CD161+ CD4+ T Cells Harbor Clonally Expanded Replication-Competent HIV-1 in Antiretroviral Therapy-Suppressed Individuals

The latent reservoir continues to be the major obstacle to curing HIV-1 infection. The clonal expansion of latently infected cells adds another layer maintaining the long-term stability of the reservoir, but its mechanism remains unclear. Here, we report that CD161+ CD4+ T cells serve as an important compartment of the HIV-1 latent reservoir and contain a significant amount of clonally expanded proviruses. In our study, we describe a feasible strategy that may reduce the size of the latent reservoir to a certain extent by counterbalancing the repopulation and dissemination of latently infected cells.

Therefore, the latent reservoir is widely considered the major barrier to curing HIV-1 infection (5,6). The dynamic mechanisms that maintain long-term immune memory within the CD4 ϩ T cell compartment may play an important role in the stability and longevity of the HIV-1 latent reservoir (7). Recently, several reports have suggested that the latent reservoir is maintained through cellular proliferation and the clonal expansion of infected CD4 ϩ T cells in ART-treated HIV-1-infected individuals (8,9). Further studies also indicated that infected CD4 ϩ T cells harboring replication-competent provirus can undergo homeostatic proliferation in response to T cell receptor (TCR) stimulation or certain cytokines without viral reactivation (10,11). Mathematical modeling showed that clonal expansion represents a critical factor contributing to the slow decay of the latent reservoir (12). However, the mechanisms underlying the expansion of latent HIV-1 remain largely unknown. Considering that HIV-1 proviruses are not equally distributed across memory CD4 ϩ T cell subsets (11), we are curious to know and it is important to know whether a certain CD4 ϩ T cell subset acts as a major driving force for the clonal expansion of latently infected CD4 ϩ T cells.
Previous reports have suggested that the relatively long-lived Th17 cells may contribute to HIV-1 persistence under ART (13,14). As Th17 cells are derived from CD161 ϩ blood precursors (15,16), CD161 would be an efficient marker to easily identify Th17 cells (15,17,18). Another study proposed that a subset of memory CD8 ϩ T cells defined phenotypically by the expression of high levels of CD161 has stem-like properties (19). Interestingly, studies have also identified the existence of CD4 ϩ T cells with stem-like properties among Th17 cells, suggesting that cell fate diversification results in the generation of T cells with a stem-like phenotype, even within more differentiated T cell subsets (20,21). In addition, CD161 can act as a costimulatory receptor to increase the response to TCR stimulation and is a marker for long-lived antigen-specific memory T cells (17,(22)(23)(24). Based on the aforementioned characteristics of CD161 ϩ CD4 ϩ T cells, we hypothesize that these cells may play a critical role in the expansion of the HIV-1 latent reservoir. In this study, we describe the existence of a specialized subset of CD4 ϩ T cells defined phenotypically by the expression of CD161 on their surfaces. These cells are more susceptible to HIV-1 infection and have a higher proliferative ability. Additionally, larger amounts of latent HIV-1 are presented within this subset than in CD161 Ϫ CD4 ϩ T cells. More importantly, CD4 ϩ T cells harboring clonally expanded HIV-1 proviral sequences are significantly enriched within the CD161 ϩ subset from ART-treated infected individuals than within other CD4 ϩ T cells, suggesting that HIV-1-infected CD161 ϩ CD4 ϩ T cells may indeed drive the clonal expansion of latently infected cells.

