Update on Merck's AIDS Vaccine Program

By Patricia Kahn

Since launching its first HIV vaccine clinical trials in 1999, Merck has emerged as a major player in the field, with over 600 people now enrolled in its preventive vaccine studies. So far the company has focused on two candidates—one based on naked DNA, the other on a vector made by modifying adenovirus-5 (Ad5), a common virus that causes colds in humans. Both vaccines exclusively target the cellular immune system.

In a data-packed presentation, Emilio Emini, who heads Merck’s vaccine research, reviewed results from ongoing preventive trials and described work in some new areas—including improved vectors, multi-gene vaccines and novel prime-boost combinations. Primate studies with Ad5-based vaccines will be covered in the next IAVI Report.

As the company evaluates data from its trials and weighs candidates and strategies for the next wave, several questions are at the forefront. One is whether its DNA vaccine is performing well enough to be kept on the A-list of candidates. Another is how to best overcome pre-existing immunity to Ad5 (seen in about 70% of people in most populations, due to the widespread distribution of wildtype virus) and to develop more immunogenic vaccine regimens. Some approaches being studied: Using either higher doses of Ad5 vaccine or different adenovirus serotypes, combining Ad5 with a DNA prime or pairing it with a different type of vaccine.

Ongoing Trials

Clinical studies so far have focused on testing the DNA and Ad5 vaccines individually and in a prime-boost combination. Both of these initial constructs were developed as “proof-of-concept” candidates carrying the HIV-gag gene, with additional genes to be added at a later stage.

Emini emphasized that the trials are still ongoing, so any conclusions remain provisional. The data shown were mostly evaluations of immune responses to Gag using Elispot assays for interferon-gamma producing cells, with results expressed as the number of responding cells (spot-forming cells, or SFC) per million white blood cells.

DNA-gag: On its own, DNA-gag has shown low immunogenicity in volunteers vaccinated four times (at weeks 0, 4, 8 and 26). Both 1mg and 5mg doses elicited weak responses in 21/130 (16%) volunteers at 12 weeks and in 36/117 (31%) at the 30-week timepoint (with geometric means of 64-140 SFC and a range of 36-431 across all groups). Neither alum nor the CRL-1005 adjuvant improved responses substantially over those seen with DNA in saline.

Ad5-gag and pre-existing immunity: The Ad5-gag constructs are proving to be more immunogenic, with about 60% of all volunteers responding at both 8 and 30 weeks. But the data also show a clear blunting of responses in people with pre-existing immunity (PEI) to Ad5.

These preliminary conclusions come from an ongoing study in four groups of 18-29 volunteers, with each group receiving an escalating dose of Ad5-gag (ranging from 108 to 1011 viral particles) at weeks 0, 4 and 26. In presenting the results, Emini subdivided the data according to volunteers’ level of PEI: High (defined as neutralizing antibody titers over 200); mid-range (18-200) and none, with roughly equal numbers of participants in each group.

Combining the low- and mid-range PEI groups, at week 30 there were 24/35 (69%) responders across all dosage groups; surprisingly, the lowest dose worked about as well as the highest. The mean number of SFC ranged from 224 to 412, while individual responses varied from 25 to 1381.

The group with high PEI responded less well, showing only 5/18 (28%) responders and in most cases requiring a higher vaccine dose (4 of the 5 responders were in the two highest dosage groups). Stated another way, while high PEI blunted responses at the lower doses, it appears that higher doses can at least partially overcome this immunity.

DNA-gag and Ad5-gag prime-boost: Emini also reported on immunogenicity of the DNA/Ad5 combination, comparing it with an Ad5-only regimen. Besides its possible effect on improving overall immune responses, DNA priming is seen as a potential strategy for reducing the Ad5 dose (and/or number of doses) and thereby helping to overcome PEI. The study was done by priming twice (in the Ad5 group) or three times (for DNA), then boosting at week 26 with a low dose of Ad5 (107 particles). Disappointingly, DNA does not look substantially better than Ad5 as a prime for an Ad5 boost, based on 30-week data. The high PEI group had 7/20 (35%) responders to DNA/Ad5, compared with 5/18 (28%) for Ad5/Ad5. Mean SFC counts were 187 versus 125, with overlapping ranges of 10-485. The trend was similar for the low- and mid-range PEI volunteers, although DNA/Ad5 induced higher responses in some individuals (reaching 2,300-2,800 SFC, compared with 1,400 for Ad5/Ad5). Emini says that final decisions about the fate of the DNA vaccines will be made once the complete data are in.

But these data—with their implication that DNA may not add significant value to an Ad5 vaccine—have led Merck to change plans for its upcoming international Phase I trial. Instead of testing DNA/Ad5, as originally anticipated, the 435-person study will use Ad5 alone. The study will be done through the US HIV Vaccine Trials Network (HVTN) at sites in the US, Brazil, Thailand, Malawi, South Africa, Haiti and Peru, and should start within the next few months.

New Combinations, New Vector Strains

In the meantime, Merck is evaluating other possible candidates to combine with Ad5.

Boosting with canarypox: Just days before the Banff meeting, Merck announced a partnership with Aventis Pasteur to test that company’s canarypox-based HIV vaccine (ALVAC vCP205). In his talk Emini summarized the monkey data behind this decision, with more complete results shown in a poster by Danilo Casimiro and colleagues. Another factor favoring ALVAC is its well-established safety record, with more than 2,000 volunteers immunized in clinical trials over the past decade.

The crucial study looked at animals primed with a low dose of Ad5 (107-109 particles) at weeks 0,4 and 26, followed at 56 weeks by a boost with either ALVAC vCP1606 or MVA carrying HIV-gag. Canarypox boosting gave the highest responses, which ranged from 1,200 to 2,300 SFC at two weeks post-boost and dropped off by about two-thirds at 8 weeks. An MVA boost induced 800-1,354 SFC at week 2, down to 100-300 at week 8. Interestingly, the synergy between Ad5 and canarypox disappeared when the vaccines were used in the reverse order, with canarypox as a prime and Ad5 as a boost. Challenge data are not yet available. But the decision to move an Ad5/ALVAC combination into Phase I trial is firm, with Merck due to start a trial imminently. The study will recruit volunteers who were previously vaccinated with Ad5 and then boost them with ALVAC.

Other adenovirus serotypes: Given the dampening effect of PEI on responses to Ad5-based vaccines, Merck is also investigating less common strains of adenoviruses as potential as vaccine vectors. Emini briefly described early studies with Ad24, 34 and 35—three strains that do not cross-react immunologically with Ad5 (or one another) and show less than 10% seroprevalence in populations that have been studied, which should largely eliminate the problem of PEI. More detailed data were shown in a poster by Andrew Betts and colleagues.

Initially, a key obstacle was that these new strains did not grow in the laboratory cell lines used for producing Ad5. But the Merck researchers were able to engineer the crucial gene from Ad5 into the other strains, enabling all three strains to grow efficiently. Next, the gag gene was added to vectors made from each strain and used to immunize six macaques (1011 viral particles) at 0 and 4 weeks.

All three constructs were somewhat less immunogenic than Ad5 alone, but Ad24 and Ad34 both seem to work well in prime-boost combinations with Ad5, according to Emini. As more data on these and other serotypes is collected, the company will decide whether to move any of them forward into human trials.

More antigens in an improved Ad5 vector: Another trial due to launch soon will add the HIV pol and nef genes into the gag-containing vaccine. These genes have been built into an improved (more genetically stable) version of the Ad5 vector called MRK5Ad5, which an ongoing clinical study shows to be at least as good as the original vector in eliciting antigen-specific responses.