Transient T Cell Expansion, Activation, and Proliferation in Therapeutically Vaccinated Simian Immunodeficiency Virus-Positive Macaques Treated with N-803

ABSTRACT

Vaccine strategies aimed at eliciting human immunodeficiency virus (HIV)-specific CD8⁺ T cells are one major target of interest in HIV functional cure strategies. We hypothesized that CD8⁺ T cells elicited by therapeutic vaccination during antiretroviral therapy (ART) would be recalled and boosted by treatment with the interleukin 15 (IL-15) superagonist N-803 after ART discontinuation. We intravenously immunized four simian immunodeficiency virus-positive (SIV⁺) Mauritian cynomolgus macaques receiving ART with vesicular stomatitis virus (VSV), modified vaccinia virus Ankara strain (MVA), and recombinant adenovirus serotype 5 (rAd-5) vectors all expressing SIVmac239 Gag. Immediately after ART cessation, these animals received three doses of N-803. Four control animals received no vaccines or N-803. The vaccine regimen generated a high-magnitude response involving Gag-specific CD8⁺ T cells that were proliferative and biased toward an effector memory phenotype. We then compared cells elicited by vaccination (Gag specific) to cells elicited by SIV infection and unaffected by vaccination (Nef specific). We found that N-803 treatment enhanced the frequencies of both bulk and proliferating antigen-specific CD8⁺ T cells elicited by vaccination and the antigen-specific CD8⁺ T cells elicited by SIV infection. In sum, we demonstrate that a therapeutic heterologous prime-boost-boost (HPBB) vaccine can elicit antigen-specific effector memory CD8⁺ T cells that are boosted by N-803.

IMPORTANCE While antiretroviral therapy (ART) can suppress HIV replication, it is not a cure. It is therefore essential to develop therapeutic strategies to enhance the immune system to better become activated and recognize virus-infected cells. Here, we evaluated a novel therapeutic vaccination strategy delivered to SIV⁺ Mauritian cynomolgus macaques receiving ART. ART was then discontinued and we delivered an immunotherapeutic agent (N-803) after ART withdrawal with the goal of eliciting and boosting anti-SIV cellular immunity. Immunologic and virologic analysis of peripheral blood and lymph nodes collected from these animals revealed transient boosts in the frequency, activation, proliferation, and memory phenotype of CD4⁺ and CD8⁺ T cells following each intervention. Overall, these results are important in educating the field of the transient nature of the immunological responses to this particular therapeutic regimen and the similar effects of N-803 on boosting T cells elicited by vaccination or elicited naturally by infection.

INTRODUCTION

CD8⁺ T cells play a critical role in controlling replication of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) through cytolytic and noncytolytic means, both in the presence and in the absence of antiretroviral therapy (ART) (1, 2). Indeed, spontaneous control of HIV/SIV replication without ART, known as elite control (EC), is associated with the higher frequency and prolonged survival of CD8⁺ T cells and with noncytolytic virus inhibition by CD8⁺ T cells (3, 4). While HIV and SIV replication can be durably suppressed by ART, under most circumstances virus replication rebounds when ART is discontinued (5, 6). Thus, a core strategy of HIV cure initiatives is inducing potent CD8⁺ T cells through therapeutic vaccinations to enable people with HIV to stop ART and durably control virus replication.

Expansion of CD8⁺ T cells during acute HIV/SIV infection is associated with peak viral load decline (7–9). During the progression of untreated infection, however, chronic antigen exposure leads to the loss of CD8⁺ T cell function and increased CD8⁺ T cell apoptosis (4, 10). Potent CD8⁺ T cell responses can also put pressure on viral evolution and, unfortunately, lead to viral escape (11, 12). One major challenge for therapeutic vaccines is to supersede naturally elicited cellular immunity, suppress virus replication without ART, and avoid the selection of viral variants.

Many current HIV vaccine strategies can induce neutralizing antibodies and/or CD8⁺ T cells, but it is essential to further understand the features of protective cellular immune responses and determine how best to elicit those responses (13–17). Heterologous prime-boost (HPB) vaccines induce superior cellular immunity compared to single-dose strategies and homologous prime-boost strategies (18, 19). HPB vaccine strategies elicit abundant CD8⁺ T cells with enhanced virus suppression, proliferative capabilities, and long-term survival. These vaccines also expand the populations of central memory (T_CM) and effector memory (T_EM) T cells (15, 19–22). Antigen-specific CD8⁺ T cells with T_EM phenotypes have exhibited high proliferative capacities and inhibitory effects against HIV and have specifically been implicated in EC (13, 21).

Numerous therapeutic vaccine regimens have demonstrated the potential to elicit anti-SIV cellular immunity in macaque models, particularly both T_CM and T_EM responses (23–25). Transient antigen exposure provided by nonreplicating vectors favors the induction of T_CM responses. T_CM expansion and production of effector cells are important for the anti-HIV immune response, but T_CM populations require time to respond via anamnestic expansion upon antigen recognition (26). Replicating vectors that provide recurrent antigen exposure, on the other hand, favor the induction of protective T_EM responses, which can deliver prompt effector function without antigen restimulation (27).

Therapeutic HIV vaccines could improve the host antiviral immune response and benefit those already infected with HIV (28). In ART-treated SIV⁺ macaques, several vaccine modalities have induced high-frequency cellular immune responses against multiple SIV proteins (i.e., Gag, Pol, and Env) (23, 29, 30). High-magnitude Gag-specific responses have also been detected in therapeutic vaccine strategies delivering Gag immunogens only (24, 25). T cell recognition of multiple epitopes in Gag has been associated with lower viremia, suggesting that Gag may be a promising immune target for vaccine strategies (24, 31–33). Some vaccine strategies have conferred improved viral control after suspending ART, but success is limited (23, 24, 29). This may be a consequence of eliciting too few functional antigen-specific T cells in the correct locations to suppress the rebounding virus when ART is interrupted. A primary objective of therapeutic vaccines is to educate the host cellular immune system to develop characteristics associated with EC, such as T_EM phenotype, proliferative capacity, high-magnitude responses, and antigen specificity. A heterologous prime-boost-boost (HPBB) strategy is one novel approach that may achieve this goal.

Cytokine administration can be used as an adjuvant for T cell-based therapies (34, 35). The proinflammatory cytokine interleukin 15 (IL-15) is a critical regulator and promoter of memory T cell development and maintenance (36, 37). IL-15 is involved in NK cell and T cell generation, homeostasis, maturation, and support (38, 39). N-803 (formerly Alt-803) is an IL-15 superagonist complex exhibiting superior half-life, bioactivity, efficacy, and tissue retention compared to IL-15. N-803 has been shown to increase the frequency, activation, proliferation, and function of CD4⁺ T cells, CD8⁺ T cells, and NK cells in mice and macaques (40–42). In SIV⁺ macaques, N-803 can activate and expand CD8⁺ T cells and NK cells and direct these cells to the lymph nodes (LNs) (43–45). However, the impact of N-803 on CD8⁺ T cells elicited by therapeutic HIV/SIV vaccines is unknown.

