; this could indicate the cyclic nature of the phenomenon [44]. During the ‘bloom years’, and particularly during the summer of 1983, considerable proliferation of Pelagia noctiluca was recorded also in Sicily and Sardinia Channels [20].On the whole, salinity was indicated to exert an influence on jellyfish behaviour and distribution as evidenced in field observations [45]. An univocal explanation of the jellyfish bloom and of the subsequent jellyfish decrease was not supplied and several scientists provided different explanations for the blooms, but it seemed acceptable that neither hydrologic changes (as water circulation didn't show clear alterations during last decades) nor pollutant inputs supported jellyfish proliferation. It was suggested that coastal aggregation of jellyfish could be due to wind action [26], to natural [41] or cyclic [42] fluctuations of jellyfish populations or to water movements [46]. Furthermore, a relationship between jellyfish proliferation and environmental factors [23, 28] or water pollution [47, 48] was suggested. Also ‘hormesis’, which involves stimulation of growth by low concentrations of toxicants [49] and evolution processes connected to the appearance of favourable mutations [50] were suggested to induce the proliferation. Recently, the cumulative effects of man-caused and climate changes has been indicated to be involved in the mechanisms which can promote the increasing of jellyfish occurrence [51, 52]. In particular, it was stated that as Cnidaria feed high on marine food chains and therefore they can compete with fishes for food, massive removal of top-predator fishes by commercial fishing efforts could open up food resources for jellyfish [51]. In addition, models indicated that the predicted pH decrease in oceans with rising CO2 could induce long-term jellyfish increase over the next 100 years [52].Jellyfish blooms have also stimulated research on the biological, ecological and chemical aspects of Pelagia noctiluca, which has been carried out in order to evaluate its role in the marine ecosystem and in the food web. The biological cycle of Pelagia noctiluca is annual [3]; for this reason it shows high natural mortality, typical of short life-cycle species [32]. Its reproductive period was also extensively studied and it was observed that Pelagia noctiluca reproduces throughout the year [53]. In studies concerning nutrition the gastrovascular content was examined and the diet of this organism was shown to be composed mainly (around 90%) by cladocerans and copepods [54]; other indications about the behaviour and the trophic ecology of Pelagia noctiluca were provided [55], observing that at 16–19°C Pelagia noctiluca maintenance and growth requires approximately 3 mg C/day indicating a substantial predatory impact of jellyfish when they are in aggregations.In studies concerning the lipid composition of Pelagia noctiluca [56] it was observed that total lipids are 0.19 % of wet weight; within lipids these Authors found 73.8 % of neutral lipids and 26.2 % of phospholipids. Among neutral lipids mainly sterol esters (11.3 %), triglycerides (20.7 %), free fatty acids (56.6 %) and sterols (7.6 %) occur. Among phospholipids, phosphatidylcholine was particularly abundant (36 % was recorded). A high amount of free fatty acids (41.7 %) was found in the total lipid fraction; these results were comparable to data obtained from other jellyfish (Rhizostoma pulmo and Cotylorhiza tuberculata) collected in the Ligurian Sea [57]. Proteins in Pelagia noctiluca range from 10.9% to 19.8% of dry weight and lipids, expressed as percent of dry weight, range from 1.3% to 2.9%; only 0.2–0.3% of dry weight consisted of phosphorus [58]. Preliminary results of HPLC analyses on Pelagia noctiluca crude extracts partially separated by gradient density showed that a noticeable amount of the extract was of protein nature; this result was also confirmed by protein analyses [59].Analyses of trace elements carried out on tentacles of Pelagia noctiluca showed that metal content does not differ significantly in different periods [60] and zinc amount was 46 μg/g dry weight in specimens collected in NW Mediterranean [61]. C and N composition of Pelagia noctiluca ephyrulae was experimentally evaluated as % of dry weight: values of 11.4 and 3.3 for C and N, respectively, and a C:N ratio of 4.2 were found [62]. C isotope analysis showed values of −18.51±1.12 δ13C‰ [63, 64]. Sphingophosphonolipid composition of Pelagia noctiluca was recently defined and a suite of two ceramide 2-aminoethylphosphonic acids (CAEP) was quantified at 2.0 and 1.3% of phospholipids [65]. The pH of body fluids taken from Pelagia noctiluca collected in the intertidal zone from Laguna Beach (California) was also performed; the results showed a pH ranging from 7.304 to 7.307 [4].Nematocyst MorphologyThe morphology of nematocysts of Pelagia noctiluca was observed during early twentieth century [66, 67]; subsequently [6] three types of nematocysts were described: heterotrichous microbasic eurytele, holotrichous isorhiza and atrichous isorhiza [68]. Subsequently two predominant types of nematocysts were referred to be found on Pelagia tentacles: a round holotrichous isorhiza (20–25 μm diameter) containing multiple spiralled barbed nema inside and a microbasic mastigophore, approximately one-fourth the size the former [69]. On the whole, P. noctiluca holotrichous isorhiza are longer than similar nematocysts of other Scyphozoan medusae such as Catostylus mosaicus and Phyllorhiza punctata [70].More recent studies showed five morphological types that have been differentiated according to the classification of Mariscal [71, 72]: heterotrichous microbasic eurytele, holotrichous isorhiza haploneme (type I), atrichous isorhiza haploneme (type I), holotrichous isorhiza haploneme (type II) and atrichous isorhiza haploneme (type II).Subsequently, these types were better defined by ultrastructural studies [68] as follows: heterotrichous microbasic eurytele, holotrichous O-isorhiza, heterotrichous isorhiza, atrichous a-isorhiza and another type resembling the microbasic p-mastigophore.Recently, the morphology of Pelagia noctiluca nematocysts was re-examined [59] and the nematocysts found were separated in three groups including different morphological types: Group 1: great-spherical nematocysts corresponding to the holotrichous isorhiza and large atrichous isorhiza, that previously were stated to be undistinguishable when the capsule is undischarged [68]; Group 2: smaller and elliptical nematocysts corresponding to the heterotrichous microbasic eurytele and the heterotrichous isorhiza [68]; Group 3: smallest elliptical nematocysts with protruded operculum, similar to group 2 nematocysts, excepting the size.The nematocysts of Pelagia noctiluca can be found on umbrella, oral arms and tentacles; thus all the jellyfish is venomous and discharge can be induced, aside from encounters with the living animal, even from handling stranded or dead organisms [73].Nematocyst DischargeThe nematocysts of Pelagia noctiluca can be maintained isolated in distilled water where they retain their discharging capacity [74]. The discharge of isolated nematocysts can be induced by treatment with trypsin, while Ca2+ ions prevent discharge [74]. Extreme values of pH (<2 and >11) promote discharging of isolated nematocysts; furthermore, aqueous solutions at pH 1.0–3.5 cause collapse of the capsular wall of undischarged nematocysts [75]. The nematocysts of Pelagia noctiluca may be triggered by vinegar [76].Nematocyst discharge in Pelagia noctiluca, as in other Cnidaria such as in some Anthozoa [77, 78], is a Ca2+-dependent process [79]. Previous treatment of oral arms with La3+ inhibits nematocyst discharge and a similar result was obtained by treatment with Gd3+, a powerful blocker of mechanosensitive ion channels; therefore, nematocyte activation seems to be Ca2+ dependent and Ca2+ permeable mechanosensitive channels were demonstrated to be involved in the activation of nematocytes [79]. The anions, such as I−, Cl−, SO42− play an important role to promote discharge of nematocysts. In particular, I− was the most effective and discharging activity decreased by using Cl− and was very low with SO42−; from these results it was stated that ion effect is probably exerted on protein conformation of the capsular wall/fluid [80]. The hypothesis that a conformational change of a capsular protein is involved in the discharging process was confirmed also by studies which evaluated the discharging effectiveness of Hofmeister anions, in which SO42− showed the lowest discharging potency and SCN− the highest one [81]; nematocyst discharge can be induced also by thioglycolate with reduction of –S–S– bridges [82].Jellyfish Stings and EpidemiologyIt is well known that jellyfish stings can induce both local and general symptoms [83] and sometimes can be lethal to humans [84]. In general, though, jellyfish stings usually cause a mild local dermatitis; so serious or fatal systemic reactions are uncommon [85]. After envenomation some neuromuscular manifestations such as localised neuropathy and mononeuritis multiplex [86–88] as well as neurological manifestations such as delirium, stupor, central respiratory failure and muscular weakness have been reported [85]. Guillain-Barré syndrome was also described, presumably a consequence of an aberrant immune response [89].Pelagia noctiluca stings are usually limited to the skin surface and cause only erythematous, edematous, and vesicular topical lesions [90, 91], with local pain which persists for 1–2 weeks [91], while systemic complications or cutaneous infections are infrequent [92].Nevertheless, dramatic immediate reactions have been observed after Pelagia noctiluca stings, even though they are rarely severe and prolonged; the lesions appear circinate or irregularly shaped, the venom is also able to cause severe generalized allergy with bronchospasm, pruritus and postinflammatory hyperpigmentation [93]. Immediate pain, distress, occurrence of urticaria-like lesions and dyspnoea after massive stingings were reported [94]. Serious consequences may be produced by immunological and/or toxic mechanisms [95, 96]. Pelagia noctiluca stings can leave scars and hyperpigmentation; the persistence of pigmentation can result from tattooing of the mauve stinger pigment into the skin or from post-inflammatory events [83]. Hyperpigmentation can remain for some years after envenomation causing aestethical problems [73, 92]. It was reported that jellyfish probably identificated as Pelagia noctiluca caused recurrent and more severe cutaneous eruptions over ten days after envenomation [97].Stinging by Pelagia noctiluca can provoke relapse of the eruption after some years, also without further contact with jellyfish; thus was hypothesized the venom could react with dermal collagen and produce an active antigen which stimulates the immunological response [91]. Cross-reactions against venom of Physalia physalis was observed in patients showing significant titers of IgG against crude extract of nematocysts of Pelagia noctiluca [98]. Anyhow, the seriousness of envenomation is due to the eventual allergic characteristics of the patient or to previous envenomations [73].The jellyfish bloom gave rise to epidemiological studies, which were carried out mostly in the Adriatic area [99–101] and in the Eastern Mediterranean region [102]. In the Northern Adriatic during 1983 67 subjects suffered from jellyfish stings with pain, swelling and erythema along the littoral of Portorož (Slovenia); most of injuries were local, on arms and chest, and limited to small areas and systemic reactions were observed in three cases [99]. In the zone of Trieste (Italy) from 1978 to 1983 127 subjects come into contact with Pelagia noctiluca and suffered from skin and systemic injuries mainly with local symptoms; only five cases, whose atopy was ascertained through family and personal history, had more severe general symptoms [103]. In Pula (Croatia) 52% of bathers were stung by jellyfish during summer 1978, some of them several times [100, 101]. In Greek waters 762 subjects, mainly adults, required medical advice from 1981 to 1984, particularly in July and August during the swimming period; 82.8% of the cases were observed in the coastal area of Attica and most of them (92.3%) suffered local symptoms (redness, pain, itching, burning, vesicles), 7.7% had general symptoms (dizziness, vomiting, fall of blood pressure and diarrhea, one case shock) [102]. On the whole in Adriatic coastal zones bathing was significantly influenced by the bloom [101]. Otherwise, a scarce incidence of dermatitis caused by Pelagia noctiluca was emphasized in the Ligurian region where the observed cases were of scarce clinical relevance: the available data, concerning only stung children from 1984 to 1989 indicate only 20 cases were observed and none of them was hospitalized [40, 104].Nematocyst Isolation and Toxicity of the VenomThe problem of the separation of nematocysts from tissue material has been dealt with by several scientists. As already reported for other Cnidaria [72, 104–110], in the case of Pelagia noctiluca the toxicity is not exclusively due to nematocysts, but is also ascribable to tissue components [111].An attempt to obtain pure suspensions of nematocysts from Pelagia noctiluca was made by soaking of tissues in distilled water and subsequent centrifugations in saccarose solution at 4°C with lyophylisation of the obtained undischarged nematocysts which can thus maintain their discharging capacity when put at room temperature [112].Recent laboratory data showed the nematocysts of Pelagia noctiluca can be separated by centrifugation using the discontinuous density gradient of Percoll [59]. Furthermore, nematocysts were recently isolated by chemical (treating them with SCN−) and physical methods (heat dissociation), and it was observed that heat dissociation promotes the release of nematocysts from tissues, but causes damage of nematocytes, so, it is scarcely useful; on the other hand, treatment of tentacles with SCN− yields 90% of intact nematocytes [113]. In addition, the preservation of the nematocysts of Pelagia noctiluca depends on maintenance method and particularly on temperature (freezing) and pH (neutral values are optimal) [114].In other studies the nematocysts from marginal tentacles of Pelagia noctiluca were isolated and partially purified; then lyophilised fractions of homogenate preparations of nematocysts were fractionated on Sephadex G-75 columns, the molecular weight was determined and fractions were tested to evaluate the effect on heart activity of rats and on neuromuscular activity of frogs. The results showed an evident toxic activity of unfractioned pools on neuromuscular synapses and less myocardial effects with only few variations of cardiac frequency [72].Preparations of intact nematocysts were tested on hairless mice and on human skin to evaluate their irritant effects: mice were seen to develop erytema and papules after contact with the preparation; intradermal injection produced erytema, oedema, leukocyte infiltrate and nodular lesions with central necrosis [115]. On human skin the most irritant effects were obtained with the scratch test which showed the occurrence of erytema after 30 minutes accompanied by pruritus in less than 50% of cases; these symptoms decreased progressively after 48 and 72 hours [90].A partial purification of a cardiotoxin in crude Pelagia venom was obtained using anti-Chrysaora or anti-Physalia monoclonal antibody–Sepharose columns. In these experiments protein bands with molecular weights of 54,000, 92,000, 130,000 and 150,000 were revealed and was also reported that both crude and partially purified Pelagia venom contained active fractions against cultured chick embryo cardiocytes [116].The venom of Pelagia noctiluca is of protein nature and contains peptides; it is antigenic and possesses dermonecrotic and hemolytic properties; electrophoretical analyses recognized eight different fractions, distinguished by molecular mass [101].The capsule fluid and the capsule wall of the nematocysts of Pelagia noctiluca were widely studied from the point of view of their protein content: it was observed that in both structures glutamic acid is the most frequent aminoacid (80% in the proteins of the capsule fluid and 90% in that of the capsule wall); these glutamate-rich proteins are probably stabilized by Ca2+; discharging agents (lyotropic anions, proteolytic enzymes, Ca2+ chelating agents) could act on capsule wall proteins inducing their conformational change [117].As referred above, a cross-reactivity between Pelagia venom and monoclonal antibodies to Physalia and Chrysaora venoms [116] was reported; this aspect was clinically verified by the release of histamine after exposure to Chrysaora venom by basophils from a patient who had clinical anaphylaxis after a Pelagia sting [96]. So, also clinical evidences could demonstrate the cross-reactivity between the venoms of these jellyfish; Pelagia venom has got more antigenic potential for man than several other jellyfish venoms [69].The cytotoxic properties of Pelagia noctiluca crude venom have been experimentally assessed by short-term [111] and long-term [104] tests on cultured cells by trypan blue dye exclusion, neutral red, colony forming efficiency and genotoxicity assay. Crude P. noctiluca venom highly affected cultured cells, producing severe survival decrease but, despite its well-known in vivo irritating properties, it showed lower effects in comparison to that evidenced by venoms of other jellyfish and anemones apparently less venomous in vivo [107–110]. Anyhow, the venom of Pelagia noctiluca showed remarkable cytotoxicity, and killed all treated cells at highest tested concentration within two hours. The protein nature of venom [96] was further confirmed by the absence of effects on DNA of treated cells [111]. Pelagia noctiluca venom caused also an increase of ATP levels in treated cells within 1 hour of treatment and a following moderate decrease [111]; this is a strange behaviour, because in general toxicity studies record decrease of ATP in stress-exposed cells [118, 119] and organisms [120]. Also long term cell proliferation tests showed that the venom of Pelagia noctiluca has less effect on cells than venom of other jellyfish [104].The hemolytic properties of the crude venom of Pelagia noctiluca were recently assessed on fish, chicken, rabbit and human erythrocytes; results showed a significant hemolysis of chicken and rabbit erythrocytes and a good resistance of fish ones. It was also observed that crude venom maintains its hemolytic properties even after freezing at −20°C and −80°C and lyophilization [121].ConclusionsDistribution and bloomA satisfactory explanation of the bloom of Pelagia noctiluca in the Mediterranean, and partly in Atlantic waters has not been provided, even though several causes have been indicated as responsible of the phenomenon; a number of scientists have supposed that it could be the result of natural cyclic fluctuations, already described in several species, correlated with environmental or trophic factors [23, 28, 36, 41, 42, 46]. On the other hand, lack of alteration in lipid content of jellyfish during [57] and after the bloom also prevented scientists from hypothesizing about the implications of food quantity and quality in supporting the phenomenon.On the whole, outbreaks of Pelagia noctiluca can have an important ecological impact on plankton dynamics and on trophic relationships; this was seen particularly in the Adriatic Sea, where the bloom showed its maximum [122]; as a matter of fact, events of massive occurrence of planktonic organisms could cause alterations of community structures and functioning as well as lack of biodiversity [123]. As Pelagia noctiluca is a top predator and its feeding activity is exerted on several zooplankters, including eggs and larvae of nektonic and benthic organisms, this impact could have affected remarkably prey populations and consequently caused lower production [122]. In this connection, from the biological and ecological point of view the chemical analyses carried out on Pelagia noctiluca specimens [62, 64, 65] are important in order to evaluate the transfer of energy to high trophic levels as well as the amount of C and N available for microorganisms in consequence of the decomposition, which becomes more intense during outbreaks.Nematocyst morphology and dischargeThe nematocysts of Pelagia noctiluca can be found on the umbrella, oral arms and tentacles; they have been classified into five morphological types [68, 72] and recently they were separated into three groups [59]. It is known that the activation of nematocysts is calcium dependent and discharge is promoted by anions [75] and inhibited by lantanium and gadolinium [79]. Furthermore, the venom of this jellyfish is cytotoxic [111] and studies performed on cell cultures, that are an efficient alternative method to the utilization of living organisms, showed also cell growth decrease after treatment.At a cellular level it was suggested that Cnidarian venoms affect the plasmalemma by binding to membrane phospholipids and increasing permeability with consequent water uptake and damage of external and intracellular membranes [124]; the phospholipase activity of Cnidarian venoms, which affects cell membrane permeability and ion exchange, was also suggested [125–127].Toxicity and epidemiologyPelagia noctiluca has not caused human fatalities, but in spite of this it can be a nuisance and a health and economical problem when it appears in huge numbers during outbreaks. As a matter of fact, the contact with all body portions – bell, tentacles and oral arms – causes in humans local pain, burning, swelling, hyperpigmentation and other local symptoms; repeated contacts with this jellyfish can produce recurrent skin eruptions, which were also observed without further stingings.During the bloom the epidemiological studies on Pelagia noctiluca stings were carried out mostly both in Italian and Croatian Adriatic localities [99–101] owing to the high impact on bathers and fishermen in these zones, while data from other Mediterranean regions are sporadic [40, 102, 104]. On the whole, data concerning the impact on human health are scarce and not many subjects come under medical observation, even though presumptively several cases have not been seen in medical first-aid stations.Lack of knowledgeA lot remains unknown, both about the ecology and toxicity of Pelagia noctiluca and several aspects remain to be clarified to explain its complex biological and ecological role in the marine environment. The causes and the dynamics of blooms in relation to the variations of environmental factors, to climate and to man-caused changes are to date greatly unknown; in this connection, the alterations induced on marine food chains [51] and the consequences of CO2 rising on pH of oceans, which was seen to be correlated with jellyfish frequency [52], are stimulating research subjects.From the toxicological point of view, further research is also required to better separate the nematocyst content from tissue components of Pelagia noctiluca, in order to characterize the toxic compounds which produce the clinical symptoms and the observed cytotoxicity in vitro. It should be also important to explain how venom acts against cells and to clarify the mechanisms which damage biomembranes and alter membrane permeability.Finally, as several substances extracted from marine organisms have bioactive and pharmacological properties it should be interesting and useful to know if compounds with such characteristics occur in Pelagia noctiluca too.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2579922.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2579922",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2579922\nAUTHORS: Steven G Ralph, Hye Jung E Chun, Natalia Kolosova, Dawn Cooper, Claire Oddy, Carol E Ritland, Robert Kirkpatrick, Richard Moore, Sarah Barber, Robert A Holt, Steven JM Jones, Marco A Marra, Carl J Douglas, Kermit Ritland, Jörg Bohlmann\n\nABSTRACT:\nBackgroundMembers of the pine family (Pinaceae), especially species of spruce (Picea spp.) and pine (Pinus spp.), dominate many of the world's temperate and boreal forests. These conifer forests are of critical importance for global ecosystem stability and biodiversity. They also provide the majority of the world's wood and fiber supply and serve as a renewable resource for other industrial biomaterials. In contrast to angiosperms, functional and comparative genomics research on conifers, or other gymnosperms, is limited by the lack of a relevant reference genome sequence. Sequence-finished full-length (FL)cDNAs and large collections of expressed sequence tags (ESTs) are essential for gene discovery, functional genomics, and for future efforts of conifer genome annotation.ResultsAs part of a conifer genomics program to characterize defense against insects and adaptation to local environments, and to discover genes for the production of biomaterials, we developed 20 standard, normalized or full-length enriched cDNA libraries from Sitka spruce (P. sitchensis), white spruce (P. glauca), and interior spruce (P. glauca-engelmannii complex). We sequenced and analyzed 206,875 3'- or 5'-end ESTs from these libraries, and developed a resource of 6,464 high-quality sequence-finished FLcDNAs from Sitka spruce. Clustering and assembly of 147,146 3'-end ESTs resulted in 19,941 contigs and 26,804 singletons, representing 46,745 putative unique transcripts (PUTs). The 6,464 FLcDNAs were all obtained from a single Sitka spruce genotype and represent 5,718 PUTs.ConclusionThis paper provides detailed annotation and quality assessment of a large EST and FLcDNA resource for spruce. The 6,464 Sitka spruce FLcDNAs represent the third largest sequence-verified FLcDNA resource for any plant species, behind only rice (Oryza sativa) and Arabidopsis (Arabidopsis thaliana), and the only substantial FLcDNA resource for a gymnosperm. Our emphasis on capturing FLcDNAs and ESTs from cDNA libraries representing herbivore-, wound- or elicitor-treated induced spruce tissues, along with incorporating normalization to capture rare transcripts, resulted in a rich resource for functional genomics and proteomics studies. Sequence comparisons against five plant genomes and the non-redundant GenBank protein database revealed that a substantial number of spruce transcripts have no obvious similarity to known angiosperm gene sequences. Opportunities for future applications of the sequence and clone resources for comparative and functional genomics are discussed.\n\nBODY:\nBackgroundConifers (members of the pine family) have very large genomes (10 to 40 Gb, [1]), and this poses difficulties for both structural and functional genomic studies. In addition, their generation times are long and their habitual out-breeding nature prevents the development of inbred strains useful for genetics research. A further difficulty in conifer genomics is the large evolutionary distance between conifers and angiosperms (i.e., flowering plants), separated by 300 million years of evolution [2], which severely restricts gene comparisons of conifers with angiosperms. While there are several completely sequenced angiosperm genomes, as well as high-quality sequence-finished full-length (FL)cDNA resources, for Arabidopsis [3,4], rice [5-7], poplar (Populus trichocarpa; [8,9]), grapevine (Vitis vinifera; [10]), and a moss (Physcomitrella patens; [11]), these basic genomics resources have not yet been developed for the conifer phyla or for any other gymnosperm.In species with large genomes, a critical first step for genome characterization is to survey the expressed genes. A common approach to characterize the expressed genome is to sequence cDNA libraries and to assemble large collections of expressed sequence tags (ESTs) [12]. In the absence of a conifer genome sequence, large and deep EST collections are particularly useful. Sequencing of cDNA libraries constructed from diverse tissues and developmental stages, and from materials subjected to diverse environmental conditions or treatments, enhances the diversity of genes captured in EST populations. In addition, normalization techniques reduce the frequency of highly expressed genes and increase the rate of rare gene discovery [13,14], thus providing more comprehensive coverage of the expressed genome.In conifers, gene discovery via EST sequencing was first conducted in loblolly pine (Pinus taeda; [15]), the most economically important tree species in the southeastern USA. The early emphasis in loblolly pine was on wood forming tissues [16], but newer projects have involved treatments such as drought stress [17] and embryogenesis [18]. As of May 2008, the loblolly pine EST collection contains more than 328,000 sequences [19]. Recent EST projects with species of spruce have used tissues related to shoot growth and xylem development in white spruce [20,21], wound treatment in interior spruce [21], root development in Sitka spruce [21], and xylem development and bud burst in Norway spruce (P. abies; [22,23]). EST resources have also been developed for a few other gymnosperm species outside of the pine family, such as cycas (Cycas rumphii; [24]), ginkgo (Ginkgo biloba; [25]), Japanese yew (Taxus cuspidata; [26]), Japanese cedar (Cryptomeria japonica; [27,28]) and Hinoki cypress (Chamaecyparis obtusa; [28]).In addition to deep EST sampling, other important components of a cDNA sequence resource are the quality and length of sequence coverage for a given gene. Ideally, FLcDNA clones that capture the entire mature transcript of a gene should be identified and completely sequenced with high accuracy. FLcDNA sequences should span not only the protein-coding open reading frame (ORF) region but also the non-coding 5' and 3' untranslated regions (UTRs). Most importantly, true FLcDNA sequences should be derived from a single individual FLcDNA clone. Using individual clones prevents the assembly of chimeric FLcDNA sequences consisting of ESTs from multiple cDNA clones representing closely related genes. Furthermore, allelic nucleotide polymorphisms and alternatively spliced variants of a gene are difficult to detect using in silico assembled sequence contigs from multiple clones. To further discriminate among closely related genes, the authenticity of sequences should be verified by re-sequencing of the same clone (sequence verification). Compared to single-pass ESTs or in silico assembled sequence contigs originating from multiple clones, sequence-verified FLcDNA clones offer several advantages for comparative, structural, and functional genome analyses, in particular for conifers with their great evolutionary distance from angiosperms. First, the complete protein-coding regions of FLcDNAs can be unambiguously identified. An accurate prediction of full-length protein sequences aides in the correct identification of distant angiosperm homologues. Second, in anticipation of a future conifer genome sequence, FLcDNAs can be used to improve gene prediction from genomic sequences as demonstrated in Arabidopsis [29-31] and poplar [8,9]. Third, FLcDNA clones can be used for functional characterization of conifer genes using biochemical approaches [e.g., [32,33]] or for functional complementation of mutants in heterologous systems. Given the lack of knock-out mutants in conifers and the slow process of generating knock-down mutants in conifers, biochemical approaches and heterologous complementation that rely on FLcDNA clones are essential tools for functional genomics in conifers. Finally, FLcDNAs can be used to accurately identify peptides in large-scale conifer proteome analyses [34,35].Despite their immense value, sequence-verified FLcDNA clones have not been generated in most plant species subjected to genome analysis. Only a few resources of large and sequence-verified FLcDNA data sets have been generated for angiosperm plant species; namely, for Arabidopsis [4], rice [7], and poplar [9]. In contrast, no substantial FLcDNA resource has been reported for a conifer or any other gymnosperm species. The Conifer Forest Health genomics project \"Treenomix\" [36] aims to develop genomic resources for spruce, characterize mechanisms of resistance against insect pests and adaptation to local environments, and identify genes for the formation of oleoresin-based terpenoid biomaterials [37-43]. Here, we report on a comprehensive spruce EST and FLcDNA resource and discuss its utility for conifer genomics. A total of 206,875 ESTs were obtained by sequencing 20 standard, normalized or full-length cDNA libraries derived from Sitka spruce, white spruce, and interior spruce. Analysis of ESTs identified 46,745 putative unique transcripts (PUTs). We describe advantages covered by the first large set of 6,464 sequence-verified, high-quality FLcDNAs obtained from a single clonally propagated tree of Sitka spruce.ResultsSequencing and assembly of spruce ESTsWe constructed 20 unidirectional standard, normalized or full-length enriched cDNA libraries from various tissues, developmental stages, and stress treatments of Sitka spruce, white spruce and interior spruce (Table 1). Several libraries were made from trees subjected to insect feeding by white pine weevils (Pissodes strobi) or spruce budworms (Choristoneura occidentalis), or to herbivory-simulation treatments such as mechanical wounding or methyl jasmonate application. From these libraries, we obtained 206,875 EST sequences, consisting of 165,403 3'-end EST sequences and 41,472 5'-end EST sequences (Table 2). We initially focused on 3'-end sequencing. Subsequent sequence reads from 5'-ends were performed as paired end reads, primarily from clones derived from FLcDNA libraries, to support the identification of a non-redundant FLcDNA set for complete insert sequencing. Removing low-quality and vector sequences (see Table 2 for criteria), as well as any obvious contaminant sequences, provided a database containing 147,146 high-quality (hq) 3' ESTs (88.9% success rate) with an average read length of 656 bp (Table 2). When we analyzed the 147,146 hq 3'-end ESTs using the CAP3 program ([44]; assembly criteria: 95% identity, 40 bp window), 120,342 ESTs assembled into 19,941 contigs and the remaining 26,804 ESTs were classified as singletons, suggesting a combined total of 46,745 PUTs across Sitka spruce, white spruce and interior spruce (Table 2). On average, contigs contained six assembled EST sequences. Only 88 contigs consisted of greater than 50 ESTs. The five largest contigs contain 618 (aspartyl protease), 229 (ribulose biphosphate carboxylase small subunit), 222 (metallothionein), 209 (translationally controlled tumor protein) and 172 (no significant match) ESTs. The proportion of EST sequences from organelles was small. Known and putative mitochondrial and chloroplast sequences contribute only 285 (0.19%) and 787 (0.53%) ESTs to the entire data set, respectively. In separate species-specific assemblies using ESTs from only white spruce or Sitka spruce, we identified 23,963 PUTs (72,649 3'-end EST sequences, 10,948 contigs and 13,015 singletons) and 17,988 PUTs (49,198 3'-end EST sequences, 6,918 contigs and 11,070 singletons), respectively.Table 1Libraries, tissue sources and spruce species for sequences described in this studycDNA LibraryTissue/Developmental stageSpecies (genotype)WS-ES-A-1aYoung shoots harvested from 25-year old treesd.P. glauca (PG-29)WS-PS-A-2aFlushing buds, young shoots and mature shoots harvested from 25-year old treesd.P. glauca (PG-29)WS-X-A-3aEarly (June 15th), mid (July 10th) and late (August 17th) season outer xylem harvested from 25-year old treesd.P. glauca (PG-29)IS-B-A-4aBark tissue (with phloem and cambium) harvested after razor blade wounding and treatment with 0.01% methyl jasmonate. Tissue was collected 0 (untreated), 3, 6 and 12 h post-treatmente.P. glauca × P. engelmannii (Fal-1028)SS-R-A-5aYoung growth (terminal 1–3 cm) and mature growth (distal to terminal 1–3 cm) rootse.P. sitchensis (Gb2-229)WS-PP-A-6aEarly (June 15th), mid (July 10th) and late (August 17th) season phloem harvested from 25-year old treesd.P. glauca (PG-29)IS-B-A-7aBark tissue (with phloem and cambium) harvested after razor blade wounding and treatment with 0.01% methyl jasmonate. Tissue was collected 24 h, 2 d, 4 d and 8 d post-treatmente.P. glauca × P. engelmannii (Fal-1028)WS-PS-N-A-8bFlushing buds, young shoots and mature shoots harvested from 25-year old treesd.P. glauca (PG-29)WS-X-N-A-9bEarly (June 15th), mid (July 10th) and late (August 17th) season outer xylem harvested from 25-year old treesd.P. glauca (PG-29)IS-B-N-A-10bBark tissue (with phloem and cambium attached) harvested after razor blade wounding and treatment with 0.01% methyl jasmonate. Tissue was collected 0 h (untreated), 3 h, 6 h, 12 h, 24 h, 2 d, 4 d and 8 d post-treatmente.P. glauca × P. engelmannii (Fal-1028)SS-R-N-A-11bYoung growth (terminal 1–3 cm) and mature growth (distal to terminal 1–3 cm) rootse.P. sitchensis (Gb2-229)WS-PP-N-A-12bEarly (June 15th), mid (July 10th) and late (August 17th) season phloem harvested from 25-year old treesd.P. glauca (PG-29)SS-IB-A-FL-13cBark tissue (with phloem and cambium