HIV-1–Infected CD4⁺ T Cells Present MHC Class II–Restricted Epitope via Endogenous Processing

Despite >40 years of intensive research, HIV-1 continues to pose a significant global health threat. HIV-1 attacks the human immune system, eventually leading to a state of profound immune deficiency known as AIDS. In 2020, an estimated 37.7 million people were living with HIV, 1.5 million were newly infected, and 680,000 died of AIDS-related illnesses (1). HIV-1 mainly infects activated CD4⁺ T cells (T_CD4+s), macrophages, and microglia, and to a lesser extent, dendritic cells (DCs) and osteoclasts (2–4). Infection causes profound immune dysregulation and T_CD4+ death, leaving infected individuals vulnerable to opportunistic infections and death in the absence of therapeutic intervention (5).

T_CD4+s play a critical role in the immune response against a wide variety of viral pathogens (6–8). Previous studies indicate that vigorous, HIV-1–specific T_CD4+ responses play an important role in controlling viremia (9) and are also correlated with broader neutralizing Ab responses in infected individuals (10). Data from chronically infected individuals indicate a link between strong HIV-1–specific T_CD4+ responses and effective HIV-1–specific CD8⁺ T cell (T_CD8+) responses (11), and IL-21–producing T_CD4+s have been shown to be important for maintaining HIV-1–specific T_CD8+s (12). In addition, HIV-1–specific cytotoxic T_CD4+s work cooperatively with HIV-1–specific T_CD8+s to control HIV-1 viremia (13), and expansion of cytolytic T_CD4+s during acute HIV-1 infection predicts disease outcome (14). T follicular helper cells (Tfhs) also play a key role in the anti–HIV-1 immune response. Work by Cohen et al. (15) demonstrated an association between the frequency of Tfh cells during early, untreated infection and the future development of broadly neutralizing Abs. In addition, a longitudinal study in rhesus macaques found that continuous HIV-1 envelope Ag production is required for driving Tfh activation, leading to more effective broadly neutralizing Ab responses (16).

Given the importance of T_CD4+s in host responses to HIV-1, it is crucial to understand the basis for HIV-1–specific T_CD4+ activation. By convention, T_CD4+ activation is mediated by professional APCs—DCs, macrophages, and B cells— that (1) internalize extracellular Ag, such as whole virions; (2) proteolyze internalized material within the endocytic network; (3) load the resulting peptides onto MHC class II (MHC II) molecules in the late endosome; and (4) transport resulting peptide:MHC II complexes to the APC plasma membrane, where they are recognized by TCRs on the surfaces of Ag-specific T_CD4+s (17). Peptide:MHC II–TCR contact, along with costimulatory signals, leads to T_CD4+ activation.

Several studies have shown that MHC II–restricted processing and presentation is more complex than this conventional presentation pathway (18). For example, endogenous processing has been described as a robust alternative to classical processing (19–24), coming into play when an APC becomes productively infected, the nascent viral proteins within the APC providing Ag processing substrates for presentation via MHC II. Endogenous processing has been demonstrated in several viral systems, including measles (25), influenza (24), and, notably, HIV-1 (26). Indeed, we have reported that endogenous processing and presentation is the primary driver of the T_CD4+ response to influenza infection (24).

The cell types capable of MHC II–restricted presentation may also be more extensive than originally appreciated. Many cell types besides professional APCs constitutively express MHC II (23). In addition, many other cell types express MHC II in response to inflammatory signals such as IFN-γ (27, 28). This includes human T_CD4+s, which, unlike their murine counterparts, express MHC II, as well as the costimulatory molecule CD86, on activation (29). Historically, these cells have not been considered as APCs, likely because of their limited ability to internalize extracellular material (a prerequisite for effective classical processing) and their nonconstitutive expression of MHC II. However, the function of MHC II in these cells has been the subject of scrutiny for some time. In addition to their induced MHC II expression, activated human T_CD4+s are a main cellular target of HIV-1, providing a large intracellular depot of newly synthesized viral Ag that could be accessed for endogenous MHC II processing and presentation. A limited number of studies have shown that, under the right circumstances, activated T_CD4+s can present HIV-1–derived Ag on MHC II (30, 31), although T_CD4+s were determined to be inferior to professional APC types, such as B cell lines and monocytes, at stimulating T_CD4+ responses. However, it is worth noting that these studies were performed using (1) inert purified proteins, not infectious HIV-1; and (2) T_CD4+ clones as APCs instead of primary ex vivo T_CD4+s, potentially limiting their relevance to the study of endogenous processing in primary T_CD4+s.

Despite the key role that HIV-1–specific T_CD4+ responses play, there are very few studies dedicated to the details of MHC II–restricted presentation of HIV-1 epitopes in any cell type. We were intrigued by the possibility of endogenous MHC II–restricted processing of HIV-1 Ags by activated human T_CD4+s. Infection would obviate the low internalization capacity of T_CD4+s, and endogenous processing of abundantly produced Ags might lead to potent presentation of HIV-1–derived epitopes. In this study, we asked whether activated human T_CD4+s could present Ag derived from infectious HIV-1 to HIV-1–specific T_CD4+, using DCs, the prototypical professional APC, as a comparator.

