Research Briefs

By Philip Cohen, PhD

Attacking HIV from Inside and Out

Two recent reports suggest that the use of the same antiviral taken orally or applied vaginally as a microbicide can prevent immunodeficiency virus infection in a rhesus macaque model. While either treatment was able to lower the frequency of infection, the authors conclude that effective prevention may require a multi-pronged approach with ARVs introduced by both routes, perhaps in combination with a partially-effective vaccine.

In the first report, John Moore of Weill Medical College of Cornell University and Ronald Veazey of Tulane National Primate Research Center and their colleagues explored the use of three different types of compounds in a vaginal microbicide: BMS-378806 (produced by Bristol-Myers Squibb), a gp120-binding small molecule inhibitor that prevents viral attachment to CD4 and CCR5 receptors; another small molecule called CMPD167 (produced by Merck), which clogs the gp120 binding site on CCR5; and C52L, a fusion-inhibiting peptide similar in sequence to the drug T20 (or enfuvirtide).

The researchers used a high virus dose vaginal challenge model of infection in macaques using SHIV-162P3, a chimeric virus with an HIV envelope and an SIV core. All nine control animals became infected. In contrast, all three ARVs delivered individually or in combinations were able to lower infection rate—only 11 out of 51 animals receiving some combination of the drugs became infected. The data also suggested that the two drugs could be applied hours before and still provide some protection (Nature 438, 99, 2005).

In a separate report, the same research groups used this macaque model to test whether oral delivery of CMPD167 could prevent vaginal infection by SHIV. The drug was given either twice a day for 4 days prior to viral challenge, or 10 days after challenge, or both before and after. While the orally-delivered drug appears to prevent infection, the results were less clear cut than when the same drug was used in a microbicide.

Only the treatment group where animals received no drug before infection and twice daily doses for 10 days after infection had a statistically-significant reduction in infection rate. But taking into account all animals that received the drug post-challenge (including those which received it pre-challenge as well), the reduction in infection rate was much more significant. Only 10 out of 20 animals receiving drug after SHIV challenge became infected, as opposed to 16 out of 18 of control animals (Nature Medicine 11, 1293, 2005).

Whether either route of delivery will provide protection from infection in humans isn't clear. But as a step towards taking one of these strategies to the clinic, Bristol-Myers Squibb and Merck recently announced they had granted the International Partnership for Microbicides a royalty-free license to use BMS-378806 and CMPD167 or closely-related compounds in microbicide trials in developing countries.

Hooking Up to Stimulate Immunity

Using drugs that activate Toll-like receptors (TLRs) to stimulate adaptive immunity is a hot topic among vaccine researchers (see Toll bridge to immunity). Now Robert Seder of the Vaccine Research Center at the NIH and his colleagues report that such TLR agonists can help stimulate both cellular and humoral immunity against HIV Gag protein in nonhuman primates—and they see dramatic improvement in the cellular response when the drug and protein are chemically linked.

Experiments in mice and cultured human cells have shown that TLRs influence adaptive immunity by activating dendritic cells (DCs), which present antigen to T cells and release factors that stimulate T cell differentiation and expansion. But whether TLRs can do the same in humans or nonhuman primates hasn't been clear. Seder's group has now tested the ability of three different TLR agonists to elicit T cell responses to HIV Gag protein in Indian rhesus macaques (Proc Natl Acad. Sci. USA 102, 15190, 2005).

The animals received an injection of Gag alone or mixed with a solution containing a different agonist against either TLR7/8 (3M-012), TLR8 (3M-002), or TLR9 (CpG oligodeoxynucleotide) four times at four week intervals. In addition, one group received an injection of Gag crosslinked to TLR7/8 agonist with ultraviolet light. At different time points, Seder's team performed ELISPOT assays to determine the frequency of IL-2 and IFN-g producing T cells. Animals receiving the TLR7/8 or TLR9 agonists had significantly higher levels of IL-2+ and IFN-g+ T cells compared to macaques receiving only Gag protein. In contrast, the TLR8 agonist had little effect. But linking the TLR7/8 agonist to the protein had a dramatic effect. Six weeks after the final injection, IL-2+ and IFN-g+ T cells were six-fold higher for the protein-linked agonist compared to the same agonist delivered with the protein as separate molecules.

The researchers then used nine-color flow cyotmetry to more finely characterize the Gag-specific memory T cell response. They measured production of IL-2, which is important in sustaining memory, and IFN-g and TNF-a, which mediate effector function. Here the effect of the conjugation was just as striking. The highest average frequency of cytokine-producing CD4+ and CD8+ T cells was elicited by the protein-linked agonist. The quality of the response was also significantly altered. For the TLR7/8 or TLR 9 agonist in solution, about 40% of CD4+ T cells were producing only IFN-g and less than 25% were producing all three cytokines. But in response to the protein-TLR7/8 agonist conjugate, 40% of CD4+ T cells were producing all three cytokines and about 25% were producing TNF-a and IFN-g.

There was an even stronger contrast in the results for CD8+ memory T cells. On average, the unlinked TLR agonists had no effect or barely increased the frequency of cytokine-producing CD8+ T cells over the protein alone, while the linked agonist had a significant effect. Flow cytometry demonstrated that about 25% of these cells were producing all three cytokines, while 40% produced TNF-a and IFN-g. With the exception of the TLR8 agonist, all the drugs elicited high titer antibodies against Gag.

