Limits...
Graded gene expression changes determine phenotype severity in mouse models of CRX-associated retinopathies.

Ruzycki PA, Tran NM, Kefalov VJ, Kolesnikov AV, Chen S - Genome Biol. (2015)

Bottom Line: Unlike down-regulated genes, which show a high degree of CRX binding and dynamic epigenetic profiles in normal retinas, the up-regulated cone-enriched genes do not correlate with direct activity of CRX, but instead likely reflect a change in rod cell-fate integrity.Furthermore, these analyses describe the impact of minor gene expression changes on the phenotype, as two mutants showed marginally distinguishable expression patterns but huge phenotypic differences, including distinct mechanisms of retinal degeneration.Our results implicate a threshold effect of gene expression level on photoreceptor function and survival, highlight the importance of CRX in photoreceptor subtype development and maintenance, and provide a molecular basis for phenotype variability in CRX-associated retinopathies.

View Article: PubMed Central - PubMed

Affiliation: Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, Saint Louis, MO, USA.

ABSTRACT

Background: Mutations in the cone-rod-homeobox protein CRX are typically associated with dominant blinding retinopathies with variable age of onset and severity. Five well-characterized mouse models carrying different Crx mutations show a wide range of disease phenotypes. To determine if the phenotype variability correlates with distinct changes in CRX target gene expression, we perform RNA-seq analyses on three of these models and compare the results with published data.

Results: Despite dramatic phenotypic differences between the three models tested, graded expression changes in shared sets of genes are detected. Phenotype severity correlates with the down-regulation of genes encoding key rod and cone phototransduction proteins. Interestingly, in increasingly severe mouse models, the transcription of many rod-enriched genes decreases decrementally, whereas that of cone-enriched genes increases incrementally. Unlike down-regulated genes, which show a high degree of CRX binding and dynamic epigenetic profiles in normal retinas, the up-regulated cone-enriched genes do not correlate with direct activity of CRX, but instead likely reflect a change in rod cell-fate integrity. Furthermore, these analyses describe the impact of minor gene expression changes on the phenotype, as two mutants showed marginally distinguishable expression patterns but huge phenotypic differences, including distinct mechanisms of retinal degeneration.

Conclusions: Our results implicate a threshold effect of gene expression level on photoreceptor function and survival, highlight the importance of CRX in photoreceptor subtype development and maintenance, and provide a molecular basis for phenotype variability in CRX-associated retinopathies.

No MeSH data available.


Related in: MedlinePlus

Hierarchical clustering and epigenetic data reveal groups of similarly regulated genes. a Hierarchical clustering analysis of all genes that showed significantly altered expression (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in at least one mutant genotype relative to age-matched WT expression. Expression levels in the indicated genotypes at the indicated ages are indicated by the blue–red heatmap. Eight groups of genes (indicated by the bars on the right) are clearly defined by the results (see Additional file 11 for designations). b Heatmaps showing the epigenetic landscape near the transcription start site (TSS) of genes in groups 1, 2, 3, and 6. Each row represents ±1 kb from the TSS (at center of panel) of individual genes contained within the groups as noted on the left; rows are ordered within each group by decreasing DNase I hypersensitivity (DHS) at the TSS in the 8-week dataset. Columns 1–2, CRX binding determined by ChIP-seq in adult WT and Nrl−/− retinas, respectively; column 3, NRL ChIP-seq in adult WT retinas; columns 4–6, DHS of WT retinas at three indicated ages; columns 7–8, H3K4me2 ChIP-seq in WT retinas at P1 and P15, columns 9–10, H3K27me3 ChIP-seq in WT retinas at P1 and P15. Quantification is presented in Additional file 13
© Copyright Policy - OpenAccess
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4556057&req=5