RESULTS
CD161 ؉ CD4 ؉ T cells are highly permissive for HIV-1 infection. To investigate whether HIV-1 infection affects CD161 ϩ CD4 ϩ T cells, we measured the percentage of cells expressing CD161 among CD4 ϩ T cells from the blood or lymph node (LN) of HIV-1-negative and HIV-1-positive individuals receiving or not receiving ART. The sociodemographic, clinical, and behavioral characteristics of the individuals studied are presented in Table 1. In healthy individuals, a median of 26% of total CD4 ϩ T cells (interquartile range [IQR], 21 to 30%) expressed CD161, which was 1.5 times higher than the proportion in HIV-1-infected individuals not on ART (median, 16%; IQR, 12 to 22%) and 1.8 times higher than that in HIV-1-infected donors on ART (median, 14%; IQR, 11 to 17%) (Fig. 1A), suggesting that HIV-1 infection may deplete or downregulate CD161 ϩ CD4 ϩ T cells. The depletion or downregulation of CD4 ϩ T cells may be associated with high levels of immune activation in HIV-1-infected subjects (25,26). Our results showed that the activation level (CD25, CD69, CD38, or HLA-DR expression) on CD161 ϩ CD4 ϩ T cells was higher in HIV-1-infected subjects than in healthy donors (see Fig. S1A in the supplemental material). To further clarify whether the decreased CD161 expression in HIV-1-infected individuals was really caused by HIV-1 infection and to determine the proportion of CD161 ϩ CD4 ϩ T cells that could be recovered after ART, we followed CD161 expression on CD4 ϩ T cells from 13 subjects at different time points from the beginning of ART to after 1 year of ART. Our results showed that the proportion of CD161-positive cells increased significantly after 1 year of ART ( Fig. 1B; clinical information is presented in Table 2), suggesting that these cells may indeed be affected by HIV-1 replication. To check whether the expression of CD161 is stable, we applied TCR activation and a T cell homeostasis signal to CD161-positive and -negative cells. Our data showed that CD161 expression on CD4 ϩ T cells was relatively stable, as CD3/CD28 activation or the level of the interleukin-7 (IL-7)/IL-15 signal did not change for either subset (Fig. 1C), supporting the suggestion that CD161 may be used as a marker to identify this specific cell population.
To determine the susceptibility of CD161 ϩ CD4 ϩ T cells to HIV-1 infection, we first quantified the expression level of the coreceptors CCR5 and CXCR4 from HIV-1-negative and -positive individuals receiving or not receiving ART. CXCR4 was expressed at high and similar levels on these two T cell subsets in the three groups (Fig. 1D) and also in the memory subsets of CD161-positive and -negative CD4 ϩ T cells from healthy donor peripheral blood mononuclear cells (PBMCs) and lymph nodes (Fig. S1B). However, CCR5 expression was significantly higher on CD161 ϩ CD4 ϩ T cells than on CD161negative cells both in healthy individuals and in HIV-1-infected donors (Fig. 1D). Memory CD161 ϩ CD4 ϩ T cells from HIV-1-negative PBMCs and lymph nodes also expressed higher levels of CCR5 than CD161 Ϫ CD4 ϩ T cells (Fig. S1B). We further investigated the susceptibility of these two T cell subsets to R5-or X4-tropic HIV-1 infection in vitro. Regardless of whether the cells were prestimulated with anti-CD3 and anti-CD28 antibodies or not, infectivity in CD161 ϩ CD4 ϩ T cells was significantly higher than that in CD161 Ϫ CD4 ϩ T cells at various time points postinfection ( Fig. 1E and F), and this was also the case for CD161-positive and -negative CD45RO ϩ CD4 ϩ T cells from healthy donor PBMCs (Fig. S2). Taken together, our data clearly indicate that CD161 ϩ CD4 ϩ T cells are highly permissive for HIV-1 infection and that the reduced amount of these cells in infected individuals may be due to the death of the infected cells.
