Antibodies to gp120 and PD-1 Expression on Virus-Specific CD8⁺ T Cells in Protection from Simian AIDS

Abstract

We compared the relative efficacies against simian immunodeficiency virus (SIV) challenge of three vaccine regimens that elicited similar frequencies of SIV-specific CD4⁺ and CD8⁺ T-cell responses but differed in the level of antibody responses to the gp120 envelope protein. All macaques were primed with DNA plasmids expressing SIV gag, pol, env, and Retanef genes and were boosted with recombinant modified vaccinia Ankara virus (MVA) expressing the same genes, either once (1 × MVA) or twice (2 × MVA), or were boosted once with MVA followed by a single boost with replication-competent adenovirus (Ad) type 5 host range mutant (Ad5 h) expressing SIV gag and nef genes but not Retanef or env (1 × MVA/Ad5). While two of the vaccine regimens (1 × MVA and 1 × MVA/Ad5) protected from high levels of SIV replication only during the acute phase of infection, the 2 × MVA regimen, with the highest anti-SIV gp120 titers, protected during the acute phase and transiently during the chronic phase of infection. Mamu-A*01 macaques of this third group exhibited persistent Gag CD8⁺CM9⁺ effector memory T cells with low expression of surface Programmed death-1 (PD-1) receptor and high levels of expression of genes associated with major histocompatibility complex class I (MHC-I) and MHC-II antigen. The fact that control of SIV replication was associated with both high titers of antibodies to the SIV envelope protein and durable effector SIV-specific CD8⁺ T cells suggests the hypothesis that the presence of antibodies at the time of challenge may increase innate immune recruiting activity by enhancing antigen uptake and may result in improvement of the quality and potency of secondary SIV-specific CD8⁺ T-cell responses.

INTRODUCTION

The human immunodeficiency virus type 1 (HIV-1) vaccine platforms brought to the clinic so far have induced either T-cell responses alone, nonneutralizing antibodies to primary HIV isolates, or both (1–4).

Recombinant poxvirus vector-based HIV vaccine candidates are among the most intensively studied live vectors (5). Numerous studies have demonstrated their ability to induce balanced CD4⁺ and CD8⁺ T-cell responses as well as mucosal immune responses against heterologous antigens. Live recombinant poxvirus vectors such as modified vaccinia Ankara virus (MVA) (6–11), and genetically attenuated NYVAC or ALVAC (8, 12, 13) have been evaluated in combination with DNA immunization in preclinical studies in macaques (5). This approach has generated high-frequency CD4⁺ and CD8⁺ T-cell responses (8, 10, 11, 14–16) and high titers of nonneutralizing antibodies when combined with viral recombinant proteins (13; our unpublished data). In the simian immunodeficiency virus (SIV) macaque model, the combination of DNA and these viral vector-based vaccine modalities has induced the expression and mobilization of proinflammatory cytokines and reduced by 1 to 2 log the viral load in plasma during primary and chronic infection, depending on the virus used in the challenges (6–13, 15–19; Vaccari et al., unpublished). Protection from high viral replication afforded by these modalities has been correlated not only with vaccine-induced central memory (CM) CD8⁺ T-cell responses but also with CD4⁺ T-cell responses (14). In addition, a direct role of vaccine-induced CD8⁺ and CD4⁺ T-cell responses in the control of SIV replication has been demonstrated by depleting CD4⁺ T cells during vaccination and by depleting CD8⁺ T cells in vaccinated animals before challenge exposure (11, 20). The protection afforded by poxvirus-based SIV vaccines in the preclinical setting using a high-dose challenge modality is, however, not long lasting, and eventually viral replication returns to the levels observed in control animals (14).

Recombinant adenoviruses (Ad) represent a second group of prominent vaccine vectors. Replication-defective Ad vectors have been the most extensively studied (21), but replication-competent Ad vectors are also highly promising, especially as they target multiple mucosal sites regardless of immunization route (22). In the chimpanzee model, replicating Ad-HIV recombinants in combination with envelope protein boosts have elicited potent, long-lasting neutralizing antibodies (23) and protection against homologous and heterologous HIV challenges (24). In the rhesus macaque model, similar prime/boost approaches using replicating Ad type 5 host range mutant (Ad5 h) viruses carrying HIV and/or SIV genes have elicited strong humoral and cellular systemic and mucosal immune responses and sterilizing protection (25) or strongly reduced viremia (26–29) following high-dose intrarectal or intravenous simian/human immunodeficiency virus (SHIV) or SIV_mac251 challenges, as well as delayed acquisition following repeated low-dose SIV_mac251 intrarectal challenges (30).

Nevertheless, in the phase 2b efficacy trial (termed Step), replication-defective Ad5-HIV vaccines expressing Gag, Pol, and Nef failed to prevent HIV acquisition. Moreover, an early vaccine-enhanced risk of HIV acquisition, seen predominantly in uncircumcised men with preexisting Ad5 antibody, was recently confirmed, although the risk waned over time (31). Though the results of this clinical trial were discouraging, a “sieve effect” was confirmed, showing that vaccine-elicited T-cell responses exerted selective pressure on the HIV-infecting founder virus population, although they were not potent enough to prevent infection or decrease viral loads significantly (32). The absence of the envelope gene in this vaccine approach may also account for the lack of protection. The recent results of the RV144 vaccine trial in Thailand suggest that vaccine-elicited antibodies, together with T-cell responses to the HIV-1 envelope, confer some degree of protection against HIV infection (4).

Here, we evaluated the ability of a vaccination approach that induces a combination of T-cell and antibody responses to protect against a high-dose SIV_mac251 challenge. We wished to test (i) whether vaccination with DNA followed by recombinant modified vaccinia Ankara virus twice (2 × MVA) could induce T-cell immune responses quantitatively different from those obtained with DNA followed by 1 × MVA or 1 × MVA/Ad5; (ii) whether the omission of the second boost with MVA (1 × MVA) or the absence of the envelope in the Ad5 boost (1 × MVA/Ad5) affected the antibody response to the envelope protein and protection; (iii) whether boosting with replication-competent Ad could provide a higher frequency of long-lasting virus-specific CD8⁺ and CD4⁺ T-cell responses in systemic compartments; and (iv) which combination of DNA and live vector-based vaccines would best protect against a high dose of the pathogenic SIV_mac251. Our results support the concept that vaccine-induced antibodies to the envelope protein are important in protection and may contribute to the maintenance of effector CD8⁺ T-cell responses in infected animals.