Fig5: Hierarchical clustering and epigenetic data reveal groups of similarly regulated genes. a Hierarchical clustering analysis of all genes that showed significantly altered expression (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in at least one mutant genotype relative to age-matched WT expression. Expression levels in the indicated genotypes at the indicated ages are indicated by the blue–red heatmap. Eight groups of genes (indicated by the bars on the right) are clearly defined by the results (see Additional file 11 for designations). b Heatmaps showing the epigenetic landscape near the transcription start site (TSS) of genes in groups 1, 2, 3, and 6. Each row represents ±1 kb from the TSS (at center of panel) of individual genes contained within the groups as noted on the left; rows are ordered within each group by decreasing DNase I hypersensitivity (DHS) at the TSS in the 8-week dataset. Columns 1–2, CRX binding determined by ChIP-seq in adult WT and Nrl−/− retinas, respectively; column 3, NRL ChIP-seq in adult WT retinas; columns 4–6, DHS of WT retinas at three indicated ages; columns 7–8, H3K4me2 ChIP-seq in WT retinas at P1 and P15, columns 9–10, H3K27me3 ChIP-seq in WT retinas at P1 and P15. Quantification is presented in Additional file 13

Mentions: To determine the modality of CRX’s regulation of differentially expressed genes, we further investigated their expression patterns and the epigenetic landscape of their proximal cis-regulatory regions in WT mice. We first used hierarchical cluster analysis on P10 and P21 datasets to find sets of genes that were similarly affected in all mutants. Figure 5a shows a heatmap representing log2 FC relative to age-matched WT samples for any gene that displayed significant change from WT (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in any single genotype; the data are arranged by hierarchical clustering (clustering branches are shown to the left of the heatmap). By visual inspection of clustered data, we further subdivided affected genes into eight groups based on similarity of altered expression patterns. These are designated as groups 1–8 (shown to the right of the heatmap; see Additional files 10 and 11 for lists and order of genes). Further analyses focused on groups 1, 2, 3 and 6, as these represented the largest and most consistent clusters. Visual inspection of biological replicate data also confirmed the consistency of the expression changes in these groups (Additional file 12). Group 1 genes were the most down-regulated genes across all genotypes. Group 2 genes were decreased compared with WT levels, but across the board were less affected than those in group 1. Groups 3 and 6 were composed of genes that were up-regulated in many of the genotypes. Group 6 genes generally were up-regulated to a greater extent.Fig. 5


Graded gene expression changes determine phenotype severity in mouse models of CRX-associated retinopathies.

Ruzycki PA, Tran NM, Kefalov VJ, Kolesnikov AV, Chen S - Genome Biol. (2015)

Hierarchical clustering and epigenetic data reveal groups of similarly regulated genes. a Hierarchical clustering analysis of all genes that showed significantly altered expression (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in at least one mutant genotype relative to age-matched WT expression. Expression levels in the indicated genotypes at the indicated ages are indicated by the blue–red heatmap. Eight groups of genes (indicated by the bars on the right) are clearly defined by the results (see Additional file 11 for designations). b Heatmaps showing the epigenetic landscape near the transcription start site (TSS) of genes in groups 1, 2, 3, and 6. Each row represents ±1 kb from the TSS (at center of panel) of individual genes contained within the groups as noted on the left; rows are ordered within each group by decreasing DNase I hypersensitivity (DHS) at the TSS in the 8-week dataset. Columns 1–2, CRX binding determined by ChIP-seq in adult WT and Nrl−/− retinas, respectively; column 3, NRL ChIP-seq in adult WT retinas; columns 4–6, DHS of WT retinas at three indicated ages; columns 7–8, H3K4me2 ChIP-seq in WT retinas at P1 and P15, columns 9–10, H3K27me3 ChIP-seq in WT retinas at P1 and P15. Quantification is presented in Additional file 13
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4556057&req=5

