Safety and immunogenicity of an HIV subtype B and E prime-boost vaccine combination in HIV-negative Thai adults

The HIV epidemic in Thailand consists predominantly of the circulating recombinant form(CRF) 01_AE, along with subtype B [1, 2]. To match these prevalent strains, candidate vaccines have been developed based on the gp120 from macrophage-tropic (R5) strains of CRF01_AE, in combination with antigens from subtype B HIV-1. These candidate vaccines, ALVAC-HIV (vCP1521) and AIDSVAX B/E, were used in combination and were assessed for safety and immunogenicity in adult Thai volunteers. Here we report the results of this prime-boost trial, which was performed at 2 academic centers in Bangkok, Thailand, where the volunteers were screened, immunized, and followed to determine safety and immune responses.

Volunteers, materials, and methods. Volunteers were healthy, HIV-negative Thai adults (20–50 years old). Women agreed to practice effective contraception before the first vaccination and to continue practicing for at least 3 months after the last vaccination. Informed consent was obtained from each volunteer; the human experimentation guidelines of the US Army Surgeon General's Human Subjects Research Review Board, the ethics review committee of the Ministry of Public Health of Thailand, and the institutional review boards of the Royal Thai Army Medical Department and of Mahidol University were followed in the conduct of this clinical research.

ALVAC-HIV (vCP1521), produced by Aventis Pasteur, is a recombinant canarypox vector vaccine expressing CRF01_AE HIV-1 gp120 (92TH023) linked to the transmembrane-anchoring portion of subtype B gp41 (strain LAI) with a deletion in the immunodominant region and also expressing HIV-1 Gag and protease (strain LAI). vCP1521 was formulated at a dose of 10^6.5 TCID⁵⁰. ALVAC placebo contained the excipient without canarypox vector.

AIDSVAX B/E, produced by VaxGen, is a bivalent HIV gp120 vaccine containing a B envelope from strain MN and a CRF01_AE envelope from strain A244. The recombinant gp120s were produced in CHO cell lines and were coformulated and administered in alum at a dose of either 200 µg (100 µg each) or 600 µg (300 µg each). AIDSVAX placebo consisted of alum in a dose amount that was equivalent to that of the vaccine dose.

Phase 1 of the study assessed the acute safety and tolerability of ALVAC-HIV alone, given at weeks 0, 4, 12, and 24. Phase 2 was a randomized, double-blind, placebo-controlled trial, with 2 groups. Group 1 volunteers were given ALVAC-HIV at weeks 0, 4, 12, and 24 and 200 µg of AIDSVAX B/E at weeks 12 and 24; group 2 varied in that volunteers received 600 µg of AIDSVAX B/E. In each group, 1 volunteer received placebo for every 3 volunteers who received vaccine. Study visits took place as follows: at 2 weeks before the first vaccination (V0), at the first vaccination (V1), at the second vaccination (V2), 2 weeks after the second vaccination (V3), at the third vaccination (V4), 2 weeks after the third vaccination (V5), at the fourth vaccination (V6), 2 weeks after the fourth vaccination (V7), 4 weeks after the fourth vaccination (V8), 12 weeks after the fourth vaccination (V9), and 24 weeks after the fourth vaccination (V10). Immunizations were given separately into deltoid muscles. Volunteers were assessed for vaccine reactogenicity for 7 days.

Clinical laboratory evaluations for safety were performed at V1, V2, V4, V6, V7, V9, and V10. Female volunteers had urine pregnancy tests performed before each immunization and at V9 and V10. HIV serologic testing was performed at V0, V5, V8, and V10, by use of standard methods. HIV serologic testing was followed by nucleic-acid testing (Roche Amplicor; version 1.5), if the serologic results were suggestive of infection.