MATERIALS AND METHODS

Animals, treatments, and SIV_mac251 challenge.

All animals used in this study were colony-bred rhesus macaques (Macaca mulatta), obtained from Covance Research Products (Alice, TX). The animals were housed and handled in accordance with the standards of the Association for the Assessment and Accreditation of Laboratory Animal Care International. The care and use of the animals were in compliance with all relevant institutional (NIH) guidelines.

A total of 27 macaques were enrolled in the study. Twenty-one macaques were divided into three groups of seven animals each (Fig. 1). The 1 × MVA group included two Mamu-A*01⁺ animals, while each of the remaining groups had three. The 21 macaques were immunized at weeks 0 and 4 with intramuscular (3-mg) and intradermal (1-mg) inoculations of DNA-SIV Retanef (8) and individual plasmid DNAs encoding codon-optimized SIV gag, pol, and env genes (33), all given at different sites as previously described (8, 12, 34). All animals were boosted at week 26 with 10⁸ PFU of recombinant MVA expressing SIV_mac251 env and gag-pol genes and SIV_mac239 Retanef genes delivered intramuscularly at two sites for each animal. Seven animals received no further immunization (1 × MVA group), and seven animals received a second boost with the two MVA recombinants at week 53 (2 × MVA group). The remaining seven macaques were vaccinated with 5 × 10⁸ PFU each of two replication-competent Ad5 h recombinants expressing SIV_239gag (35) and SIV_{239nefΔ1–14} (36) by the intranasal route at week 53 (1 × MVA/Ad5 group). Six control macaques were mock vaccinated with the equivalent amount of nonrecombinant MVA (3 animals, 1 Mamu-A*01⁺) or Ad5 h (3 animals, 1 Mamu-A*01⁺) at weeks 26 and 53. All macaques were challenged intrarectally with a high dose (1:10 stock dilution) of pathogenic SIV_mac251 21 or 26 weeks after the last immunization (Fig. 1A).

Fig 1.

Clinical and virological outcome of vaccinations. (A) Schematic representation of the immunization regimens. The arrows indicate times of immunizations and of challenge exposure to SIV_mac251, performed by the intrarectal route at week 74 in all animals and at week 52 in the 1 × MVA group. (B, C, D, E) Plasma virus levels are presented for each animal. *, identified Mamu-A*01⁺ macaques. (F, G, H) Comparison of mean virus load in plasma in macaques from the different groups.

Binding antibody assay.

To detect anti-SIV_mac251-binding antibodies, serial dilutions of plasma were incubated with SIV_mac251 lysate spiked with native purified gp120 or p27 Gag protein of SIV_mac251 bound to microtiter enzyme-linked immunosorbent assay plates (19). Endpoint titers were defined as the reciprocal of the highest serum dilution that gave an optical absorbency at 450 nm at least 2 standard deviations greater than average values obtained with negative-control serum.

Preparation of lymphocytes from blood and tissues.

Mononuclear cells from blood and lymph nodes (LNs) were isolated by density gradient centrifugation on Ficoll and resuspended in RPMI 1640 medium (Gibco BRL, Gaithersburg, MD) containing 10% fetal bovine serum (R-10).

Measurement of viral RNA.

SIV_mac251 in plasma was quantified by nucleic acid sequence-based amplification as previously described (23). Total RNA was extracted from isolated cells using the guanidium thiocyanate-phenol-chloroform method, modified for TRIzol (Invitrogen, Carlsbad, CA). RNA (1 μg) was reverse transcribed into first-strand cDNA using random hexanucleotide primers, oligo(dT), and Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI). cDNA quantification was performed by real-time PCR, conducted with the ABI Prism 7900HT (Applied Biosystems, Foster City, CA). All reactions were performed using a SYBR green PCR mix (Qiagen, Valencia, CA) according to the following thermal profile: denaturation at 95°C for 15 s, annealing at 60°C for 15 s, and extension at 72°C for 15 s (data collection was performed during the extension step). Primer sequences were as follows: GAPDH forward, 5-GTCTGGAAAAACCTGCCAAG-3, and reverse, 5-ACCTGGTGCTCAGTGTAGCC-3; SIVgag forward, 5-GCAGAGGAGGAAATTACCCAGTAC-3, and reverse, 5-CAATTTTACCCAGGCATTTAATGTT-3.

Lymphocyte proliferation assays.

Peripheral blood mononuclear cells (PBMC) were incubated with antigens (SIVp27 Gag or gp120 envelope proteins) at a concentration of 5 μg/ml for 5 days at 37°C in 96-well plates. Concanavalin-A (ConA; Sigma-Aldrich) at a concentration of 2 μg/ml was used as polyclonal stimulator, while cultures without ConA or Gag or Envelope proteins were used to determine the background proliferative response. For the in vitro Programmed death-1 (PD-1) and its natural ligand PD-L1 blockade experiments, blocking antibody specific to PD-L1 (eBiosciences) was added to cell cultures at a concentration of 10 μg/ml. A standard surface-staining protocol was followed for CD4⁺ and CD8⁺ T cells using anti-human CD4-PE (BD-Pharmingen) and CD8-phycoerythrin (PE)-Texas Red (ECD) (Cedarlane) monoclonal antibodies (MAbs). Between 250,000 and 1 × 10⁶ events were acquired for each condition. Data were analyzed using DIVA software (BD Biosciences). Alternatively, PBMC were cultured at 10⁵ cells per well in triplicate for 3 days in the absence or presence of native high-performance liquid chromatography-purified SIV p27 Gag or gp120 Env protein (Advanced BioScience Laboratories, Kensington, MD) or with ConA as a positive control. The cells were then pulsed overnight with 1 μCi of [³H]thymidine before harvest. Proliferation is expressed as the stimulation index, calculated as the fold increase in thymidine incorporation into cellular DNA of stimulated versus unstimulated (medium only) cells.