Human PBMCs were isolated from healthy HLA-DR1, -DR11, and -DR15⁺ donors by the University of Pennsylvania Human Immunology Core (HIC) by density gradient centrifugation using Lymphoprep (Stemcell Technologies, Vancouver, BC, Canada). T_CD4+s and monocytes were further isolated by the HIC using the RosetteSep Human CD4⁺ T Cell Enrichment Cocktail and EasySep Human Monocyte Enrichment kits (Stemcell Technologies). All studies were conducted with the permission of the University of Pennsylvania Institutional Review Board.

T_CD4+s were cultured in RPMI 1640 (Invitrogen, Waltham, MA) supplemented with 100 U/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), 2 mM l-glutamine (Corning, Corning, NY), 10% heat-inactivated FBS (R&D Systems, Minneapolis, MN), and 10 mM HEPES (Invitrogen). Media were further supplemented with 100 IU/ml recombinant human IL-2 (BioLegend, San Diego, CA). T_CD4+s were activated using K562-based artificial APCs (aAPCs) expressing CD64 and CD86 (32) that were additionally engineered to express a chimeric Ag receptor based on OKT3 Ab irradiated at 100 Gy. T_CD4+s and aAPCs were plated at a 2:1 ratio at a concentration of 1 × 10⁶ cells/ml in a 24-well plate. On day 2, the media volume was quadrupled, and cells were moved to a T25 cell culture flask. Cells were counted at days 4 and 6 and fed with fresh media + 100 IU/ml IL-2 to maintain a cell concentration of 2.5 × 10⁵ cells/ml. We confirmed that K562-activated T_CD4+s expressed typical activation markers using flow cytometry (Supplemental Fig. 1A). To generate DCs, we cultured 5 × 10⁶ monocytes in 3 ml of RPMI 1640 (Invitrogen) supplemented with 100 U/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), 2 mM l-glutamine (Corning), 10% FBS (R&D Systems), and 0.05 mM 2-ME (Sigma-Aldrich, St. Louis, MO). Media were further supplemented with 100 ng/ml recombinant human GM-CSF (BioLegend) and 50 ng/ml recombinant human IL-4 (BioLegend). On days 3 and 6, fresh media supplemented with the same concentration of GM-CSF and IL-4 were added to DC cultures. Unless otherwise indicated, T_CD4+s and DCs were used on day 7.

NL4-R3A (+Nef) has been previously described (33). Nef/Vpu knockout (KO) NL4-R3A and K574D NL4-R3A viruses were also generated by J.A.H.’s laboratory. HIV-1 R3A gp120/41 Env was mutated at position K574D (34) with QuikChange XL Site-Directed Mutagenesis (Agilent Technologies) with primers 5′-CTCACAGTCTGGGGCATCGATCAGCTCCAGGCAAGAGTC-3′ (forward) and 5′-GACTCTTGCCTGGAGCTGATCGATGCCCCAGACTGTGAG-3′ (reverse) to make the K574D fusion-deficient mutant. Mutations were confirmed by sequencing. The Env was then cloned into the pNL4-3 HIV-1 provirus. This virus was used to inoculate T_CD4+s in vitro, and no infected cells were detected by microscopy using an anti-p24 Ab. To make the Nef⁻/Vpu double-KO R3A, the R3A Env was cloned into the pNL4-3 vector (which is Nef⁻ by design); then the Vpu gene was mutated with QuikChange XL (Invitrogen) at the amino acid L12 with the primer introducing a premature stop codon forward 5′-GCAATAGTAGCATAAGTAGTAGCAGCAATAATAGCAATAGTAG-3′. For clarity, the amino acid is underlined: MQSLQILAIVALVVAAIIAIVVWSIALIEYRKILRQRKIDRIINRIIERAEDSGNESEGDQEELSALVEMGHHAPWDINDL*. Aldrithiol-2 (AT2)-inactivated NL4-R3A (+Nef) was obtained from the University of Pennsylvania Center for AIDS Research (CFAR) Viral and Molecular Core. All viruses were made in HEK 293T cells transfected with Lipofectamine 2000 (Invitrogen) and were titered for p24 content by the University of Pennsylvania CFAR Viral and Molecular Core using Alliance HIV-1 p24 Ag ELISA kits (PerkinElmer). Viruses were aliquoted and stored at −80°C until use.

On day 4 of cell culture, T_CD4+s and DCs were infected in their original media with NL4-R3A or Nef/Vpu KO NL4-R3A (100 ng of p24) per 1 × 10⁶ cells in a 48-well plate. Infection was allowed to proceed for 3 d before cell staining or Ag presentation assays were performed. A total of 100 ng p24 AT2-treated NL4-R3A and 69 ng p24 K574D NL4-R3A per 1 × 10⁶ cells were added to T_CD4+s and DCs 6 d postculture, ∼12 h before Ag presentation assays were performed.

