Limits...
Genetic, molecular and physiological basis of variation in Drosophila gut immunocompetence.

Bou Sleiman MS, Osman D, Massouras A, Hoffmann AA, Lemaitre B, Deplancke B - Nat Commun (2015)

Bottom Line: Gut immunocompetence involves immune, stress and regenerative processes.Using genome-wide association analysis, we identify several novel immune modulators.This genetic and molecular variation is physiologically manifested in lower ROS activity, lower susceptibility to ROS-inducing agent, faster pathogen clearance and higher stem cell activity in resistant versus susceptible lines.

View Article: PubMed Central - PubMed

Affiliation: 1] Global Health Institute, School of Life Sciences, Station 19, EPFL, 1015 Lausanne, Switzerland [2] Institute of Bioengineering, School of Life Sciences, Station 19, EPFL, 1015 Lausanne, Switzerland.

ABSTRACT
Gut immunocompetence involves immune, stress and regenerative processes. To investigate the determinants underlying inter-individual variation in gut immunocompetence, we perform enteric infection of 140 Drosophila lines with the entomopathogenic bacterium Pseudomonas entomophila and observe extensive variation in survival. Using genome-wide association analysis, we identify several novel immune modulators. Transcriptional profiling further shows that the intestinal molecular state differs between resistant and susceptible lines, already before infection, with one transcriptional module involving genes linked to reactive oxygen species (ROS) metabolism contributing to this difference. This genetic and molecular variation is physiologically manifested in lower ROS activity, lower susceptibility to ROS-inducing agent, faster pathogen clearance and higher stem cell activity in resistant versus susceptible lines. This study provides novel insights into the determinants underlying population-level variability in gut immunocompetence, revealing how relatively minor, but systematic genetic and transcriptional variation can mediate overt physiological differences that determine enteric infection susceptibility.

No MeSH data available.


Related in: MedlinePlus

Specific gene expression signatures define susceptibility to bacterial enteric infection.(a) Venn diagram showing differentially expressed genes (as revealed by RNA-seq experiments) between four resistant and four susceptible DGRP lines, in the unchallenged condition and 4 h post Pseudomonas entomophila infection (q-value<0.2, two-fold change). Genes in red and green have higher levels in susceptible and resistant lines respectively. The number of genes (black) indicated in the intersections represents the total number of non-differentially expressed genes. (b) Principal component analysis (PCA) on the top 2,000 varying genes between the 16 samples reveals that resistant lines cluster separately from susceptible lines, before (UC) and post-P. entomophila infection. PC1 separates samples based on treatment whereas PC2 separates them based on susceptibility class. (c) Modulated modularity clustering analysis on the top 2,000 varying genes identifies 24 correlated transcriptional modules (n≥15 genes). Each coloured point represents the spearman correlation (rs) between two genes. (d) A selection of functional categories identified by gene ontology (GO) analysis of genes belonging to the different modules identified in c (excluding the largest module with n=523, Supplementary Data 3). For the GO analysis, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID). (e) PCA using the expression levels of genes within each of the 24 modules identifies module #96 as the only module for which the lines are clearly separated on the first principal component according to treatment and susceptibility. (f) Heat map of gene expression levels in module #96 reveals important differences across susceptibility classes and treatment conditions.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4525169&req=5

