Figure 1. Transcriptional modulation of glycolytic enzymes in productively or latently HIV‐1‐infected cells.

• A–C
  Primary CD4⁺ T cells were activated with α‐CD3‐CD28 beads and infected with HIV‐1_pNL4‐3 or mock‐infected. Cells were cultured for 2 weeks post‐infection (p.i.) to model different infection stages (days 3–9 p.i., i.e., productive infection; day 14 p.i., i.e., latent infection) and subjected to microarray (A, B; n = 2) or RNA‐Seq (C; n = 3) analysis. (A) Gene set enrichment analysis (GSEA) of the expression of the glycolytic pathway (HUMAN‐GLYCOLYSIS) in mock‐infected or HIV‐1‐infected cells. (B,C) Heatmaps of the relative expression of glycolytic enzymes upon HIV‐1 infection. Data are expressed as Log₂ fold change in HIV‐1‐infected vs mock‐infected cells. For microarray data (B), expression values of infected and mock‐infected cells at different time points were pooled. For RNA‐Seq data (C), expression values in infected cells were normalized using the corresponding time point in mock‐infected cells. Adjusted P‐values (q values) to account for multiple testing were calculated by Significance Analysis of Microarrays [SAM (Tusher et al, 2001)] and Deseq2 for RNA‐Seq data (Love et al, 2014).
• D–F
  scRNA‐Seq of the expression of the entire glycolytic pathway or of glucose phosphate isomerase only (GPI) in primary CD4⁺ T cells infected in vitro (D,E) or CD4⁺ T cells of PLWH (F). In panels (D, E), cells were infected with VSVG‐HIV‐1‐GFP and sorted for viral expression as detailed in Golumbeanu et al (2018). Following latency establishment, cells were left untreated or HIV‐1 expression was reactivated through suberoylanilide hydroxamic acid (SAHA) or α‐CD3‐CD28 engagement. Clusters 1 and 2 were identified by principal component analysis as described in Golumbeanu et al (2018). In panel (F), CD4⁺ T cells were isolated from total blood of PLWH under ART as described in Cohn et al (2018). Viral expression was reactivated by treatment with phytohemagglutinin (PHA), and cells were sorted using antibodies against Env and Gag. Sorted cells were then subjected to scRNA‐Seq analysis. The expression level of the HUMAN‐GLYCOLYSIS pathway in (D) was calculated as the average expression of genes comprising the gene list; expression levels in clusters 1 and 2 were compared using Wilcoxon rank‐sum test. For panels (E, F), significance of GPI differential expression level between clusters (E) or between control and Env⁺ Gag⁺ conditions (F) was assessed by the Wilcoxon rank‐sum test encoded in FindMarkers Seurat R function. **P < 0.01, ***P < 0.001; ***P < 0.0001. For panels (D, E) n = 1 donor, 43 cells (untreated), 90 cells (SAHA), 91 cells (TCR), for panel (F) n = 3 donors, 109 cells (control) and 85 cells (Gag⁺ Env⁺).