Research ArticleTuberculosis

A Human Type 5 Adenovirus–Based Tuberculosis Vaccine Induces Robust T Cell Responses in Humans Despite Preexisting Anti-Adenovirus Immunity

See allHide authors and affiliations

Science Translational Medicine  02 Oct 2013:
Vol. 5, Issue 205, pp. 205ra134
DOI: 10.1126/scitranslmed.3006843

Abstract

There is an urgent need to develop new tuberculosis (TB) vaccines to safely and effectively boost Bacille Calmette-Guérin (BCG)–triggered T cell immunity in humans. AdHu5Ag85A is a recombinant human type 5 adenovirus (AdHu5)–based TB vaccine with demonstrated efficacy in a number of animal species, yet it remains to be translated to human applications. In this phase 1 study, we evaluated the safety and immunogenicity of AdHu5Ag85A in both BCG-naïve and previously BCG-immunized healthy adults. Intramuscular immunization of AdHu5Ag85A was safe and well tolerated in both trial volunteer groups. Moreover, although AdHu5Ag85A was immunogenic in both trial volunteer groups, it much more potently boosted polyfunctional CD4+ and CD8+ T cell immunity in previously BCG-vaccinated volunteers. Furthermore, despite prevalent preexisting anti-AdHu5 humoral immunity in most of the trial volunteers, we found little evidence that such preexisting anti-AdHu5 immunity significantly dampened the potency of AdHu5Ag85A vaccine. This study supports further clinical investigations of the AdHu5Ag85A vaccine for human applications. It also suggests that the widely perceived negative effect of preexisting anti-AdHu5 immunity may not be universally applied to all AdHu5-based vaccines against different types of human pathogens.

INTRODUCTION

Tuberculosis (TB) remains one of the leading infectious causes of death globally. About 1.4 million people die of TB each year. One-third of the world population is latently infected by Mycobacterium tuberculosis (M.tb), and there are 8 million to 9 million new TB cases every year. HIV+ persons are much more susceptible to TB, and one-third of people with AIDS succumb to TB (15).

The only anti-TB vaccine, Bacille Calmette-Guérin (BCG), has been used globally for more than 50 years. Although BCG is effective in protecting from disseminated childhood TB, it has failed to effectively control pulmonary TB, which is the major burden of the global TB epidemic. Because repeated BCG immunizations are unable to improve protection triggered by the initial BCG priming vaccination, there is an urgent need to develop TB vaccines that can be used to boost protective T cell immunity by BCG vaccination (1, 4, 5).

Among the most promising candidates to boost BCG vaccination is the recombinant human type 5 adenovirus (AdHu5)–vectored vaccine engineered to express an immune dominant M.tb antigen, Ag85A (AdHu5Ag85A) (6). The AdHu5 vector has proven to be the most robust gene transfer vehicle for in vivo T cell activation and is among the most promising vaccine platforms for immunization against infections in humans (713). Indeed, AdHu5Ag85A has been successfully shown to be effective, when used as either a stand-alone or a booster vaccine after BCG priming, in murine, guinea pig, goat, and bovine models of pulmonary TB (6, 1417). However, its potential remains to be translated to human applications. Furthermore, it is unclear whether an AdHu5-based TB vaccine can overcome the widely preexisting anti-AdHu5 humoral immunity to potently activate TB antigen-specific T cells in humans.

Here, we report on the safety and immunogenicity of AdHu5Ag85A vaccine in healthy adult humans in a phase 1 clinical study. Both BCG-naïve and previously BCG-vaccinated human volunteers were enrolled into the study. AdHu5Ag85A was administered intramuscularly to these volunteers both as a stand-alone priming and as a BCG boosting vaccine. We show that AdHu5Ag85A is safe and robustly immunogenic, particularly in BCG-vaccinated humans, and that preexisting anti-AdHu5 antibodies have a minimal negative effect on Ag85A-specific T cell reactivity in healthy adult humans.

RESULTS

Study participants and follow-up

Twenty-six human volunteers were screened for the trial. One dropped out before the baseline visit and was not vaccinated, and one was positive for QuantiFERON-TB Gold In-Tube test (QFT) and thus excluded from the trial. A total of 12 male and 12 female healthy volunteers were enrolled (table S1). The ages of these volunteers ranged from 21 to 49 and 22 to 51 years in the BCG and the previously BCG-vaccinated (BCG+) groups, respectively (table S1). Demographic details were recorded (table S1). Most of the BCG volunteers were born in Canada. One BCG+ volunteer did not complete the week 24 visit.

Safety of AdHu5Ag85A vaccination in healthy human volunteers

The vaccine was found to be safe and well tolerated (Table 1). There were no vaccine-related serious adverse effects. The most common adverse effects were local reactions at the site of injection including pain and redness, which were judged mild (grade 1) and resolved within 24 hours (Table 1). One-third to half of the volunteers experienced grade 1 systemic reactions such as headache, fatigue, or malaise. One subject reported fever (37.7°C on day 5 after vaccination) and mild arthralgia (Table 1). All systemic reactions resolved within 24 hours. There was one serious adverse event (hospitalization for pancreatitis) judged not related to the vaccination. One subject who had a history of food allergy developed grade 2 (moderate) blood eosinophilia after vaccination, which resolved within 14 days. There were six upper respiratory tract infections reported that were judged not related to the vaccination (one was confirmed by polymerase chain reaction to be rhinovirus, and one was probably H1N1 influenza infection).

Table 1 Adverse events after vaccination with AdHu5Ag85A.

There was one serious adverse reaction (hospitalization secondary to pancreatitis) judged not related to vaccination.