Intracellular staining and enzyme-linked immunospot (ELISpot) assays.

SIV-specific CD4⁺ and CD8⁺ T-cell responses were detected using pools of 15-meric peptides overlapping by 11 amino acids covering entire Gag, Env, and Pol proteins of SIV_mac239. Cells (1 × 10⁶) in RPMI 1640 medium (containing 10% human serum and antibiotics) were incubated in the absence or the presence of a specific peptide pool at 1 μg/ml of each peptide for 1 h or in the presence of the superantigen staphylococcal enterotoxin B (Sigma, St. Louis, MO) at a 1-mg/ml final concentration as a positive control. The costimulatory MAbs CD28 and CD49d (0.5 μg/ml; BD Pharmingen, San Diego, CA) were added to all the samples to maximize the detection of T cells with higher activation thresholds. The positive control was treated with staphylococcal enterotoxin B, brefeldin A (GolgiPlug; BD Biosciences) at a final concentration of 10 μg/ml was added, and the cells were incubated for an additional 5 h. The cells were washed, stained for the surface antigens with anti-human CD3alt epsilon (clone SP34; BD Pharmingen) and anti-human CD8beta Abs (clone 2ST8.5H7; Beckman Coulter), permeabilized by incubation in Cytofix/Cytoperm solution (BD Biosciences), and stained with anti-human interleukin-2 (IL-2)-fluorescein isothiocyanate (FITC) (clone MQ1-17H12) and anti-gamma interferon (IFN-γ)-allophycocyanin (APC) MAbs (clone B27; BD Pharmingen). The results were calculated as the total number of cytokine-positive cells with background subtracted.

T-cell ELISpot assay was performed using Mabtech ELISpot plus for Monkey IFN-γ kits in triplicate on freshly purified PBMC. Cells (3 × 10⁵ cells/well) were used with or without stimulation with the SIV_mac251 Gag or Env overlapping peptides as described previously (11). ConA was used as a positive control (final concentration, 1 μg/ml). The cells were then incubated at 37°C in 5% CO₂ atmosphere for 24 h and washed. The plates were assayed for IFN-γ-producing cells per the manufacturer's protocol. The spot-forming cells (SFC) were counted on an ELISpot reader. The results are expressed as number of spots per 10⁶ cells after subtracting the number of spots in unstimulated wells.

Flow cytometry and isolation of Tetramer⁺ and memory T-cell subpopulations.

Mononuclear cells from blood were separated and stained as previously described (11).

The frequency of CCR5-positive CD4⁺ T-cells was assessed by staining mononuclear cells with anti-human CD3 antibody (clone SP34), anti-human CD4 antibody (clone L200), and anti-human CCR5 antibody (clone 3A9; BD Biosciences). SIV-specific CD8⁺ T cells staining was performed with anti-human CD8β antibody (clone 2ST8.5H7; Beckman Coulter), anti-human CD28 antibody (clone CD28.2; Pharmingen, San Diego, CA), anti-human CD95 antibody (clone DX2; BD Pharmingen), and Gag181-189 CM9 (p11C; CTPYDINQM)-Mamu-A*01 tetrameric complexes (Beckman Coulter) for 30 min at room temperature.

Lymph nodes isolated from the macaques at week 24 following challenge were labeled as follows. Cells isolated from Mamu-A*01⁺ animals were labeled with CD8 (CD8-PE-Texas Red [ECD]; BD), CM9 (Tet-PE; BD), and 7-AAD (PE-Cy5; BD). Cells isolated from non-Mamu-A*01⁺ animals were labeled with CD4 (clone L200-PerCP-Cy5.5; BD Biosciences), CD8 (CD8-PE-Texas Red [ECD]; Cedarlane), CD95 (clone DX2-PE-Cy5; BD Biosciences), CD28 (clone CD28.2-Pacific Blue; custom made, BD Biosciences), CCR7 (clone 3D12-PE-Cy7; BD Biosciences), 7-AAD (PE-Cy5; BD Biosciences). For Mamu-A*01⁺ animals, cells were sorted for CD8⁺CM9⁺ cells. For non-Mamu-A*01⁺ animals, cells were sorted into CM and effector memory (EM) T cells. Sorting was performed using a flow cytometer (FACSAria; BD Biosciences). The purity of the CM and EM subpopulations ranged from 96 to 99%. Cells were sorted directly into RLT buffer and followed by RNA extraction. All procedures were done at 4°C to minimize any changes in cell phenotype or gene expression.

In vitro infection of human DCs with MVA and Ad5.

PBMC, isolated from 5 cytomegalovirus (CMV)-negative subjects, were infected in vitro with MVA or nonreplicating Ad5, and PD-L1 levels of expression on dendritic cells (DCs) were monitored following infection. A noninfected group (Mock) was also monitored. Briefly, 3 × 10⁶ cells were infected, or not, with either MVA (gpe) or Ad5 (gag) at a multiplicity of infection (MOI) of 10 and then incubated at 37°C for 24 h. Cells were then labeled with anti-CD3-Alexa 700, anti-CD19-Alexa 700, anti-CD14-Alexa 700 (BD Biosciences), anti-CD16-Alexa 700 (Biolegend), anti-HLADR-APC-Cy7, anti-CD11c-APC, anti-CD123-PE, anti-CD86-PE-cy5, and anti-PD-L1:PE-Cy7 (BD Biosciences) and analyzed with a BD LSRII flow cytometer. Myeloid DCs (mDCs) were identified as CD3⁻, CD19⁻, CD16⁻, CD14⁻, HLADR⁺, CD11c⁺, and CD123⁻; and plasmacytoid dendritic cells (pDCs) as CD3⁻, CD19⁻, CD16⁻, CD14⁻, HLADR⁺, CD11⁻, and CD123⁺. The level of activation, and hence infection, of DCs was monitored by measuring CD86 expression (n = 5 for the three groups, MVA, Adv, and Mock).