Expression of Ag processing and presentation machinery was assessed as follows: DCs were stained extracellularly with anti–HLA-DR (1:100 dilution, G46-6; BD) and anti-CD86 (1:100 dilution, FUN-1; BD). Activated T_CD4+s were stained extracellularly with anti-CD3 (1:200 dilution, UCHT1; BioLegend) and anti-CD4 (1:100 dilution, OKT4; BioLegend), anti–HLA-DR (G46-6; BD), and anti-CD86 (1:100 dilution, FUN-1; BD) for 20 min at 4°C. For intracellular staining, both DCs and T_CD4+s permeabilized with fixation/permeabilization solution (BD Biosciences) for 15 min at room temperature and were then stained with anti-CD74 (1:100 dilution, Pin.1; BioLegend) and anti–HLA-DM (1:100 dilution, MaP.DM1; BioLegend) for 25 min at room temperature. To assess the activation state of aAPC-activated T_CD4+s as compared with anti-CD3/CD28–stimulated T_CD4+s, we stained cells extracellularly with anti-CD4 (OKT4; BioLegend), anti-CD69 (1:100 dilution, FN50; BioLegend), anti–HLA-DR (G46-6; BD), anti-CD25 (1:100 dilution, BC96; BioLegend), and anti-CD62L (1:100 dilution, DREG-56; BioLegend) for 20 min at 4°C. Cells were then stained intracellularly with anti-CD3 (UCHT1; BioLegend) and anti–TNF-α (1:100 dilution, MAb1; BioLegend). For DQ-OVA internalization and processing assays, DCs were stained extracellularly with anti–HLA-DR (G46-6; BioLegend), and activated T_CD4+s were stained with anti–HLA-DR (G46-6; BioLegend), anti-CD4 (OKT4; BioLegend), and anti-CD3 (UCHT1; BioLegend). To validate the HIV-specific TCR expression in transduced T_CD4+s, we stained cells extracellularly with anti-CD4 (OKT4; BioLegend) and anti-TCR Vβ22 (1:5 dilution, IMMU 546; Beckman Coulter, Brea, CA) for 20 min at 4°C. For Ag presentation and viral spread assays, cocultures were stained extracellularly with anti-CD4 (OKT4; BioLegend) and anti-CD69 (FN50; BioLegend) for 20 min at 4°C. Cocultures were stained intracellularly with anti-CD3 (UCHT1; BioLegend), anti–TNF-α (MAb1; BioLegend), anti–IL-2 (1:75 dilution, MQ1-17H12; BD), anti–IFN-γ (1:100 dilution, B27; BD Biosciences), and anti-p24 (1:500 dilution, Kc57; Beckman Coulter). All samples were incubated with LIVE/DEAD Fixable Near-IR (Invitrogen) to exclude dead cells. Cells were fixed with 4% paraformaldehyde before analysis. Cells were acquired on a BD LSR Fortessa (BD Biosciences), and data were analyzed using FlowJo software (Tree Star, Ashland, OR).

To assess the ability of DCs and activated T_CD4+s to internalize and process the fluorescent DQ-OVA substrate, we plated 200,000 DCs or T_CD4+s in 200 μl R10 in a 96-well U-bottom plate. DQ-OVA (Invitrogen) was then added at a concentration of 10 μg/ml. After 2 h at 37°C, cells were washed three times in sterile PBS and stained for flow cytometry as described. To assess the ability of DCs and activated T_CD4+s to internalize FITC-Dextran, we plated 200,000 DCs or T_CD4+s in 200 μl R10 in a 96-well U-bottom plate. FITC-Dextran (Sigma-Aldrich) was then added at a concentration of 1 mg/ml. After 2 h at either 4°C or 37°C, cells were washed three times in sterile PBS and stained for flow cytometry as described.

To determine the cathepsin D, L, and S activities of DCs and activated T_CD4+s, we cultured DCs and T_CD4+s for 7 d as described earlier. DC and T_CD4+ lysates were then made according to the instructions in the SensoLyte 520 Cathepsin D Assay Kit, SensoLyte 520 Cathepsin L Assay Kit, and SensoLyte 520 Cathepsin S Assay Kit (all by Anaspec). Lysates were then assessed for their cathepsin D and L activities according to the manufacturer’s instructions. Lysate from 100,000 cells was assayed per well in a 96-well plate. Fluorescence was measured after 30 min using an Infinite M200 Pro plate reader (Tecan).