attached) harvested after continuous feeding by Pissodes strobi weevils. Tissue was collected 2, 6 and 48 h post-treatmente.P. sitchensis (FB3-425)SS-IL-A-FL-14cGreen portion of leader tissue harvested after continuous feeding by Choristoneura occidentalis budworms. Tissue was collected 3 h, 6 h, 12 h, 24 h, 52 h, 4 d, 6 d, 8 d and 10 d post-treatmente.P. sitchensis (FB3-425)SS-IB-A-FL-15cBark tissue (with phloem and cambium attached) harvested after continuous feeding by P. strobi weevils. Tissue was collected 2, 6 and 48 h post-treatmente.P. sitchensis (FB3-425)WS-SE-A-16aSomatic embryo tissue harvested at the callus stage, and after 2, 4 and 6 weeks of growth on media supplemented with abscisic acid and indole-3-butyric acid.P. glauca (I-1026)WS-MC-A-17aCones harvested from 25-year old treesdP. glauca (11)WS-SE-N-A-18bSomatic embryo tissue harvested at the callus stage, and after 2, 4 and 6 weeks of growth on media supplemented with abscisic acid and indole-3-butyric acid.P. glauca (I-1026)WS-SE-N-A-19bSomatic embryo tissue harvested at the callus stage, and after 2, 4 and 6 weeks of growth on media supplemented with abscisic acid and indole-3-butyric acid.P. glauca (I-1026)WS-MC-N-A-20bCones harvested from 25-year old treesdP. glauca (11)aStandard cDNA library; bNormalized cDNA library; cFull-length cDNA library; dField site located at Kalamalka Research Station in Vernon, British Columbia; eOne- or two-year old trees grown in potted soil under greenhouse conditions at the University of British ColumbiaTable 2Spruce EST summaryTotal sequences206,875Number of 5' sequences41,472Number of 3' sequences165,403Average assembled 3' EST length (bp)a656.4Number of high-quality 3' sequencesb147,146Number of contigsc19,941Number of singletons26,804Number of putative unique transcriptsd46,745Number of assembled 3' ESTs withe Significant BLASTX match96,454 No significant BLASTX match50,692Average number of contig members6.03Number of contigs containing 2 ESTs6,050 3–5 ESTs7,449 6–10 ESTs3,841 11–20 ESTs1,941 21–50 ESTs572 >50 ESTs88aHigh-quality (hq) sequences only.bA sequence is considered of hq if it is not derived from contaminant species and its vector-trimmed and poor-quality-trimmed PHRED 20 length is >100 bases.cA contig (contiguous sequence) contains two or more ESTs; 3' sequences only.dNumber of putative unique transcripts (PUTs) among assembled 3' ESTs equals the number of contigs plus the number of singletons.eThreshold for BLASTX significance versus the non-redundant (NR) database of GenBank is a score value > 50.Gene discovery in normalized and non-normalized cDNA librariesFrom each of the 20 cDNA libraries, between 1,536 and 24,959 clones were 3'-end sequenced, with the rate of hq sequences ranging from 77.1% to 94.1% and an average EST length of 532 bp to 756 bp in each library (Additional File 1). The rate of gene discovery for each library was assessed from: (1) the number of unique transcripts sequenced from each library; (2) the average number of EST sequences forming contigs; (3) the percentage of ESTs with no similarity to protein sequences in the non-redundant (NR) database of GenBank using BLASTX; (4) the percentage of singleton ESTs; and (5) the percentage of library-specific transcripts. Based on these criteria, all but two of the normalized libraries (i.e., WS-SE-N-A-18 and WS-SE-N-A-19) showed considerably higher rates of gene discovery, and hence higher complexity, than the corresponding non-normalized libraries (Additional File 1). For example, among the six successfully normalized EST libraries, the percentage of unique transcripts identified within the first 1,000 reads averaged 94.7% (92.7% to 95.9%), whereas among the seven corresponding standard EST libraries made from the same RNA samples, the average was only 78.8% (73.8% to 85.6%). The diversity of starting biological materials combined with normalization resulted in low sequence redundancy demonstrated by the presence of only three PUTs (derived from 3'-end ESTs) sequenced in all of the 20 cDNA libraries (Table 3). These three transcripts were identified as translationally controlled tumor protein (209 ESTs), eukaryotic translation initiation factor 5A (115 ESTs) and S-adenosylmethionine synthase (104 ESTs).Table 3Distribution of ESTs in multiple cDNA librariesNumber of librariesNumber of putative unique transcripts with ESTs in all libraries compared2031921821712161615221436134112661110110175Quality assessment of FLcDNAsFLcDNAs are defined as individual cDNA clones that contain the complete ORF coding sequence as well as at least partial 5' and 3' UTRs for a given transcript. We prepared three FLcDNA libraries using the biotinylated cap trapper method [45]. All FLcDNA libraries were made from insect-induced tissues of a single Sitka spruce genotype (Table 1). From these libraries, we identified 8,127 cDNA candidate clones for complete insert sequencing, which resulted in 6,464 hq sequence-verified FLcDNA clones (Additional File 2). Analysis of the 6,464 FLcDNA sequences using the CAP3 program ([44]; assembly criteria: 95% identity, 40 bp window) identified 5,197 FLcDNAs as singletons, with the remaining 1,267 grouping into 521 contigs, suggesting a total of 5,718 PUTs represented with finished FLcDNA sequences. The high rate (88.5%) of unique transcript discovery resulted from a successful strategy for selection of a low-redundancy FLcDNA clone set prior to sequence finishing (Figure 1).Figure 1Clone selection and complete insert sequencing of 6,464 Sitka spruce FLcDNAs. A total of 20,469 candidate FL transcripts were identified in two consecutive rounds of clone selection involving initially 32,980 and then 46,745 putative unique transcripts (PUTs) derived from a total of 147,146 high-quality 3'-end ESTs. See Methods for complete details of candidate clone selection criteria. Among the 8,127 candidates selected for complete insert sequencing, 5,298 were finished by end reads only, and another 1,166 were finished by end reads plus gap closing using primer walking, yielding a total of 6,464 sequence-verified finished FLcDNAs. An additional 1,396 clones (17.1%) from the starting set of 8,127 will be finished in future work. Only 267 clones (3.2%) were aborted, which supports the success of our strategy for FLcDNA clone selection.All 6,464 sequence-verified FLcDNAs achieved a minimum of Phred30 sequence quality at every base (i.e., no more than one error in 103 bases). The majority were of even higher quality with the minimum and average quality values exceeding Phred45 (less than one error in approximately 3 × 104 bases) and Phred80 (less than one error in 108 bases), respectively (Figure 2). We predicted the complete protein-coding ORFs for all 6,464 FLcDNAs (Additional File 2). The average sequenced FLcDNA length (from beginning of the 5' UTR to the end of the polyA tail) was 1,088 ± 404 bp (mean ± SD), and ranged from 401 to 3,003 bp, whereas the average predicted ORF was 616 ± 374 bp and ranged from 30 to 2,583 bp (Figure 3). ORFs could not be detected (i.e., less than 30 bp) for 11 FLcDNAs. The 5' and 3' UTRs averaged 154 ± 164 bp and 301 ± 174 bp, respectively (Figure 3).Figure 2Validation of sequence quality of FLcDNAs. Sequence accuracy was measured as the percentage of the 6,464 FLcDNAs which, with 100%, 95.0–99.9%, 90.0–94.9% or <90.0% of their sequence length, exceeded Phred30, Phred40, Phred50 or Phred60 sequence quality thresholds. All 6,464 FLcDNAs exceeded the Phred30 quality thresholds (less than 1 error in 103 sequenced nucleotides) over 100% of their sequence length. Even at the threshold level of Phred60 (less than 1 error in 106 sequenced nucleotides) the majority (74.1%) of the FLcDNA sequences met this very high sequence quality score over > 95.0% of their length.Figure 3Distribution of open reading frame (ORF) and 5' and 3' untranslated region (UTR) sizes among the finished 6,464 FLcDNAs (A), and the mean ORF and UTR length (± standard deviation) (B). Each finished FLcDNA sequence was examined for the presence of ORFs using the EMBOSS getorf program (version 2.5.0; [69,70]). In each case, the longest stretch of uninterrupted sequence between a start (ATG) and stop codon (TGA, TAG, TAA) in the 5' to 3' direction was taken as the predicted ORF. The presence and coordinates of the 5' second strand primer adaptor sequence (SSPA) and polyA tail were also noted. The regions between the 5'SSPA and the predicted ORF start and between the predicted ORF stop and the polyA tail were taken to be the 5' and 3' UTRs, respectively. The 5' SSPA and 3' polyA tail lengths were not included when determining UTR length.To further assess the quality of the FLcDNAs, we performed reciprocal BLAST analysis using 872 known FL sequences from other conifer and gymnosperm species identified in previous entries in the NR database of GenBank. Using a stringent similarity threshold [identity ≥ 50%; BLASTX score value ≥ 95, where alignment scores are calculated based on match, mismatch and gaps in alignments using the default BLAST scoring matrices and parameters] we identified 297 pairs of Sitka spruce and other gymnosperm FLcDNAs. Of these pairs, 244 (82.1%) agreed well with regard to their ORF lengths (Figure 4) and positions of their starting methionine and stop codons (± ten amino acids). For the remaining pairs, the predicted 5' and/or 3' ORF ends did not match, suggesting alternative start or stop codons, splice variants, or the possibility that one of the pair members was truncated or had an incorrectly predicted ORF. Despite the relatively small number of other gymnosperm FL sequences available for pairwise comparison, the high sequence similarity within this dataset indicates that most of the 6,464 FLcDNAs represent true FL transcripts with complete ORFs and correctly annotated start and stop codons.Figure 4Validation of spruce FLcDNAs by comparison of ORF lengths (A) and cDNA lengths (B) of 297 spruce FLcDNAs with matching gymnosperm FLcDNAs in the public domain. The 6,464 FLcDNAs were compared to a collection of 872 gymnosperm sequences from SwissProt using BLASTX ([71]; release 50.1 of June 13th, 2006) annotated as full-length (excluding predicted proteins derived from genomic DNA). This comparison identified 297 homologous pairs. A spruce-gymnosperm FLcDNA pair was considered homologous if (1) the best gymnosperm protein BLASTX match exceeded a stringent threshold (% identity ≥ 50%; score value > 95) and (2) the reciprocal TBLASTN analysis identified the same spruce FLcDNA with a score value equal to or within 10% of the best match. ORF and cDNA lengths for gymnosperm sequences were extracted from the SwissProt records, and spruce ORF lengths were predicted using the EMBOSS getorf program. Strong correlations were observed for both ORF and cDNA lengths between spruce and gymnosperm sequences for the available test set of 297 homologous pairs.Most spruce ESTs have low similarity with angiosperm sequencesSince conifers and other gymnosperms are difficult experimental systems with few functionally characterized proteins, in silico annotation of spruce ESTs was performed against predicted peptides from sequenced genomes of four angiosperms (Arabidopsis, rice, poplar, and grapevine) and the moss Physcomitrella patens, together with all protein sequences in the NR database of GenBank. Among hq 3'-end ESTs > 400 bases in length (N = 133,065), between 60.5% and 68.6% have matches against each of the five plant genomes with a low stringency BLASTX score of > 50 (Figure 5A and Additional File 3). Using a more stringent threshold of score > 200, between 16.1% and 21.4% of spruce 3'-end ESTs match peptides from each of the five plant genomes of this comparison. BLASTX matches with hq 3'-end ESTs were slightly higher (72.8% and 24.5% at score > 50 and > 200, respectively) when compared to the more comprehensive collection of proteins in the NR database (Figure 5A and Additional File 3). Similar results were obtained using the assembled contig set of 46,745 spruce PUTs derived from 3'-end ESTs (Figure 5C and Additional File 5). Among hq 5'-end ESTs > 400 bases in length (N = 36,505), sequence similarity with proteins predicted from the five plant genome sequences was higher compared to 3' ESTs and PUTs, with between 74.3% and 82.6% (low stringency) and 30.7% and 40.2% (high stringency) of 5'-end ESTs matching each of the plant genomes (Figure 5B and Additional File 4). As observed with 3'-end ESTs and PUTs, an even higher proportion of 5'-end ESTs had BLASTX matches against the NR database (85.9% and 43.8% at score > 50 and > 200, respectively). These results illustrate the challenge of in silico annotation of conifer ESTs, even with hq sequences averaging > 650 bases in length.Figure 5Sequence annotation of 3' and 5' ESTs and putative unique transcripts (PUTs) against published databases. Panels A, B and C show the percentage of 3' ESTs, 5' ESTs and PUTs (derived from 3'-end ESTs), respectively, with sequence similarity to entries in nine databases including BLASTX searches against peptides from five sequenced plant genomes (i.e., Arabidopsis thaliana, Populus trichocarpa, Oryza sativa, Vitis vinifera, and Physcomitrella patens), and all peptides in the non-redundant (NR) database of GenBank; as well as BLASTN searches against 1) all gymnosperm ESTs in dbEST database of GenBank, 2) all Picea glauca ESTs in dbEST, and 3) all Pinus taeda ESTs in dbEST. Matches were identified using low (score > 50) medium (score > 200) or high (score > 1,000) BLAST stringency thresholds.We also compared the spruce ESTs and PUTs against ESTs from all gymnosperm species combined (dbEST database of GenBank, excluding ESTs reported in this study) using BLASTN. As expected, sequence similarity between the spruce ESTs and published gymnosperm ESTs was high (Figure 5 and Additional Files 3, 4, 5). Among PUTs (derived from 3'-end ESTs), hq 3'-end and 5'-end ESTs > 400 bases in length, 88.6%, 95.4% and 96.9%, respectively, have matches with scores > 50. At higher BLASTN stringency levels (i.e., scores > 200 and > 1,000), sequence matches for PUTs, 3'-end and 5'-end ESTs remain consistently high. Among those PUTs, 3'-end and 5'-end ESTs > 400 bases in length and with no obvious similarity to proteins from the five sequenced plant genomes (at score ≤ 50), 60.0%, 79.1%, and 82.0%, respectively, have BLASTN scores > 200 versus published gymnosperm ESTs (Additional Files 3, 4, 5). When the spruce ESTs are compared against published ESTs from white spruce and loblolly pine, the two gymnosperm species with the most substantial EST collections, a higher proportion of PUTs, and 3'-end and 5'-end ESTs show sequence similarity to white spruce compared to loblolly pine, especially at the highest BLASTN threshold (Figure 5 and Additional Files 3, 4, 5).Utility of spruce FLcDNAs for comparative sequence annotationAs might be expected, sequence similarity between the 6,464 Sitka spruce FLcDNAs and other gymnosperm ESTs is very high, with 96.5%, 94.6% and 78.7% of FLcDNAs matching published gymnosperm ESTs at low, medium, and high sequence similarity thresholds, respectively (Figure 6A and Additional File 2). As observed with spruce ESTs, sequence similarity was highest between spruce FLcDNAs and white spruce ESTs, with lower similarity observed with loblolly pine ESTs (Figure 6A). Next, the spruce FLcDNAs were compared against predicted proteins from five plant genome sequences and protein sequences in the complete NR database of GenBank. At a low sequence similarity threshold of score > 50, between 76.5% and 84.2% of FLcDNAs matched proteins from each of the plant genomes of this comparison, whereas at a higher threshold of score > 200 the percentages of FLcDNAs with matches in the plant genome sequences ranged from 38.1% to 44.9% (Figure 6A and Additional File 2). Overall, the Sitka spruce FLcDNAs show greater similarity to predicted proteins from sequenced plant genomes compared to the spruce ESTs. The proportion of spruce FLcDNAs with similarity to proteins in the NR database was also higher than spruce ESTs at 87.7% and 47.9% at score > 50 and score > 200, respectively (Figure 6A and Additional File 2).Figure 6Sequence annotation of 6,464 high-quality spruce FLcDNAs against published databases. Panel A shows the percentage of FLcDNAs with sequence similarity to entries in nine databases including BLASTX searches against peptides from five sequenced plant genomes (i.e., Arabidopsis thaliana, Populus trichocarpa, Oryza sativa, Vitis vinifera, and Physcomitrella patens), and all peptides in the non-redundant (NR) database of GenBank; as well as BLASTN searches against 1) all gymnosperm ESTs in dbEST database of GenBank, 2) all Picea glauca ESTs in dbEST, and 3) all Pinus taeda ESTs in dbEST. Matches were identified using low (score > 50) medium (score > 200) or high (score > 1,000) BLAST stringency thresholds. Panels B and C show the non-overlapping distribution of matches of spruce FLcDNAs against seven databases (peptides from A. thaliana, P. trichocarpa, O. sativa, V. vinifera, P. patens, and the NR database of GenBank; and gymnosperm ESTs) at BLAST score thresholds of > 50 and > 200, respectively. Panels D and E show the database source in cases where spruce FLcDNAs matched only a single database in panels C and D at BLAST score thresholds of > 50 and > 200, respectively.These results show that FLcDNAs provide a clear advantage over ESTs for large scale in silico annotation of spruce sequences. Nevertheless, when using high stringency criteria relevant for in silico functional annotation (score values > 200), the comparison of spruce FLcDNAs against the five plant genomes, as well as all plant species in the NR database, still identifies a substantial number of sequences that only show significant matches with other gymnosperms, as opposed to angiosperms. Among the 6,464 spruce FLcDNAs, we found 927 (14.3%) without a reliable match to angiosperm sequences at a low stringency (i.e., BLASTX score ≤ 50), of which 743 (80.1%) match with high sequence similarity (i.e., BLASTN score > 200) to a published gymnosperm EST sequence (Additional File 2). A very small number of spruce FLcDNAs lack sequence similarity to angiosperm or gymnosperm sequences (at score ≤ 50) and display a best match with non-plant species in the NR database of GenBank; 1.0% at score > 50 and 0.3% at score > 200 (Additional File 2). In these cases, the best match is often an insect sequence suggesting small amounts of contaminants in the cDNA libraries.Comparing the entire spruce FLcDNA dataset against sequences from all species identified that 71.9% (at score > 50) or 34.2% (at score > 200) have matches in all seven datasets (i.e., five plant genomes, the NR database of GenBank, and gymnosperm ESTs) (Figure 6B and 6C). It is notable that at the higher threshold of score > 200, 47.2% of spruce FLcDNAs match only to a single database, and in the vast majority of cases this is a gymnosperm sequence (Figure 6E). Another 1.0% (at score ≤ 50) or 3.8% (at score ≤ 200) of spruce FLcDNA sequences do not align to any sequences in available databases. These sequences could represent genes from spruce (or genes from other contaminant organisms) that have not been sequenced before in any source.DiscussionSpruce ESTs and FLcDNAs enhance conifer genomics resourcesGenomics research on conifers has been limited by the lack of a relevant gymnosperm reference genome sequence. The very large size of conifer genomes (10 to 40 Gb; [1]), dominated by repetitive DNA, has been a roadblock to a conifer genome sequence project. Furthermore, the phylogenetic distance between conifers and the well-studied angiosperms is more than 300 million years [2], limiting the utility of angiosperm genome information for research in conifers. To overcome these obstacles to conifer genome research, we have developed two new valuable components for the \"conifer genomics toolbox\".First, we have assembled a large collection of high-quality, sequence-verified FLcDNA clones from Sitka spruce, along with a corresponding database of in silico annotations (Additional File 2). These FLcDNAs are of very low redundancy. They represent the third largest sequence-verified FLcDNA resource for any plant species, behind only rice [7] and Arabidopsis [4], and are the only substantial FLcDNA resource for a conifer or any other gymnosperm.Second, we have added a large number of new EST sequences to the public spruce EST collection in GenBank, along with corresponding databases of in silico annotations (Additional Files 3, 4, 5). This resource, which was developed from Sitka, white and interior spruce (interior spruce has varying degrees of admixture between white and Engelmann spruce), substantially improves the size and quality of the previously described spruce EST collections [20-23]. The spruce EST collection, along with the ESTs from loblolly pine [15-18], is now one of the two largest EST resources for any conifer species. To enhance gene discovery, we strategically employed library normalization, which had previously not been applied to a conifer EST program. Also, we have added sequences from an until now poorly represented class of tissues representing a biologically important component of conifer defense: insect-, wound- or elicitor-induced tissues.We identified 46,745 PUTs (19,941 contigs, 26,804 singletons; derived from 3'-end ESTs) in the three species groups surveyed here; Sitka spruce, white spruce, and interior spruce. The rates of PUT discovery for all species combined (31.8%), white spruce only (33.0%) and Sitka spruce only (36.6%) are comparable, as are the ratios of singletons to contigs in each collection. Among contigs from the combined analysis of white and Sitka spruce ESTs, 26.7% contained ESTs from both species, suggesting that ESTs derived from different spruce species representing the same spruce gene often cluster together. The PUTs identified here may represent a substantial portion of the expressed gene catalogue for species of spruce, but a complete genome sequence is needed for assessment of true gene numbers in conifers.The spruce ESTs described here have already provided the foundation for functional and comparative genomics research on conifer defense against insects, adaptation to the environment, somatic embryogenesis and wood formation, via both transcriptome and proteome analyses [21,34,35,42,46]. They have also allowed development of three types of genetic markers: microsatellites [47,48], single nucleotide polymorphisms (SNP) and conserved orthologous sequences (COS) [41]. The FLcDNA sequences enable rigorous large-scale comparisons of evolutionary patterns at large evolutionary scales (K. Ritland et al., manuscript in preparation).Utility of spruce FLcDNAs for functional characterization of gene families including nearly identical paralogous genesPrior to this work, only a few dozen complete spruce protein sequences were available in the SwissProt database, and no substantial FLcDNA resource was available for any gymnosperm. Using FLcDNAs, detailed pathway annotation, gene expression analysis, and biochemical functional characterization of individual genes and gene families are now possible (S.G. Ralph and J. Bohlmann, manuscript in preparation). The Sitka spruce FLcDNAs have already advanced the discovery and the characterization of conifer defense genes [49-53]. Importantly, Sitka spruce FLcDNAs allow for accurate analysis of closely related members of gene families such as cytochrome P450-dependent monooxygenases or terpenoid synthases (TPS) involved in defense against insects or pathogens [40,54]. For example, TPSs represent a gene family containing many pairs or groups of nearly identical paralogous genes each with a potentially different biochemical function [32]. Our recent mutational analysis of two closely related paralogous Norway spruce di-TPS illustrated that a single amino acid mutation in a background of more than 800 amino acids completely alters biochemical product profiles [55]. Similarly, in rice, the functional divergence of two distinct TPS of primary and secondary metabolism was due to a single amino acid substitution [56]. These examples illustrate the utility of true FLcDNAs for discovery of nearly identical paralogous genes and for functional assessment of gene evolution that is now possible in Sitka spruce.Utility of FLcDNAs for conifer proteome and genome characterizationBeyond their importance for functional characterization of individual genes and the analysis of gene families, on an even larger scale, FLcDNAs are also superior to ESTs for overall proteome and genome characterization in a conifer. Because the Sitka spruce FLcDNAs allow for a much more reliable prediction of the complete protein-coding ORF than ESTs, they have been invaluable for proteome predictions and practical proteome analyses [35]. In expectation of future efforts to sequence a conifer genome, FLcDNAs and their ORFs will be essential for the development and training of gene prediction software, as has recently been demonstrated for poplar [8,9].Spruce FLcDNAs from insect-induced libraries reveal genes not detected in angiospermsComparison of Sitka spruce sequences against angiosperm plants suggests that there are likely a substantial number of genes in the collection of 6,464 FLcDNAs that are either absent in other species, or lack significant sequence similarity for unambiguous identification. In earlier work, Kirst et al. [16] suggested that less than 10% of loblolly pine transcripts lack a related gene in Arabidopsis (defined at a BLASTX E value cutoff of 1e-10 or ca. score 60). When we analyzed the spruce FLcDNAs, we found that approximately 14% had no similarity to any angiosperm at a BLASTX stringency of score 50 (slightly lower than that applied by Kirst et al. [16]), based on comparisons to four sequenced angiosperm genomes and all angiosperm sequences in the NR database. This slightly higher rate may be the result of sequencing libraries made from tissues induced by insect attack, which may disproportionally represent genes with specialized functions in conifer defense that are subject to high levels of natural selection due to biotic interaction. By contrast, genes involved in xylem development and wood formation appear to be well conserved in angiosperms and conifers [16,46].ConclusionThe 206,875 ESTs and 6,464 FLcDNAs and the corresponding in silico annotated sequence databases provide a new and valuable genomics resource for species of spruce, as well as for gymnosperms in general. Our emphasis on FLcDNAs and ESTs from cDNA libraries constructed from herbivore-, wound- or elicitor-treated induced spruce tissues, along with incorporating normalization to capture rare transcripts, gives a rich conifer EST resource which also apparently contains a substantial number of transcripts with no obvious sequence similarity to known angiosperm sequences. Recent research has begun to fully realize the application of these EST and FLcDNA sequences, and FLcDNA clones.MethodscDNA library constructionDetails of the isolation of total and poly(A)+ RNA are described in Additional File 6. Standard cDNA libraries were directionally constructed (5' EcoRI and 3' XhoI) using 5 μg poly(A)+ RNA and the pBluescript II XR cDNA Library Kit, following manufacturer's instructions (Stratagene, La Jolla, USA) with modifications. First-strand synthesis was performed using Superscript II reverse transcriptase (Invitrogen, Carlsbad, USA) and an anchored oligo d(T) primer [5'-(GA)10ACTAGTCTCGAG(T)18VN-3']. Size fractionation was performed on XhoI-digested cDNA prior to ligation into vector using a 1% NuSieve GTG low melting point agarose gel (BioWhittaker Molecular Applications, Walkersville, USA) and β-agarase (New England Biolabs, Ipswich, USA) to isolate cDNAs from 300 bp to 5 kb. Select cDNA libraries were normalized to Cot = 5 using established protocols [13,14]. Library plasmids were propagated in ElectroMAX DH10B T1 Phage Resistant Cells (Invitrogen). FLcDNA libraries were directionally constructed (5' XhoI and 3' BamHI) according to methods of Carninci and Hayashizaki [57] and Carninci et al. [58], with modifications described in Additional File 6.DNA sequencing and sequence filteringDetails of bacterial transformation with plasmids, clone handling, DNA purification and evaluation, and DNA sequencing are provided in Additional File 6. Sequences from each cDNA library were closely monitored to assess library complexity and sequence quality. DNA sequence chromatograms were processed using the PHRED software (versions 0.000925.c and 0.020425.c) [59,60]. Sequences were quality-trimmed according to the high-quality (hq) contiguous region determined by PHRED and vector-trimmed using CROSS_MATCH software [61]. Sequences with less than 100 high quality bases (Phred20 or better) after trimming and sequences with polyA tails of ≥ 100 bases were removed from the analysis. Also removed were sequences representing bacterial, yeast or fungal contaminations identified by sequence alignments using BLAST [62,63] to E. coli K12 DNA sequence (GI: 6626251), Saccharomyces cerevisiae (GenBank, ), Aspergillus nidulans (TIGR ANGI.060302), and Agrobacterium tumefaciens (custom database generated using SRS, Lion Biosciences). Sequences were also compared to the NR protein database [67]. Top ranked BLAST matches to species other than plants with score values > 60 were flagged as contaminants and were removed from the EST dataset. EST sequences have been deposited in the dbEST database of GenBank [DR448912 to DR451924; DR463975 to DR595214; CV720218 to CV720219; CO203067 to CO245079; CO250245 to CO252887; CO252989 to CO253183; CO253265 to CO257405; CO257513 to CO258618; CN480886 to CN480910].Selection of candidate FLcDNA clones and sequencing strategyAll 3'-end ESTs remaining after filtering were clustered and assembled using CAP3 ([44]; assembly criteria: 95% identity, 40 bp window). The resulting contigs and singletons were defined as the putative unique transcript (PUT) set. PUTs with a cDNA clone from a FLcDNA library were selected as candidates for complete insert sequencing. Candidate clones from FLcDNA libraries were single-pass sequenced from both 3'- and 5'-ends, and both sequences were used for subsequent clone selection. Clones were screened for the presence of a polyA tail (3'-end EST) and the second-strand primer adaptor (SSPA; 5'-ACTAGTTTAATTAAATTAATCCCCCCCCCCC-3'; 5'-end EST). Clones lacking either of these features were eliminated. A polyA tail was defined as at least 12 consecutive, or 14 of 15 \"A\" residues within the first 30 bases of the 3'-end EST (5' to 3'). The presence of the SSPA was detected using the Needleman-Wunsch algorithm limiting the search to the first 30 bases of the 5'end EST (5' to 3'). The SSPA was defined as eight consecutive \"C\" residues and a ≥ 80% match to the remaining sequence (5'-ACTAGTTTAATTAAATTAAT-3'). In each case, the algorithms used to detect the 5' and 3' clone features were set to produce maximal sensitivity while maintaining a 0% false positive rate, as determined using test data sets. Candidate clones for which either of the initial 5'-end or 3'-end EST sequences had a Phred20 quality length of < 400 bases were also excluded. Finally, any clone with a 5'-end EST which had a BLASTN match (score value > 300) to a gymnosperm EST in the public domain (excluding ESTs from this collection) and was > 100 bases shorter at the 5' end than the matching EST was flagged as truncated at the 5' end and was excluded. For each PUT represented by multiple candidate clones after filtering, the clone with the longest 5' sequence was selected for complete insert sequencing. Insert sizing using colony PCR and vector primers was performed on 1,634 cDNA clones with an average insert size of ca. 1,250 bp. Based on this information, a sequencing strategy emphasizing the use of end reads was chosen. Using end reads only, 5,298 clones were complete insert sequenced to a high quality. Among this set, the average sequenced insert size was 1,005 ± 282 bp (average ± SD) with an average of 5.93 ± 0.51 end reads required to finish. Using a combination of end sequencing and primer walking, an additional 1,166 clones were complete insert sequenced, with an average insert size of 1,653 ± 447 bp, and requiring six end reads and 2.62 ± 1.51 internal primer reads per clone.Sequence finishing of FLcDNA clonesFLcDNA clones selected for complete sequence finishing were rearrayed into 384-well plates, followed by two additional rounds of 5'-end and 3'-end sequencing using vector primers. All sequences from an individual clone were then assembled using PHRAP (version 0990329) [59,60]. To meet our hq criteria, the resulting clone consensus sequence was required to achieve a minimum average score of Phred35, with each base position having a minimum score of Phred30. Each base position also required at least two sequences, each with a minimum quality of Phred20, that were in agreement with the consensus sequence (i.e., no high-quality discrepancies). Clones that did not meet these finishing criteria or that had gaps after three rounds of end sequencing were then subjected to successive rounds of sequencing using custom primers designed using the Consed graphical tool version 14 [64] until the required quality levels were achieved. Regardless of the finishing strategy, all clones that did not meet the minimum finishing criteria according to an automated pipeline were manually examined. Clones were aborted if they were manually verified to lack the minimum finishing criteria, did not possess the cloning structures, were identified as chimeric, were refractory to sequence finishing due to the presence of a \"hard-stop\", or if errors were identified in the re-array of glycerol stocks. FLcDNA sequences have been deposited in GenBank [EF081469 to EF087932].Comparative sequence annotationThe following databases were used to perform BLAST analyses for EST and FLcDNA annotation: 1) Arabidopsis thaliana, The Arabidopsis Information Resource version 7, release date April 25th, 2007, 31,921 peptides [65]; 2) Populus trichocarpa, Joint Genomes Institute (JGI) version 1.1, release date September 16th, 2006, 45,555 peptides [66]; 3) Oryza sativa, National Center for Biotechnology Information (NCBI), download date April 8th, 2008, 177,254 peptides [67]; 4) Vitis vinifera, NCBI, download date April 8th, 2008, 55,851 peptides [67]; 5) Physcomitrella patens, JGI version 1.1, release date January 4th, 2008, 35,938 peptides [68]; 6) NR database of GenBank, NCBI release 162, release date October 15th, 2007, 5,372,238 peptides [67]; 7) gymnosperm ESTs in NCBI (excluding ESTs reported in this study), download date April 8th, 2008, 622,923 ESTs [67]; 8) Picea glauca ESTs in NCBI (excluding ESTs reported in this study), download date April 8th, 2008, 197,042 ESTs [67]; 9) Pinus taeda ESTs in NCBI, download date April 8th, 2008, 328,628 ESTs [67].Authors' contributionsJB and SGR conceived and directed this study. SGR, NK, DC, and CO developed full-length cDNA and EST libraries. SGR, HJEC, RK and JB analyzed data with assistance from the coauthors. RAH, SJMJ and MM directed sequencing and bioinformatics work at the GSC. JB and SGR wrote the paper. All authors read and approved the final manuscript.Supplementary MaterialAdditional File 1cDNA library summary statistics. Sequencing statistics organized by cDNA library source for spruce expressed sequence tags.Click here for fileAdditional File 2Full-length cDNA inventory. Predicted protein-coding features, annotation, and GenBank accession numbers for the Sitka spruce full-length cDNA collection.Click here for fileAdditional File 33'-end EST inventory. Detailed annotation and GenBank accession numbers for the complete set of spruce 3'-end ESTs.Click here for fileAdditional File 45'-end EST inventory. Detailed annotation and GenBank accession numbers for the complete set of spruce 5'-end ESTs.Click here for fileAdditional File 5PUT inventory. Detailed annotation for the complete set of putative unique transcripts.Click here for fileAdditional File 6Supplemental methods. Detailed methods for RNA isolation, full-length cDNA library construction, bacterial transformation with plasmids, clone handling, DNA purification and evaluation, and DNA sequencing.Click here for file\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2581729.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2581729",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2581729\nAUTHORS: C. Bettocchi, P. Ditonno, F. Palumbo, G. Lucarelli, G. Garaffa, B. Giammusso, M. Battaglia\n\nABSTRACT:\nEven in the era of phoshodiesterase type 5 inhibitors, penile implants are considered the definitive solution for the treatment of organic erectile disfunction. The advent of new surgical tools and new infection-resistant materials has significantly reduced the risk of intra and post-operative complications and the need for revision surgery. Various companies have also improved their mechanical systems in order to reduce the risk of failures, and their products are now so good they may last lifelong. In this article, we evaluate the intraoperative and postoperative complications recorded in our experience and in literature reports, and make some suggestions as to how to prevent or correct them.