RNA isolation, amplification, and microarray hybridization.

CM9⁺ cells within LN of Ad5 h and MVA-vaccinated macaques were sorted 24 weeks following challenge using a BD FACSARIA I. Sample RNA was extracted using an RNA extraction kit (Qiagen) and amplified using an RNA kit (MessageAmp; Ambion) according to the manufacturer's instructions. Quantification was performed using a spectrophotometer (NanoDrop Technologies), and RNA quality was assessed using the Experion automated electrophoresis system (Bio-Rad). Total RNA was then amplified and labeled using the Illumina TotalPrep RNA amplification kit, which is based on the Eberwine amplification protocol. This protocol involves an initial cDNA synthesis step followed by in vitro transcription for cRNA synthesis. The biotinylated cRNA was hybridized onto Illumina Human RefSeq-8 BeadChips at 58°C for 20 h and quantified using an Illumina BeadStation 500GX scanner and Illumina BeadStudio v3 software. We optimized the use of the Illumina microarray human RefSeq-8 BeadChips for the analysis of macaque gene expression.

Microarray data preprocessing.

Illumina probe results were exported from BeadStudio as raw data and screened for quality. Samples failing chip visual inspection and control examination were removed. Gene expression data were analyzed using Bioconductor, an open-source software library for the analysis of genomic data based on R, a language and environment for statistical computing and graphics (www.r-project.org). Genes were then filtered by intensity and by variance filters to allow a reduction in the number of tests and a corresponding increase in power of the differential gene expression analysis (37). The resulting matrix showing filtered genes as rows and samples as columns was log₂ transformed and used as input for linear modeling using Bioconductor's limma package. P values from the resulting comparison were adjusted for multiple testing. This method controls the false-discovery rate, which was set to 0.05 in this analysis. Data were then analyzed using gene set enrichment analysis (GSEA) and ingenuity pathway analysis. GSEA is a computational method that determines whether a previously defined set of genes shows statistically significant, concordant differences between two biological states (e.g., phenotypes). GSEA was performed using the C2 pathways collection from the Molecular Signatures Database (MSigDB), which is a collection of gene sets for use with GSEA software. Microarrays were scanned at 16 bits using the ScanArray Express (Packard Instrument Co.). The microarrays were then screened for quality, first by visual inspection of the array with flagging of poor-quality spots and then with automated scripts that scanned the quantified output files and measured overall density distribution on each channel and the number of flagged spots. Lowess normalization was then applied on the scanned chips.

RESULTS

Vaccination with 2 × MVA, but not 1 × MVA or 1 × MVA/Ad5, is associated with durable control of viral replication.

All vaccinated and control macaques were exposed intrarectally to a single high dose of SIV_mac251 as previously described (10, 11, 38) and outlined in Fig. 1A. All animals became infected. Plasma SIV RNA copies were measured from the time of exposure up to 24 weeks postchallenge (Fig. 1B to E). A peak of viral replication was observed at 2 weeks from exposure in all animals, as previously observed with this SIV_mac251 challenge stock (10, 11, 38). All vaccinated groups experienced significantly lower levels in plasma at week 2 during primary infection than the mock-vaccinated animals (for all groups, P = 0.0051; Fig. 1B to H). The mean virus levels over weeks 2 to 16 postinfection were significantly lower than in the control group in the 2 × MVA group only (P = 0.0070, by the Wilcoxon rank-sum test) (Fig. 1G). No significant differences were observed in mean virus levels in plasma among the 2 × MVA, 1 × MVA/Ad5, and 1 × MVA groups at any time point analyzed (data not shown). The frequency of CCR5⁺CD4⁺ T cells was assessed in mononuclear cells isolated from blood, peripheral lymph nodes, rectal mucosal pinch biopsy specimens, and bronchoalveolar lavage fluids collected at week 24 postinfection from 5, 6, and 4 animals in the 2 × MVA and 1 × MVA/Ad5 groups (data not shown). Despite the lack of significant difference in virus levels in plasma among the groups of vaccinated animals, there was a significantly higher percentage of CCR5⁺CD4⁺ T cells in the rectal mucosa of the animals in the 2 × MVA group than in the controls (P = 0.042), while no differences were found for the 1 × MVA/Ad5 group (data not shown). The frequency of CD4⁺ T cells in the 2 × MVA group was also significantly higher in the bronchoalveolar lavage fluids than in the 1 × MVA/Ad5 group and the controls (P = 0.017 and 0.007, respectively, by the Wilcoxon rank sum test) (data not shown). These data suggest a better regeneration of CD4⁺ T cells at mucosal sites in animals vaccinated with the 2 × MVA vaccine.

The 2 × MVA vaccine induced significantly higher anti-gp120 titers than 1 × MVA or 1 × MVA/Ad5 regimens but equivalent frequencies of T-cell responses.

Serum binding antibodies to the Gag and Envelope proteins were measured by enzyme-linked immunosorbent assay (ELISA). At 3 to 4 weeks after the last immunization, antibody titers to p27 Gag did not differ significantly among the vaccinated groups (Fig. 2A). In contrast, we found significantly higher antibody titers to gp120 in the 2 × MVA-immunized group (P = 0.044 across the three groups (P = 0.029 between 2 × MVA and 1 × MVA/Ad5), whereas the difference only approached statistical significance between the 2 × MVA and 1 × MVA groups (P = 0.07; Fig. 2B). The titers of binding antibodies to SIV gp120, however, did not correlate significantly with the virus levels in plasma (Fig. 2C).

Fig 2.

