Phase 1 Human Immunodeficiency Virus (HIV) Vaccine Trial to Evaluate the Safety and Immunogenicity of HIV Subtype C DNA and MF59-Adjuvanted Subtype C Envelope Protein

Abstract Background The Pox-Protein Public-Private Partnership is performing a suite of trials to evaluate the bivalent subtype C envelope protein (TV1.C and 1086.C glycoprotein 120) vaccine in the context of different adjuvants and priming agents for human immunodeficiency virus (HIV) type 1 (HIV-1) prevention. Methods In the HIV Vaccine Trials Network 111 trial, we compared the safety and immunogenicity of DNA prime followed by DNA/protein boost with DNA/protein coadministration injected intramuscularly via either needle/syringe or a needle-free injection device (Biojector). One hundred thirty-two healthy, HIV-1–uninfected adults were enrolled from Zambia, South Africa, and Tanzania and were randomized to 1 of 6 arms: DNA prime, protein boost by needle/syringe; DNA and protein coadministration by needle/syringe; placebo by needle/syringe; DNA prime, protein boost with DNA given by Biojector; DNA and protein coadministration with DNA given by Biojector; and placebo by Biojector. Results All vaccinations were safe and well tolerated. DNA and protein coadministration was associated with increased HIV-1 V1/V2 antibody response rate, a known correlate of decreased HIV-1 infection risk. DNA administration by Biojector elicited significantly higher CD4+ T-cell response rates to HIV envelope protein than administration by needle/syringe in the prime/boost regimen (85.7% vs 55.6%; P = .02), but not in the coadministration regimen (43.3% vs 48.3%; P = .61). Conclusions Both the prime/boost and coadministration regimens are safe and may be promising for advancement into efficacy trials depending on whether cellular or humoral responses are desired. Clinical Trials Registration South African National Clinical Trials Registry (application 3947; Department of Health [DoH] no. DOH-27–0715–4917) and ClinicalTrials.gov (NCT02997969).

Despite progress in human immunodeficiency virus (HIV) treatment, an estimated 1.8 million individuals were infected with HIV in 2017 [1], highlighting the need for an effective HIV vaccine. The RV144 vaccine trial is to date the only HIV vaccine trial that has demonstrated any efficacy [2]. The Pox-Protein Public-Private Partnership (P5) was established to improve on RV144 to create a more efficacious and durable vaccine in the predominately subtype C region of sub-Saharan Africa [3], where half the world's HIV-infected population (36.9 million) resides [1]. The RV144 regimen, designed to protect against subtype B/E HIV type 1 (HIV-1) strains, was reformulated for the subtype C virus [4] and has demonstrated adequate immunogenicity [5]. Comprised of subtype C-adapted ALVAC and bivalent glycoprotein (gp) 120 protein with MF59 adjuvant (gp120/MF59), this vaccine regimen is currently under efficacy evaluation in South Africa (HIV Vaccine Trials Network [HVTN] 702). In parallel, the P5 correlates program is conducting several phase 1/2a trials designed to provide insights into potentially superior vaccine candidates and/or regimens based on favorable immune profiles. These phase 1/2a trials evaluate alternative strategies, adjuvants, and products for vaccine delivery, using common immunological end points.
HVTN 049 compared a DNA prime/gp140 protein boost regimen with a protein-only regimen for inducing humoral and cellular responses to HIV antigens [6]. The prime/boost strategy resulted in higher levels of envelope (Env) binding antibodies (bAbs) and homologous neutralizing antibodies (nAbs), greater CD4 + T-cell responses to Env antigens, and greater polyfunctionality of CD4 + T-cell responses. In nonhuman primate and mouse models, coadministration of DNA/protein elicited more robust humoral immunity than DNA alone or a prime/boost strategy [7,8].
The most common vaccine administration method is via needle and syringe (needle/syringe). An alternative to needle/ syringe is the Biojector, a needle-free injection system that results in improved interaction with antigen-presenting cells and enhanced immunological responses [9,10]. The first of the P5 correlates program trials, HVTN 111, was designed to determine whether DNA vaccine administration method and schedule influence immune responses. Specifically, HVTN 111 evaluated the safety and immunogenicity of subtype C DNA-HIV-PT123 administered via needle/syringe or Biojector, and either boosted or coadministered with bivalent subtype C gp120/MF59 in HIV-uninfected, healthy adults in sub-Saharan Africa.

