Microbicides 2008: First hint of efficacy in rectal microbicide trial, thanks to new biopsy assay

Preliminary, still blinded data presented to the 2008 Microbicides Conference in Delhi today offered a strong hint that a rectal microbicide gel containing a non-nucleoside HIV drug may prove to be effective at stopping infection in humans.

The same preliminary analysis also seems to indicate that the drug may be safe, at least if volunteer reports, and the results of a battery of tests of inflammation marker chemicals, are reliable indicators of likely harm.

The trial

Peter Anton of the ‘U19’ rectal microbicide programme was presenting data from the first 19 of 28 subjects enrolled into a phase I safety study of two doses of a microbicide gel containing the NNRTI drug UC-781. This programme, funded by the US National Institutes of Health, is the first series of studies to trial rectal microbicides in human subjects. It involves a series of double-blinded trials, currently comparing two doses of UC-781 (1% and 2.5%) with placebo; as the programme progresses this may be narrowed down to one dose.

This initial study is primarily designed as an acceptability study. However the use of some innovative inflammation markers and assays allows early educated guesses to be made about toxicity and even eventual efficacy.

In this study 36 male and female volunteers were screened at the beginning of the trial for exclusion criteria such as HIV and STI infection. Those enrolled inserted a measured dose of the UC-781 gel with a rectal applicator at two separate timepoints. At baseline and both times after gel application, biopsies – small sections of rectal tissue – were taken for the safety and drug efficacy assays – of which more below. At the same time samples of rectal fluid/mucus were taken for evaluation of soluble proteins.

On the second week of the study volunteers applied a single dose of UC-781 gel or placebo under supervised conditions, and biopsies were taken immediately afterwards (within 30 minutes).

After a rest period, volunteers were then issued with a week’s worth of gel doses, again in applicators, which they used every day at home on the sixth week of the study. At the end of this week they returned to the research centre for more biopsies to be taken. Anton emphasised that each biopsy, of which typically 15 were taken from each subject at distances of 10cm and 30cm into the rectum, were small and painless procedures.

Adverse events and toxicity

Anton was presenting data from the 19 out of 28 subjects who have completed all visits. Importantly, all the data he was presenting was blinded: so we cannot be sure whether results are due to the microbicide, the placebo, the participants, or chance. However they do describe general changes within the trial population over time.

In terms of safety there were no serious adverse events (grade 3 or 4). There were seven minor events reported from four participants, but four of these (diarrhoea, bloating and abdominal discomfort) were reported by a single participant at the single-dose visit – and were probably due to his previous meal.

There were also preliminary, blinded safety data available in terms of laboratory markers of toxicity. The U-19 programme will use a whole battery of tests to try and find any indication of cellular toxicity or inflammation in gel recipients. Conscious that the rectum has been largely ‘unknown territory’ in terms of what constitutes a ‘normal’ immune picture, this had to involve a lot of preliminary studies to assess benchmarks – see this report.

The early data also showed no difference between the trial groups in terms of macroscopic or microscopic cell damage, nor any signal of an inflammatory response in terms of the levels of a large number of different cytokines and chemokines (soluble inflammatory proteins produced in response to injury).

A hint of efficacy, and a way of estimating it

The really innovative aspect of this trial is the way the biopsies are being used as ‘surrogate markers’ for HIV seroconversion.

Microbicide studies have suffered from the huge disadvantage that up till now there has been no early way of estimating their likely protective effect in humans. Animal models can give unreliable results and there have been no ‘correlates of protection’ such as the immunogenicity markers that have been used in vaccine trials to select promising candidates (though those are currently being called into question). The only way up until now has been to put on a huge and expensive efficacy trial on the basis of results in a handful of monkeys, or even in vitro data, and hope the candidate proves to be effective.

The U19 Programme is getting round this by using an ‘in vivo-ex vivo’ HIV infection model. In this, volunteers use the microbicide or placebo gel as instructed, and biopsies are taken. These are then cultivated as cellular explants – small pieces of tissue kept alive on gel rafts in a nutrient medium. Two hours after being set up, the explants are then incubated with HIV. The proportion that get infected with HIV is then determined by measuring the amount of the HIV p24 protein in the culture medium over the next month (which typically grows exponentially in cases of infection).

In this study, explants were infected with two different doses of HIV, one containing 100 times more virus than the other. The stronger of the two doses should in theory infect 100% of biopsies containing cells expressing the CCR5 HIV co-receptor.

Ninety-six per cent of volunteers’ biopsies taken at baseline, before any gel application, were infected with HIV (i.e. 18 out of 19 subjects – one person’s biopsies stubbornly resisted HIV infection throughout, despite his having CCR5-expressing cells).

At the three-week visit, one-third of the explants grown from biopsies taken immediately after gel application (six subjects) resisted infection with the higher dose of virus, while another third showed reduced p24 expression. The final third were infected just as easily.

This pattern was not observed at week seven. This is probably, Anton commented, because biopsies were taken the morning after the last day of gel use, in other words anything between ten and 24 hours after the last possible exposure to UC-781.

“It would be convenient to anticipate from this blinded data,” said Anton, “that the high drug dose group will turn out to be the ones whose cells showed no response.” And, he might have added, the lower dose group those with a blunted response.

However, this being blinded data, with full unblinded results not available untilthe end of this year, anything could turn out to be the case, ranging from complete concordance with microbicide use to a worst-case perverse response in which the lower rate of infections turn out to be in those who received placebo. At the moment all we can say is that the 33/33/33 split in cellular response is suggestive.