Antibody titers and kinetics of CD8⁺CM9⁺ T-cell response in the vaccinated macaques. Endpoint antibody titers to p27 Gag (A) and gp120 (B) in the sera of the immunized monkeys. (C) Correlative analysis between the titers of the gp120-binding antibodies in all the vaccinated animals and plasma viral loads at week 6 postinfection. (D to G) Frequency and kinetics of CM9⁺CD8⁺ T cells in blood of the macaques during immunization and following challenge exposure to SIV_mac251. (H to K) Relative frequency of CD95⁺CD28⁻ (effector memory [EM] in solid black bars) and CD95⁺CD28⁺ (central memory [CM] in white bars) CM9⁺CD8⁺ T cells in blood of vaccinated and control macaques following challenge exposure to SIV_mac251. The data are shown as means for each group, calculated using the individual frequency of EM or CM on the total Gag-CM9⁺ memory population.

Prior to the final boost with either MVA-SIV or Ad5-SIV at week 53 (Fig. 1A), the DNA/MVA regimens elicited equivalent T-cell responses, as shown by measurement of T-cell proliferation to gp120 and Gag in the 2 × MVA group and 1 × MVA/Ad5 group on freshly collected blood 1 week following the week 26 immunization and up to the time of the last boost (see Fig. S1A in the supplemental material). Similarly, the ability of PBMC to produce IFN-γ in an ELISpot assay or of CD4⁺ or CD8⁺ T cells to produce IFN-γ and IL-2 in an intracellular cytokine staining (ICS) assay following in vitro stimulation with overlapping Gag, Env, Nef, and Pol peptides did not differ between the 2 × MVA and 1 × MVA/Ad groups (see Fig. S1B to D in the supplemental material). None of these T-cell responses correlated significantly with protection from high virus levels in plasma (data not shown).

Sustained expansion of Gag tetramer CM9⁺CD8⁺ T effector cells in 2 × MVA-vaccinated macaques.

We studied the frequency of CD8⁺CM9⁺ Gag-specific T cells in the blood of the Mamu-A*01⁺ animals included in each group and found that the overall frequency of the CM9⁺CD8⁺ T cells did not differ significantly in the three vaccinated groups after the first MVA boost (week 26), as expected (Fig. 2D to F). The administration of either an additional MVA or an additional Ad5 vaccine did not further increase these responses (Fig. 2E and F). Following challenge exposure, the overall frequency of CM9⁺ Gag-specific tetramer cells (area under the curve) did not differ among the three vaccinated groups. However, the CD8⁺CM9⁺ population that was still expanding at week 16 or 18 postchallenge in the 1 × MVA and 2 × MVA groups (Fig. 2D and E) had already begun to contract in the 1 × MVA-Ad5 group (Fig. 2F). As expected, the kinetics of expansion of CD8⁺CM9⁺ T cells was delayed in the mock-vaccinated group and was already contracted by week 16 (Fig. 2G).

The kinetics of expansion of CD8⁺CM9⁺ T cells in the blood suggested either differences in trafficking to tissues or that the vaccine regimens, in addition to eliciting different levels of antibody responses to gp120 (Fig. 2B), may have elicited CD8⁺CM9⁺ T cells with a different proliferative potential. To investigate this hypothesis, we analyzed the memory phenotype of the CM9⁺CD8⁺ T-cell population in the blood of the vaccinated animals. We found a marginally significant larger and more-sustained expansion of effector memory (EM; CD28⁻CD95⁺) CM9⁺ T cells in macaques immunized with the 2 × MVA than in macaques immunized with the 1 × MVA/Ad5 regimen vaccine (P = 0.051, by repeated measures analysis of variance at all postchallenge time points tested) (Fig. 2I and J). The 1 × MVA/Ad5 group macaques showed a more balanced response of both EM and central memory (CM; CD28⁺CD95⁺) CM9⁺ T cells (Fig. 2H). Thus, boosting with MVA or adenovirus did not induce immune responses that could be differentiated by the frequency of IFN-γ or by the overall frequency of total tetramer⁺ CD8⁺ T cells in Mamu-A*01⁺ macaques; rather, the data suggested a possible difference in the ability of antigen-specific cells to proliferate and expand following challenge exposure.

Differential PD-1 expression on antigen-specific cells from vaccinated macaques.

To investigate the mechanism behind this observation, we analyzed Gag CM9⁺ CD8⁺ T cells for the expression of Programmed death-1 (PD-1), a critical mediator of virus-specific CD8⁺ T-cell exhaustion (39), in the six Mamu-A*01⁺ macaques included in the 2 × MVA and the 1 × MVA/Ad5 groups.

Gag-CM9⁺ T cells were collected from the Mamu-A*01⁺ animals in 2 × MVA and 1 × MVA/Ad-5 groups at week 74 (before the challenge) and at weeks 7 and 24 following challenge exposure. Cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) and stimulated with a pool of Gag peptides. The level of expression of PD-1 on nonproliferating CFSE^hi CM9⁺ cells from macaque 906 in the 2 × MVA group and macaque 221 in the 1 × MVA/Ad5 group is shown in Fig. 3A. The levels of expression of PD-1 in both proliferating (CFSE^lo) and nonproliferating (CFSE^hi) CM9 tetramer⁺ T cells before (week 74, day 0) and after (weeks 7 and 24) challenge exposure to SIV_mac251 (Fig. 3B to D) were lower in macaques immunized with 2 × MVA than in macaques immunized with 1 × MVA/Ad5 (Fig. 3B to D) when the mean fluorescence intensity (MFI) values for PD-1 were compared at all time points (P = 0.003 for CFSE^hi and P = 0.05 for CFSE^lo; Fig. 3E). PD-1 levels on CFSE^hi and CFSE^lo cells from mock-vaccinated macaques were comparable to those of 1 × MVA/Ad5-vaccinated macaques (data not shown). Importantly, in vitro treatment with blocking antibody for PD-L1 (40, 41), the natural ligand for PD-1, significantly restored proliferation in Gag-specific CD8⁺ (Fig. 3F) and CD4⁺ (Fig. 3G) T cells isolated from lymph nodes of Mamu-A*01 and non-Mamu-A*01 macaques (4 from the 2 × MVA group, 3 from the 1 × MVA/Ad5 group, and 4 from the control group) at week 24 postinfection. These results confirm that the interaction of PD-L1 with PD-1 negatively regulates T-cell proliferation.