Study Design
HVTN 111 was a phase 1 randomized, placebo-controlled study with 6 arms according to administration method (needle/ syringe or Biojector) and prime/boost versus coadministration approaches. Participants were randomized to 1 of 6 regimens, according to the schema in Table 1. The DNA vaccine was administered via needle/syringe or Biojector, according to the regimen; adjuvanted protein vaccinations were administered via needle/syringe. When coadministered, the DNA and protein vaccines were administered contralaterally. The administration schedule in prime/boost via needle/syringe and prime/boost via Biojector matched the primary regimens of RV144 and HVTN 702. Coadministration via needle/syringe and coadministration via Biojector received 3 coadministrations of the 2 products, following a classic month 0-1-6 schedule (tetanus, diphtheria, hepatitis B). This approach is based on the observation that peak humoral immunogenicity is reached after 3 doses of protein, with no additional benefit of a fourth dose [11], and dampening of immunoglobulinin (Ig) G3 responses with additional protein administrations (unpublished data). The longer rest between the second and third vaccinations is preferred for antibody maturation.

Study Participants
HVTN 111 enrolled 132 healthy, HIV-negative adults, aged 18-40 years, assessed as low risk for HIV infection, who agreed to study requirements and provided written informed consent in their preferred language. Good general health was determined by medical history, physical examination, and laboratory tests, including hematology, chemistry, and hepatitis serology. All female participants agreed to consistent contraception; pregnant and breastfeeding women were excluded. Eligibility criteria are described in Supplementary Table 1. Participants were enrolled at 1 site in Zambia (Matero), 3 sites in South Africa (Isipingo, Klerksdorp, and Tembisa), and 1 site in Tanzania (Mbeya). HVTN 111 was approved by the research ethics committee of the participating sites. The study was registered with the South African National Clinical Trials Registry (application 3947; Department of Health [DoH] no. DOH-27-0715-4917) and ClinicalTrials.gov (NCT02997969).

Study Products
The DNA vaccine, designated DNA-HIV-PT123 (IPPOX Foundation), comprises a mixture of 3 DNA plasmids in a 1:1:1 ratio, each at 1.33 mg: (1) subtype C ZM96 gag, (2) subtype C ZM96 gp140, and (3) subtype C CN54 pol-nef. The DNA vaccine was delivered at a total dose of 4 mg. The bivalent subtype C gp120 Env protein vaccine comprises subtype C TV1.C gp120 Env and subtype C 1086.C gp120 Env (GlaxoSmithKline Biologicals), each at a dose of 100 µg. The protein vaccine was mixed with MF59 adjuvant (Seqirus). The placebo was 0.9% sodium chloride for injection.

Study Procedures
Participants were followed up for 12 months after the initial vaccination with safety evaluations and procedures as per Supplementary Methods. Adverse events (AEs) were reported over 30 days after each vaccination visit, with a subset of AEs being reported for the duration of the study.

Immunogenicity Assays
All laboratory assays were performed blinded to treatment group with validated and qualified methods published elsewhere. Measurements included bAb, nAb and T-cell responses at the peak immunogenicity time point (2 weeks after final vaccination). A list of the specific antigens used in all immunogenicity assays is found in Supplementary Table 2.

Intracellular Cytokine Staining Assay
Peripheral blood mononuclear cells, collected at baseline and the peak immunogenicity time point, were isolated and cryopreserved from whole blood, as described elsewhere [17]. T-cell responses to vaccine-matched antigens were measured by intracellular cytokine staining (ICS), as described elsewhere [18,19] (Supplementary Methods). The 17-color ICS panel is described in Supplementary Table 3.