View this table:

The QFT was repeated on all of initially the QFT-negative and subsequently vaccinated volunteers at week 24 visit. All but one remained QFT-negative. The one volunteer that turned QFT-positive had neither clinical evidence of TB infection nor a history of TB exposure.

The purified protein derivative (PPD) skin test was not performed on the enrolled volunteers before vaccination. However, it was performed at week 26 after vaccination on the 10 volunteers that reported a negative PPD test before vaccination. Vaccination with AdHu5Ag85A did not convert PPD test outcome except in one volunteer with a history of BCG vaccination. This conversion was judged to indicate a false-negative result from the previous PPD test.

Markedly increased type 1 cytokine responses in whole-blood culture by AdHu5Ag85A vaccination

To investigate the immunogenicity of AdHu5Ag85A vaccine, we first evaluated the overall antigen-specific reactivity of T cells before and after vaccination in BCG and BCG-vaccinated (BCG+) human volunteers by using a whole-blood culture approach. The whole-blood culture was performed at the baseline (week 0) and at weeks 2, 4, and 24 after vaccination. The whole-blood samples were incubated and stimulated with M.tb culture filtrate proteins (M.tbCF), recombinant Ag85A protein (rAg85A), or a single pool of Ag85A peptides (single Ag85A p. pool). Production of interferon-γ (IFN-γ), tumor necrosis factor–α (TNF-α), and interleukin-2 (IL-2) in whole-blood cultures was quantified by enzyme-linked immunosorbent assay (ELISA). In the BCG group, the baseline production of these cytokines before vaccination was low overall. However, upon AdHu5Ag85A vaccination, the amounts of these cytokines significantly increased and mostly peaked around week 2 (Fig. 1, A to I). Such increases were more obvious in rAg85A- and single Ag85A p. pool–stimulated samples than M.tbCF stimulation. In the BCG+ group, the production of these cytokines markedly increased over baseline and peaked also around week 2 after vaccination (Fig. 1, A to I). Furthermore, compared to the BCG group, the BCG+ group appeared to have a higher baseline of cytokine production (Fig. 1G, BCG versus BCG+, 0 week; P = 0.005, Mann-Whitney test). The BCG+ group also responded to vaccination to a higher extent (Fig. 1A, BCG versus BCG+, 2 weeks, P = 0.05; Fig. 1B, BCG versus BCG+, 2 weeks, P = 0.061; Fig. 1C, BCG versus BCG+, 2 weeks, P = 0.05; Fig. 1G, BCG versus BCG+, 2/4 weeks, P = 0.026 and 0.006, respectively; Fig. 1H, BCG versus BCG+, 2 weeks, P = 0.069) (Mann-Whitney U test). These data indicate that AdHu5Ag85A is immunogenic in both BCG and BCG+ healthy human volunteers and that it can robustly boost T cell immunity triggered by previous BCG immunization.

Fig. 1 AdHu5Ag85A vaccination increases type 1 cytokine responses.

Cytokine responses in whole-blood culture were performed at the baseline (0 week) and at 2, 4, and 24 weeks after vaccination in BCG and BCG+ human volunteers. Blood samples were cultured in the absence or presence of M.tbCF, rAg85A, or single Ag85A p. pool. (A to I) The production of IFN-γ (A to C), TNF-α (D to F), and IL-2 (G to I) was quantified by ELISA. The measurements were subtracted from the unstimulated control values at each time point and are expressed as the median (horizontal line), the interquartile range (box), and the range (whiskers) from n = 12 subjects per time point per group, except BCG+ 24 weeks, where n = 11 because of the unavailability of one subject. The comparison within the vaccination group (BCG or BCG+ groups) was made between 0 week and various time points and analyzed by Wilcoxon signed rank test. Only values of P ≤ 0.05 were displayed.

Markedly increased antigen-specific IFN-γ–secreting T cells by AdHu5Ag85A vaccination

To further investigate the immunogenicity of AdHu5Ag85A vaccine, we examined the antigen-specific reactivity of T cells before and after vaccination in BCG and BCG+ human volunteers by using freshly isolated peripheral blood mononuclear cells (PBMCs) and an IFN-γ enzyme-linked immunospot (ELISPOT) assay. The PBMCs were isolated at the baseline (week 0) and at weeks 2, 4, 8, and 24 after vaccination and subjected to ELISPOT upon stimulation with M.tbCF proteins, rAg85A, or six individual Ag85A peptide pools.

In the BCG group, at the baseline, there was a minimal amount of ELISPOT T cell responses (Fig. 2, A to C). Compared to the baseline week 0 data, most BCG volunteers (10 of 12) responded to vaccination and demonstrated significantly increased T cell responses, which peaked at either week 2 or 4 (Fig. 2, A and B). Raised T cell responses declined after week 4. On the other hand, all BCG+ volunteers (12 of 12) responded to vaccination, demonstrating significantly boosted T cell responses over the baseline control; these responses peaked mostly at week 2 and gradually declined thereafter (Fig. 2, A to C). Furthermore, compared to the BCG group, the baseline ELISPOT T cell responses in the BCG+ group, particularly upon M.tbCF stimulation, were higher (Fig. 2A, BCG versus BCG+, 0 week; P = 0.031, Mann-Whitney test). In accord with the pattern of whole-blood cytokine responses (Fig. 1, A to I), the overall ELISPOT T cell reactivity in BCG+ volunteers was significantly higher than that in BCG counterparts (Fig. 2A, BCG versus BCG+, 2/8/24 weeks, P = 0.028, 0.021, and 0.007, respectively; Fig. 2B, BCG versus BCG+, 2/8 weeks, P = 0.021 and 0.014, respectively; Fig. 2C, BCG versus BCG+, 8 weeks, P = 0.056) (Mann-Whitney test). Exceptionally good responders were only seen in the BCG+ group and produced high numbers of IFN-γ+ ELISPOTs stimulated by M.tbCF or rAg85A at peak times, ranging from 750 to 5780 (Fig. 2, A and B).

