Follicular CD8⁺ T Cells Are Elevated in HIV Infection and Induce PD-L1 on B Cells

CD8⁺ T cells are essential for the control of viral infections (1). Ag-specific CD8⁺ T cells differentiate into effector CTLs, which can detect and kill virus-infected cells (1). However, in chronic HIV infection, CTLs become dysfunctional and exhausted, and the accumulation of defective CD8⁺ T cells is thought to contribute to the systemic inflammation that is the hallmark of chronic HIV infection (2). A novel population of CD8⁺ T cells expressing CXCR5 present in secondary lymphoid tissues has been recently described (3–5). Functionally, CD8⁺CXCR5⁺ T cells seem analogous to CD4⁺ T follicular helper (T_FH) cells. T_FH cells (CD4⁺CXCR5⁺programmed cell death protein 1 (PD-1)⁺BCL-6⁺) are present in germinal centers in lymph nodes, where they interact with B cells to refine Ab responses (6). T_FH cells are part of the HIV reservoir: the pool of cells where HIV remains latent until reactivation (7). Follicular germinal centers, where T_FH cells reside, were once thought to lack CD8⁺ T cells, thereby allowing HIV-infected T_FH cells to escape CTL killing and to establish an HIV reservoir that evades immune system control (8). CD8⁺ T cells that express CXCR5, that is, follicular CD8⁺CXCR5⁺ T cells, have been shown to be present in secondary lymphoid tissues (4, 5) and interact with B cells and promote Ab responses (9). CD8⁺CXCR5⁺ T cells have been detected in lymphoid tissues of HIV-infected individuals, thus suggesting that they play a role in controlling HIV infection (4, 5, 10). The frequency of circulating HIV-specific CD8⁺CXCR5⁺ T cells inversely correlates with HIV serum viral load, further suggesting a protective role (4, 5). Moreover, CD8⁺CXCR5⁺ T cells display potent cytolytic activity to target and kill HIV-infected cells via bispecific Abs in vitro (10).

During chronic viral infections, a sizable population of viral-specific CD8⁺ T cell clones acquires an exhausted profile characterized by reduced IL-2 production, impaired cytolytic function and proliferation, and increased expression of coinhibitory receptors, such as the programmed cell death protein 1 (PD-1), Tim3, KLRG1, CD160, LAG3, and other markers of exhaustion (2, 11). Some evidence supports that CD8⁺CXCR5⁺ T cells express high levels of PD-1 compared with CD8⁺CXCR5⁻ T cells (4, 12, 13), whereas other reports suggest that CD8⁺CXCR5⁻ T cells express slightly higher levels of PD-1 (5, 14). Interestingly, it has been shown that CD8⁺CXCR5⁺ T cells maintain cytolytic capacity and are likely not exhausted (3, 15). Furthermore, blocking the PD-1/PD ligand 1 (PD-L1) pathway during HIV infection diminished the function of CD8⁺CXCR5⁺ T cells (12), which further suggests a protective role.

CD8⁺CXCR5⁺ T cells play a number of roles during infection by promoting cytotoxic responses, supporting B cell stimulating functions and displaying CD8⁺ T effector memory characteristics (3). CD8⁺CXCR5⁺ T cells present in human tonsils and follicular lymphoma samples reside in B cell follicles, exhibit strong cytotoxic effects against tumor cells in vitro, can inhibit T_FH-mediated B cell differentiation, and exert strong antitumor activity after adoptive transfer of CD8⁺CXCR5⁺ T cells into an animal model of lymphoma compared with its naive and CD8⁺CXCR5⁻ counterparts (16). Recently, a colorectal cancer study revealed that CD8⁺CXCR5⁺ T cells maintain cytolytic capacity to directly lyse tumor cells but also influence B cell secretion of IgG, suggesting multiple mechanisms for tumor control by these cells (17). Others provide evidence of clinical correlations where high CD8⁺CXCR5⁺ T cells in tumors or tumor-draining lymph nodes correlate with increased colorectal cancer-free survival (18), disease-free pancreatic cancer survival after tumor resection (19), or high tumor expression levels of CD8⁺CXCR5⁺ T cell signature genes that correlate with prolonged survival in follicular lymphoma (20).

