606 Cell-specific human endogenous retrovirus expression, host gene expression and SLE phenotypes

Background/purposeHuman endogenous retroviruses (HERVs) and long interspersed nuclear elements (LINEs) make up 5-8% and 21% of the human genome. Their expression may contribute to production of type I interferon and the generation of autoantibodies. The objective of this study was to detectHERVs and LINEs in 4 cell-types in SLE patients and characterize their relationship to host gene expression and SLE phenotypes.MethodsPeripheral blood mononuclear cells were isolated from 120 deeply-phenotyped SLE participants. Cells were sorted utilizing magnetic beads (CD14+ monocytes, B cells, CD4+ T cells, and NK cells) and STEM cell technologies for a total of 480 samples. Libraries were sequenced on a HiSeq4000 PE150. Trimmed fastq files were aligned to GRCh38 release 104 using default settings with STAR to generate alignment files. Alignment files were converted to gene counts using featureCounts. Raw counts from Telescope were normalized using DESeq2 and summed per patient; patients were then separated into tertiles based on the summed counts for HERVs and LINEs. DESeq2 was used to perform differential gene expression analysis using gene counts from featureCounts, comparing the third to the first tertile. Gene set enrichment analysis was performed using genes with adjusted p values < 0.05, ranking genes by log2FoldChange, and running WebGestalt. For clinical outcomes, outliers were identified and dropped per cell type and differential expression analysis was run using raw counts from Telescope with DESeq2 per cell type, adjusting for race, lane, sex, and immunosuppressant use at the time of blood draw.Outcomes studied included disease activity (SLEDAI score), autoantibody production (dsDNA, RNP, Sm), ACR renal criteria and disease severity as defined by clinical clusters previously described in the same SLE participants, (Lanata et al, Nat Commun. Aug 29 2019;10(1):3902).ResultsA total of 26,768 HERVs/LINEs were detected across the 480 samples. These were mostly cell-specific (figure 1). High HERVs/LINEs expression correlated with host gene transcription in a cell specific manner. Significant associations with retroviral load include differentially expressed genes in pathways of: olfactory signaling pathway, regulation of IFNA signaling, and interferon alpha/beta signaling in CD14 cells; DNA repair and host response of HIV factors in CD4 cells; activation of HOX genes and antimicrobial peptides in CD19 cells; and regulation of complement cascade, neutrophil degra