f4: Specific gene expression signatures define susceptibility to bacterial enteric infection.(a) Venn diagram showing differentially expressed genes (as revealed by RNA-seq experiments) between four resistant and four susceptible DGRP lines, in the unchallenged condition and 4 h post Pseudomonas entomophila infection (q-value<0.2, two-fold change). Genes in red and green have higher levels in susceptible and resistant lines respectively. The number of genes (black) indicated in the intersections represents the total number of non-differentially expressed genes. (b) Principal component analysis (PCA) on the top 2,000 varying genes between the 16 samples reveals that resistant lines cluster separately from susceptible lines, before (UC) and post-P. entomophila infection. PC1 separates samples based on treatment whereas PC2 separates them based on susceptibility class. (c) Modulated modularity clustering analysis on the top 2,000 varying genes identifies 24 correlated transcriptional modules (n≥15 genes). Each coloured point represents the spearman correlation (rs) between two genes. (d) A selection of functional categories identified by gene ontology (GO) analysis of genes belonging to the different modules identified in c (excluding the largest module with n=523, Supplementary Data 3). For the GO analysis, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID). (e) PCA using the expression levels of genes within each of the 24 modules identifies module #96 as the only module for which the lines are clearly separated on the first principal component according to treatment and susceptibility. (f) Heat map of gene expression levels in module #96 reveals important differences across susceptibility classes and treatment conditions.

Mentions: Variability in survival and physiology among DGRP lines could in part be explained by system-specific transcriptional differences. We therefore performed RNA-seq on 16 gut samples comprising the same four susceptible and four resistant lines as introduced above in the unchallenged condition and 4 h after P. entomophila infection (Supplementary Fig. 7a). Genes (1287) were differentially expressed 4 h post infection compared with the unchallenged condition when all eight lines were treated as replicates (false discovery rate (FDR) adjusted P-value<0.05 and two-fold change, Supplementary Data 1). This set of genes overlaps with what we have previously shown when characterizing the gut transcriptional response to P. entomophila infection, even though that analysis was carried out using microarrays and on a different genetic background (OregonR)31. However, when we looked for differences in gene expression between the four resistant and four susceptible lines by pooling the samples of each susceptibility class, very few genes exhibited significant differential gene expression. Specifically, the expression of only 5 and 34 genes were changed in the unchallenged and challenged guts, respectively, when comparing phenotypic classes (Fig. 4a; Supplementary Data 2). This may reflect reduced statistical power given the large number of genes that are compared. In addition, it is possible that small but systematic differences in gene expression collectively differentiate resistant from susceptible profiles. We therefore performed principal component analysis (PCA) on 2000 genes with the highest expression variance in the 16 transcriptomes. Since infection status has a large impact on the transcriptome, expression profiles derived from infected samples were separated from those of unchallenged samples on the first principal component (PC), which explains 53% of the variance (Fig. 4b). Strikingly, even before infection, profiles of resistant lines were separated from those of susceptible lines based on the second PC, which explains 7.3% of the variance (Fig. 4b). This separation implies that the basal intestinal transcriptional state of resistant lines is distinct from that of susceptible lines, which may either define or reflect a molecular pre-disposition to enteric infection susceptibility. To dissect the molecular signatures that underlie this transcriptional stratification of the two phenotypic classes, we performed modulated modularity clustering37 on the same 2000 genes. We identified 24 transcriptional modules including >15 correlated genes (Fig. 4c; Supplementary Data 3). On the basis of Gene Ontology analysis and manual annotation38, we assigned the genes within the modules to six functional groups (Fig. 4d). To identify those modules whose gene levels clearly separate the lines according to treatment and phenotypic class, we systematically performed PCA on each module by taking the expression levels of its genes (Fig. 4e). We found that in module #96, samples are clearly separated on the first PC, even though the probability for such a separation to spuriously occur is <3 in 10,000 (Fig. 4e; Supplementary Fig. 7b,c). This module contains 20 genes, of which nine are related to stress response and most notably to ROS metabolism (Fig. 4e,f) and collectively explains 29% of the observed phenotypic variation (Supplementary Table 5). Other modules such as #102 (16 genes) also separated the samples on the first two PCs (Supplementary Fig. 8). Interestingly, module #102 likewise contains several ROS-related genes such as Cyp6a9 and Thioredoxin-2 (Trx-2)39. ROS are essential signalling molecules and immune effectors that are induced by the infected gut to neutralize pathogens13 and promote intestinal renewal14. However, a high ROS load can also cause inhibition of protein translation and consequently severe intestinal damage31, necessitating a finely tuned regulation of ROS production and metabolism40.