Fig 3.

PD-1 expression on CD8⁺CM9⁺ T-cell response in the vaccinated macaques. (A) CFSE gating strategy and mean fluorescence intensity (MFI) of PD-1 expression on nonproliferating CFSE^hi CM9⁺CD8⁺ T cells from representative macaques. (B to D) PD-1 level of expression on CFSE^lo or CFSE^hi CM9⁺ CD8⁺ T cells from animals from the 2 × MVA and 1 × MVA/Ad5 groups were analyzed at week 74 (before challenge) and at weeks 7 and 24 after challenge exposure to SIV_mac251. (E) PD-1 expression levels on low and high proliferating CD8⁺CM9⁺ T cells in the two groups of vaccinated animals at all time points. The plot includes all the data points, and each dot represents one macaque. (F and G) Frequency of Gag-specific T cells 5 days following stimulation with Gag peptides in the presence (or absence) of anti-PD-L1 blocking antibody; CD8⁺ T cells (F) and CD4⁺ T cells (G) are gated on CFSE^low cells. (H to J) PD-L1 expression on human mDCs (H) and on pDCs (J) and CD86 expression on human mDCs (I) following in vitro infection with Ad5 or MVA.

Next, we investigated the ability of the vaccine vectors used in this study to increase the expression of PD-L1 on dendritic cells. Infection of in vitro-matured human myeloid dendritic cells (mDC) by nonreplicating Ad5 significantly enhanced the level of PD-L1 expression compared to MVA-infected mDCs (Fig. 3H), and the CD86 maturation marker was significantly increased by infection with MVA but not Ad5 (Fig. 3I). No difference in PD-L1 expression was observed following exposure of pDC to either MVA or Ad5 (Fig. 3J).

SIV-specific CD8⁺ T cells elicited by the 1 × MVA/Ad5 vaccine exhibit a more proapoptotic gene expression profile than do those elicited by the 2 × MVA regimen.

To better define and confirm the qualitative differences in the SIV responses induced by these two vaccine regimens, we determined the gene expression profile of SIV-specific CM9⁺ CD8⁺ T cells. We obtained total RNA from CD8⁺CM9⁺ cells sorted from lymph nodes collected from 2 Mamu-A*01⁺ animals in the 1 × MVA/Ad5 group and 3 animals in the 2 × MVA group at week 24 postinfection and analyzed it by microarray (Table 1). The expression of genes involved in survival and apoptosis was analyzed using both the gene set enrichment analysis (GSEA) and the ingenuity pathway analysis. The significantly upregulated and downregulated pathways are presented in Fig. 4 and in Fig. S2 in the supplemental material. Figure 4 (upper panel) illustrates 8 significant pathways (P values < 0.05), 6 upregulated in 1 × MVA/Ad5 CM9⁺ cells and 2 upregulated in 2 × MVA CM9⁺ cells. Within those pathways, we have selected the most up- and downregulated genes; 27 were upregulated in 1 × MVA/Ad5 CM9⁺ cells, and 8 were upregulated in 2 × MVA CM9⁺ cells. The genes within the 6 upregulated pathways in the 1 × MVA/Ad5 group were involved mostly in apoptosis, cellular anergy and inactivation, or transcriptional suppression as indicated in Fig. 4 (upper panel) and Table 1. Of particular importance are genes like PPP2R5C, ATF2, ATF4, PTEN, MAP3K4, MAP3K7, MAPK1, NIN, RAP1, and YWHAZ, which are known to have proapoptotic activity and to be implicated in the negative control of cell growth and division (42–44). On the other hand, genes within pathways upregulated in 2 × MVA such as TAPBP, HLA-DQB1, HLA-DMB, CIITA, IRAK1, IGBP1, and CCND2 are involved in major histocompatibility complex (MHC) formation and presentation, indicating immune activation as well as cell persistence (45–47). The upregulation of genes involved in MHC class II maturation and peptide presentation were also observed independently when we analyzed the data using the ingenuity approach (Fig. 4, lower panel). Collectively, the data from two independent analysts suggest that the CM9⁺ cells in the 1 × MVA/Ad5-immunized animals could be more prone to apoptosis than those in the 2 × MVA-immunized animals.

Table 1.

Definition of pathways and genes differentially expressed in Adv versus MVA Tet⁺ cells

Pathway or gene name	Function or description
Pathways
HSA04010_MAPK_SIGNALING_PATHWAY	Genes involved in MAPK signaling pathway
HSA04540_GAP_JUNCTION	Genes involved in gap junction
M_2.9_ERK_TRANSACTIVATION_CYTOSKELETAL_MAPK_JNK	Progressively downregulated through 12 h following treatment of WS1 human skin fibroblasts with UVC at a high dose
UVC_HIGH_D4_DN	Downregulated at any time point following treatment of XPB/CS fibroblasts with 3 J/m2 UVC
UVC_XPCS_ALL_DN	Downregulated at 4 h following treatment of XPB/CS fibroblasts with 3 J/m2 UVC (genes involved in apoptosis)
UVC_XPCS_4HR_DN	Upregulated in mouse hematopoietic stem cells
HSC_HSCANDPROGENITORS_FETAL	Genes involved in antigen processing and presentation
HSA04612_ANTIGEN_PROCESSING_AND_PRESENTATION	Genes involved in MAPK signaling pathway
Genes
Upregulated in Tet+ MVA compared to Tet+ Adv
TAPBP	Required for MHC formation
CIITA	Involved in transcription of HLA class II and important for antigen presentation
HLA-DQB1	Involved in transcription of HLA class II and important for antigen presentation
HLA-DMB	Helps to load HLA class II
IRAK1
IGBP1	B-cell activation
PSMC5	Important for MHC-I processing
CCND2	Important for cell division
Uppregulated in Tet+ Adv compared to Tet+ MVA
MAP3K4	Apoptosis
JUN	Apoptosis
ATF4	Apoptosis
ATF2	Apoptosis
RAP1A	Plays a role in apoptosis
DUSP5	Apoptosis
TAOK1	Apoptosis
PTEN	Tumor suppressor
ITPR1	Apoptotic gene
TUBB2B
PRKACB	Apoptosis
PRKCB1	Apoptotic gene
SOS1	Differential role
UBE2G1	Ubiquitination
POGZ	Gene suppressor
ARID5B	Gene repression
PPP2R5C	Negative control of cell division
TRIM33	Gene repression
SPEN	Gene transcription repression
CDYL	Oncogene repression
CNOT4	Transcription repression
CBLB	Anti-inflammatory

Fig 4.