In the present study, we hypothesized that combining HPBB vaccination with N-803 treatment could elicit and recall a high volume of Gag-specific CD8⁺ T cells and target them to the LNs. We vaccinated four SIV⁺ ART-treated Mauritian cynomolgus macaques (MCMs) with an HPBB regimen consisting of a vesicular stomatitis virus (VSV) prime followed by two boosts with modified vaccinia virus Ankara strain (MVA) and recombinant adenovirus serotype 5 (rAd-5), all expressing SIVmac239 Gag. Four unvaccinated SIV⁺ ART-treated macaques served as controls. Six weeks after the third vaccine, we discontinued ART for all eight animals. The vaccinated animals received three doses of N-803. We report that this therapeutic regimen enhanced the frequency of Gag-specific lymphocytes with phenotypes associated with activation (CD69), proliferation (Ki-67), and memory (T_EM) in the peripheral blood and LNs of the vaccinated animals.

RESULTS

Animal study design, infection, animals, and viral loads.

We infected eight MCMs intravenously (i.v.) with SIVmac239M (Fig. 1A). All animals expressed at least one copy of the M3 major histocompatibility complex (MHC) haplotype, which is associated with a high SIV viral load set point, and none possessed the M1 haplotype, which is associated with SIV control (Table 1) (46, 47). Animals exhibited peak viremia between 3 × 10⁵ and 4 × 10⁷ viral copy equivalents (cEq)/mL on day 11 postinfection (Fig. 1B). Animals began a daily ART regimen of dolutegravir (DTG), tenofovir disoproxil fumarate (TDF), and emtricitabine (FTC) 2 weeks postinfection. The 2-week point was chosen because in rhesus macaque studies, this allows time for seroconversion, reservoir establishment, and the generation of a T cell response prior to virus suppression but not enough time to enable viral immune escape and immune exhaustion (48, 49). Viremia was suppressed below the limit of detection in all 8 animals within 42 days of starting ART. Four animals (cy1035, cy1036, cy1039, and cy1043) were therapeutically vaccinated i.v. with an HPBB vaccine strategy under the control of ART, beginning ~4.5 months after infection (Fig. 1A, blue). Animals first received VSV expressing SIVmac239 Gag, followed by boosts with MVA expressing SIVmac239 Gag-Pol and rAd-5 expressing SIVmac239 Gag. Inoculations were separated by 4 weeks. These vectors have been shown to elicit robust antigen-specific CD8⁺ T cells (15, 29). Four animals (cy1037, cy1040, cy1044, and cy1045) were left unvaccinated as controls and received saline injections rather than vaccine vectors (Fig. 1A, red).

FIG 1.

Experimental design and longitudinal plasma viral loads. (A) Study design depicting timeline of SIV infection, therapeutic vaccination, ART release, and N-803 delivery. Eight male MCMs were infected i.v. with 10,000 IU of SIVmac239M. All eight began receiving ART 14 days postinfection. Four MCMs (blue; vaccinated animals) were sequentially immunized with recombinant heterologous viral vectors VSV, MVA, and rAd-5, each encoding SIVmac239 Gag. ART was discontinued in all eight animals 6 weeks after the final boost. The vaccinated animals, but not the control animals, received 3 doses of N-803 beginning 3 days after ART release. (B) Individual plasma viral loads from SIV inoculation through vaccination.

TABLE 1.

Animals used in this study

Animal ID	MHC haplotype	Treatment group
cy1035	M3/M4	Vaccinated + N-803 treated
cy1036	M3/M5	Vaccinated + N-803 treated
cy1039	M3/M3	Vaccinated + N-803 treated
cy1043	M2/recM3M4a	Vaccinated + N-803 treated
cy1037	M3/recM3M4a	Control
cy1040	M3/M3	Control
cy1044	M2/M3	Control
cy1045	M3/M5	Control

^aExpressed the major MHC class I A and B and MHC class II DRB and DQ alleles present in the M3 MHC haplotype, but also expressed minor MHC class II DP alleles of the M4 MHC haplotype.

Therapeutic i.v. HPBB vaccination elicits robust and broad Gag-specific T cell responses.

To assess the immunogenicity of the HPBB vaccination, we performed gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assays using PBMCs collected before the first inoculation and 1 week after each vaccination, as well as 3 weeks after the rAd-5 vaccination. We stimulated the peripheral blood mononuclear cells (PBMCs) with Mafa-A1*063-restricted Gag_386-394GW9 or Nef_103-111RM9 peptides, or a Gag peptide pool spanning the SIVmac239 Gag proteome. PBMCs from vaccinated animals responded to stimulation with the Gag pool and Gag_386-394GW9 peptide throughout the vaccine phase (Fig. 2A, left). The immunodominant CD8⁺ T cell response targeting the Gag_386-394GW9 epitope was consistent with studies of other SIV⁺ Mafa-A1*063⁺ MCMs (47, 49). As anticipated, the CD8⁺ T cell responses against the Gag pool or Gag_386-394GW9 were substantially lower in the control group (Fig. 2A, right). CD8⁺ T cell responses specific for the Nef_103-111RM9 peptide were low across both groups of animals. Further, the Gag_386-394GW9 IFN-γ ELISPOT assay responses in the vaccinated animals were ~10- to 50-fold greater than the Nef_103-111RM9 responses, while the Gag_386-394GW9 responses in the control animals were approximately the same as the Nef_103-111RM9 responses (Fig. 2A; also, see Table S1 in the supplemental material). These data demonstrate the difference between the frequency of vaccine-elicited versus naturally elicited antigen-specific T cells in ART-suppressed SIV⁺ animals.

FIG 2.

High response magnitude, breadth, and proliferation of Gag-specific CD8⁺ T cell in the PBMC induced by HPBB vaccination. (A) IFN-γ ELISPOT assays were performed longitudinally at the indicated time points to assess responses to Gag_386-394GW9, Nef_103-111RM9, and a Gag peptide pool spanning the SIVmac239 Gag proteome in the vaccinated (left) and control (right) animals. Values are medians. (B) IFN-γ ELISPOT assays were performed longitudinally at the indicated time points to assess responses to Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9 in the vaccinated (left) and control (right) animals. Values are medians. Empty bars indicate time points where the median was zero; there are no missing data points. (C) Number of Gag_386-394GW9 tetramer-positive CD8⁺ T cells per microliter of blood throughout the vaccine phase in the vaccinated (blue) and control (red) animals. (D) Frequency of Gag_386-394GW9 tetramer-positive CD8⁺ T cells (within the parent CD8⁺ T cell population) throughout the vaccine phase in the vaccinated (blue) and control (red) animals. (E) Frequency of CD8⁺ T cells that are both Gag GW9 specific and Ki-67⁺ (within the parent CD8⁺ T cell population) throughout the vaccine phase in the vaccinated (blue) and control (red) animals. (F) Flow cytometry dot plots illustrating expansion of proliferating (Ki-67⁺) Gag GW9-specific CD8⁺ T cells (within the parent CD8⁺ T cell population) after VSV-, MVA-, and rAd-5-SIVmac239 Gag immunizations. Results are displayed for each animal individually. *, P = 0.0286. P values were calculated using Mann-Whitney U tests comparing the vaccinated and control groups at each indicated point.