CD161 ؉ CD4 ؉ T cells are primarily of the memory phenotype with typical of Th17 and pTFH characteristics. The HIV-1 latent reservoir has been shown to be composed of mainly resting memory CD4 ϩ T cells (5,27). We characterized the phenotype of CD161 ϩ CD4 ϩ T cells to determine if it was in line with the features of latently infected cells. Roughly 90% of CD161 ϩ CD4 ϩ T cells expressed CD45RO but not CD45RA (Fig. S3A). Using CCR7 and CD27 expression to further distinguish memory subsets, we found that nearly 50% of CD161 ϩ CD4 ϩ T cells were of the central memory T cell (T CM ) phenotype, with slightly fewer cells exhibiting the transitional memory T cell (T TM ) and effector memory T cell (T EM ) phenotypes ( Fig. 2A). On the contrary, the CD161-negative population was primarily of the naive phenotype, with only 20% of it being T CM cells ( Fig. 2A). We also found that the composition of memory cell subsets of CD161 ϩ cells in HIV-1-infected donors under ART was similar to that in healthy individuals (Fig. S3B). To gain insight into the lineage commitment and homing properties of CD161 ϩ CD4 ϩ T cells, we examined their chemokine receptor profile. CCR6 was expressed on most (72%) CD161-positive CD4 ϩ T cells, in contrast to the findings for the CD161-negative subsets (19%) ( Fig. 2B and C). CD161 ϩ CD4 ϩ T cells from HIV-1-infected and healthy donor PBMCs produced larger amounts of IL-17A, (E and F) Infectivity of X4-or R5-tropic HIV-1 for bulk, CD161 ϩ , or CD161 Ϫ CD4 ϩ T cells. Subsets from HIV-negative subjects were sorted and then exposed directly (without prestimulation) to R5-tropic NL4-3 (BaL env) (n ϭ 6) or X4-tropic NL4-3 (n ϭ 11) or were first prestimulated via anti-CD3/CD28 for 3 days and then exposed to NL4-3 (BaL env) (n ϭ 4) or NL4-3 (n ϭ 6). Intracellular HIV-1 p24 levels were quantified by FACS (E) at the indicated time points postinfection (F). FSC-A, forward scatter area. Significant intergroup differences were determined using the rank Kruskal-Wallis test incorporating Dunn's tests for multiple comparisons. The Mann-Whitney U test or Student's t test was used to compare the statistical significance between cell subsets. P values less than 0.05 were considered significant. *, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001.
IL-22, gamma interferon, and tumor necrosis factor alpha than CD161 Ϫ CD4 ϩ T cells after ex vivo stimulation with phorbol-12-myristate-13-acetate (PMA) and ionomycin (Fig. 2D), as did the memory subset of CD161 Ϫ CD4 ϩ T cells (Fig. S3C). CD161 ϩ CD4 ϩ T cells from healthy donor LN cells also secreted more IL-17A and IL-22 than CD161 Ϫ CD4 ϩ T cells (Fig. S3C). To further investigate the expression of CD161 in different T helper cell subsets, we used CCR4, CXCR3, CCR6, and CD45RO to identify Th1, Th2, Th17, and Th1Th17 cells. The frequency of Th17 and Th2 cells was higher among CCR6-positive and -negative cells, respectively. The expression of CD161 was higher in Th17 and Th1Th17 cells than in Th1 or Th2 cells ( Fig. 2E and F). Peripheral follicular T helper (pTFH) cells have recently been shown to be a major viral replication cellular compartment and harbor a significant amount of intracellular HIV-1 proviral DNA (28). We found that CD161 ϩ CD4 ϩ T cells expressed higher levels of CXCR5 than CD161 Ϫ CD4 ϩ T cells both in blood and in LN from healthy donors ( Fig. 2G; Fig. S3D). The production of IL-21 was also significantly higher in CD161 ϩ CD4 ϩ T cells than in CD161 Ϫ CD4 ϩ T cells after being stimulated in vitro at the RNA and protein levels ( Fig. 2G).
To further confirm whether the CD161 ϩ CD4 ϩ T cells among the Th17 and Th1Th17 cell subsets were highly permissive to HIV-1 infection, we performed an in vitro infectivity assay. Our results showed that CCR5 expression was higher on CD161positive cells than on CD161-negative cells for the Th1, Th2, and Th1Th17 subsets (Fig. S3E). The Th1Th17 subset was more permissive to R5-tropic HIV-1 than the other subsets (Fig. S3F). Finally, the infectivity of CD161 ϩ cells was higher than that of CD161 Ϫ cells in these four cell populations (Fig. S3G). Together, our data demonstrated that CD161 ϩ CD4 ϩ T cells are primarily of the memory phenotype with typical characteristics of Th17 and pTFH cells; thus, they possess the potential to be latently infected with HIV-1.
