Distinct gene expression by expanded clones of quiescent memory CD4+ T cells harboring intact latent HIV-1 proviruses

Abstract

Antiretroviral therapy controls, but does not cure, HIV-1 infection due to a reservoir of rare CD4⁺ T cells harboring latent proviruses. Little is known about the transcriptional program of latent cells. Here, we report a strategy to enrich clones of latent cells carrying intact, replication-competent HIV-1 proviruses from blood based on their expression of unique T cell receptors. Latent cell enrichment enabled single-cell transcriptomic analysis of 1,050 CD4⁺ T cells belonging to expanded clones harboring intact HIV-1 proviruses from 6 different individuals. The analysis reveals that most of these cells are T effector memory cells that are enriched for expression of HLA-DR, HLA-DP, CD74, CCL5, granzymes A and K, cystatin F, LYAR, and DUSP2. We conclude that expanded clones of latent cells carrying intact HIV-1 proviruses persist preferentially in a distinct CD4⁺ T cell population, opening possibilities for eradication.

Keywords: CP: Immunology.

Graphical abstract

Figure 1

Screening for intact latently infected cell enrichment by sorting for CD45RA and TRBC Red outlines indicate the population containing the clone of interest. Maximum-likelihood phylogenetic trees show env gene of the latent clone of interest marked in red. Each dot represents a recovered env sequence from the respective subpopulation of CD4⁺ T cells. The scale bars indicate the number of substitutions per site. Each sort was performed once. See also Figures S2 and S3 and Table S3. (A) CD4⁺ T cells from individual 5125 were magnetically sorted into CD45RA⁺ and CD45RA⁻ populations. Subsequent gDNA extraction, limiting dilution, and env sequencing revealed that the infected clone of interest in was enriched in the CD45RA⁻ memory compartment. (B) CD4⁺ T cells from individual 5125 were stained with anti-TRBC1 antibody and sorted into TRBC1⁺ and TRBC1⁻ populations. The latent clone was enriched in the TRBC1⁺ compartment. (C) CD4⁺ T cells from individual 9247 were stained with anti-TRBC1 antibody and sorted into TRBC1⁺ and TRBC1⁻ populations. The latent clone was enriched in the TRBC1⁻ compartment.

Figure 2

Screening for the TRBV expressed by the latent clone in individuals 5104, 5125, and 9247 Red outlines indicate the population containing the clone of interest. Maximum-likelihood phylogenetic trees of env sequences derived from sorted population, as indicated. Env gene of the clone of interest marked in red. The scale bars indicate the number of substitutions per site. TRBV^A−H indicate the grouping of monoclonal antibodies used in the cocktails (see STAR Methods). (A–C) Flow cytometry plots show TRBV staining using a combination of 24 anti-TRBV antibodies. Percentage of TRBV⁻ cells is indicated for individual 5104 and TRBV⁺ cells for 9247 and 5125. (D) Flow cytometry plots show TRBV staining with one of 2 groups of 12 antibodies for individual 9247. Percentages of TRBV⁺ cells are indicated. The sorted population is indicated. The positive group was further subdivided in (E). (E) Flow cytometry plots show TRBV staining with one of 4 groups of 3 TRBV antibodies from individual 9247. Percentages indicate the fraction of TRBV⁺ cells in each group. The sorted population is indicated. The clone of interest was only found in cells stained with the TRBV^H cocktail indicated in red and was further subdivided in (F). (F) Flow cytometry plots show TRBV staining with 3 different monoclonal antibodies from individual 9247. Percentage indicates the fraction of TRBV4-3⁺ cells. (G) Flow cytometry plots show TRBV staining with antibody cocktails A–H from individual 5125. The clone of interest was only found in the TRBV^G cocktail, which was further subdivided in (H). Percentage indicates the fraction of TRBV⁺ cells for each cocktail. (H) Flow cytometry plots show TRBV staining with the 3 different antibodies in the TRBV^G cocktail from individual 5125. The clone of interest was found in the TRBV2⁺ population. Percentage indicates the fraction of TRBV2⁺ cells.

