Article Text
Abstract
Background Multiple desperate results support the host immune response to Epstein-Barr virus (EBV) infection as a causal etiology for lupus with evidence supporting autoantibody generation through a molecular mimicry mechanism originating from EBNA1 (Panel A). If true, then this conclusion would be supported by independent tests. We asked whether the epigenetics of lupus in East Asian (EAS) and European (EU) ancestries would independently support this hypothesis, given their large number of now identified risk loci.
Methods Literature review identified SLE risk loci at p<5x10–8 under quality control. We curated 18,075 ChIP- seq datasets for 1,339 transcription factors and co-factors (TFs) (9,862 datasets) and 81 histone marks (HMs) (8,213 datasets), from 3,531 cellular types, evaluated with modified RELI and MARIO from PMID: 2966164.
Results 330 SLE risk loci with p<5x10–8 were published prior to 2023. KEGG analysis of the 330 loci shows strong associations with IFNγ, IL-12, IL-23, B-cell receptor, & EBV (6.1<OR<16, 10–30<pa<10–8). Of the 330 loci, 255 are found in East Asians (EAS) and 106 in Europeans (EU), with fewer in other ancestries. At p<5x10–8, only 40 loci are shared by EAS & EU. Evaluating ChIP-seq datasets separately in EAS & EU, reveals that 200 (14.9%) of 1,339 tested TFs found in 1,871 (19.0%) of the 9,862 TF ChIP-seq datasets are associated with SLE loci at p<10–6 in both ancestries. The TF dataset associations are highly similar in EAS & EU (r=0.73 (Spearman), p<0.0001) (Panel B). TF DNA binding at SLE loci in EBV-infected B cells is enriched relative to other cell types in both EAS & EU (OR>19, p<10–300). Removing the SLE risk loci separately in EAS & EU bound by the EBV-encoded EBNA2, EBNA3C, and EBNALP TFs reduces the significantly associated human TF datasets at p<10–6 by >95%, as is also found by removing the loci bound by many human-encoded TFs.
Of the associated host DNA binding of TFs at p<10–6, EBNA2, EBNALP, & EBNA3C, all EBV-encoded Latency III expression products, are separately associated (2.1<OR<2.6, p<10–11) with the SLE risk loci in both EAS & EU ancestries (Panel C). At OR>3.2, p≤10–4, human TFs, known to form super-enhancers in EBV-infected, Latency III expressing transformed B cells (LCLs), are enriched compared to uninfected B cells independently at the EAS & EU loci.
Of the 81 HM types, H3K27ac and H3K36me3 are the most closely associated with both the EAS & EU SLE loci (2.43<OR<2.6, 10–56<p<10–24 and 2.55<OR<2.73, 10–56<p<10–34), binding a majority (>66%) of EAS & EU SLE loci (Panel C). The same pattern of significant association for H3K27ac as found in LCLs, is found in other cell types, but the proportion of significant H3K27ac and H3K36me3 LCL datasets is far higher than any other cell type (125<OR<882, 10–14<p<10–10), including compared to B cells that are not EBV-infected (45<OR<248, 10–4<p<10–5). >60% of both EAS & EU SLE loci are bound by H3K27ac or H3K36me3 more frequently in LCLs than in B cells that are not EBV-infected.
H3K27ac and H3K36me3 DNA binding is distorted at heterozygotes for 58 & 16 EAS loci and 44 & 18 EU loci (by MARIO with score>0.4).
Conclusions These results complement previous conclusions that the host response to EBV infection generates autoantibodies in lupus, by showing that histone marks, H3K27ac and H3K36me3, possibly poise chromatin to support the regulatory changes induced by the SLE risk variants in EBV-infected B cells expressing Latency III EBV-encoded TFs (LCLs). These findings are independently present in the genetic mechanisms altering the risk of lupus for both EAS and EU ancestries, therefore, potentially important in the pathogenesis of lupus.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.