Chronic B cell activation associated with HIV infection is believed to contribute to the development of non-Hodgkin lymphoma (NHL) (21–23) and plays an important role in lymphomagenesis. Our prior work has shown that serum levels of inflammatory cytokines, such as IL-6, IL-10, IP-10/CXCL10, TNF-α, soluble CD23, and CXCL13 (the ligand for CXCR5), are associated with increased risk for AIDS-associated NHL (AIDS-NHL) (24–28). We have also shown that regulatory B cells are elevated prior to AIDS-NHL diagnosis, and that PD-L1–expressing B cells comprise a subpopulation of regulatory B cells (29). We are investigating chronic HIV infection and the early events leading to AIDS-NHL by elucidating the role of CD8⁺CXCR5⁺ T cells on B cell activation and differentiation and the early events of lymphomagenesis. In this study, we show that CD8⁺CXCR5⁺ T cells are significantly elevated in peripheral blood of HIV-infected individuals part of the Multicenter AIDS Cohort Study (MACS). HIV+ individuals also show elevated absolute numbers of CD8⁺CXCR5⁺BCL-6⁺PD-1⁺ T cells compared with HIV-negative individuals. Moreover, we show that activated CD8⁺CXCR5⁺ T cells can activate B cells and promote the production of PD-L1⁺ and PD-L1⁺IgG⁺ B cells.

Viable frozen PBMCs were obtained from the UCLA MACS repository. We obtained archival samples obtained from combination antiretroviral therapy (cART) naive HIV+ participants (n = 15), as well as from HIV-negative controls (n = 15) whose race/ethnicity and age matched the HIV+ participants studied (Table I). Samples acquired from HIV+ participants were >4 y prior to AIDS-NHL diagnosis.

Multicolor flow cytometry was performed on 10⁶ viable, thawed PBMCs collected from HIV+ and HIV-negative participants from the UCLA MACS repository. PBMCs were stained using a panel of extracellular immune cell markers, such as anti-CD3 eFluor 650 (clone OKT3, eBioscience), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), anti-CXCR5 (CD185) PerCP-Cy5.5 (clone RF8B2, BD Pharmingen), anti-CD8 allophycocyanin-eFluor 780 (or allophycocyanin-Cy7) (clone RPA-T8, Invitrogen), anti-CD62L eFluor 605 (clone Dreg-56, Invitrogen), anti–PD-1 (CD279) allophycocyanin (clone MIH4, eBioscience), anti-CD71 FITC (clone OKT9, eBioscience), anti-CD27 Pacific Blue (clone O323, BioLegend), and anti-ICOS (CD278) PE (clone ISA-3, eBioscience) for 20 min at 4°C, washed once in FACS buffer (0.5% BSA in PBS; Life Technologies), and then fixed with 1% paraformaldehyde solution. A separate tube of cells was stained with the above Ab panel of extracellular immune cell markers that included anti–BCL-6 PE (clone K112-91, BD Pharmingen) for intracellular staining of BCL-6 (replacing anti-ICOS PE). For intracellular staining, a FOXP3 transcription factor staining buffer set (eBioscience/Invitrogen) was used. Briefly, 1 ml of FOXP3 fixation/permeabilization working solution was added to each sample and incubated for 30–60 min at 2–8°C. After the incubation period, 2 ml of 1× permeabilization buffer was added to each sample and centrifuged at 1500 rpm for 5 min at room temperature, washed once more, and then stained with BCL-6 PE for 1 h at room temperature. Cells were then washed once with FACS buffer and analyzed for flow cytometry using a BD LSRFortessa cell analyzer. Data presented in Figs. 1, 2, and 5 were analyzed with the FCS Express software (v3, De Novo Software), and the gating strategy shown in Fig. 1A was completed using the FCS Express software (v7.04.0004, De Novo Software). Absolute CD3⁺ T cell numbers were provided by the UCLA MACS, and absolute CD4⁺ T cell numbers for each HIV-negative or HIV+ participant were determined from the percentage of gated CD3⁺ T cells. Median CD4⁺ T cell numbers are provided for each group in Table I.

