Interferon inducible X-linked gene CXorf21 may contribute to sexual dimorphism in Systemic Lupus Erythematosus

Abstract

Systemic lupus erythematosus (SLE) is an autoimmune disease, characterised by increased expression of type I interferon (IFN)-regulated genes and a striking sex imbalance towards females. Through combined genetic, in silico, in vitro, and ex vivo approaches, we define CXorf21, a gene of hitherto unknown function, which escapes X-chromosome inactivation, as a candidate underlying the Xp21.2 SLE association. We demonstrate that CXorf21 is an IFN-response gene and that the sexual dimorphism in expression is magnified by immunological challenge. Fine-mapping reveals a single haplotype as a potential causal cis-eQTL for CXorf21. We propose that expression is amplified through modification of promoter and 3'-UTR chromatin interactions. Finally, we show that the CXORF21 protein colocalises with TLR7, a pathway implicated in SLE pathogenesis. Our study reveals modulation in gene expression affected by the combination of two hallmarks of SLE: CXorf21 expression increases in a both an IFN-inducible and sex-specific manner.

Fig. 1

Genetic refinement of the Xp21.2 (rs887369) SLE susceptibility locus. a Association plot of the 1 Mb region (X:30,077,846–31,077,845) of SLE-associated region Xp21.2 following genotype imputation to the level of UK10K-1000G Phase III and association testing as described (n = 10,995 individuals of European ancestry). rs887369 is shown as the most significantly associated lead SNP. Genetic association plots were generated using LocusZoom. b Association plot of the 1 Mb region following conditional analysis on lead SNP rs887369. c Haplotype construction and visualisation of the Xp21.2 SLE susceptibility locus conducted in Haploview 4.2. The top panel shows the structure of the three blocks and haplotypes surrounding the lead SNP rs887369 (highlighted in red, block B, SNP #15). Blocks are separated by regions of high recombination as specified by D′ and r². The frequency of each haplotype is denoted. The middle panel presents the colour-coded haplotypes and individual SNPs by their genomic coordinates around CXorf21. The bottom panel shows the LD structure and pair-wise correlation (r²) of SNPs, and length of each block. d Right table: case–control association analysis of each haplotype using Haploview 4.2

Fig. 2

eQTL association analysis of SLE associated risk SNP rs887369 in immune cell types. a Cis-eQTL analysis of rs887369 using male samples of the Geuvadis RNA-Seq expression cohort profiled in LCLs. The MAGEB family of genes and NR0B1 were not expressed in LCLs (RPKM < 1). Allele [C] of rs887369 tags the risk haplotype. The number underneath each box-plot represents the mean of the group and the number underneath the x-axis refers to the number of inviduals in each group. b Cis-eQTL analysis performed for all SNPs in cis (±1 Mb) to rs887369 against CXorf21 expression using the males of the Geuvadis cohort. The coordinate of each SNP is plotted on the x-axis and the –log₁₀(P) value of association on the y-axis; rs887369 is highlighted as the best eQTL. c cis-eQTL analysis of rs887369 against CXorf21 expression in LCLs from the TwinsUK cohort using only females who exhibit non-skewed patterns of X-chromosome inactivation (see methods). d Relative protein abundance of CXORF21 in LCLs from females stratified on genotype at the rs887369 SNP. Relative abundance normalised against beta-actin loading control. Source data are provided in the Source Data file (e) Cis-eQTL analysis of rs887369 using the microarray data from the Fairfax et al.^, and Naranbhai et al. cohorts in primary ex vivo immune cell types (see Methods). The remaining cis-genes did not pass quality control. Box-plots show minimum (Q1–1.5*IQR), 25th percentile (Q1), Median, 75th percentile (Q3) and maximum Q3+1.5*IQR