Genetic, molecular and physiological basis of variation in Drosophila gut immunocompetence.

Bou Sleiman MS, Osman D, Massouras A, Hoffmann AA, Lemaitre B, Deplancke B - Nat Commun (2015)

Specific gene expression signatures define susceptibility to bacterial enteric infection.(a) Venn diagram showing differentially expressed genes (as revealed by RNA-seq experiments) between four resistant and four susceptible DGRP lines, in the unchallenged condition and 4 h post Pseudomonas entomophila infection (q-value<0.2, two-fold change). Genes in red and green have higher levels in susceptible and resistant lines respectively. The number of genes (black) indicated in the intersections represents the total number of non-differentially expressed genes. (b) Principal component analysis (PCA) on the top 2,000 varying genes between the 16 samples reveals that resistant lines cluster separately from susceptible lines, before (UC) and post-P. entomophila infection. PC1 separates samples based on treatment whereas PC2 separates them based on susceptibility class. (c) Modulated modularity clustering analysis on the top 2,000 varying genes identifies 24 correlated transcriptional modules (n≥15 genes). Each coloured point represents the spearman correlation (rs) between two genes. (d) A selection of functional categories identified by gene ontology (GO) analysis of genes belonging to the different modules identified in c (excluding the largest module with n=523, Supplementary Data 3). For the GO analysis, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID). (e) PCA using the expression levels of genes within each of the 24 modules identifies module #96 as the only module for which the lines are clearly separated on the first principal component according to treatment and susceptibility. (f) Heat map of gene expression levels in module #96 reveals important differences across susceptibility classes and treatment conditions.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4525169&req=5