Randomization
The randomization allocation sequence was obtained by computer-generated random numbers and provided to each site through the HVTN statistics and data monitoring center's Webbased randomization system.

Positivity Calls and Response Comparisons
Barnard exact and Wilcoxon rank sum tests were used to compare the response rates and magnitudes for responders, respectively, between 2 groups [20]. Two-sided 95% confidence intervals for binomial proportions were calculated using the Wilson score method [21]. All tests were 2 sided with no adjustment for multiple comparisons, and differences were considered statistically significant at P <.05. For positivity analyses of ICS and bAb multiplex assays, see the Supplementary Methods.

T-cell Polyfunctionality Analyses
Combinatorial Polyfunctionality Analysis of Single Cells (COMPASS) analysis was used to analyze antigen-specific T-cell subsets [22]. The functionality score (FS) is defined as the estimated proportion of antigen-specific subsets detected among all possible ones. The polyfunctionality score (PFS) is similar but weighs the different subsets by their degree of functionality, naturally favoring subsets with higher degrees of functions, motivated by the observation that higher-degree function has been correlated with good outcomes in certain vaccine studies (Supplementary Methods).

Study Population and Schema
One hundred thirty-two healthy, HIV-uninfected, low-risk participants were enrolled at 5 sites between 21 June 2016 and 13 July 2017 (Table 2). Thirty participants were allocated to each of 4 vaccine arms and 6 participants to each placebo arm. Overall, 123 of 132 participants (93%) completed all vaccinations and follow-up ( Figure 1).

Safety and Tolerability
All 132 participants received the first vaccinations, 127 received the second and third vaccinations, and 128 received the fourth vaccinations. All vaccinations were well tolerated. No related serious AEs, related severe AEs, or severe local reactogenicity symptoms were reported. Apart from 1 case of grade 3 elevated temperature, no severe systemic reactogenicity symptoms were reported (Supplementary Figure 1) No significant differences were observed in local and systemic reactogenicity symptoms between vaccine and placebo recipients, overall. When the DNA and protein injections were analyzed separately, pain and/or tenderness were more common in some of the vaccine groups when compared with placebo for DNA injections (Supplementary Figure 2); the same was true for protein injections. Pain, tenderness, and pain and/or tenderness were more common with Biojector than needle/syringe administration. For DNA injections via needle/syringe, the median severity grade was "none, " whereas for DNA injections via Biojector, the median severity grade was "mild" (Figure 2). Four participants discontinued receiving vaccinations, 3 because of relocation or refusal, and 1 because of an undisclosed pre-existing condition of diabetes. Overall, 81 AEs were reported; 78 of them mild or moderate. One grade 3 AE for type 1 diabetes mellitus, 1 grade 4 AE for increased aspartate aminotransferase, and 1 grade 4 serious AE for intentional self-injury were reported, all deemed not related to study product. Two AEs were deemed related to study product, both mild in severity: 1 case of decreased neutrophil count and 1 case of malaise. One participant reported injection site blistering at the site of DNA administration via Biojector. Blistering occurred after the second, third, and fourth vaccinations, and resolved spontaneously.