Gag293-specific TCR (F5) sequences have been previously described (35, 36) and were synthesized by Genscript (Piscataway, NJ). These TCR sequences were cloned into pTRPE transfer plasmid (37) with the TCRβ and TCRα sequences separated by a furin domain serine-glycine-serine-glycine + T2A (38). A total of 27 μg pTRPE transfer plasmid was mixed with 3 μg codon-optimized Cocal-g plasmid (ATUM, Newark, CA) (39). A total of 18 μg HIV gag-pol packaging plasmid and 18 μg HIV_REV expression plasmid (pTRP Rev) were transfected using Lipofectamine 2000 (Invitrogen) into 293T cells. At 24 and 48 h, supernatants were collected and ultracentrifuged for 2.5 h at 25,000 rpm at 4°C, split into four aliquots, and stored at −80°C until use. A ZnT8-specific TCR was obtained from R. Mallone (INSERM) (40), used as a control TCR, and made according to the earlier protocol.

To generate HIV-specific T_CD4+s, we stimulated T_CD4+s from healthy donors with irradiated aAPCs and IL-2 (BioLegend) as described earlier. Twenty-four hours after stimulation, T_CD4+s were transduced with lentiviruses encoding either a ZnT8 TCR or the Gag293-specific TCR. Cells were expanded and fed as described earlier but were allowed to expand for 10 d, with an additional feeding on day 8. Successful transduction was validated by flow cytometric analysis for TCR Vβ22 (Supplemental Fig. 1B) and by measuring T_CD4+ responses to Gag293 peptide pulsed DCs (Supplemental Fig. 1C). Transduced T_CD4+s were then frozen in freezing media consisting of 10% DMSO and 90% FBS and stored at −80°C until needed.

To assess the ability of DCs and activated T_CD4+s to present HIV-1–derived, MHC II–restricted epitope, we cultured and infected cells as described earlier. On the day of the assay, both uninfected and infected DCs and T_CD4+s were collected, washed with PBS, and incubated with 12.5 μg/ml brefeldin A (BFA; BioLegend) in fresh R10 media for a total of 3.5 h. After an initial 1.5 h of BFA treatment, 0.5 mg/ml purified Gag293 peptide (FRDYVDRFYKTLRAEQASQE) (Genscript) was added to the appropriate wells of uninfected DCs and T_CD4+s. Cells were then washed with BFA-containing PBS and replated in fresh R10 with 12.5 μg/ml BFA. Lentivirus-transduced T_CD4+s expressing either a ZnT8 TCR or a Gag293-specific TCR were labeled with CellTrace Violet (Invitrogen) according to manufacturer’s recommendations. A total of 1 × 10⁵ transduced T_CD4+s and 5 × 10⁴ candidate 50,000 potential APCs were then cocultured for 8 h at 37°C. Cells were then stained for flow cytometry as described earlier. To confirm MHC II restriction, we collected DCs and T_CD4+s, incubated them with BFA, and treated them with purified Gag293 peptide as described earlier. However, after 1 h of peptide incubation, 50 μg/ml MHC II blocking Ab (Tü39; BD Biosciences) or CD71 blocking Ab (CY1G4; BioLegend) was added to relevant wells. After 1 h, cells were washed with BFA-containing PBS and cocultured with transduced T_CD4+s as described earlier. To test the ability of DCs and T_CD4+s to present epitope derived from NL4-R3A (+Nef)-infected T_CD4+ cultures, we collected supernatant from wild-type (WT) HIV-1–infected T_CD4+ cultures 48 h postinfection. This supernatant was either treated with UV light for 1 min to inactivate any infectious virions (validated by a lack of syncytia formation in Sup R5 cells) or was filtered through a 100-kDa Ultra-0.5 ml Centrifugal Filter (MilliporeSigma, Burlington, MA) to remove any infectious virions and large subcellular particles. This treated supernatant was then mixed at a 1:1 ratio with fresh media supplemented with IL-2 and added to uninfected DCs and T_CD4+s 12 h before Ag presentation assays.

For all experiments, except where noted, significance was evaluated by two-tailed paired t test or one-way ANOVA test using PRISM software (GraphPad, San Diego, CA). The Tukey method was used to correct for multiple comparisons. A p value <0.05 was considered significant.

To convert Ag presentation data from percent cytokine secretion to fold induction, we loaded the data into R (41). Data were grouped by cell type, cytokine, and experiment replicate, and summary statistics were calculated. Within each summary statistic group, the mean response rate of the negative condition (DMSO) was determined. Fold changes were calculated as the response to experimental conditions divided by the response to DMSO. The mean and SD of the fold changes were calculated in R.

