Circulating integrin α₄β₇⁺ CD4 T cells are enriched for proliferative transcriptional programs in HIV infection

Abstract

HIV preferentially infects α₄β₇⁺ CD4 T cells, forming latent reservoirs that contribute to HIV persistence during antiretroviral therapy. However, the properties of α₄β₇⁺ CD4 T cells in blood and mucosal compartments remain understudied. Employing two distinct models of HIV infection, HIV-infected humans and simian–human immunodeficiency virus (SHIV)-infected rhesus macaques, we show that α₄β₇⁺ CD4 T cells in blood are enriched for genes regulating cell cycle progression and cellular metabolism. Unlike their circulating counterparts, rectal α₄β₇⁺ CD4 T cells exhibited a core tissue-residency gene expression program. These features were conserved across primate species, indicating that the environment influences memory T-cell transcriptional networks. Our findings provide an important molecular foundation for understanding the role of α₄β₇ in HIV infection.

Abbreviations

SIV, simian immunodeficiency virus

SHIV, simian–human immunodeficiency virus

Trm, tissue-resident memory cells

FACS, fluorescence-activated cell sorting

Memory CD4 T cells facilitate HIV establishment following transmission and contribute to HIV persistence during suppressive antiretroviral therapy (ART). Within the memory pool, cells expressing the surface integrin heterodimer, α₄β_7, are of particular interest for several reasons. First, in vitro studies show preferential infection of α₄β₇^hi CD4 T cells due in part to high affinity of α₄β₇ to a conserved epitope on gp120 Env in some HIV strains [[1, 2]]. In addition, cellular activation by α₄β₇ signaling together with gp120-mediated induction of lymphocyte function-induced antigen 1 is posited to favor HIV replication and potentiate cell-to-cell spread of HIV [[3, 4]]. Moreover, α₄β₇-dependent migration of CD4 T cells to the gastrointestinal (GI) tract, the major site of HIV replication, precipitates systemic HIV dissemination from genital and rectal portals of entry. Several studies demonstrate that blockade of α₄β₇ interaction with its ligand mucosal vascular addressin cell adhesion molecule 1 attenuates mucosal simian immunodeficiency virus (SIV) transmission in nonhuman primates [[5, 6]], but the clinical efficacy of α₄β₇ blockade as an approach to eliminate latent HIV reservoirs has been controversial [[7, 8]].

The association between α₄β₇ and HIV infection has also been extended to human studies. In women enrolled in the CAPRISA 004/002 cohorts, pre-infection frequencies of circulating integrin β₇^hi memory CD4 T cells were associated with CD4 decline following HIV infection [[9]]. The same study found that frequencies of α₄β₇^hi memory CD4 T cells in blood were associated with α₄β₇^hi CD4 T cells in the cervix in female sex workers in Nairobi and Kenya. A study in men who have sex with men found that blood α₄β₇ CD4 cells correlated with higher percentage of proliferating α₄β₇ CD4 T cells in rectal tissue [[10]]. Collectively, these data indicate that α₄β₇^hi CD4 T cells in blood may be representative of counterparts seeding the mucosal compartment.

In an acute HIV-1 infection cohort, β7^hi memory CD4 T cells in blood were enriched for integrated and total HIV-1 DNA in Fiebig stages II/III and during chronic infection [[11]]. A recent study in six HIV-infected patients on ART showed that α₄β₇ blockade decreased frequency of activated CD38⁺ CD4 T cells and led to attrition of lymphoid aggregates, key sanctuaries for HIV persistence, in the terminal ileum [[8]]. However, α₄β₇ blockade during analytical treatment interruption in a phase I clinical trial of twenty HIV-infected patients failed to induce viral suppression indicating that targeting α₄β₇ alone may be insufficient to eliminate HIV reservoirs [[7]]. Collectively, the data suggest that intrinsic molecular and cellular pathways within α₄β₇^hi CD4 T cells together with their ability to recirculate, traffic to secondary lymphoid organs, and access mucosal portals influence HIV infection and associated immune activation.

Here, we focused on delineating the molecular characteristics of α₄β₇⁺ CD4 T cells and determining how transcriptional profiles of circulating α₄β₇⁺ CD4 T cells corresponded to that within the lower GI tract during HIV infection. Our approach revealed novel insights into the immunobiology of α₄β₇⁺ CD4 T cells and the data support three main conclusions; first, that α₄β₇⁺ CD4 T cells in peripheral blood express a predominantly effector T-cell signature with enrichment of genes regulating cell cycle progression and cellular metabolism. Second, induction of α₄β₇ occurs following TCR stimulation during the effector CD4 T-cell response. Third, that α₄β₇⁺ CD4 T cells within the colorectal mucosa are transcriptionally distinct from α₄β₇⁺ CD4 T cells in blood and exhibit a core tissue-resident gene expression signature denoting that environment impacts differentiation program of α₄β₇-expressing cells. These data provide an important framework to understand the role of α₄β₇ in HIV and disentangle the relative importance of α₄β₇ CD4 T cells in HIV transmission versus functional cure.