PBMCs from healthy HIV-negative blood donors were obtained from the UCLA CFAR Virology Core. Cells (10⁸) were used to sort CD3⁺CD4⁺CD8⁻CXCR5⁺ (follicular CD4⁺ or T_FH) and CD3⁺CD4⁻CD8⁺CXCR5⁺ (follicular CD8⁺) T cells. PBMCs were stained with a panel of anti-human Abs that included anti-CXCR5 Alexa Fluor 488 (clone RF8B2, BD Biosciences), anti-CD3 eVolve655 (clone OKT3, eBioscience), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), and anti-CD8 allophycocyanin-eFluor 780 (clone RPA-T8, eBioscience). After staining cells for 20 min at 4°C, cells were washed once with FACS buffer and resuspended in PBS. Cells were then sorted using a FACSAria II cell sorter at the UCLA Jonsson Comprehensive Cancer Center Janis V. Giorgi Flow Cytometry Core Laboratory. For sorting of PBMCs and tonsil cells (below), the BD FACSDiva software program (v8.02) was used. FlowJo (v10.8.1) was used to generate the gating strategy plots shown in Supplemental Figs. 1 and 2.

CD19⁺ B cells were isolated from PBMCs obtained from the same donor as used to obtain T_FH or CD8⁺CXCR5⁺ T cells. B cells were isolated from a starting total of 5 × 10⁷ PBMCs using an EasySep human pan B cell enrichment kit according to the manufacturer’s instructions (STEMCELL Technologies). Sorted T_FH or CD8⁺CXCR5⁺ T cells were plated at a density of 10⁵ cells per well in duplicate of a 96-well plate, and 10⁵ CD19⁺ B cells were added to each well for a 1:1 T cell/B cell ratio in a final volume of 200 µl per well of complete RPMI 1640 medium (RPMI 1640 medium supplemented with 10% FBS, 100 U/ml penicillin [Life Technologies, Thermo Fisher Scientific], and 100 μg/ml streptomycin [Life Technologies, Thermo Fisher Scientific]). Cells were cultured in the presence of IL-2 (20 U/ml), with or without anti-CD3– and anti-CD28–coated Dynabeads (Life Technologies) (1:1 bead/cell ratio) for 2 d at 37°C, 5% CO₂. After 2 d of stimulation, cells were washed once with FACS buffer and then resuspended in fresh FACS buffer. Cells were then stained for flow cytometry using the following anti-human Ab panel: anti-CXCR5 Alexa Fluor 488 (clone RF8B2, BD Biosciences), anti-CD3 eVolve655 (clone OKT3, eBioscience), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), anti-CD8 allophycocyanin-eFluor 780 (clone RPA-T8, eBioscience), anti-CD19 PerCP-Cy5.5 (clone SJ25C1, eBioscience), and anti–PD-L1 allophycocyanin (clone M1H1, eBioscience).

Tonsil specimens were obtained from healthy donors from the UCLA Ronald Reagan Medical Center. Tonsils were processed by grinding the tissue and passing through a 70-µm cell strainer (Corning) with 1× PBS. The resulting cell suspension was mixed with 1× RBC lysis buffer (Invitrogen) and incubated for 5 min at room temperature. Cells were then washed once in complete RPMI 1640 medium at 1500 rpm for 5 min. Cells were counted and placed in two separate aliquots of 35 × 10⁶ cells each in 5-ml polystyrene tubes, and then washed with FACS buffer and centrifuged at 1500 rpm for 5 min. The two tubes consisting of 35 × 10⁶ cells each were then stained with a panel of anti-human Abs that included anti-CXCR5 Alexa Fluor 488 (clone RF8B2, BD Biosciences), anti-CD3 Super Bright 645 (clone OKT3, Invitrogen), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), anti-CD8 Alexa Fluor 700 (clone RPA-T8, eBioscience), anti-CD19 PerCP-Cy5.5 (clone SJ25C1, eBioscience), anti-IgG PE (eBioscience), and IgM eFluor 450 (clone SA-DA4, eBioscience). After staining cells for 20 min at 4°C, cells were washed once with FACS buffer and resuspended in a final volume of 1.5 ml of 1× PBS. Cells were sorted using the FACSAria II cell sorter to separate the following three cell populations: CD3⁺CD4⁻CD8⁺CXCR5⁻ T cells, CD3⁺CD4⁻CD8⁺CXCR5⁺ T cells, and CD3⁻CD19⁺IgM⁺IgG⁻ B cells. The purities of these populations after sorting were as follows: CD3⁺CD4⁻CD8⁺CXCR5⁻ T cells (93–95%), CD3⁺CD4⁻CD8⁺CXCR5⁺ T cells (93–96%), and CD3⁻CD19⁺IgM⁺IgG⁻ B cells (96–97%) across four independent experiments.