Heat map representation of gene expression within CM9⁺ CD8⁺ T cells at 24 weeks following challenge. (Upper panel) Genes from representative pathways analyzed with GSEA and displayed as a heat map to demonstrate differential expression between Adv CM9⁺ and MVA CM9⁺ cells. Each pathway comprising selected top upregulated genes is represented by a different color. The false color expression in log₂ scale is depicted on the right side of the figure. The heat map shows the expression level of each gene (red, upregulated; blue, downregulated). Eight significant pathways are shown (P values < 0.05) (6 upregulated in Ad5 CM9⁺ samples and 2 upregulated in MVA CM9⁺ cells), as is a subset of 30 most up- and downregulated genes (22 upregulated in Ad5 CM9⁺ samples and 8 upregulated in MVA CM9⁺ samples). (Lower panel) Genes upregulated and downregulated in antigen presentation pathways.

To determine if the differences observed in proapoptotic gene expression among the Mamu-A*01⁺ macaques could explain the greater level of sustained viral control observed in all the 2 × MVA-vaccinated animals (Fig. 1G) but not in all the 1 × MVA/Ad5-vaccinated macaques (Fig. 1H), we examined viral loads in just the Mamu-A*01⁺ animals (see Fig. S1E in the supplemental material). No difference in viremia control was seen between groups (Fig. S1E). Moreover, 1 × MVA-Ad5-immunized macaques that exhibited proapoptotic gene expression (Fig. 4, upper panel) were not correlated with viremia control. Macaque 907 exhibited a high persistent viral load, whereas macaque 221 controlled viremia. Similarly, among the 2 × MVA-vaccinated macaques that displayed profiles indicative of immune activation and cell persistence, only macaque 902 controlled viremia, whereas macaques 906 and 314 exhibited high viral loads (see Fig. S1E in the supplemental material). Thus, more-complex interactions must be involved in the control of viremia in these macaques, as discussed below.

DISCUSSION

Heterologous prime-boost vaccination has been shown to enhance immune responses and control viral replication (48). MVA and nonreplicating Ad5 vectors have the capability to infect immature dendritic cells, but they may differently influence the ability of dendritic cells to present antigens (49–51). In the case of nonreplicating Ad5, dendritic cell activation has been demonstrated to depend on the recognition of a heparin-sensitive receptor recognized by the Ad5 fiber, the Shaft (52), suggesting that the same mechanisms of activation of dendritic cells may also be used by the replicating Ad5 h used in our study. We reasoned that because the DNA/MVA platform induces a high frequency of CD4⁺ T-cell responses and Ad5 induces potent CD8⁺ T-cell responses (8, 33, 49–51), the combination of a DNA prime followed by boosts with the two vectors could increase immune responses to Gag and enhance protection (53). In this study, we demonstrated that boosting DNA vaccines with two consecutive doses of MVA that express the Envelope protein (2 × MVA) elicited a significantly higher binding antibody response to the SIV envelope protein and conferred more-durable protection against SIV_mac251 replication than boosting with 1 × MVA followed by an Ad5 h vector that did not express the Envelope protein (1 × MVA/Ad5). This result was expected, since repeated inoculation with MVA vectors expressing the envelope protein boosts antibody responses (7).

T-cell frequency and functionality, measured by ELISpot following in vitro stimulation with SIV Gag or by proliferative responses to Gag and Env, were comparable between the two groups before and after the second boosts with either MVA or Ad5 h. However, following challenge exposure and infection, there was a clear difference in the kinetics of expansion and durability of CM9⁺ T memory cells in blood of the vaccinated Mamu-A*01⁺ macaque groups. While SIV-specific cells within the 1 × MVA/Ad5 group peaked at 2 weeks following challenge exposure to SIV_mac251, those of the 2 × MVA group were still expanding and did not peak until week 16 following challenge exposure. The memory CM9⁺ T cells appeared to decrease faster in the 1 × MVA-Ad5 group than in the 2 × MVA or the 1 × MVA group.

The 2 × MVA vaccination regimen induced a significantly lower level of PD-1 expression on SIV-specific cells when all the time points were taken together (Fig. 2E). But most striking were the distinctly different gene expression profiles associated with apoptosis seen between the SIV-specific T cells in the 2 × MVA compared to the 1 × MVA/Ad5 macaques that carried the Mamu-A*01 allele. The decreased activation of several proapoptotic pathways in the 2 × MVA animals suggested a reasonable explanation for the different rate of expansion and contraction of Gag-specific effector cells in this group.

PD-1 is believed to be a major regulator of apoptosis (39) that can impact the frequency of HIV-specific CD8⁺ T cells and could be manipulated to improve CD8⁺ T-cell survival and function (28, 54, 55). Upregulation of PD-1 expression on T cells and PD-L1 expression on DCs (28, 55) could be the consequence of SIV_mac251 continuous replication in the infected animals, since at these time points, all animals were viremic, although the 2 × MVA animals controlled viremia better than did the control macaques. However, because the low level of PD-1 on SIV-specific cells was also observed before exposure to SIV_mac251 (prechallenge time point week 74 [day 0]; Fig. 4, lower panel) and persisted after infection in the 1 × MVA/Ad5 group, we believe that differential PD-1 expression may be due to immunogenic differences between Ad5 and MVA. Indeed, our data using the nonreplicating Ad5 demonstrated the elicitation of higher expression of PD-L1 on human myeloid dendritic cells in vitro than MVA.