Materials and methods

Human subjects

All study procedures were approved by The Institutional Review Board at Emory University, and experiments were performed in accordance with relevant guidelines and regulations. Participants were HIV⁺ men who have sex with men aged 24–32 years from the Atlanta community (Table 1). All biopsies were collected ˜ 3–10 cm from the anal verge via rigid sigmoidoscopy with no prior bowel preparation. An informed consent was obtained from all participants, according to IRB guidelines. Single-cell suspensions from peripheral blood and rectal biopsies were obtained as previously described [[12]].

Table. Table. 1. Clinical characteristics of human donors and rhesus macaques for RNA sequencing. TAF, tenofovir alafenamide; FTC, emtricitabine; DTG, dolutegravir; EVG, elvitegravir; TDF, tenofovir disoproxil fumarate; cobicistat; RM, rhesus macaque. Rhesus macaques were sampled at 32 weeks post-infection with the tier 2 clade C SHIV1157ipd3N4 virus and were not on ART. Plasma viral load is expressed as viral RNA copies/ml

ID	Sex	Age	Plasma Viral load (time of Bx)	Time on ART	CD4 counts/	ART Regimen
Human Donor #1	M	31	ND	On and off > 5 years (mostly on)	514 cells/μL/22%	TAF/FTC/DTG
Human Donor #2	M	32	ND	7 years	528 cells/μL/26%	TAF/FTC/EVG/c
Human Donor #3	M	24	ND 4 mo prior to bx; VL= 3700 1 mo post bx	9 mos	361 cells/μL/25%	TDF/FTC/EVG/c

ID	Sex	Plasma Viral load (at time of Bx)	Lymphocyte counts (μL blood)
RM#1	M	4007	5280
RM#2	M	878	2800
RM#3	M	1690	6030
RM#4	M	1380	1810

Rhesus macaques

Rhesus macaques for the RNA sequencing arm of the study were housed at the Yerkes National Primate Research Center (YNPRC) in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC). Animals at the YNPRC were infected by the intrarectal route with the tier 2 clade C SHIV1157ipdN4 virus (1 : 400 dilution of stock containing 9.8 × 10^6 TCID50/mL), and viral loads were determined using quantitative real-time PCR, as described [[13]]. (Table 1). Animals were sampled at week 32 post-infection (during viral set point) for RNA sequencing analysis of T cells from blood and rectal biopsies. Studies in immunized monkeys, SARS-CoV-2-infected monkeys, and SHIV C.CH505-infected monkeys were performed using rhesus macaques housed at California National Primate Research Center (CNPRC) in accordance with AAALAC guidelines (Table S1). For the immunization arm, we assessed peripheral blood mononuclear cells (PBMCs) at days 0, 7, and 14 following a SHIV C.1086 protein immunization (week 30) after a SHIV C.1086 DNA prime (weeks 0, 8, 16), as described previously [[14]]. Splenocytes were obtained from animals infected with 2 × 10 ^ 6 plaque-forming units of SARS-CoV-2 (SARS-CoV-2/human/USA/CA-CZB; GenBank accession number: MT394528), at day 14 post-infection [[15]]. Animals infected with SHIV.C.CH505 via the intravaginal route were euthanized 3 weeks post-infection, and splenocytes were assessed for expression of α₄β₇ [[14]]. Single-cell suspensions from peripheral blood and tissues were obtained as previously described [[14]].

RNA sequencing and bioinformatics

Sorted cells were preserved in RLT buffer with BME, flash-frozen, and subsequently processed by the Yerkes NHP Genomics Core Laboratory. RNA integrity and quality were assessed by microcapillary electrophoresis on an Agilent Bioanalyzer. Libraries from samples meeting quality control criterion were sequenced using 3'-Tag-RNA-Seq library prep protocol at Emory University using the Illumina HiSeq 4000 platform (Table S2, S3). The data are deposited to Gene Expression Omnibus database with the accession number GSE165213. Bioinformatics analyses including gene count normalization and differential expression analysis, PCA plots, and heatmaps were created using R (version 4.0.0). The following R packages were used: ComplexHeatmap 2.5.5 for heatmaps, stats/prcomp 3.0.0 and ggplot2 3.3.2 for PCA plots, EdgeR 3.30.3 for differential expression analysis, and Enhanced Volcano 1.8.0 for volcano plots. All packages were run with default settings unless otherwise noted.