CD161 ؉ CD4 ؉ T cells exhibit higher survival and proliferative abilities than CD161 ؊ CD4 ؉ T cells. Having demonstrated that CD161 ϩ CD4 ϩ T cells are highly permissive to HIV-1 infection, we went on to further investigate whether these cells have the potential to survive longer and proliferate in vitro, which is a critical prerequisite for promoting the clonal expansion of latent HIV-1. c-kit and Bcl-2 are key signaling molecules for cellular survival (29,30). We found that the percentage of c-kit expression was significantly higher on CD161 ϩ CD4 ϩ T cells than on the negative populations in both healthy and infected individuals. The mean fluorescence intensity (MFI) of Bcl-2 was also higher on CD161-positive CD4 ϩ T cells than CD161-negative CD4 ϩ T cells from healthy donors (Fig. 3A to D). This pattern remained the same in the memory subsets of CD161-positive and -negative CD4 ϩ T cells from healthy donors (Fig. S4A). OX40 was previously shown to be associated with the long-term survival and the clonal proliferation of CD4 ϩ T cells and may denote cells harboring a higher frequency of latent HIV-1 (31, 32). Here, we found that either without stimulus or with anti-CD3/CD28, IL-15, or IL-7 treatment, both bulk and memory CD161 ϩ CD4 ϩ T cells expressed higher levels of OX40 than their CD161-negative counterparts (Fig. 3E). IL-7 or IL-15 is an important stimulus for the homeostatic proliferation of memory CD4 ϩ T cells (33)(34)(35). To explore the potential to receive IL-7 signaling, we measured IL-7 receptor (CD127) expression on CD161-positive and -negative CD4 ϩ T cells. Although the percentage of cells expressing CD127 was comparable between the two groups, the MFI of CD127 was significantly higher on CD161 ϩ CD4 ϩ T cells than on CD161 Ϫ CD4 ϩ T cells from both healthy and infected individuals (Fig. 3F). This pattern remained the same with the memory subsets of CD161-positive and -negative CD4 ϩ T cells (Fig. S4B). We further examined the proliferative ability of CD161 ϩ CD4 ϩ T cells in response to in vitro anti-CD3/CD28, IL-15, or IL-7 treatment. Compared with their CD161-negative counterparts, CD161 ϩ CD4 ϩ T cells proliferated significantly better with TCR or homeostatic cytokine signals (Fig. 3G). Even within the memory subset, CD161 ϩ CD4 ϩ T cells showed a higher proliferative ability than CD161 Ϫ CD4 ϩ T cells under the same treatment conditions with or without IL-2 (Fig. S4C). Taken together, CD161 ϩ CD4 ϩ T cells possess characteristics of long-lived memory T cells and undergo significant proliferation in response to in vitro antigenic or homeostatic signaling, suggesting that these cells have the potential to promote the clonal expansion of latent HIV-1 if they become infected. CD161 ؉ CD4 ؉ T cells harbor more replication-competent latent HIV-1 and clonal expanded proviruses than CD161 ؊ CD4 ؉ T cells. If infected CD161 ϩ CD4 ϩ T Here, we used a quantitative PCR (qPCR) assay, an intact proviral DNA assay (IPDA) (36), and a quantitative viral outgrowth assay (QVOA) (37) to measure the reservoir size in purified CD161-positive and -negative CD4 ϩ T cells from ART-suppressed infected individuals. The data clearly showed that the frequencies of both proviral DNA and replication-competent latent HIV-1 were significantly higher in CD161 ϩ CD4 ϩ T cells than in CD161 Ϫ CD4 ϩ T cells by the three different approaches, with 6.7-fold, 13.0-fold, and 2.1-fold enrichment in the positive population, respectively ( Fig. 4A to C). HIV-1 latently infected cells possess a remarkable stability and persist long term in HIV-1-infected subjects despite successful ART (4,38,39). The mechanism of memory cell homeostasis suggests that the stability of the latent reservoir could be partially dependent upon the ability of infected cells to proliferate. Phylogenetic analysis established that identical proviral sequences in HIV-1-infected individuals under ART can reflect expanded cellular clones (40). Each set of identical env sequences is most likely to represent a clonal population of infected cells derived from a single cell initially infected by HIV-1 (10).