Binding antibody was measured in serum samples collected at V1, V3, V5, V7, and V10, with p24 and Env proteins of interest; an EIA was used as described elsewhere [3]. End-point titers were determined as the highest serum dilution with EIA optical-density signals >2 × mean + 2 SDs of the individual preimmune serum samples (typical positive cutoff, ODof >l0.10). If a sample was nonreactive at the 1:50 dilution, a value of 25 was assigned for analysis.

Neutralizing antibody was measured in serum samples collected at V1 and V7, was heat inactivated, and was centrifuged for 10 min. Cell lines were infected with viral stocks and were propagated in tissue culture medium, as described elsewhere [4–6]. The B strains HIV-1 MN and SF2 and E strain NP03 (all syncytium-inducing viruses) were propagated in H9 cells, whereas E strain HIV-1 CM244 (a non-syncytium-inducing virus) was adapted to and grown in A3R5 T cells. Neutralization was performed as described elsewhere [4–6] at a final serum dilution of 1:10, with the percent reduction of p24 production calculated on the basis of a comparison between V7 and V1 serum activity. A neutralization titer of ⩾50% was defined as positive; each positive serum sample was titrated against that virus, and end points were estimated by quadratic projection.

Cytotoxic T lymphocyte (CTL) assays were used to determined chromium release from peripheral-blood mononuclear cells (PBMCs), as described elsewhere [7]. Autologous Epstein- Barr virus-transformed B cell lines (B-LCL) were established for each volunteer. CTL activity was assessed at V1, V3, V5, V7, V9, and V10, by use of recombinant vaccinia vectors expressing HIV-1 CRF01_AE TH023 gp160 (vP1536) and subtype B HIV-1 HXB2 Gag/Pol (vVK-1). Assays were performed after 13–15 days of culture with autologous B-LCL simultaneously infected with vP1536, vVK-1, or control vaccinia virus (vP1170) at an MOI of 5 as targets. Vaccinia vectors were commercially prepared (Advanced BioScience Laboratories). Vaccinia construct vT142 expressing subtype A Gag (90CR402.1; obtained through NIAID) was used for the assessment of cross-clade HIV Gag responses. Acceptable spontaneous release values were <33% maximal detergent lysis. A positive assay was defined as HIV-1-specific lysis of ⩾10% at 2 effector-to-target cell ratios. The result was considered to be CD8 specific when ⩾50% of the lytic activity was lost after CD8 depletion and there was ⩾5% specific lysis after CD4 depletion.

Lymphoproliferative responses of volunteers' PBMCs were measured at V1 and V7, as described elsewhere [5, 8]. Data are expressed as a stimulation index, calculated as (PBMC counts per minute + antigen/mitogen)/(PBMC counts per minute + medium). A positive stimulation index was defined as ⩾5.

Statistical tests were conducted as 2-sided tests at the 5% level of significance. Demographic, safety, and reactogenicity comparisons included all volunteers in an intent-to-treat analysis, although the percentages reported for reactogenicities include only volunteers who received at least 1 vaccination. To be included in immunogenicity results, volunteers must have completed all 4 vaccinations. Demographic characteristics were evaluated by use of a χ² test, to evaluate differences in sex and age between groups.

Results. Of the 133 volunteers who enrolled in the study, 122 completed it; 2 volunteers were diagnosed with HIV infection before vaccination, and 9 did not complete the vaccination series. The reasons for withdrawal were death (1 volunteer), medical problems (4 volunteers), withdrawn consent (2 volunteers), lost to follow-up (1 volunteer), and schedule conflict (1 volunteer). Of the volunteers, 36% were female and 64% were male; 70% were 20–30 years old, 24% were 31–40 years old, and 6% were 41–50 years old. Groups 1 and 2 were similar with regard to sex and age and when analyzed by vaccine versus placebo recipients.

