Abstract Body

HIV eradication is hindered by the existence of latent HIV reservoirs in CD4 T cells. T follicular helper (Tfh) cells are a subset of CD4 T cells present in lymph nodes (LN) which harbor high frequencies of HIV-infected cells. Here, we used a primary cell model of HIV latency and the dual reporter virus, HIV DuoFluo (provided by E. Verdin, UCSF) to investigate latency establishment in CD4 and Tfh cells.

Tonsils from HIV-neg donors (n=8) were obtained as a source of LN CD4 T cells. Purified CD4 were activated with anti-CD3/CD28 beads for 3d prior to infection with HIV DuoFluo, containing GFP under HIV-LTR promoter and mKO2 under EF1α promoter. Live, uninfected (U, GFP-mKO2-), latent (L, GFP-mKO2+), and productive (P, GFP+mKO2+/-) infected cells were analyzed 3d later for surface expression of CD4, CXCR5, and PD1 and were sorted in bulk for gene expression (GE) analysis of 96 genes involved in T cell function and metabolism by Fluidigm Biomark assay. Single-cell sorting of U, L, and P cells (62 cells per group) was performed using cells from one tonsil donor for 3′ whole transcriptome amplification using BD Precise platform. Limma and FDR correction were applied to determine significant differences in GE.

Tfh (PD1+CXCR5+) were more permissive to latent infection compared to Non-Tfh (PD1-CXCR5-) cells by comparing ratios of L: P (Mean: 0.4 vs. 1.0, respectively, p=0.02 student’s t test). Biomark analysis of Tfh cells showed differential GE of 36 genes between P and U and 8 unique genes between P and L cells (ANOVA, p<0.05). 0 genes were exclusively differentially expressed (DEGs) between L and U, though IL21 was downregulated in P and L Tfh compared to U which may affect overall function of Tfh. P had reduced expression of genes involved in transcriptional regulation and metabolism compared to U, and L had higher expression of TNF superfamily members and restriction factor, APOBEC3G compared to P. In single cell analysis, U and L overlapped with 0 DEGs while P showed 202 DEGs compared to U and 155 compared to L (121 overlapped between the 2 comparisons). FCGR2A, the proposed marker of latent cells in blood was not detectable in sorted single cells suggesting differences between blood and LN markers of HIV latency.

Our methods using an established HIV latency model and bulk and single cell transcriptomic analysis in LN cells shed light on biology of latency in a crucial anatomical site for HIV persistence and are valuable tools for investigating cure strategies.