CD3⁺CD4⁻CD8⁺CXCR5⁻ or CD3⁺CD4⁻CD8⁺CXCR5⁺ T cells were then cocultured with CD19⁺IgM⁺IgG⁻ B cells at a 1:1 T cell/B cell ratio in complete RPMI 1640 medium and in the presence of IL-2 (20 U/ml) plus anti-CD3– and anti-CD28–coated Dynabeads (1:1 bead/cell ratio) for 3 and 5 d (from the same sorted subsets) at 37°C, 5% CO₂. After 3 and 5 d, cells were harvested, split in thirds, and stained with three separate anti-human Ab panels for T and B cells. Panel 1 included anti-CXCR5 Alexa Fluor 488 (clone RF8B2, BD Biosciences), anti-CD3 Super Bright 645 (clone OKT3, Invitrogen), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), anti-CD8 Alexa Fluor 700 (clone RPA-T8, eBioscience), anti-CD19 PerCP-Cy5.5 (clone SJ25C1, eBioscience), anti–PD-1 allophycocyanin (clone EH12.2H7, BioLegend), and anti-CD40L PE (clone 89-76, BD Biosciences). Panel 2 included the above Ab panel without anti-CD40L and with BCL-6 PE (clone K112-91, BD Biosciences) for intracellular staining. Panel 3 included anti-CXCR5 Alexa Fluor 488 (clone RF8B2, BD Biosciences), anti-CD3 Super Bright 645 (clone OKT3; Invitrogen), anti-CD4 PE-Cy7 (clone SK3, BD Biosciences), anti-CD8 Alexa Fluor 700 (clone RPA-T8, eBioscience), anti-CD19 PerCP-Cy5.5 (clone SJ25C1, eBioscience), anti–PD-L1 allophycocyanin (clone M1H1, eBioscience), anti-IgG PE (eBioscience), and anti-IgM eFluor 450 (clone SA-DA4, eBioscience). Cells were stained for 20 min at 4°C, washed once in FACS buffer, and then resuspended in 2% paraformaldehyde. For PBMC and tonsil cell experiments described above, cells were analyzed for flow cytometry using a BD LSRFortessa cell analyzer. Data were analyzed with the FCS Express software program (v7.04.0004, De Novo Software).

Graphical representations and statistical analysis were performed with GraphPad Prism 9.4.0. For data comparing HIV+ and HIV-negative study populations, unpaired, nonparametric Mann–Whitney tests (two-tailed) were performed (Figs. 1, 2). For data acquired from PBMC experiments (B cells cocultured with T_FH or CD8⁺CXCR5⁺ T cells), an unpaired, nonparametric Kruskal–Wallis test was performed (Fig. 3). For data acquired from tonsil cell experiments of B cells cocultured with CD8⁺CXCR5⁻ or CD8⁺CXCR5⁺ T cells, an unpaired, nonparametric Kruskal–Wallis test was conducted (Fig. 4). In one case (Fig. 4D), a statistical comparison was made using an unpaired, nonparametric Mann–Whitney U test (two-tailed). For correlation analysis, two-tailed tests with a 95% confidence interval were performed, where the Pearson correlation coefficient and p value were determined (Fig. 5). A p value of ≤0.05 was considered statistically significant in all cases.