To further evaluate whether the cellular proliferation of CD161 ϩ CD4 ϩ T cells can act as an important driving force for HIV-1 persistence, we analyzed the proviral env sequence by limiting dilution PCR and single-genome proviral sequencing. We recovered identical proviral env sequences from 6 out of 8 subjects analyzed. The CD161 ϩ CD4 ϩ T cells in all 6 subjects harbored identical proviral env sequences, whereas the CD161 Ϫ CD4 ϩ T cells from only 3 subjects harbored clonal expanded sequences. A total (D) Clonality of the proviral env sequence recovered from CD161-positive or -negative CD4 ϩ T cells within the total number of sequences. Significance was calculated using a chi-square test; the P value is indicated. (E) Representative neighbor-joining phylogenetic tree constructed with env sequences recovered from CD161-positive or -negative CD4 ϩ T cells by single-genome proviral sequencing. Red arrows, identical clonal expansion sequences; blue symbols, sequences detected in CD161-positive CD4 ϩ T cells; purple symbols, sequences detected in CD161-negative CD4 ϩ T cells. See Table 3 for patient characteristics. of 32.7% of the proviral env sequences recovered from CD161 ϩ CD4 ϩ T cells but only 8.9% of those from their CD161-negative counterparts were clonal (Fig. 4D). A summary of the results of phylogenetic analyses of proviral HIV-1 env sequences is shown in Table 3. A representative tree of one infected individual is shown in Fig. 4E, and the results for the other donors are shown in Fig. S5. Together, this direct evidence from ART-treated individuals strongly indicates that HIV-1-infected CD161 ϩ CD4 ϩ T cells undergo clonal expansion in vivo, therefore composing a significantly larger HIV-1 latent reservoir in these cells than in other cells.

DISCUSSION
Previous works showed that the distribution of latently infected CD4 ϩ T cells across different subsets is not equal (41). For example, central memory T cells (11), Th17 cells (13), and follicular T helper cells (42) largely contribute to viral persistence. The documented role of CD161 ϩ CD4 ϩ T cells in pathogenic autoimmunity (17,43) and their lineage connection to Th17 cells make them a possible candidate to serve as HIV-1 targets at the portal of entry (44). We observed that CD161 ϩ CD4 ϩ T cells were highly permissive to X4-or R5-tropic HIV-1 infection and harbored total HIV-1 DNA at significantly higher frequencies than the CD161-negative population. IPDA and QVOA further demonstrated that CD161 ϩ CD4 ϩ T cells make up a substantial portion of the replication-competent HIV-1 latent reservoir. Latently infected T cells that persist under successful ART do not express viral antigens, making it nearly impossible to target these cells for clearance (41). The identification of viable markers of latently infected cells is therefore urgent. Some molecules, such as immune checkpoint inhibitors (45), CD32a (46), or CD30 (47), have recently been reported to be markers of the latent reservoir, but the use of these as markers is still under debate. It has also been reported that CCR5-expressing cells enrich latently infected cells, but CCR5 is not an ideal marker because its expression turns on and off (48). Our experiments showed that CD161 expression on CD4 ϩ T cells is relatively stable, and when this characteristic is combined with the fact that latent HIV-1 is highly enriched in these cells, CD161 is a promising marker that may help to identify latently infected cells and a potential marker for targeting the latent reservoir.