\n\nBODY:\n1. INTRODUCTIONNowadays implanting a penile prosthesis is the definitive solution for the treatment of\norganic erectile dysfunction (ED), even in the era of effective and safe oral\nmedications [1]. The types of prosthesis most commonly implanted are the three-piece\ninflatable device, the two-piece inflatable device, and the soft and malleable\nprosthesis. In the last few\nyears, the three-piece inflatable device has been used for preference, as it\nimproves the erection, the flaccid, and\nappearance of the penis and as it yields a more acceptable and cosmetical functional\nresults [2]. On the other hand, the relative complexity of this last device is\nalso the source of mechanical failures and patients' difficulties in managing\nthe device. In the last decade, there has been a continuous improvement in the mechanical\nfunction of the devices and in the composition of the materials used but device-related complications still occur.Some complications can be prevented by a correct preoperative assessment. The\nsurgeon has to understand the patient's real needs and expectations, as well as\nthose of his partner in order to be able to choose the right device. The\ncounselling must also include a complete, clear explanation of how the device\nfunctions and the obvious changes that will arise in the sexual life of the\ncouple. Informed consent to the procedure is mandatory, and when discussing the\noption of a penile implant with the patient, issues such as complications and\nthe irreversibility of the procedure should be exhaustively discussed.In this paper, we evaluate the\nintraoperative and postoperative complications recorded in our experience and in\nliterature reports, and make some suggestions as to how to prevent or correct\nthem.2. INTRAOPERATIVE COMPLICATIONS2.1. Cylinders positioningDuring the implant procedure, after having exposed the corpora cavernosa\nand performed the corporotomy, the first critical step is dilating the corpora.\nIn most patients the corpus cavernosum cavity is\ndilated to the maximum capacity using Hegar dilators of various sizes. The\ndilator must be introduced through the corpus by pushing it in an outward direction in order to avoid\ncross-over perforation. In cases of fibrotic corpora, special dilators may be\nuseful to create an appropriate space (Rossello dilators or Otis urethrotome)\nbecause perforation is especially risky in this case.A distal corpora perforation can be\ncorrected first of all by exposing the damaged corpus apex. Then, if it is only\na small hole, the tip can be closed with separate PDS stitches. The way to\nmanage distal perforation in cases of larger holes is by covering the damaged\napex with a dacron or gore-tex sleeve.Proximal corpora perforation usually\noccurs during dilatation of the corpus cavernosum crura. A possible way to\nevaluate a proximal perforation intraoperatively is by positioning dilators in\nboth crura and checking whether they are at different heights, showing that one\nhas penetrated too deeply inside the corpus. If not discovered during the\noperation, a postoperative MRI scan is the best evaluation to confirm a\nproximal perforation. One of the two ways of managing this complication is by\ncreating dacron or gore-tex socks, especially in cases of a malleable or soft\nprosthesis (Figure 1). The other possibility, indicated for inflatable devices,\nis to fix the cylinders to the surrounding corpora tissue, placing stitches\nabove and below the tubes input. The anchored cylinder tends not to protrude,\nallowing healing of the perforation. Another similar solution involves fashioning\na sling through the tip extender using nonabsorbable sutures.Incorrect introduction of the\ndilators is the main cause of cross-over perforation. It is important to\nrecognise this kind of perforation as soon as possible so as to implant two\ncylinders in the same corpus. Usually a redo correct ipsilateral dilatation is\nsufficient to correct the cross-over perforation.Another consequence of incorrect\ndilators introduction is urethral perforation. To check for urethral injuries,\nit is always best to irrigate the corpora with a saline plus antibiotic\nsolution: if the fluid leaks through the urethral meatus, a perforation has\noccurred. The diagnosis can be confirmed by cystoscopy. The treatment option in\nsuch cases is urethral repair for proximal perforations. If the laceration\ninvolves the urethral meatus, it is advisable to postpone the procedure. It is\npossible to position a urethral catheter if necessary with a suprapubic\ncatheter, delaying insertion of the cylinder or positioning of a malleable\nprosthesis until the damaged urethra has healed. The malleable prosthesis will\nbe replaced by the inflatable cylinder at a later date during a second\noperation.A rare complication has been\ndescribed by Hatzimouratidis et al. [3]; it occurred during dilation of the\ncorpora cavernosa with Brooks dilators: the \nhead detached and stuck to the tip of the corpus cavernosum. The case\nwas managed by incising the distal lateral part of the corpora cavernosa and\nthen removing the head of the dilators. In any case, we strongly recommend\nexamining all surgical tools carefully before using them.2.2. Reservoir positioningThe possible complications occurring\nduring the reservoir positioning step are mostly due to this peculiar blind\nprocedure. If the fascia is not completely opened, the reservoir may not pass\nthrough, remaining outside: this is a typical postoperative complication.\nAnother possibility is to open the peritoneum: in this case, it is mandatory to\ncheck for bowel injuries.During reservoir positioning, it is\nvery important to have positioned a urethral catheter and ensured that the\npatient has completely emptied his bladder. If not, the risk of bladder\nperforation is high. This complication can also occur in patients who have previously\nundergone pelvic surgery, such as radical prostatectomy. If a bladder\nperforation occurs, cystoscopy can confirm the damage severity; usually leaving\na catheter in place for a few days is sufficient to treat such complications.\nIn rare cases of wide perforation, an open bladder repair can be performed.2.3. Component failure/breakageIn order to avoid a malfunctioning\ndevice, it is always advisable to check correct device functioning before\nplacement and to activate the pump with cylinders connected after the\nplacement. At this surgical stage, it is easy to substitute a nonfunctioning\ndevice.Another possible complication is\nbreakage of device components during cavernotomy closure or during\nrepositioning of Scott retractor's hooks during the operation. One way to\nprevent device perforation is to put the stitches in before performing the\ncorporotomy and before positioning the cylinders.3. POSTOPERATIVE COMPLICATIONS3.1. Cylinders complications\nInfectionsInfection is one of the most fearsome complications, having an incidence\nof 8 to 20%, as reported in large series of implants [2–4]. Infections can\noccur a few months after surgery and a typical sign is persistent, unchanging,\nor even increasing pain. The pain could be exacerbated by activating the\ndevice. Other signs of infection are penile or scrotal erythema, fever,\npurulent drainage from the wound, or skin erosion. Diabetic patients are more\nlikely to develop an infection, even if the previous concept that poor glycemic\ncontrol increases the risk has not been confirmed [5]. Moreover, insulin\ndependency and hemoglobin A1C serum levels are not considered additional risk\nfactors. Other conditions, possibly associated with an increased risk of\ninfection, are the use of immunosuppressive drugs and steroids, and the\npresence of spinal cord injury.When the presence of infection is confirmed, the use of systemic\nantibiotics therapy is not sufficient in the vast majority of cases. This is\ndue to the infectious agent's ability to create a biofilm surrounding the\nprosthesis components, protecting bacteria from the antibiotic action. In most\ncases, the infection is sustained by opportunistic bacteria such as Staphylococcus\nepidermidis or Streptococcus agalactie; more rarely, toxic bacteria\nlike Escherichia coli, Staphylococcus\naureus, Enterococcus faecalis, or Pseudomonas are involved. The\nlatter agents tend to present early in the postoperative period, with fever,\ndeep tissue penetration, and abundant purulent drainage.The classical approach to an infected device is the immediate removal of\nall the components and placement of a new implant after some delay for healing.\nThe advantage of this solution is that the new implant is scheduled only when\nthe infection has completely cleared. The main disadvantage is the scarring\nprocess that occurs inside the penis and hence penile retraction causing more\ndifficult surgery later. In the last years salvage procedures have been\nproposed that allow positioning of a new penile prosthesis at the same time as\nremoval of the infected one [6, 7]. The immediate salvage procedure consists of\nremoval of the infected prosthesis and wound irrigation with seven different\nantiseptic solutions including antibiotics (Kanamycin, Bacitracin, Vancomycin,\nand Gentamycin), hydrogen peroxide, and betadine. A new prosthesis is then\neasily placed, and the overall success rate is more than 80%. The delayed\nsalvage procedure consists of placement of a drainage tube after removal of the\nprosthesis; antibiotic solution is irrigated through the drain and a new\nprosthesis is placed about 3 days later. Actually, no advantage has been\ndemonstrated for the delayed salvage procedure over the immediate one. A few\nyears ago, based on the evidence that some antibiotics are particularly\nindicated to protect silicone graft materials, the American Medical System Company\ndeveloped a minocycline-rifampicin-coated penile prosthesis called Inhibizone\n[8]. Early experiences with this new device have demonstrated an evident\nreduction of overall infections, and no infections at all in primary implanted\npatients [9]. Another local approach to prevent device infection has been\nproposed by the Mentor Corporation Company and consists of applying a special\nhydrophilic coating that seems to inhibit bacterial adherence. The prosthesis is then soaked in antibiotics\nand the combined effect should reduce the risk of infection. In an initial experience, the Mentor Titan\nprosthesis has also demonstrated effectiveness in reducing the infection rate\n[10].In some patients, the infection could be associated with important\ntissue necrosis: in this case, a salvage procedure is not advisable. Severe\ndistal tissue necrosis is a dramatic event that may even require penile\nglansectomy or amputation (Figures 2 and 3) after prosthesis removal.\n\nWrong sizingUsing an oversized cylinder can lead to an S-shaped deformity and\nbuckling. As reported by Moncada et al. [11], an oversized cylinder is\nresponsible for constant pain and exposes the patient to the risk of erosion.\nThe solution in such cases is to replace the device. The opposite problem is\nundersizing, which will have the effect of a so-called “concorde deformity” (Figure 4) with excess mobility of the glans. In this case, cylinder removal is not\nnecessary and it is possible to mobilize the glans with a subcoronal incision.\nWhen the cylinder tip becomes visible, nonabsorbable sutures can be used to\nhitch the glans and anchor it to the tunica albuginea, in order to completely\ncover the head of the prosthesis.\n\nErosionIn the era of hydraulic inflatable devices, erosions are considered a\nrare complication. Distal erosion can be due to an excessive intraoperative\ncorpora cavernosa dilatation, when oversized cylinders are used, in patients\nwith loss of penile sensation (cold glans syndrome) and in patients unable to\ndeflate the device when not in use. To manage distal erosion, it is necessary\nto remove the cylinder if oversized and replace it with a smaller prosthesis.\nThe new device has to be placed far from the scar tissue, performing a new dilatation. Cavernosa\nreconstruction can be performed with albugineal surgery, as proposed by Mulcahy\n[12]. The cylinder can usually be readily reseated in an area of spongy\ntissue behind the back wall of the sheath containing the extruded cylinder.\nThis is done by making a corporotomy over the cylinder laterally, about half\nthe distance towards the penoscrotal junction, retracting the cylinder to the\nside, incising the back wall of the cylinder sheath, and dilating a new cavity\nbehind this back wall up to the subglandular area. The cylinder can then be\nreseated in this new cavity and the back wall of the cylinder sheath will act\nas the outer covering of the cylinder. A second layer consisting of the outer\nwall of the cylinder sheath can also be closed to create a more secure barrier\nagainst the extrusion of parts. The corporotomy is closed with long-term\nadsorbable suture. The cylinder is now secured in its proper location by two\ntough layers comprising the back wall of the original sheath and the\ncorporotomy closure. Cavernosa reconstruction can also be made using synthetic\nmaterials like dacron or Gore-Tex.A peculiar kind of distal erosion is urethral erosion. A possible\nsolution is to remove the cylinder and to position a suprapubic catheter to\nallow healing of the urethral perforation. A single-stage procedure has been\ndescribed by Shaeer [13]: having mobilized the glans off the tip of the corpus\ncavernosum, the caverno-urethral fistula is disconnected and sealed by primary\nsutures. The perforation on the corpus cavernosum side is corrected by double\nbreasting or by grafting. The prosthesis\nis then reimplanted.Proximal erosion and cross-over erosion are usually intraoperative\ncomplications. MRI will confirm the diagnosis: the management consists of\nremoval of the protruded cylinder. A cavernosa reconstruction with a dacron\nsock is necessary before inserting a new prosthesis.\n\nMechanical failureCylinders mechanical failure would involve loss of fluid due to\nbreakage, bulging, or aneurysmatic dilatation. The only solution to manage such\ncases is to remove the broken device and replace it with a new penile\nprosthesis. The introduction of new covering materials like Parylene has\ndramatically reduced the risk of cylinders bulging.\n3.2. Pump complicationsPump infections require the same management as described above for\ncylinders. Prevention of hematoma and swelling with closed-suction drains has\nbeen shown not to increase the infection rate and to promote an earlier\nrecovery time. In a large series of 425 consecutive primary three-piece penile\nprosthesis implantations, there were a total of 14 (3.3%) infections and three\nhematomas (0.7%) during a mean follow-up of 18 months [14].Pump or connecting tubes erosion is usually associated with infections.\nIf the infection is not extensive and not associated with severe tissue\nnecrosis, a salvage procedure can be performed locally and a new pump can be\ninserted. In cases of considerable loss of tissue, poor patient conditions, and\nfever, it is advisable to remove the prosthesis and delay the reimplant.Pump migration or incorrect positioning is mainly due to insufficient\nclosure of the scrotal space. If the pump is no longer useful because of its\nincorrect position, a new operation is required to fix it in the correct\nscrotal place.3.3. Reservoir complicationsReservoir complications are not frequent but include positioning of the\nreservoir over the fascia. Migration is a rare event and usually occurs when a too\nbig space is created through the fascia to access the Retzius space. With a\nsuprapubic incision, the reservoir can be replaced in the correct paravesical\nspace.A difficult or failed device deflation can be due to pseudocapsule\nformation around a partially emptied reservoir. To prevent capsule formation, it\nis usually sufficient to leave the reservoir half-filled for 24 hours after the\noperation. Early hospital testing of the\nprosthesis function is also advisable. When a pseudocapsule is present, surgical\nrevision will be needed to access the Retzius space once more, to break the\ncapsule, and to replace the reservoir. If the previous side is no longer\navailable, it is best to replace the reservoir in the other paravesical space\nor, if necessary, in the peritoneum.4. CONCLUSIONSPenile prosthesis implantation is a fascinating surgical technique that\nhas gained an important role in the treatment of severe erectile dysfunction.\nThe advent of new surgical tools and new infection-resistant materials has\nsignificantly reduced the risk of intra- and postoperative complications and\nthe need for revision surgery. Various companies have also improved their\nmechanical systems in order to reduce the risk of failures, and their products\nare now very good as they may last lifelong. Nevertheless, surgical skill and a meticulous respect\nfor sterility rules remain fundamental requirements to guarantee the success of\na penile prosthesis implant.\n\nREFERENCES:\n1. Sadeghi-NejadHPenile prosthesis surgery: a review of prosthetic devices and associated complicationsJournal of Sexual Medicine20074229630917367425\n2. MinerviniARalphDJPryorJPOutcome of penile prosthesis implantation for treating erectile dysfunction: experience with 504 proceduresBJU International200697112913316336342\n3. HatzimouratidisKKoliakosNKoutsogiannisIMoisidisKGiakoumelosAHatzichristouDRemoval of a detached head of the Brooks dilator from the corpora cavernosa during penile prosthesis implantationJournal of Sexual Medicine200744ii1179118117484773\n4. NataliAOlianasRFischMPenile implantation in Europe: successes and complications with 253 implants in Italy and GermanyJournal of Sexual Medicine2008561503151218410306\n5. WilsonSKCarsonCCClevesMADelkJRIIQuantifying risk of penile prosthesis infection with elevated glycosylated hemoglobinThe Journal of Urology19981595153715409554349\n6. MulcahyJJLong-term experience with salvage of infected penile implantsThe Journal of Urology2000163248148210647660\n7. KnollLDPenile prosthetic infection: management by delayed and immediate salvage techniquesUrology19985222872909697796\n8. BrockGBochinskiDMahoneyCBInhibiZone treatment: the first antibiotic treatment impregnated into the tissue-contacting surface of an inflatable penile prosthesisThe Journal of Urology2001165p. A1047\n9. WilsonSKDelkJRHenryGDShort-term follow-up for enhanced American medical systems 700 CX prosthesisThe Journal of Urology2002167supplement, p. 150, abstract 600.\n10. WolterCEHellstromWJGEarly experience with the hydrophilic-coated three-piece IPP (Titan alpha-1 IPP)The Journal of Urology2003169supplement, abstract 1333.\n11. MoncadaIHernándezCJaraJBuckling of cylinders may cause prolonged penile pain after prosthesis implantation: a case control study using magnetic resonance imaging of the penisThe Journal of Urology1998160167719628607\n12. MulcahyJJDistal corporoplasty for lateral extrusion of penile prosthesis cylindersThe Journal of Urology1999161119319510037396\n13. ShaeerOManagement of distal extrusion of penile prosthesis: partial disassembly and tip reinforcement by double breasting or graftingJournal of Sexual Medicine2008551257126218331264\n14. Sadeghi-NejadHIlbeigiPWilsonSKMulti-institutional outcome study on the efficacy of closed-suction drainage of the scrotum in three-piece inflatable penile prosthesis surgeryInternational Journal of Impotence Research200517653553815988544"
+}

batch_13/PMC2583043.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2583043",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2583043\nAUTHORS: Liraz Kantor, Dalia Pinchasi, Michelle Mintz, Yetrib Hathout, Adeline Vanderver, Orna Elroy-Stein\n\nABSTRACT:\nBackgroundMutations in eukaryotic translation initiation factor 2B (eIF2B) cause Childhood Ataxia with CNS Hypomyelination (CACH), also known as Vanishing White Matter disease (VWM). The disease is manifested by loss of brain myelin upon physiological stress. In a previous study, we showed that fibroblasts isolated from CACH/VWM patients are hypersensitive to pharmacologically-induced endoplasmic reticulum (ER) stress. Since brain cells from affected individuals are not available for research, we wished to assess the effect of eIF2B mutation on oligodendroglial-derived cells.Methodology/Principal FindingsA rat oligodendroglial-derived cell line was used for a stable knock-down of eIF2B5 followed by stable expression of mutated eIF2B5(R195H) cDNA. In response to a pharmacological ER-stress agent, eIF2B5(R195H) expressing cells exhibited heightened ER-stress response demonstrated by hyper induction of ATF4, GADD34, Bip, PDIA1, PDIA3, PDIA4 and PDIA6 proteins. Moreover, even in the absence of a pharmacological stress agent, eIF2B5(R195H)-expressing cells exhibited high basal levels of ATF4, GADD34 and ER-associated Bip, PDIA1 and PDIA3.SignificanceThe data provide evidence that oligodendroglial-derived cells expressing a mutated eIF2B constantly use their stress response mechanism as an adaptation mean in order to survive. The current study is the first to demonstrate the effects of eIF2B5 mutation on ER homeostasis in oligodendroglial-derived cells.\n\nBODY:\nIntroductioneIF2B is a major evolutionary conserved eukaryotic translation initiation factor. It consists of five different subunits, α, β, γ, δ and ε, (also referred to as subunits 1–5) at a 1∶1 ratio. Subunit 5 is catalytic, whereas subunits 1–4 are regulatory [1]. The eIF2B complex serves as the guanine nucleotide exchange factor of eIF2, another major translation initiation factor that is activated by binding GTP. eIF2-GTP binds and loads the initiator Met-tRNAi onto the small ribosomal subunit, to form the 43S pre-initiation complex that binds to the mRNA. Upon AUG recognition at each round of translation initiation, GTP is hydrolyzed and eIF2-GDP is released. Inactive eIF2-GDP is then recycled by eIF2B to eIF2-GTP, which can serve again to establish another initiation event [2]. eIF2B activity governs the rate of global protein synthesis in the cell. Upon a variety of stress conditions such as starvation, viral infection, oxidative and ER-stress, eIF2 is phosphorylated by one of four kinases (HRI, PKR, GCN2 or PERK) on Ser51 of its alpha subunit, generating a phosphorylated form of eIF2 that functions as a competitive inhibitor of eIF2B. The ratio of eIF-2B to eIF-2 is approximately 0.6 and 0.3 in rat liver and reticulocytes, respectively [3] and 0.3 in rat oligodendroglial-derived cells (Elroy-Stein, unpublished). Therefore, phosphorylation of only part of the total cellular eIF2 could potentially sequester all of the eIF-2B into an inactive eIF2-eIF2B complex. Due to its essential role in protein synthesis under normal and stress conditions, it is surprising that mutations in each eIF2B subunit, rather than being lethal, specifically lead to a neurodegenerative disease in humans. This disease, termed eIF2B-related leukodystrophy, is also known as Childhood Ataxia with CNS Hypomyelination (CACH) or Vanishing White Matter (VWM). The classical form of the disease is characterized by progressive loss of myelin in the CNS, leading to motor and cognitive neurological symptoms that deteriorate upon physiological stress, such as fever and mild head trauma [4], [5]. For obvious reasons, brain glial cells from patients are not available for research. Therefore, cultured primary fibroblasts from patients were used instead, revealing that the eIF2B-mutated fibroblasts are hypersensitive to ER-stress induced by a pharmacological agent [6]. The current study focused on the effect of eIF2B5 mutation on the ER-stress response of an oligodendroglial-derived cell line. The mutation used here (R195H in eIF2B5) is associated with a particularly severe form of the disease, prevalent in the Cree Native American population [7]. The generated oligodendroglial-derived cells expressing eIF2B5(R195H) exhibited heightened ER-stress response demonstrated by hyper-induction of ATF4, GADD34, Bip, PDIA1, PDIA3, PDIA4 and PDIA6 proteins, in response to Thapsigargin, a pharmacological ER-stress agent. The current study provides evidence that oligodendroglial-derived cells are forced to elicit their adaptation capacity in order to survive in the face of a mutation in eIF2B5, since even under normal conditions they express high basal levels of ATF4, GADD34 and ER-associated Bip, PDIA1 and PDIA3.ResultsTo study the molecular consequences of eIF2B5 mutation in a cell type that is physiologically relevant to the CNS, we generated a cellular model for CACH/VWM disease using DDR1 cells, a rat cell line of the oligodendroglial lineage. This goal was achieved by a two-step approach. First, the expression of endogenous rat eIF2B5 gene was down-regulated by stable expression of siRNA directed against its 3′UTR. DDR1 stably expressing the pSuperRetro/si2B5-3′UTR plasmid were termed sh2B5 cells. This cell line was then stably transfected with a plasmid expressing a mutated eIF2B5 cDNA lacking the 3′UTR. Fig. 1A shows the lower level of eIF2B5 protein in sh2B5 cells compared to DDR1 controls. Down-regulation of eIF2B5 expression did not exhibit a significant inhibitory effect on global protein synthesis as demonstrated by 35S]L-methionine/[35S]L-cysteine incorporation rate and polysomal profile analyses (Figs 1B and 2B, sh2B5 cells), indicating that eIF2B5 is expressed in excess amounts by the parental DDR1 cells. To investigate the effect of eIF2B5 down-regulation on cell viability, a time-course experiment was performed in the presence of Tunicamycin (Tun), which blocks the first step of glycoprotein synthesis, thus inhibiting the synthesis of all N-linked glycoproteins and leading to ER-stress. Fig. 1C shows the slight decrease in survival rate of sh2B5 cells compared to DDR1 controls upon 3–24 hr exposure to Tun.10.1371/journal.pone.0003783.g001Figure 1Effect of eIF2B5 down-regulation.A. Identical amounts of total cell protein extracted from DDR1 and sh2B5 cells (stably expressing shRNA against eIF2B5 3′UTR) were subjected to Western blot analysis using antibodies specific for eIF2B5 and p38. B. 5×105 cells were labeled with [35S]-Met/Cys mix for 20 minutes followed by protein extraction, TCA-precipitation and scintillation counting of equal amounts of protein. The data represent average of three independent experiments performed in triplicates+/−SE. C. Control (open bars) or sh2B5 cells (dark bars) were incubated with 10 µg/ml Tunicamycin (Tun) for the indicated times, followed by XTT viability assay. Cell viability is expressed as percentages of viable cells grown in Tun-free medium. The data represent average of three independent experiments performed in triplicates+/−SE.10.1371/journal.pone.0003783.g002Figure 2Effect of eIF2B5 overexpression.A. Identical amounts of total cell protein extracted from DDR1, sh2B5 and sh2B5+2B5(R195H) cells (1, 2, or 3, respectively) were subjected to Western blot analysis using antibodies specific for eIF2B5, Flag and p38. B. Polysomal profiles of the indicated cell lines harvested at their logarithmic growth phase. Top and bottom of the sucrose gradients, free ribosomal subunits (40S, 60S), monosomes (80S) and heavy polysomes are indicated.Since the siRNA was targeted to the 3′UTR of the endogenous eIF2B5, a cDNA of eIF2B5 without the 3′UTR should be resistant to down-regulation by constitutive expression of siRNA in sh2B5 cells. Therefore, sh2B5 cells were stably co-transfected with a vector expressing the Zeocin resistance gene (Zeo) and a vector expressing flag-tagged mutated eIF2B5 cDNA. The R195H mutation of eIF2B5 was chosen because it is associated with a severe form of CACH/VWM disease in human patients [7]. The generated Zeo-resistant stable cell line was termed sh2B5+2B5(R195H). Fig. 2A shows high expression of flag-tagged mutated eIF2B5 protein compared to endogenous eIF2B5 in DDR1 controls and down-regulation thereof in sh2B5 cells. Expression of mutated eIF2B resulted in reduced protein synthesis, as evident from the lower proportions of heavy polysomes in sh2B5+2B5(R195H) cells compared to sh2B5 and DDR1 controls (Fig. 2B). The effect of eIF2B5 mutation on ER-stress response was tested by monitoring ATF4 and GADD34 protein levels at different time points after treatment with Thapsigargin (Tg), which inhibits intracellular calcium pumps and thereby depletes ER calcium stores. In agreement with the kinetics of the unfolded protein response (UPR) [8], [9], an increase in ATF4 and GADD34 protein levels in DDR1 controls and sh2B5 cells was preceded by an increase in eIF2α phosphorylation. However, a more rapid increase in both ATF4 and GADD34 was observed in sh2B5+2B5(R195H) cells, despite similar eIF2α phosphorylation kinetics throughout the three cell lines (Fig. 3A). Also noteworthy is the presence of ATF4 and GADD34 proteins in sh2B5+2B5(R195H) cells even in the absence of a pharmacological ER-stress agent, indicating that mutated eIF2B5 enforces a constant stress state in cells of the oligodendroglial lineage. To further demonstrate the detrimental effect of R195H mutation, the cells were exposed to a low concentration of Tg for 24 h and PARP cleavage was monitored as a marker of apoptosis. Whereas PARP level did not change in DDR1 controls and sh2B5 cells, it was markedly reduced in the mutated cells (Fig. 3B), further demonstrating the increased vulnerability of oligodendroglial-derived cells that express mutated eIF2B5.10.1371/journal.pone.0003783.g003Figure 3ER-stress response.A. The indicated cells were treated with 3 µM Thapsigargin (Tg) for the indicated time points followed by Western blot analysis using antibodies specific for eIF2B5, ATF4, GADD34, phosphorylated eIF2α (eIF2α -P), total eIF2α and p38. B. Cells were treated with 1 µg/ml Thapsigargin (Tg) for 24 h followed by Western blot analysis using antibodies specific for PARP and total eIF2α.We then determined the ER-associated protein profile under normal and stress conditions using the stable isotope labeling by amino acids in cell culture (SILAC) method. All three cell types were cultured under identical conditions except that the entire protein population of the DDR1 control cells was stably labeled by 13C6-Arginine and 15N2, 13C6-Lysine. All cells were analyzed in the absence of exogenous pharmacological stress as well as following 12 and 24 hours incubation with 1 µM Tg. Pairs were created such that experimental cells and SILAC labeled DDR1 control cells were matched at 0, 12, and 24 hours of Tg incubation, under identical experimental conditions. Cells were harvested and mixed with the corresponding labeled DDR1 control cells at a 1∶1 ratio for subsequent ER fractionation. The ER fractions containing labeled and unlabeled proteins were further separated by SDS-PAGE, followed by trypsin digestion of sliced gel bands. The resulting peptides were analyzed by LC-MS/MS for identification and quantification of proteins. Identified peptides were required to have a matched peptide partner of 6 Da greater for arginine-terminating peptides and 8 Da greater for lysine-containing peptides, in addition to other proper combinations for missed cleavage peptides. Only proteins with two or more identified peptides in at least one data set were retained for analysis. Fig. 4 reflects the proteins status under normal conditions, e.g. shows the basal level of some ER resident proteins in sh2B5 and sh2B5+2B5(R195H) cells compared to DDR1 controls. In sh2B5 cells, the level of Bip (also known as glucose related protein 78, GRP78) and four members of the protein disulfide isomerase family (PDIA1, PDIA3, PDIA4 and PDIA6) was not dramatically changed. However, in sh2B5+2B5(R195H) cells, Bip and PDIA3 were significantly above basal levels as compared to DDR1 control cells (2.0±0.09 and 1.54±0.06 fold up-regulation, respectively). Fig. 5 reflects the ER-stress response, i.e. shows the change in the levels of the above proteins at 12 and 24 hours of mild ER-stress (1 µM Tg, compared to 3 µM Tg used in Fig. 3). In sh2B5 cells, the levels of Bip as well as the four PDI proteins were similar to that of the control cells at each of the three time points, indicating that eIF2B down-regulation did not have a significant effect on the response to long-term mild ER stress. However, although the basal levels of Bip, PDIA1 and PDIA3 were high to begin within sh2B5+2B5(R195H) cells, it further increased compared to DDR1 cells under mild ER stress (2.65±0.06, 2.2±0.03 and 1.7±0.08 fold up-regulation at 24 h; 2.1±0.16 and 2.0±0.01 fold up-regulation at 12 h for Bip and PDIA1, respectively) (Fig. 5). Moreover, the levels of PDIA4 and PDIA6 were also higher than their ER-stress induced levels in control cells (1.7±0.09 and 1.9±0.09 fold up-regulation at 24 h, respectively). The hyper-expression of ER-related molecular chaperons in sh2B5+2B5(R195H) cells under mild stress conditions further demonstrates the remarkable effect of this eIF2B mutation on the sensitivity and vulnerability of cells of the oligodendroglial lineage to ER stress.10.1371/journal.pone.0003783.g004Figure 4Proteome profiling of ER proteins at baseline.The SILAC methodology followed by mass spectrometry of microsomal preparations was processed as described in Materials and Methods. Untreated labeled DDR1 cells were mixed at a 1∶1 ratio with unlabeled sh2B5 or sh2B5+2B5(R195H) cells. The level of Bip, PDIA1, PDIA3, PDIA4 and PDIA6 relative to DDR1 control cells is shown.10.1371/journal.pone.0003783.g005Figure 5Proteome profiling of ER proteins following stress.The SILAC methodology followed by mass spectrometry of microsomal preparations was processed as described in Materials and Methods. Labeled DDR1 cells and unlabeled sh2B5 or sh2B5+2B5(R195H) were treated with 1 µM Tg for 0, 12 and 24 h. The unlabeled cells at each time point were mixed at a 1∶1 ratio with the labeled DDR1 controls. The level of Bip, PDIA1, PDIA3, PDIA4 and PDIA6 at each time point (except for PDIA6 at baseline in sh2B5+2B5(R195H) cells) relative to DDR1 control cells is shown.DiscussionThe current study demonstrates a connection between a mutation in eIF2B and abnormal ER stress state in glial cells. The ER is an essential cellular compartment for protein synthesis and maturation, and a Ca2+ storage organelle. Interference with Ca2+ homeostasis, inhibition of disulfide bond formation or protein glycosylation, as well as hypoxia and oxidative stress can all result in accumulation of misfolded or unfolded proteins, leading to ER stress. An impaired ER-stress response is associated with neurodegeneration and other diseases [10]. The unfolded protein response (UPR) is a pro-survival adaptive pathway responsible for restoring perturbed ER homeostasis. The ER chaperone Bip is involved in protein folding and assembly and serves as a key guardian for ER disturbances. It monitors the folding status of proteins, thus controlling the activation state of the three UPR transducers PERK, IRE-1 and ATF6. Active PERK phosphorylates eIF2α, thereby leading to inhibition of global translation. ATF4 mRNA is specifically translated under these inhibitory conditions due to regulatory sequences in its 5′-untranslated region. The induced ATF4 protein, together with the products of activated IRE-1 and ATF6, trans-activate the transcription of ER-stress target genes, including a variety of ER chaperones [11]. Bip and other ER-resident quality control proteins, such as the protein disulfide isomerase (PDI) family that catalyzes the formation, cleavage, and rearrangement of disulfide bonds, are induced during ER stress to handle the accumulation of misfolded proteins within the ER [12]–[14]. The comparative proteomic SILAC approach used in this study showed that, in response to induced ER-stress, the ER-resident proteins Bip and PDIs were induced to a higher degree in sh2B5+2B5(R195H) cells compared to control oligodendrocytes (Fig. 5) and primary fibroblasts from two CACH/VWM patients homozygous for the same mutation (Mintz et al., submitted for publication). PERK−/− cells and cells expressing eIF2α(S51A) [15], [16] are more sensitive to agents that perturb protein folding in the ER, implying that reduced level of eIF2-GTP-tRNAMet\ni has an important role in activation of ER chaperones towards adaptation to ER stress. Our data support this notion as they shows that mutation in eIF2B leads to increased levels of ER-chaperones.In the absence of a pharmacological stress agent, high basal levels of ER-associated Bip, PDIA1 and PDIA3 were detected in eIF2B-mutated oligodendroglial-derived cells (Fig. 4) in contrast to eIF2B-mutated fibroblasts (Mintz et al., submitted for publication). Since cells occasionally experience mild ER stress under normal physiological conditions, the increased basal levels of Bip and PDIs presumably compensate for the occasional increased protein load in the ER. The protective role of Bip in neurodegeneration is supported by several studies [17], [18]. Transient cerebral ischemic pre-activation of the ER-stress response has been shown to be associated with a several fold rise in Bip protein levels and less PERK activation, leading to delayed neuronal cell death [19]. IFN-γ activation of PERK in mature oligodendrocytes was shown to prevent demyelination in a mouse model for multiple sclerosis [20]. Similar to