The researchers demonstrated that this four injection protocol with the Gag-linked TLR agonist produced a T cell response comparable to replication-defective adenovirus expressing Gag, suggesting it could be used instead of or in conjunction with such viral vectors as a prime or boost. This team is now repeating their experiment with the TLR7/8 agonist linked to SIV Gag in order to test whether this vaccination protocol by itself can protect animals against challenge by SIV.

A New Way to Snuff the Viral Fuse

A crucial step in HIV's replication cycle is when the virus fuses with its target cells. Inhibiting the process of viral fusion has become a promising approach for HIV therapeutics and is the mode of action of T20 (or enfuvirtide), a drug that works by binding the membrane fusion-promoting gp41 protein. Now John Shiver's team at Merck and their colleagues report that a monoclonal antibody that binds the same gp41 region as T20 is able to inhibit viral fusion of diverse HIV-1 clinical isolates, suggesting a novel strategy for eliciting broadly neutralizing antibodies in a preventive vaccine (Proc Natl Acad. Sci. USA 102, 14759, 2005).

T20 blocks HIV fusion by binding the heptad repeat 1 (HR1) region of gp41. After HIV attaches to a cell through gp120, the HR1 region plays an important structural role in a series of rapid and dramatic shufflings of protein domains that culminates in fusion of the viral and cellular membranes. In particular, T20 stops a transition of HR1 from a so-called prehairpin structure to a bundle of six alpha-helices. Shiver and his team began their search by selecting antibodies from a phage library of single chain antigen-binding Fv antibody regions derived from normal human B cells using two peptides designed to mimic the HR1 structure in the prehairpin as their target antigen. This process identified 100 different candidates.

The researchers then tested these in an HIV fusion assay. The report describes full characterization of one promising Fv, named 5H, which inhibited fusion in a dose dependent manner. They demonstrated that this antigen-binding region maintained its ability to inhibit fusion even when it was converted to a full IgG molecule. They also confirmed through mutational, biochemical, and structural analysis that the antibody binding site on gp41 overlapped with highly conserved amino acids in the HR1 region.

This monoclonal antibody was able to neutralize diverse HIV isolates, 9 out of 19 viruses tested, including examples from subtypes B, C, and F. However, compared to broadly neutralizing antibodies which have been isolated from HIV-infected individuals—IgG1b12 and 2F5—D5 was at least 10 times less potent and neutralized fewer HIV isolates. And T20, which binds the same region and presumably acts by the same mechanism, was an extremely potent inhibitor of all 19 strains. Shiver's group is now trying to select more potent variants of the antibody by using in vitro antibody evolution.

The team also evaluated the ability of their prehairpin mimics to elicit HIV neutralizing antibodies when injected into rhesus macaques. Unfortunately this has repeatedly resulted in high-titer non-neutralizing antibodies. But the researchers point out that by studying the binding properties of D5 they may be able to design immunogens that elicit more powerful neutralizing antibodies to the same epitope.

For Whom the B Cell Tolls

The role of Toll-like receptor (TLR) activation in generating a humoral immune response goes beyond their ability to stimulate dendritic cells (DCs), concludes a new report by Chandrashekhar Pasare and Ruslan Medzhitov of Yale University. They have shown in mice that direct activation of TLRs on B cells is necessary for robust production of some classes of antigen-specific antibodies (Nature 438, 364, 2005).

This work is part of a growing body of evidence supporting multiple roles for TLRs in generating adaptive immunity, making them of great interest to immunologists and vaccinologists (see Toll bridge to immunity). For example, when the gene for MyD88—a signaling adaptor protein crucial to the function of many TLRs—is knocked out in mice, the ability of these animals to mount strong antigen-specific antibody responses is severely compromised. This makes sense given the role of TLRs in stimulating DC maturation and activating T helper cells. However it wasn't clear if TLRs on B cells also played an important role.

To get at this question, the researchers used mouse genetics and cell transplants to test the effect of TLR signaling on B cells specifically. They started with mice with a defect in the immunoglobulin m chain gene, which renders the rodents deficient in B cells. These animals were immunized with human serum albumin (HSA) and lipopolysaccharide (LPS), a ligand for TLR4, to generate HSA-specific memory T helper cells. After 90 days the animals received an infusion of purified naïve B cells from wild-type mice and were immunized again with either HSA alone or HSA with LPS. While memory T helper cell responses were comparable in both groups, the HSA-specific IgG1 response dropped by about half if LPS was not included in the immunization, suggesting TLR signaling on B cells boosted the response.

Indeed, when the researchers transferred B cells from mice with MyD88 or TLR4 knockouts into B cell deficient mice and then immunized with HSA and LPS, levels of HSA-specific IgM and IgG were produced at no greater than 25 per cent of levels produced by wild-type B cells. The defect appeared to be very specific to the production of certain Ig classes. The production of IgE antibodies was not significantly impaired in the MyD88 knockout B cells and there was no difference in the homing or survival between the MyD88 and TLR4 knockout B cells and wild-type.