Fig. 2 AdHu5Ag85A activates antigen-specific IFN-γ–producing T cells.

(A to C) Freshly isolated PBMCs were obtained from BCG and BCG+ human volunteers at the baseline (0 week) and at 2, 4, 8, or 24 weeks after vaccination and cultured in the absence or presence of M.tbCF (A), rAg85A (B), or one of the six pools of Ag85A peptides (C) (data presented as summed ELISPOT responses across all peptide pools). The measurements were subtracted from the unstimulated control values. Horizontal lines represent the median values from n = 12 subjects per time point per group, except BCG+ 24 weeks, where n = 11 because of the unavailability of one subject. The comparison within the vaccination group (BCG or BCG+ groups) was made between 0 week and various time points and analyzed by Wilcoxon signed rank test. Only values of P ≤ 0.05 were displayed. SFCs, spot-forming cells.

On the basis of the summed-up T cell reactivity to six Ag85A peptide pools (Fig. 2C), we separately analyzed the individual T cell reactivity to each of the six Ag85A peptide pools (Fig. 3, A to F). When compared to T cell responses to stimulation by M.tbCF and rAg85A, this approach provided an assessment on whether the human volunteers enrolled into the trial reacted preferentially to certain segments of the Ag85A protein expressed by AdHu5 vaccine carrier. As reflected in Fig. 2C, collectively, there were more BCG+ volunteers that responded well to each of the six peptide pools (Fig. 3, A to F). On the other hand, both BCG and BCG+ volunteers responded better to stimulation by the peptide pool 1, 2, 3, or 4 (Fig. 3, A to D) than by the peptide pool 5 or 6 (Fig. 3, E and F). A small number of BCG+ subjects still had raised T cell reactivity, particularly to pools 2 and 3 (Fig. 3, B and C), at 24 weeks. Although not part of the original study design, we had an opportunity to obtain fresh PBMCs from a high BCG+ vaccine responder at week 44 after vaccination and found persistently raised ELISPOT reactivity to M.tbCF, rAg85A, or Ag85A peptide stimulation (fig. S1).

Fig. 3 AdHu5Ag85A activates antigen-specific IFN-γ–producing T cells.

(A to F) Freshly isolated PBMCs were obtained from BCG and BCG+ human volunteers at the baseline (0 week) and at 2, 4, 8, or 24 weeks after vaccination and cultured in the absence or presence of one of the six individual Ag85A peptide pools (pools 1 to 6). The measurements were subtracted from the unstimulated control values. Horizontal lines represent the median values from n = 12 subjects per time point per group, except BCG+ 24 weeks, where n = 11 because of the unavailability of one subject. The comparison within the vaccination group (BCG or BCG+ groups) was made between the baseline 0 week and various time points and analyzed by Wilcoxon signed rank test. Only values of P ≤ 0.05 were displayed.

The above data together suggest that AdHu5Ag85A vaccination can activate antigen-specific T cells in BCG humans and robustly boost memory T cell responses in humans triggered by previous BCG priming immunization.

Activation of polyfunctional CD4+ and CD8+ T cell subsets by AdHu5Ag85A vaccination

The above T cell analyses provided information on the overall activation of total T cells after AdHu5Ag85A vaccination. To examine the relative amounts of activation of CD4+ and CD8+ T cells by vaccination, we performed intracellular cytokine staining (ICS) for IFN-γ, TNF-α, and IL-2 by using frozen PBMC samples collected at the baseline (week 0) and at weeks 2, 4, 8, and 24 after vaccination. At first, the T cells positive for IFN-γ, TNF-α, or IL-2 were assessed independently upon stimulation with the single Ag85A p. pool or M.tbCF. Compared to the baseline, AdHu5Ag85A vaccination activated both CD4+ (Fig. 4A) and CD8+ (Fig. 5A) T cells. Notably, whereas AdHu5Ag85A primarily activated CD4+ T cells in BCG subjects, it activated both CD4+ and CD8+ T cell subsets in BCG+ subjects (Figs. 4A and 5A). Consistent with the whole-blood cytokine (Fig. 1) and fresh PBMC ELISPOT (Figs. 2 and 3) data, both CD4+ and CD8+ T cell responses peaked at 2 and 4 weeks after vaccination (Figs. 4A and 5A), and the magnitude of CD4+ and CD8+ T cell responses in BCG+ subjects was markedly higher than that in the BCG counterparts (Figs. 4A and 5A and fig. S2). This was particularly true of CD4+ T cell responses to M.tbCF stimulation (P ≤ 0.01 at weeks 2, 4, 8, and 24, Mann-Whitney test) (Fig. 4A) and CD8+ T cell responses to the peptide stimulation (P ≤ 0.02 at week 4 for total cytokine+ and IFN-γ+ CD8+, and at week 24 for IL-2+ CD8+, Mann-Whitney test) (Fig. 5A). The CD8+ T cells in either BCG or BCG+ subjects responded poorly to M.tbCF stimulation (fig. S3). Upon close examination, in BCG+ subjects, the frequencies of CD4+ T cells, particularly those positive for TNF-α or IL-2, at 24 weeks still remained raised above the baseline (Fig. 4A, the peptide stimulation). On the other hand, although IFN-γ+ CD8+ T cells at week 24 in BCG+ subjects declined near 0-week levels, the increased frequencies of TNF-α+ or IL-2+ CD8 T cells at week 24 in these subjects sustained (Fig. 5A).

