Abstract Body

As there is currently no cure for HIV, achieving antiretroviral treatment (ART)-free viral suppression has been made the focus of scientific efforts. In the absence of biomarkers that predict time to viral rebound, the evaluation of new therapeutic strategies requires treatment interruption (TI), which may confer risk. All but a rare subpopulation of people living with HIV experience viral rebound shortly after TI2. However, the drivers of this phenotype have not yet been determined. If remission is determined by aspects of host immunity this may be evident by studying gene expression.

To explore the molecular differences between early and late rebounders, we present a longitudinal analysis of the host transcriptome that explores key genetic signatures associated with time to HIV rebound. We sequenced expressed host mRNA from South African women enrolled in the SPARTAC clinical trial, and who received treatment for up to 48 weeks, starting in primary infection. We studied CD4+ T-cells sampled at treatment interruption (wk 48). We used DESeq2 to quantify and transform our data, Gene Set Enrichment Analysis (GSEA) with the Reactome database and Weighted Gene Correlation Network Analysis (WGCNA) to identify putative genetic signatures associated with clinical outcomes. Genes identified by WGCNA to be correlated with time to rebound were screened using Cox Regression with LASSO to determine a gene signature to predict time to viral rebound.

GSEA showed an IFN-I response pathway enrichment in sustained controllers (SC) (>500 days to rebound) versus non-sustained controllers (non-SC) (<500 days to rebound). WGCNA identified three modules of genes associated with the SC phenotype and one of them was found to be associated with both SC phenotype and time to rebound, as a continuous variable. IFN-I response was found to be enriched in this module. Two of five genes that were found to be correlated with time to rebound, were selected to construct a signature of Risk Score, correlating with smaller likelihood of rebounding earlier. A Kaplan-Meier curve of gene expression-based Risk Score (RS), predicted the likelihood of early and late post-TI rebound.

IFN-I response gene signature is enhanced in people with HIV stopping ART who rebound >500 days. The expression of TRIM25 and ISG15 may predict time to rebound at the point of TI.