All gene counts were imported into the R/Bioconductor package edgeR, and TMM normalization size factors were calculated to adjust for samples for differences in library size. Low expressed genes were excluded from further analysis [[16]]. Compared to reference genes, 25 977 genes were detected, and filtering of lowly expressed genes left 12, 522 genes before normalization. PCA was performed on log2-transformed TMM-normalized gene counts. EdgeR test was used to estimate differentially expressed genes (DEGs).

Cell proliferation assay

For cell trace violet assays (CTV), cryopreserved PBMCs from unvaccinated/ uninfected animals housed at the CNRPC were labeled (Thermo Fisher Scientific) according to the manufacturer’s instructions. A total of 1 × 10⁶ CTV-labeled cells were stimulated in media containing CD3/CD28 and interleukin (IL)-2 (nonhuman primate CD3/CD28 T-cell activation/expansion kit; Miltenyi Biotec). Cells were harvested every 24 h for up to 5 days and surface stained to identify coexpression of integrins with dye dilution. Stained cells were acquired immediately on the BD Fortessa (BD Biosciences). Antibodies used for phenotyping and CTV staining are listed in Table S1.

Results

RNA sequencing of α₄β_7-expressing CD4 T cells in human and rhesus blood

To identify major transcriptional features of α₄β₇⁺ CD4 T cells during HIV infection, we sampled three HIV-infected donors (24–31 years, Table 1, Fig. 1A) and four SHIV-infected rhesus macaques (32 weeks post-intrarectal infection with clade C SHIV1157ipdN4, Table 1, Supporting information S1A–B). Paired blood and rectal pinch biopsies were collected, processed to obtain single-cell suspensions, stained fresh, and FACS sorted into α₄β₇⁻ and α₄β₇⁺ subsets (approximately 2500 CD4⁺ and CD8⁺ T cells) for whole transcriptome profiling by RNA sequencing (Fig. 1A). To delineate molecular signatures of α₄β₇⁺ CD4 T cells, we first undertook a comparison of CD45RO⁺ α₄β₇⁻ and α₄β₇⁺ CD4 T-cell subsets to naive CD45RO⁻ CD4 T cells within peripheral blood.

Fig. 1

RNA sequencing of α₄β₇ expressing CD4 T cells in human and rhesus blood. (A) Experimental design and gating strategy for sorting naive, α₄β₇⁻, and α₄β₇⁺ CD4 T-cell subsets in blood (n = 3 HIV+ MSMs on ART). (B) Principal component analysis (PCA) shows distinct clustering of transcriptomes from naive, α₄β₇⁻_, and α₄β₇⁺ subsets from blood of three donors (˜12 000 genes). (C) Scatterplots display log₂ fold change of α₄β₇⁻ and α₄β₇⁺ subsets versus naive CD4 T cells and negative log₁₀-transformed p-values obtained from edgeR test. Horizontal dotted line representing unadjusted P-value with an FDR < 0.05. (D–E) Plots show normalized read counts (NRC) of select differentially expressed genes with adjusted P values.

Principal component analysis (PCA) revealed that the transcriptome of memory α₄β₇⁺ CD4 T cells was distinct from α₄β₇⁻ CD4 T cells (Fig. 1B). Applying the criteria for significance (P < 0.001 threshold for FDR and absolute value of log2 fold change ≥ 1), we identified 823 genes differentially expressed between α₄β₇⁺ and naive CD4 T cells, 518 differentially expressed genes between α₄β₇⁻ and naive CD4 T cells, and 257 genes differentially expressed between α₄β₇⁻ and α₄β₇⁺ CD4 T cells (Fig. 1C).

Consistent, with their differentiation status, both α₄β₇⁻ and α₄β₇⁺ CD4 subsets expressed significantly higher levels of the death receptor Fas/CD95; adhesion molecules CD58 and CD99; and the IL-2Rα chain, CD25 (Supporting information S1C). Both subsets expressed CD28, showed markedly decreased gene expression of the lymph node homing receptor, CCR7, while the interleukin (IL)-7 receptor, CD127, was significantly downregulated in α₄β₇⁺ CD4 T cells suggestive of recent T-cell activation (Fig. 1D). Studies show decreased CD127 expression in HIV-infected patients as a result of IL-7 and HIV Tat-mediated downregulation of CD127 [[17]]. However, we did not observe evidence for CD127 downregulation in α₄β₇⁺ CD8 T cells relative to naive CD8 T cells (data not shown). We observed that expression of the lectin receptors, killer-cell lectin-like receptor G1 (KLRG1), and CD69 was comparable across all populations. Consistent with β₁ integrin-mediated negative regulation of α₄ pairing with β₇, α₄β₇⁺ CD4 T cells expressed lower relative levels of integrin β₁ [[18]].