Fig. 4 AdHu5Ag85A activates polyfunctional CD4 T cells measured by ICS.

(A) Kinetics of combined (total cytokine) or single expression of IFN-γ, TNF-α, and IL-2 by CD4 T cells in response to stimulation by single Ag85A p. pool or M.tbCF in BCG and BCG+ subjects. Data are expressed as the median points and the interquartile ranges (whiskers) of cytokine-expressing CD4 T cell frequencies at various time points with background unstimulated values subtracted. The comparison within the vaccination group (BCG or BCG+ groups) was made between 0 week and various time points and analyzed by Wilcoxon signed rank test. Only values of P ≤ 0.05 were displayed. (B) Polyfunctional CD4 T cells in response to stimulation by single Ag85A p. pool or M.tbCF in BCG and BCG+ subjects. Data are shown as the frequencies of single (1+)–, double (2+)–, and triple (3+)–cytokine–positive CD4 T cells at 0 week and various time points after vaccination, and expressed as the median (horizontal line), the interquartile range (box), and the range (whiskers). *P ≤ 0.05, **P ≤ 0.005, ***P ≤ 0.001, τP ≤ 0.07 compared to 0 week within the same group, Wilcoxon signed rank test. (C) Average proportions displayed in pie chart of the CD4 T cells expressing a specific single cytokine, a specific combination of two cytokines, and a combination of three cytokines in BCG and BCG+ subjects. All data are from n = 12 subjects per time point per group, except BCG+ 24 weeks, where n = 11 because of the unavailability of one subject.

Fig. 5 AdHu5Ag85A activates polyfunctional CD8+ T cells measured by ICS.

(A) The kinetics of combined (total cytokine) or single expression of IFN-γ, TNF-α, and IL-2 by CD8 T cells in response to stimulation by single Ag85A p. pool in BCG and BCG+ subjects. Data are expressed as the median points and the interquartile ranges (whiskers) of cytokine-expressing CD8 T cell frequencies at various time points with background unstimulated values subtracted. The comparison within the vaccination group (BCG or BCG+ groups) was made between 0 week and various time points and analyzed by Wilcoxon signed rank test. Only values of P ≤ 0.05 were displayed. (B) Polyfunctional CD8+ T cells in response to stimulation by single Ag85A p. pool in BCG and BCG+ subjects. Data are shown as the frequencies of single (1+)–, double (2+)–, and triple (3+)–cytokine–positive CD8+ T cells at 0 week and various time points after vaccination, and expressed as the median (horizontal line), the interquartile range (box), and the range (whiskers). *P ≤ 0.05, **P ≤ 0.005, ***P ≤ 0.001, τP ≤ 0.07 compared to 0 week within the same group, Wilcoxon signed rank test. (C) Average proportions displayed in pie chart of the CD8 T cells expressing a specific single cytokine, a specific combination of two cytokines, and a combination of three cytokines at various time points in BCG and BCG+ subjects. All data are from n = 12 subjects per time point per group, except BCG+ 24 weeks, where n = 11 because of the unavailability of one subject.

Because the “polyfunctional” T cells that coexpress IFN-γ, TNF-α, and IL-2 have been associated with immune protection (18), we further assessed the T cells that coexpressed one, two, or three of these cytokines. Compared to the single-cytokine–positive (1+) CD4+ T cells, vaccination significantly increased the double (2+)– and triple (3+)– cytokine–positive CD4+ T cell responses to the peptide pool stimulation at various time points in both BCG and BCG+ subjects (Fig. 4B and fig. S2). A similar pattern of activation was also observed with the CD4+ T cells reactive to M.tbCF stimulation (Fig. 4B). Similarly, vaccination also significantly increased the polyfunctionality of CD8+ T cells at various time points (Fig. 5B and fig. S2).

We further examined the proportions of specific cytokine-expressing polyfunctional CD4+ and CD8+ T cells in BCG and BCG+ subjects. At weeks 2, 4, and 8, BCG+ subjects had greater proportions of the peptide pool–stimulated polyfunctional CD4+ T cells (IFN-γ+TNF-α+IL-2+, IFN-γ+TNF-α+, IFN-γ+IL-2+, or TNF-α+IL-2+), whereas BCG subjects had greater proportions of the peptide pool–stimulated single-cytokine–positive CD4+ T cells (IFN-γ+, TNF-α+, or IL-2+) (Fig. 4C). By week 24, both BCG and BCG+ subjects had comparable proportions of polyfunctional and single-cytokine–expressing CD4+ T cells. Similar to the peptide pool–stimulated CD4+ T cells, BCG+ subjects had greater proportions of IFN-γ+TNF-α+IL-2+, IFN-γ+TNF-α+, IFN-γ+IL-2+, or TNF-α+IL-2+ CD4+ T cells reactive to M.tbCF stimulation at various time points than their BCG counterparts (Fig. 4C). Compared with CD4+ T cells, vaccine-activated, peptide pool–stimulated CD8 T cells demonstrated lack of TNF-α+IL-2+ CD8+ T cells in both BCG and BCG+ subjects (Fig. 5C). Furthermore, in BCG+ subjects, there were progressively increased proportions of IFN-γ+TNF-α+ polyfunctional CD8+ T cells, and by week 24, >50% of cytokine+ CD8+ T cells were IFN-γ+TNF-α+, whereas the single IFN-γ+ cell population markedly shrank (Fig. 5C). BCG+ subjects also maintained a larger IFN-γ+TNF-α+IL-2+ CD8+ T cell population at various time points than their BCG counterparts (Fig. 5C).