Expression of proteins associated with Ag processing and presentation, such as HLA-DR, CD86, and invariant chain (Ii), has been previously reported in activated human T_CD4+s (42–44). To verify that activated human T_CD4+s express these proteins in our experimental system, and to compare their expression in T_CD4+s versus DCs from the same donor, we analyzed activated human T_CD4+s and DCs by flow cytometry. DCs express high levels of HLA-DR, CD86, and HLA-DM (both by percent positive and median fluorescence intensity [MFI]) (Fig. 1A, 1B, gating strategies are shown in Supplemental Fig. 4A–C), an unsurprising finding given their role as professional APCs. Activated T_CD4+s also express all these proteins, although at lower levels, supporting the potential for APC function. Because HIV-1 infection has been previously shown to downmodulate Ag processing and presentation machinery in myeloid-derived cells (45, 46), we were also interested to determine whether expression levels of these proteins are maintained in HIV-1–infected T_CD4+s. To identify HIV-1–infected cells, we gated on CD3⁺ cells with downregulated levels of surface CD4, because HIV-1 infection induces CD4 downregulation on the cell surface (47). When uninfected and infected T_CD4+s were compared for expression levels of these proteins, a greater percentage of uninfected T_CD4+s expressed HLA-DR and CD86, but there were no significant differences in the percentage of cells expressing HLA-DM and invariant chain (Fig. 1C). Furthermore, WT HIV-1 infection did not lead to decreased expression levels of these markers as measured by MFI. Indeed, infection slightly increased HLA-DM expression as assessed by MFI (Fig. 1D). We further confirmed that expression of these markers is upregulated on activation, because a far lower proportion of freshly thawed, nonactivated T_CD4+s express these markers (Fig. 1E) when compared with activated T_CD4+s, although MFI was high in the few cells that did express these proteins (Fig. 1F). Based on these data, we conclude that activated T_CD4+s, both uninfected and HIV-1 infected, express well-known components of the Ag processing and presentation machinery.

Endocytic proteolysis is integral to classical Ag processing (48). Therefore, we investigated the activities of representative endosomal proteases, cathepsin D, L, and S, in the lysates of DCs and activated T_CD4+s (Fig. 1E–G). Cathepsin D, a lysosomal aspartyl protease, and cathepsin L, a lysosomal endopeptidase, are necessary for intracellular protein breakdown and turnover and have been shown to play a role in Ag presentation (49–51). Cathepsin S is responsible for the degradation of invariant chain associated with mature MHC II molecules, allowing for the binding of antigenic peptides (52). Although DCs had significantly greater cathepsin L activity across three different donors, in two of three of donors (Fig. 1H), activated T_CD4+s had greater cathepsin D activity than did DCs (Fig. 1G). DCs and T_CD4+s had nearly indistinguishable cathepsin S activities (Fig. 1I). These activities, coupled with the expression of Ag processing and presentation components in (Fig. 1A–D, suggests that activated human T_CD4+s have the potential to be effective APCs during an HIV-1 infection.

We next investigated the ability of DCs and activated T_CD4+s to internalize and proteolyze the fluorescent protease substrate DQ-OVA and the endocytic tracer FITC-Dextran, because internalization of extracellular Ag has been functionally linked to conventional MHC II–restricted processing and presentation (17, 53–56). DQ-OVA is internalized through receptor-mediated endocytosis (57) and fluid-phase micropinocytosis (58), and subsequent endocytic hydrolysis results in a fluorescent signal. Both DCs and activated T_CD4+s internalize and proteolyze DQ-OVA (Fig. 1J, 1K). However, DCs internalize and proteolyze DQ-OVA at far greater levels than activated T_CD4+s, an unsurprising result given the well-documented ability of DCs to sample the external environment at a high rate (17, 53, 55, 56). Given that the fluorescent signal from DQ-OVA requires both internalization and proteolysis, and we can detect protease activity in activated T_CD4+, the low signal likely reflects the modest internalization capabilities of this cell type. We confirmed these findings by examining the internalization of FITC-Dextran, another fluorescent substrate (59), by these two cell types. Similar to our DQ-OVA results, we found that although both DCs and activated T_CD4+s are capable of internalizing FITC-Dextran (Fig. 1L), DCs do so at a much greater level (Fig. 1M). Taken together, these data suggest that activated T_CD4+s have a limited ability to internalize and process exogenous Ags.

Results thus far suggest that T_CD4+s are limited in conventional processing of exogenous Ags because of their modest ability to internalize exogenous Ag. However, we reasoned that, because activated T_CD4+s are highly susceptible to HIV-1 infection, they might be effective at endogenous processing after synthesis of viral proteins (19–24). To explore this possibility, we needed to establish a reliable system to read out epitope production. In initial experiments, we attempted to use activation of HIV-1–specific T_CD4+ clones as the readout; however, we were unable to generate the cell numbers needed for sufficient biological and technical replicates and the experimental variations outlined later. We therefore turned to a lentiviral transduction system similar to that described by Benati et al. (35), in which we attained HIV-1 specificity by transducing T_CD4+s from HIV-1–negative donors with an HIV-1–specific TCR. Extensive searches yielded only three full-length, MHC II–restricted, HIV-1–specific TCR sequences, all specific for the immunodominant Gag293 epitope located within the HIV-1 capsid protein. The HIV-1–specific public TCR used in these studies, termed F5 (35), recognizes the Gag293 epitope in complex with HLA-DR1, -DR11, and -DR15 molecules. HLA-DR1⁺, -DR11⁺, or -DR15⁺ DCs and activated T_CD4+s were infected with WT HIV-1 3 d before the assay. To minimize any nonspecific cytokine secretion from uninfected and HIV-1–infected T_CD4+s over the course of infection, these cells were treated with BFA for 4 h before coculture with HIV-1–specific T_CD4+s and subsequently washed to remove any secreted cytokines. Synthetic Gag293 peptide was added to DCs and activated T_CD4+s 2 h before coculture (2 h into the BFA pretreatment) and was washed out before coculture. HIV-1–specific T_CD4+s were labeled with CellTrace Violet dye to differentiate between the two T_CD4+ populations. Candidate APCs (DCs and activated T_CD4+s) and HIV-1–specific T_CD4+s were cocultured for 8 h in the presence of BFA. Cocultures were then stained for flow cytometric analysis to detect production of IL-2, TNF-α, and IFN-γ by responding T_CD4+s. Notably, we did not observe any differences in presentation patterns depending on identity of the presenting molecule (HLA-DR1, -DR11, or -DR15), although we did note occasional variability in the magnitude of responses depending on the donor and experiment. We measured three different indicators of activation (IL-2, TNF-α, and IFN-γ) to maximize the chances of detecting T_CD4+ activation, because it has been observed that T_CD4+ cytokine profiles may differ depending on the activating conditions used (60).

