Identification of Mechanisms by Which Genetic Susceptibility Loci Influence Systemic Sclerosis Risk Using Functional Genomics in Primary T Cells and Monocytes

Abstract

Objective: Systemic sclerosis (SSc) is a complex autoimmune disease with a strong genetic component. However, most of the genes associated with the disease are still unknown because associated variants affect mostly noncoding intergenic elements of the genome. We used functional genomics to translate the genetic findings into a better understanding of the disease.

Methods: Promoter capture Hi-C and RNA-sequencing experiments were performed in CD4+ T cells and CD14+ monocytes from 10 SSc patients and 5 healthy controls to link SSc-associated variants with their target genes, followed by differential expression and differential interaction analyses between cell types.

Results: We linked SSc-associated loci to 39 new potential target genes and confirmed 7 previously known SSc-associated genes. We highlight novel causal genes, such as CXCR5, as the most probable candidate gene for the DDX6 locus. Some previously known SSc-associated genes, such as IRF8, STAT4, and CD247, showed cell type-specific interactions. We also identified 15 potential drug targets already in use in other similar immune-mediated diseases that could be repurposed for SSc treatment. Furthermore, we observed that interactions were directly correlated with the expression of important genes implicated in cell type-specific pathways and found evidence that chromatin conformation is associated with genotype.

Conclusion: Our study revealed potential causal genes for SSc-associated loci, some of them acting in a cell type-specific manner, suggesting novel biologic mechanisms that might mediate SSc pathogenesis.

Figure 1

Promoter capture Hi‐C (pCHi‐C) interactions and gene expression in the rs11117420 (IRF8) genome‐wide association study (GWAS) locus. A, Genomic coordinates (GRCh38) are shown at the top of the panel. The tracks include NCBI RefSeq genes, systemic sclerosis (SSc)–associated GWAS single‐nucleotide polymorphisms (SNPs) from López‐Isac et al (ref. 9) and those in high linkage disequilibrium (LD) (r² > 0.8), transcription activation domains (TADs) (shown as bars), SNPs overlapping promoter interacting regions and enhancer regions, enhancer regions as defined by ChromHMM software, H3K27ac signal, and pCHi‐C significant interactions (CHiCAGO score >5) (shown as arcs) in CD4+ T cells (blue) and CD14+ monocytes (red). The red highlighted region includes the block of all the SSc‐associated SNPs in LD. B, Box plot of IRF8 expression level in CD4+ T cells and CD14+ monocytes in count per million (CPM). Each box represents the 25th to 75th percentiles. Lines inside the boxes represent the median. Lines outside the boxes represent the 10th and 90th percentiles. The dot represents an outlier. C, Chicdiff software bait profiles for IRF8. The plot shows the raw read counts versus linear distance from the bait fragment as mirror images for CD4+ T cells (top) and CD14+ monocytes (bottom). Other‐end interacting fragments are pooled and color‐coded by their weighted adjusted P value. Significant differentially interacting regions detected by Chicdiff overlapping SSc‐associated GWAS SNPs and enhancer regions are depicted as red blocks. Chr = chromosome.

Figure 2

Promoter capture Hi‐C (pCHi‐C) interactions and gene expression in the rs11117420 (STAT4) GWAS locus. A, Genomic coordinates (GRCh38) are shown at the top of the panel. The tracks include NCBI RefSeq genes, SSc‐associated GWAS SNPs from López‐Isac et al (ref. 9) and those in high LD (r² > 0.8), TADs (shown as bars), SNPs overlapping promoter interacting regions and enhancer regions, enhancer regions as defined by ChromHMM, H3K27ac signal, and pCHi‐C significant interactions (CHiCAGO score >5) (shown as arcs) in CD4+ T cells (blue) and CD14+ monocytes (red). The red highlighted region includes the block of all the SSc‐associated SNPs in LD. B, Box plots of STAT4 (left) and NABP1 (right) expression levels in CD4+ T cells and CD14+ monocytes in count per million (CPM). Each box represents the 25th to 75th percentiles. Lines inside the boxes represent the median. Lines outside the boxes represent the 10th and 90th percentiles. The dot represents an outlier. C, Chicdiff bait profiles for STAT4 (left) and NABP1 (right). Plots show the raw read counts versus linear distance from the bait fragment as mirror images for CD4+ T cells (top) and CD14+ monocytes (bottom). Other‐end interacting fragments are pooled and color‐coded by their weighted adjusted P value. Significant differentially interacting regions detected by Chicdiff overlapping SSc‐associated GWAS SNPs and enhancer regions are depicted as red blocks. See Figure 1 for other definitions.

Figure 3

Promoter capture Hi‐C (pCHi‐C) interactions and gene expression in the rs2056626 (CD247) GWAS locus. A, Genomic coordinates (GRCh38) are shown at the top of the panel. The tracks include NCBI RefSeq genes, SSc‐associated GWAS SNPs from López‐Isac et al (ref. 9) and those in high LD (r² > 0.8), TADs (shown as bars), SNPs overlapping promoter interacting regions and enhancer regions, H3K27ac signal, enhancer regions as defined by ChromHMM, and pCHi‐C significant interactions (CHiCAGO score >5) (shown as arcs) in CD4+ T cells (blue) and CD14+ monocytes (red). The red highlighted region includes the block of all the SSc‐associated SNPs in LD. B, Box plots of CD247 (left) and CREG1 (right) expression levels in CD4+ T cells and CD14+ monocytes in count per million (CPM). Each box represents the 25th to 75th percentiles. Lines inside the boxes represent the median. Lines outside the boxes represent the 10th and 90th percentiles. The dot represents an outlier. C, Chicdiff bait profiles for CD247 (left) and CREG1 (right). Plots show the raw read counts versus linear distance from the bait fragment as mirror images for CD4+ T cells (top) and CD14+ monocytes (bottom). Other‐end interacting fragments are pooled and color‐coded by their weighted adjusted P value. Significant differentially interacting regions detected by Chicdiff overlapping SSc‐associated GWAS SNPs and enhancer regions are depicted as red blocks. See Figure 1 for other definitions.

Figure 4

Promoter capture Hi‐C (pCHi‐C) interactions and gene expression in the rs11217020 (DDX6) GWAS locus. A, Genomic coordinates (GRCh38) are shown at the top of the panel. The tracks include NCBI RefSeq genes, SSc‐associated GWAS SNPs from López‐Isac et al (ref. 9) and those in high LD (r² > 0.8), TADs (shown as bars), SNPs overlapping promoter interacting regions and enhancer regions, enhancer regions as defined by ChromHMM, H3K27ac signal, and pCHi‐C significant interactions (CHiCAGO score >5) (shown as arcs) in CD4+ T cells (blue) and CD14+ monocytes (red). The red highlighted region includes the block of all the SSc‐associated SNPs in LD. B, Box plots of CXCR5, DDX6, ARCN1, and IFT46 expression levels in CD4+ T cells and CD14+ monocytes in count per million (CPM). Each box represents the 25th to 75th percentiles. Lines inside the boxes represent the median. Lines outside the boxes represent the 10th and 90th percentiles. Dots represents outliers. C, Chicdiff bait profiles for CXCR5, DDX6, IFT46/ARCN1 (shared capture bait), and UPK2. Plots show the raw read counts versus linear distance from the bait fragment as mirror images for CD4+ T cells (top) and CD14+ monocytes (bottom). Other‐end interacting fragments are pooled and color‐coded by their weighted adjusted P value. Significant differentially interacting regions detected by Chicdiff overlapping SSc‐associated GWAS SNPs and enhancer regions are depicted as red blocks. See Figure 1 for other definitions.