Next, we used cryopreserved PBMCs to assess T cell responses against SIV epitopes restricted by other MHC class I alleles expressed by the M3 MHC haplotype (Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9) (50). Some control animals exhibited low-level (median = 0 spot-forming cells [SFCs] per 10⁶ PBMCs after background subtraction) responses to Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9 at each time point examined (Fig. 2B, right; Table S2). The vaccinated animals exhibited a slightly higher frequency of IFN-γ⁺ responses (median = 14 to 73 SFCs per 10⁶ PBMCs) to Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9 than the control animals throughout the vaccine phase, suggesting that vaccination elicited a population of T cells targeting a greater diversity of Gag peptides (Fig. 2B, left; Table S2).

Therapeutic i.v. HPBB vaccination increases the frequency of Gag-specific CD8⁺ T cells.

We next assessed the frequency and phenotypes (see Fig. S1) of Gag_386-394GW9 tetramer-positive CD8⁺ T cells (referred to here as Gag GW9-specific CD8⁺ T cells) longitudinally in PBMCs collected prior to and during vaccination. The number of Gag GW9-specific CD8⁺ T cells was significantly higher in the vaccinated animals after the MVA and rAd-5 vaccines than in the control animals (Fig. 2C, asterisks). The frequency of Gag GW9-specific CD8⁺ T cells in some of the vaccinated animals was higher after each vaccine than in the control animals, but this was not statistically significant, likely because the vaccines also drastically increased the number and frequency of bulk CD8⁺ T cell populations (Fig. 2D and Fig. S2). We also compared the frequency of Ki-67 and Gag GW9 tetramer double-positive CD8⁺ T cells (Fig. 2F, top right quadrants) between the vaccinated and control animals after each vaccine. The frequency of these cells was significantly higher in the vaccinated animals after the MVA and rAd-5 vaccines than in the control animals (Fig. 2E). The Ki-67⁺ Gag GW9-specific CD8⁺ T cells remained less than 0.25% in the control animals throughout the study (Fig. 2E, red). Similarly, the frequency of bulk CD8⁺ T cells and CD8⁺ T_EM expressing the proliferation marker Ki-67 alone or in combination with the activation marker CD69 increased in the vaccinated animals 1 week after each of the three vaccines, and the frequency of bulk CD4⁺ T cells expressing Ki-67 or Ki-67 and CD69 increased in the vaccinated animals 1 week after the second vaccine (Fig. S2, blue animals).

T_EM phenotype bias in Gag-specific CD8⁺ T cells elicited by therapeutic i.v. HPBB vaccination.

The frequencies of Gag GW9-specific CD8⁺ T_CM and T_EM cells (Fig. 3A) were compared between the vaccinated and control animals 1 week after each of the three vaccinations. The frequencies of Gag GW9-specific T_CM and T_EM cells were similar between the treatment and control animals 1 week after administration of VSV or the first saline dose (Fig. 3B). However, 1 week after the second and third injections, we detected an increased frequency of Gag GW9-specific T_EM cells in the vaccinated animals, while the proportion of Gag GW9-specific T_EM cells declined in the control group. These results suggest that the vaccine regimen biased antigen-specific CD8⁺ T cells to an effector memory phenotype. This T_EM bias plateaued 3 weeks after the final boost. The proportion of Gag GW9-specific T_EM cells in the vaccinated animals waned slightly, remaining similar to the T_EM proportion after the second vaccination, but there were too few Gag GW9-specific CD8⁺ T cells in the control animals to be evaluated for memory phenotype markers 3 weeks after the final boost (data not shown).

FIG 3.

Gag GW9-specific CD8⁺ T_CM and T_EM cells in PBMCs throughout vaccination. (A) Representative flow cytometry dot plots from a vaccinated animal (top) and a control animal (bottom) illustrating the frequency of Gag GW9-specific CD8⁺ T_CM cells (top gate, CD28⁺ CD95⁺) and T_EM cells (bottom gate, CD28⁻ CD95⁺) (within the parent Gag GW9 tetramer-positive CD8⁺ T cell population) after rAd-5-SIVmac239 Gag immunization. (B) Stacked bar plots showing the median frequencies of Gag GW9-specific T_CM cells (gray) and Gag GW9-specific T_EM cells (black) in the peripheral blood of the vaccinated (left) and control (right) animals 1 week after each vaccination.

Gag-specific AIM expression increases following therapeutic i.v. HPBB vaccination.

To further characterize the functional profile of vaccine-elicited Gag-specific T cells, we used the activation-induced marker (AIM) assay to measure the expression of effector molecules by CD4⁺ and CD8⁺ T cells after in vitro Gag stimulation. In this assay, we used cells collected before the first vaccination and 3 weeks after the last vaccination (Fig. 4) and quantified the expression of CD25 and CD69 (which indicate early activation), CD137 (which marks antigen recognition), and CD107a (which is associated with cytokine secretion and degranulation) (51–55). Prior to vaccination, there were no significant differences in the frequency of cells expressing activation markers between unstimulated and Gag-stimulated conditions (Fig. 4, left), though values from the unstimulated conditions tended to be slightly higher than those from Gag-stimulated conditions. The frequency of cells, particularly CD4⁺ T cells, producing activation markers in the unstimulated conditions was also generally higher before vaccination than after vaccination. We suspect that these observations were due to residual immune activation from acute SIV infection (56, 57). After vaccination, however, we observed distinct populations of CD4⁺ and CD8⁺ T cells expressing activation markers after being stimulated with the Gag peptide pool (Fig. 4, right). These results support the idea that the vaccine regimen expanded the repertoire of Gag-specific T cells with the potential to respond to antigen and exhibit antiviral immune functions.

FIG 4.

Vaccination enhances the frequency of CD4⁺ and CD8⁺ T cells in PBMCs that express activation-induced markers in response to Gag stimulation. The frequency of CD4⁺ T cells double-positive for CD25 and CD69 (within the parent CD4⁺ T cell population), and CD8⁺ T cells double-positive for CD25 and CD69, CD25 and CD107a, and CD25 and CD137 (within the parent CD8⁺ T cell population) are plotted from time points before vaccination (left) and 3 weeks after the final vaccination (right). Frequencies are compared between unstimulated conditions (open circles) and stimulation with a Gag peptide pool (closed circles) in vitro. P values were calculated using paired T-tests.