The above data suggest that AdHu5Ag85A can potently boost antigen-specific responses of both polyfunctional CD4+ and CD8+ T cells in previously BCG-vaccinated subjects. The polyfunctional cytokine profile differs between CD4+ and CD8+ T cells and between BCG and BCG+ subjects.

Substantial amounts of anti-AdHu5 antibody titers before and after AdHu5Ag85A vaccination

AdHu5 is a common cold virus (7, 9). Abundant evidence suggests that anti-AdHu5 antibodies are varyingly prevalent in humans depending on geographic regions (9). The preexisting anti-AdHu5 humoral immunity may severely dampen the potency of AdHu5-based gene vectors in both humans and animals (9, 13, 19). To investigate whether this was the case in our trial participants, we first examined the circulating levels of AdHu5-specific total immunoglobulin G (IgG) antibodies before and at week 8 after vaccination by ELISA. There was a high concentration of circulating AdHu5-specific total IgG in most of the trial participants (>105 titers) (Fig. 6A), and AdHu5Ag85A vaccination markedly increased these titers by 0.5 to 1 log in both BCG and BCG+ volunteers (Fig. 6A). There was no marked difference in AdHu5-specific total IgG before and after vaccination between BCG and BCG+ volunteers.

Fig. 6 Substantial amounts of anti-AdHu5 antibodies in BCG and BCG+ human volunteers before and after AdHu5Ag85A vaccination.

(A) Total anti-AdHu5 IgG titers in sera at 0 and 8 weeks after vaccination were determined by ELISA. (B) Anti–AdHu5-neutralizing antibody titers in sera at 0 and 8 weeks after vaccination were determined by a biologic AdHu5 neutralization assay. Horizontal lines represent the median values from n = 12 subjects per time point.

Because the titers of anti-AdHu5 IgG may not correlate with their AdHu5-neutralizing capacity (20), we further evaluated the circulating AdHu5-neutralizing antibody titers before and at week 8 after vaccination by using a biologic assay. About 50% of both BCG and BCG+ volunteers had comparably high amounts of preexisting AdHu5-neutralizing antibody titers (>102) (Fig. 6B). Vaccination with AdHu5Ag85A significantly increased AdHu5-neutralizing antibody amounts by an average of 2 logs in both BCG and BCG+ volunteers (Fig. 6B). Because the preexisting anti-AdHu5 antibody titers might potentially affect the potency of AdHu5Ag85A vaccination (9), we further analyzed the relationship between preexisting (week 0) anti-AdHu5 total IgG and preexisting anti–AdHu5-neutralizing antibody titers. As shown in Fig. 7A, the titers of anti-AdHu5 total IgG did not significantly correlate with the levels of AdHu5-neutralizing antibodies.

Fig. 7 Lack of significant correlation between preexisting anti-AdHu5 total IgG and neutralizing antibody titers or between the magnitude of vaccine-induced T cell activation and preexisting AdHu5-neutralizing antibody titers.

(A) Total anti-AdHu5 IgG titers at 0 week were plotted against the corresponding anti-AdHu5 neutralization titers. (B to D) The peak ELISPOT T cell reactivity values (2 or 4 weeks) to M.tbCF (B), rAg85A (C), or Ag85A peptide pools (D) were plotted against the corresponding anti-AdHu5 neutralization titers. The correlation analysis was performed with Pearson correlation coefficient test (n = 24 subjects in all panels).

The above data together indicate that most of the trial participants have medium to high levels of anti-AdHu5 antibody titers before AdHu5Ag85A vaccination.

Minimal effects of preexisting anti-AdHu5 antibodies on the magnitude of AdHu5Ag85A-activated T cell responses

Because all of our trial volunteers had anti-AdHu5 total IgG and neutralizing antibodies before vaccination (Fig. 6, A and B), we addressed the question of whether these antibodies negatively affected the potency of intramuscular AdHu5Ag85A vaccination. To this end, we first specifically examined the relationship by plotting the highest ELISPOT T cell reactivity value from each volunteer (both BCG and BCG+ groups) according to individual antigenic stimulation against the corresponding preexisting AdHu5 neutralization titer. There was no significant correlation between the titer of preexisting AdHu5-neutralizing antibodies and the amount of vaccine-triggered T cell reactivity, regardless of the type of antigenic stimulation by M.tbCF, rAg85A, or Ag85A peptides (Pearson correlation coefficient test for their linear relationship) (Fig. 7, B to D). Because these variables were not normally distributed due to the small sample size, we also examined their monotonic relationship by Spearman rank correlation test. We still found a lack of correlation in Fig. 7A (P = 0.34) and Fig. 7B (P = 0.24), and found only a borderline correlation in Fig. 7C (P = 0.04) and Fig. 7D (P = 0.02). Furthermore, there was no significant correlation found between the amount of preexisting anti-AdHu5 total IgG and the amount of vaccine-triggered T cell reactivity (fig. S4). These data together suggest that despite high prevalence of preexisting antibodies against the vaccine backbone AdHu5 in our trial participants, it has a minimal negative impact on the potency of AdHu5-based TB vaccine.