Fig. 3

Functional prioritisation of causal variants at the Xp21.2 SLE susceptibility locus. The five SNPs carried on the risk haplotype attributed to SLE susceptibility and modulation of CXorf21 gene expression were epigenetically fine-mapped using chromatin data from the Roadmap Epigenomes Project (twelve different marks across 127 cell/tissue types). a The five SNPs localised to significant H3K36me3 modification sites in five immune cell types. The heatmap shows the fold-enrichment of H3K36me3 between cell-types across SNP positions. b Signal tracks of H3K36me3 in primary monocytes (blue) and primary neutrophils (red) from peripheral blood across the CXorf21 susceptibility locus. Only rs887369 localises to the binding site summit of H3K36me3 in these two cell types. c Promoter-capture Hi-C interaction of the rs887369 target locus (chrX :30576528–30582605) with four bait loci across 17 primary immune cell types from healthy human donors (Note that the majority of the samples are pooled from multiple donors making it impossible to deconvoloute the sex and genotypes of the individuals). Interaction #1 is the interaction between the association target region (at the 3′ end of CXorf21) and the CXorf21 promoter region. d Heatmap of strength of interaction (CHiCAGO score) of the four interactions across immune cell types. e Correlation of interaction score for interaction #1 (3′ of CXorf21 and CXorf21 promoter) with Roadmap Epigenomes Project chromatin marks found at the CXorf21 promoter across different immune cell types. Higher interactions are correlated with greater enrichment of active chromatin marks suggesting the interaction to regulate gene expression is cell-type specific

Fig. 4

Expression of CXorf21 in primary ex vivo cells stratified on sex and cellular activation. CXorf21 expression data in resting and stimulated monocytes from healthy individuals of European ancestry derived from the Fairfax et al. studies (see Methods; n = 322 biologically independent samples). Samples were separated based on sex and activiation condition: following stimulation with interferon gamma (IFN-γ) or with lipopolysaccharide (LPS) and harvested after 24 h. For each group, the mean is reported in the corners and the effect size (Cohen's d) is reported along the corresponding three-dimensional regression plane. Plots were constructed using plot3D for R. Source data are provided as a Source Data file

Fig. 5

CXorf21 as an interferon response gene. a Differential gene expression of X chromosome genes in response to IFN-α stimulation (harvested after 6 h) in primary ex vivo B cells from healthy females of European ancestry (in-house data). Genes highlighted in red are significantly differentially expressed (q < 0.01; absolute fold-change >2). b Epigenetic landscape of CXorf21 using ENCODE transcription factor binding data in LCLs (GM12878 cell line). All five transcription factors have genome-wide significant binding sites at the CXorf21 promoter. Heat colour is a function of signal strength (fold-change over input)

Fig. 6

Super resolution microscopy of CXORF21 and TLR7. Structured Illumination Microscopy data showing colocalisation of TLR7 and CXORF21 in ex vivo B cells. Representative results on individual cells are shown in a through D with TLR7 staining in the first column, CXORF21 in the second column, DAPI nuclear staining in column three, and in the fourth column all three stains are merged: TLR7 (magenta), CXORF21 (green) and DAPI (blue). The B cells are under different conditions in the panels: a resting, b resting and IFN-α treated (1000 U/ml), c Ig/CD40 stimulated and d Ig/CD40 stimulated and IFN-α treated ex vivo B cells at 20 h. Maximum intensity projections are shown. Scale bar in white on bottom left hand corner is 2 µm. e Plot showing the correlation co-efficients (ρ) between TLR7 and CXORF21 staining of multiple B cells quantified using the results from Z-stack images from individual cells (represented as open circles). From left to right: unstimulated cells (n = 84), cells stimulated with IFN-α (n = 60), B cells stimulated with Ig/CD40 (n = 32), B cells stimulated with Ig/CD40 and IFN-α (n = 22). The horizontal bar represents the mean correlation co-efficient (μ^ρ) and the bars above and below this denote the standard deviation of the distribution. f Mander’s colocalisation coefficient (M2) between TLR7 and CXORF21 are shown from Z-stack images from single B cells (represented as open circles). From left to right: unstimulated cells (n = 84), cells stimulated with IFN-α (n = 60), B cells stimulated with Ig/CD40 (n = 32), B cells stimulated with Ig/CD40 and IFN-α (n = 22). The horizontal bar represents the mean colocalisation co-efficient (μ^M2) and the bars above and below this denote the standard deviation of the distribution. Source data are provided as a Source Data file