A series of recent studies has indicated that the clonal proliferation of infected T cells carrying replication-competent virus is a major mechanism underlying the long-term stability of the HIV-1 latent reservoir (7,10,49,50). It was reported that functionally polarized Th1 CD4 ϩ T cells carry more clonally expanded genome-intact HIV-1 (7). Th17 cells were shown to possess the characteristics of proliferative self-renewal, survival, and apoptotic resistance (21). Besides, a recent study supported the concept that HIV-1 takes advantage of the long-lived properties of specific Th17 cell subsets to ensure its persistence under ART treatment (13). According to previously recognized hallmarks of Th17 cells, analyses of such Th17 clones revealed the specific expression of CD161 (15). IL-17-producing cells differentiate from CD161 precursors and maintain the expression  of CD161 throughout their life cycle. The C-type lectin-like receptor CD161, which has been described to promote T cell expansion (24), is expressed on a discrete subset of human CD4 ϩ T cells (22). Our results showed that CD161 ϩ CD4 ϩ T cells had higher rates of survival and a higher proliferative ability than their CD161-negative counterparts with TCR or homeostatic signals. Considering the enrichment of replicationcompetent latent HIV-1 and clonal expanded proviral sequences, CD161 ϩ CD4 ϩ T cells may act as a major cellular subset that drives clonal expansion. Computational analysis predicted that the inhibition of infected CD4 ϩ T cell proliferation may decrease the half-life of latently infected cells by up to 20-fold (51). Therefore, perturbation of the clonal expansion of latently infected cells or the CD161 ϩ CD4 ϩ T cell population may provide a promising strategy to counterbalance the repopulation and dissemination of latently infected cells and reduce the size of the latent reservoir, as we have demonstrated here. On the other hand, the application of latency-reversing agents (LRAs) needs to be revisited, since some LRAs, like IL-15 or IL-15 agonist, could actually promote the expansion of the HIV-1 latent reservoir, and we reported this effect on CD161 ϩ CD4 ϩ T cells here. Further evidence is still needed to understand the intrinsic cellular program by which these cells maintain self-renewal and survival. Pharmaceutical inhibition of CD161 ϩ CD4 ϩ T cells or specific signaling pathways needs to be investigated for its ability to restrict the clonal proliferation of infected cells, which may be translated into improved clinical strategies for HIV-1 functional cure and eradication.

MATERIALS AND METHODS
Study subjects. Peripheral blood for the isolation of PBMCs or CD4 ϩ T cells was obtained from 71 HIV-1-infected patients and 20 healthy adult participants. Among them, 53 patients had been on antiretroviral therapy (ART) for at least 2 years and had maintained undetectable plasma HIV-1 RNA levels (Ͻ50 copies per ml) for at least 1 year before blood collection, and their CD4 ϩ counts were more than 450 cells per l. In addition, 18 patients who had recently been diagnosed with HIV-1 infection but who had not started treatment were recruited and followed for 1 year. This study was approved by the Ethics Review Boards of Sun Yat-sen University and the Eighth People's Hospital of Guangzhou, Guangzhou, China. HIV-1-infected patients were recruited at The Eighth People's Hospital, and all were written consent informed. Normal human lymph node tissue obtained during exploration of the abdominal cavity in patients with nonneoplastic acute abdomen at Guangzhou First People's Hospital was used for the cytoimmunofluorescence staining technique under a study protocol approved by the Institutional Review Board of Guangzhou First People's Hospital. The use of PBMCs from healthy adult donors was approved by the Institutional Review Board of Guangzhou Blood Center. The age of the HIV-negative donors included in our study ranged from 18 to 65 years. CD4 and CD8 counts were in the normal range of clinical test values, and none of the donors was infected with HIV-1.
Isolation and culture of primary human T lymphocytes. PBMCs derived from HIV-1-infected patients or healthy donors were isolated by Ficoll-Paque gradient separation. Primary CD4 ϩ T cells were obtained from PBMCs by negative magnetic selection with human CD4 ϩ T lymphocyte enrichment set DM (BD IMag cell separation system). Then, CD161-positive or -negative and bulk CD4 ϩ T cells were sorted by use of a FACSAria cell sorter (BD Biosciences) upon staining with CD3, CD4, and CD161 surface-staining antibodies. Cells were sorted and resulted in the isolation of lymphocytes with the defined phenotypic characteristic of Ͼ95% purity. Data were analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA). The isolated T cells were cultured in basal medium or were stimulated for 3 days with anti-CD3 antibody at 1 g ml Ϫ1 (BioLegend) or anti-CD28 antibody at 1 g ml Ϫ1 (BioLegend) before wild-type HIV-1 infection. CD4 ϩ T cells or PBMCs were stimulated with anti-CD3 (1 g ml Ϫ1 ) or anti-CD3 and anti-CD28 (1 g ml Ϫ1 ) monoclonal antibodies (MAbs), IL-7 (10 ng ml Ϫ1 ), and IL-15 (10 ng ml Ϫ1 ) or with IL-7 and IL-15 only with or without IL-2 (100 U ml Ϫ1 ). All cell culture basal media contained 90% RPMI 1640 supplemented with 10% fetal bovine serum, 100 U ml Ϫ1 penicillin, and 100 g ml Ϫ1 streptomycin (Gibco, Invitrogen, Carlsbad, CA), and cell cultures were maintained in an environment of 37°C and 5% CO 2 .