We next asked whether α₄β₇⁻ versus α₄β₇⁺ CD4 T cells were differentially polarized based on transcript levels of key chemokine receptors (Fig. 1E). Induction of transcripts regulating expression of C-C motif chemokine receptor 4 (CCR4), C-X-C motif chemokine receptor 3 (CXCR3)/CCR5, CCR6, and CXCR5 in both subsets suggested that signaling via α₄β₇ was not a key determinant of T helper (T_h)1, T_h2, T_h17, and T_fh differentiation. The higher relative expression of CXCR3 within α₄β₇⁺ CD4 T cells is in line with the observation that T_h1-polarizing conditions promote α₄β₇ expression [[19]]. Indeed, transcript levels of the T_h1 fate-determining molecule, T-box transcription factor Tbx21(α₄β₇⁺/naive: 1.67-fold; P value = 0.01), and interferon gamma (IFN-γ) (α₄β₇⁺/naive: 3.34-fold, P value = 0.0006) were significantly higher in α₄β₇⁺ CD4 T cells but not α₄β₇⁻ CD4 T cells relative to naive CD4 T cells.

In contrast, expression of the intestinal homing receptor CCR9 was unique to α₄β₇⁺ CD4 T cells while α₄β₇⁻ CD4 T cells exclusively expressed the skin-homing receptor CCR10 supporting a model where signaling via α₄β₇ controls recirculation of T cells to either cutaneous or intestinal tissues, but not both [[20]]. α₄β₇⁺ CD4 T cells also expressed markedly lower levels of the bone marrow homing receptor, CXCR4. The transcriptional repressor regulating effector T-cell differentiation, Blimp1, and the master transcription factor of T regulatory cell development, Foxp3, were expressed by both subsets. Notably, parallel RNA sequencing analysis of the peripheral blood compartment in rhesus macaques revealed a strikingly similar pattern of gene expression indicating strong conservation of these signatures across α₄β₇⁻ versus α₄β₇⁺ CD4 T cells between humans and macaques (Supporting information S1D).

α₄β₇ defines CD4 T cells enriched for cell cycle transcriptional signatures

While most genes showed a similar pattern of gene expression within α₄β₇⁻ and α₄β₇⁺ CD4 T cells (Fig. 2A), gene set enrichment analysis (GSEA) revealed higher expression of genes regulating cell cycle progression and cellular metabolism in α₄β₇⁺ CD4 T cells. The top 5 pathways enriched in α₄β₇⁺ CD4 T cells were related to mitosis, comprising of genes encoding cell cycle-related targets of E2F transcription factors (E2F1-3, 7–8) which control cell division (Fig. 2B). The dual-specificity protein kinase, TTK, transcriptionally induced during cell cycle progression and regulated by IL-2 [[21]] showed a 3.7-log-fold increase in α₄β₇⁺ CD4 T cells. Correspondingly, the cell division control proteins CDC6, CDC7, essential regulators of DNA replication, were also expressed at higher levels. The dual-specificity phosphatases CDC25A and CDC25C which remove inhibitory phosphorylation in the cyclin-dependent kinases (CDK) were highly expressed and CDK2, in turn, was also induced [[22]]. Additionally, the licensing factor for DNA replication—mini-chromosome maintenance proteins involved in the initiation of replication—MCM2-7, and MCM10 was highly expressed by α₄β₇⁺ CD4 T cells [[23]]. The polo-like kinase (PLK1)—an important regulator of M phase in T cells—was further enriched within the α₄β₇⁺ subset [[24]].