Results reproducibly showed that DCs present synthetic Gag293 peptide but are unable to present the epitope derived from infectious HIV-1 (Fig. 2A). Conversely, activated T_CD4+s present the Gag293 epitope derived from infectious HIV-1 but are unable to present synthetic Gag293 peptide (Fig. 2B, representative flow plots shown in Supplemental Fig. 4D), even at much higher concentrations of synthetic peptide (Supplemental Fig. 3C) despite high levels of MHC II expression. This phenomenon has been previously observed in other Ag systems (23). Two potential explanations are limited peptide exchange at the cell surface (61–63), or the need for peptide to be internalized by the cell for successful presentation (64), a mechanism that is likely severely compromised in T_CD4+s considering the results shown in (Fig. 1.

We confirmed that the activation of T_CD4+s was epitope specific, because T_CD4+s transduced with a TCR specific for ZnT8, a zinc transporter implicated in type 1 diabetes (40), demonstrated either marginal or undetectable activation in response to infected T_CD4+s (Supplemental Fig. 2A, 2B). In addition, we used an MHC II blocking Ab to confirm that presentation to HIV-1–specific T_CD4+s is MHC II dependent (Fig. 2C). Treatment with a control Ab against the transferrin receptor CD71 did not block presentation by infected T_CD4+s (Supplemental Fig. 2C). Notably, HIV-1 infection leads to CD4 downregulation at the infected T_CD4+ surface, preventing superinfection (47) and likely preventing infected T_CD4+s from presenting material derived from newly internalized virions. These results indicate that T_CD4+s, but not DCs, are capable of presenting Gag293 from infectious HIV-1 in an MHC II–restricted manner, leading to the activation of HIV-1–specific T_CD4+s. These data strongly suggest that activation is a direct consequence of MHC II–restricted epitope presentation.

Given the weak internalization capabilities of activated T_CD4+s (Fig. 1G, 1H) and their susceptibility to productive HIV-1 infection, presentation seemed most attributable to endogenous processing, in which nascent viral Ag, not input virions, constitutes the major processing substrate (24, 26, 65). Thus, we hypothesized that activated T_CD4+s would be unable to present epitope derived from structurally intact, chemically inactivated HIV-1. The same experimental setup as in (Fig. 2 was used in these experiments, but a new experimental condition was introduced: AT2-inactivated HIV-1. AT2-treated HIV-1 was added to uninfected DCs and activated T_CD4+s 12 h before the start of the assay. AT2-treated HIV-1 is still capable of fusing to its receptor, CD4, and its coreceptor, CCR5 or CXCR4, at the cell surface and inserting its nucleocapsid into the cytoplasm (66). However, the virus is unable to uncoat and initiate productive infection. If input virus was responsible for the presentation observed in (Fig. 2, then activated T_CD4+s should be able to present this inactivated virus. As shown in (Fig. 3A and 3B (representative flow plots shown in Supplemental Fig. 4F, 4G), neither DCs nor activated T_CD4+s can detectably present the Gag293 epitope derived from AT2-treated HIV-1, even at a much higher dose (Supplemental Fig. 3A, 3B), supporting the notion that activated HIV-1–infected T_CD4+s use endogenous presentation to activate HIV-1–specific T_CD4+s.