this rationale, increases in the rates of Bip and PDI mRNA synthesis in rat exocrine pancreatic cells precede extensive mRNA expression of secretory proteins induced by glycocorticoid hormones [21]. Neuronal overexpression of PDI has also been shown to promote survival in response to stresses that induce ER dysfunction [22].Perfect timing of translational recovery along the stress response is crucial for stress resistance [23]. Stress-induced gene expression requires programmed recovery from translational repression, a process induced by GADD34, which recruits protein phosphatase 1 catalytic subunit (PP1c) for removal of the inhibitory phosphate group on eIF2α, thereby reversing the shutoff of protein synthesis [8]. The high basal level of GADD34 in the absence of exogenous stress in sh2B5+2B5(R195H) glial cells (Fig. 3A, time zero) is consistent with the notion that these cells suffer from an intrinsically low threshold of protein synthesis, which is supported by lower proportions of heavy polysomes, as indicated in Fig. 2B. Therefore, the activation of GADD34 as a cellular rescue mechanism may serve to maintain the minimal protein translation rate critical for survival. The induction of GADD34 expression by transient cerebral ischemia in resistant cells, but not in specific vulnerable neurons, is in line with this notion [24].Although the cell line of the oligodendroglial lineage used in this study do not myelinate nerve axons, the increased levels of proteins responsible for ER quality control under normal conditions strongly suggest that sh2B5+2B5(R195H) cells suffer from elevated burden of occasional ER overload with client proteins. Although their adaptation capacity enables survival even in the face of mutation in eIF2B, their hyper ER-stress response (Figs 3A and 5) suggests that increased ER burden imposes a bona fide threat. The vulnerability of the eIF2B-mutated glial cells to such threat is evident by their propensity to apoptosis (Fig. 3B) albeit a heightened stress response. The consequence of such a feature on actively myelinating oligodendrocytes in the brain may be deleterious. Supporting the above concept is the active unfolded protein response (UPR) state demonstrated in mixed glia of CACH/VWM patients [25], [26].Materials and MethodsPlasmidsOligoengine™ [27] was employed to design oligonucleotides for long-term gene knockdown following cloning into the pSuperRetro vector (Invitrogen). The designed oligonucleotides 5′GATCCCCCGGAAGTTGCAACTACAGTTTCAAGAGAACTGTAGTTGCAACTTCCGTTTTTGGAAA3′ and 5′AGCTTTTCCAAAAACGGAAGTTGCAACTACAGTTCTCTTGAAACTGTAGTTGCAACTTCCGGGG3′ were annealed to generate a fragment that matches 19 nucleotides region within the 3′UTR of rat eIF2B5 cDNA (region 2361–2379, NM_138866). The fragment was cloned into HindIII and BglII sites of pSuperRetro under the RNA pol1 promoter, to generate pSuperRetro/si2B5-3′UTR. The oligonucleotides 5′CAGCTAGATCTTCACTTGTCCCAGCCC3′ and 5′CCGTGAGAGAAGCTTATGGCGGCC3′ were used to amplify the 2.2 Kbp human eIF2B5 cDNA followed by its cloning into BglII and HindIII sites of pCMV Flag2 (Sigma) to generate pNflag-2B5(wt) which expresses N-terminal Flag-tagged human eIF2B5 from the CMV promoter. This plasmid was used as template for further PCR reactions: the first product was generated using oligonucleotide primers 2b5 Mut sense2 (5′CGGTAGGCGTGTACGGTGGGAG3′) and MutG584A antisense (5′CGTGGCAATGAGTTGGGTGGC3′); the second product was generated using oligonucleotide primers 2b5 Mut antisense (5′TCCTCTGCATCTGGAGGGTGCA3′) and MutG584A sense (5′GCCACCCAACTCATTGCCACG3′). The two PCR products were then annealed and used as template for a third PCR using 2b5 Mut sense2 and 2b5 Mut antisense primers to generate a fragment harboring the human eIF2B5 cDNA with the G584A point mutation. This fragment was then cloned into the ApaI - HindIII 5.57 Kbp fragment of pNflag-2B5(wt) to generate pNflag-2B5(R195H).Cells and stable transfectionsDDR1 rat oligodendrocytes cell line [28] kindly obtained from Bernard Attali (Tel Aviv University) was grown on Dulbeco Modified Eagles Medium (DMEM, Biological Industries) supplemented with 10% Fetal Calf Serum, 100 U/ml penicillin and 100 µg/ml streptomycin (Biological Industries). DDR1 cells were stably transfected using the standard calcium-phosphate precipitation technique. To generate control and sh2B5 cells, DDR1 cells were transfected with the plasmids pSuperRetro (Invitrogen) or pSuperRetro/si2B5-3′UTR, respectively, and selected using medium containing 1 µg/ml puromycin. To generate sh2B5+2B5(R195H) cells stably overexpressing the mutated human eIF2B5 on the background of endogenous rat eIF2B5 knocked-down, sh2B5 cells were co-transfected with pCDNA4/To (Invitrogen) and pNflag-2B5(R195H) followed by selection for Zeocine resistance in medium containing 100 µg/ml Zeocine and 1 µg/ml puromycin.Protein synthesis rate and polysomal profiles analysis6×105 cells were plated in 60 mm plates and labeled for 20 min with 20 µCi/ ml 35S-L-methionine, 35S-L-cysteine mix (NEN, #NEG072) in their growth medium, followed by two washes with cold PBS. Extraction and processing of labeled proteins were performed as previously described [29]. Polysomal profiles analyses were performed according to Sivan et al [30] using extract prepared from 2×107 cells (30 OD260 nm units) per each sucrose gradient. Teledyne ISCO UA-6 UA6 Absorbance Detector was used for analysis.XTT viability Assays5×103 cells were plated in each well of a 96-well plate. Cells were seeded in triplicates. The following day, the medium was replaced with fresh medium containing 10 µg/ml Tunicamycin (Tun). The plates were incubated for different time periods before addition of 50 µl of 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt (XTT) reaction solution (Biological Industries) to each well; Following incubation of 2 hr the optical density was measured at 450 nm. Viability was calculated relative to cells incubated in the absence of Tun.AntibodiesThe following antibodies were used: anti p38 (Sigma, M0800), anti FLAG M2 (Sigma, F3165), anti mouse ATF4 and GADD34 (gifts from David Ron, NYU), anti eIF2α phosphorylated at serine 51 (Research Genetics, Inc.), anti human eIF2B5 (Santa Cruz #28854) and monoclonal antibody specific for eIF2α [31].SILAC Technique and ER stress inductionDDR1 control cells were seeded in 75 cm2 flasks and subcultured in custom-made DMEM medium (Atlanta Biologicals, Lawrenceville, GA) where Arg and Lys were replaced by 13C6-Arg (147.5 µg/ mL), and 15N2, 13C6-Lys (91.25 µg/mL) (Cambridge Isotope Laboratories, Inc., Andover, MA). Cellular proteins were fully labeled after incubation for at least 7 cell doublings. In parallel, the same amounts of experimental cells (sh2B5 and sh2B5+2B5(R195H)) were grown in unlabeled DMEM medium. Both the unlabeled experimental cells and the labeled DDR1 control cells were treated with thapsigargin (final concentration 1 µM) (Tg). At time points 0, 12 and 24 hours cells were harvested followed by mixing of each experimental unlabeled cells at 1∶1 ratio with the labeled DDR1 controls for subsequent subcellular fractionation.Preparation of microsomal fractions, mass spectrometry analysis and database searchThis experiment was performed as thoroughly described by Mintz et al. [14]. Briefly, ER fractions containing labeled and unlabeled proteins were further separated by SDS-PAGE. Thereafter the resulting gel bands were sliced, digested with trypsin and analyzed by LC-MS/MS using LTQ instrument (Thermo Fisher Scientific, Waltham, MA) coupled to Dionex LC-Packing system (Sunnyvale, CA). Protein identification was performed and filtered using BioWorks 3.1 software (Thermo Fisher Scientific). Each file was searched against a subset of the SwissProt database. Full Sequest search results, prior to any filtration, were loaded into ZoomQuant Software. Identified peptides were required to have a matched peptide partner 6 Da greater for arginine terminating peptides, 8 Da greater for lysine containing peptides and other proper combinations for missed cleavage peptides. ZoomQuant software was used to determine protein ratios between 13C6, Arg/, 15N2, 13C6-Lys labeled and unlabeled cells. Peptide ratios were normalized to the median expression in each time point of the experimental group. An average ratio with standard deviation was determined using the multiple peptides detected per protein. Only proteins with two or more identified peptides in at least one data set were retained for analysis. The data was analyzed using GeneSpring GX Analysis Platform (Agilent technologies Inc, SantaClara, Ca). A two sided p-value or z-score was generated for each data point. Proteins with a z-score<0.005 were retained for further analysis. Fold changes of specific proteins were compared at the different time points between the various cells.\n\nREFERENCES:\n1. KimballSRMellorHFlowersKMJeffersonLS\n1996\nRole of translation initiation factor eIF-2B in the regulation of protein synthesis in mammalian cells.\nProg Nucleic Acid Res Mol Biol\n54\n165\n196\n8768075\n2. SonenbergNDeverTE\n2003\nEukaryotic translation initiation factors and regulators.\nCurr Opin Struct Biol\n13\n56\n63\n12581660\n3. KimballSRKarinchAMFeldhoffRCMellorHJeffersonLS\n1994\nPurification and characterization of eukaryotic translational initiation factor eIF-2B from liver.\nBiochim Biophys Acta\n1201\n473\n481\n7803480\n4. SchiffmannRElroy-SteinO\n2006\nChildhood ataxia with CNS hypomyelination/vanishing white matter disease–a common leukodystrophy caused by abnormal control of protein synthesis.\nMol Genet Metab\n88\n7\n15\n16378743\n5. van der KnaapMSPronkJCScheperGC\n2006\nVanishing white matter disease.\nLancet Neurol\n5\n413\n423\n16632312\n6. KantorLHardingHPRonDSchiffmannRKaneskiCR\n2005\nHeightened stress response in primary fibroblasts expressing mutant eIF2B genes from CACH/VWM leukodystrophy patients.\nHum Genet\n118\n99\n106\n16041584\n7. FogliAWongKEymard-PierreEWengerJBouffardJP\n2002\nCree leukoencephalopathy and CACH/VWM disease are allelic at the EIF2B5 locus.\nAnn Neurol\n52\n506\n510\n12325082\n8. NovoaIZhangYZengHJungreisRHardingHP\n2003\nStress-induced gene expression requires programmed recovery from translational repression.\nEmbo J\n22\n1180\n1187\n12606582\n9. KojimaETakeuchiAHanedaMYagiAHasegawaT\n2003\nThe function of GADD34 is a recovery from a shutoff of protein synthesis induced by ER stress: elucidation by GADD34-deficient mice.\nFaseb J\n17\n1573\n1575\n12824288\n10. MarciniakSJRonD\n2006\nEndoplasmic reticulum stress signaling in disease.\nPhysiol Rev\n86\n1133\n1149\n17015486\n11. HardingHPCalfonMUranoFNovoaIRonD\n2002\nTranscriptional and translational control in the Mammalian unfolded protein response.\nAnnu Rev Cell Dev Biol\n18\n575\n599\n12142265\n12. HendershotLM\n2004\nThe ER function BiP is a master regulator of ER function.\nMt Sinai J Med\n71\n289\n297\n15543429\n13. LeeAS\n2005\nThe ER chaperone and signaling regulator GRP78/BiP as a monitor of endoplasmic reticulum stress.\nMethods\n35\n373\n381\n15804610\n14. MintzMVanderverABrownKJLinJWangZ\n2008\nTime series proteome profiling to study endoplasmic reticulum stress response.\nJ Proteome Res\n7\n2435\n2444\n18435558\n15. HardingHPZhangYZengHNovoaILuPD\n2003\nAn integrated stress response regulates amino acid metabolism and resistance to oxidative stress.\nMol Cell\n11\n619\n633\n12667446\n16. ScheunerDSongBMcEwenELiuCLaybuttR\n2001\nTranslational control is required for the unfolded protein response and in vivo glucose homeostasis.\nMol Cell\n7\n1165\n1176\n11430820\n17. ZhaoLLongo-GuessCHarrisBSLeeJWAckermanSL\n2005\nProtein accumulation and neurodegeneration in the woozy mutant mouse is caused by disruption of SIL1, a cochaperone of BiP.\nNat Genet\n37\n974\n979\n16116427\n18. SenderekJKriegerMStendelCBergmannCMoserM\n2005\nMutations in SIL1 cause Marinesco-Sjogren syndrome, a cerebellar ataxia with cataract and myopathy.\nNat Genet\n37\n1312\n1314\n16282977\n19. HayashiTSaitoAOkunoSFerrand-DrakeMChanPH\n2003\nInduction of GRP78 by ischemic preconditioning reduces endoplasmic reticulum stress and prevents delayed neuronal cell death.\nJ Cereb Blood Flow Metab\n23\n949\n961\n12902839\n20. LinWBaileySLHoHHardingHPRonD\n2007\nThe integrated stress response prevents demyelination by protecting oligodendrocytes against immune-mediated damage.\nJ Clin Invest\n117\n448\n456\n17273557\n21. HenselGAssmannVKernHF\n1994\nHormonal regulation of protein disulfide isomerase and chaperone synthesis in the rat exocrine pancreas.\nEur J Cell Biol\n63\n208\n218\n7915986\n22. TanakaSUeharaTNomuraY\n2000\nUp-regulation of protein-disulfide isomerase in response to hypoxia/brain ischemia and its protective effect against apoptotic cell death.\nJ Biol Chem\n275\n10388\n10393\n10744727\n23. BoyceMBryantKFJousseCLongKHardingHP\n2005\nA selective inhibitor of eIF2alpha dephosphorylation protects cells from ER stress.\nScience\n307\n935\n939\n15705855\n24. PaschenWHayashiTSaitoAChanPH\n2004\nGADD34 protein levels increase after transient ischemia in the cortex but not in the CA1 subfield: implications for post-ischemic recovery of protein synthesis in ischemia-resistant cells.\nJ Neurochem\n90\n694\n701\n15255948\n25. van der VoornJPvan KollenburgBBertrandGVan HarenKScheperGC\n2005\nThe unfolded protein response in vanishing white matter disease.\nJ Neuropathol Exp Neurol\n64\n770\n775\n16141786\n26. van KollenburgBvan DijkJGarbernJThomasAAScheperGC\n2006\nGlia-specific activation of all pathways of the unfolded protein response in vanishing white matter disease.\nJ Neuropathol Exp Neurol\n65\n707\n715\n16825957\n27. BrummelkampTRBernardsRAgamiR\n2002\nA system for stable expression of short interfering RNAs in mammalian cells.\nScience\n296\n550\n553\n11910072\n28. EizenbergOFaber-ElmanAGottliebEOrenMRotterV\n1995\nDirect involvement of p53 in programmed cell death of oligodendrocytes.\nEmbo J\n14\n1136\n1144\n7720704\n29. GerlitzGJagusRElroy-SteinO\n2002\nPhosphorylation of initiation factor-2 alpha is required for activation of internal translation initiation during cell differentiation.\nEur J Biochem\n269\n2810\n2819\n12047392\n30. SivanGKedershaNElroy-SteinO\n2007\nRibosomal slowdown mediates translational arrest during cellular division.\nMol Cell Biol\n27\n6639\n6646\n17664278\n31. ScorsoneKAPanniersRRowlandsAGHenshawEC\n1987\nPhosphorylation of eukaryotic initiation factor 2 during physiological stresses which affect protein synthesis.\nJ Biol Chem\n262\n14538\n14543\n3667588"
+}

batch_13/PMC2583339.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2583339",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2583339\nPatients with Unresectable or Metastatic Leiomyosarcoma\n\nAUTHORS: Michelle S. Boyar, Mary Hesdorffer, Mary Louise Keohan, Zhezhen Jin, Robert N. Taub\n\nABSTRACT:\nWe assessed the efficacy of combined temozolomide and thalidomide in patients with unresectable or metastatic leiomyosarcoma in a phase II single-institution trial. Twenty-four patients were enrolled. Temozolomide (150 mg/m2/day for 7 days every other week) was administered with concomitant thalidomide (200 mg/day), and continued until unacceptable toxicity or disease progression. There were no complete responses and two (10%) partial responses. Five patients (24%) had stable disease for at least six months. Fourteen patients (67%) progressed after a median of two-month treatment. The median overall survival (twenty-two assessable patients) was 9.5 months [95% CI 7–28 months]. There were no treatment-related deaths or CTC grade 4 toxicities. Thirteen patients were dose-reduced or discontinued thalidomide due to toxicity. In conclusion, this combination of temozolomide and thalidomide provided disease stabilization in a subset of patients with advanced leiomyosarcoma. We hypothesize that temozolomide is the active agent in this regimen, and should be further studied.\n\nBODY:\n1. INTRODUCTIONSoft tissue sarcomas are rare tumors, representing\nless than 1% of all new cancers diagnosed in the United States\neach year [1]. Complete \nsurgical resection offers the best chance of cure for localized soft tissue\nsarcoma. Patients with unresectable or\nmetastatic disease have a median survival of approximately 12 months. Chemotherapy is the main treatment option for these patients. Doxorubicin, ifosfamide,\nand dacarbazine (DTIC) each have single-agent activity with response rates\napproaching 20% [2–4]. When DTIC was\nadded to doxorubicin, the response rate and overall survival increased at the\ncost of increased toxicity [5].The use of immunohistochemistry and genetic markers\nto better define subsets of soft tissue sarcomas is changing the approach to\nclinical trial design in soft tissue sarcoma. \nThere is increasing evidence that the different subtypes of soft tissue\nsarcoma have distinct biologic characteristics which define metastatic\npotential and response to therapy. The\nremarkable activity of imatinib in patients with gastrointestinal stromal\ntumors is proof of this principle [6]. Instead of large clinical trials which include a variety of histologic subtypes, newer trials are limiting enrollment\nto specific subtypes to assess response rates in a subset of patients. This has led to the identification of active regimens for certain subtypes of soft tissue sarcoma. However, new therapies\nare clearly needed for patients with advanced or metastatic leiomyosarcoma.Over 50 subtypes of soft tissue tumors have been\ndescribed in adults, and leiomyosarcomas are one of the most common malignant\nsoft tissue sarcomas in adults. Leiomyosarcomas\nare derived from smooth muscle cells and can arise in any location. However,\nmore than half are located in retroperitoneal or intraabdominal sites [7].Temozolomide is a cytotoxic alkylating agent that\nwas developed as an oral and less toxic alternative to DTIC. Both temozolomide and DTIC exert their\nantitumor effects through the formation of 5-3-methyl-1-triazenolimidazole-4\ncarboxamide (MTIC), the putative active chemical metabolite of DTIC [8, 9].\nTemozolomide has activity against malignant gliomas and metastatic melanoma\n[10, 11]. Based on its similar mechanism of action to dacarbazine, temozolomide\nhas been evaluated in soft tissue sarcoma using a variety of dosing\nschedules. In several studies which\nincluded a variety of STS subtypes, all of the clinical benefit was seen in\npatients with leiomyosarcoma [12–14].Thalidomide is an agent shown to be useful in a\nvariety of tumors. Its mechanisms of\naction in cancer may be multiple, including direct cytotoxic, antiangiogenic,\nand anti-inflammatory effects [15]. The combination of temozolomide and\nthalidomide has shown promising activity in metastatic melanoma [16] and\nmetastatic neuroendocrine tumors [17], where it led to increased response rates\nwhen compared to temozolomide as a single agent. We conducted this\nsingle-institution phase II trial to assess the efficacy of a combination\nregimen of temozolomide and thalidomide in patients with unresectable or\nmetastatic leiomyosarcoma.2. PATIENTS AND METHODS2.1. Study populationThe study population consisted of adult patients\n(age ≥18) with histologically confirmed, locally advanced,\nunresectable or metastatic leiomyosarcoma. \nPrior treatment with up to three prior systemic chemotherapy regimens\nfor advanced disease was permitted, as was previous dacarbazine treatment.\nPrior radiation therapy was allowed if completed at least 4 weeks prior to\nstudy drug administration. Patients were required to have at least one\nunidimensional measurable lesion documented on computed tomography (CT).\nPreviously radiated lesions were excluded unless there was evidence of disease\nprogression at that site prior to enrollment. Further inclusion criteria\nincluded: SWOG performance status of 0–2, life\nexpectancy >2 months, adequate bone marrow function (absolute neutrophil\ncount [ANC] ≥1500/μL, platelets >70 000/μL, and hemoglobin ≥10 g/dL), adequate hepatic function (bilirubin < upper limit of\nnormal [ULN], AST or ALT <1.5 X ULN,\nalkaline phosphatase <2 X ULN or ≤5 X ULN with documented liver\nmetastases), and adequate renal function (creatinine <1.5 X ULN, BUN <1.5\nX ULN). Exclusion criteria included\nbrain metastases, more than 3 prior chemotherapy regimens for treatment of\nleiomyosarcoma, insufficient recovery from toxicities of prior therapies, other\nserious medical or psychiatric illness, inability to take oral medications,\nprior malignancy other than curatively treated carcinoma in situ of the cervix\nor skin cancer, and pregnant or nursing women. \nPrior chemotherapy, radiation therapy or surgery must have been\ncompleted at least four weeks prior to enrollment. All patients and physicians participated in\nthe Enhanced STEPS program (Enhanced System for Thalidomide Education and\nPrescribing Safety). All patients gave\nwritten informed consent before study entry. \nThe study was approved by the local institutional review board committee\nand was conducted in accordance with the ethical principles stated in the\nDeclaration of Helsinki and the guidelines on good clinical practice.2.2. Study designPatients received treatment with temozolomide at a\ndose of 150mg/m2/day orally on days 1 to 7 and days 15 to 21.\nThalidomide was administered daily at a dose of 200 mg. No dose escalations were permitted. Cycles\nwere repeated every 28 days. Temozolomide was held if ANC remained <1500/μL or platelet count <100 000/μL, and was\nnot resumed until hematologic recovery. When treatment resumed, dose reductions\nfor temozolomide were made based on nadir platelet count or ANC. For nonhematologic toxicities of grade 3 or higher as defined by the National Cancer Institute (NCI) Common Toxicity\nCriteria (CTC) grading system, dosage for the subsequent cycle of temozolomide\nwas reduced to 125mg/m2 for grade 3 toxicity and to 100mg/m2 for grade 4 toxicity. No dose adjustments were required if the toxicity was\njudged to be non-drug-related. Treatment was discontinued if the patient was\nunable to resume therapy within three weeks or if they experienced unacceptable\ntoxicity levels. Dose modifications for thalidomide were made based on\nthalidomide-related toxicity. When thalidomide-related toxicity was noted, the\ndose of thalidomide was reduced to 100 mg daily; patients who experienced\ntoxicity at a dose of 100 mg were taken off thalidomide.Radiologic assessments of all sites of measurable disease with CT scan were performed on\nenrollment and every 8 weeks after starting treatment. Therapy was continued\nuntil evidence of disease progression on CT scan, unacceptable drug toxicity,\ndelay of both drugs for >21 days, or patient-initiated withdrawal for any\nreason. Standard World Health Organization (WHO) response criteria were used. A\ncomplete response (CR) required complete disappearance of all clinically\ndetectable malignant disease for at least four weeks. A partial response (PR)\nrequired ≥50% decrease in the sum of the products of the largest\nperpendicular diameters of a measurable lesion.\nNot all lesions had to show regression to qualify for partial response, but no lesion should have increased by ≥25%, and no new lesions should have appeared. Stable disease (SD) was defined as <50% decrease or <25% increase in the sum of the products of the\nlargest perpendicular diameters of all measurable lesions. Progression of disease (PD) was defined as a ≥25% increase in the sum of products of measurable lesions, clear worsening of\nany evaluable disease, or appearance of any new lesions.2.3. Statistical analysisThis was an open-label, phase II study conducted at Columbia University Medical Center. The primary end point in this trial is\nresponse (complete and partial response) rate. A response probability of 25%\nwould be of interest, while further testing would not be pursued if the\nresponse probability was 5% or lower. \nInitially, 15 patients would be entered. If at least one response was\nobserved, an additional 10 patients will be entered into the trial. Four or more responses out of a total of 25\npatients would be considered as evidence warranting further study of this\nregimen, provided that other factors, such as toxicity and survival were\nfavorable. This design has a\nsignificance level (probability of falsely declaring an agent with a 5%\nresponse probability to warrant further study) of 5%, and a power (probability\nof correctly declaring an agent with a 25% response probability to warrant\nfurther study) of 90%.Time to progression, overall survival, safety, and\ntoxicity were assessed as secondary outcomes. \nTime to progression (date of initiation of treatment to date of\nprogression or death) and overall survival (time from treatment initiation to\nthe date of death) were assessed with Kaplan-Meier estimator. Toxicity and\ncomplications of treatment were assessed based on patient reports of adverse\nevents, physical examination, and laboratory measurements.3. RESULTS AND DISCUSSION3.1. Patient characteristicsA total of 25 patients were entered on the study\nfrom February 2002 to November 2004. One\npatient was found to be ineligible and withdrawn from the study before\nreceiving therapy; that patient was excluded from the analysis. Demographics and baseline disease\ncharacteristics for the 24 treated patients are presented in Table 1. Patients had a median age of 60 years (range from\n27 to 75 years); 22 (92%) were females and 2 (8%) males. Median SWOG performance status was 1. All patients had biopsy-proven leiomyosarcoma. \nThe primary site was uterus in 11 (46%) of patients. Other primary sites included retroperitoneum\n(n = 3, 13%), small or large intestine (n = 3, 13%), ovarian (n = 1, 4%), and unknown\n(n = 3, 13%). All patients had evidence of\nmetastatic disease. Fifteen patients\n(63%) had lung metastases and fourteen patients (58%) had liver\nmetastases. Eight patients (33%) had\nmetastases to bone or soft tissue. \nTwenty patients (83%) had received prior chemotherapy, including treatment\nwith doxorubicin (n = 16, 67%) and/or dacarbazine (n = 4, 17%) and others. The median number of prior regimens was\n2. Fourteen patients (58%) had prior\nsurgery, and three (13%) had prior radiation therapy.3.2. Duration of treatmentOf the 25 patients consented for the trial, 24\npatients received treatment for a median of two months (range from 0.5 to 25\nmonths) (Table 2). Five patients received treatment for 6 months or more. Three patients received treatment for less than 1 cycle due to rapidly progressive\ndisease (n = 2), and withdrawal of consent (n = 1). \nTwo patients required dose reduction of temozolomide due to neutropenia\nand thrombocytopenia. Thirteen patients\nrequired dose reduction of thalidomide (n = 8) or discontinued thalidomide (n = 5) due to neuropathy or fatigue.3.3. ToxicityTwenty one patients were assessable for toxicity, which\nis summarized in Table 3. Grade 2 or 3\nneutropenia developed in three patients (12%), and grade 2 or 3\nthrombocytopenia developed in 4 patients (17%). Dose reduction of temozolomide\nwas required in 3 patients. Grade 2 or 3\nanemia was seen in 3 patients. There\nwere no grade 4 hematologic toxicities.Fatigue was the most common nonhematologic toxicity\nwith grade 2 or 3 fatigue occurring in 13 (54%) patients, and attributed\nprimarily to thalidomide toxicity. Grade 2 or 3 nausea and emesis occurred in 8\n(33%) and 10 (41%) of patients. Neurologic toxicity occurred in 8 patients and\nwas primarily neuropathy, however, one patient had grade 3 vision loss, and two\npatients had grade 3 ataxia. Other\ntoxicities were relatively mild and consisted of grade 2 anorexia in five\npatients (21%), grade 2 constipation in 10 patients (42%), and grade 2 edema in\nthree patients (13%). There were no\ninfectious complications due to treatment.3.4. EfficacyTwenty one patients completed at least one cycle and\nwere assessable for treatment response. \nThere were no complete responses. \nTwo patients experienced durable partial responses. The overall\nradiologic response rate was 10%. Five (24%) patients experienced stable\ndisease, and 14 (67%) had disease progression. The two patients with radiographic responses had durable responses lasting 24 and 25 months before disease progression. For the five patients who had disease stabilization, the median duration of stable disease was 15 months (range from 6\nto 24 months). The median\nfollow-up time for the patient cohort is 41 months (range from 18 to 51\nmonths). Twenty three patients developed\nprogressive disease while receiving therapy, and the median progression-free\nsurvival was 2 months with 95% CI (2 months–6 months) as\nshown in Figure 1. Twenty two patients are assessable for survival and the\nmedian overall survival for the cohort is 9.5 months with 95% CI (7 months–28 months). The one-year survival rate was 40.9% with 95%\nCI (24.8%, 67.6%), and the 2-year survival rate was 26.5% with 95% CI (13.1%,\n53.9%) (Figure 2).The results of this study suggest that the\ncombination of temozolomide given on an alternating weekly schedule with daily\nthalidomide has minimal clinical activity in patients with locally advanced or\nmetastatic leiomyosarcoma. Two (10%) out\nof 21 patients evaluable for response had a partial response by WHO response\ncriteria. Five patients (24%) had\ndisease stabilization for at least 6 months while receiving treatment. The overall response rate in this study is\nlower than what has been reported with DTIC alone [4]. Despite the low response\nrate, both patients on this trial with radiographic responses had durable\nresponses lasting 24 and 25 months before disease progression. The median survival of 9.5 months is typical\nfor this advanced stage, heavily pretreated patient population.Temozolomide has been evaluated in soft tissue\nsarcoma using several different dosing schedules. In our institution, a phase II trial enrolled\n26 patients with unresectable or metastatic soft tissue sarcoma who were\ntreated with temozolomide administered twice daily on a 12-hour schedule for 5\ndays as an oral bolus dose of 200 mg/m2 followed by 9 doses of 90 mg/m2 every four weeks. There were 2 partial\nresponses, 2 mixed responses, and 3 patients with stable disease lasting >6\nmonths, for an overall objective response rate of 8%. All of the patients with clinical benefits\nhad leiomyosarcoma, none of the other soft tissue sarcoma histologic subtypes\nhad any benefit [12]. Another phase II study was conducted in 60 soft tissue\nsarcoma patients (19 with GIST and 41 with other STS histologies). There were no responses seen in the patients\nwith GIST, and 22% had stable disease. \nOf the evaluable patients with other soft tissue sarcomas, there was 1\nCR and 1 PR for a total response rate of 5%, another 33% had stable\ndisease. The median time to progression\nand median overall survival time in patients with other STS was 3.3 months and\n11 months [13]. A phase II study by the\nEORTC treated 31 patients with advanced STS with temozolomide dosed at 750 mg/m2 over 5 days during cycle 1 and then 1000 mg/m2 over 5 days at cycle\n2. There was only 1 partial response for\nan overall response rate of 3.33%. The\nmedian TTP was 2 months and the median OS was 6.75 months [14].The Spanish Group for Research on Sarcomas conducted\na phase II trial of temozolomide given as daily for 6 weeks at a dose of 75 mg/m2/day–100 mg/m2/day. They enrolled 49 patients with pretreated STS\nand 18 patients with GIST. Among the\npatients in the STS arm, there were 7 PR for an overall response rate of\n15.5%. There were 11 patients with\ngynecologic leiomyosarcoma enrolled, 5 of which showed response. The median\nresponse duration was 12.5 months (range from 3.9 to 58 months). In 4 patients the response lasted over one\nyear. The median TTP was 2.2 months and\nmedian OS was 8.1 months. The drug was well tolerated at this dose and grade 3-4 hematologic\ntoxicities were seen in 10–15% of\npatients. This suggested that the\nextended daily dosing schedule had activity in patients with gynecologic\nleiomyosarcoma [18]. Memorial Sloan-Kettering published their experience with\ntemozolomide in patients with pretreated leiomyosarcoma from 2001 to 2004. Twelve patients were treated with continuous\ndaily dose temozolomide, there was one PR which lasted 4 cycles and 4 patients\nhad stabilization of disease from 2 to 5 cycles. Seven patients were treated with bolus dose\ntemozolomide. One patient had a near CR\nwhich lasted 13 cycles and four patients had disease stabilization lasting from\n3 to 16 cycles [19]. The collective\nevidence from these phase II trials suggested that leiomyosarcomas,\nparticularly uterine leiomyosarcomas may be more responsive to treatment with\ntemozolomide than other STS histologic subtypes.In our study, enrollment was limited to patients\nwith leiomyosarcoma, but patients with leiomyosarcoma originating from any site\nwere eligible. Although there were only\ntwo partial radiographic responses among the evaluable patients (10%), the\nduration of response for both patients was prolonged, lasting 24 and 25 months,\nrespectively. Other groups have also\nreported patients with durable responses lasting over 12 months [18]. The\nsignificance of this finding in a disease where the median progression-free\nsurvival ranges from two to three months suggests there is a cohort of\nleiomyosarcoma patients who may derive significant benefit from treatment with\ntemozolomide.The relative contribution of thalidomide to the\nantitumor efficacy in this study is difficult to determine. Although the reported toxicities were\ngenerally mild, grade 2 thalidomide-related toxicities such as fatigue,\nconstipation, and neurologic toxicity were seen in up to 50% of patients. Dose reduction or discontinuation of\nthalidomide was required in 13 (54%) of patients. Patients who continued treatment with\ntemozolomide had continued benefit even after stopping the thalidomide. We\nconcluded that thalidomide was poorly tolerated, and it is unlikely to add\nadditional antitumor efficacy in this patient population. Our results and those of other groups suggest\nthat temozolomide is the active agent in this regimen.4. CONCLUSIONSIn conclusion, temozolomide offers another option to\nconsider in the treatment of patients with advanced or metastatic\nleiomyosarcoma. Its benefits include the\nconvenience of oral administration and an improved side-effect profile compared\nto traditional chemotherapy regimens for advanced soft tissue sarcoma. Although it does not produce high response\nrates, there are a number of patients who have disease stabilization from\nmonths to years even without radiographic responses. There are very few chemotherapy regimens that\noffer the possibility of benefit in this patient population, and our data support\nthe consideration of temozolomide in the treatment of progressive disease. Although several dosing schedules have been\ntested in trials (continuous daily dosing, bolus dosing and biweekly dosing),\nthere is not sufficient evidence to support adopting an alternative dosing\nschedule, therefore, we recommend using the 5-day bolus dosing regimen for\ntemozolomide.Future investigation should focus on determining the\nbiologic and molecular characteristics of leiomyosarcomas that can predict\nresponse to temozolomide. Methylation of the DNA repair protein\n06-methylguanine-DNA-methyltransferase (MGMT) has been identified as a\npredictor of response to temozolomide treatment in patients with glioblastoma\nmultiforme [20]. Epigenetic silencing\nof MGMT and/or other key regulatory genes in tumor cells may play a role in\ntemozolomide resistance, and in the pathogenesis of soft tissue sarcomas. Because of the potential therapeutic benefits\nto those patients who may respond to temozolomide, investigation of mechanisms\nof response and resistance to this drug warrants further investigation in\nleiomyosarcomas.\n\nREFERENCES:\n1. JemalAMurrayTWardECancer statistics, 2005CA: A Cancer Journal for Clinicians2005551103015661684\n2. EdmonsonJHRyanLMBlumRHRandomized comparison of doxorubicin alone versus ifosfamide plus doxorubicin or mitomycin, doxorubicin, and cisplatin against advanced soft tissue sarcomasJournal of Clinical Oncology1993117126912758315424\n3. AntmanKHRyanLEliasAShermanDGrierHEResponse to ifosfamide and mesna: 124 previously treated patients with metastatic or unresectable sarcomaJournal of Clinical Oncology1989711261312491883\n4. GottliebJABenjaminRSBakerLHRole of DTIC (NSC-45388) in the chemotherapy of sarcomasCancer Treatment Reports1976602199203769974\n5. BordenECAmatoDARosenbaumCRandomized comparison of three adriamycin regimens for metastatic soft tissue sarcomasJournal of Clinical Oncology1987568408503585441\n6. DemetriGDvon MehrenMBlankeCDEfficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumorsThe New England Journal of Medicine2002347747248012181401\n7. BrennanMFCasperESHarrisonLBDeVitaVTJr.HellmanSRosenbergSASoft tissue sarcomaCancer Principles and Practice of Oncology19975th editionPhiladelphia, Pa, USALippincott-Raven17381788\n8. TsangLLHQuartermanCPGescherASlackJAComparison of the cytotoxicity in vitro of temozolomide and dacarbazine, prodrugs of 3-methyl-(triazen-1-yl)imidazole-4-carboxamideCancer Chemotherapy and Pharmacology19912753423461998993\n9. StevensMFGNewlandsESFrom triazines and triazenes to temozolomideEuropean Journal of Cancer Part A199329710451047\n10. BowerMNewlandsESBleehenNMMulticentre CRC phase II trial of temozolomide in recurrent or progressive high-grade gliomaCancer Chemotherapy and Pharmacology19974064844889332462\n11. MiddletonMRGrobJJAaronsonNRandomized phase III study of temozolomide versus dacarbazine in the treatment of patients with advanced metastatic malignant melanomaJournal of Clinical Oncology200018115816610623706\n12. TalbotSMKeohanMLHesdorfferMA phase II trial of temozolomide in patients with unresectable or metastatic soft tissue sarcomaCancer20039891942194614584078\n13. TrentJCBeachJBurgessMAA two-arm phase II study of temozolomide in patients with advanced gastrointestinal stromal tumors and other soft tissue sarcomasCancer200398122693269914669291\n14. WollPJJudsonILeeSMTemozolomide in adult patients with advanced soft tissue sarcoma: a phase II study of the EORTC Soft Tissue and Bone Sarcoma GroupEuropean Journal of Cancer199935341041210448291\n15. KumarSRajkumarSVThalidomide and lenalidomide in the treatment of multiple myelomaEuropean Journal of Cancer200642111612162216750621\n16. HwuW-JKrownSEMenellJHPhase II study of temozolomide plus thalidomide for the treatment of metastatic melanomaJournal of Clinical Oncology200321173351335612947072\n17. KulkeMHStuartKEnzingerPCPhase II study of temozolomide and thalidomide in patients with metastatic neuroendocrine tumorsJournal of Clinical Oncology200624340140616421420\n18. Garcia del MuroXLopez-PousaAMartinJA phase II trial of temozolomide as a 6-week, continuous, oral schedule in patients with advanced soft tissue sarcoma: a study by the spanish group for research on sarcomasCancer200510481706171216134177\n19. AndersonSAghajanianCTemozolomide in uterine leiomyosarcomasGynecologic Oncology20059819910315916799\n20. HegiMEDiserensA-CGorliaTMGMT gene silencing and benefit from temozolomide in glioblastomaThe New England Journal of Medicine200535210997100315758010"