HIV+ participants and HIV-negative controls, who were participants in the UCLA MACS, were comparable in their demographic characteristics such as recruitment year, race, and ethnicity profiles (93% white, non-Hispanic and 7% Hispanic for HIV-negative controls versus 90% white, non-Hispanic and 5% Hispanic for HIV+ individuals), as well as age group, which had medians of 34 and 38 y of age. All of the HIV+ individuals were cART naive (see Table I).

We stained PBMCs from individuals living with untreated HIV infection to determine the absolute numbers of CD8⁺CXCR5⁺ T cells and CD8⁺CXCR5⁺PD-1⁺ T cells at a single MACS visit (n = 15). CD8⁺CXCR5⁺, CD8⁺CXCR5⁺PD-1⁺, and CD8⁺CXCR5⁺PD-1⁻ T cells were each gated from CD3⁺CD4⁻CD8⁺ T cells, as shown in (Fig. 1A. Follicular CD8⁺ T cells accumulate in HIV-infected lymphoid follicles and germinal centers (4, 5, 9, 10), with a minor fraction of these cells in HIV-negative individuals. In addition, others have shown that CD8⁺CXCR5⁺ T cell quantity is elevated in peripheral blood of HIV-infected individuals compared with healthy controls (12). We observed that HIV+ participants had significantly elevated absolute numbers of CD3⁺CD4⁻CD8⁺CXCR5⁺ T cells in blood when compared with HIV-negative individuals (Fig. 1B). The average percentage of CD3⁺CD4⁻CD8⁺ T cells that express CXCR5⁺ T cells in HIV-negative individuals (n = 15) was 3.21 ± 1.45 (n = 15) and for those who were HIV+ was 2.42 ± 1.71 (n = 15). The overall percentages were not significantly different. We found that the absolute numbers of CD3⁺CD4⁻CD8⁺CXCR5⁺PD-1⁺ T cells were also significantly elevated in HIV+ individuals compared with HIV-negative controls (Fig. 1C), whereas no significant differences in the absolute numbers of CD3⁺CD4⁻CD8⁺CXCR5⁺PD-1⁻ T cells were observed (Fig. 1D).

We then quantified the expression of BCL-6, a transcriptional factor that is characteristically expressed in T_FH cells. We observed that CD8⁺CXCR5⁺ T cells isolated from HIV+ and HIV-negative individuals expressed PD-1, CD27, and BCL-6 (Fig. 2A). HIV+ individuals had elevated absolute numbers of CD8⁺CXCR5⁺BCL-6⁺PD-1⁺ T cells compared with HIV-negative individuals (p = 0.021) (Fig. 2B), and similar results were observed for the percentage of CD8⁺CXCR5⁺BCL-6⁺ T cells that express PD-1 (median percentage of 59.8% for HIV+ individuals versus median percentage of 35.2% for HIV-negative individuals; p = 0.015).

Because CD8⁺CXCR5⁺ T cells share phenotypic characteristics with T_FH cells, it is possible that they may also share functional features. To test this hypothesis, we determined whether CD8⁺CXCR5⁺ T cells can induce B cells to express PD-L1, as T_FH cells do. PD-L1⁺ B cells are known to be elevated during HIV infection (29). We isolated PBMCs from healthy donors and sorted CD3⁺CD4⁺CD8⁻CXCR5⁺ (T_FH cells) (positive control) and CD3⁺CD4⁻CD8⁺CXCR5⁺ (CD8⁺CXCR5⁺) T cells from the same donor (Supplemental Fig. 1). T_FH or CD8⁺CXCR5⁺ T cells were cocultured with naive CD3⁻CD19⁺ B cells isolated from the same donor (1:1 T cell/B cell ratio) in the presence of IL-2 and in the presence or absence of anti-CD3/anti-CD28 coated Dynabeads. As shown in (Fig. 3A, T_FH cells induced expression of PD-L1 on B cells and this induction was more pronounced when T_FH cells were activated with both IL-2 and anti-CD3/anti-CD28 (Fig. 3A), as has been shown by others (30, 31). Activated CD8⁺CXCR5⁺ T cells were also capable of significantly inducing PD-L1 expression on B cells, and this induction was also more pronounced when CD8⁺CXCR5⁺ T cells were activated with both IL-2 and anti-CD3/anti-CD28 (Fig. 3B).

