Microbicides 2008: Cellulose sulphate has unexpected tissue toxicity

Investigations into the candidate microbicide cellulose sulphate (Ushercell) using the substance on vaginal tissue samples have revealed that it causes loss of tissue integrity by destroying proteins that bind cells together. This allows HIV to leak into the underlying tissues.

Trials of Ushercell were halted in 2007 after women who received the microbicide in one study became infected more frequently than those who received the inactive placebo gel.

A study presented at the Microbicides 2008 conference last week in Delhi found that this unexpected toxicity is also observed to a greater degree with the spermicide nonoxynol-9, and to a very limited degree with the microbicide PRO 2000, which is still involved in clinical trials, though in this case the tissue damage appears reversible.

This effect causes neither macroscopic tissue damage that can be seen by colposcopy nor a significant inflammatory immune response (though repeated application can cause inflammation), and would have been unobserved by standard ways of assessing microbicide toxicity.

The CONRAD trial of cellulose sulphate was stopped in January 2007. It involved 1428 high-risk women at four sites in Benin, South Africa, Uganda and India, and was stopped when it was noted that there were significantly more (p=

The eventual total of infections for the whole study was 60 on cellulose sulphate (8.4% of participants) and 27 in placebo (3.8%), though this did not quite reach statistical significance. Another trial run by Family Health International in Nigeria was stopped at the same time – see this report for more.

The Microbicides 2008 conference heard more data from the Durban site (Govinden). At this site 606 women were enrolled before the trial was halted. There was an extremely high local HIV prevalence (49.7%) amongst women screened for the trial and only 46.6% reported regular condom use.

There were 28 new infections at the Durban site, 7.56% in cellulose sulphate recipients and 5.91% in placebo recipients. This 28% difference was not statistically significant. However there were significantly more infections in patients diagnosed with an STI – 18.31%.

There was a more than threefold greater risk of infection in patients diagnosed with chlamydia, bacterial vaginosis or trichonomiasis, and a more than fourfold higher risk of infection in participants who reported more than one new partner between the quarterly trial visits (p=0.01).

There were however no seroconversions in married women and marriage conferred an estimated fourfold protective effect against infection.

Delegates were reminded why cellulose sulphate was considered as a microbicide by a report on an animal model trial using monkeys (Saifuddin). In this trial twelve rhesus macaques were randomised to receive either cellulose sulphate or placebo while being challenged with two viruses, one using the CCR5 receptor and one the CXCR4 receptor.

The former is the type most often transmitted, and over the course of 13 weeks and 13 challenges five out of the six placebo-receiving monkeys were infected (though only four produced antibodies to HIV, i.e. seroconverted) against zero on cellulose sulphate (p=0.005). Only two monkeys became co-infected with CXCR4 virus.

However it was observed that there were transient viral ‘blips’ in two of the six cellulose sulphate-receiving animals (to 3,000 and 1,000 copies) and four out of the six showed signs of integrated proviral DNA in the lymphocytes in their blood, as did the two non-seroconverting placebo recipients. Four out of these six monkeys showed virus-specific immune activity in their CD4 cells.

Microbicides researcher Robin Shattock commented after this presentation that the presence of proviral DNA and transient viremia “was not indicative of a clean response” to the candidate microbicide.

The reason cellulose sulphate may have increased vulnerability to HIV was revealed by Pedro Mesquita of Albert Einstein College of Medicine at Yeshiva University, New York. His findings were also reported at the recent CROI Conference – see here for that abstract.

Mesquita incubated uterine epithelial cells and reconstituted vaginal tissue for 18 hours with one dose of nonoxynol-9 (N-9), cellulose sulphate, PRO 2000, and tenofovir gel, and also for repeated periods of two hours apiece. He measured tissue integrity by a technique called transepithelial electric resistance (TER), whereby resistance drops as cell architecture is damaged, and imaged the tissues using confocal microscopy, in which different proteins are stained with different fluorescent dyes.

In cells repeatedly incubated with N-9, TER dropped to zero after one day, and with cellulose sulphate after two days. PRO 2000 produced a 40% drop in TER by the time of microbicide removal, but levels recovered to normal after it was removed, unlike with the other two gels. Tenofovir gel produced no drop in TER.

Confocal microscopy showed that cellulose sulphate and N-9 selectively destroyed a protein called desmoglein, This is one of a group of hook-like proteins called cadherins that literally stitch cells together, and loss of these proteins causes tissues to become ‘leaky’. PRO 2000 and tenofovir gel did not have this effect.

This was shown to enhance HIV transmission by allowing diffusion of virus across the epithelium (the mucous membrane), as shown by measuring the HIV p24 protein. In cells treated with N-9, 25 nanograms per millilitre (billionths of a gram) of p24 diffused across cells and in cells with cellulose sulphate, ten nanograms. There was no viral translocation with the other two microbicides.

Prolonged incubation over 18 hours with cellulose sulphate also produced a previously-missed inflammatory response. Cells treated with cellulose sulphate developed threefold higher levels of the cell activation marker nuclear factor kappa B (NFkB), a molecule also associated with HIV proliferation. There was even a small degree of inflammation caused by PRO 2000, as indicated by other cellular markers, though not NFkB.

These changes would have been unlikely to be observed using the toxicity assays in safety trials prior to the current phase III trials, and Dr Mesquita recommended that his assays become standard practice in future.