Fig. 2

α₄β₇ defines CD4 T cells enriched for cell cycle transcriptional signatures. (A) Scatterplots display log counts per million (CPM) of genes expressed by CD4 α₄β₇⁻ T cells on y-axis versus CD4 α₄β₇⁺ T cells on x-axis. Gray dots represent genes not differentially expressed between subsets. Red dots represent select genes differentially expressed by α₄β₇⁺ T cells. Gene Set Enrichment Analysis (GSEA) shows enrichment of (B) cell cycle genes. In GSEA plot, x-axis shows genes ranked and y-axis shows enrichment scores. Corresponding heat map next to GSEA plot illustrates top 25 genes within each pathway. (C) Expression of α₄β₇ following CD3/CD28 stimulation in the presence of IP-10 or IL-6, as indicated. (D) Histograms show expression of surface CD69, PD-1, and Ki-67(intracellular) on CD95⁺ CD4 T cells stratified based on α₄β₇ expression. (E) Histogram shows α₄β₇ surface expression on T-cell effectors (indicated by loss of cell trace violet (CTV) dye) generated ex vivo. (F) CD3/CD28 stimulated cells analyzed every 24 h for surface CD69, α₄β_7, integrin β₇ expression versus CTV dye loss.

Congruent with increased metabolic demands of activated CD4 T cells, genes regulating oxidative phosphorylation such as NADH dehydrogenases, cytochrome C oxidases, and ATP synthases were also overexpressed by α₄β₇⁺ CD4 T cells (Supporting information S2A). A similar gene expression program in rhesus PBMCs confirmed that α₄β₇⁺ CD4 T cells within peripheral blood were enriched for highly metabolically active cells in cell cycle (Supporting information S2B–C). We found that except for lymphocyte-activation gene 3 (LAG3) which showed a log 2.3-fold increase in the α₄β₇⁺ subset (P = 7.13e20) transcript levels of programmed death (PD)-1, T-cell immunoreceptor with Ig and ITIM domains (TIGIT), and B and T lymphocyte-associated (BTLA) associated with persistent TCR stimulation, were not differentially expressed by α₄β₇⁺ cells (Supporting information S3A).

Based on these transcriptional profiles, we hypothesized that α₄β₇⁺ expression was indicative of recent TCR stimulation. To test our hypothesis, we stimulated PBMCs with anti-CD3/CD28 and interrogated activated cells for α₄β₇ expression. α₄β₇ was induced upon stimulation, but not in response to inflammatory cytokines (Fig. 2C). The majority of both the α₄β₇⁺⁺ and α₄β₇⁺ subsets expressed CD69, PD-1, and Ki-67 (Fig. 2D). Consistent with retinoic acid-dependent induction of α₄β₇ [[25]], retinoic acid metabolites amplified α₄β₇ expression on Ki-67⁺ cells (Supporting information S3B). Inhibiting cell proliferation with cyclosporine, however, did not decrease α₄β₇ expression suggesting that while α₄β₇ induction was a proximal event following TCR stimulation, it was uncoupled from cell proliferation, as previously reported for IL-7-mediated induction of α₄β₇ on naive CD4 T cells [[26]].

We further confirmed our hypothesis using a standard cell trace violet (CTV) dye dilution assay. Phenotypic analysis of in vitro generated effectors at day 5 showed a greater proportion of α₄β₇⁺/ α₄β₇⁺⁺ cells in cell cycle (Fig. 2E). Time course studies at days 2, 3, 4, and 5 showed relatively stable expression of α₄β₇ and the integrin monomer β₇ over successive rounds of cell division with a significantly higher frequency of α₄β₇⁺ cells having undergone cell division relative to α₄β₇⁻ cells (Fig. 2F). Taken together, these data suggest that human memory CD4 T cells expressing α₄β₇ comprise of recently activated effector cells.

Induction of α₄β₇⁺ Ki-67⁺ CD4 T cells in blood following subcutaneous immunization

We sought to confirm the aforementioned observations in vivo by determining whether α₄β₇ is induced on CD4 T-cell effectors. To capture the kinetics of α₄β₇ induction, we analyzed CD4 T-cell response following subcutaneous HIV-1 Env protein immunization in rhesus macaques primed with HIV-1 Env DNA [[14]]. We observed a significant increase in frequencies of Ki-67⁺ α₄β₇⁺ CD4 T cells at day 7, corresponding to the peak effector response indicating that nonmucosal antigen exposure induced CD4 T cells with the potential to seed the GI tract. Within blood, Ki-67⁺ α₄β₇⁺ CD4 cells were predominantly CD28⁺, heterogeneous for expression of CXCR3, and were largely negative for CCR5 and expressed PD-1 and CCR7 (Fig. 3A–B),

Fig. 3

Induction of α₄β₇⁺ Ki-67⁺ CD4 T cells in blood following subcutaneous immunization. (A) Flow plot illustrating gating strategy to identify T-cell effectors expressing α₄β₇ and Ki-67. (B) shows induction of α₄β₇⁺ Ki-67⁺ CD4 T cells following subcutaneous immunization. (C) Gating strategy for identifying CD4 T cell effectors in spleen, scatter plot shows α₄β₇⁺ Ki-67⁺ cells (% CD4 T cells) in spleen following SARS and SHIV infection.