Intracellular cytokine or HIV-1 p24 staining. To assess intracellular staining for cytokines, PBMCs stimulated with PMA (50 ng ml Ϫ1 )-ionomycin (1 g ml Ϫ1 ) (8 h) and brefeldin A (10 g ml Ϫ1 ) were added at the last 5 h of stimulation to inhibit cytokine release.
Cells were stained with monoclonal antibodies to the surface markers for 30 min on ice in the dark and washed with wash medium twice and then fixed by the use of fixation/permeabilization buffer (BD Biosciences), followed by intracellular cytokine or HIV-1 p24 staining with antibodies directed against intracellular antigen. Details of the antibodies used for flow cytometry are provided in Table S1 in the supplemental material. Cells were analyzed by fluorescence-activated cell sorting (FACS) using a BD LSR Fortessa flow cytometer and FlowJo software (Tree Star Inc., Ashland, OR, USA).
Virus production and in vitro HIV-1 infection. On the day before transduction, HEK293T cells were seeded at 8 ϫ 10 6 cells per 100-mm dish. Twenty-four hours later, virus was generated by transfecting HEK293T cells with a plasmid encoding NL4-3 or NL4-3 BaL, using a polyethylenimine transfection system CD161 ϩ CD4 ϩ T Cells Harbor Expanded Latent HIV-1 ® and following the manufacturer's instructions. Supernatants were harvested after 48 h, centrifuged (10 min, 500 ϫ g, room temperature), and filtered through a 0.45-m-pore-size membrane to remove the cell debris. Viruses were concentrated by centrifuging with a 25% volume of 50% polyethylene glycol 6000 and a 10% volume of 4 M NaCl. Concentrated virions were resuspended in complete medium and stored at Ϫ80°C. The virus concentration was estimated by p24 titration using an enzyme-linked immunosorbent assay (ELISA). Different CD4 ϩ T cell subsets cultured in basal medium or stimulated for 3 days with anti-CD3 antibody at 1 g ml Ϫ1 (BioLegend) and anti-CD28 antibody at 1 g ml Ϫ1 (BioLegend) were infected with NL4-3 Bal or NL4-3 virus (p24 titer, 50 ng ml Ϫ1 ). Infected CD4 ϩ T cells were further cultured in basal medium and incubated at 37°C in a humidified incubator with 5% CO 2 . Then, infected cells were analyzed by flow cytometry at 3 to 7 days postinfection.
Assessment of HIV-1 proviral DNA and replication-competent HIV-1 in patient primary CD161positive or -negative CD4 ؉ T cells. To compare the frequency of proviral DNA, two subsets of cells, including CD161-positive and -negative CD4 ϩ T cells, were purified by FACS. Sorted CD4 ϩ T cell populations were subjected to DNA and RNA extraction using commercial kits purchased from Magen (catalog number R5111-02). Genomic DNA was collected for quantitative PCR using Gag primers and probe, which are described in Table S2. qPCR was performed using the following program: 50°C for 2 min, 95°C for 10 min, and 40 cycles of 94°C for 15 s and 60°C for 1 min. An intact proviral DNA assay (IPDA), which is a quantitative approach for measuring the reservoir of latent HIV-1 proviruses, was used to measure the frequency of latent HIV-1 proviral DNA between CD161-positive and -negative cells (36). Briefly, genomic DNA from these two types of T cells was detected by droplet digital PCR (ddPCR) using a primer/probe mix consisting of oligonucleotides for two independent hydrolysis probe reactions that interrogate conserved regions of the HIV-1 genome to discriminate intact from defective proviruses, The ddPCR was performed on a Bio-Rad QX200 AutoDG digital droplet PCR system (Bio-Rad Laboratories). Simultaneous quantification of DNA shearing and input human genome equivalents was performed using another aliquot of the same DNA sample and two independent hydrolysis probe reactions that interrogate the human RPP30 gene (Table S3). The number of proviral copies per 10 6 CD4 ϩ T cells can be estimated by using the DNA shearing index to correct the raw ddPCR output for RPP30 and HIV-1. To measure the frequency of latent HIV-1, CD161-positive and -negative CD4 ϩ T cells were obtained by FACS and then used for a limiting dilution virus outgrowth assay (37). Briefly, these two T cell subsets were cultured with irradiated PBMCs in the presence of phytohemagglutinin (PHA) and IL-2. The PHA was removed 18 h later. Fresh CD8-negative PBMCs from healthy donors were added to the culture on day 1 and day 7. The culture supernatant was collected at 14 and 21 days for the HIV-1 p24 ELISA.