Fig5: Hierarchical clustering and epigenetic data reveal groups of similarly regulated genes. a Hierarchical clustering analysis of all genes that showed significantly altered expression (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in at least one mutant genotype relative to age-matched WT expression. Expression levels in the indicated genotypes at the indicated ages are indicated by the blue–red heatmap. Eight groups of genes (indicated by the bars on the right) are clearly defined by the results (see Additional file 11 for designations). b Heatmaps showing the epigenetic landscape near the transcription start site (TSS) of genes in groups 1, 2, 3, and 6. Each row represents ±1 kb from the TSS (at center of panel) of individual genes contained within the groups as noted on the left; rows are ordered within each group by decreasing DNase I hypersensitivity (DHS) at the TSS in the 8-week dataset. Columns 1–2, CRX binding determined by ChIP-seq in adult WT and Nrl−/− retinas, respectively; column 3, NRL ChIP-seq in adult WT retinas; columns 4–6, DHS of WT retinas at three indicated ages; columns 7–8, H3K4me2 ChIP-seq in WT retinas at P1 and P15, columns 9–10, H3K27me3 ChIP-seq in WT retinas at P1 and P15. Quantification is presented in Additional file 13
Mentions: To determine the modality of CRX’s regulation of differentially expressed genes, we further investigated their expression patterns and the epigenetic landscape of their proximal cis-regulatory regions in WT mice. We first used hierarchical cluster analysis on P10 and P21 datasets to find sets of genes that were similarly affected in all mutants. Figure 5a shows a heatmap representing log2 FC relative to age-matched WT samples for any gene that displayed significant change from WT (FC ≥ 2 or ≤ −2, FDR ≤ 0.05) in any single genotype; the data are arranged by hierarchical clustering (clustering branches are shown to the left of the heatmap). By visual inspection of clustered data, we further subdivided affected genes into eight groups based on similarity of altered expression patterns. These are designated as groups 1–8 (shown to the right of the heatmap; see Additional files 10 and 11 for lists and order of genes). Further analyses focused on groups 1, 2, 3 and 6, as these represented the largest and most consistent clusters. Visual inspection of biological replicate data also confirmed the consistency of the expression changes in these groups (Additional file 12). Group 1 genes were the most down-regulated genes across all genotypes. Group 2 genes were decreased compared with WT levels, but across the board were less affected than those in group 1. Groups 3 and 6 were composed of genes that were up-regulated in many of the genotypes. Group 6 genes generally were up-regulated to a greater extent.Fig. 5

Bottom Line: Unlike down-regulated genes, which show a high degree of CRX binding and dynamic epigenetic profiles in normal retinas, the up-regulated cone-enriched genes do not correlate with direct activity of CRX, but instead likely reflect a change in rod cell-fate integrity.Furthermore, these analyses describe the impact of minor gene expression changes on the phenotype, as two mutants showed marginally distinguishable expression patterns but huge phenotypic differences, including distinct mechanisms of retinal degeneration.Our results implicate a threshold effect of gene expression level on photoreceptor function and survival, highlight the importance of CRX in photoreceptor subtype development and maintenance, and provide a molecular basis for phenotype variability in CRX-associated retinopathies.

View Article: PubMed Central - PubMed

Affiliation: Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, Saint Louis, MO, USA.

ABSTRACT

Background: Mutations in the cone-rod-homeobox protein CRX are typically associated with dominant blinding retinopathies with variable age of onset and severity. Five well-characterized mouse models carrying different Crx mutations show a wide range of disease phenotypes. To determine if the phenotype variability correlates with distinct changes in CRX target gene expression, we perform RNA-seq analyses on three of these models and compare the results with published data.

Results: Despite dramatic phenotypic differences between the three models tested, graded expression changes in shared sets of genes are detected. Phenotype severity correlates with the down-regulation of genes encoding key rod and cone phototransduction proteins. Interestingly, in increasingly severe mouse models, the transcription of many rod-enriched genes decreases decrementally, whereas that of cone-enriched genes increases incrementally. Unlike down-regulated genes, which show a high degree of CRX binding and dynamic epigenetic profiles in normal retinas, the up-regulated cone-enriched genes do not correlate with direct activity of CRX, but instead likely reflect a change in rod cell-fate integrity. Furthermore, these analyses describe the impact of minor gene expression changes on the phenotype, as two mutants showed marginally distinguishable expression patterns but huge phenotypic differences, including distinct mechanisms of retinal degeneration.

Conclusions: Our results implicate a threshold effect of gene expression level on photoreceptor function and survival, highlight the importance of CRX in photoreceptor subtype development and maintenance, and provide a molecular basis for phenotype variability in CRX-associated retinopathies.

No MeSH data available.


Related in: MedlinePlus