Next, we sought to determine whether α₄β₇ was induced within lymphoid tissue during the effector phase of the immune response. For this purpose, we examined splenocytes from macaques 3 weeks post-SHIV C.CH505 infection. To capture the phenotype of CD4 T-cell effectors within the spleen, we assessed Ki-67⁺ PD-1⁺ CD4 T cells (excluding the PD-1⁺⁺ T_fh cells) for expression of α₄β₇ (Fig. 3C). We furthermore examined a cohort of macaques infected with SARS-CoV-2. The data showed that effector cells seeding the spleen at days 10–14 following SARS-CoV-2 infection also expressed α₄β₇, with higher relative expression in SHIV-infected animals. Thus, the data support the conclusion that α₄β₇ ⁺ CD4 T cells in circulation represent cells recently exposed to antigen.

α₄β₇⁺ CD4 T cells in rectal mucosa express a tissue-resident transcriptional program

We next asked whether the highly activated gene expression profile was retained in α₄β₇⁺ CD4 T cells seeding the lower GI tract. As illustrated in the flow plots, distinct populations CD3⁺ CD8⁻ CD4⁺ cells were identifiable in the rectum with the majority of CD4 T cells expressing CD45RO (RO) (Fig. 4A). Median fluorescence intensity of α₄β₇ was significantly lower within the rectum consistent with downregulation of α₄β₇ after trafficking to the gut (Fig. 4B). Unlike distinct differences between the α₄β₇⁺ versus α₄β₇⁻ subsets in PBMCs, subset-related differences at the level of individual genes in the gut were minimal (Fig. 4C) indicating that anatomic location, not cellular phenotype, was a key driver of CD4 T-cell biology in the rectum with respect to α₄β₇ expression. Evaluation of key cell cycle genes overexpressed by α₄β₇⁺ CD4 T cells in blood showed that these genes were not differentially expressed between gut subsets (Fig. 4D).

Fig. 4

α₄β₇⁺ CD4 T cells in rectal mucosa are not enriched for cell cycle transcriptional signatures. (A) Flow plot illustrates gating strategy for sorting CD4 T cell subsets in rectal biopsies. In (B), histograms depict decreased expression of α₄β₇ on CD4 T cells in rectum relative to blood. (C) PCA plots show distinct clustering of CD4 T cell transcriptomes across blood and rectal compartments. (D) Plots show normalized read counts (NRC) of cell cycle genes across peripheral blood (PBMC) and rectal (RB) compartments.

To elucidate the transcriptional program of α₄β₇⁺ CD4 T cells within the rectal compartment, we compared the pattern of gene expression across compartments. PCA revealed that blood subsets clustered together and were distinct from the rectal compartment in both primate species (Supporting information S4A–B). We first assessed whether genes regulating migration and retention of CD4 T cells within tissues were among the differentially expressed genes and observed that rectal CD4 T cells downregulated integrin β₇ with a corresponding upregulation of integrin α₄E (CD103), granzyme B, and CD69—classical markers of tissue-resident memory cells, a profile shared by rhesus rectal CD4 T cells (Fig. 5A–B).

Fig. 5

α₄β₇⁺ CD4 T cells in rectal mucosa express a tissue-resident transcriptional program. (A) Heatmap presents relative levels of differentially regulated genes across tissue compartments within each CD4 subset. Gene names are color-coded based on representation of T-cell differentiation programs. Low to high gene expression is represented by a change of color from green to red, respectively. The color key scale bar shows z-score values for the heatmap. Each row (gene) is scaled to have mean zero and standard deviation one. (B) Corresponding scatter plots show log₂ fold change of genes representing tissue-resident memory cells in rectum relative to blood in humans and rhesus (C) Violin plots show expression of key genes across tissue compartments in humans with adjusted P values.