N-803 treatment is associated with increased Ki-67⁺ SIV-specific CD8⁺ T cells and continued T_EM bias of vaccine-stimulated CD8⁺ T cells.

We wanted to determine whether N-803 could enhance SIV-specific vaccine-elicited T cells compared to those naturally elicited by infection. ART was discontinued in all eight animals 6 weeks after the last vaccine or saline administration. The four vaccinated animals, but not the control animals, received three subcutaneous doses of N-803, 0.1 mg/kg of body weight, separated by 2 weeks each, beginning 3 days after ART cessation (Fig. 1A). We used Gag_386-394GW9 and Nef_103-111RM9 tetramers to compare vaccine-elicited to naturally elicited CD8⁺ T cells, respectively. The Nef_103-111RM9 tetramer-positive CD8⁺ T cells (referred to here as Nef RM9-specific CD8⁺ T cells) could be elicited only by SIV infection, as Nef was not included in the vaccine. One clear caveat is that some Gag GW9-specific CD8⁺ T cells were also elicited by SIV infection, in addition to the vaccine-induced Gag-responsive cells. Nonetheless, distinguishing between Gag GW9- and Nef RM9-specific CD8⁺ T cells represents one way to compare vaccine-elicited and naturally elicited CD8⁺ T cells within the same animals.

In the peripheral blood, the frequency of Gag GW9-specific CD8⁺ T cells was generally higher than that of Nef RM9-specific CD8⁺ T cells following vaccination, both before and after ART withdrawal and throughout N-803 treatment (Fig. 5A), as measured by the total area under the curve (AUC) for the two populations throughout N-803 treatment. Because small sample sizes precluded statistical analyses of these paired data, AUC was used to visually represent the biological trends despite interanimal variability. The AUC for the Ki-67⁺ Gag GW9-specific CD8⁺ T cells was similar to the AUC for the Ki-67⁺ Nef RM9-specific CD8⁺ T cells (Fig. 5B). This suggests that N-803 enhances the proliferation of antigen-specific CD8⁺ T cells indiscriminately whether those cells were elicited by vaccination or by natural infection. We then compared the total AUC of the frequencies of the Gag GW9- and Nef RM9-specific CD8⁺ T_CM and T_EM cells (gated on CD28⁺ CD95⁺ or CD28⁻ CD95⁺ cells within the tetramer-positive parent population, respectively). Although some time points had to be excluded due to low numbers of cells, the AUC of Nef RM9-specific CD8⁺ T_CM cells was higher than that of Gag GW9-specific CD8⁺ T_CM cells (Fig. 5C) and the AUC of Gag GW9-specific CD8⁺ T_EM cells was higher than that of the Nef RM9-specific CD8⁺ T_EM cells (Fig. 5D). While in general the Gag-specific CD8⁺ T cells remained T_EM and the Nef-specific CD8⁺ T cells remained T_CM, we detected no N-803-mediated changes in Gag- or Nef-specific CD8⁺ T cell frequencies or the distribution of T_CM or T_EM phenotypes. N-803 also transiently increased the number and frequency of bulk, Ki-67⁺, CD69⁺, and Ki-67⁺ CD69⁺ CD4⁺ T cells, CD8⁺ T cells, and CD4⁺ and CD8⁺ T_CM, T_TM, and T_EM populations in the peripheral blood of the vaccinated animals (Fig. S3 and S4).

FIG 5.

Changes in Gag GW9- and Nef RM9-specific CD8⁺ T cells in the PBMC of the vaccinated animals (blue shades) or control animals (red shades) following ART interruption. (A) Frequency and AUC of Gag GW9-specific CD8⁺ T cells (light blue) compared to Nef RM9-specific CD8⁺ T cells (dark blue) (gated on tetramer-positive cells within the parent CD8⁺ population) throughout the N-803 phase in the vaccinated animals. (B) Frequency and AUC of proliferating (Ki-67⁺) Gag GW9- or Nef RM9-specific CD8⁺ T cells throughout the N-803 phase in the vaccinated animals (gated on Ki-67⁺ cells within the tetramer-positive parent population). (C) Frequency and AUC of Gag GW9- compared to Nef RM9-specific CD8⁺ T_CM (gated on CD28⁺ CD95⁺ cells within the tetramer-positive parent population) throughout the N-803 phase. (D) Frequency and AUC of Gag GW9- compared to Nef RM9-specific CD8⁺ T_EM (gated on CD28⁻ CD95⁺ cells within the tetramer-positive parent population) throughout the N-803 phase. (E) Frequency and AUC of Gag GW9-specific CD8⁺ T cells (light red) compared to Nef RM9-specific CD8⁺ T cells (dark red) (gated on tetramer-positive cells within the parent CD8⁺ population) in the control animals following ART interruption. Results are displayed for each animal individually.

In the control animals, the Gag GW9- and Nef RM9-specific CD8⁺ T cells (Fig. 5E) remained at similar frequencies to each other prior to and following ART interruption and exhibited comparable AUCs. There was a slight increase in the frequency of both these populations ~3 weeks after ART release. This was likely attributable to increased antigen exposure when ART was no longer present to suppress virus replication.

Gag GW9- and Nef RM9-specific CD8⁺ T cell frequencies in the LNs are similar during N-803 treatment.

Because N-803 treatment can cause migration of CD8⁺ T cells to the LNs (44, 45), we investigated whether vaccine-elicited T cells were found at higher frequencies than naturally elicited T cells in the LNs. We again used Gag_386-394GW9 and Nef_103-111RM9 tetramers to measure the phenotypes and frequencies of these vaccine- and naturally elicited populations of antigen-specific T cells, respectively.

In the LNs of the treatment animals, the frequency of Gag GW9-specific CD8⁺ cells (range of 0.11 to 5.13%) was up to 10-fold higher than the frequency of Nef RM9-specific CD8⁺ T cells (range of 0.11 to 0.54%), and the frequencies of bulk Gag- and Nef-specific CD8⁺ T cells in the LNs were unaffected by N-803 treatment (Fig. 6A). The Gag GW9- and Nef RM9-specific T_CM and T_EM in the LNs of the treatment animals were also unchanged throughout the N-803 phase. Like in the peripheral blood, the frequency of Gag GW9-specific T_CM was lower than that of Nef RM9-specific T_CM and the frequency of Gag GW9-specific T_EM was higher than that of Nef RM9-specific T_EM cells in the LNs (Fig. 6B and C).

FIG 6.