To further investigate whether CD8⁺CXCR5⁺ T cells from secondary lymphoid organs are capable of inducing PD-L1 expression on B cells, tonsil cells were stained and sorted for CD3⁺CD4⁻CD8⁺CXCR5⁻ and CD3⁻CD4⁻CD8⁺CXCR5⁺ T cells, as well as CD3⁻CD19⁺IgM⁺IgG⁻ B cell populations (Supplemental Fig. 2). It has been reported that CXCR5⁺PD-1⁺CD8⁺ T cells provide help to B cells for the production of Igs, such as IgG1 (32). Thus, CD8⁺CXCR5⁻ or CD8⁺CXCR5⁺ T cells were cocultured with CD19⁺IgM⁺IgG⁻ B cells at a 1:1 ratio in the presence of IL-2 plus anti-CD3/anti-CD28 Dynabeads for 3 and 5 d. We found that activated CD8⁺CXCR5⁺ T cells significantly induced the percentage of CD19⁺PD-L1⁺ B cells at 3 (Fig. 4A) and 5 d of coculture (Supplemental Fig. 3A). Moreover, we determined a significantly higher percentage of CD19⁺PD-L1⁺IgG⁺ B cells after culture with activated CD8⁺CXCR5⁺ T cells compared with cocultures with activated CD8⁺CXCR5⁻ T cells or B cells alone and in the absence of activated CD8⁺ T cells at 3 (Fig. 4B) and 5 d (Supplemental Fig. 3B).

The ability of CXCR5⁺PD-1⁺CD8⁺ T cells to help B cells depends on the CD40/CD40L interaction (32). Thus, we further measured the expression of CD40L on CD8⁺CXCR5⁻ and CD8⁺CXCR5⁺ T cells after coculture with B cells. We found that in most cases, a higher percentage of CD40L⁺ cells was observed among activated CD8⁺CXCR5⁺ T cells after 3 (Fig. 4C, 4D) and 5 d of coculture (Supplemental Fig. 3).

Considering that CD8⁺CXCR5⁺ T cells can induce PD-L1 expression on B cells and PD-L1⁺ B cells are also elevated during HIV infection (29), we further investigated whether CD8⁺CXCR5⁺ T cells correlated with the numbers of PD-L1–expressing B cells in peripheral blood of HIV-infected individuals. We found that the absolute numbers of CD8⁺CXCR5⁺ T cells were significantly and positively correlated with the absolute numbers of PD-L1⁺ B cells (Fig. 5).

In this study, we describe the presence of CD8⁺CXCR5⁺ T cells in peripheral blood during HIV infection. Utilizing flow cytometry, we analyzed these cells in the peripheral blood of 15 HIV-infected participants, who were naive to cART, as well as in a group of 15 HIV-negative controls. Both groups were matched for race/ethnicity and age (Table I).

We observed that the numbers of CD8⁺CXCR5⁺ T cells were elevated in the peripheral blood of HIV-infected individuals. CD8⁺CXCR5⁺ T cells have been previously shown to be elevated in lymphoid tissue of HIV-infected individuals (4, 5). During HIV infection, CD8⁺CXCR5⁺ T cells in lymph nodes have been described to be HIV specific and to play an important role in HIV clearance (4, 5). The higher number of CD8⁺CXCR5⁺ T cells in the blood of HIV-infected individuals is of importance, as this suggests that these cells migrate from the lymphoid tissue during HIV infection. Although the reason for the presence of these cells in the peripheral circulation is unknown, it is known that CXCL13, the natural ligand of CXCR5, is elevated in the serum of HIV-infected individuals (33). Therefore, it is possible that the elevated levels of CXCL13 seen in blood of HIV-infected individuals may be leading CD8⁺CXCR5⁺ T cells to traffic out of the lymphoid tissue into the peripheral circulation. Additionally, the architecture of secondary lymphoid tissues is disrupted in progressive untreated HIV infection (34), which may result in the relocation to peripheral circulation of cells that would normally reside in these tissues. We also observed a higher percentage of CD8⁺CXCR5⁺PD-1⁺ T cells in HIV-infected individuals.