Humoral Responses
In all vaccine recipients (100%) IgG responses developed to vaccine-matched Env gp120/gp140 (1086.C gp120, TV1c8.2.C gp120, 96ZM651.C gp140) and the group M consensus gp120/ gp140 (Con 6 gp120, Con S gp140 CFI) at high magnitudes ( Figure 3 and data not shown). As expected, no positive responses were seen among placebo recipients. Responses to vaccine-matched 96ZM651.C gp140 were significantly higher in the coadministration group versus prime/boost group via needle/syringe (P = .007) ( Figure 3A); there were no significant differences between other treatment groups. The magnitude of the IgG response to both subtype B and C V1V2 antigens, known correlates of decreased HIV-1 risk in RV144 [12,15], trended higher in the coadministration group versus prime/ boost group via needle/syringe ( Figure 3B and 3C).
All vaccine recipients and no placebo recipients developed nAb responses against tier 1A strains TV1c8.2.C and MW965.26.C ( Figure 4). There were no positive responders against tier 2 vaccine strains in any group (data not shown). nAb responses in the coadministration via needle/syringe group were significantly higher in magnitude than those in the prime/boost via need/syringe group for TV1c8.2.C (P = .0004) ( Figure 4A) and MW965.26.C (P = .03) ( Figure 4B). The same pattern emerged in the Biojector groups between coadministration and prime/ boost: TV1c8.2.C (P = .002) and MW965.26.C (P = .03). There were no significant differences between nAb response magnitudes for TV1c8.2.C or MW965.26.C for vaccine groups of the same regimen but with different methods of administration.
CD8 + T-cell responses were infrequently induced in all groups (Supplementary Figure 3) with no statistically significant differences in response rates or magnitudes with needle/syringe or Biojector administration. When the prime/boost regimens were compared with each other, Biojector resulted in significantly higher response rates than needle/syringe for any HIV (33.3% vs 7.4%, respectively; P = .02) and Gag-ZM96.C (16.7% vs 0.0%; P = .03). There were no significant differences in CD8 + T-cell response rates between coadministration vaccines via the 2 administration methods, nor were there any differences in response magnitudes between prime/boost vaccinees (via needle/syringe or Biojector) or coadministration vaccinees (via needle/syringe or Biojector).
We performed a polyfunctionality analysis to assess IFN-γ, IL-2, tumor necrosis factor (TNF) α, CD40L, interleukin 4, interleukin 17, and granzyme B coexpression using COMPASS analysis [22]. Comparisons of FS and PFS between prime/ boost and coadministration via needle/syringe revealed significant differences only for the combined vaccine-matched Env ZM96 ("any Env ZM96") (P = .04 for FS) and the Env-1-ZM96.C (P = .03 for FS and P = .02 for PFS), with prime/boost via needle/syringe having higher FS and PFS ( Figure 6A). When prime/boost was compared with coadministration delivered by Biojector, FS and PFS were significantly higher in the prime/ boost group for any Env ZM96 (P < .001 for FS and P = .09 for PFS) ( Figure 6A) and for each of the individual Env peptide pools (Env.1086.C [P = .005 for FS and P = .004 for PFS], Env-1-ZM96.C [P = .008 and P = .01, respectively] and Env-2-ZM96.C [both P = .02]; data not shown). There were no significant differences in FS and PFS comparisons by administration route.
In Figure 6B, a heat map of the COMPASS posterior probabilities allows visualization of the different populations contributing to the FS. All vaccine regimens induced polyfunctional CD4 + T cells expressing 2, 3, and 4 functional combinations of IFN-γ, IL-2, TNF-α, and CD40L. Triple functional CD4 + T cells coexpressing IL-2, TNF-α, and CD40L or IFN-γ, TNF-α, and CD40L were the 2 dominant populations responding to any Env-ZM96 in all vaccine groups, regardless of administration method. In contrast, the 4-function subset expressing IFN-γ IL-2, TNF-α, and CD40L was more likely to be expressed in the prime/boost regimens. CD40L single-expressing cells were most common in the prime/boost group vaccinated via Biojector.

Vaccine-induced Seroreactivity
Vaccine-induced seroreactivity, assessed by means of commercial HIV serological methods, occurred in 1 vaccine recipient (0.8%). This individual was in the coadministration via Biojector group and tested reactive only with the Alere Determine HIV-1/2 Ag/Ab Combo test.