N-803-mediated changes in Gag GW9- and Nef RM9-specific CD8⁺ T cells in the LNs of the vaccinated animals. (A) Frequency of Gag GW9-specific CD8⁺ T cells (light blue) compared to Nef RM9-specific CD8⁺ T cells (dark blue) (gated on tetramer-positive cells within the parent CD8⁺ population) throughout the N-803 phase. (B) Frequency of Gag GW9- compared to Nef RM9-specific CD8⁺ T_CM (gated on CD28⁺ CD95⁺ cells within the tetramer-positive parent population) throughout the N-803 phase. (C) Frequency of Gag GW9- compared to Nef RM9-specific CD8⁺ T_EM (gated on CD28⁻ CD95⁺ cells within the tetramer-positive parent population) throughout the N-803 phase. (D to F) Frequency of Gag GW9- or Nef RM9-specific CD8⁺ T cells expressing the proliferation marker Ki-67 alone (D), the activation marker CD69 alone (E), or Ki-67 and CD69 together (F) throughout the N-803 phase, each gated within the parent tetramer-positive population. Results are displayed for each animal individually.

We then examined the frequency of Gag GW9- and Nef RM9-specific CD8⁺ T cells expressing either Ki-67, CD69, or both Ki-67 and CD69 in the LNs. The effect of N-803 on the subpopulations of each of these tetramer-positive parent populations was similar (Fig. 6D to F). Thus, in the LNs, vaccine-elicited and naturally elicited antigen-specific CD8⁺ T cells responded similarly to N-803 treatment. Unfortunately, LNs could not be collected 1 week after the third dose, so we were unable to evaluate whether the proliferative responses were enhanced to a higher degree after the third N-803 dose. The fold change of bulk CD4⁺ and CD8⁺ T cells, as well as the frequency of CD4⁺ and CD8⁺ T_CM, T_TM, and T_EM expressing Ki-67 and/or CD69 in the LNs typically increased in the vaccinated animals after each of the first two doses of N-803, mirroring the results of the Gag GW9- and Nef RM9-specific populations in the LNs (Fig. S5). As expected, the bulk populations of T cells were unaffected in the control animals over time.

SIV was not consistently detected in the plasma during N-803 treatment.

While N-803 treatment in vivo increased the number, frequency, activation, and proliferation of CD4⁺ and CD8⁺ T cell subsets (Fig. S2 to S4), these cellular changes had no apparent effect on SIV plasma viremia, which remained below 10⁴ copies/mL following ART discontinuation. Up to 8 weeks after suspending ART, one vaccinated animal (cy1035) and two control animals (cy1037 and cy1040) had transient, detectable low-level viremia (between 1.2 × 10² and 6 × 10³ cEq/mL) (Fig. 7A). Viremia was undetectable in all animals 8 weeks after ART interruption. Further studies to understand the lack of rebound in both groups of animals following ART cessation are under way (our unpublished data). There was no significant difference in the AUCs of the viral loads between the vaccinated and control animals during the 8-week period following ART interruption evaluated in this study (Fig. 7B). All animals had low frequencies of Gag GW9- and Nef RM9- specific CD8⁺ T cells (Fig. 5A and E), and there was no clear relationship between the frequency of Gag GW9- or Nef RM9-specific CD8⁺ T cells and the presence or absence of plasma viremia.

FIG 7.

Detection of SIV in vivo and in vitro following N-803 treatment. (A) Plasma viral loads following ART interruption. (B) Log₁₀ viral load AUC analysis from day 0 to 2 months after ART interruption. Results are displayed as median and individual values. P value was calculated using a Mann-Whitney U test. (C) Table of supernatant p27 Gag concentrations from CD4⁺ T cells treated with anti-CD3/CD28 beads (center column) or N-803 (right column).

We also treated isolated CD4⁺ T cells from both groups of animals with N-803 in vitro. Consistent with our in vivo observations, N-803 did not induce the production of p27 Gag (Fig. 7C). As a positive control, treatment of cells with anti-CD3/anti-CD28 beads readily induced virus production, leading to p27 Gag concentrations ranging from 2,379.17 to 2,932.41 pg/mL in the supernatants (Fig. 7C). We thus observed no direct increase in SIV replication due to N-803 in vivo or in vitro, despite profound N-803-mediated changes in CD4⁺ and CD8⁺ lymphocyte populations and an absence of ART.

DISCUSSION

Here, we evaluated an SIV therapeutic vaccine strategy using heterologous viral vectors delivering Gag immunogens to SIV⁺ MCMs receiving ART, followed by immunotherapy with N-803 after ART discontinuation, as a novel therapeutic regimen to induce anti-SIV cellular immunity. We hypothesized that this regimen would generate and recall Gag-specific CD8⁺ T cells. To our knowledge, this is the first study to evaluate the combination of therapeutic i.v. HPBB vaccine administration with N-803 treatment after suspending ART. We found that this vaccine regimen elicited a high frequency of Gag-specific CD8⁺ T cells that contracted following each vaccine dose. These vaccine-elicited Gag-specific cells were boosted by N-803 to a similar degree as naturally elicited Nef-specific CD8⁺ T cells (Fig. 5 and 6). Because durable rebound was not detected in any animal up to 8 weeks after ART cessation, it remains unclear whether therapeutically elicited immune responses would have been sufficient to suppress SIV replication. This lack of rebound will be investigated in further studies.

One goal of the vaccine regimen presented in this study was to elicit broad and abundant Gag-specific CD8⁺ T cell responses. Broader responses to Gag have been associated with strong antiviral activity and decreased viral loads (8, 31, 32). High-magnitude Gag-specific CD8⁺ T cell responses have been associated with CD4⁺ T cell preservation and lower HIV loads (31). Our vaccine regimen modestly increased the breadth of Gag-specific T cells targeting Gag_386-394GW9 and four additional Gag epitopes (Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9) compared to the unvaccinated animals. After vaccination, there was also an increase in the frequency of Gag-specific CD4⁺ and CD8⁺ T cells expressing AIM markers, suggesting that vaccination enhanced the frequency of cells able to recognize and respond to Gag by way of antigen recognition, activation, and degranulation. While cytokine production and polyfunctionality are commonly used metrics of antigen-specific CD8⁺ T cells (30, 58), cells from MCMs are incredibly poor in vitro cytokine producers (our unpublished observations) (59). This precluded the analysis of cytokine production after each vaccination. However, these assays identify only antigen-specific T cells producing a predetermined set of cytokines while excluding antigen-specific T cells with other functions (51). For these reasons, we used AIM assays to measure the frequency of vaccine-induced antigen-specific CD4⁺ and CD8⁺ T cells via the expression of surface activation and degranulation markers.