To reinforce this finding, we tested the abilities of DCs and activated T_CD4+s to generate epitope from infectious, but fusion-deficient, HIV-1 (K574D HIV-1). This virus was tested for two main reasons. First, chemical inactivation might have negatively impacted virion processibility, and thus use of a fusion-deficient virus provided an orthogonal approach to reinforce the finding. Second, because HIV-1 fuses at the target cell surface and injects its viral material directly into the cytoplasm (67), HIV-1 proteins may have limited opportunity to access endocytic compartments where conventional MHC II Ag processing occurs (17). Thus, absence of fusion at the cell surface may drive more input virions into endocytic compartments. Again, the same experimental setup was used in these experiments as in (Fig. 2, except that fusion-deficient HIV-1, quantified by p24 content, was compared with fusion-competent HIV-1, both being added to DCs and activated T_CD4+s 12 h before addition of Gag293-specific T_CD4+s. As indicated in (Fig. 3C and 3D (representative flow plots shown in Supplemental Fig. 4H, 4I), activated T_CD4+s are unable to present fusion-deficient HIV-1. DCs are able to present fusion-deficient HIV-1, although the resulting HIV-1–specific T_CD4+ responses are not robust. These data, combined with the lack of presentation of AT2-inactivated HIV-1, strongly suggest that HIV-1 must (1) fuse at the cell surface and (2) initiate productive infection to be processed and presented by activated T_CD4+s. However, in our hands, DCs are reproducibly unable to present the Gag293 epitope from infectious or chemically inactivated HIV-1 and are capable of only modest presentation of fusion-deficient virus.

We have so far observed presentation by activated T_CD4+s only when infectious HIV-1 was used. This could be because of direct, endogenous processing and presentation by the infected T_CD4+s or via “indirect presentation” (68, 69), in which infected T_CD4+s release viral material that is taken up and presented by uninfected, activated T_CD4+s. We subjected supernatant from HIV-1–infected T_CD4+s to either (1) UV light to render the supernatant noninfectious or (2) 100-kDa filtration to remove whole virions and larger subcellular material. DCs and activated T_CD4+s were incubated with treated supernatant for 12 h before the assay. We found that DCs were able to present epitope from UV-treated supernatant, indicating that they are capable of MHC II–restricted indirect presentation (Fig. 4A, representative flow plots shown in Supplemental Fig. 4J, 4K). This result is consistent with previous studies showing DCs are capable of indirect presentation (68, 69). Interestingly, DCs were unable to present filtered supernatant (Fig. 4A), and thus may require relatively large subcellular material, such as exosomes (69) or apoptotic bodies (70), for successful indirect presentation of the Gag293 epitope. In contrast, activated T_CD4+s are capable of only modest presentation of UV-treated supernatant (Fig. 4B), indicating that indirect presentation does not play a major role in the presentation of HIV-1–derived epitope by activated T_CD4+s.

In addition to addressing the question of indirect presentation, the results provide additional support for our conclusion that activated T_CD4+s carry out endogenous processing. Supernatants were collected from activated T_CD4+s 72 h postinfection, the same duration of infection before addition of responding T_CD4+s. Lack of a robust level of presentation of the UV-treated supernatant by activated T_CD4+s (Fig. 4) argues against the presentation of exogenous material in the form of extracellular virions or subviral material and in favor of bona fide endogenous processing and presentation.

The viral accessory proteins Nef and Vpu have been reported to disrupt MHC II–restricted Ag presentation in myeloid-derived cells (45, 71, 72). We investigated whether these proteins contribute to the lack of detectable presentation of live HIV-1 by DCs. To test this, and to determine the impact of these proteins on MHC II–restricted presentation by activated T_CD4+s, we used a Nef/Vpu double-KO HIV-1. Infection with the Nef/Vpu double-KO virus was performed using the same method as in (Fig. 2. Using this virus, we could not detect any impact of Nef or Vpu on presentation of the Gag293 epitope by either DCs or activated T_CD4+s (Fig. 5A, 5B, representative flow plots are shown in Supplemental Fig. 4L, 4M). DCs were still unable to present the Gag293 epitope from infectious Nef/Vpu KO virus (Fig. 5A), and there were no significant differences in the presentation of WT and double-KO HIV-1 by activated T_CD4+s (Fig. 5B).

Our study reveals that activated T_CD4+s express several key components of the Ag processing and presentation machinery (Fig. 1A, 1B), and this expression remains intact in HIV-1–infected cells (Fig. 1C, 1D). We also show that activated T_CD4+s express functional endosomal proteases (Fig. 1E, 1F), although their ability to internalize exogenous Ag is restricted, as demonstrated by their limited ability to unquench DQ-OVA (Fig. 1G, 1H) and internalize FITC-Dextran (Fig. 1L, 1M). Taken together, these data suggest that activated T_CD4+s have the ability to process and present Ag on MHC II but are limited with respect to classical processing and presentation by their modest ability to internalize extracellular Ag.