DISCUSSION
In the first reported clinical trial from the P5 correlates program, HVTN 111, the subtype C DNA vaccine (DNA-HIV-PT123) and the bivalent subtype C gp120/MF59 were generally safe and well tolerated in healthy volunteers in sub-Saharan Africa. Humoral responses were robust in all vaccination arms. Some cellular responses were enhanced with the prime/boost approach  compared with the coadministration regimen when administered via Biojector. Given that humoral responses were higher after coadministration and cellular responses higher after prime/ boosting, the potential use of these approaches may depend on whether humoral or cellular responses are more desirable. Notably, despite the lower CD4 + T-cell responses in the coadministration arms, the potent humoral responses indicate that T cells are sufficiently stimulated by vaccination to support antibody responses.
Local pain and/or tenderness were more common with administration via Biojector than with administration via needle/ syringe, albeit these reactions were mostly mild. One participant experienced self-limiting blistering at the injection site after Biojector administration, something that has been reported elsewhere at a low frequency [9]. Vaccinations with the bivalent subtype C gp120/MF59 were also safe and well tolerated.
The HVTN 111 candidate vaccine regimens elicited robust humoral responses in response rate and magnitude. The antibody response magnitude increased with coadministration of the protein and DNA components. The coadministration arms received 3 gp120/MF59 dosages compared with 2 in the prime/ boost approach. One objective of this study was to determine whether the frequency and magnitude of the V1V2 IgG correlate of decreased HIV-1 risk [12,14,15] could be differentially modulated. We observed that the response magnitude did trend higher in the coadministration arm compared with the prime/ boost arm. Consistent with expectations that Biojector administration largely influences cellular responses, we saw no difference in humoral responses with the different administration methods.
The prime/boost approach elicited higher cellular response rates than the coadministration regimen. The prime/boost strategy provides 1 additional DNA vaccine administration compared with the coadministration regimen, which may explain some of the observed differences. With respect to administration method, the Biojector enhanced CD4 + T-cell response rate and magnitude when compared to standard vaccination via needle/syringe, but only in the prime/boost regimen. These CD4 + T-cell response rates exceed those seen when the canarypox vector ALVAC is given with the same protein boost (Moodie Z et al, submitted).
The current study presented no safety concerns, which supports advancement of the suite of P5 vaccines. However, Biojector administration requires the Biojector device, carbon dioxide cartridges, and specific syringes that are not widely available and will require a more complicated supply chain management plan in a public health setting than needle/syringe strategies. The modest gains in cellular responses may be of questionable added value given the necessary complexities for implementation [12,15]. Coadministration may induce earlier protective humoral responses. A public health advantage to coadministration would include simplicity in dosing, with the same vaccine (DNA and protein) administered at each vaccination time point. This ease of use would need to be balanced with potential reduction in the cellular responses seen with the standard prime/boost approach. Importantly, both approaches were performed with the same bivalent subtype C proteins used in the ongoing efficacy trial HVTN 702. HVTN 111 represents the first time these particular proteins have been assessed with a different priming CD4 + T-cell responses, as measured by intracellular cytokine staining. Response rate (bar charts) and magnitude (box plots) 2 weeks after the final vaccination by treatment arm are shown for the following vaccine-matched peptide pools: any human immunodeficiency virus (HIV) (A), Any envelope (Env) (B), Env.1086.C (C), and Env-1-ZM96.C (D). Any HIV is the sum of any Pol, any Env, Nef-CN54, and Gag-ZM96.C, where any Pol is the sum of Pol-1-CN54 and Pol-2-CN54. Any Env is the maximum of Env ZM96, Env.1086.C, and Env.TV1.C, where Env ZM96 is the sum of Env-1-ZM96.C and Env-2-ZM96.C. Bar charts show positive response rates. Box plots show responses and are based on positive responders only (shown as colored circles); negative responders are shown as gray triangles. *P ≤ .05; †P ≤ .01; ‡P ≤ .001. Abbreviations: B, Biojector; IFN, interferon; IL-2, interleukin; S, needle/syringe; Pol, polyfunctionality. agent, and in different administration regimens. In this study, we confirmed the differentiated skewing of the humoral and cellular responses in the prime/boost and coadministration regimens. The results of this trial help pave the road to potential refinements of the HVTN 702 regimen with the aim to fine-tune the immunological responses deemed most critical for vaccine efficacy based on correlates of risk/protection. Ideally these targeted refinements can help shape a future vaccine to enhance levels of efficacy beyond those seen in RV144. Thus, this study sets the foundation for optimizing vaccine regimens based on observed correlates of risk/protection from existing HIV-1 efficacy trials, in concert with upcoming efficacy trial results.

Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.