Our vaccine regimen transiently increased the frequency of Gag GW9-specific CD8⁺ T cells in the PBMCs. These cells had a T_EM phenotype and expressed the proliferation marker Ki-67. Whereas both longer and shorter intervals (2 weeks to multiple months between doses) can elicit high frequencies of CD8⁺ T cells with a similar potential to proliferate and respond to antigen re-exposure, very short intervals (2 weeks) elicit CD8⁺ T cell populations that are more susceptible to contraction in mice (20). Petitdemange et al. found that proliferating Gag-specific CD8⁺ T cells elicited by HPBB vaccination delivered to macaques in longer intervals (weeks 1, 9, and 37) expanded to a higher, more sustained frequency after the final boost compared to our HPBB vaccine regimen, which was given at 4-week intervals (15). While populations of Gag GW9-specific CD8⁺ T cells did expand in our study, they subsequently contracted to nearly prevaccine levels after each dose (Fig. 2). The differences in response magnitude and persistence of Gag-specific CD8⁺ T cells that we observed and those detected in the study by Petitdemange et al. could stem from a variety of sources. First, an HPBB vaccine regimen may require longer intervals than we used to elicit longevous cell populations. Second, we used a nonreplicating MVA vector for the second immunization, and Petitdemange et al. used a replicating vaccinia virus (VV) vector. The inclusion of a more pathogenic VV vector rather than the attenuated MVA vector may have improved the efficacy of this boost. Finally, although our animals were receiving ART, the initial insult of SIV infection and underlying impact on the host cellular immunity could weaken the responsiveness to vaccination. It would be interesting to evaluate this HPBB vaccine regimen in SIV-naive MCMs to determine if it elicits a persistently higher frequency of GW9-specific CD8⁺ T cells.

Because all the vaccines delivered SIVmac239 Gag in this study, we were able to use Gag_386-394GW9 and Nef_103-111RM9 tetramers to compare vaccine-elicited (Gag GW9-specific) to naturally elicited (Nef RM9-specific) CD8⁺ T cells within each animal. Many therapeutic vaccine regimens deliver multiple SIV immunogens, such as Gag, Pol, and Env (23, 29, 30), making any distinction between vaccine-elicited cells and cells naturally elicited by HIV/SIV infection difficult. Prophylactic vaccine studies delivering multiple immunogens similarly lose the ability to distinguish between antigen-specific CD8⁺ T cells elicited by vaccination and those from SIV infection (60, 61). Further, there is limited availability of tetrameric reagents to track defined antigen-specific CD8⁺ T cells in individuals even when their MHC genetics are known (17, 45, 62). These peptide-loaded MHC class I tetramers are required to distinguish between CD8⁺ T cells targeting immunogens present in the vaccine or challenge virus. By using MCMs expressing the Mafa-A1*063 MHC class I molecule, which restricts the Gag_386-394GW9 epitope in the vaccine immunogen and the Nef_103-111RM9 epitope in the SIV challenge virus, we could distinguish between cells elicited by vaccination versus SIV infection within the same animals.

As N-803 has been shown to enhance the frequency and LN trafficking of SIV-specific CD8⁺ T cells elicited by SIV infection (44, 45), we tested the hypothesis that N-803 would boost the magnitude of vaccine-elicited CD8⁺ T cell responses to a higher degree than that of CD8⁺ T cells elicited naturally by SIV infection. However, we actually found that N-803 treatment increased the frequency of Ki-67⁺ Gag GW9- and Nef RM9-specific CD8⁺ T cells similarly in PBMCs, suggesting a role for N-803 in the proliferation of all antigen-specific CD8⁺ T cells. This is consistent with defined contributions of IL-15 in CD4⁺ and CD8⁺ T cell activation, proliferation, and maintenance (63, 64). Notably, since the vaccine regimen elicited a high frequency of Gag GW9-specific CD8⁺ T cells, N-803 treatment further magnified the frequency of CD8⁺ T cells elicited by the vaccine compared to those elicited by natural infection. In the context of antiviral immunity, the potential of N-803 to drive activation and proliferation of vaccine-elicited CD8⁺ T cells may support the use of N-803 as an adjuvant in HIV/SIV therapeutic strategies.

We were surprised that N-803 treatment did not dramatically increase the frequency of Gag GW9-specific CD8⁺ T cells in the LNs. It is possible that N-803 directed these antigen-specific CD8⁺ T cells to different tissue sites like the gut, which we did not sample. Alternatively, the presence of antigen may be required for N-803-mediated LN localization of antigen-specific T cells. As only one vaccinated animal exhibited low-level plasma viremia and the other three vaccinated animals were aviremic at the time of N-803 administration, it is unclear whether higher viremia and therefore higher levels of antigen-stimulation would have enabled N-803 to direct more antigen-specific T cells to the LNs. It would be interesting to deliver N-803 and vaccines concomitantly to determine if combining N-803 with antigen stimulation would enhance the frequency of SIV-specific CD8⁺ T cells directed to the LNs.

Antigen-specific CD8⁺ T_EM are critical for HIV/SIV therapeutic strategies because lymphocytes differentiated into T_EM phenotypes have been associated with enhanced effector function, expression of IL-2, IFN-γ, and TNF, progressive clearance of SIV reservoirs, and detection of replication-competent virus in resting CD4⁺ T cells (65–67). The vaccinated animals here developed a higher frequency of Gag GW9-specific T_EM than the control animals. The vaccinated animals also developed a higher frequency of Gag GW9-specific CD8⁺ T_EM than Gag GW9-specific T_CM, consistent with reports that replicating vectors like VSV provide recurrent antigen exposure that favors the induction of protective T_EM responses (27). This represents one potentially advantageous feature of this vaccine regimen. Interestingly, Hansen and colleagues found that protective SIV-specific T_EM populations elicited by persistent rhesus cytomegalovirus (RhCMV) vectors prior to SIV infection may have played a role in preventing SIV infection (65). In our study, the T_EM bias of vaccine-elicited Gag-specific CD8⁺ T cells and T_CM bias of Nef-specific CD8⁺ T cells elicited by SIV infection were maintained and generally unaltered throughout N-803 treatment. Whether N-803 would boost T_EM cells in the case of an RhCMV vaccination strategy or whether the HPBB vaccine strategy described here combined with N-803 treatment would be protective if delivered prophylactically remains unclear. However, we propose that alternative ways to elicit and expand antigen-specific CD8⁺ T_EM cells that do not require vaccination with a variant strain of CMV exist.

One drawback of this study is the lack of additional single-intervention animal cohorts. Ideally, the inclusion of one cohort treated with the vaccine regimen but not N-803 and another cohort treated with N-803 and no vaccinations would have enabled a more thorough evaluation of the immune responses of each intervention individually. These single-intervention cohorts would allow us to define how vaccine-elicited CD8⁺ T cells respond to ART interruption in the absence of additional cytokine modulation and how N-803 boosts lymphocyte populations untouched by a therapeutic vaccine intervention when delivered after ART release. However, macaques are a valuable and costly resource. Thus, rather than add two additional cohorts of animals, we evaluated the impact of N-803 on vaccine-elicited CD8⁺ T cells versus CD8⁺ T cells elicited only by SIV infection within the vaccinated group.

