At the Keystone Symposia on HIV Prevention that took place last week, Dan Barouch presented some new and disturbing data from a study exploring DNA and adenovirus vector immunizations in the macaque SIV challenge model. The vaccine regimen used in the study is similar to - but not exactly the same as - the Vaccine Research Center’s DNA/Ad5 combination, which is to be evaluated for efficacy in the next proposed human HIV vaccine trial, HVTN 505. Barouch’s results have provoked controversy regarding whether HVTN 505 should proceed, not only due to the findings themselves but because they have highlighted the lack of evidence that the VRC’s DNA/Ad5 regimen is superior to Ad5 alone. Ad5 alone, of course, failed to show any efficacy in the recent STEP study.
Barouch’s study involved 30 rhesus macaques, split into five different groups. Animals with MHC genes known to be associated with enhanced control of SIV replication (Mamu A*01, B*08 and B*17) were excluded. The five groups in the study were:
1) DNA prime including Gag, Pol, Nef & Env. Booster immunization with an adenovirus vector called Ad5HVR48 encoding the same antigens. This vector is almost entirely the Ad5 serotype, but with the hexon protein swapped with the same protein from the less common Ad48 serotype.
2) DNA prime including Gag, Pol, Nef & Env, given with DNA adjuvants encoding the chemokine MIP-1 alpha and flt-3 ligand. Booster immunization with Ad5HVR48 encoding the same antigens.
3) Ad5HVR48 encoding Gag, Pol, Nef & Env.
4) Ad5HVR48 encoding Gag, Pol & Nef.
5) Sham (placebo) vaccination.
Six months after the last immunization, all animals were given a high dose intravenous challenge with SIVmac251 (the same virus from which the vaccine antigens were derived, which is called a homologous challenge).
Prior to the challenge, Barouch noted that vaccine-induced T cell responses were five-fold higher in the macaques that received the DNA prime versus those that received Ad5HVR48 alone. However, post-challenge outcomes did not mirror the immunogenicity data. Set point viral loads were lowest in the animals that received Ad5HVR48, averaging 4.4 logs, compared to 5.8 logs in recipients of the DNA prime plus adjuvants (group 2), 5.2 logs in the DNA prime arm (group 1) and 5.5 logs in the placebo arm. In an exploratory analysis comparing the Ad5HVR48 and DNA/ Ad5HVR48 groups, viral loads were significantly higher (by around 0.75 logs) in the latter. In terms of survival, after 500 days of follow up, only 4/12 animals in the DNA groups were alive, compared to 10/12 in the Ad5HVR48 groups. Among the placebo recipients, 1/6 animals survived, comparable with the DNA groups.
Barouch speculated that the results might be explained by the preponderance of Env-specific CD4 T cell responses induced by the DNA prime; animals in these groups showed 4-5 fold higher responses to Gag, Pol & Nef compared to the Ad5HVR48 recipients, but 10-fold higher responses to Env. Barouch cited the possibility that Env-specific CD4 responses might provide additional targets for SIV and thereby enhance viral replication (as was seen in a prior monkey vaccine study that induced only Env-specific CD4 T cell responses using a VZV vector). However, Dr. Barouch took pains to emphasize that the negative impact of DNA priming in his study was only statistically significant in the exploratory combined analysis, and thus the results must be viewed as hypothesis-generating rather than conclusive.
It’s also important to stress that there are differences between the vaccines used in this study and the VRC constructs. The VRC adenovirus vector is entirely Ad5, while Barouch’s has a substituted hexon protein. The VRC DNA vaccine being used in humans consists of six different plasmids, one for each antigen (Gag, Pol, Nef and Envs from clades A, B & C), and splitting the vaccine in this way has been shown to reduce the bias toward Env-specific CD4 T cell responses by enhancing Gag-specific and Nef-specific CD4 T cell responses. However, in terms of the published and presented data on the performance of VRC constructs in the macaque challenge model, it is perhaps surprising to note that there is essentially no evidence supporting the superiority of the DNA/Ad5 approach compared to Ad5 alone. In the most-cited macaque study, published in Science in 2006, the authors write that:
“The vaccination regimens elicited high-frequency SIV-specific cellular immune responses, with greater responses detected in the monkeys that received the DNA prime/rAd boost than those that received only rAd immunizations (Fig. 1). After virus challenge, there were no statistically significant differences in any parameters of clinical outcome between the various groups of experimentally vaccinated monkeys. Therefore, all 24 experimentally vaccinated monkeys were subsequently treated as a single group and compared to the 6 sham-vaccinated control monkeys."