f4: Specific gene expression signatures define susceptibility to bacterial enteric infection.(a) Venn diagram showing differentially expressed genes (as revealed by RNA-seq experiments) between four resistant and four susceptible DGRP lines, in the unchallenged condition and 4 h post Pseudomonas entomophila infection (q-value<0.2, two-fold change). Genes in red and green have higher levels in susceptible and resistant lines respectively. The number of genes (black) indicated in the intersections represents the total number of non-differentially expressed genes. (b) Principal component analysis (PCA) on the top 2,000 varying genes between the 16 samples reveals that resistant lines cluster separately from susceptible lines, before (UC) and post-P. entomophila infection. PC1 separates samples based on treatment whereas PC2 separates them based on susceptibility class. (c) Modulated modularity clustering analysis on the top 2,000 varying genes identifies 24 correlated transcriptional modules (n≥15 genes). Each coloured point represents the spearman correlation (rs) between two genes. (d) A selection of functional categories identified by gene ontology (GO) analysis of genes belonging to the different modules identified in c (excluding the largest module with n=523, Supplementary Data 3). For the GO analysis, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID). (e) PCA using the expression levels of genes within each of the 24 modules identifies module #96 as the only module for which the lines are clearly separated on the first principal component according to treatment and susceptibility. (f) Heat map of gene expression levels in module #96 reveals important differences across susceptibility classes and treatment conditions.
Mentions: Variability in survival and physiology among DGRP lines could in part be explained by system-specific transcriptional differences. We therefore performed RNA-seq on 16 gut samples comprising the same four susceptible and four resistant lines as introduced above in the unchallenged condition and 4 h after P. entomophila infection (Supplementary Fig. 7a). Genes (1287) were differentially expressed 4 h post infection compared with the unchallenged condition when all eight lines were treated as replicates (false discovery rate (FDR) adjusted P-value<0.05 and two-fold change, Supplementary Data 1). This set of genes overlaps with what we have previously shown when characterizing the gut transcriptional response to P. entomophila infection, even though that analysis was carried out using microarrays and on a different genetic background (OregonR)31. However, when we looked for differences in gene expression between the four resistant and four susceptible lines by pooling the samples of each susceptibility class, very few genes exhibited significant differential gene expression. Specifically, the expression of only 5 and 34 genes were changed in the unchallenged and challenged guts, respectively, when comparing phenotypic classes (Fig. 4a; Supplementary Data 2). This may reflect reduced statistical power given the large number of genes that are compared. In addition, it is possible that small but systematic differences in gene expression collectively differentiate resistant from susceptible profiles. We therefore performed principal component analysis (PCA) on 2000 genes with the highest expression variance in the 16 transcriptomes. Since infection status has a large impact on the transcriptome, expression profiles derived from infected samples were separated from those of unchallenged samples on the first principal component (PC), which explains 53% of the variance (Fig. 4b). Strikingly, even before infection, profiles of resistant lines were separated from those of susceptible lines based on the second PC, which explains 7.3% of the variance (Fig. 4b). This separation implies that the basal intestinal transcriptional state of resistant lines is distinct from that of susceptible lines, which may either define or reflect a molecular pre-disposition to enteric infection susceptibility. To dissect the molecular signatures that underlie this transcriptional stratification of the two phenotypic classes, we performed modulated modularity clustering37 on the same 2000 genes. We identified 24 transcriptional modules including >15 correlated genes (Fig. 4c; Supplementary Data 3). On the basis of Gene Ontology analysis and manual annotation38, we assigned the genes within the modules to six functional groups (Fig. 4d). To identify those modules whose gene levels clearly separate the lines according to treatment and phenotypic class, we systematically performed PCA on each module by taking the expression levels of its genes (Fig. 4e). We found that in module #96, samples are clearly separated on the first PC, even though the probability for such a separation to spuriously occur is <3 in 10,000 (Fig. 4e; Supplementary Fig. 7b,c). This module contains 20 genes, of which nine are related to stress response and most notably to ROS metabolism (Fig. 4e,f) and collectively explains 29% of the observed phenotypic variation (Supplementary Table 5). Other modules such as #102 (16 genes) also separated the samples on the first two PCs (Supplementary Fig. 8). Interestingly, module #102 likewise contains several ROS-related genes such as Cyp6a9 and Thioredoxin-2 (Trx-2)39. ROS are essential signalling molecules and immune effectors that are induced by the infected gut to neutralize pathogens13 and promote intestinal renewal14. However, a high ROS load can also cause inhibition of protein translation and consequently severe intestinal damage31, necessitating a finely tuned regulation of ROS production and metabolism40.

Bottom Line: Gut immunocompetence involves immune, stress and regenerative processes.Using genome-wide association analysis, we identify several novel immune modulators.This genetic and molecular variation is physiologically manifested in lower ROS activity, lower susceptibility to ROS-inducing agent, faster pathogen clearance and higher stem cell activity in resistant versus susceptible lines.

View Article: PubMed Central - PubMed

Affiliation: 1] Global Health Institute, School of Life Sciences, Station 19, EPFL, 1015 Lausanne, Switzerland [2] Institute of Bioengineering, School of Life Sciences, Station 19, EPFL, 1015 Lausanne, Switzerland.

ABSTRACT
Gut immunocompetence involves immune, stress and regenerative processes. To investigate the determinants underlying inter-individual variation in gut immunocompetence, we perform enteric infection of 140 Drosophila lines with the entomopathogenic bacterium Pseudomonas entomophila and observe extensive variation in survival. Using genome-wide association analysis, we identify several novel immune modulators. Transcriptional profiling further shows that the intestinal molecular state differs between resistant and susceptible lines, already before infection, with one transcriptional module involving genes linked to reactive oxygen species (ROS) metabolism contributing to this difference. This genetic and molecular variation is physiologically manifested in lower ROS activity, lower susceptibility to ROS-inducing agent, faster pathogen clearance and higher stem cell activity in resistant versus susceptible lines. This study provides novel insights into the determinants underlying population-level variability in gut immunocompetence, revealing how relatively minor, but systematic genetic and transcriptional variation can mediate overt physiological differences that determine enteric infection susceptibility.

No MeSH data available.


Related in: MedlinePlus