+}

batch_13/PMC2584038.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584038",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584038\nAUTHORS: Sundus A Hussein, Monalisa Sur\n\nABSTRACT:\nBackgroundMyxopapillary ependymomas (MPE) occur in the filum terminale of the spinal cord, but also present in extra-spinal locations such as subcutaneous tissue and brain. They are slow growing grade I gliomas. Areas of solid growth pattern with aggregates of cells with \"epithelioid morphology\" seen in MPE can mimic metastatic carcinoma. The presence of occasional cells with clear cytoplasm and morphology can resemble Chordoma. Diagnosis can be missed due to these morphological similarities, which could affect patient management and hence, long term survival.Case presentationWe describe two cases of MPE with cytokeratin (AE1 AE3, CAM 5.2, Cytokeratin 7 and cytokeratin 20) expression.ConclusionMPE can be positive for Cytokeratins (CAM 5.2, AE1 AE3, CK7) and focally for EMA, which could be misdiagnosed as metastatic carcinoma. In cases demonstrating epithelioid and clear cell morphology, the diagnosis of MPE should be made in conjunction with histology, proper immunohistochemical profile which includes co-expression of GFAP, S-100 protein and epithelial markers, radiologic findings and site. It is important to be aware of the cytokeratin profile in MPE to avoid erroneous diagnosis with other tumour entities.\n\nBODY:\nBackgroundMyxopapillary ependymomas (MPE) generally occur in the filum terminale of the spinal cord, however, they have been described in extra-spinal locations such as subcutaneous tissue [1] and brain [2]. They are slow growing gliomas corresponding to WHO grade I [3]. The classical morphology shows papillae embedded in a myxoid/mucoid background. Each papilla contains a central fibrovascular core and is lined by cuboidal to elongated cells, occasionally showing clear cytoplasm. Microcysts are also present. The myxoid background contains both neutral and acidic mucopolysaccharides. Prognosis depends on the completeness of excision [4]. However, areas of solid growth pattern with aggregates of cells with \"epithelioid morphology\" can also be encountered which can mimic metastatic carcinoma. In addition, the presence of cells with clear cytoplasm can also be mistaken for chordoma. Diagnosis is easily missed due to these morphological similarities, which could affect patient management and hence, long term survival. A good number of studies have reported the immunophenotype of MPE and differential diagnosis of MPE aided by immunohistochemical stains [5-7]. Several studies have reported absence of cytokeratin expression in MPE [8-10].Case presentationWe describe two cases of MPE with cytokeratin expression. Two female patients, aged 46 and 72 years respectively presented with low back pain. Magnetic resonance imaging (MRI) of the spine indicated the presence of a large intra spinal mass in the central canal extending from L3–L5 in the former patient and a tumour attached to filum terminale in the latter. Both underwent neurosurgical removal of the tumor.Pathological findingsMicroscopic examination of the tissue obtained showed the classical morphological features of MPE with formation of pseudopapillae and pseudorosettes embedded in a myxoid stroma. The cells, which made up the pseudorosettes, had epithelioid morphology with occasional cells showing clear cytoplasm (figure 1). In addition, cribriform areas (figure 2), solid sheets and cords of cells resembling a carcinoma were also present (figure 3). The differential diagnosis was, MPE, metastatic carcinoma and chordoma.Figure 1MPE with areas of epithelioid morphology with occasional cytoplasmic clearing. ×200; H&E.Figure 2MPE with areas showing a cribriform pattern of cells resembling a carcinoma. ×400; H&E.Figure 3MPE demonstrating solid sheets and cords of cells. ×200; H&E.Immunohistochemically, the neoplastic cells showed strong, diffuse positive reaction with S-100 protein (figure 4a) and glial fibrillary acidic protein (GFAP) (figure 4b), indicating the glial nature of the lesion. The tumour cells in both cases showed strong positivity for cytokeratin markers, AE1AE3 (figure 5a), CAM 5.2 (figure 5b) and focally for cytokeratin 7. Both cases demonstrated a low proliferative index (< 2%) with Ki-67. The neoplastic cells were negative for cytokeratin 20, NSE, synaptophysin and neurofilament. Epithelial membrane antigen (EMA) focally stained luminal tips of occasional tumor cells. Histochemically, areas of mucoid degeneration were positive for alcian blue and periodic-acid-schiff (PAS) stains.Figure 4A: Positive S-100 protein stain in the neoplastic cells indicating glial nature of the lesion. ×200. 4B: GFAP positivity in the neoplastic cells. ×200.Figure 5Positive epithelial markers; AE1AE3 (A) and CAM5.2 (B) in the neoplastic cells.DiscussionMPE is a glial tumour occurring almost exclusively in the region of the cauda equina and considered to be one of the most frequent primary tumours to occur in this location [3,11]. MPE was first described as a separate entity by Kernohan in 1931 [11] with isolated case reports subsequently appearing in the literature [[12]&[13]]. The clinical presentation depends on the location of the tumour. The majority of cauda equina and filum terminale tumours present with low back pain due to nerve root compression similar to our cases. Lower limb weakness and sphincter dysfunction are the two other common clinical manifestations. Intramedullary MPE arise from the ependymal lined cells of the filum terminale. The proposed histogenesis of extramedullary MPE in this location is the presence of ependymal rest of the neural tube during canalization and retrogressive differentiation [14,15] or from ectopic ependymal cell [16]. Recent study has suggested that radial glia is the cell of origin of ependymoma [17,18].The classic morphology is easily recognizable, comprising multiple papillae covered by flattened to cuboidal cells embedded in a myxoid stroma and forming pseudorosettes. However, there are cases where the tumour obtains a solid growth pattern with aggregates of cells with \"epithelial morphology\" which in addition may show clearing of the cytoplasm. In such instances, metastatic carcinomas of renal origin and chordoma have to be ruled out by adequate clinical history and immunohistochemical stains.It is well known that MPE stain positive with GFAP and S-100 protein. Cytokeratin positivity in MPE has been a subject of controversy with few cases reported showing positive cytokeratin expression in MPE [5,19,20]. Both our cases were positive for CAM5.2, AE1AE3, EMA and CK7 but negative for CK20.Morphological variations in MPE can resemble metastatic carcinoma. An erroneous diagnosis may subject the patients to unnecessary metastatic work-up and additional adjuvant therapy. In addition, one should also consider the psychological implication of diagnosing a carcinoma. MPE has a better prognosis with tendency for late recurrence, except for some cases with aggressive behavior and seeding to the CNS [21]. MPE with clear cell changes and positive staining for S-100 protein, keratins and EMA can be misdiagnosed as chordoma, the latter being positive for Cytokeratin markers and S-100 protein but negative for GFAP. Although, metastatic carcinomas are positive for cytokeratins, these are consistently negative for GFAP.MPE can be positive for Cytokeratins (CAM 5.2, AE1 AE3, and CK7) and focally for EMA. Given the similarities in morphology between metastatic carcinoma, chordoma and MPE, the diagnosis of MPE should be made in conjunction with clinical history including tumour location, immunohistochemical profile (co-expression of GFAP, S-100 protein and epithelial markers in MPE) and radiological findings. The recommended panel should include Cytokeratins (positive in MPE, metastatic carcinoma and chordoma), GFAP (positive in MPE and negative in metastatic carcinoma and chordoma), and S-100 protein (positive in MPE, chordoma but negative in most metastatic carcinomas).The majority of MPE are slow growing gliomas with a tendency for local recurrence. Prognosis depends on complete surgical resection of the tumor. The risk of metastasis is very low and rare. The overall prognosis of MPE is much better than metastatic carcinoma and the latter has to be ruled out as they can have similar morphological features.In conclusion, we reemphasize the importance of using a panel of immunohistochemical stains to differentiate between MPE and other tumour entities to avoid misdiagnosis.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsBoth authors had contributed equally to the preparation of the manuscript. Both authors read and approved the final manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2584046.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584046",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584046\nAUTHORS: Tapio Mappes, Alessandro Grapputo, Harri Hakkarainen, Esa Huhta, Esa Koskela, Raimo Saunanen, Petri Suorsa\n\nABSTRACT:\nBackgroundSince Darwin's pioneering work, evolutionary changes in isolated island populations of vertebrates have continued to provide the strongest evidence for the theory of natural selection. Besides macro-evolutionary changes, micro-evolutionary changes and the relative importance of natural selection vs. genetic drift are under intense investigation. Our study focuses on the genetic differentiation in morphological and life-history traits in insular populations of a small mammal the bank vole Myodes glareolus.ResultsOur results do not support the earlier findings for larger adult size or lower reproductive effort in insular populations of small mammals. However, the individuals living on islands produced larger offspring than individuals living on the mainland. Genetic differentiation in offspring size was further confirmed by the analyses of quantitative genetics in lab. In insular populations, genetic differentiation in offspring size simultaneously decreases the additive genetic variation (VA) for that trait. Furthermore, our analyses of differentiation in neutral marker loci (Fst) indicate that VA is less than expected on the basis of genetic drift alone, and thus, a lower VA in insular populations could be caused by natural selection.ConclusionWe believe that different selection pressures (e.g. higher intraspecific competition) in an insular environment might favour larger offspring size in small mammals. Island selection for larger offspring could be the preliminary mechanism in a process which could eventually lead to a smaller litter size and lower reproductive effort frequently found in insular vertebrates.\n\nBODY:\nBackgroundPopulation genetics models [1] emphasise the importance of different stochastic processes related to geographical isolation, such as the founder effect and genetic drift, on the differentiation of small populations. Severe reductions of genetic variability and population size are suggested to favour drift and constrain natural selection [2]. However, recent studies [3,4] have shown that natural selection could be the dominant diversifying agent in the evolution of quantitative traits. Irrespective of whether we consider natural selection or random drift as the major causative agent in evolution, together they may allow populations on isolated islands to evolve a collection of traits that distinguish them from their mainland relatives.Differences between mainland and island populations of mammals have often been referred to as the Island rule or Island Syndrome [5-11]. The most familiar pattern on islands is the evolution of larger-bodied species towards a smaller size and smaller-bodied species towards a larger size [12]. In small mammals, the pattern also includes reduced reproductive output, higher survival rate, and differences in behaviour (see reviews in [9,13]). Recently the generality of the island rule has been criticized by Meiri and colleagues [14,15]. According to their phylogenetic analyses, the increase in body size might only hold true in some mammalian groups (e.g. murid rodents). They argued that earlier reviews were biased by a few extreme examples in some mammalian groups (e.g. elephants), and these reviews might have ignored many examples where body size has not changed. Furthermore, it has been suggested [15,16] that future studies should be focused more clearly on the possible differences in natural selection caused by island characters (size and isolation) [17], ecological mechanisms (e.g. predation rate and inter/intraspecific competition) [18] and species specific mechanisms (e.g. evolutionary constraints caused by additive genetic variation).Here we focused on the possible genetic differentiation in morphological and life-history traits between insular and mainland populations of a small mammal, the bank vole Myodes glareolus. Our previous studies have indicated a large additive genetic variation in reproductive traits (e.g size and number of offspring) in the mainland population of our study species [19]. Furthermore, we have shown that the rapid selection caused by intraspecific competition can significantly regulate the proportion of genetic reproductive tactics (high or low reproductive effort) in mainland populations [20]. In the present study, we tested the hypothesis that the selection for reproductive tactics might differ in an insular environment, causing genetic differentiation of insular populations from mainland ones. We also tested the relative importance of natural selection and random drift on genetic differentiation. In these analyses fitness-related additive genetic variation was compared to neutral genetic variation (neutral genetic markers) [21].ResultsWe observed neither morphological differentiation in the breeding females, nor a significant difference in their reproductive effort between insular and mainland populations of the bank vole. By contrast, females that originated from the islands produced significantly larger offspring than those from the mainland (Fig 1, Table 1), indicating either environmental or genetic responses of mothers to the insular environment. The size of island or distance to mainland did not affect the breeding characters of insular females (see Additional file 1).Figure 1Island offspring were significantly heavier than mainland ones in the field (see statistics in Table 1). In further lab analyses, males originating from island populations fathered significantly heavier offspring than mainland fathers (paternal effects) (see statistics in Table 2). (■, island; □, mainland).Table 1Characteristics (mean ± SE) of breeding females from mainland and island populations.island (n = 51)mainland (n = 33)Fndf, ddfp-valueBody mass of offspring (g)1.85 ± 0.011.74 ± 0.024.541,23.30.044Litter size5.0 ± 0.25.6 ± 0.33.091,820.083Litter mass (g)9.25 ± 0.379.74 ± 0.440.8131,800.370Reproductive effort (1)0.76 ± 0.030.81 ± 0.041.201,820.277Reproductive effort (2)0.40 ± 0.020.42 ± 0.21.201,29.10.392Post-partum head width of mother (mm)13.2 ± 0.113.3 ± 0.10.401,33.90.529Post-partum body mass of mother (g)23.2 ± 0.423.2 ± 0.40.021,32.10.896Only the test statistics of origin are presented from the mixed model analyses (SPSS 14.0). Origin of mother was used as a fixed factor and population as a random factor in the analyses. Mother was also used as a random factor in the analysis of offspring body mass. Reproductive effort (1) = L × Mo0.75/Mm0.75, and Reproductive effort (2) = L × Mo/Mm; where L is litter size; Mo is mean pup mass at birth, and Mm is weight of the female after delivery. ndf = numerator degrees of freedom, ddf = denominator degrees of freedom.As the phenotypic differences in neonate size do not necessarily imply micro-evolutionary differentiation, the genetic basis of offspring size was further analysed in the laboratory using paternal half sib analyses. Males originating from island populations fathered significantly heavier offspring at birth than mainland fathers (Fig 1, Table 2), when both were mated to a common stock of females. Moreover, the analyses indicated significant additive genetic variance only among the mainland fathers (VA ± S.E. = 0.047 ± 0.020 and h2 ± S.E = 0.96 ± 0.41), whereas genetic variance was zero or very low among island fathers (Table 3). Heritability estimates differed significantly between the mainland and island fathers (t = 3.67, df = 26, P < 0.002) (Table 3).Table 2Mixed Model Analyses (SPSS) for differences in body mass of offspring fathered either by mainland or island males.SourcendfddfFWald ZPOrigin of sire16.810.4880.015Sire2748.31.8280.033Dam (sire)4827814.631< 0.001Population0.2400.810Origin of sire, sire and dam (within sire) were used as fixed factors and population as a random factor. ndf = numerator degrees of freedom, ddf = denominator degrees of freedom.Table 3Genetic basis of the birth mass of offspring sired by fathers from two different origins.SourcedfMSFPVPVA ± S.E.h2 ± S.E.MainlandSire70.2982.3840.0470.0490.047 ± 0.020*0.96 ± 0.41*Dam (sire)280.14013.010< 0.001Error1710.011IslandSire50.3281.6760.1860.0810.012 ± 0.012n.s..0.14 ± 0.24n.s..Dam (sire)190.29416.918< 0.001Error1070.017Phenotypic variances (VP), additive genetic variances (VA) and heritabilities (h2) of body mass were estimated by half-sib analyses from the variance components among sires separately for the different environments. Standard errors of VA and h2 were estimated using the formula in [59]. The effect of population (P > 0.39 in both origins) was included to the models as a random factor.* P < 0.05, n.s.= non-significant.The relative importance of natural selection and random drift on offspring size divergence can be tested by comparing among-island differences based on additive genetic variation and measures based on neutral marker genes (Fst) [21]. Here the island populations were different according to neutral markers (Fst ± S.E. = 0.177 ± 0.021; P < 0.001) but not according to the additive genetic variance (VA ≈ 0) (Table 3). This means that the VA was less than expected on the basis of genetic drift alone. Mainland populations (localities) did not differ according to neutral markers (Fst ± S.E. = 0.008 ± 0.006; n.s.). Fst values differed significantly between mainland and island populations (P = 0.001).Discussions and conclusionThe present results are in agreement with earlier suggestions that offspring size might be the first life-history characteristic to evolve in insular populations of vertebrates [22,23]. Large neonate size can be a local adaptation to an insular environment, where many ecological selection pressures, e.g. intra- and inter-specific competition and predation, differ from larger mainland populations [9]. Here we could not directly test the biological significance of larger offspring size (0.11 g difference between the island and mainland populations), but according to our previous research, the increase in female offspring size from 1.74 g to 1.85 g may have an important impact on the future fitness of bank vole offspring [19]. For example, the age of first breeding would decreased by 5 days, calculated by a linear regression model (y = -46.11x + 181; Fig 1b in [19]), and the probability of breeding was increased from 0.68 to 0.78, calculated by a logistic regression model (ln(y/1-y) = 4.87x - 7.70; Fig. 2a in [19]).The breeding density of territorial bank vole females seems to be lower in open populations (approx. 10 females/ha maximum) [24] compared to the artificially enclosed populations (e.g. over 20 females/ha) [25]. If intra-specific competition between breeding females is similarly increased on (enclosed) islands, it could be an important selective force for the larger offspring size at birth. In general, the precise mechanisms of selection and when intra-specific competition may favour the fitness of larger offspring or adult size have not yet been tested in insular populations [18]. There exists only indirect evidence that competition is lower on larger islands, decreasing selection for body size in mammals [17]. Here we did not find any effect of island area or isolation on the measured morphological or life-history traits. We suggest that compared to the earlier studies, the present islands were relatively small and variation in size might be too low to find the significant effects of island characters. However, our findings are in agreement with most of the earlier studies, which do not support the importance of island characters for the differentiation of insular individuals from their mainland descendants [26-30].The phenomenon to produce large offspring can also be linked to the genetic dispersal tactics of individuals [22,31]. The individuals which are more prone to take risks, e.g. by dispersing over large open ice to islands, might also genetically differ according to many other traits besides offspring size. The change in offspring size could then be a by-product of selection on other traits. This possibility cannot be ruled out before knowing more about population genetics and the behaviour of individuals (e.g. extinctions, dispersal and mutation rates) in our island system. Even the loss of additive genetic variation in offspring body mass that we have shown here can be explained by non-adaptive genetic processes, e.g. random drift. However, our analyses of differentiation in neutral marker loci (Fst) indicate that VA is less than expected on the basis of genetic drift alone [21]. Here, we were not able to analyse additive genetic variance or variance in neutral markers within single islands, as our estimates of additive genetic variance were based on a few individuals per location. Therefore, comprehensive comparisons of differentiation in neutral marker loci and additive genetic variation in quantitative traits are still lacking [21].In contrast to the island populations, additive genetic variation in offspring body mass has been observed in mainland populations [[19] and here]. Additive genetic variation was also higher than expected by the variation in neutral markers. Similarly, a large additive genetic variation has also been show in several other life-history traits in different systems [32-34]. In agreement with our findings, the additive genetic variation is also usually higher than the variation in neutral markers [21]. Additive genetic variation in life-history traits can be maintained by a trade-off (negative genetic correlation) between two traits [35], in this case between the size and number of offspring [19], especially when natural selection favours one trait under current conditions and another at a later date [36]. Strong annual and multi-annual density fluctuations (cyclicity) are suggested to maintain additive genetic variation and even genetic polymorphism in life-history traits, particularly in rodent populations [37]. In fact, our recent findings with the bank vole indicate that density- and negative frequency-dependent selection favour the genetically different allocation tactics between the size and number of offspring [20].We supposed that if density fluctuations and other ecological parameters related to them are more stable in insular environments, selection for large offspring size could also be stable long term. Theoretically, strong selection might decrease additive genetic variation found in our islands [34,38,39]. Moreover, selection for large offspring size might simultaneously decrease litter size [19]. A future goal would be to show whether large neonate size is an adaptation to insular environments, and also how genetic differentiation in this particular trait is related to other life-history and behavioral traits (e.g. litter size, reproductive effort, adult size, longevity, disperal) as well as their evolution.MethodsStudy speciesThe bank vole is a common mammal in coniferous forests of northern Europe [40]. The breeding period in central Finland lasts from May to September [25]. Pregnancy lasts for 19–20 days and pups are weaned until the age of three weeks [41]. In addition to remarkably large phenotypic [25] and additive genetic variation [19] in litter size (2–10) and offspring size (1.3–2.5 g), a trade-off (i.e. both negative phenotypic and genetic correlations) also exists between these traits [19]. Furthermore, a larger size at birth [19,24] and at weaning [41] increases the probability of maturation (i.e. breeding in summer they are born) in juvenile females. Reproducing bank vole females are territorial, while home ranges of males and non-breeding individuals overlap [42-44]. The density of breeding females is limited due to their territoriality [45].Field sampleThe study was carried out in central Finland (62° 37'N, 26° 20'E). The data are based on 898 individuals caught from 37 islands (0.12–70 ha) in lake Konnevesi and from 20 mainland localities within 5 km of the lake during the summer 1999. Pregnant females were caught from 20 islands and all mainland areas, and thus only these populations were included in the present study. The shortest distance between neighbouring study islands varied between 50 to 500 m and the mean distance from islands to the mainland was 631 m (S.E. = 75 m). Dispersal between islands during summer is very low, indicated by a separate study during the autumn where we recaptured 106 island individuals (three to five months from the first capture); none of these individuals left their home islands. The mainland trapping areas were sufficiently far apart (mean ± S.E = 832 ± 129 m, range 300 – 2 000 m) to decrease dispersal between different mainland localities. We were not able to estimate dispersal rate between mainland localities, but according to earlier studies [25], it should be very low especially among territorial breeding females. The individuals were caught using Ugglan multiple-capture live-traps. On the smaller islands (< 3.5 ha), trap lines were set at c. 20 m intervals (25 traps/ha). On the larger islands (> 3.5 ha) and in the mainland localities, individuals were trapped using the small quadrat sampling method (modified from [46]: each quadrat area (side = 15 m) contained four trap sites (4 traps/area). Pre-baited traps were left open for two nights, after which they were set and checked over three consecutive days. Trappings were carried out from early May to the end of the breeding season in September.All trapped voles were taken into the laboratory where each individual was sexed, weighed to the nearest 0.1 g, and measured for maximum head width to the nearest 0.1 mm with a digital calliper [47]. Males and non-pregnant females were then released back to the field. 51 pregnant females from 20 islands and 33 pregnant females from 20 mainland localities were kept in the laboratory prior to producing a litter [41]. Immediately after birth, pups were weighed with an electronic balance to the nearest 0.01 g, and width of head was measured using a stereomicroscope. The mothers were released with their pups at their point of capture [41]. The proportion of breeding females did not differ between the islands (21.2%, n = 241) and mainland (26.8%, n = 123)(G = 1.45, df = 1, P = 0.229).The reproductive effort of females was estimated using two formulas. First, we used the formula:RE(1) = L × Mo0.75/Mm0.75, where L is litter size; Mo is pup mass at birth, and Mm is weight of the female after delivery [22,48]. In this formula, energy requirements to produce offspring is calculated relative to the allometric requirement of the mother (assuming standard metabolism increases to the 0.75 power of mass for mammals) [48-51]. The mothers were released with their pups at their point of capture [41]. Since the theoretical and empirical basis of 0.75 scaling is still under debate in the literature of animal metabolism (see e.g. [52,53]), we also used a more simple formula: RE(2) = L × Mo/Mm, where litter mass was divided by mother body mass.Analyses of quantitative genetics and neutral markersThe genetic basis of offspring characteristics was analyzed in the laboratory. In the analyses, we compared the effect of male origin (island/mainland) on the characteristic of their offspring. Bank vole males do not rear their offspring and male quality does not affect the amount of maternal care [54], hence we can assume that the genetic analyses are not biased by covariances between non-genetic maternal effects and genetic paternal effects.We mated a random sample of males from the mainland and island populations with two to three randomly chosen females that originated from a separate lab colony (Fig. 1, Table 2). The females mated with the island males did not differ from the females mated with mainland males (head width: (t = 0.255, df = 73, P = 0.800; body mass: t = 1.651, df = 73, P = 0.103). Body mass and head width of the males that originated from islands did not differ from the males that originated from the mainland (Mixed model analyses (SPSS);origin fixed and population random factor; head width: F1,15.7 = 1.21, P = 0.287; body mass: F1,10.9 = 1.13, P = 0.311). To obtain an estimate of heritability (h2) and additive genetic variance (VA) for the body mass of offspring in the mainland and island populations, we performed standard half-sib analyses (Table 3). The analyses included 16 sires, 44 dams and 216 progeny from mainland populations and 12 sires, 31 dams and 140 progeny from island populations. We were only able to include one to three males per island or mainland location to the analyses, so our estimates of genetic variance indicate the variance among the whole island system, not variance within single islands.Individuals were genotyped with six microsatellite loci, which are highly variable in the bank vole [55]. To obtain comparable analyses, we used mature males from the same islands (62 males from 9 islands) and mainland localities (75 males from 10 areas) as we used in the analyses of quantitative genetics. An estimate of population structure was obtained using Fst [56], calculated using FSTAT ver 2.9.3 [57]. Standard error was obtained with Jacknifing over loci. The significance of population differentiation was tested by log-likelihood G-statistics and the test was based on 1000 randomization of genotypes within samples [58].Authors' contributionsTM, HH, EH, EK, RS and PS planned the study and performed the field trappings. TM performed the analyses of quantitative genetics. AG was responsible for the analyses of molecular genetics.Supplementary MaterialAdditional file 1Mixed Model Analyses (SPSS) for the effects of island size and island distance to mainland on the breeding characters of insular females. Size of island and distance to mainland were used as fixed factors (covariates) and population as a random factor in the analyses. ndf = numerator degrees of freedom, ddf = denominator degrees of freedom.Click here for file\n\nREFERENCES:\n1. WrightSEvolution of Mendelian populationsGenetics1931169715917246615\n2. FrankhamRBallouJDBriscoeDAIntroduction to conservation genetics2002Cambridge: Cambridge University Press\n3. CleggSMDegnanSMMoritzCEstoupAKikkawaJOwensPFMicroevolution in island forms: the roles of drift and directional selection in morphological divergence of passerine birdEvolution2002562090209912449495\n4. KoskinenMTHaugenTOPrimmerCRContemporary Fisherian life-history evolution in small salmonid populationNature200241982683010.1038/nature0102912397355\n5. CopeEPrimary factors of organic evolution1886Chicago: University Press\n6. DamuthJCope's rule, the island rule and scaling of mammalian population densityNature199336574875010.1038/365748a08413651\n7. FosterJBThe evolution of mammals in islandsNature196420223423510.1038/202234a0\n8. LomolinoMVBody size of mammals on islands: the island rule re-examinedAmerican Naturalist198512531031610.1086/284343\n9. AdlerGHLevinsRThe island syndrome in rodent populationsThe Quarterly Review of Biology19946947349010.1086/4187447855232\n10. LomolinoMVBody size evolution in insular vertebrates: generality of the island ruleJournal of Biogeography200532101683169910.1111/j.1365-2699.2005.01314.x\n11. van ValenLA new evolutionary lawEvolutionary Theory19731130\n12. WhittakerRJFernández-PalaciosJMIsland Biogeography2007Oxford University Press\n13. BerryRJGrant PREvolution of small mammalsEvolution on islands1998Oxford: Oxford University Press3550\n14. MeiriSDayanTSimberloffDThe generality of the island rule reexaminedJournal of Biogeography20063391571157710.1111/j.1365-2699.2006.01523.x\n15. MeiriSCooperNPurvisAThe island rule: made to be broken?Proceedings: Biological Sciences2008275163114114810.1098/rspb.2007.1056\n16. LomolinoMVSaxDFRiddleBRBrownJHThe island rule and a research agenda for studying ecogeographical patternsJournal of Biogeography20063391503151010.1111/j.1365-2699.2006.01593.x\n17. HeaneyLRIsland Area and Body Size of Insular Mammals: Evidence from the Tri-Colored Squirrel (Callosciurus prevosti) of Southeast AsiaEvolution1978321294410.2307/2407408\n18. DayanTSimberloffDSize patterns among competitors: ecological character displacement and character release in mammals, with special reference to island populationsMammal Review19982839912410.1046/j.1365-2907.1998.00029.x\n19. MappesTKoskelaEGenetic basis of the trade-off between offspring number and quality in the bank voleEvolution20045864565015119447\n20. MappesTKoivulaMKoskelaEOksanenTASinervoBFrequency- and density dependent selection on reproductive effort strategiesPLoS ONE20083e168710.1371/journal.pone.000168718301764\n21. MeriläJCrnokrakPComparison of genetic differentiation at marker loci and quantitative traitsJournal of Evolutionary Biology20011489290310.1046/j.1420-9101.2001.00348.x\n22. EbenhardTA colonization strategy in the field vole (Microtus agrestis): reproductive traits and body sizeEcology1990711833184810.2307/1937592\n23. WigginsDAMøllerAPSørensenMFLBrandLAIsland biogeography and the reproductive ecology of great tits Parus majorOecologia199811547848210.1007/s004420050544\n24. OksanenTAKoivulaMKoskelaEMappesTThe cost of reproduction induced by body size at birth and breeding densityEvolution2007612822283110.1111/j.1558-5646.2007.00245.x17924957\n25. KoivulaMKoskelaEMappesTOksanenTACosts of reproduction in the wild: manipulation of reproductive effort in the bank voleEcology20038439840510.1890/0012-9658(2003)084[0398:CORITW]2.0.CO;2\n26. MillienVDamuthJClimate change and size evolution in an island rodent species: New perspectives on the island ruleEvolution20045861353136015266983\n27. LawlorTEComparative biogeography of mammals on islandsBiological Journal of the Linnean Society1986281–29912510.1111/j.1095-8312.1986.tb01751.x\n28. RaiaPMeiriSThe island rule in large mammals: Paleontology meets ecologyEvolution20066081731174217017072\n29. MeiriSDayanTSimberloffDArea, isolation and body size evolution in insular carnivoresEcology Letters20058111211121710.1111/j.1461-0248.2005.00825.x21352445\n30. MeltonRHBody size and island Peromyscus: a pattern and a hypothesisEvol Theor19826113126\n31. EbenhardTHigh activity in Bank Voles in relation to colonization abilityOikos19874929730210.2307/3565764\n32. MeriläJKruukLEBSheldonBCCryptic evolution in a wild bird populationNature2001412767910.1038/3508358011452309\n33. MousseauTARoffDANatural selection and the heritability of fitness componentsHeredity19875918119810.1038/hdy.1987.1133316130\n34. RoffDRLife history evolution2002U.S.A.: Sinauer\n35. RoseMRLife history evolution with antagonistic pleiotropy and overlapping generationsTheor Popul Biol19852834235810.1016/0040-5809(85)90034-6\n36. CoulsonTKruukLEBTavecchiaGPembertonJMClutton-BrockTHEstimating selection on neonatal traits in red deer using elasticity path analysisEvolution2003572879289214761065\n37. ChittyDDo lemmings commit suicide? Beautiful hypotheses and ugly facts1996Oxford Univ Press\n38. FisherRAThe Genetical Theory of Natural Selection1930Oxford: Clarendon Press\n39. GustafssonLLifetime reproductive success and heritability: empirical support for Fisher's fundamental theoremAm Nat198612876176410.1086/284601\n40. StensethNCGeographic distribution of Clethrionomys speciesAnnales Zoologici Fennici198522215219\n41. MappesTKoskelaEYlönenHReproductive costs and litter size in the bank voleProceedings of Royal Society of London, Biological Sciences1995261192410.1098/rspb.1995.0111\n42. Bondrup-NielsenSKarlssonFMovements and spatial patterns in populations of Clethrionomys species: A reviewAnnales Zoologici Fennici198522385392\n43. JonssonPHartikainenTKoskelaEMappesTDeterminants of reproductive success in voles: space use in relation to food and litter size manipulationEvolutionary Ecology20021645546710.1023/A:1020854525220\n44. KoskelaEMappesTYlönenHTerritorial behaviour and reproductive success of bank vole Clethrionomys glareolus femalesJournal of Animal Ecology19976634134910.2307/5980\n45. BujalskaGReproduction stabilizing elements in island populations of Clethrionomys glareolusActa Theriologica197015381412\n46. MyllymäkiAPaasikallioAPankakoskiEKanervoVRemoval experiments on small quadrats as a means of rapid assessment of the abundance of small mammalsAnnales Zoologici Fennici19718177185\n47. KoskelaEJonssonPHartikainenTMappesTLimitation of reproductive success by food availability and litter size in the bank vole Clethrionomys glareolusProceedings of the Royal Society of London, Biological Sciences19982651129113410.1098/rspb.1998.0408\n48. MillarJSAdaptive features of mammalian reproductionEvolution19773137038610.2307/2407759\n49. KleiberMThe fire of life1961New York: John Wiley & Sons Publ\n50. GouldSJOn the Scaling of Tooth Size in MammalsAmer Zool1975152353362\n51. SymondsMREElgarMAPhylogeny affects estimation of metabolic scaling in mammalsEvolution200256112330233312487362\n52. WhiteCRSeymourRSAllometric scaling of mammalian metabolismJournal of Experimental Biology20052081611161910.1242/jeb.0150115855392\n53. GlazierDSEffects of metabolic level on the body size scaling of metabolic rate in birds and mammalsProc R Soc B200827516411405141010.1098/rspb.2008.0118\n54. OksanenTAAlataloRVHorneTJKoskelaEMappesJMappesTMaternal effort and male quality in the bank vole, Clethrionomys glareolusProceedings of Royal Society of London, Biological Sciences19992661495149910.1098/rspb.1999.0806\n55. GockelJHarrBSchöttererCArnoldWGerlachGTautzDIsolation and characterization of microsatellite loci from Apodemus flavicollis (rodentia, muridae) and Clethrionomys glareolus (rodentia, cricetidae)Molecular Ecology1997659759910.1046/j.1365-294X.1997.00222.x9200832\n56. WrightSThe genetical structure of populationsAnnals of Eugenics195115323354\n57. GoudetJFstat (vers. 1.2): a computer program to calculate F-statisticsJournal of Heredity199586485486\n58. GoudetJRaymondMDe MeeuesTRoussetFTesting differentiation in diploid populationsGenetics1996144193319408978076\n59. RobertsonAExperimental design in the evaluation of genetic parametersBiometrics19591521922610.2307/2527670"