Further investigation revealed that rectal CD4 T cells (both α₄β₇⁻ and α₄β₇⁺ subsets) expressed a core Trm gene signature. This was exemplified by expression of the immunoglobulin-superfamily transmembrane protein, cytotoxic and regulatory T-cell molecule (CRTAM), an important negative regulator of T-cell proliferation and positive regulator of IFN-γ and IL-22, also highly expressed by gut CD4 T cells [[27]]. The dual-specificity phosphatase, DUSP6, was also overexpressed in rectal CD4 T cells as were the integrins ITGA1 (CD49a), chemokine receptor CXCR6, inhibitory receptors PDCD1, LAG3, CD101 as reported for human Trm cells [[28]]. Downregulation of KLRG1 and CX3CR1 was further emblematic of the Trm gene signature. The runt-related transcription factor 3 (RUNX3), a transcription factor mediating CD8⁺TRM formation in mice [[29]], was also overexpressed. Most gene signatures were enriched within rhesus rectal CD4 T cells (Fig. 5C). In addition to the Trm signature, we noted that genes regulating the T_fh program were enriched within gut CD4 T cells—B-cell lymphoma 6 (Bcl-6), showed a log 3.6-fold increase (P = 1.28e13) with a corresponding upregulation of signal transducer and activator of transcription (STAT4), CXCR5, inducible T-cell costimulator (ICOS), and CD200. Additionally, IL23R, interferon regulatory factor (IRF)3, IRF4, CCR5-genes corresponding to T_h1/T_h17 fate were induced in gut CD4 T cells. Interestingly, gut resident cells showed a gene expression pattern of type 1 regulatory cells with expression of transcription factors IRF1, IRF8, basic leucine zipper ATF-like transcription factor (BATF), and early growth response 2(EGR2). However, Foxp3 expression was also noted. Collectively, these analyses demonstrate that transcriptional heterogeneity between α₄β₇⁺ CD4 subsets in blood does not persist in the GI tract, and α₄β₇ expression does not delineate transcriptionally distinct CD4 subsets within the rectal compartment.

Discussion

While HIV infects α₄β₇⁺ CD4 T cells forming latent reservoirs that contribute to HIV persistence, the properties of α₄β₇⁺ CD4 T cells in blood and mucosal compartments and the relationship between these subsets remain understudied. Gaining insights into the differentiation programs of these cells in tissue niches relevant to HIV transmission and dissemination will inform HIV vaccine and functional cure approaches. In this study, we provide important insights into the immunobiology of α₄β₇ integrin-expressing CD4 T cells through systematic transcriptional and phenotypic analysis of human and macaque CD4 T-cell subsets in peripheral blood and rectal mucosa. Our results establish that human recirculating α₄β₇⁺ CD4 T cells exhibit an effector T-cell signature which parallels the transcriptional profile of rhesus CD4 T cells. We demonstrate that within the rectal mucosa, α₄β₇⁺ CD4 T cells adopt a tissue-resident transcriptional signature which significantly distinguishes them from their circulating counterparts. The data provide a biological basis for the association between integrin β₇⁺ CD4 T cells in blood and CD4 depletion reported in clinical studies and suggest that Trm depletion may underlie HIV-associated GI dysfunction.

Antigen experienced T cells trafficking through blood mainly comprise of central memory, effector memory, or effector cells [[30]]. Our finding that α₄β₇⁺ circulating CD4 T cells in humans exhibit an effector gene expression program indicates that they represent effectors and effector memory pools trafficking through blood following priming/ reactivation within SLOs. Mouse studies show that α₄β₇ is induced transiently following systemic or subcutaneous antigen exposure [[31]]. The extent to which this activation gene signature in our studies was influenced by HIV and/or gut microbial translocation is an open question. However, our findings from flow cytometry of in vitro activated PBMCs from SIV unexposed macaques demonstrated that α₄β₇ is induced upon TCR stimulation and is sustained over successive cell divisions. In concert with data in macaques following immunization, the data support a model whereby α₄β₇ is induced following TCR stimulation [[19]], imparting effector cells with the ability to migrate to the GI lymphoid tissues and to the lamina propria.

Within nonlymphoid tissues such as the rectal mucosa, antigen-experienced T cells mainly comprise of effectors, effector memory cells, and tissue-resident memory populations [[32]]. Our data show that in striking contrast to circulating α₄β₇⁺ CD4 T cells, CD4 T cells isolated from rectal biopsies in both humans and macaques did not overexpress genes regulating cell cycle progression akin to effectors or effector memory cells but instead demonstrated a core Trm gene signature (CD69+, Gzm B+ LAG3+ PDCD1+ CXCR6+, KLRG1^lo, CX3CR1^lo). This gene expression profile delineates Trms across multiple mucosal and lymphoid tissues in humans and mice [[28, 30]]. Because programs regulating the maintenance of Trms are dictated by the tissue environment, not antigen stimulation, these data suggest that suppressive ART may partially restore HIV-induced mucosal immune dysfunction [[33, 34]].