N-803 treatment did not induce SIV replication in vivo or in vitro. The role of N-803 as a latency-reversing agent (LRA) remains unclear, despite in vitro and in vivo evaluation (45, 68–70). Jones et al. described N-803 treatment of cell cultures in vitro induced antigen presentation and CD8⁺ T cell recognition in primary cell cultures and ex vivo cultures from HIV⁺ individuals (68). McBrien and colleagues found that N-803 reactivated HIV expression from CD4⁺ T cells infected in vitro, but that reactivation could be inhibited by the presence of activated CD8⁺ T cells (70). These results stand in contrast to our observations that in vitro N-803 treatment of CD4⁺ T cells isolated from the PBMCs of SIV⁺ MCMs did not induce p27 Gag expression. It is possible that our p27 enzyme-linked immunosorbent assay (ELISA) was not sensitive enough to detect very low levels of N-803-mediated latency reversal in vitro. In vivo, N-803 appears to act as an LRA only when CD8⁺ T cells are absent (45, 69, 70). Whether this is due to CD8⁺ T cells assisting in latency maintenance or CD8⁺ T cells suppressing virus from reactivated target cells remains unclear. Webb et al. also showed that in vivo N-803 administration was insufficient to perturb the viral reservoir (45). These findings are consistent with our observation that N-803 did not induce viremia in vivo following ART withdrawal in animals with an intact CD8 compartment. The failure of N-803 to induce p27 Gag expression in isolated CD4⁺ T cells in vitro suggests that this agent did not perturb the reservoir in CD4⁺ T cells isolated from MCMs. This could be attributed to a low frequency of infected CD4⁺ T cells used in this assay, or it could be that N-803, alone, cannot reactivate SIV from MCM CD4⁺ T cells. It may therefore be necessary to combine N-803 with a different LRA for a “shock-and-kill” strategy to effectively reduce the size of the viral reservoir.

In conclusion, we describe the immunogenicity and immunomodulation of a novel therapeutic regimen combining HPBB vaccination delivering Gag immunogens during ART combined with N-803 after suspending ART. The vaccine regimen elicited a high frequency of proliferative and activated CD8⁺ T cells driven toward a T_EM phenotype. N-803 enhanced both the vaccine-elicited and naturally elicited SIV-specific CD8⁺ cells. Thus, N-803 remains an attractive immunomodulatory agent for HIV/SIV to enhance cellular immune responses elicited both by infection and by other therapeutic interventions. Any functional cure for HIV will surely involve a combination of multiple therapeutic interventions, and we show here that N-803 combined with vaccination generates and recalls cellular immune responses against SIV.

MATERIALS AND METHODS

Animal care and use.

Eight male Mauritian cynomolgus macaques (MCMs) were purchased from Bioculture, Ltd., and were housed and cared for by the Wisconsin National Primate Research Center (WNPRC) according to protocols approved by the University of Wisconsin Graduate School Animal Care and Use Committee (IACUC; protocol number G005507). The animals were chosen based on the presence of at least one copy of the M3 MHC haplotype and the absence of the M1 haplotype, which is associated with viral control (Table 1) (46, 47, 71). All eight MCMs were infected intravenously (i.v.) with 10,000 IU of SIVmac239M (72). An antiretroviral therapy (ART) regimen consisting of 2.5 mg/kg dolutegravir (DTG; ViiV Healthcare, Research Triangle, NC), 5.1 mg/kg tenofovir disoproxil fumarate (TDF; Gilead, Foster City, CA), and 40 mg/kg emtricitabine (FTC; Gilead) in 15% Kleptose (Roquette America) in water was delivered subcutaneously daily beginning 2 weeks postinfection. Four animals (the vaccinated group) received Gag or Gag-Pol proteins encoded by three heterologous viral vectors. Recombinant vesicular stomatitis virus New Jersey strain (VSV) expressed full-length SIVmac239 Gag (61, 73). The VSV-Gag vector was given i.v. at a dose of 5 × 10⁷ PFU per animal at week 23 after SIV infection. Recombinant modified vaccinia virus Ankara strain (MVA) expressed SIVmac239 Gag-Pol (74). The MVA-Gag-Pol vector was given i.v. at a dose of 1 × 10⁸ PFU per animal at week 27 post-SIV. Recombinant adenovirus serotype 5 (rAd-5) expressed full-length SIVmac239 Gag (ViraQuest, Inc.). The rAd-5-Gag was given i.v. at a dose of 7 × 10¹⁰ particles per animal at week 31 after SIV infection. All immunogens were injected i.v. in a 1-mL total volume, diluted in sterile 1× phosphate-buffered saline (PBS), where necessary. Four animals (the control group) received 1 mL sterile 1× PBS injections only. Three doses of N-803, separated by 2 weeks each, were delivered subcutaneously to the four treatment animals at a dose of 0.1 mg/kg beginning 3 days after ART interruption.

Plasma viral load analysis.

Plasma was isolated from undiluted whole blood by Ficoll-based density centrifugation and cryopreserved at −80°C. Plasma viral loads were quantified as previously described (75). Briefly, the Maxwell viral total nucleic acid purification kit (Promega, Madison, WI) was used to isolate viral RNA (vRNA) from plasma samples. vRNA was then reverse transcribed using the TaqMan Fast virus one-step quantitative reverse transcription-PCR (qRT-PCR) kit (Invitrogen) and quantified on a LightCycler 480 (Roche, Indianapolis, IN) instrument.

IFN-γ ELISPOT assays.

IFN-γ ELISPOT assays were performed using fresh and cryopreserved PBMCs, as previously described (76). Peptides (Gag_386-394GW9, Nef_103-111RM9, Gag_459-467TV9, Gag_146-154HL9, Gag_28-37KA10, and Gag_221-229PR9, and a Gag peptide pool containing 15-mer peptides spanning the full SIVmac239 Gag proteome, each overlapping by 11 amino acids [NIH HIV Reagent Program, managed by ATCC]) were selected from epitopes known to be restricted by the Mafa-A1*063 MHC class I allele expressed on the M3 MHC haplotype (50). PBMCs were isolated from EDTA-anticoagulated blood by Ficoll-based density centrifugation. Precoated monkey IFN-γ ELISPOTplus plates (Mabtech, Cincinnati, OH) were blocked with R10 (RPMI 1640 supplemented with 10% fetal bovine serum (FBS), 1% antibiotic-antimycotic [Thermo Fisher Scientific, Waltham, MA], and 1% l-glutamine [Thermo Fisher Scientific]), and individual peptides were added to each well at a final concentration of 10 μM. The Gag peptide pool was added to cells at a final concentration of 625 μg/mL (5 μg/mL of each peptide). Each peptide or peptide pool was tested in duplicate. Concanavalin A (10 μM) was used as a positive control and was tested in duplicate as well. Four wells per animal received no peptides as a negative control to calculate background reactivity. Plates were incubated overnight at 37°C in 5% CO₂. Assays were performed according to the manufacturer’s protocol, and wells were imaged with an ELISPOT plate reader (AID Autoimmun Diagnostika GmbH). Positive responses were determined using a one-tailed t test at an α level of 0.05, where the null hypothesis was that the background level would be greater than or equal to the treatment level (60, 76). Statistically positive responses were considered valid only if both duplicate wells contained 50 or more spot-forming cells (SFCs) per 10⁶ PBMCs. If responses were statistically positive and were ≥50 SFCs per 10⁶ PBMC, values were reported as the average of the two test wells minus the average of all four negative-control wells.