+}

batch_13/PMC2584077.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584077",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584077\nAUTHORS: KB Sriram, HPA Jersmann\n\nABSTRACT:\nMinimally invasive investigations, such as pleural fluid cytological assessment and closed percutaneous pleural biopsy, are often performed first in the investigation of suspected malignant pleural effusions. Malignant pleural effusions can be diagnosed with pleural fluid cytology alone in most cases; however, closed pleural biopsy is performed to increase the diagnostic yield when pleural fluid cytology is negative. This additional yield is at the expense of increased complication rates. We report a 64-year old man with a negative pleural fluid cytology but suspected malignant pleural effusion who underwent a closed pleural biopsy, which was complicated by pneumothorax, pneumomediastinum and severe subcutaneous emphysema. Pulmonary laceration by the pleural biopsy needle is the most likely aetiology of these complications. Our case report highlights an infrequent but significant complication of closed percutaneous pleural biopsy.\n\nBODY:\nBackgroundMalignant pleural effusions typically indicate disseminated disease, with an expected median survival of 3–12 months [1,2]. Prognosis is particularly poor when the pleural effusion is due to primary lung cancer [1,2]. There are several techniques available to investigate suspected malignant pleural effusions, ranging from simple procedures, such as pleural fluid aspiration and closed percutaneous pleural biopsy (CPB), to more complicated procedures, such as thoracoscopy. Guidelines for the diagnosis of unilateral pleural effusions typically recommend using minimally invasive investigations to obtain a diagnosis when malignancy is suspected [1]. We report a patient with an exudative unilateral pleural effusion who underwent a CPB and suffered a severe complication of this procedure.Case reportA 64-year-old-man was referred to our institution for investigation of a large left pleural effusion. Over the preceding two months, he had developed progressive dyspnea, left pleuritic chest pain and 6 kg weight loss. He was a reformed smoker with a 40-pack-year smoking history. He was admitted to the hospital, and a thoracic CT scan (Figure 1) was performed which showed a large left pleural effusion, left lung collapse and a left lower lobe mass. Pleural fluid cytology did not contain any malignant cells. Subsequently, CPB using Abrams needle was performed by a senior medical officer. Three out of four specimens contained normal pleural tissue without any malignant features.Figure 1Coronal view of thoracic computed tomography scan showing large left pleural effusion and mass in the left lower lobe.Twelve hours after the procedure, he developed a high pitched dysphonia, severe dyspnea, and pleuritic left chest pain. On examination, there was extensive subcutaneous emphysema along the chest wall extending to the neck. An emergency thoracic CT scan (Figure 2) was performed, which revealed subcutaneous emphysema, pneumothorax and pneumomediastinum. An intercostal chest tube was inserted, resulting in complete resolution of the pneumothorax. The subcutaneous emphysema, dysphonia and pneumomediastinum were resolved in seven days. The chest tube was successfully removed after seven days.Figure 2Coronal view of thoracic computed tomography scan showing pneumothorax, pneumomediastinum and subcutaneous emphysema which developed after iatrogenic pulmonary injury caused by Abrams pleural biopsy needle.Subsequently, a CT guided FNA of the left lower lobe mass was obtained. This revealed non-small cell lung cancer, subtype adenocarcinoma. The patient is currently receiving palliative chemotherapy with symptomatic improvement and no recurrence of the pleural effusion.DiscussionMalignant pleural effusions can be diagnosed by pleural fluid cytology alone in 60% of cases, and CPB can increase the diagnostic yield by 7%–27% [1]. Pleural biopsies are performed using either an Abrams' or Cope needle and can be successfully performed with limited training [[1,3], and [4]]. Pneumothorax develops in 3–15% of patients undergoing CPB, and of these patients, 1% requires chest tube drainage [1,3,4]. Pneumomediastinum and subcutaneous emphysema after CPB has not been commonly reported.Pneumothorax most commonly develops due to entry of air through the biopsy needle or laceration to the lung parenchyma. We hypothesize that in our patient, laceration to the lung parenchyma rather than air entry through biopsy needle is the most likely pathogenetic mechanism. The injury probably produced a bronchopleural fistula, allowing air to track along the continuum between the endothoracic fascia of outer chest wall, pleural space and mediastinum, resulting in subcutaneous emphysema, pneumothorax and pneumomediastinum [5]. We believe that air in the subcutaneous tissues of the neck caused the patient's dysphonia, because as the subcutaneous emphysema resolved so did the dysphonia.Because of the limited yield and risk of complications, CPB has been superseded by thoracoscopy in some centers. Thoracoscopy is diagnostic in 95% of malignant pleural effusions [6,7]. The high yield is due to the direct visualisation of the pleura and the ability to obtain large biopsy specimens from abnormal pleura. Additionally, thoracoscopy can be used to perform talc pleurodesis, producing effective symptom palliation [6,7]. However, the procedure is expensive, requires highly trained operators and is not widely available.Our report illustrates the complications of pneumothorax, pneumomediastinum and subcutaneous emphysema after a non-diagnostic CPB. We believe that by being aware of the limitations of CPB, clinicians can better choose the appropriate test in the investigation of exudative unilateral pleural effusions.ConsentWritten informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsKBS reviewed the case notes and prepared the manuscript. HPAJ read and approved the final manuscript.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2584372.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584372",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584372\nAUTHORS: Philippe J. Guerin, Lisbeth M. Næss, Carole Fogg, Einar Rosenqvist, Loretxu Pinoges, Francis Bajunirwe, Carolyn Nabasumba, Ray Borrow, Leif O. Frøholm, Salah Ghabri, Vincent Batwala, Rogers Twesigye, Ingeborg S. Aaberge, John-Arne Røttingen, Patrice Piola, Dominique A. Caugant\n\nABSTRACT:\nBackground\nNeisseria meningitidis serogroup A is the main causative pathogen of meningitis epidemics in sub-Saharan Africa. In recent years, serogroup W135 has also been the cause of epidemics. Mass vaccination campaigns with polysaccharide vaccines are key elements in controlling these epidemics. Facing global vaccine shortage, we explored the use of fractional doses of a licensed A/C/Y/W135 polysaccharide meningococcal vaccine.Methods and FindingsWe conducted a randomized, non-inferiority trial in 750 healthy volunteers 2–19 years old in Mbarara, Uganda, to compare the immune response of the full dose of the vaccine versus fractional doses (1/5 or 1/10). Safety and tolerability data were collected for all subjects during the 4 weeks following the injection. Pre- and post-vaccination sera were analyzed by measuring serum bactericidal activity (SBA) with baby rabbit complement. A responder was defined as a subject with a ≥4-fold increase in SBA against a target strain from each serogroup and SBA titer ≥128. For serogroup W135, 94% and 97% of the vaccinees in the 1/5- and 1/10-dose arms, respectively, were responders, versus 94% in the full-dose arm; for serogroup A, 92% and 88% were responders, respectively, versus 95%. Non-inferiority was demonstrated between the full dose and both fractional doses in SBA seroresponse against serogroups W135 and Y, in total population analysis. Non-inferiority was shown between the full and 1/5 doses for serogroup A in the population non-immune prior to vaccination. Non-inferiority was not shown for any of the fractionate doses for serogroup C. Safety and tolerability data were favourable, as observed in other studies.ConclusionsWhile the advent of conjugate A vaccine is anticipated to largely contribute to control serogroup A outbreaks in Africa, the scale-up of its production will not cover the entire “Meningitis Belt” target population for at least the next 3 to 5 years. In view of the current shortage of meningococcal vaccines for Africa, the use of 1/5 fractional doses should be considered as an alternative in mass vaccination campaigns.Trial RegistrationClinicalTrials.gov NCT00271479\n\nBODY:\nIntroductionSub-Saharan African countries in the “Meningitis Belt,” situated between Ethiopia and Senegal, face epidemics of meningococcal meningitis almost every year [1]. Following the current World Health Organization (WHO) recommendation, mass vaccination campaigns with polysaccharide meningococcal vaccine are implemented solely to control the spread of the epidemic [2]. Until recently, Neisseria meningitidis serogroup A has been the main organism causing those epidemics, while other serogroups play a minor epidemiological role.Following W135 outbreaks in Saudi Arabia in 2000 and 2001, cases of N. meningitidis serogroup W135 were reported in Burkina Faso in 2001, resulting in the first large W135 epidemic in that country in 2002 [3],[4]. This outbreak raised serious concerns regarding the availability of a vaccine protecting against that serogroup, i.e., a tetravalent A/C/Y/W135 polysaccharide vaccine (PSV). Mass vaccination of the population in Burkina Faso with the tetravalent PSV was not possible because of the global shortage in supply, in addition to its cost. In 2003, GlaxoSmithKline began producing a trivalent A/C/W135 polysaccharide vaccine for approximately USD1.50/dose, which was used in Burkina Faso in another epidemic the same year [5].Since then however, availability and affordability of the tetravalent or trivalent polysaccharide vaccines remain uncertain every year. The production of the bivalent A/C polysaccharide vaccine has been considerably reduced since 2005 and the quantity of vaccines to be produced in the next 3 to 5 years is uncertain [6],[7]. In case of simultaneous large outbreaks in different countries, the supply of meningococcal PSV for the coming meningitis seasons is unlikely to be sufficient to cover vaccination needs (Perea W., WHO, personal communication, March 2008). Conjugate meningococcal vaccines, are not expected to be available and affordable in large quantities to cover the need for Africa over the next several years [7]–[9].The current dose of the licensed tetravalent PSV developed in the 1970s contains 50 µg of each polysaccharide component. Studies in the 1970s and 1980s have shown that lower doses of polysaccharide were as effective as 50 µg in inducing bactericidal antibody levels that should be protective against disease in adults in the US [10]–[14].To test if fractionate doses might also be protective in an African population and in younger age groups, we conducted a clinical vaccine trial in Uganda to evaluate the potential use of fractional doses of meningococcal tetravalent PSV to control disease outbreak caused by N. meningitidis. The study population selected for the trial was 2–19 years of age, i.e., the population at highest risk of the disease and the primary target of mass vaccination campaigns in Africa during epidemics [15].MethodsStudy design and populationThe study design was a randomized, single-blind controlled trial. Three arms were defined in the trial: group 1 received a dose of 50 µg of each component of tetravalent PSV, i.e., a full dose of the licensed vaccine; group 2 received a 1/5 volume of tetravalent PSV (10 µg of each component); and group 3 received a 1/10 volume of tetravalent PSV (5 µg of each component).The study was conducted in the rural area of Kinoni, Rwampara County, Mbarara District, Uganda. This location was chosen on the basis of the following criteria: i, this area had not experienced recent epidemics of meningococcal meningitis; ii, the study population was considered to be stable; iii, the health subdistrict was considered a suitable site for this interventional study because it has a long-standing collaboration with Mbarara University, Department of Community Health.The recruitment of participants for the clinical trial was done on a voluntary basis. Volunteers aged 2–19 years old were recruited in proportions matching the Ugandan age distribution of the 2–19 years old extracted from the “2002/03 Uganda National Household Survey.” Volunteers were residents of Mbarara district, living within a 15-km radius of the vaccination site, with no plans of moving from the area during the study period. Community awareness meetings were held with local leaders and field workers from the study team, who then went house to house to get a list of people who were willing to participate. Participants came to the study site on a planned date. Refusal rates were not recorded in order to avoid unnecessary pressure on the communities.Objectives and outcomesThis study aimed to demonstrate non-inferiority in the immune response of doses corresponding to 1/5 and/or 1/10 of the amount of the full dose of a licensed A/C/Y/W135 polysaccharide vaccine (Menomune, Sanofi Aventis) and to evaluate the tolerability of these vaccinations. The primary endpoint was the proportion of responders defined by immunogenicity criteria at four weeks after vaccination based on SBA titers. The secondary endpoint considered the IgG response (Elisa).Sample size calculationThe sample size was calculated by choosing a one-sided 0.05 level of significance and power of 80%. Expecting equal proportions of responders in all groups given the vaccine being 80%, and assuming a non-inferiority margin of 10%, this gave a required sample size n of 198 persons in each group. Because the reference group (full dose) was used for two comparisons, a correction of () was applied [16], bringing that group to 280. The calculations have been performed using nQuery Advisor.Randomization and allocationsFollowing consent and a clinical examination, each subject was randomly allocated to one of the 3 dosage groups. The allocation schedule was computer-generated, using a block randomization method, stratified by age group (2 to 4; 5 to 9; 10 to 14 and 15 to 19 years). The researchers responsible for seeing the volunteers allocated the next available number on entry into the trial. The vaccination was given subcutaneously using low-volume syringes (0.5mL BD Micro-Fine insulin syringes), by the same nurse throughout the study, without participant knowledge of the dosage received. A single dose vaccine Menomune vial was used per volunteer, numbered with the study number and stored after vaccination. A full dose injection corresponded to 0.5ml of the vaccine, 1/5 of the dose corresponded to 0.1ml and 1/10 of the dose to 0.05ml.SafetyVolunteers were observed for 1 hour following vaccination for adverse events. Safety and tolerability data were collected for all volunteers during the 4 weeks following the injection. Safety data were collected during weekly interviews. The intensity of the adverse events was evaluated by clinicians, members of the study team and classified as “mild,” “moderate,” or “severe” using the Common Toxicity Criteria (CTC) grading (http://ctep.cancer.gov/reporting/CTC-3.html, US National Cancer Institute).Laboratory analysisSerum samples (10 mL of whole blood) were collected from each volunteer immediately before vaccination and 4 weeks later, stored at −80°C from the trial to the laboratories. Assays were carried out blinded at the Norwegian Institute of Public Health (NIPH). Immune responses to the different doses of the TPSV were analyzed in serum bactericidal assays (SBA) and enzyme-linked immunosorbent assays (ELISA). SBA was performed against four target strains of the A, C, W135, and Y serogroups: A: F8238 (4/21:P1.20,9); C: C11 (16:P1.7-1,1); W135: M01240070 (NT:P1.18-1,3); and Y: M00242975 (2a:P1.5,2). Heat-inactivated test sera were diluted 2-fold in microtiter plates (starting at serum dilution of 1:4) and incubated for 60 min with bacteria and baby rabbit complement (Pel-Freeze) before plating onto agar plates [17]. Colony-forming units were counted (Sorcerer colony counter, Perceptive Instruments), and bactericidal antibody titers were expressed as the reciprocal of the final serum dilution giving ≥50% killing compared with controls (inactive complement/no test serum). External quality control of SBA measurements was performed by Manchester Health Protection Agency (HPA) by analyzing in parallel approximately 10% of samples taken before vaccination and four weeks later. IgG antibodies to each separate polysaccharide A, C, Y, and W135 were measured in ELISA as described by Carlone et al.\n[18] and modified according to Joseph et al. using the CDC 1992 standard (NIBSC code 99/706) [19].Carriage studyTonsillo-pharyngeal samples were collected from the volunteers before vaccination and four weeks later. The technique and results of this carriage study are published elsewhere [20]. Volunteers found to be carriers of N. meningitidis of a homologous serogroup at any time between the vaccination and four weeks later were excluded from the analysis of response to that polysaccharide.Statistical analysisFor computational purposes, titers <4 were assigned a value of 2. A subject with SBA titer ≥128 was defined as putatively protected [21]. The Modified Intention To Treat (MITT) population included all randomized and exposed subjects with a defined SBA titer before vaccination and four weeks later. The Per Protocol (PP) population excluded subjects from the MITT presenting protocol violation. Some immunogenicity measures were not planned and described in the statistical analysis of the protocol. For the benefit of the study, the scientific committee coordinating the trial suggested additional statistical analyses: i, the principal criteria to define a responder was reinforced, as not only a 4-fold or greater increase in antibody titer between pre- and post-immunization sera, but also an SBA titer ≥128 four weeks after vaccination; ii, we also considered an exploratory population of the MITT, namely the “non-immune population” before vaccination, defined as individuals with SBA titers <128 before vaccination, which is considered the threshold of non-immunity [21]–[25]. Baseline characteristics were summarized by treatment groups using descriptive statistics (Geometric Mean Titer [GMT] and Geometric Mean Concentration [GMC)] were used for the analysis of the SBA titers and IgG concentrations). McNemar's test was used to compare matched pair titer data before vaccination and four weeks later.The proportion of adequate responses in each group was expressed as a percentage (“response to vaccine rate”). A 95% confidence interval was calculated for the observed difference in response proportion (full versus fractional dose), and if the upper limit was <10%, the fractional dose was considered non-inferior to the full dose. These analyses were performed on MITT, PP, and non-immune subsets of the MITT.We performed a logistic regression to look at the impact of age among responders by serogroup and by arm. Age was considered in two groups of interest (≤5 and >5 years of age) knowing that in previous studies, eliciting an immune response under 5 was the most critical [11],[26].Data were double-entered using Epidata 3.0 (The EpiData Association, Odense, Denmark). Statistical analyses were performed using STATA 9 (College Station, Texas, USA).EthicsWritten informed consent in the local language was obtained from the parents or guardians of every volunteer <18 years of age or by the volunteers themselves if >18 years. The study was approved by the Faculty Research and Ethics Committee of the Mbarara University of Science and Technology (MUST), the MUST Institutional Review Board, the Uganda National Committee of Science and Technology, and the Regional Committee for Medical Research Ethics in Norway. The trial was registered at Clinicaltrials.gov (NCT00271479).ResultsStudy groupsBetween 5 July 2004 and 22 September 2004, 763 volunteers from the Kinoni community in Mbarara, Uganda were screened (Figure 1). Among them, 750 volunteers were included, with 291 randomized to the full-dose vaccine arm, 225 to the 1/5-dose arm, and 234 to the 1/10-dose arm.10.1371/journal.pntd.0000342.g001Figure 1Consort flowchart.The demographic and serological baseline characteristics of the population at inclusion before vaccination are displayed in Table 1. For each serogroup, volunteers were not considered in the analyses if an SBA value was missing for either before vaccination or four weeks later. No differences were observed between arms for demographic and serological data.10.1371/journal.pntd.0000342.t001Table 1Demographic and serological baseline characteristicsFull dose (n = 291)1/5 dose (n = 224)*\n1/10 dose (n = 235)*\n\nAge (years)\nMedian (IQR)9.1 (5.1–13.1)9.1 (5.1–13.1)9.1 (5.1–14.1)\nSex\nRatio (M/F)0.89 (137/154)1.06 (115/109)0.90 (111/124)\nWeight (kg)\nMedian (IQR)25 (16–37)25 (17–37)25 (17–39)\nHeight (cm)\nMedian (IQR)128.0 (107–146.5)127.1 (110–146.5)129.8 (108.5–149.4)\nSBA A titers\nGMT (GSD)41.6 (21.6)58.7 (20.9)48.8 (21.0)\nGMC IgG µg/mL ELISA A\nGMC (GSD)2.6 (2.4)2.8 (2.3)2.4 (2.4)\nSBA W135 titers\nGMT (GSD)7.1 (10.0)6.0 (8.6)7.1 (10.1)\nGMC IgG µg/mL ELISA W135\nGMC (GSD)2.7 (2.5)2.9 (2.4)2.9 (2.3)\nSBA C titers\nGMT (GSD)2.7 (3.5)2.8 (3.9)3.5 (5.3)\nGMC IgG µg/mL ELISA C\nGMC (GSD)0.9 (2.4)1.0 (2.4)1.0 (2.3)\nSBA Y titers\nGMT (GSD)2.2 (1.9)2.7 (3.8)2.2 (2.0)\nGMC IgG µg/mL ELISA Y\nGMC (GSD)2.9 (2.4)3.1 (2.4)3.1 (2.3)*: One patient randomized in the 1/5-dose group received 1/10 of the dose.GSD: Geometric Standard Deviation.Natural immunity toward N. meningitidis serogroups A, C, Y, and W135 before vaccination in the study population was measured by the proportion of volunteers with SBA titers ≥128 before vaccination: 51.4% (382/743) for serogroup A; 22.6% (168/744) for serogroup W135; 6.2% (45/729) for serogroup C, and 2.3% (17/741) for serogroup Y.Immunologic responseProtocol deviations leading to exclusion of population are described in Table 2. The primary end point, i.e. proportions of responders per arm and per analyses are reported in Table 3. For serogroup W135, 94.4% (168/178) of the non-immune, vaccinated subjects in the 1/5-dose arm, and 97.2% (172/177) in the 1/10-dose arm, were responders, compared with 93.7% (207/221) in the full-dose arm. For serogroup A, 92.2% (94/102) and 88.3% (98/111) of non-immune vaccinees in the 1/5- and 1/10-dose arms, respectively, were responders, compared with 94.6% (140/148) in the full-dose arm.10.1371/journal.pntd.0000342.t002Table 2Description of individual exclusions by population (MITT, PP, Safety population) and randomized groupReason for exclusionNumber of volunteersRandomized groupExcluded forReceived 1/10 instead of 1/511/5 dosePP all serogroupsWithdrew consent before vaccination11/10 doseMITT, PP, Safety Analysis all serogroupsMis-stratified*\n32 in 1/5 dose and 1 in 1/10PP all serogroupsCarrier of W135**\n21in Full dose and 1 in 1/10 dosePP for W135Malnourished (weight/height <−2 Z-score)1Full dosePP all serogroups*: Allocated to the wrong age group.**: Volunteers presenting a carriage of W135 between the vaccination and four weeks later.10.1371/journal.pntd.0000342.t003Table 3Proportion of responders per serogroup and per population and SBA GMT at 4 weeks after vaccinationAnalysesDose GroupProportion of RespondersSBA Titers four weeks after vaccinationn/N%95% CIGMT95% CISerogroup AMITTFull249/28986.282.2–90.23607.12952.8–4406.31/5 dose173/22477.271.7–82.72035.41600.1–2589.01/10 dose159/23069.163.2 -75.01367.61083.1–1726.8PPFull247/28786.182.1–90.13612.62953.3–4419.11/5 dose172/22277.572.0–83.02054.41612.2–2618.01/10 dose159/22969.463.5–75.31369.31083.3–1730.7Non immuneFull140/14894.691.0–98.21918.01426.0–2579.81/5 dose94/10292.287.0–97.4852.3573.2–1267.51/10 dose98/11188.382.4–94.2754.1495.3–1148.1Serogroup W135MITTFull269/28993.190.2–96.02190.31728.9–2774.61/5 dose212/22494.691.7–97.52029.11573.7–2616.21/10 dose220/23195.292.5–97.92422.71979.9–2964.6PPFull267/28693.490.5–96.32175.91714.2–2762.01/5 dose210/22294.691.6–97.62041.61582.2–2634.51/10 dose219/22995.693.0–98.22426.31979.7–2973.7Non immuneFull207/22193.790.5–96.91539.51160.0–2043.21/5 dose168/17894.491.0–97.81517.41129.3–2039.01/10 dose172/17797.294.8–99.62008.31583.0–2547.9Serogroup CMITTFull259/28491.287.9–94.51168.3911.0–1498.21/5 dose179/22280.675.4–85.8472.1332.3–670.61/10 dose171/22376.771.2–82.2399.3277.1–575.4PPFull257/28291.187.8–94.41175.1914.9–1509.31/5 dose177/22080.475.2–85.6467.3328.0–665.81/10 dose170/22276.671.1–82.1396.3274.6–572.0Non immuneFull252/27193.090.0–96.01108.5858.4–1431.51/5 dose172/21181.576.3–86.7412.2288.2–589.61/10 dose156/20277.271.4–83.0315.6214.8–463.8Serogroup YMITTFull242/28684.680.4–88.8936.2673.7–1301.01/5 dose185/22482.677.7–87.5772.7529.2–1128.21/10 dose194/23184.079.3–88.7822.6569.8–1187.5PPFull240/28484.580.3–88.7924.2663.8–1286.71/5 dose183/22282.477.4–87.4768.3524.4–1125.61/10 dose193/23083.979.2–88.6816.8565.1–1180.8Non immuneFull238/28284.480.2–88.6916.8657.3–1278.81/5 dose175/21481.876.7–86.9687.5466.1–1014.11/10 dose191/22883.879.0–88.6798.1551.1–1155.7Non-inferiority was demonstrated for serogroups W135 and Y (full dose versus each fractional dose in MITT analyses), but was statistically rejected for serogroups A and C (Table 4). When analyzing only the non-immune population, non-inferiority was also demonstrated for full versus 1/5 doses for serogroups A (2.4% [95% confidence interval, −3.9 to 8.8%]), W135 (−0.7% [95% confidence interval, −5.4 to 3.9%]), and Y (2.6% [95% confidence interval, −4.1 to 9.3%]), but not for serogroup C (11.5% [95% confidence interval, 5.4 to 17.5%]) (Table 4). When considering the response by age group (logistic regression), children under 5 had a lower chance of positive response compared to older ones for serogroup W135 (significant only for full dose arm), serogroup C (significant for full dose and 1/5 dose arms) and for serogroup Y (significant for 1/10 dose arm) (Table 5). For serogroup A, although not significant, fractional doses seem to elicit a better response in children under 5.10.1371/journal.pntd.0000342.t004Table 4Non-inferiority analysis results of SBA responders per serogroup and analysesTotal populationNon-immune population\nSerogroup A\n\nMITT (n = 743)(n = 361)Diff.95%CIDiff.95%CI\nFull dose vs 1/5\n+8.9%[+2.1%, +15.7%]+2.4%[−3.9%, +8.8%]\nFull dose vs 1/10\n+17.0%[+9.8%, +24.2%]+6.3%[−0.7%, +13.3%]\nSerogroup W135\n\nMITT (n = 744)(n = 576)Diff.95%CIDiff.95%CI\nFull dose vs 1/5\n−1.6%[−5.7%, +2.6%]−0.7%[−5.4%, +3.9%]\nFull dose vs 1/10\n−2.2%[−6.2%, +1.9%]−3.5%[−7.5%, +0.5%]\nSerogroup C\n\nMITT (n = 729)(n = 684)Diff.95%CIDiff.95%CI\nFull dose vs 1/5\n+10.6%[+4.4%;+16.7%]+11.5%[+5.4%;+17.5%]\nFull dose vs 1/10\n+14.5%[+8.1%;+21.0%]+15.8%[+9.2% ; +22.3%]\nSerogroup Y\n\nMITT (n = 741)(n = 724)Diff.95%CIDiff.95%CI\nFull dose vs 1/5\n+2.0%[−4.5%;+8.5%]+2.6%[−4.1%;+9.3%]\nFull dose vs 1/10\n+0.6%[−5.7%;+6.9%]+0.6%[−5.8%;+7.0%]10.1371/journal.pntd.0000342.t005Table 5Logistic regression results of age effect on responder per serogroup and per arm–MITT populationSerogroupsFull dose1/5 dose1/10 doseAge>5 vs. ≤5 years oldOR(95% CI)OR(95% CI)OR(95% CI)\nSerogroup A\n2.07(0.99;4.31)0.62(0.27; 1.42)0.79(0.39; 1.61)\nSerogroup W135\n2.78(1.08; 7.15)2.81(0.85; 9.28)0.37(0.05; 2.95)\nSerogroup C\n3.37(1.44; 7.87)2.83(1.36; 5.87)1.66(0.81; 3.44)\nSerogroup Y\n1.99(0.98; 4.05)1.84(0.85; 3.98)2.46(1.14; 5.30)The secondary immunogenicity criterion based on ELISA data is reported on Figure 2. For each serogroup and each dose of vaccine, the geometric means of IgG concentrations showed no difference between arms before vaccination but a significant difference four weeks later with full dose greater than both 1/5 and 1/10 doses. Statistically significant differences were observed between the vaccination and four weeks later for each dose and each serogroup (p<0.0001 for all comparisons).10.1371/journal.pntd.0000342.g002Figure 2ELISA IgG concentrations per serogroup and per arm before the vaccination and four weeks later in the MITT population (GMC with superior limit of 95%CI).GMC = geometric mean concentration. P<0.001 for all comparisons between the vaccination and four weeks later for all serogroups.Adverse eventsA total of 158 volunteers reported at least one adverse event during the 4 weeks after vaccination (171 total adverse events, Table 6). No significant statistical difference was observed among the three dose arms (χ2 test, p = 0.42). The most commonly reported adverse events were upper respiratory tract infections (URTI) (57%) and malaria (20%). Five severe adverse events were recorded: one severe case of malaria, one severe episode of seizures, and 3 severe URTI, but these events were not considered to be related to the vaccination. Three adverse events considered “probably related” were reported and classified as mild (2 subjects with fever and 1 with headache).10.1371/journal.pntd.0000342.t006Table 6Distribution of adverse events following vaccinationDistribution of adverse events per week post-vaccinationAttendance at schedule visits n/N (%)Full dose1/5 dose1/10 doseTotalFirst week745/749 (99.5%)25161859 (34.5%)Second week746/749 (99.6%)22111851 (29.8%)Third week746/749 (99.6%)125623 (13.5%)Fourth week749/749 (100%)17111038 (22.2%)\nTotal\n76 (44.4%)43 (25.1%)52 (30.4%)171 (100%)Quality controlExternal quality control of the SBA titer measurements showed no significant difference with regard to responders for serogroup A (McNemar pair matched test, p =  0.63), serogroup C (p = 0.06), and serogroup Y (p = 0.41). For serogroup W135, the difference was statistically significant (p<0.001).DiscussionSBA is the accepted correlate of protection for meningococcal disease. In the MITT analysis of this study, non-inferiority was demonstrated between full and 1/5 and 1/10 fractional doses of TPSV in SBA response against the meningococcal serogroups W135 and Y. Non-inferiority was only shown between the full and 1/5 doses for serogroup A in the pre-vaccination, non-immune population. Non-inferiority was rejected for serogroup C in all analyses. Safety and tolerability data were favourable, as observed with TPSV in other studies [27],[28].In analyzing the proportion of responders per serogroup, we observed a decline in response for serogroup A and C from the full versus 1/5 dose, and this decrease was accentuated versus the 1/10 dose. For serogroup A, which is the most important serogroup to protect against in sub-Saharan Africa, the response in the MITT analysis decreased from 86% to 77%. Several elements must be considered in the interpretation of these results. A notable proportion of volunteers (51.4%) had high SBA titers against serogroup A prior to vaccination, presumably resulting from natural immunity. In demonstrating non-inferiority between the full and 1/5 dose groups in the non-immune population, the difference in responses occurred mainly in the naturally immune subgroup. These results suggest that the full dose may elicit higher increase in SBA titers for subjects with pre-vaccination SBA titers ≥128 compared with 1/5 of the dose. However, assuming that a post-vaccination SBA titer ≥128 is a proxy for vaccine efficacy, we believe that 1/5 of the dose induced an acceptable increase of SBA for non-immune populations, although it did not strictly meet the criteria we designed for the total population. When considering the response for children under five, overall fractional doses do not affect the chance of response compared to full dose. For serogroup A, the response could be possibly better in children under five with fractional doses, though the study was not powered to demonstrate this hypothesis.For all serogroups, the IgG concentrations decreased with fractional doses. However, the SBA titer/IgG ratios showed similar results between arms for all serogroups (data not shown), indicating a higher proportion of bactericidal antibodies in fractional doses. This could be due to differences in antibody avidity, though this hypothesis would require further studies. In an epidemic response setting, the goal of a mass vaccination campaign is short term immunity-basically protection through to the end of the epidemic season. Therefore, longer duration of protection (presumably predicted by higher titers) is a less important issue.Licensed meningococcal polysaccharide vaccines are known to confer an immunity of short duration (2–3 years) and are therefore not recommended in expanded vaccination programs [6],[29]. But this characteristic may not impact the use of fractional dosing in a reactive mass vaccination campaign aimed at preventing further new cases during an ongoing epidemic. Study subjects in this trial were followed up to 2 years, and the duration of protection will be addressed later on.Several potential limitations of this study must be addressed. Tolerability data were excellent; however, the weekly visits between the vaccination and four weeks later may not have been optimal to capture adverse events often occurring in the first days after vaccination. HIV testing was not systematically performed. Considering the epidemiological indicators of HIV in the adult population aged 15–49 years (HIV prevalence rate 6.7% [5.7–7.6]) [30], and the exclusion criteria of known or suspected cases in our study population, the impact of HIV is unlikely to be noticeable. Injections of fractional doses with “insulin syringes” were considered relatively simple to perform in the field for the 1/5 (0.1 mL) dose, but the 1/10 (0.05 mL) dose was more difficult to inject. Such difficulty may have hampered the delivery of the 1/10 fractional dose. This evaluation was based on the informal evaluation from the study team. Considering the absence of difficulties to inject 1/5 of the dose providing the use of appropriate syringes and training, health workers engaged in an outbreak response during an epidemic should not faced major problems to implement this vaccination. The unexpected high background rate of immunity to serogroup A in the study population has been a constraint to demonstrate the impact of the vaccination for this serogroup. Despite the fact that no large outbreak of meningococcal meningitis due to serogroup A had been declared in southern Uganda in the years prior to the study, it is likely that the strain was circulating in the region, following the outbreaks of serogroup A in neighbouring countries, Burundi and Rwanda in 2002 [31].Quality control of the SBA titers showed satisfactory results for serogroups A, C, and Y. However, a discrepancy was found for the W135 serogroup. This discrepancy was found to be due to the use of a different strain between the two laboratories. Once repeated with same strain, there was no significant difference between the results of the two laboratories (p = 0.31). As the proportion of responders for serogroup W135 was the same in the two laboratories and the source of the discrepancy was identified, we believe that our overall results of serogroup W135 are validated.Baby rabbit complement was used in the SBA assays in accordance with international standard protocols to evaluate polysaccharide vaccines against meningococcal disease, but SBA with human complement might be more relevant to elucidate the immune response after disease and vaccination. Additional insight would be gained by assaying these sera in a human complement SBA assay, and such analyses are ongoing.The two prevailing serogroups that cause N. meningitidis epidemics in the African Meningitis Belt are A and W135, and serogroups C and Y are not presently reported as the causal agent of meningitis epidemics in the region [6]. The WHO states that problems regarding the availability and affordability of protective meningococcal vaccines over the coming years need to be addressed urgently [6]. A risk-benefit analysis of the use of fractional doses should guide decision-makers. Similar strategies with other vaccines have already proved successful [32]. Assuming 90%, short-term protection by the licensed meningococcal polysaccharide vaccines, and a conservative protection of 80% using a reduced 1/5 dose, the same amount of resources invested in vaccine purchase would protect 4.4 times more subjects. Although the cost of immunization is not a primary interest of this strategy in the context of a global shortage, the use of a fractional dose would decrease the cost per person vaccinated by approximately half (data not shown). While the advent of conjugate A vaccine will largely contribute to control serogroup A outbreaks in Africa, the scale-up of its production will not cover the entire “meningitis belt” target population over the next 3 to 5 years (Laforce M., Meningitis Vaccine Project, personal communication January 2008). Considering the current shortage of meningococcal vaccines for Africa and the prevalence of serogroups A and W135, the use of 1/5 fractional doses should be explored as an alternative strategy in mass vaccination campaigns.Supporting InformationAlternative Language Abstract S1Translation of the Abstract into French by Philippe J. Guerin(0.03 MB DOC)Click here for additional data file.Protocol S1Final approved protocol(1.03 MB PDF)Click here for additional data file.Checklist S1Consort Checklist(0.06 MB DOC)Click here for additional data file.