Intestinal CD4 T cells are seeded as both intraepithelial lymphocytes (IELs) and lamina propria lymphocytes (LPL) but constitute the most predominant T lymphocyte subset within the LP while CD8 T cells are the more prevalent IELs [[35]]. Studies show that the LP harbors diverse CD4 helper subsets and we found that colorectal CD4 T cells exhibited features of all major T helper subsets. We did not, however, identify a T_h2 gene expression program. Studies show that following oral listeria infection in mice, listeria-specific CD4 T cells express α₄β₇ and accumulate in the intestinal LP and epithelium as CD69⁺ cells with production of T_h1 cytokines IFNγ and tumor necrosis factor alpha (TNFα) [[36]]. These data suggest that T_h effector subsets induced following antigen stimulation give rise to respective Trm cells. Whether T_fh cells primed in nonmucosal lymphoid tissue have the capacity to migrate to the LP or the GALT is unclear. We did, however, observe a T_fh-like transcriptome (Bcl-6, CXCR5, ICOS, and CD200) within the rectal mucosa. This was suggestive of follicular cell homing to the GI tract as has been described in models of autoimmunity; however, we did not observe expression of IL-21 transcripts [[37]].

Intriguingly, rectal CD4 T cells also showed a gene expression signature of Tr1 cells with expression of LAG3, CD49b, Egr2, IL-10, IFN-γ raising the possibility of Tr1 seeding within the intestinal mucosa. A recent study in HIV-infected aviremic patients on ART demonstrated increase in CD49b/LAG3+ cells ascribed as Tr1 cells [[38]]. To conclusively capture the diversity of helper subsets within the gut, single-cell analysis or extensive multiparameter flow analysis is needed to identify which T helper subsets contribute to the bonafide Trm pool in the GI tract, and importantly how HIV infection impacts this population. The data clearly indicate that in the gut, α₄β₇ expression within the lower GI tract does not distinguish between CD4 T cells of divergent differentiation programs or functionalities at the transcriptional level suggesting that downregulation of α₄β₇ following ingress into the gut is a stochastic process.

While our data provide novel insights into transcriptional heterogeneity between subsets in blood and striking compartmental differences, our study has several drawbacks. First, we performed our studies in relatively young men, and whether the observed gene expression profiles hold true across the sexes, in people without HIV, and in older adults needs to be determined. Second, longitudinal analysis of α₄β₇-expressing CD4 T subsets prior to infection and over the course of HIV infection and ART initiation would significantly add to our understanding of HIV infection-induced dynamic changes within CD4 subsets and across compartments. Third, analysis of these cells within the female reproductive tract would yield critical insights into target cells relevant for HIV acquisition in the female genital mucosa. Finally, how the biology of α₄β₇⁺ cells is modulated following α₄β₇ blockade is an intriguing question which may have direct implications for understanding strategies for HIV prevention and functional cure. Thus, additional well-controlled studies are needed to effectively determine if defining traits of α₄β₇⁺ CD4 T cells identified here hold true across other settings.

In summary, our study elucidates for the first time the immunobiology of α₄β₇-expressing CD4 T cells at the transcriptional level in HIV-infected humans and SIV-infected macaques. We identify major unifying features of these cells that are conserved across species. These results will serve as a valuable resource to understand the role of α₄β₇ in HIV infection and pathogenesis.

Acknowledgments

We thank the study volunteers for their participation in this research. We thank the Yerkes Genomics Core for processing sorted human and rhesus T-cell subsets. The authors are grateful to Brian Schmidt for depositing files on GEO. We thank George Shaw and Nancy Miller for the SHIV.C.CH505 virus. The authors are grateful to Rachel Rutishauser for scientific discussions. The α₄β₇ reagent used in these studies was provided by the NIH Nonhuman Primate Reagent Resource (R24 OD010976 and NIAID contract HHSN 272201300031C). This work was supported by the Center for AIDS Research at Emory University (P30AI050409). RNA sequencing analysis was supported by a supplement to SSI by the Women’s Interagency HIV Study (WIHS; U01-AI-103408). This work was supported by the NIAID grants K01 OD023034, R03 AI138792, 1R21AI143454 (SSI).

Conflict of interest

The authors have no conflicts of interest to declare.

Author contributions

Conceptualization – ANS, RRA, CFK, IO, SSI; Experimentation/Sampling – YSL, VV, ANS, CFK, SSI; Analysis and Interpretation – YSL, JWR, NNR, ARD, DDT, EW, SSI; Manuscript writing – SSI.

Open Research

Data accessibility

RNA-seq data that support the findings in this study are openly available in NCBI’s Gene Expression Omnibus and are accessible through https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE165213