Immunizations were well tolerated; reactogenicity is summarized in table 1. The majority of adverse events were mild to moderate in intensity. Four adverse events (none serious) were considered to be vaccine related: there were 2 cases of myalgia and 1 case of pruritis, and 1 volunteer had an erythematous rash. Seven serious adverse events occurred, none of them vaccine related (among vaccinees, death [suicide], hepatitis, vertigo, suicide attempt, and fracture of clavicle; among placebo recipients, appendectomy and gastroenteritis). There were no cases of HIV infection among immunized volunteers.

Assays for binding antibodies to gp120 MN or A244 were negative at V3. Of group 1 vaccinees, 95% had anti-MN and 86% had anti-A244 antibodies at V7, with geometric mean titers (GMT) of antibodies to gp120 MN of 3744 (95% confidence interval [CI], 2167–6469) and to gp120 A244 of 596 (95% CI, 301–1180). Of group 2 vaccinees, 100% had anti- MN and 96% had anti-A244 antibodies at V7, with GMTs of antibodies to gp120 MN of 7730 (95% CI, 4961–12,045) and to A244 of 1691 (95% CI, 918–3114). There were no sex-based differences. Responses were not induced in placebo recipients. At V1, anti-p24 antibody was detected in 2 volunteers; at V7, 47% of 92 ALVAC-HIV recipients were positive for anti-p24 antibody, with a GMT of 63.

No neutralizing antibodies were detected in volunteers receiving ALVAC alone or in placebo recipients. Of low- and high-dose AIDSVAX B/E recipients, 100% and 98% developed antibodies to MN/H9, respectively. Regarding the 2 CRF01_AE virus/cell targets, the low-and high-dose recipients had antibody frequencies of 23% and 31% to NP03/H9, respectively, and 44% and 64% to CM244/A3R5, respectively. When responses to either CRF01_AE target were considered, the frequencies in groups 1 and 2 were 47% and 71%, respectively. Similar to those for binding antibodies, these results provide evidence for a dose-dependent humoral response to AIDSVAX B/E. The geometric mean 50% neutralization titers of antibody to MN/ H9 for groups 1 and 2 were 109.4 (95% CI, 74.3–161.3) and 163.3 (95% CI 109.4–243.6), respectively. These values were 12.3 (95% CI, 3.2–46.5) and 14.8 (95% CI, 4.8–45.8) for antibody to NP03/H9, respectively, and 7.0 (95% CI, 3.1–15.5) and 5.4 (95% CI, 2.6–11.3) for antibody to CM244/A3R5, respectively. The titers of binding and neutralization antibody to MN antigen and virus were strongly correlated (r=0.55; P< .001); those to E strains were not.

No placebo recipient had HIV-specific CD8 CTL activity at any time point (173 assays were performed). Results for ALVAC vaccinees are shown in table 2. Positive CTL responses were seen as early as V3 and continued through V10. There were similar point prevalences in CTL activity at V7, V9, and V10 (range, 8%–11%). Three volunteers developed an initial CTL response at V10. The cumulative frequency of positive CTL response in vaccinees was 24%. Repeat positive CTL responses were observed in 9 (41%) of the 22 responders. HIV Gag activity to subtype A was assessed in 8 responders; 3 (38%) showed cross-clade CTL activity. The CD8 CTL response to Gag was independent of p24 antibody response.

Lymphoproliferative responses were measured in 88 volunteers who completed all 4 vaccination visits and in 29 individuals who received placebo at those 4 visits. Phytohemagglutinin responses were positive for all assays; the overall frequency of tetanus toxoid responses was 74%. The frequency of background responses to the envelope proteins before any vaccination was 9%. Overall, responses at V7 to gp120 E occurred in 63% of vaccinees and 7% of placebo recipients; responses at V7 to gp120 MN occurred in 61% and 24%, respectively. This high frequency of background response to gp120 MN complicated interpretation.