Sequencing of the env sequence and construction of phylogenetic trees. Genomic DNA was extracted from the cell populations indicated above using a tissue DNA kit (Mega). The DNA was serially diluted 1:3, 1:9, 1:27, and 1:81, and for each dilution, 20 reactions with two rounds of nested PCR were performed using Invitrogen Platinum Taq high-fidelity polymerase and primers specific to the HIV-1 env regions. Briefly, for the first round of PCR, 2 l diluted DNA was amplified in a 40-l reaction mixture containing 1 M of the primers indicated in Table S2, 1ϫ high-fidelity buffer [180 mM (NH 4 ) 2 SO 4 , 2 mM MgSO 4 , 600 mM Tris-SO 4 , pH 8.9, 0.2 mM deoxynucleoside triphosphates; Invitrogen), and 0.025 U/ml Platinum Taq high-fidelity polymerase (Invitrogen) (52). PCR conditions for the first round were 94°C for 2 min and then 94°C for 30 s, 60°C for 1 min, and 68°C for 5 min for 3 cycles; 94°C for 15 s, 60°C for 30 s, and 68°C for 4.5 min for 32 cycles; and then 68°C for 10 min. The first-round PCR product was diluted 1:3 in Tris-HCl (5 mM, pH 8), and 2 l of the diluted reaction mixture was transferred to the 30-l second-round reaction mixture using the primers listed in Table S2. Second-round PCR conditions were 94°C for 2 min; then 94°C for 15 s, 55°C for 30 s, and 68°C for 3 min for 35 cycles; and then 68°C for 10 min. Wells positive for amplified HIV-1 proviruses were identified by diluting the second-round PCR product 1:3 with Tris-HCl, followed by visualization on a 1% agarose gel. According to the Poisson distribution, the dilution at which 30% of the PCRs were positive has an 80% probability of containing a single amplified provirus. Therefore, the dilution at which approximately 30% of the amplicons were positive was selected, and additional PCRs were completed until all the DNA sample from two T cell subsets was amplified. The PCR products were then purified and either sequenced by regular Sanger sequencing or deep sequenced by use of an Illumina MiSeq platform. Neighbor-joining distance analysis was performed in MEGA (version 7) software. The average pairwise distances (APDs) were calculated in MEGA (version 7) software from env proviral single-genome sequences.
Real-time RT-qPCR analysis. Total RNA was isolated with the TRIzol reagent (Life Technologies) and then subjected to cDNA synthesis using an EasyScript One-Step genomic DNA removal and cDNA synthesis supermix kit (TransGen Biotech). Primers were obtained from PrimerBank and are shown in Table S2. The primers were annealed at 37°C, and reverse transcriptase (RT) was processed at 42°C. Quantitative PCR was performed with ChaQ SYBR qPCR master mix (Vazyme Biotech) by following the manufacturer's instructions. Quantification was performed by normalization of the amount to that of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene.
Statistical analysis. Data are presented as the mean Ϯ standard error of the mean (SEM) from at least 4 independent experiments unless indicated otherwise in the figure legends. A Student's t test (paired) was applied to normally distributed data. One-way analysis of variance followed by Bonferroni's correction (when two groups were compared), Dunnett's test (when all experimental groups were compared to one control group), or Tukey's multiple-comparison test (when all groups were compared to each other) was applied in the multiple comparisons after one-way analysis of variance (one-way Li et al. mbio.asm.org 12 ANOVA). Statistical significance was accepted at a P value of Ͻ0.05. Statistical analyses were performed using GraphPad Prism (version 7) software.