Figure 3

Combined enrichment based on CD45RA, TRBC, and TRBV expression in individuals 603, 605, B207, 5104, 5125, and 9247 Red outlines indicate the population containing the clone of interest. Maximum-likelihood phylogenetic trees of env sequences derived from sorted population, as indicated. Env gene of the clone of interest marked in red. The scale bars indicate the number of substitutions per site. Each sort was performed once for limiting dilution Env PCR amplification and sequencing and confirmed the previous screening for the respective marker of the latent clone of interest. See also Tables S3 and S4. (A) Enrichment in individual 603 after purification of CD45RA⁻ memory cells followed by staining and sorting for TRBC1⁺ and TRBV19⁺ cells. Flow cytometry plots show TRBC1 and TRBV19 staining. Percentages of CD45RA⁻TRBC1⁺ and CD45RA⁻TRBC1⁺TRBV19⁺ cells are indicated. (B) Enrichment in individual 605 after purification CD45RA⁻ memory cells followed by staining TRBV11-2⁺ cells, which were >90% TRBC1⁺ (Figure S3B). Flow cytometry plot shows TRBV11-2 staining. Percentage of CD45RA⁻TRBV11-2⁺ cells is indicated. (C) Enrichment in individual B207 after purification of CD45RA⁻ memory cells followed by staining and sorting for TRBC1⁺ and TRBV⁻ cells. Flow cytometry plots show TRBC1 and combined 24 TRBV staining. Percentages of CD45RA⁻TRBC1⁺ and CD45RA⁻TRBC1⁺TRBV⁻ cells are indicated. (D) Enrichment in individual 5104 after purification CD45RA⁻ memory cells followed by staining and sorting for TRBC1⁺ and TRBV⁻ cells. Flow cytometry plots show TRBC1 and TRBV staining. Percentages of CD45RA⁻TRBC1⁺ and CD45RA⁻TRBC1⁺TRBV⁻ cells are indicated. (E) Enrichment in individual 5125 after purification CD45RA⁻ memory cells followed by staining and sorting for TRBC1⁺ and TRBV2⁺ cells. Flow cytometry plots show TRBC1 and TRBV2 staining. Percentages of CD45RA⁻TRBC1⁺ and CD45RA⁻TRBC1⁺TRBV2⁺ cells are indicated. (F) Enrichment in individual 9247 after purification of CD45RA⁻ memory cells followed by staining and sorting for TRBC1- and TRBV4-3⁺ cells. Flow cytometry plots show TRBC1 and TRBV4-3 staining. Percentages of CD45RA⁻TRBC1⁻ and CD45RA⁻TRBC1⁻TRBV4-3⁺ cells are indicated.

Figure 4

Identification of latent clone TCR sequences (A) Limiting dilution sorting strategy to identify the TCR expressed by the CD4⁺ T cells harboring the clone of interest. Each sample was enriched based on CD45RA, TRBC, TRBV expression, and 5 cells sorted per well (c/w) into microwell plates. Env sequencing identified wells that contained a cell of the latent clone of interest. TCRs were amplified and sequenced from all Env⁺ and a selection of Env⁻ wells in technical duplicates. Each sort was performed once. (B) Bar graph shows the relevant TCR clonotypes identified in Env⁺ and Env⁻ wells. 603: Env⁺ n = 8 and Env⁻ n = 77; 605: Env⁺ n = 21 and Env⁻ n = 62; 5104: Env⁺ n = 7 and Env⁻ n = 36; 5125: Env⁺ n = 4 and Env⁻ n = 31; 9247: Env⁺ n = 9 and Env⁻ n = 42. See also Table S5. (C) Pie charts show the relative size of TCR clones as slices. The areas indicated in white represent unique TCR sequences. The number on the left above the pie chart is the donor ID for each individual. The number in the center of the pie chart represents the number of cells assayed for each individual. The clone of interest is indicated by a red arrow and pie slice. See also Table S4.

Figure 5

Uniform manifold approximation and projection (UMAP) of 10x gene-expression data Data representing mRNA expression by 109,217 individual cells is shown. The latent clone of interest as well as the next biggest and next smallest clone in size were located in the UMAP by their TCR sequence. Enrichment and 10x Genomics gene expression and TCR sequencing were performed once for individuals 603, 605, and 5125 and twice for individuals B207, 5104, and 9247. See also Tables S6 and S7. (A) UMAPs show the position of the cells expressing the latent clone TCR for each of the 6 individuals as red dots. Underneath, UMAPs show the position of the next biggest (yellow triangles) and the next smallest (blue squares) clones in size to the clone of interest. For individual 605, the latent clone of interest was the biggest clone, and only the next smallest clone is shown. (B) The bar graphs show the fraction of cells in the latent clone (red bars) in each of the 15 UMAP clusters, and the fraction of cells of the next biggest (yellow bars) and the next smallest (blue bars) clones in size to the clone of interest in each of the 15 UMAP clusters (Table S6).

Figure 6

Heatmap of cluster-defining, differentially expressed genes (A) Heatmap shows up to 10 of the most upregulated genes per UMAP cluster compared with all other clusters. Genes are indicated on the left and clusters above. Yellow shows relatively highly expressed genes and purple relatively downregulated genes. See also Table S8. (B) Heatmap shows unsupervised clustering based on the expression of differentially expressed genes in cytotoxic CD4⁺ T cells (Appay et al., 2002; Cachot et al., 2021; Hashimoto et al., 2019; Juno et al., 2017; Takeuchi and Saito, 2017; Zaunders et al., 2004). Genes are indicated on the left and clusters below. Yellow shows relatively highly expressed genes, and black shows relatively downregulated genes.

Figure 7

Projection of 10x gene-expression data on UMAP of CD4⁺ T cells from multimodal single-cell sequencing Projection of data representing mRNA expression by 109,217 individual cells on a multimodal UMAP of CD4⁺ T cells from HIV-negative individuals (Hao et al., 2021). The latent clone of interest in each individual was identified by its TCR sequence and is represented as red dots. Underneath each UMAP, the bar graph shows the fraction of cells in the latent clone in each T cell subpopulation as indicated by the number above each bar. See also Table S9.