The macaques in this experiment all received the same high-dose intravenous SIVmac251 challenge that was used in Dan Barouch’s study. The vaccine groups in (six animals in each) were as follows:
1) DNA-SIVmac239 gag/pol/nef prime, rAd5-SIVmac239 gag/pol boost
2) DNA-SIVmac239 env prime, rAd5-SIVmac239 env boost
3) DNA-SIVmac239 gag/pol/nef + DNA-SIVmac239 env prime, rAd-SIVmac239 gag/pol + rAd-SIVmac239 env boost
4) rAd5-SIVmac239 gag/pol + rAd5-SIVmac239 env only
5) sham DNA + empty Ad5
One other challenge study presented at the Keystone meeting by Diane Bolton also included a comparison between the VRC’s DNA/Ad5 and Ad5 alone; while both performed significantly better than placebo in terms of viral load reductions, no significant difference between the vaccine strategies could be discerned.
The implications for the HVTN 505 trial are troubling, to say the least. It appears that there is little – if any – evidence suggesting that the VRC’s DNA/Ad5 approach has a better chance of success than Merck’s Ad5 vaccine. In addition to the safety issues with Ad5 raised by the Merck vaccine results, there is also cause for concern that Env-specific CD4 responses induced by the DNA priming could lead to increased target cells for HIV infection. TAG has previously opposed proceeding with a larger antecedent of HVTN 505 called PAVE100A, and takes the position that HVTN 505 should not go ahead.
Keystone Symposia: Prevention of HIV/AIDS (X3), Keystone, Colorado, March 22 - 27, 2009. Abstract #017.
Novel Adenovirus Vector-Based Vaccines for HIV-1
Dan H. Barouch
Beth Israel Deaconess Medical Center, Boston, MA, USA
We have recently shown that heterologous rAd prime-boost regimens utilizing two serologically distinct rAd vectors can elicit potent and polyfunctional cellular immune responses and can afford superior protective efficacy as compared with a homologous rAd5 regimen against SIVmac251 challenges in rhesus monkeys. We will present an update regarding our preclinical and clinical studies involving rare serotype rAd vectors. We will also describe our efforts to develop novel HIV-1 immunogens aimed at optimizing cellular immune breadth. Finally, we will address important safety considerations regarding the use of rare serotype rAd vectors in humans.
Keystone Symposia: HIV Immunobiology: From Infection to Immune Control (X4), Keystone, Colorado, March 22 - 27, 2009. Abstract #125.
Aerosol Adenovirus Immunization Controls Early Viremia
Diane L. Bolton(1), Kamei Song(1), Pam Kozlowski(2), Srinivas Rao(3), and Mario Roederer(1).
1) Immunotechnology Section, Vaccine Research Center (VRC), NIAID, NIH, Bethesda, MD 20892; 2) Gene Therapy Program, LSU Health Sciences Center, New Orleans, LA 70012; 3) Laboratory Animal Medicine, VRC.
Objective: Immunization in the lung may prevent diseases caused by airborne pathogens as well as confer immunity at distal mucosal sites, where HIV and most sexually transmitted pathogens initially replicate. We induced strong immune responses in the respiratory tract of rhesus macaques (RM) using fine aerosol droplets of recombinant adenovirus (serotype 5, rAd5) generated by the PARI eFlow nebulizer. Indian-origin RM were immunized twice either intramuscularly (i.m.) or by aerosol with rAd5 encoding SIV gag-pol and env (5x1010 pfu each). Two additional groups were first primed with three doses of gag-pol-nef and env DNA (4mg each), while a sham group received only the empty aerosolized empty rAd5. SIVmac251 challenge was intravenous.
Results: The aerosol route induced much stronger cellular immune responses in the lung (represented by the bronchoalveolar lavage, BAL) than i.m. (p<0.05), with up to 20% of CD4 and 45% of CD8 T lymphocytes specific for the SIV immunogens. By contrast, PBMC T cell responses were greater following i.m. immunization and even undetectable in the aerosol group at the time of challenge (9 weeks post-rAd). Equivalent systemic humoral IgG and IgA responses were elicited to SIV Env, despite much lower anti-vector antibody responses generated by aerosol. Aerosol rAd5 induced mucosal IgA in the BAL and nasal washes at titers greater than in serum, suggesting local (secretory) IgA production. It also primed for an anamnestic CD8 response upon SIV challenge, albeit weaker than systemic vaccination. Strikingly, all immunization groups significantly controlled acute viremia by 1-2 logs; control was maintained for ~5 months. Lung CD4 T cells were also preserved in immunized animals compared to controls.
Conclusions: Aerosolized adenovirus vaccines generate strong and polyfunctional lung cellular immune responses and persistent systemic and mucosal humoral responses. In the RM SIV model, aerosol immunization affords the same protection against viremia and pathogenicity as i.m. delivery despite the lack of peripheral T cell responses and neutralizing antibodies at the time of challenge. This modality represents a powerful immunization vehicle for respiratory pathogens such as influenza respiratory syncytial virus and tuberculosis, and it may complement vaccine strategies against HIV and other pathogens by inducing excellent mucosal humoral immunity.