\n\nREFERENCES:\n1. LapeyssonnieL\n1968\n[Comparative epidemiologic study of meningococcic cerebrospinal meningitis in temperate regions and in the meningitis belt in Africa. Attempt at synthesis].\nMed Trop\n28\n709\n720\n2. World Health Organization\n2000\nDetecting meningococcal meningitis epidemics in highly-endemic African countries.\nWkly Epidemiol Rec\n75\n306\n309\n11045076\n3. Ouedraogo-TraoreRHoibyEASanouISangareLKyelemN\n2002\nMolecular characteristics of Neisseria meningitidis strains isolated in Burkina Faso in 2001.\nScand J Infect Dis\n34\n804\n807\n12578146\n4. World Health Organization\n2002\nMeningococcal disease, serogroup W135, Burkina Faso.\nWkly Epidemiol Rec\n77\n152\n155\n12037945\n5. ZombreSHacenMMOuangoGSanouSAdamouY\n2007\nThe outbreak of meningitis due to Neisseria meningitidis W135 in 2003 in Burkina Faso and the national response: main lessons learnt.\nVaccine\n25\nSuppl 1\nA69\nA71\n17630052\n6. World Health Organization\n2007\nRisk of epidemic meningitis in Africa: a cause for concern.\nWkly Epidemiol Rec\n82\n79\n87\n17348094\n7. GreenwoodB\n2006\nEditorial: 100 years of epidemic meningitis in West Africa-has anything changed?\nTrop Med Int Health\n11\n773\n780\n16771997\n8. JodarLFeaversIMSalisburyDGranoffDM\n2002\nDevelopment of vaccines against meningococcal disease.\nLancet\n359\n1499\n1508\n11988262\n9. LaForceFMKondeKVivianiSPreziosiMP\n2007\nThe Meningitis Vaccine Project.\nVaccine\n25\nSuppl 1\nA97\n100\n17521780\n10. ArtensteinMSGoldRZimmerlyJGWyleFABrancheWCJr\n1970\nCutaneous reactions and antibody response to meningococcal group C polysaccharide vaccines in man.\nJ Infect Dis\n121\n372\n377\n4985950\n11. GoldRLepowMLGoldschneiderIDraperTLGotschlichEC\n1975\nClinical evaluation of group A and group C meningococcal polysaccharide vaccines in infants.\nJ Clin Invest\n56\n1536\n1547\n1202084\n12. GriffissJMBrandtBLBroudDD\n1982\nHuman immune response to various doses of group Y and W135 meningococcal polysaccharide vaccines.\nInfect Immun\n37\n205\n208\n6809627\n13. GriffissJMBrandtBLBroudDDAltieriPLBermanSL\n1985\nRelationship of dose to the reactogenicity and immunogenicity of meningococcal polysaccharide vaccines in adults.\nMil Med\n150\n529\n533\n3934585\n14. BrandtBLGriffissJMBroudDDBoslegoJW\nKluwer Academic Publishers, editors\n1988\nTetravalent meningococcal polysaccharide vaccine. Relashionship of dose to immunogenicity and reactogenicity in young adults.\nGonococci and Meningococci\nDordrecht, The Netherlands\n33\n37\n15. CampagneGSchuchatADjiboSOusseiniACisseL\n1999\nEpidemiology of bacterial meningitis in Niamey, Niger, 1981–96.\nBull World Health Organ\n77\n499\n508\n10427935\n16. LellouchJLazarP\n1974\nMéthodes statistiques en expérimentation biologique. Paris, Flammarion.\n17. BorrowRCarloneGM\n2001\nSerogroup B and C serum bactericidal assays.\nPollardAJ MMe\nMeningococcal vaccines: methods and protocols. Methods in molecular medicine\nTotowa (New Jersey)\nHumana Press\n289\n304\n18. CarloneGMFraschCESiberGRQuataertSGheeslingLL\n1992\nMulticenter comparison of levels of antibody to the Neisseria meningitidis group A capsular polysaccharide measured by using an enzyme-linked immunosorbent assay.\nJ Clin Microbiol\n30\n154\n159\n1734048\n19. JosephHBalmerPBybelMPapaTRyallR\n2004\nAssignment of Neisseria meningitidis serogroups A, C, W135, and Y anticapsular total immunoglobulin G (IgG), IgG1, and IgG2 concentrations to reference sera.\nClin Diagn Lab Immunol\n11\n1\n5\n14715537\n20. CaugantDAFoggCBajunirweFPiolaPTwesigyeR\n2006\nPharyngeal carriage of Neisseria meningitidis in 2-19-year-old individuals in Uganda.\nTrans R Soc Trop Med Hyg\n100\n1159\n1163\n16765397\n21. AndrewsNBorrowRMillerE\n2003\nValidation of serological correlate of protection for meningococcal C conjugate vaccine by using efficacy estimates from postlicensure surveillance in England.\nClin Diagn Lab Immunol\n10\n780\n786\n12965904\n22. GoldschneiderIGotschlichECArtensteinMS\n1969\nHuman immunity to the meningococcus. I. The role of humoral antibodies.\nJ Exp Med\n129\n1307\n1326\n4977280\n23. SantosGFDeckRRDonnellyJBlackwelderWGranoffDM\n2001\nImportance of complement source in measuring meningococcal bactericidal titers.\nClin Diagn Lab Immunol\n8\n616\n623\n11329468\n24. BorrowRBalmerPMillerE\n2005\nMeningococcal surrogates of protection–serum bactericidal antibody activity.\nVaccine\n23\n2222\n2227\n15755600\n25. JodarLStephensDFeaversIM\n2002\nAssay parameters and methods of data analysis for the comparison of complement sources in the Neisseria meningitidis serogroup C serum bactericidal assay.\nBiologicals\n30\n323\n329\n12474864\n26. KayhtyHKarankoVPeltolaHSarnaSMakelaPH\n1980\nSerum antibodies to capsular polysaccharide vaccine of group A Neissera meningitidis followed for three years in infants and children.\nJ Infect Dis\n142\n861\n868\n6780634\n27. GoldRArtensteinMS\n1971\nMeningococcal infections. 2. Field trial of group C meningococcal polysaccharide vaccine in 1969-70.\nBull World Health Organ\n45\n279\n282\n5316907\n28. YergeauAAlainLPlessRRobertY\n1996\nAdverse events temporally associated with meningococcal vaccines.\nCMAJ\n154\n503\n507\n8630839\n29. PatelMLeeCK\n2005\nPolysaccharide vaccines for preventing serogroup A meningococcal meningitis.\nCochrane Database Syst Rev\nCD001093\n30. Joint United Nations Programme on HIV/AIDS\n2008\nCountry situation analysis, Uganda.\nAvailable: http://www.unaids.org/en/CountryResponses/Countries/Uganda.asp. Accessed 11 January 2008\n31. Meningococcal disease, Great Lakes area (Burundi, Rwanda, United Republic of Tanzania)—update\n2002\nWkly Epidemiol Rec\n77\n317\n32. ObaroSKOtaMO\n2006\nSense and the science of childhood immunization: can we achieve more with less?\nVaccine\n24\n6460\n6467\n16911848"
+}

batch_13/PMC2584648.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584648",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584648\nAUTHORS: Andreas Hansch, Mieczyslaw Gajda, Joachim Boettcher, Alexander Pfeil, Werner A Kaiser\n\nABSTRACT:\nBackgroundLiposarcomas are classified into four subtypes, with different malignancy potential and characteristic imaging appearances. Well-differentiated liposarcomas have imaging characteristics similar to those of benign lipomas, however they can be usually distinguished from lipomas because of the larger size and broader fibrous septa, with a more nodular appearance.Case presentationThis paper presents a case of atypical lipoma (well-differentiated liposarcoma) of the pelvis, leading to partial involvement of the sciatic nerve. In our case, computed tomography (CT) showed a low-density lesion. In magnetic resonance imaging (MRI), T1 and T2-weighted sequences revealed a fatty appearance with signal loss on fat saturation pulse sequences.ConclusionThe lesion was successfully resected and no other similar lesions have been found within one year of follow-up.\n\nBODY:\nBackgroundLiposarcoma, a malignant tumor of mesenchymal origin, is one of the most common primary neoplasms in the retroperitoneum [1]. Histologically, liposarcomas are classified in four subtypes with different potential of malignancy: well-differentiated or atypical, myxoid, pleomorphic, and round-cell subtypes. The surgical strategy and patient prognosis largely depends on these different subtypes [2].Case presentationA 60-year old woman was admitted to the outpatient department of our hospital due to deep pain of the pelvis and distal lumbar spine. Progressive pain of the posterior right leg also developed. Neurological examination revealed a reduction of strength of the right leg and incomplete paresis of the right tibial and common peroneal nerve (both branches of the sciatic nerve). The patellar reflexes were evocable on both sides, the Achilles tendon reflex was positive on the left side but absent contralaterally. Additionally, a hypoaesthesia of the S1 dermatome was detected on the right leg. The clinical picture was therefore that of a partial impairment of the right sciatic nerve. Initially, a local irritation of the sacral spinal nerve 1 was suspected, however a magnetic resonance imaging (MRI) showed the presence of a massive tumor in the right pelvis. An extension of the MRI to the whole pelvis revealed a lipomatous appearance and the presence of septa (Figure 1A–C). The diameter of the lesion was 18 × 17 × 8 cm. The enormous tumor displaced adjacent anatomic structures, for example the right obturator and piriformis muscles were displaced to the right, the uterus and the rectum were displaced to the left. The sciatic foramen was completely occupied. The neoplasm spread out to the gluteal region, with displacement of the right gluteal muscles (minimus, medius, and maximus).Figure 1Pelvic MR and CT imaging. A) Massive fatty tumor with hyperintense appearance in the noncontrast T1-weighted (TR/TE, 168/5) axial image. B) The sagittal noncontrast T1-weighted (TR/TE, 600/10) image demonstrated a thin septum inside the tumor. C) The tumor showed a complete signal loss in the fat saturation sequences (TR/TE 5660/48). D) Axial noncontrast multidetector CT image demonstrated a massive soft-tissue pelvic tumor occupying the sciatic foramen. The mean density was -100HE and identified the fatty tissue characteristics. E) Magnification of the axial noncontrast multidetector CT image revealed nodular appearance of the septa (see arrows). F) Coronal reconstructed noncontrast multidetector CT image showed the expansive soft-tissue mass inside and outside the pelvic. Adjacent organs und muscles were displaced.The lesion presented high signal intensities on T1-weighted sequences (Figure 1A, B) and a complete signal loss on fat-saturated T1-weighted images (Figure 1C). The fat equivalent density was verified by computed tomography (CT), with measured densities of -100 HE (Figure 1D). There was no pathological enhancement of contrast media (Figure 1E) nor any signs of bone destruction (Figure 1D–F). The CT and MRI findings, and especially the signal loss on fat-saturated T1-weighted images, indicated a lipomatous tumor, e.g. a lipoma. In addition, the huge size of the lesion and the presence of broad fibrous septa (Figure 1E) were considered indicative of a well-differentiated liposarcoma. The lesion was successfully resected.After surgical removal of the tumor, the histopathology confirmed the diagnosis of a well-differentiated liposarcoma (Figure 2). Nearly all neurological symptoms disappeared postoperatively. Only the hypoaesthesia of the S1 dermatome of the right leg remained.Figure 2A) Massive tumor with fatty appearance after surgery. B) and C) Adipocytes and some atypical nuclei (see arrows, B: H&E 200× and C: H&E 400×) as a typical finding of a well-differentiated liposarcoma (atypical lipoma). D) Assessment of cell proliferation by detection of Ki67 antigen shows the low proliferation index of the atypical lipoma (see arrows, Ki67 staining 400×).DiscussionWe present a case of well-differentiated liposarcoma (atypical lipoma) in the pelvis with incomplete involvement of the sciatic nerve. About 16–18% of all malignant soft-tissue tumors in the general population are liposarcomas [3], which are of mesenchymal origin [4] and are classified into four different histological entities: well-differentiated or atypical, myxoid, pleomorphic, and round-cell subtypes [5].Best to our knowledge, this is the first description of a case with paresis of a nerve as clinical manifestation of a lipomatous lesion. Local compression in the sciatic foramen due to the hudge size of the tumor caused impairment of the nerve.The presented massive pelvic tumor demonstrated typical imaging findings of a lipomatous lesion. In CT imaging, the neoplasm showed a low density reaching -100 HE, except for the septa, which showed densities of soft tissue. MRI also enables a precise diagnosis of the tumor and the adjacent anatomic structures. Fat tissue demonstrates a short T1 and a relatively long T2 relaxation and therefore appears hyperintense on T1-weighted and intermediately isointense to hyperintense on T2-weighted fast-spin-echo and gradient-echo images. The presented case showed the characteristics of a benign lipoma on both CT and MR imaging [2,6]. Lipomas and well-differentiated liposarcomas consist both of lipocytes and therefore comprise a large amount of fat tissue, usually more than 75% of their volume [4]. Lipoma does not cause any symptoms. The size of the described tumor indicate an atypical lipoma. Additionally, in well-differentiated liposarcomas, however, fibrous septa are usually broader, with a more nodular appearance than seen in lipomas. If only marginally excised, atypical lipomas can recur.Interestingly, only well-differentiated liposarcomas are predominantly fatty. The other histological subtypes usually demonstrate less than 25% or no fat tissue [2]. The most common type of liposarcoma is the myxoid one, which is considered of intermediate-grade malignancy [7]. Its signal intensities in MRI and CT imaging differ from those of lipomas and well-differentiated liposarcomas because the myxoid tumor consists of less than 10% fat tissue [7]. In CT imaging, myxoid liposarcomas often show an inhomogeneous appearance, with a density less than that of muscle tissue. In MRI, myxoid tumors demonstrate very long, homogeneous T1 and T2 relaxation times similar to those of water, low signal intensities on native T1-weighted images and high signal intensities on T2-weighted sequences [2]. However, some areas of the neoplasm consist of fat tissue with a high signal intensity on T1-weighted images and an intermediate signal intensity on T2-weighted sequences, enabling the correct diagnosis [1]. Intravenous administration of contrast media characteristically results in marked enhancement of the myxoid liposarcoma, revealing the solid nature of this tumor [4].Pleomorphic and round cell liposarcomas are high-grade sarcomas, consisting of only small (or no) amounts of fat tissue. Both tumors show a heterogeneous appearance, also occurring in other malignant soft tissue tumors.Preoperative diagnosis of the subtype has an impact on the surgical planning [2]. If the lesion presents a low metastatic potential (i.e., is well differentiated) a marginal excision (as performed in this case) may be possible. For high-grade tumors (round-cell or pleomorphic liposarcomas) a marginal excision is not indicated because of their high metastatic potential.ConclusionIn conclusion, well-differentiated liposarcomas present imaging characteristics similar to benign lipomas because both lesions consist of large amounts of fat tissue (more than 75%). However, a differentiation is usually possible because well-differentiated liposarcomas generally demonstrate a larger size and broader fibrous septa with a more nodular appearance compared to benign lipomas. Neurological symptoms as clinical manifestation of a well-differentiated liposarcoma are firstly described in this case.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsAH conceived the study. JB did research the literature review. AH, MG and JB prepared the manuscript. WAK edit and coordinated the manuscript.All authors read and approved the final manuscript.ConsentWritten informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.\n\nREFERENCES:\nNo References"
+}

batch_13/PMC2584958.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2584958",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2584958\nAUTHORS: R Torrisi, V Bagnardi, A Cardillo, F Bertolini, E Scarano, L Orlando, P Mancuso, A Luini, A Calleri, G Viale, A Goldhirsch, M Colleoni\n\nABSTRACT:\nThe antiangiogenic agent bevacizumab showed synergistic effects when combined with chemotherapy in advanced breast cancer. We presently investigated the activity of bevacizumab in combination with chemotherapy, including capecitabine and vinorelbine, and endocrine therapy, including letrozole (+triptorelin in premenopausal women), as primary therapy for patients with ER and/or PgR ⩾10% T2–T4a-c, N0–N2, M0 breast cancer. Biological end point included the proliferative activity (Ki67), whereas clinical end points were clinical response rate, pathological complete response (pCR) and tolerability. Circulating endothelial cells (CECs) and their progenitors, as surrogate markers of antiangiogenic activity, were measured at baseline and at surgery.Thirty-six women are evaluable. A clinical response rate of 86% (95% CI, 70–95) and no pCR were observed; Ki67 was significantly decreased by 71% (interquartile range, −82%, −62%). Toxicity was manageable: two grade 3 hypertension, four grade 3 deep venous thrombosis and no grade >2 proteinuria were observed. Treatment significantly decreased the percentage of viable CECs and prevented the chemotherapy-induced mobilisation of circulating progenitors. Basal circulating progenitors were positively associated with clinical response. In conclusion, bevacizumab is feasible and active in association with primary chemoendocrine therapy for ER-positive tumours in terms of proliferation inhibition, clinical response and antiangiogenic activity.\n\nBODY:\nPreoperative treatment of endocrine-responsive breast cancer is a matter of debate within the scientific community. The striking improvement in terms of pathological complete response (pCR) obtained in endocrine-unresponsive tumours with new schedules of anthracyclines and taxanes has had a lesser impact in ER- and PgR-positive tumours, which obtained pCR up to 10% (Mazouni et al, 2007).On the other hand, endocrine therapy, although associated with a high clinical response rate, almost rarely yields pCR (up to 3–6%) (Kaufmann et al, 2006). The association of chemoendocrine therapy, not systematically investigated in an appropriate population, has obtained discouraging results too (von Minckwitz et al, 2001; Bottini et al, 2005).We have investigated earlier the activity of the combination of a non-anthracycline- and taxane-based chemotherapy with endocrine therapy in a series of patients with ER- and PgR-positive tumours (Torrisi et al, 2008). Chemotherapy included six courses of capecitabine and oral vinorelbine administered in association with letrozole (+3-monthly triptorelin, if premenopausal). The results were rather disappointing, with a response rate of 62% and no patient obtaining a pCR (Torrisi et al, 2008).Angiogenesis represents a key process in the development and growth of cancer cells at any stage (Folkman, 1971). Vascular endothelial growth factor (VEGF) is the most potent and specific angiogenic factor and has been identified as a crucial regulator of both normal and pathological angiogenesis (Banerjee et al, 2007). Vascular endothelial growth factor acts as a mitogen for vascular endothelial cells and stimulates the secretion of enzymes involved in extracellular matrix degradation. Moreover, VEGF stimulates paracrine and autocrine signalling in endothelial cells (Banerjee et al, 2007). The recombinant humanised anti-human VEGF monoclonal antibody (rhuMAb, bevacizumab) inhibits several activities of VEGF, including endothelial cell growth, vascular permeability and angiogenesis (Kim et al, 1993; Willett et al, 2004). Bevacizumab has shown substantial activity in breast cancer and synergism with some chemotherapeutic agents (Sweeney et al, 2001; Cobleigh et al, 2003). The combination of bevacizumab and vinorelbine was investigated in a phase II study in patients with refractory advanced breast cancer, and a 31% objective response rate was observed (Burstein et al, 2002). In a phase III randomised study in breast cancer patients pretreated with anthracyclines and taxanes, the addition of bevacizumab significantly increased the response rate as compared with capecitabine alone (Miller et al, 2005). In untreated metastatic patients, the association of bevacizumab and paclitaxel significantly improved either response rate or progression-free survival as compared with paclitaxel alone (Miller et al, 2007). On the other hand, oestrogens are potent modulators of angiogenesis. It is thus conceivable that the combination of an antiangiogenic agent and a hormonal manipulation may result in an increased antitumour activity on an endocrine-sensitive tumour. In preclinical models, oestrogens increase endothelial cell proliferation and migration, whereas the effect of aromatase inhibition on angiogenic factors is less clear. In fact, in preclinical models, aromatase inhibitors downregulate VEGF expression, whereas no significant change of serum VEGF has been observed after preoperative anastrozole (Traina et al, 2007; Banerjee et al, 2008). Preclinical data also suggest that VEGF is involved in precocious stages of angiogenesis (Relf et al, 1997). It may thus be speculated that the earlier the angiogenic pathway is blocked, the greater the clinical effect that may be expected. Given the encouraging results obtained by the combination of capecitabine and vinorelbine in advanced breast cancer and the synergism with bevacizumab shown by both drugs, we thereby decided to investigate the combination of bevacizumab, capecitabine and vinorelbine and endocrine therapy in patients with ER- and/or PgR-positive locally advanced breast cancer. We defined as the principal measure of activity, the decrease of proliferation by Ki67, which has been shown to correlate better with prognosis than the absolute baseline value (Burcombe et al, 2005; Dowsett et al, 2005). Clinical activity in terms of either objective response rate, pCR or the tolerability of the combination was also evaluated.Finally, with the aim of identifying putative predictors and biomarkers of antiangiogenic activity, we determined baseline and post-treatment levels of circulating endothelial cells (CECs), either viable or apoptotic, and their progenitors and investigated their correlation with clinical outcome variables.Materials and methodsPatientsPatients with histologically proven T2–T4a-c, N0–N2, M0, ER and/or PgR ⩾10% and HER2-negative breast cancer were considered eligible for the study. Eligibility criteria also included no earlier chemotherapy/hormonotherapy, Eastern Cooperative Oncology Group performance status 0–2, measurable lesions, age between 18 and 70 years, white blood cells ⩾3000 per mm3, platelets ⩾100 000 per mm3, aspartate aminotransferase, alanine aminotransferase, ⩽2.5 × upper limit of normal and bilirubin ⩽3 mg per 100 ml.Patients with cardiac disease (congestive heart failure, history of myocardial infarction within the previous 3 months), severe vascular disease or uncontrolled concomitant infections were excluded. In addition, patients with a prior history of bleeding diathesis or coagulopathy, including deep venous thrombosis (DVT) or pulmonary embolism, recent (within last 6 months) or current history of gastrointestinal bleeding and current use of full-dose or parenteral anticoagulants or chronic daily treatment with aspirin (greater than 325 mg day−1) were excluded as were patients with 24 h urine protein greater than or equal to 500 mg or any active primary renal disease (excluding infection).Patients had baseline liver and renal function tests, electrolyte studies and complete blood count and urine analysis performed within 2 weeks of inclusion in the study. In addition, bilateral mammography and breast ultrasound, chest X-ray, abdominal ultrasound, bone scan or FDG-PET, serum CA 15.3 determination and electrocardiography were performed within 2 weeks from the start of treatment. Before starting the treatment and providing a signed informed consent, patients were submitted to a trucut to obtain a tumour sample to be stored for gene profiling and further molecular determination and a blood sample for the determination of circulating biomarkers (VEGF, CEC and their progenitors). Blood samples were also obtained before (day −1) and immediately after surgery (week +1).Willing patients were submitted to a trucut after 3 weeks (first cycle) for assessment of early change in Ki67.Written informed consent from all patients was obtained. The protocol was approved by the Ethical Committee.TreatmentPatients received chemotherapy containing capecitabine 2000 mg m−2 orally on days 1–14, vinorelbine 20 mg m−2 i.v. on days 1 and 3 (Nolè et al, 2006), and bevacizumab 15 mg kg−1 i.v. on day 1 every 3 weeks for eight courses. Bevacizumab (Avastin®; Roche, Basel, Switzerland) was administered as a 90′ i.v. infusion at the first administration and as 60′ and then 30′ infusions in the following courses. Bevacizumab was provided at no cost by Roche. A central venous catheter (CVC) in the subclavian or in the jugular vein contralateral to the site of the tumour was implanted in all patients before starting bevacizumab.Endocrine therapy consisted of letrozole 2.5 mg day−1. Treatment started on day 1 of the first course of chemotherapy in postmenopausal patients. In premenopausal patients, GnRH analogue (triptorelin 11.25 mg every 3 months) started on day 1 of chemotherapy and letrozole was added when oestradiol levels were in the postmenopausal range according to the IEO laboratory reference values. The median time to start of letrozole in premenopausal women was 50 days (interquartile (IQ) range, 42–69 days).Endocrine treatment was continued until the day of surgery. Surgery was planned 4 weeks after the last dose of bevacizumab.Toxicity and dose modificationsToxicity was evaluated according to NCIC-CTG 3.0 criteria by clinical and laboratory evaluations at day 21 of each cycle.Chemotherapy was postponed by 1 week if the blood count on day 21 showed a neutrophil count <1000 per mm3 and/or a platelet count <100 000 per mm3. If on day 28 the neutrophil count was >1000 per mm3 and platelet count >100 000 per mm3, the treatment was re-administered. If after 2 weeks of treatment delay (on day 35), haematologic recovery (neutrophils >1000 per mm3 and platelets >100 000 per mm3) was not obtained, chemotherapy was discontinued.Bevacizumab was held or definitively suspended in case of three to four haemorrhagic events, venous or arterial thromboembolic events, uncontrolled hypertension, grade 4 proteinuria, gastrointestinal perforations, wound healing complications and infusion-related allergic reactions.Response criteriaTumour was evaluated at baseline by physical measurement, with callipers, of the two largest diameters and by means of mammography and ultrasound. After four and eight cycles, patients also had mammography and ultrasound breast examination to assess response. Clinical responses were evaluated according to both radiological (breast ultrasound or mammography) and clinical evaluation, by measuring the largest diameters of the tumour, and graded according to standard RECIST criteria (Therasse et al, 2000).Patients with stable disease, partial remission or complete remission were candidates for four more courses of therapy. Pathological complete remissions were evaluated according to Kuerer et al (1999). A pCR was defined as the total disappearance of invasive tumour either in the breast or in the axilla. The presence of intraductal carcinoma qualified for pCR.Pathology and immunohistochemistryAll included patients had pathological evaluation performed at the EIO. Surgical specimens were extensively sampled for the evaluation of residual tumour after primary chemotherapy as published earlier.The immunostained slides were evaluated independently by two of the authors, as reported earlier. Only nuclear reactivity was taken into account for ER, PgR and Ki67 antigens. The results were recorded as the percentage of immunoreactive cells over at least 2000 neoplastic cells.HER2 status was defined at immunohistochemistry (IHC) as negative (faint and partial staining in >10% of cells=1+) and equivocal (faint and complete staining in >10% of cells=2+). In the latter cases, fluorescence in situ hybridisation (FISH) was performed to assess the amplification of the HER2 gene.Biological measurementsBlood samples were collected at baseline, immediately before surgery, and 1 week after surgery to estimate any change in CECs and their progenitors (CEPs).Peripheral blood samples were collected for measurement of circulating cells by six flow cytometry. Cell suspensions were evaluated by FACSCanto (Becton Dickinson, San Jose, CA, USA). The antibodies used were CD31 and CD146 (EC marker), CD45 (pan-haematopoietic marker), CD133 (AC133, progenitor/stem cell marker), CD34 (progenitor/stem cells, EC), VEGFR-1, VEGFR-2 and VEGFR-3.Fluorescently labelled isotype-matched IgG1 antibodies were used as control for analysis.Appropriate analysis gates were used to enumerate viable and apoptotic CECs and CEPs.After acquisition of at least 1 × 106 cells per blood sample, analyses were considered as informative when adequate numbers of events (i.e. >100) were collected in the CEC enumeration gates. Circulating endothelial cells were defined as negative for the haematopoietic marker CD45, positive for the endothelial markers CD31 and CD146, and negative for the progenitor marker CD133. Circulating endothelial progenitor cells were depicted by the expression of CD133. 