Tetramerization of Gag_386-394GW9 and Nef_103-111RM9.

The Gag_386-394GW9 and Nef_103-111RM9 peptides were purchased from GenScript (Piscataway, NJ). The NIH Tetramer Core Facility at Emory University (Atlanta, GA) produced biotinylated Mafa-A1*063 MHC class I monomers loaded with these peptides. The Mafa-A1*063 Gag_386-394GW9 and Mafa-A1*063 Nef_103-111RM9 monomers were tetramerized with streptavidin-phycoerythrin (PE) (0.5 mg/mL; BD Biosciences) and streptavidin-BV421 (0.1 mg/mL; BD Biosciences), respectively, at a 4:1 molar ratio of monomer to streptavidin in the presence of a 1× protease inhibitor cocktail solution (Calbiochem, Millipore Sigma). Then, 1/5 volumes of streptavidin-PE or streptavidin-BV421 were added to each monomer every 20 min and incubated, rotating in the dark at 4°C until the full streptavidin volume was added.

Phenotype staining of T cells by flow cytometry.

Previously frozen PBMCs isolated from whole blood and previously frozen lymph node (LN) mononuclear cells (LNMC) isolated from LN biopsy specimens were used to assess the quantity and phenotype of T cell populations longitudinally. Briefly, cells were thawed, washed once with R10, and rested for 30 min at room temperature in a buffer consisting of 2% FBS in 1× PBS (2% fluorescence-activated cell sorting [FACS] buffer) with 50 nM dasatinib (Thermo Fisher Scientific). Cells were washed once with 2% FACS buffer with 50 nM dasatinib and incubated with the Gag_386-394GW9 and Nef_103-111RM9 tetramers for 45 min at room temperature. Cells were then washed once with 2% FACS buffer with 50 nM dasatinib and incubated with the remaining surface markers (Table 2) for 20 min at room temperature. Cells were washed twice with 2% FACS buffer with 50 nM dasatinib and fixed using fixation/permeabilization solution (Cytofix/Cytoperm fixation and permeabilization kit; BD Biosciences) for 20 min at 4°C. Cells were next washed twice with cold 1× Perm/Wash buffer (Cytofix/Cytoperm fixation and permeabilization kit; BD Biosciences) and incubated with a master mix containing 95 μL of 1× Perm/Wash buffer and 5 μL of the intracellular marker Ki-67 (Table 2) for 20 min at 4°C. Cells were then washed twice with 1× Perm/Wash buffer and acquired immediately using a FACS Symphony A3 instrument (BD Biosciences). The data were analyzed using FlowJo software for Macintosh (BD Biosciences, version 10.8.0). Subpopulations of cells were excluded from analysis when the parent population contained <50 events.

TABLE 2.

Antibodies used for T cell phenotyping

Antibody	Clone	Fluorochrome	Location
Live/Dead		Near-infrared	Surface
CD3	SP34-2	BUV563	Surface
CD4	SK3	BUV737	Surface
CD8	RPA-T8	BUV395	Surface
CD28	CD28.2	BUV661	Surface
CD69	TP1.55.3	PE-Texas Red (ECD)	Surface
CD95	DX2	BV711	Surface
CCR7	150503	FITCa	Surface
CXCR5	MU5UBEE	PE-Cy7	Surface
Ki-67	B56	AF-647	Intracellular
Gag GW9 tetramer		PE	Surface
Nef RM9 tetramer		BV421	Surface

^aFluorescein isothiocyanate.

AIM assays.

AIM assays were performed to characterize the antigen-specific markers of activation similarly to previously published work (51, 54, 77). Previously frozen PBMCs isolated from whole blood and previously frozen LNMC isolated from LN biopsy specimens were thawed, washed twice with R10, and incubated for ~20 h in Gibco AIM V serum-free medium (Thermo Fisher Scientific) at 37°C in 5% CO₂ either with AIM V medium alone (unstimulated) or with a Gag peptide pool containing 15-mer peptides spanning the full SIVmac239 Gag proteome, each overlapping by 11 amino acids (provided by the HIV Reagent Program), at a final concentration of 62.5 μg/mL (0.5 μg/mL of each peptide) (Gag stimulated). Two wells stimulated with 5 μg/mL concanavalin A (ConA) were included in each batch of staining as a positive control. Anti-CD107a and anti-CD154 antibodies (Table 3) were added to all cells during stimulation. Following stimulation, cells were washed twice with 2% FACS buffer and stained with antibodies to the indicated surface markers (Table 3) for 20 min at room temperature. Cells then were washed twice with 2% FACS buffer, fixed with 2% paraformaldehyde for 20 min at room temperature, washed twice more with 2% FACS buffer, and resuspended in 2% FACS buffer. Flow cytometry was performed as described above.

TABLE 3.

Antibodies used for AIM assay

Antibody	Clone	Fluorochrome
Live/Dead		Near-infrared
CD3	SP34-2	AF-700
CD4	SK3	BUV737
CD8	RPA-T8	BUV395
CD25	M-A251	BV786
CD69	FN50	PE-Cy7
CD107a	H4A3	APCa
CD137	4B4-1	PE CF594
CD154	24-31	BV605

^aAllophycocyanin.

In vitro latency reactivation and p27 ELISA.

Previously cryopreserved PBMCs isolated from whole blood were thawed. After thawing, CD4⁺ T cells were isolated from the PBMCs by negative selection using nonhuman primate CD4 Microbeads according to the manufacturer’s protocol (Miltenyi Biotec). CD4⁺ T cells were incubated for 72 h at 37°C with either anti-CD3/anti-CD28 beads (1:2 bead-to-cell ratio; Miltenyi Biotec) or 200 ng/mL N-803 in the presence of 50 μM raltegravir and 5 μM saquinavir. After 72 h, the supernatant was then collected, frozen, and subjected to SIV p27 ELISA per the manufacturer’s protocol (ZeptoMetrix). ELISA plates were immediately read using a GloMax-Multi detection system microplate reader (Promega) at an absorbance of 450 nm.

Statistical analysis.

AUC analyses were performed using GraphPad Prism. For statistical analyses in which animal groups were compared to each other at the same time point, Mann-Whitney U tests were performed. Comparisons between unstimulated conditions and stimulated conditions within the same animal were tested using paired t tests.

Study approval.

This study was approved by the University of Wisconsin Graduate School Animal Care and Use Committee (IACUC; protocol number G005507).