Background: DNA methylation (DNAm) predicts disease progression, frailty, and mortality in people with HIV (PWH). No data on this topic are available in PWH with 4-class drug resistance (4DR), characterized by a high burden of disease. Our aim was to investigate DNAm in PWH-4DR, as compared to PWH without drug resistance (no-DR).

Methods: Cross-sectional, propensity-score-matched cohort study on PWH on antiretroviral treatment (ART), with viral load (VL) <200 copies/mL, classified as: i) 4DR (defined as documented resistance to NRTIs, NNRTIs, PIs, and INSTIs) from the PRESTIGIO Registry (n=27); ii) no-DR (n=27). Groups matched by age (±1.5 years) and sex. Genome-wide methylation patterns were determined by reduced representation bisulfite sequencing from peripheral blood mononuclear cells. Differential DNAm evaluated on cytosines covered by ≥10 reads in ≥10 PWH for each group: bases with a minimum difference of 25% in medium methylation and a false discovery rate <0.01 were identified as differentially methylated. The number of significant cytosines per chromosomes (chrs) and genes were normalized according to chr and gene length, respectively. Enrichment analysis was performed for genes with ≥10 differentially methylated cytosines in absolute count and ≥0.005% after normalization for gene length.

Results: Overall, 54 individuals evaluated: characteristics reported in Table 1. Comparing PWH-4DR with PWH-no-DR, there were 27060 differentially methylated cytosines distributed throughout all chrs, with a notable enrichment in chr 19 (Figure 1). The top 10 hypermethylated genes in PWH-4DR included 5 long non-coding RNAs (lncRNAs), 2 pseudogenes, C19orf13, IDH3B (coding for a subunit of isocitrate dehydrogenase, known to be downregulated in striatal cells of murine models of HIV), and MGAT4B (coding for a glycosyltransferase); the top 10 hypomethylated genes included 4 lncRNAs, 2 microRNAs, 2 pseudogenes, BHLHE22 (coding for a transcription factor involved in neural and retinal development, known to be upregulated in macrophages with integrated HIV proviruses), and B3GALT6 (coding for a galactosyltransferase involved in synthesis of glycosaminoglycans; it is a ligand of VISTA, a novel monocytic checkpoint regulator involved in immune tolerance and increased in HIV infection). Enrichment analysis of differentially methylated genes highlighted two pathways: ‘estrogen response early’ and ‘serine/threonine kinases’; the last one was confirmed in the analysis with hypomethylated genes only, suggesting a possible upregulation of the genes in the pathway in PWH-4DR.

Conclusions: Even when HIV replication is controlled by ART, multidrug resistance is associated with a different DNAm, a proxy for different gene expression. DNAm might play a role in the high inflammation and disease burden observed in the fragile population with 4DR.