DISCUSSION

We have evaluated the safety and immunogenicity of a recombinant replication-incompetent AdHu5-based TB vaccine after a single dose of intramuscular injection. AdHu5Ag85A was evaluated both in BCG-naïve humans as a priming vaccine and in BCG+ humans as a boosting vaccine. We found AdHu5Ag85A to be safe and well tolerated in both trial populations. Although it is immunogenic in both BCG-naïve and BCG-vaccinated subjects, AdHu5Ag85A much more robustly boosted polyfunctional CD4+ and CD8+ T cell immunity in previously BCG-vaccinated humans. Furthermore, we found preexisting anti-AdHu5 humoral immunity in almost all volunteers. However, such anti-AdHu5 humoral immunity only had a minimal negative effect on the potency of AdHu5Ag85A vaccine, as measured by T cell reactivity.

It is widely believed that viral vector–based TB vaccines hold great potential to be robust boosters for BCG-vaccinated humans (1, 4, 7). The viral-vectored TB vaccines also have an advantage over mycobacterial- and protein-based candidates for effective respiratory mucosal vaccination (7, 21, 22). Currently, 12 TB vaccine candidates are globally undergoing clinical evaluations (1, 2). Among these candidates, three are viral vector–based vaccines: MVA-Ag85A, AdHu35-Ag85A/B-TB10.4, and our AdHu5Ag85A. Each of the three candidates has its own pros and cons. For instance, although preexisting anti-AdHu35 humoral immunity is much less prevalent in humans than that for AdHu5 (9), the former vector is much less immunogenic than the latter (23). Furthermore, compared to AdHu35-based vaccines, AdHu5-based vaccines appear more effective for respiratory mucosal immunization (6, 24). In this regard, intranasal administration of AdHu5-based HIV and influenza vaccines has been shown to be safe and immunogenic in nonhuman primates and humans (11, 25). Note that a fair comparison between these viral vectors may be difficult due to different infectivity. The current conviction is that multiple vaccine candidates are required for clinical evaluation to identify the best candidates for ultimate use in different human scenarios (1). The importance of exploring multiple vaccine candidates is further highlighted by the findings from a recently published MVAAg85A phase 2b trial (26). Thus, our current report on AdHu5Ag85A vaccine in healthy humans represents a critical step going forward and contributes important knowledge to the current global effort of clinical TB vaccine development.

AdHu5Ag85A has been extensively evaluated in various animal models of pulmonary TB (6, 1417). Although we have shown that intramuscular AdHu5Ag85A vaccination potently activates peripheral T cells but provides poor lung protection in murine models (6, 14), it provides robust protection in larger animal models of TB (1517). These lines of evidence suggest that the relative efficacy by parenteral AdHu5Ag85A-activated T cells differs between different animal species, and the parenteral AdHu5Ag85A vaccination remains a promising modality of boosting immunization for BCG-vaccinated humans.

The dose of AdHu5Ag85A [108 plaque-forming units (PFU) or 3.2 × 109 viral particles] used in our study is considered rather small compared to the doses of up to 1010 and 1011 viral particles of AdHu5-based HIV and malaria vaccines in recent clinical trials (12, 13, 27). However, despite the relatively small dose used, AdHu5Ag85A was shown to be potently immunogenic in our current study. Most trial volunteers had varying levels of preexisting AdHu5-neutralizing antibodies, and it may be expected that small doses of AdHu5Ag85A would be more prone to the blocking effect of neutralizing antibodies.

Indeed, there is a pervasive notion that high prevalence of neutralizing antibodies in humans is a major disadvantage to AdHu5-based vaccines (79), which has led to the use of less prevalent serotypes of human adenovirus such as AdHu35 as a vaccine carrier (9, 28). However, the caveat for using such human adenoviruses is that they are inherently less immunogenic (23). Moreover, recently published evidence suggests that the magnitude of preimmunization neutralizing antibodies does not always correlate with the potency of AdHu5-based vaccines, depending on the nature of transgene-encoded antigen and the dose of vaccination. For instance, preexisting neutralizing antibodies did not affect the immunogenicity of an AdHu5-based malaria vaccine in humans (12), and preexisting neutralizing antibodies did not affect T cell reactivity of AdHu5-based HIV vaccine to the env but it did to the gag antigen (29). Furthermore, although preexisting neutralizing antibodies negatively affected the immunogenicity of small and high doses of AdHu5-based HIV vaccine, it did not affect the medium dose (13).

The AdHu5-based TB vaccine was developed primarily for boosting immunization in BCG-vaccinated humans. Although we have observed that this vaccine is immunogenic in BCG-naïve humans, the much higher magnitude of T cell activation observed in BCG+ humans supports its application for boosting BCG-primed T cell immunity. Furthermore, as anti-AdHu5 antibodies increase with age, the potential of anti-AdHu5 immunity to negatively affect AdHu5-based vaccination would be even less a concern for pediatric populations (20, 30).

A further advantage of using AdHu5-based TB vaccine for BCG-vaccinated humans is that it elicits the balanced activation of both polyfunctional CD4+ and CD8+ T cell subsets in BCG+ subjects, whereas it primarily activates CD4+ T cells in BCG subjects. This represents a major difference between experimental and human studies because AdHu5Ag85A mostly activates CD8 T cells in animal models (6, 14, 31). In this regard, AdHu5Ag85A differs from MVAAg85A vaccine in that the latter predominantly activates CD4+ T cells in BCG+ humans (32, 33). In comparison, AdHu35-based TB vaccine activates both CD4+ and CD8+ T cells, whereas AdHu5-based HIV and malaria vaccines primarily engage CD8+ T cells in humans (12, 13). Increasing evidence suggests that it may be advantageous for a TB vaccine to be able to engage both T cell subsets (34). Of interest, we find the vaccine-activated T cells, particularly CD8+ T cells, positive for IL-2 or TNF-α to be more durable than their IFN-γ+ counterparts. This observation supports the current notion that compared to IFN-γ+ T cells, TB vaccine–activated IL-2+ and/or TNF-α+ cells tend to be long-lived memory T cells (35). Another advantage of using AdHu5-based vaccine for BCG+ humans is that it also activates the multiclonal CD4+ T cells reactive to M.tb antigens other than Ag85A.