By diagnostic serologic testing, at V0 all volunteers were EIA nonreactive. None of the 5 phase 1 volunteers had reactive serologic tests. During phase 2, at V5 no group 1 vaccinee was EIA reactive; at V8, 18 (40%) were EIA reactive and 1 (2.2%) was Western blot (WB) positive; at V10, 2 (4.4%) were EIA reactive and none wereWB positive. In group 2, at V5 1 (2.2%) vaccinee was EIA reactive; at V8, 27 (60%) were EIA reactive and 1 (2.2%) was WB positive; at V10, 4 (8.7%) were EIA reactive and none were WB positive.

Discussion. The combination of ALVAC-HIV (vCP1521) and either dose of AIDSVAX B/E was found to be safe and immunogenic, and there was a dose response between AIDSVAX B/E and antibody response. Induced cellular immune responses were measured at baseline and after immunization (once for helper T cells and 5 times for CTLs). Positive lymphoproliferative responses to gp120 Thai E were detected in 63% of the vaccinees. This high frequency of induction of CD4 helper T cells is similar to that reported for the comparable subtype B ALVAC/gp120 prime-boost combination [9].

The cumulative CTL responses observed in the present trial were similar to, but lower than, those reported for National Institute of Allergy and Infectious Diseases AIDS Vaccine Evaluation Group trials in the United States. A multiple-arm study in the United States, with 150 volunteers using vCP205 and rgp120, assessed CTL activity at 7 time points (compared with 5 in the present study). vCP205 contains the identical gag, pro, and gp41 genes as does vCP1521. The cumulative frequency of CTL responses attributable to vaccination was 50% (95% CI, 16%–74%) [10]. In contrast, a vCP205 trial in Uganda reported a 20% cumulative frequency of CTL responses in vaccinees and 10% in placebo recipients. The assay in the present study displayed higher specificity, compared with other ALVAC-HIV trials [9–14]. In the present trial, new CTL responses continued to be detected at the last visit; the longevity of response is unknown.

Vaccine-induced positive serologic test results, which have the potential to cause psychological and social problems, was found in 60% of vaccinees receiving the 600-µg boost near the time of expected peak antibody response. In contrast to this EIA response, WB positivity was detected in only 2 of 90 vaccinees. By 6 months after immunization, EIA reactivity had decreased to 4%–9%, and no WBs were positive. On the basis of this experience, preparations for the phase 3 trial have included ensuring that both the wider medical community and trial volunteers are well informed about possible false positivity. Inclusion of nucleicacid testing in the present trial insured that infections and vaccine- induced serologic responses were distinguished, and it has been included in the phase 3 trial.

The results of the present study demonstrated that most of the vaccinees developed neutralizing antibodies to Thai E HIV and/or CD8 CTL responses to ALVAC-expressed HIV antigens. Given the safety and immunogenicity profiles reported here, experience with other ALVAC-HIV and AIDSVAX constructs, the match of the vaccine subtypes with locally circulating viruses, and the urgent need for a vaccine to aid in the control of the Thai HIV-1 epidemic, this vaccine combination was advanced to phase 3 testing in Thailand.

Thai AIDS Vaccine Evaluation Group (TAVEG) trial members. TAVEG members from the Armed Forces Research Institute of Medical Sciences (Bangkok, Thailand) include N. Sirisopana, S. Sukwit, S. Tabprasit, A. Kleebmontha, V. Kamonsin, P. Panjapornsuk, S. Akapirat, W. Kaneechit, C. Chuenchitra, P. Chanbancherd, W. Lokpicaht, R. Paris, B. Merrell, J-L. Excler, S. Wongkamhaeng, A. Triampon, P. Buapunth, S. Chinaworapong, R. Trichavaroj, S. Chantakulkij, N. Khaochalod, S. Mason, P. Srisaengchai, S. Chanthong, and Y. Poangngern; members from the Vaccine Trial Centre (Bangkok, Thailand) include W. Supamaranond, S. Naksrisuk, W. Peonim, N. Thantamnu, and R. Muanoum.

• Received December 2, 2003.
• Accepted February 10, 2004.

• © 2004 by the Infectious Diseases Society of America