7AAD was used to gain insight into CEC/CEP viability according to Philpott et al (1996).Sorted CECs, investigated by electron microscopy, were found to be bona fide endothelial cells by the presence of Weibel–Palade bodies. More than 75% of the circulating mRNAs of the endothelial-specific gene, VE-cadherin, found in the blood, were present in the sorted population. Coefficients of variation related to the CEC and CEP enumeration procedure by flow cytometry were 4±4 (intrareader), 17±4 (inter-reader) and 17±7% (variability over 0–72 h), respectively (P Mancuso et al, 2008, in press).Statistical methodsThe primary objective of the study was to estimate the effect of bevacizumab on the change in tumour cell proliferation by assessment of Ki67, a biomarker of cell proliferation, in breast cancer tissue at baseline (Ki67T0) and after 24–28 weeks (surgical resection specimen) of treatment (Ki67T1). The decrease was expressed in terms of relative reduction within each patient and calculated as ((Ki67T1–Ki67T0)/Ki67T0) × 100.The secondary efficacy objectives of the study were (a) to assess the effect of 24 weeks of bevacizumab on the change in tumour size, as assessed by ultrasound and mammography, and on the percentage of pCR as defined earlier; (b) to estimate any change in breast cell proliferation as measured by Ki67 in biopsies after 3 weeks; and (c) to estimate any change in CECs, in their sub-population and in the circulating progenitors, in blood at baseline, immediately before surgery, and 1 week after surgery.According to the historical data collected at our institute on about 400 T2–T4, N0–N2, M0 breast cancer patients receiving primary chemotherapy, about 25% of patients had a relative decrease greater than 80% (Colleoni et al, 2004).At the design phase, a sample size of 30 patients was planned, yielding an 80% power to detect a 50% proportion of patients with a relative decrease of Ki67 levels greater than 80%, twice as great as the proportion observed in historical controls (25%).The Fisher's exact test and the Wilcoxon signed-rank test were used to evaluate differences in the distribution of categorical and continuous variables, respectively. The Wilcoxon signed-rank test was used to evaluate differences within the same patients of continuous variables measured at two different time points.Subgroup analyses and/or analyses of secondary end points were exploratory in nature.All P-values were two sided. The statistical analyses were run using SAS version 8.2 (SAS Institute Inc., Cary, NC, USA).ResultsFrom May 2006 to January 2007, 37 patients were enrolled in the study. One patient was diagnosed with suspicious ischaemic cerebral alterations at the basal MRI and neither received bevacizumab nor was considered evaluable. Thirty-six patients were evaluated for clinical and biological end points.Baseline characteristics of patients and tumours are reported in Table 1. Median age was 44 years (range, 30–68). Twenty-five patients (69%) were premenopausal and 11 (31%) were postmenopausal as assessed by circulating gonadotrophins and oestradiol. Eighty-six per cent of tumours were T2–T3 and 69% were clinically nodal positive (at ultrasound and/or FDG-PET).As for protocol, all patients had ER ⩾10% tumours, whereas PgR was positive in 30 out of 36 tumours. HER2 was negative at IHC in all but one tumour, whereas FISH was negative. At baseline, Ki67 was >20% in 22 tumours (61%). At surgery, all tumours but one remained ER positive, whereas PgR switched to a negative phenotype in 25 out of 30 patients. The patient who switched to a negative ER phenotype also turned out to be HER2 positive at IHC and at FISH while remaining IHC 2+ and FISH negative at baseline.Responses after treatment are summarised in Table 2. A complete response was observed in one patient (3%) and a partial response was observed in 30 patients (83%) with a cumulative objective response rate of 86% (95% CI, 70–95). Fourteen patients had pathological negative nodes (39%; 95% CI, 23–56). Breast-conservative surgery was feasible in 64% of patients (95% CI, 46–79) and was raised to 74%, when excluding T4 tumours, who were candidated to mastectomy irrespective of response.In premenopausal patients, letrozole was started at a median of 50 days (IQ range, 43–69) after the first cycle of chemotherapy and bevacizumab at the achievement of oestradiol levels within the postmenopausal range. The overall median duration of letrozole was 139 days (IQ range, 119–168). The median duration of letrozole was 134 days (IQ range, 111–140) in premenopausal patients and 183 days (IQ range, 168–203) in postmenopausal patients, respectively (P<0.0001).Median Ki67 at baseline was 23.5 (IQ range, 14.5–29), whereas after treatment, the median value was 5 (IQ range, 3–9.5%). The median percent decrease was 71% (IQ range, −82%, −62%), with P-value <0.0001 (Wilcoxon signed-rank test) (Figure 1). Thirty-one percent of patients (95% CI, 16–48) had a decrease ⩾80%. The only factor significantly associated with Ki67 decrease >80% was the duration of letrozole treatment, in that 50% of patients receiving letrozole for more than the median time (139 days) experienced a decrease of Ki67 >80% as compared with 12% of patients treated for a shorter time (P=0.03). When the analysis was performed according to the menopausal status, this association was maintained with a borderline significance in premenopausal patients (140 vs 122 days, P=0.056), whereas it was completely lost in postmenopausal patients (191 vs 181 days, P=0.45).On the other hand, the duration of letrozole was not associated with the likelihood of obtaining a clinical response.Ki67 was determined after 3 weeks (first cycle of bevacizumab and chemotherapy) in 17 patients. In these patients, an early significant decrease was observed (median percent decrease 65% (IQ range, 44–69%) P<0.001).Toxicities grade >2 are summarised in Table 3. Bevacizumab was administered for eight courses in 26 patients. Bevacizumab was discontinued before the fourth cycle in two patients for the occurrence of DVT of subclavian (one case) and internal giugular vein (one case), in both cases corresponding to the site of insertion of the catheter of the CVC. In one further patient, bevacizumab was discontinued after the sixth cycle because of the occurrence of a seizure associated with MRI alteration referable to a leukoencephalopathy. A DVT of the giugular vein also occurred in this patient. A DVT of the femoral vein occurred in another patient after completion of the treatment before surgery. In one patient, bevacizumab was held for three cycles because of grade 3 hypertension, and two patients received six and seven cycles because of infection of the CVC and a delay in the implant of CVA, respectively. In the remaining five patients, treatment was discontinued before the eighth cycle because of grade >2 chemotherapy-related toxicity. Grade 3 hypertension was observed in two patients and grade 2 in seven patients, probably related to bevacizumab administration. One case of grade 2 proteinuria was observed that did not require drug discontinuation. Grade 1 bleeding, mainly epistaxis, was observed in five patients. No major surgical complication occurred after surgery except for one case of infection and one case of wound healing delay.Measurements of CECs and CEPs were feasible in all patients at baseline and in 29 patients at surgery, whereas only a small proportion of patients had blood samples drawn at 1 week after surgery, and statistical analyses were not performed for the latter group. Treatment did not significantly affect the total number of CEPs and CECs, although a trend towards reduction of CECs was observed (P=0.07). However, when different sub-populations of CECs were analysed separately, a significant increase in apoptotic CECs and a significant decrease in viable CECs was evident (Figure 2). At the same time, a significant increase in CD31+VEGFR-1+ was observed.When we analysed the correlation of biological parameters with clinical response, we observed that higher levels of baseline CEPs were positively correlated with the likelihood of obtaining a clinical response (P=0.026) (Figure 3) and showed a trend towards the prediction of a Ki67 decrease >80% (P=0.08). No other factor significantly correlated with clinical response, the likelihood of having pathological negative nodes at surgery or of obtaining a reduction of proliferative activity >80%.At the same time, the change of molecular parameters induced by treatment was not predictive of any clinical outcome variable except for a significant correlation between the CD31+/VEGFR-2+ sub-population of CECs and clinical response (P=0.017).DiscussionPreoperative treatment of endocrine-responsive tumours represents a challenge for medical oncologists. Although ER- and/or PgR-positive tumours respond poorly to primary chemotherapy in terms of pCR, long-term prognosis of these tumours is consistently better compared with that of ER-negative tumours (Guarneri et al, 2006; Colleoni et al, 2008). Different treatment approaches and surrogate biomarkers of activity are required in this subset of tumours.Angiogenesis represents a key process at multiple steps in breast carcinogenesis (Folkman, 1971). Different from other molecular-targeted drugs recently recognised in the therapeutic armamentarium of the medical oncologists, antiangiogenic drugs do not work through interaction with a specific target but by interfere with a pathway that is shared by all neoplastic cells. It is thus conceivable that endocrine-responsive tumours may benefit from a disruption of the angiogenic switch. Literature data confirm that VEGF expression is a significant prognostic factor of clinical outcome, irrespective of hormone receptor status (Banerjee et al, 2007). In patients receiving endocrine therapy for advanced breast cancer, tissutal VEGF levels were inversely related to clinical response (Foekens et al, 2001; Manders et al, 2003).Bevacizumab was shown to improve clinical activity of chemotherapy in advanced breast cancer, although no effect on overall survival has been observed (Miller et al, 2005, 2007), and the drug is currently investigated as adjuvant treatment of early breast cancer in association (concurrently or sequentially) with chemotherapy. Preclinical data suggest that the role of VEGF is crucial in the early stages of breast carcinogenesis, and the implication of an anti-VEGF therapy may be greater the earlier the treatment is introduced (Relf et al, 1997). The clinical and molecular activities of a single course of bevacizumab before adding chemotherapy with docetaxel were assessed in inflammatory breast cancer (Wedam et al, 2006). Significant decreases of VEGFR-2 and of vascular permeability (assessed by MRI parameters), correlated with clinical response, were observed (Wedam et al, 2006). Preliminary data on the combination of bevacizumab and chemotherapy have consistently reported a high response rate, although differences in pCR rate have been observed across studies (Lyons et al, 2006; Wedam et al, 2006; Greil et al, 2007).Short-term change of proliferative activity better than baseline levels after primary chemotherapy and endocrine therapy has been associated with clinical outcome (Chang et al, 1999; Dowsett et al, 2005). In a randomised study of preoperative therapy, the 2-week decline of Ki67 was greater after anastrozole as compared with tamoxifen and the combination of both agents, mirroring the results in terms of DFS obtained in the phase III randomised trial of adjuvant therapy with the three (Dowsett et al, 2005). The proliferative activity at surgery was also shown as a significant independent predictor of long-term outcome after primary chemotherapy (Jones et al, 2007). In our study, a dramatic decrease of Ki67 was observed after treatment, with a median post-treatment value of 5%. In addition, about one-third of patients achieved a decrease >80%. Interestingly, the likelihood of achieving a greater decrease was positively correlated with the duration of endocrine treatment, underscoring the relevance of an endocrine manipulation in patients with endocrine-responsive tumours. However, given the multiplicity of treatments and the limited number of patients considered, it is not possible to assess how the single treatment contributed to this dramatic decrease of proliferative activity. Limited data are available with the combination of bevacizumab and endocrine therapy, although preliminary data show a promising activity of the combination of bevacizumab and letrozole as primary therapy in postmenopausal women with ER-positive breast cancer (Forero-Torres et al, 2007).Although it has been proposed that assessing an antiangiogenic treatment by measuring tumour shrinkage may not fully reflect the antitumour activity of the drug, in this study, we observed that 86% of imaging confirmed clinical responses, a significant figure in this population of patients. In addition, the proportion of patients with pathological negative nodes is comparable with that reported in a population of ER-positive patients treated with anthracylines and taxanes containing chemotherapy (Guarneri et al, 2006). After treatment, 74% of T2–T3 tumours were treated with conservative surgery, which represents an appreciable figure. In an earlier study with capecitabine and oral vinorelbine in association with endocrine therapy in a similar population of patients, we observed a clinical response rate and a breast-conserving surgery rate of 62% (Torrisi et al, 2008). Although recognising the limitation of an indirect comparison and that other factors such as the longer duration of treatment (eight vs six courses) and the different routes of vinorelbine administration (intravenous vs oral) may have concurred with this difference, the addition of bevacizumab appeared to increase the clinical response and the breast-conserving rate as compared with chemoendocrine therapy, although not affecting the likelihood of obtaining pathological complete remission. The value of pCR as a surrogate end point of clinical outcome in ER-positive tumours after primary chemotherapy has been questioned, given the inconsistent results deriving from the analysis of retrospective series (Ring et al, 2004; Guarneri et al, 2006). The limited number of pCRs observed in both series among ER-positive tumours may account for this inconsistency, and the search for other markers to assess that the value of preoperative therapy in this sub-population is strongly encouraged.The identification of surrogate markers of antiangiogenic activity different from standard tumour measurement appears crucial for monitoring drug efficacy and for designing an optimal integration with conventional antitumoral strategies (Bertolini et al, 2006). An increase of apoptotic CECs was shown to positively correlate with clinical response and clinical benefit in patients treated with metronomic chemotherapy (Mancuso et al, 2006). In patients treated with neoadjuvant chemotherapy, significantly higher levels of CECs were observed in patients with ER-negative as compared with ER-positive tumours; in addition, an increase in CEPs as well as a reduction of CECs was observed after neoadjuvant chemotherapy (Furstenberger et al, 2006). An increase in CECs also predicted a worse outcome in patients with metastatic breast cancer receiving the combination of bevacizumab and letrozole (Traina et al, 2007). In our study, baseline levels of CECs in patients with newly diagnosed and localised breast cancer were lower than those observed in patients with advanced disease, supporting the hypothesis that the synthesis of proangiogenic factors increases along with tumour progression (Relf et al, 1997).As expected, treatment increased the proportion of apoptotic CECs while reducing the viable CECs. However, this effect was not correlated with any of the clinical outcome variables considered. A body of evidence suggests that CEPs are mobilised by chemotherapy (Bertolini et al, 2003; Furstenberger et al, 2006). As we did not observe any modulation of CEPs, we may speculate that bevacizumab prevented the chemotherapy-induced mobilisation of CEPs, as already shown in preclinical models (Shaked et al, 2006). On the other hand, given the positive correlation between higher baseline levels of circulating progenitors and the likelihood of obtaining a clinical response and the involvement of CEPs in neovascularisation, we may further speculate that these are patients who may benefit more from an antiangiogenic manipulation. In addition, we observed a significant correlation between the modulation of CD31+/VEGFR-2+ and the likelihood of obtaining a clinical response, in that patients who responded to treatment did not experience an increase of this sub-population of CEC. CD31+/VEGFR-2+ cells are crucial in the neovascularisation process and have recently been shown to be elevated in breast cancer patients to significantly correlate with tumour size and to rapidly decline after removal of the tumour, suggesting a possible induction by tumour-driven angiogenic stimuli (Richter-Ehrenstein et al, 2007).CD31+VEGFR-1+ CECs, a sub-population that still lacks an in-depth biological characterisation, were found to be significantly increased by treatment. Further studies have been planned to better understand this phenomenon.The manner in which bone marrow-derived progenitor cells contribute to neovascularisation either after differentiation in mature endothelial cells or after homing to sites of angiogenesis is still a matter of debate (Shaked et al, 2005). Our findings are consistent in showing that antiangiogenic treatment is also able to induce in vivo a biological response in terms of inhibition of the CEC and their progenitors involved in the angiogenic process. In addition, there is a suggestion towards the usefulness of some of these populations in monitoring biological activity, given the observed correlation with clinical response. These hypotheses should further be exploited in larger subsets of patients.We showed that bevacizumab may also be safely administered in the neoadjuvant setting with two minor postsurgical complications, which may be attributed to the antiangiogenic treatment. The rate of bevacizumab-related adverse events was consistent with previous data, except for 11% of DVT, which was greater than that reported in the phase III trials (Miller et al, 2007). However, the presence of a CVC and the concurrent use of cytotoxic agents, as vinorelbine, potentially inducing an endothelial damage may have concurred with this figure. Hypertension was in most cases grade 2 and was easily managed with therapy. It required drug discontinuation in only one case, whereas proteinuria was negligible.In conclusion, the combination of bevacizumab, endocrine therapy and tailored chemotherapy induced a high clinical response rate in ER-positive breast cancer, whereas proliferative activity of the tumour was dramatically reduced by the treatment. Moreover, the duration of endocrine treatment with letrozole correlated positively with the chance of obtaining a greater Ki67 decrease. Molecular analyses showed that treatment significantly reduced viable cells. Interestingly, it also appeared to prevent chemotherapy-induced CEP mobilisation, thus interfering with neovascularisation. In addition, patients with higher baseline levels of CEPs were more likely to experience a clinical response.Further studies investigating optimal timing, duration and combination of antiangiogenic agents with conventional antineoplastic drugs, and particularly the identification of surrogate markers of antiangiogenic activity, are warranted.\n\nREFERENCES:\n1. Banerjee S, Dowsettt M, Ashworth A, Martin L-A (2007) Mechanisms of disease: angiogenesis and the management of breast cancer. Nat Clin Pract\n4: 536–550\n2. Banerjee S, Pancholi S, A’Hern R, Ghazoui Z, Smith IE, Dowsett M, Martin L-A (2008) The effects of neoadjuvant anastrozole and tamoxifene on circulating vascular endothelial growth factor and soluble vascular endothelial growth factor receptor 1 in breast cancer. Clin Cancer Res\n14: 2656–266318451229\n3. Bertolini F, Paul S, Mancuso P, Monestiroli S, Gobbi A, Shaked Y, Kerbel RS (2003) Maximum tolerable dose and low-dose metronomic therapy have opposite effects on the mobilization and viability of circulating endothelial progenitor cells. Cancer Res\n53: 4342–4346\n4. Bertolini F, Shaked Y, Mancuso P, Kerbel RS (2006) The multifaceted circulating endothelial cell in cancer: towards marker and target identification. Nat Rev Cancer\n6: 835–84517036040\n5. Bottini A, Berruti A, Brizzi MP, Bersiga A, Generali D, Allevi G, Aguggini S, Bolsi G, Bonardi S, Tondelli B, Vana F, Tampellini M, Alquanti P, Dogliotti L (2005) Cytotoxic and antiproliferative activity of the single agent epirubicin versus epirubicin plus tamoxifen as primary chemotherapy in human breast cancer: a single-institution phase III trial. Endocr Relat Cancer\n12: 383–39215947110\n6. Burcombe RJ, Makris A, Richman PI, Daley FM, Noble S, Pittam M, Wright D, Allen SA, Dove J, Wilson GD (2005) Evaluation of ER, PgR, HER-2 and Ki-67 as predictors of response to neoadjuvant anthracycline chemotherapy for operable breast cancer. Br J Cancer\n92: 147–15515611798\n7. Burstein HJ, Parker LM, Savoie J, Younger J, Kuter I, Ryan PD, Garber JE, Campos SM, Shulman LN, Harris LN, Gelman R, Winer EP (2002) Phase II trial of the anti-VEGF antibody bevacizumab in combination with vinorelbine for refractory advanced breast cancer. Breast Cancer Res Treat\n76(Suppl 1): S115\n8. Chang J, Powles TJ, Allred DC, Ashley SE, Clark GM, Makris A, Assersohn L, Gregory RK, Osborne CK, Dowsett M (1999) Biologic markers as predictors of clinical outcome from systemic therapy for primary operable breast cancer. J Clin Oncol\n17: 3058–306310506600\n9. Cobleigh MA, Langmuir VK, Sledge GW, Miller KD, Haney L, Novotny WF, Reimann JD, Vassel A (2003) A phase I/II dose-escalation trial of bevacizumab in previously treated metastatic breast cancer. Semin Oncol\n30(Suppl 16): 117–12414613032\n10. Colleoni M, Viale G, Zahrieh D, Bottiglieri L, Gelber RD, Veronesi P, Balduzzi A, Torrisi R, Luini A, Intra M, Dellapasqua S, Cardillo A, Ghisini R, Peruzzotti G, Goldhirsch A (2008) Expression of ER, PgR, HER1, HER2 and response : a study of preoperative chemotherapy. Ann Oncol\n19: 465–47217986623\n11. Colleoni M, Viale G, Zahrieh D, Pruneri G, Gentilini O, Veronesi P, Gelber RD, Curigliano G, Torrisi R, Luini A, Intra M, Galimberti V, Renne G, Nolè F, Peruzzotti G, Goldhirsch A (2004) Chemotherapy is more effective in patients with breast cancer not expressing steroid hormone receptors: a study of preoperative treatment. Clin Cancer Res\n10: 6626–6628\n12. Dowsett M, Ebbs SR, Dixon M, Skene A, Griffith C, Boeddinghaus I, Walter J, Detre S, Hills M, Ashley S, Francis S, Walsh G, Smith IE (2005) Biomarker changes during neoadjuvant anastrozole, tamoxifen, or the combination: influence of hormonal status and HER2 in breast cancer – a study from the IMPACT trialists. J Clin Oncol\n23: 2477–249215767642\n13. Foekens JA, Peters HA, Grebenchtchickov N, Look MP, Mejier-van Gelder ME, Geurts-Moespot A, van der Kwast TH, Sweep CG, Klijn JG (2001) High tumor levels of vascular endothelial growth factor predict poor response to systemic therapy in advanced breast cancer. Cancer Res\n61: 5407–541411454684\n14. Folkman J (1971) Tumor angiogenesis: therapeutic implications. N Engl J Med\n285: 1182–11864938153\n15. Forero-Torres A, Percent J, Galleshaw L, Nabell J, Carpenter C, Falkson C, Jones H, Krontriras J, De Los Santos M, Saleh UAB, Breast Cancer Working Group (2007) A study of preoperative (neoadjuvant) letrozole in combination with bevacizumab in postmenopausal women with newly diagnosed operable breast cancer. A preliminary safety report. Proc Am Soc Clin Oncol\n25(18S): 11020\n16. Furstenberger G, von Moos R, Lucas R, Thurlimann B, Senn HJ, Hamacher J, Boneberg EM (2006) Circulating endothelial cells and angiogenic serum factors during neoadjuvant chemotherapy of primary breast cancer. Br J Cancer\n94: 524–53116450002\n17. Greil R, Moik M, Reitsamer R, Ressler S, Stoll M, Namberger K, Menzel C, Mlineritsch B (2007) Capecitabine+docetaxel+bevacizumab as neoadjuvant therapy for invasive breast cancer: results of a phase II pilot study. Breast Cancer Res Treat\n106: S188 (abstract 4064)\n18. Guarneri V, Broglio K, Kau SW, Cristofanilli M, Buzdar AU, Valero V, Buchholz T, Meric F, Middleton L, Hortobagyi GN, Gonzalez-Angulo AM (2006) Prognostic value of pathologic complete response after primary chemotherapy in relation to hormone receptor status and other factors. J Clin Oncol\n24: 1037–104416505422\n19. Jones RL, Salter A, Nerukar M, Parton M, A’Hern R, Smith IE, Dowsett M (2007) Prognostic significance of Ki67 before and after neoadjuvant chemotherapy in early breast cancer. Proc Am Soc Clin Oncol\n25(18S): 573\n20. Kaufmann M, Hortobagyi GN, Goldhirsch A, Scholl S, Makris A, Valagussa P, Blohmer JU, Eiermann W, Jackesz R, Jonat W, Lebeau A, Loibl S, Miller W, Seeber S, Semiglazov V, Smith R, Souchon R, Stearns V, Untch M, von Minckwitz G (2006) Recommendations from an International Export Panel on the use of neoadjuvant (primary) systemic treatment of operable breast cancer: an update. J Clin Oncol\n24: 1940–194916622270\n21. Kim KJ, Li B, Winer J, Armanini M, Gillett N, Phillips HS, Ferrara N (1993) Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in vivo. Nature\n362: 841–8447683111\n22. Kuerer HM, Newman LA, Smith TM, Ames FC, Hunt KK, Dhingra K, Theriault RL, Singh G, Binkley SM, Sneige N, Bucholz TA, Ross MI, McNeese MD, Buzdar AU, Hortobagyi GN, Singletary SE (1999) Clinical course of breast cancer patients with complete pathologic primary tumor and axillary lymph node response to doxorubicin-based neoadjuvant chemotherapy. J Clin Oncol\n17: 460–46910080586\n23. Lyons JA, Silverman P, Remick S, Chen H, Leeming R, Shenk R, Fu P, Dumadag L, Escuro K, Overmoyer B (2006) Toxicity results and early outcome on a randomized phase II study of docetaxel±bevacizumab for locally advanced, unresectable breast cancer. Proc Am Soc Clin Oncol\n24(18S): 3049\n24. Mancuso P, Colleoni M, Calleri A, Orlando L, Maisonneuve P, Pruneti G, Agliano A, Goldhirsch A, Shaked Y, Kerbel RS, Bertolinif F (2006) Circulating endothelial-cell kinetics and viability predict survival in breast cancer patients receiving metronomic chemotherapy. Blood\n108: 452–45916543470\n25. Manders P, Beex LV, Tjan-Helijnen VC, Span PN, Sweep CG (2003) Vascular endothelial growth factor is associated with the efficacy of endocrine therapy in patients with advanced breast carcinoma. Cancer\n98: 2125–213214601081\n26. Mazouni C, Kau SW, Frye D, Andre F, Kuere HM, Bucholz TA, Symmans WF, Anderson K, Hess KR, Gonzalez-Angulo AM, Hortobagyi GN, Buzdar AU, Pusztai L (2007) Inclusion of taxanes, particularly weekly paclitaxel, in preoperative chemotherapy improves pathologic complete response rate in estrogen receptor positive breast cancers. Ann Oncol\n18: 874–88017293601\n27. Miller K, Wang M, Gralow J, Dickler M, Cobleigh MA, Perez EA, Shenkier T, Cella D, Davidson NE (2007) Paclitaxel plus bevacizumab versus paclitaxel alone for metastatic breast cancer. N Engl J Med\n357: 2666–267618160686\n28. Miller KD, Chap LI, Holmes FA, Cobleigh MA, Marcom PK, Fehrenbacher L, Dickler M, Overmoyer BA, Reimann JD, Sing AP, Langmuir V, Rugo HS (2005) Randomized phase III trial of capecitabine compared with bevacizumab plus capecitabine in patients with previously treated metastatic breast cancer. J Clin Oncol\n23: 792–79915681523\n29. Nolè F, Catania C, Munzone E, Rocca A, Verri E, Sanna G, Ascione G, Adamoli L, Zampino MG, Minchella I, Goldhirsch A (2006) Capecitabine/vinorelbine: an effective and well-tolerated regimen for women with pretreated advanced stage breast cancer. Clin Breast Cancer\n6: 518–52416595035\n30. Philpott NJ, Turner AJ, Scopes J, Westby M, Marsh JC, Gordon-Smith EC, Dalgleish AG, Gibson FM (1996) The use of 7-amino actinomycin D in identifying apoptosis: simplicity of use and broad spectrum of application compared with other techniques. Blood\n87: 2244–22518630384\n31. Relf M, LeJeune S, Scott PA, Fox S, Smith K, Leek R, Moghaddam A, Whitehouse R, Bicknell R, Harris AL (1997) Expression of the angiogenic factprs vascular endothelial growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet derived endothelial growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and is relation to angiogenesis. Cancer Res\n57: 963–9699041202\n32. Richter-Ehrenstein C, Rentzsch J, Runkel S, Schneider A, Schonfelder G (2007) Endothelial progenitor cells in breast cancer patients. Breast Cancer Res Treat\n106: 343–34917972175\n33. Ring AE, Smith IE, Ashley S, Fulford LG, Lakhani SR (2004) Oestrogen receptor status, pathological complete response and prognosis in patients receiving neoadjuvant chemotherapy for early breast cancer. Br J Cancer\n91: 2012–201715558072\n34. Shaked Y, Bertolini F, Man S, Rogers MS, Cervi D, Foutz T, Rawn K, Voskas D, Dumont DJ, Ben-David Y, Lawler J, Henkin J, Huber J, Hicklin DJ, D’Amato RJ, Kerbel RS (2005) Genetic heterogeneity of the vasculogenic phenotype parallels angiogenesis: implications for cellular surrogate marker analysis of antiangiogenesis. Cancer Cell\n7: 101–11115652753\n35. Shaked Y, Ciarrocchi A, Franco M, Lee CR, Man S, Cheung AS, Hicklin DJ, Chaplin D, Foster FS, Benezra R, Kerbel RS (2006) Therapy-induced acute recruitment of circulating endothelial progenitor cells to tumours. Science\n313: 1785–178716990548\n36. Sweeney CJ, Miller KD, Sissons SE, Nozaki S, Heilman DK, Shen J, Sledge Jr GW (2001) The antiangiogenic property of docetaxel is synergistic with a recombinant humanized monoclonal antibody against vascular endothelial growth factor or 2-methoxiestradiol but antagonized by endothelial growth factors. Cancer Res\n61: 3369–367211309294\n37. Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, Rubinstein L, Verweji L, Van Glabbeke M, van Osteroom AT, Christian MC, Gwyther SG (2000) New guidelines to evaluate the response to treatment in solid tumours. J Natl Cancer Inst\n92: 205–21610655437\n38. Torrisi R, Dellapasqua S, Ghisini R, Viale G, Veronesi P, Luini A, Intra M, Peruzzotti G, Rocca A, Balduzzi A, Cardillo A, Goldhirsch A, Colleoni M (2008) Preoperative concurrent chemo- and endocrine therapies for women with large operable breast cancer expressing steroid hormone receptors. The Breast; (e-pub ahead of print)\n39. Traina TA, Rugo HS, Dickler M (2007) Bevacizumab for advanced breast cancer. Hematol Oncol Clin North Am\n21: 303–31917512451\n40. von Minckwitz G, Costa SD, Raab G, Blohmer JU, Eidtmann H, Hilfrich J, Merkle E, Jackisch C, Gademann G, Tulusan AH, Eiermann W, Graf G, Kaufmann M (2001) Dose-dense doxorubicin, docetaxel and granulocyte colony-stimulating factor support with or without tamoxifen as preoperative therapy in patients with operable carcinoma of the breast: a randomized, controlled open phase IIb study. J Clin Oncol\n19: 3506–351511481357\n41. Wedam SB, Low JA, Yang SX, Chow CK, Choyke P, Danforth DH, Hewitt SM, Berman A, Steinberg SM, Liewehr DJ, Plehn J, Doshi A, Thomasson D, McCarthy N, Koeppen H, Sherman M, Zujewski J, Camphausen K, Chen H, Swain SM (2006) Antiangiogenic and antitumour effects of bevacizumab in patients with inflammatory and locally advanced breast cancer. J Clin Oncol\n24: 769–77716391297\n42. Willett CG, Boucher Y, di Tomaso E, Duda DG, Munn LL, Tong RT, Chung DC, Sahani DV, Kalva SP, Kozin SV, Mino M, Cohen KS, Scadden DT, Hartford AC, Fischman AJ, Clark JW, Ryan DP, Xhu AX, Blaszkowsky LS, Chen HX, Shellito PC, Lauwers GJ, Jain RK (2004) Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med\n10: 145–14714745444"
+}

batch_13/PMC2586018.json

@@ -0,0 +1,4 @@
+{
+    "id": "PMC2586018",
+    "text": "This is an academic paper. This paper has corpus identifier PMC2586018\nAUTHORS: Lorin M Benneker, Christoph Leitner, Luca Martinolli, Kretschmer Robert, Heinz Zimmermann, Aristomenis K Exadaktylos\n\nABSTRACT:\nBackgroundSerum protein S-100B determinations have been widely proposed in the past as markers of traumatic brain injury and used as a predictor of injury severity and outcome. The purpose of this prospective observational case series was therefore to determine S-100B serum levels in patients with isolated injuries to the back.MethodsBetween 1 February and 1 May 2008, serum samples for S-100B analysis were obtained within 1 hour of injury from 285 trauma patients. All patients with a head injury, polytrauma, and intoxicated patients were excluded to select isolated injuries to the spine. 19 patients with isolated injury of the back were included. Serum samples for S-100B analysis and CT spine were obtained within 1 hours of injury.ResultsCT scans showed vertebral fractures in 12 of the 19 patients (63%). All patients with fractures had elevated S-100B levels. Amongst the remaining 7 patients without a fracture, only one patient with a severe spinal contusion had an S-100B concentration above the reference limit. The mean S-100B value of the group with fractures was more than 4 times higher than in the group without fractures (0.385 vs 0.087 μg/L, p = 0.0097).ConclusionOur data, although limited due to a very small sample size, suggest that S-100B serum levels might be useful for the diagnosis of acute vertebral body and spinal cord injury with a high negative predictive power. According to the literature, the highest levels of serum S-100B are found when large bones are fractured. If a large prospective study confirms our findings, determining the S-100B level may contribute to more selective use of CT and MRI in spinal trauma.\n\nBODY:\nBackgroundThe reliable detection of spinal injury remains difficult despite major progress in imaging techniques. Computed tomography and now MRI are recommended as the gold standard for the early detection of spinal injuries. The indication for computed tomography in spinal trauma patients is still controversial, as not all hospitals have 24-hour access to such modalities and transfers are cost-intensive and may be linked to increased morbidity and mortality. Serum protein S-100B determinations have been widely proposed in the past as markers of traumatic brain injury and used as a predictor of injury severity and outcome. [1-4] S-100 protein is a low molecular weight protein found in vertebrates characterized by two calcium binding sites of the helix-loop-helix (\"EF-hand type\") conformation. There are at least 21 different types of S-100 protein. The name is derived from the fact that the protein is 100% soluble in ammonium sulfate at neutral pH. S-100 is normally present in cells derived from the neural crest (Schwann cells, melanocytes, and glial cells), chondrocytes, adipocytes, myoepithelial cells, macrophages, Langerhans cells, dendritic cells, and keratinocytes. S-100 proteins have been implicated in a variety of intracellular and extracellular functions. S100 proteins are involved in regulation of protein phosphorylation, transcription factors, Ca++ homeostasis, the dynamics of cytoskeleton constituents, enzyme activities, cell growth and differentiation, and the inflammatory response.[5] Recent studies suggested that S-100B protein may be a biomarker for traumatic spinal cord injury in an animal model [6]. In trauma patients, S-100B shows a high negative predictive power for head injuries, but is also elevated in larger extracranial injuries, especially in fractures. [7,8] However, little is known about the effects of isolated vertebral injuries on S-100B serum concentrations.The purpose of this prospective observational case series was therefore to determine the levels of S-100B serum levels in patients with isolated injuries to the spine. Our aim was to determine whether S-100 B is a reliable serum marker for spinal injuries and, if so, whether it is reliable in spinal injury both with and without fractures of the spinal column.MethodsOur site is a Swiss level-one emergency unit with a 24-hour trauma service, seeing about 500 severely injured (ISS > 16) patients per year.Between 1 February and 1 May 2008, serum samples for S-100B analysis were obtained within 1 hour of injury from285 trauma patients. All patients with a head injury, GCS < 15, polytrauma, patients with limb or torso injuries, intoxicated or sedated patients, and patients with an unreliable history were excluded to select isolated injuries to the spine. 19 patients with isolated injuries to the spine remained for further analysis. The samples were assessed using a S-100 ECLIA (electrochemiluminescence immunoassay) method on a Modular Analytics E170 automated analyser (Roche Diagnostics AG, Rotkreuz, Switzerland). S-100B serum were designated as positive if concentrations were higher than 0.10 μg/L, the upper 95% reference limit for a normal healthy population. [7]CT scans of the head and the complete spine from all patients were obtained using a 16-row detector CT scanner (Sensation16, Siemens, Erlangen, Germany) with a collimation between 0.75 and 1.5 mm. The scans were evaluated for traumatic lesions independently by a neuroradiologist and a spine surgeon who were blinded to the results of the S-100B measures.ResultsCT scans showed vertebral fractures in 12 of the 19 patients (63%). All patients with fractures had elevated S-100B levels. Amongst the remaining 7 patients without a fracture, only one patient with a severe spinal contusion had an S-100B concentration above the reference limit. 4 of the 19 patients showed neurological symptoms; all these patients has clearly elevated S100B concentrations and all but the above mentioned patient with a spinal contusion had instable fractures (Table 1). The mean S-100B value in the group with fractures was more than 4 times higher than in the group without fractures (0.385 vs 0.087 μg/l, p = 0.0097). No correlation was found between age, gender or location of the fracture. Overall it seems that fractures of larger vertebrae and more complex fractures result in higher S-100B values. (Table 1)Table 1Patient characteristics, S 100B levels and diagnosisNogenderageS100B ug/Ltype of injury1f590.046strain trauma, no fracture2m310.658compression fracture T6, T11 (Type A 1.2)3f210.238burst fracture C7 (Type A 3.1)4m300.225bullet through massa lateralis C1 and spinal cord, no neurological symptoms5m170.115compression fracture T7- 9 (A 1.1)6m400.226burst fracture L1 (A 3.1)7m280.067strain trauma, no fracture8m320.072strain trauma, no fracture9m750.253compression fractures L4, L5 (A 1,2)10m300.108inferior articular process and arcus fracture C611m380.715flexion/distraction fracture T5/6 (B 2.3), paraplegia12m440.063distortion, no fracture13m840.201dense fracture type 2, Jefferson fracture C114m630.221contusio spinalis C3-5, transient tetraplegia15m680.481subluxation C5/6 with fracture C6 (B 1.2.1)16m220.048strain trauma, no fracture17m440.921burst-split fracture L1 (A 3.3), transient hyposthesia18f600.483luxation fracture C6-7, radiculopathy C619f170.096thoracal vertebral contusion, no fractureDiscussionS100B has a high positive and negative predictive power, and the finding of a normal S100B value shortly after trauma may exclude significant injury. Our data, although limited due to a very small sample size, suggest that S-100B serum levels might be a useful tool for the diagnosis of acute vertebral body and spinal cord injury. According to the literature, the highest levels of serum S-100B are found when large bones are fractured [7]. The clear correlation between the vertebral fractures and elevated S-100B serum concentrations and the tendency to higher values with more complex fractures of larger bones make the bone marrow a more likely source of S-100B in serum than potentially injured nerve tissue. However, spinal cord injuries without fractures can also apparently elevate S-100B serum concentrations, although our sample was too small for reliable conclusions or statistics. If a large prospective study confirms our findings, determining the S-100B level may contribute to more selective use of CT and MRI in spinal trauma. It should be borne in mind that positive S-100B results may be over interpreted as a reliable positive predictor for injuries to the spine, especially if further injuries are present. In our sample, however, normal S-100B concentration were not found in the presence of relevant injuries. A limitation of our study is that axonal injuries cannot be excluded, even in patients without history of percussive trauma to the head. Such patients may show elevated S-100B levels as may patients with large soft-tissue injuries, although our aim was to exclude such cases we tried to exclude such as far as possible [1,7].ConclusionThe clinical impact of our positive S-100B findings is limited. It can, however, be assumed that patients with reduced consciousness (intoxicated, intubated, sedated) and a negative S-100B have no significant brain or spinal injury. This might contribute to sparing patients from unnecessary exposure to radiation, saving resources and ultimately reducing costs. Although our results are promising, no firm conclusions can be drawn at present. Future efforts to develop biomarkers that predict functional outcomes in the acute phase of spinal injury should also focus on axon-specific proteins.[8]Competing interestsThe authors declare that they have no competing interests.Authors' contributionsLMB was responsible for data analysis and writing of the manuscript. CL and LM were collecting data. RK conducted all biochemical assessments. HZ performed the statistical analysis. AKE was responsible for study design and writing of the manuscript. All authors read and approved the final manuscript.\n\nREFERENCES:\nNo References"
+}