Importantly, we observed that activated T_CD4+s are able to present Ag derived from infectious HIV-1 to HIV-1–specific T_CD4+s, eliciting a strong cytokine response (Fig. 2A, 2B). This presentation is MHC II dependent, because the stimulation of IL-2 secretion by HIV-1–specific T_CD4+s can be inhibited with MHC II blocking Ab (Fig. 2C) and is Ag-specific, as ZnT8 TCR-transduced T_CD4+s experienced either far weaker or undetectable activation in response to HIV-1–infected T_CD4+s (Supplemental Fig. 2A, 2B). We further confirmed that T_CD4+s can present HIV-1–derived epitope only when they are productively infected, because these cells are unable to present chemically inactivated virus (Fig. 3B) or fusion-deficient virus (Fig. 3D) and are unable to indirectly present epitope from supernatants of HIV-1–infected cells (Fig. 4B). We also note that activated T_CD4+s are unable to present synthetic peptide, even at high concentrations (Supplemental Fig. 3C). This phenomenon has been previously observed in other Ag systems (23). Two potential explanations are limited peptide exchange at the cell surface (61–63), or the need for peptide to be internalized by the cell for successful presentation (64), which T_CD4+s appear to be incapable of performing. These data led us to conclude that activated T_CD4+s use endogenous processing to stimulate HIV-1–specific T_CD4+s. Thus, cell types with only a marginal capacity for classical processing and presentation can nevertheless be highly functional APCs via viral infection and endogenous processing and presentation. How far this paradigm extends beyond HIV-1 and T_CD4+s remains to be seen. To date, most endogenous processing studies in primary cells focus on professional APCs. However, Toulmin et al. (23) recently demonstrated that type II alveolar cells, epithelial cells located in the distal lung, express MHC II and use endogenous processing to present influenza proteins and the endogenous E protein. These findings, along with ours, suggest that nonprofessional APCs may play key roles in MHC II–restricted presentation via endogenous Ag processing during viral infections.

Previous studies have reported that activated T_CD4+s are capable of presenting HIV-1 gp120-derived epitope (30, 31). However, to our knowledge, ours appears to be the first demonstration of MHC II–restricted presentation by activated T_CD4+s of epitope derived from viable HIV-1 virions. These results suggest a novel function for T_CD4+s during the immune response to HIV-1, although these findings will need to be confirmed in vivo. In addition to orchestrating the adaptive immune response to HIV-1, our data suggest that T_CD4+s may also act as APCs, leading to the activation of HIV-1–specific T_CD4+s, although this presentation could potentially come at the price of enhanced viral spread to HIV-1–specific T_CD4+s (73).

Previous studies have shown that monocyte-derived DCs are capable of presenting the Gag293 epitope from infectious HIV-1 (36); however, our results are not necessarily in conflict with this work. Galperin et al. (36) focused on T_CD4+-mediated killing of infected DCs as a readout of T_CD4+ activation, while we focused on cytokine secretion. It is possible that the presentation of infectious HIV-1 by DCs leads to cytotoxicity and not cytokine production in responding T_CD4+s, which could explain the differences in our results. Notably, DCs can present the Gag293 epitope when exposed to UV-treated, HIV-1–infected supernatant (Fig. 4A), indicating that indirect presentation is a viable production route for this particular epitope in DCs. We focused on the Gag293 epitope because of the limited number of available full-length, MHC II–restricted HIV-1–specific TCR sequences. It will be interesting to determine whether our observations extend to other MHC II–restricted HIV-1 epitopes. Based on our previous work with influenza (24) (in which all six MHC II–restricted epitopes are endogenously presented), we anticipate that Gag293 is representative of other epitopes, although the processing pathways could be quite different.

We also showed that the HIV-1 accessory proteins Nef and Vpu do not impact presentation of the Gag293 epitope by either DCs or activated T_CD4+s (Fig. 5A, 5B). Again, the different processing requirements for different epitopes (24, 65, 74, 75) may explain the apparent contradiction with previous studies, which showed that Nef and Vpu disrupt MHC II–restricted presentation in DCs (45, 71, 72). In addition, although we can conclude that Nef and Vpu do not interfere with the presentation of the Gag293 epitope in DCs and activated T_CD4+, we are unable to speculate on the effects of these proteins on the presentation by other APC types.

In summary, our results indicate that activated, HIV-1 infected T_CD4+s are highly effective at MHC II–restricted presentation of HIV-1–derived epitopes due to the tropism of HIV-1 for T_CD4+s and the potency of endogenous MHC II–restricted processing and presentation. Given the critical importance of HIV-1–specific T_CD4+s, these results could significantly alter our understanding of the MHC II–restricted processing and presentation landscape that develops during HIV-1 infection.

We are grateful to the following groups for assistance with this project: the University of Pennsylvania HIC (supported by AI-045008 and CA-016520) for invaluable assistance in cell procurement and processing; the University of Pennsylvania CFAR Viral and Reservoirs Core (supported by P30 AI 045008) for help with HIV-1 titering and HIV-1 inactivation; (3) Roberto Mallone (INSERM) for supplying the ZnT8-specific TCR sequence; and (4) the Children’s Hospital of Philadelphia Flow Cytometry Core for expertise and patience. The authors are also grateful to John Paul Bisciotti for assistance and expertise with data transformation and statistical analysis. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

J.L.R. is an equity holder of and receives sponsored research funding from Tmunity Therapeutics. The other authors have no financial conflicts of interest.