PD-L1 has been shown to interact with B7-1 costimulatory molecule to inhibit T-cell responses. Extensive studies have shown that a PD-1/PD-L1 blockade restores exhausted T cells during chronic viral infections and in tumors (28, 55). Song et al. evaluated the effects of soluble PD-1 (sPD-1) as a blockade of PD-1 and PD-L1 on vaccine-elicited antigen-specific T-cell responses in mice (56). Coadministration of sPD-1 with an Ad-based vaccine could increase antigen-specific CD8⁺ T-cell responses, indicating a vaccine type-independent adjuvant effect of sPD-1. This coadministration of sPD-1 DNA augmented T-cell proliferation and reduced T-cell apoptosis through upregulation of Bcl-xL expression during T-cell activation.

Gene array analysis revealed the activation of proapoptotic pathways within SIV-specific T cells of the 1 × MVA/Ad5 group 24 weeks following challenge. This is in accordance with the lower frequency of the CM9⁺CD8⁺ T cells at the specified time points in the 1 × MVA/Ad5 group. This apoptotic state could be triggered by PD-1-mediated apoptosis and cell contraction (41, 57). Genes within the 5 pathways upregulated in the Ad5-vaccinated group were mostly genes involved in apoptosis, cellular anergy, and inactivation or were transcription suppressor genes. PPP2R5C is implicated in the negative control of cell growth and division (54, 57). ATF2 and ATF4 have proapoptotic activity and are involved in the upregulation of Fas and FasL (56). PTEN is known to inhibit AKT phosphorylation, involved in phosphorylation of FOXO3a in different cell models (58) as well as it negatively regulates the activation of NF-κB and Iκ kinase β (IKKβ) (59). In fact, both PTEN and DUSP5, also upregulated in the Ad5 group, are upstream of FOXO3a and are believed to inhibit its phosphorylation, leading to a reduction in cell survival. Other genes that are also upregulated and are part of the mitogen-activated protein kinase (MAPK) or extracellular signal-regulated kinase (ERK) pathway are MAP3K4 and MAP3K7 (both activate the JNK MAPK pathways in response to cellular stress). MAPK1, NIN, RAP1, and YWHAZ are all involved in many vital cellular processes such as metabolism, protein trafficking, signal transduction, apoptosis, and cell cycle regulation. Moreover, we have preliminary results indicating that the PD-1 inhibitory effect is mediated through dephosphorylation of FOXO3a, which in turn upregulates expression of proapoptotic FOXO3a target genes such as Fas ligand and Bim genes (R. P. Sekaly, unpublished data). Therefore, the high level of expression of PD-1 on SIV-specific CD8⁺ T cells in the 1 × MVA/Ad5 group may drive the upregulation of the above genes, leading to cellular anergy and death.

Nevertheless, the lack of correlation of viremia control with the proapoptotic gene expression profiles does not allow us to conclude that this mechanism contributed to the lesser protection during the chronic phase in the 1 × MVA/Ad5 group. Microarray analysis was conducted only on 2 Mamu-A*01 macaques in this group, and more extensive studies may still be needed to prove this hypothesis. However, other factors may have offset the impact of any enhanced apoptosis or exhaustion. We note that the 1 × MVA/Ad5-immunized macaques exhibited a balanced T-cell EM and CM response (Fig. 2I). Both EM (60) and CM (61) T cells have been shown to be important for vaccine-elicited protection. CM T cells are important for replenishing memory T-cell populations. It is possible that gene expression profiles of the 1 × MVA/Ad5-vaccinated macaques reflect in part the restraint needed to prevent uncontrolled CM T-cell expansion.

An alternative explanation for the enhanced protection of the 2 × MVA group was suggested by the gene expression profiles. The pathways upregulated in the 2 × MVA group included the antigen processing and presentation pathway as well as the hematopoietic stem cells (HSC) and progenitors pathway, all involved in MHC-I and -II maturation and peptide presentation. TAPBP is involved in the association of MHC-I with transporter associated with antigen processing (TAP), which is required for the assembly of MHC-I with peptide (peptide loading). Both HLA-DQB1 and HLA-DMB belong to the HLA class II beta chain paralogues and play a central role in the immune system by presenting peptides derived from extracellular proteins (62). CIITA is essential for transcriptional activity of the HLA class II promoter (46). Upregulated genes within the HSC and progenitor pathway (Fig. 4, upper panel) are involved in cell survival and persistence. IRAK1 binds to the IL-1 type I receptor following IL-1 engagement, triggering intracellular signaling cascades leading to transcriptional upregulation and mRNA stabilization. IGBP1 is involved in the proliferation and differentiation of B cells (22), while CCND2 is essential for the control of the cell cycle (45). This upregulation of genes involved in MHC-II maturation and peptide presentation was also scored when the gene data were analyzed using the ingenuity approach (Fig. 4, lower panel). In fact, the MHC-II presentation pathway was one of three gene expression pathways that were significantly differentiated using that approach (see Fig. S2 in the supplemental material). The higher titers of binding antibodies to gp120 elicited by vaccination in the 2 × MVA group, which per se do not correlate with better control of plasma virus levels, may rather have enhanced antigen uptake through complex formation and stimulation of the Fcγ receptor, thus influencing the quality of the secondary T-cell response to SIV_mac251 (63, 64). The microarray data that demonstrate upregulation of genes involved in MHC-I and II antigen presentation in the 3 Mamu-A*01 macaques of the 2 × MVA group support this hypothesis (Fig. 4, lower panel), although further studies on larger numbers of animals will be necessary to confirm it.

Poxvirus-Ad vector combinations are being increasingly explored in HIV vaccine design, and dissecting immune correlates of protection is difficult. In both mice (65) and macaques (66), the order of heterologous vector administration influences the memory profile of virus-specific T cells. As suggested here, an envelope protein boost, in addition to providing potentially protective antibodies, may also impact T-cell responses. It will be important to elucidate whether this potential mechanism influences vaccine protective efficacy.