Fig. 7

CXORF21 and the autophagosome. Structured Illumination Microscopy data showing colocalisation of LC3 and CXORF21 in ex vivo B cells. Representative results on individual cells are shown in panels a through d with LC3 staining in the first column, CXORF21 in the second column, DAPI nuclear staining in the third column, and in the fourth column all three stains are merged: LC3 (magenta), CXORF21 (green) and DAPI (blue). In panel a B cells were Ig/CD40 stimulated, b Ig/CD40 stimulated and bafilomycin-treated, c Ig and TLR7/8 stimulated and d Ig and TLR7/8 stimulated and bafilomycin-treated after 27 h. Maximum intensity projections are shown. Scale bar in white on bottom left hand corner is 2 µm. e Plot showing the correlation co-efficients (ρ) between LC3 and CXORF21 staining quantified using the results from Z-stack images, individual B cells are represented as open circles. From left to right: Ig/CD40 stimulated cells (n = 17), Ig/CD40 stimulated cells in the presence of 10 nM bafilomycin (n = 22), B cells stimulated with Ig and resiquimod (n = 21), B cells stimulated with Ig and resiquimod in the presence of 10 nM bafilomycin (n = 32). The horizontal bar represents the mean correlation co-efficient (μ^ρ) and the bars above and below this horizontal bar denote the standard deviation of the distribution. f Mander’s colocalisation coefficient (M2) between LC3 and CXORF21 are shown from Z-stack images, individual B cells are represented as open circles. From left to right: Ig/CD40 stimulated cells (n = 17), Ig/CD40 stimulated cells in the presence of 10 nM bafilomycin (n = 23), B cells stimulated with Ig and resiquimod (n = 21), B cells stimulated with Ig and resiquimod in the presence of 10 nM bafilomycin (n = 32). The horizontal bar represents the mean colocalisation co-efficient (μ^M2) and the bars above and below this denote the standard deviation of the distribution. g Western blot analysis of protein extract from starved LCL, in the left-hand blot CXORF21 is quantified in the absence of bafilomycin and after 10 nM and 100 nM treatment. The amount of CXORF21 was quantified by densitometry and the relative abundance shown against a beta actin control, using the unstimulated conditions as a reference point. In the right-hand blot sequestosome 1 (p62) is quantified in the absence of bafilomycin and after 10 nM and 100 nM treatment. The amount of Sequestosome-1 was quantified by densitometry and the relative abundance shown against a beta actin control, using the unstimulated conditions as a reference point. Source data are provided as a Source Data file

Fig. 8

Summary of factors influencing expression of CXorf21 at RNA and protein level. We summarise five factors increasing the cellular abundance of CXorf21 either at RNA level or protein level across a range of immune cell types. These are: (1) genetic variation at SLE susceptibility haplotype - tagged by SNP rs887369 - where the risk haplotype [C] may drive up-regulation by modulation of chromatin interaction and/or modification of H3K36me3 state; (2) female sex, in which escape from X-inactivation results in an increased amount of transcript and protein in females; (3) X chromosome aneuploidy; (4) type I and type II interferons, and LPS, increase the expression of CXorf21 in ex vivo B cells and monocytes; (5) ancestry – potentially linked to the minor allele frequency of rs887369 – in which higher levels of CXorf21 transcript is observed in LCLs derived from donors with European ancestry. We hypothesise that elevation of CXORF21 is a risk factor for developing SLE and that this is may be mediated through it’s role in the endosomal pathway. Figure generated by C.A.O