There are several limitations to the current study. Because the vaccine was evaluated only in healthy humans, it remains to be investigated whether AdHu5Ag85A will be safe and effective in humans with latent TB or HIV infection. It also remains to be tested in humans in TB-endemic areas. Furthermore, larger trials are required to determine whether indeed preexisting anti-AdHu5 immunity has a minimal negative impact on the potency of AdHu5-based TB vaccine. However, our findings suggest AdHu5Ag85A to be a promising TB vaccine for human applications.

MATERIALS AND METHODS

Study design

This was an open-label phase 1 clinical study conducted at McMaster University Medical Centre. After written, informed consent was obtained, subjects underwent a physical examination and chest x-ray, and blood samples were collected for hematology and biochemistry, HIV antibody test, and IFN-γ release assay for TB (QFT). The healthy male and female subjects of 18 to 55 years of age (table S1) determined by history, and physical and laboratory examinations, with or without a history of BCG vaccination and with negative HIV antibody and QFT results, were enrolled into the study. Enrolled women of childbearing potential had a negative pregnancy test and were willing to practice two forms of contraception, and enrolled men were willing to use barrier contraception for the duration of the study. Participants were excluded if they had a history of active or latent TB, had a known exposure to TB with the last 6 months, were at increased occupational risks of TB exposure, had any abnormality on chest x-ray suggestive of TB, or had received PPD skin test within 12 months. This clinical study was approved by Health Canada and the institutional Research Ethics Board and was registered on ClinicalTrials.gov (NCT00800670).

At the baseline visit, blood was drawn for immune analysis, and AdHu5Ag85A (108 PFU or 3.2 × 109 viral particles) diluted in 1 ml of sterile water was injected intramuscularly to the deltoid muscle of the right arm (BCG volunteers) and to the arm contralateral to the BCG-injected arm (BCG+ volunteers). The subjects were monitored medically for local and systemic adverse effects at 48 hours and at 1, 2, 4, 8, 16, 24, and 26 weeks after vaccination. At weeks 2, 4, 8, and 24, blood was also drawn for immune analysis. Local and systemic symptoms and temperature were recorded daily for 5 days after vaccination. Toxicity was assessed with National Cancer Institute Expanded Common Toxicity Criteria Version 3 (Table 1). At week 24, QFT test was repeated on all subjects. At week 26, PPD skin test was performed only on those who had a documented negative PPD test result.

Vaccine and manufacturing

The construction and preclinical evaluation of the replication-incompetent AdHu5Ag85A vaccine have previously been described (6, 1417, 31). Clinical-grade AdHu5Ag85A was produced under Good Manufacturing Practice conditions in the Robert E. Fitzhenry Vector Laboratory of McMaster Immunology Research Centre. The master cell bank, the master virus bank, and the final clinical vaccine lot were tested for sterility, Mycoplasma, adventitious viruses, safety, and/or toxicity by WuXi AppTec.

QFT assay

All subjects were subjected to QFT assay (Cellestis Ltd.) for potential latent TB infection or TB history at the time of screening and for those who were enrolled, at week 24 after immunization, by following the manufacturer’s instructions.

Whole-blood culture and cytokine assay

Immediately after collection, 1 ml of heparinized whole blood was added into each well of a 24-well plate (36). Each was stimulated for 18 to 24 hours with one of the following antigens: 20 μg of M.tbCF, 10 μg of rAg85A, a single pool of 57 Ag85A peptides (2 μg of each peptide), and 50 μg of phytohemagglutinin. An unstimulated well was set up as a negative control. Collected plasma was stored at −70°C. Cytokines were determined with DIF50, D2050, and DTA00C Quantikine ELISA kits (R&D) for human IFN-γ, IL-2, and TNF-α, respectively.

Peripheral blood mononuclear cells

PBMCs were isolated from ~30 ml of heparinized blood. Blood was mixed with equal volume of phosphate-buffered saline, and PBMCs were collected with Ficoll-Paque gradient solution. PBMCs were washed and resuspended in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and 1% l-glutamine. Fresh PBMCs were used for IFN-γ ELISPOT assay. The rest of PBMCs were cryopreserved.

Fresh PBMC IFN-γ ELISPOT assay

ELISPOT was performed on fresh PBMCs with human IFN-γ ELISPOT Set (BD Biosciences) according to the instructions. PBMCs (0.1 × 106 per well) were plated in the presence of antigen in duplicate and incubated for 16 to 24 hours. Antigens used for stimulation included M.tbCF (2 μg), rAg85A (1 μg), and each of six pools of 7 to 10 overlapping Ag85A peptides (10 μg of each peptide). A complete peptide library used for stimulation consisted of the 15-mer peptides with a 10-mer overlap spanning the Ag85A protein expressed by AdHu5Ag85A. Upon incubation, wells were washed and processed for ELISPOT detection. IFN-γ+ cells were enumerated with CTL-Immunospot microanalyzer (Cellular Technology Ltd.). Average number of IFN-γ+ cells, if any, in the unstimulated wells was subtracted from that of stimulated wells, and the resulted spot counts were used for analyses. The vaccine responder was defined as the one that had a two times baseline week 0 level of ELISPOT T cell reactivity or had ≥10 ELISPOT when the baseline week 0 ELISPOT was 0, in response to stimulation by rAg85A or Ag85A peptide pools (at least responded to three of six peptide pools) at 2 or 4 weeks after vaccination. Stimulation with the peptide pools might have potentially resulted in a T cell reactive to the 10-mer overlapping regions across the two adjacent pools, being counted twice.