The role of CD8⁺CXCR5⁺ T cells in the control of HIV infection can be explained, at least in part, by the fact that these cells express BCL-6. BCL-6 is the transcriptional factor characteristic of T_FH cells (6), and CD8⁺CXCR5⁺ T cells residing in the lymph nodes have been ascribed to provide help promoting B cell activation and Ab secretion (9, 35). HIV infection increases the number of CD8⁺CXCR5⁺ T cells, and these cells presumably have an active role in the maturation of specific anti-HIV B cells. During chronic HIV infection, CD8⁺CXCR5⁺ T cells can become exhausted, as indicated by the expression of PD-1 and other markers of exhaustion, thus adding to the pool of differentiated, dysfunctional CD8⁺ T cells that are characteristically seen in HIV infection.

CD8⁺CXCR5⁺ T cells are phenotypically similar to T_FH cells. Therefore, we wanted to determine whether CD8⁺CXCR5⁺ T cells can also exert a similar function. The main function of T_FH cells is to help B cells in the germinal center, where they induce B cells to undergo class switch recombination and somatic hypermutation, which eventually induce B cells to secrete specific Abs (6). The interaction between T_FH and B cells is tightly regulated. To study whether CD8⁺CXCR5⁺ T cells share functions with T_FH cells, we investigated whether CD8⁺CXCR5⁺ T cells could induce PD-L1 expression on B cells, as T_FH cells are capable of doing so. We found that CD8⁺CXCR5⁺ T cells isolated from peripheral blood and a secondary lymphoid organ, such as the tonsil, can induce the generation of PD-L1⁺IgG⁺ B cells. Our data also show that activated CD8⁺CXCR5⁺ T cells express CD40L, further suggesting a possible mechanism for activating B cells. Moreover, we determined that the number of CD8⁺CXCR5⁺ T cells correlated with the number of PD-L1⁺ B cells isolated from HIV-infected individuals. As PD-L1⁺ B cells are elevated in HIV infection, and more so in individuals who are HIV+ and go onto develop lymphoma (29), the activation of B cells by CD8⁺CXCR5⁺ T cells may contribute to immune dysfunction and pathogenesis in HIV infection. PD-L1–expressing B cells may exert regulatory functions by interacting with PD-1 and limiting T cell activation.

In summary, CD8⁺CXCR5⁺ T cells, a novel population of B helper cells, merit further study, as the function of these cells in those living with chronic, untreated HIV infection may contribute to progressive HIV disease. Therefore, more precisely determining the function of these cells may provide important insights on the immune control of HIV infection. Additionally, these cells may be important contributors to the general immune activation present during the course of HIV infection, which contributes to the many comorbidities seen in HIV-infected individuals.

We thank Dr. Zoran Galic, Dr. Alejandro Garcia, Salem Haile, Miriam Guemes-Aragon, Min Zhou, and Iris Williams at the Janis V. Giorgi Flow Cytometry Core Facility, Jonsson Comprehensive Cancer Center, and the Center for AIDS Research/AIDS Institute Shared Resource, which helped with FACS sorting and data analysis. We also thank Larry Magpantay for laboratory support, and Dr. Roger Detels for assistance in obtaining specimens from the Los Angeles Clinical Research Site for the MACS/WIHS Combined Cohort Study.

M.L.P. has a financial interest in Stellar Biosciences, Inc. The Regents of the University of California are in discussions with Stellar Biosciences to license a technology invented by M.L.PP. to this firm. In addition, M.L.P. has a financial interest in Klyss Biotech, Inc. The other authors have no financial conflicts of interest.