Peripheral blood mononuclear cell intracellular cytokine staining

Frozen PBMCs were thawed and rested overnight before stimulation in complete RPMI containing 10% FBS, 2 mM l-glutamine, penicillin G (100 U/ml), and streptomycin sulfate (100 μg/ml) (36). Multicytokine responses of PBMCs were assessed by ICS upon stimulation with single Ag85A p. pool (2 μg/ml of each peptide) or with a combination of M.tbCF (10 μg/ml) and rAg85A (10 μg/ml) (M.tbCF). During the 6-hour stimulation of single Ag85A p. pool, brefeldin A (10 μg/ml) and the costimulatory antibodies CD28 and CD49d (each at 1 μg/ml) were included. PBMCs were stimulated with M.tbCF for 12 hours in the presence of anti-CD28 and CD49d with brefeldin A added for the last 6 hours. PBMCs were then processed and stained with fluorochrome-conjugated monoclonal antibodies CD14 (V450), CD19 (V450), and CD4 (AF700). Before ICS, PBMCs were permeabilized and stained for CD3 (peridinin chlorophyll protein–Cy5.5), CD8 (phycoerythrin-Cy7), IFN-γ (phycoerythrin), TNF-α (fluorescein isothiocyanate), and IL-2 (allophycocyanin). All antibodies were from BD Biosciences. Cells were analyzed with LSRII flow cytometer and assessed with FlowJo version 9.3.2 (Tree Star Inc.).

ELISA for anti-AdHu5 total IgG antibodies

Circulating AdHu5-specific total IgG was determined by ELISA. Nunc immunoplates were coated with 50 μl of wild-type AdHu5 virus–infected HeLa cell lysate (100 μg/ml) overnight at 4°C. Serial dilutions of sera were set up in duplicates and incubated at room temperature for 2 hours. After wash, 50 μl of a 1:100 dilution of biotin-conjugated mouse anti-human IgG (Sigma) was added to each well and incubated for 2 hours at room temperature. Wells were washed and treated with streptavidin–horseradish peroxidase conjugate and subsequently with the substrate buffer. Reaction was stopped by 2 M H2SO4, and absorption was read at 450 nm.

AdHu5-neutralizing antibody assay

Sera were tested on confluent A549 cells for the presence of AdHu5-neutralizing antibodies. At the time of assay, 180 μl of culture medium with AdHu5 vector expressing β-galactosidase (500 PFU/ml) was mixed with an equal volume of serially diluted serum and incubated on a shaker for 1 hour at 37°C. The medium was removed from A549 monolayers, and 300 μl of serum-virus mixture was added and incubated overnight at 37°C. The cells were fixed with 0.5% glutaraldehyde, treated with 300 μl of X-galactosidase staining solution, and incubated for 3 hours at 37°C. Upon drying, digital images of wells were taken under a microscope, and stained cells were enumerated with AxioVision4 software. AdHu5-neutralizing antibody titers were calculated as (number of stained cells × 4 × virus dilution)/volume in well.

Data analyses

Adverse events were reported descriptively. Wilcoxon matched pairs signed rank test was used to compare the change of T cell or cytokine responses at various time points from the baseline values within the same vaccination groups (BCG or BCG+). Mann-Whitney U test was used for comparison of the difference in T cell or cytokine responses between BCG and BCG+ vaccination groups. For correlation analysis, Pearson correlation or Spearman rank coefficient test was used. A test significant level of 5% was used in the study. Data analyses were performed with Prism 5 version 5.0c software.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/5/205/205ra134/DC1

Fig. S1. Persistently raised T cell reactivity to AdHu5Ag85A vaccination.

Fig. S2. Activation of polyfunctional CD4 and CD8 T cell subsets in BCG and BCG+ human subjects by AdHu5Ag85A vaccination.

Fig. S3. Lack of CD8 T cell responsiveness to M.tbCF stimulation in vaccinated BCG and BCG+ subjects.

Fig. S4. Lack of significant correlation between the magnitude of T cell activation after AdHu5Ag85A vaccination and preexisting anti-AdHu5 total IgG antibody titers.

Table S1. Demographic details of enrolled trial volunteers.

REFERENCES AND NOTES

  1. Acknowledgments: We acknowledge the clinical assistance from S. Culgin, N. Angeline, M. Stirling, and E. Scheid, and technical assistance from U. Sankar, X. Feng, and J. Millar. Funding: This study is supported by the Canadian Institutes for Health Research, McMaster University, and McMaster Michael G. DeGroote Institute for Infectious Disease Research. Author contributions: F.S., M.S., M.F.M., R.F., J.B., J.G., and Z.X. designed the study and protocols; M.J., N.T.-D., A.Z., C.Y., A.H., D.D., L.P., J.H., and F.X. performed the assays and analyses; F.S. and Z.X. were the principal investigators for this study and wrote the manuscript. Competing interests: The commercial right of AdHu5Ag85A vaccine is licensed to Tianjin CanSino Biotechnology Inc. under a license agreement between McMaster and CanSino (L/O12-015). The authors have no other competing interests.
View Abstract

Navigate This Article