Microbicides in the Spotlight

Recent meetings highlight this young field, and its many points of convergence with AIDS vaccines

By Emily Bass*

The recent International AIDS Conference in Bangkok will be remembered as the meeting in which microbicides—topical compounds designed to prevent sexual transmission of HIV—attained the spotlight which has frequently eluded this small but growing field over the first two decades of its evolution.

Calls to hasten microbicides research started at the opening ceremony of the conference when United Nations Secretary General Kofi Annan urged the audience of thousands to do more to protect women and girls and ensure “that they have full access to the practical options that can protect them from HIV—including microbicides, as they become available.”

Throughout the week, a diverse array of speakers—many from outside the field’s small community of longtime advocates and researchers—also singled out microbicides as a crucial strategy for mitigating what Annan called the “terrifying” impact of the epidemic on women and girls. By the time the vaccine and microbicide “theme” day arrived on the last full day of the conference, Zeda Rosenberg, head of the International Partnership for Microbicides, set aside her prepared plenary text to remark, “When I wrote this presentation weeks ago, I could not have predicted that this would be a watershed conference for microbicides.”

The surge of interest in microbicides comes at a time when the field is coming into its own. In recent years the field has been energized by an influx of funding from a variety of sources including the Bill & Melinda Gates Foundation, the Rockefeller Foundation, and European governments. There has also been an awakening of interest among basic scientists who are focusing on determinants and early events of sexual transmission of HIV with an eye to developing effective microbicides.

see HIV and women: Statistics from a global crisis


As with AIDS vaccines, microbicide research faces challenges related to coordination among research entities, long-term planning for success or failure of first and second generation candidates, and the need for clinical trials capacity in resource-poor settings. The Bangkok meeting touched on these issues and they were explored in greater depth at the Microbicides 2004 (M2004) in London in March. A meeting of funders, researchers and advocates was also held in April in Washington, DC that laid the groundwork for an over-arching coordinating body that may bear some resemblance to the nascent HIV Vaccine Enterprise.

Preparing for—and debating—efficacy trials

By the end of 2005 as many as five efficacy trials involving six candidate microbicides could be underway (See table 1). Preparation for these trials has engendered creative and provocative tensions as the field has re-assessed the current candidates in light of more recent scientific techniques and understanding. Five of the candidates come from the same class (polyanions); the sixth candidate, SAVVY or C31G, is from the same class (surfactant) as nonoxynol-9 (N-9), the spermicide that failed to show efficacy—and even appeared to potentiate HIV infection in some groups—in a Phase III trial completed in 2000. Pointing to the similarity among the candidates, and concerns about the general safety profile of surfactants, critics within the field have suggested that not all of the trials should go forward.

Table 1. Planned or ongoing microbicide efficacy trials: 2004-2005

The majority view is that the five trials scheduled to begin should go ahead essentially as planned,” states Alan Stone, who chairs the International Working Group on Microbicides. “The minority view is that the case for going ahead with all six products is weak and that some of them should be put on ice.”

The current debates are informed by recent studies that have applied newer scientific techniques to these first generation candidates, many of which were developed more than a decade ago. For example, most of the candidates scheduled for testing were evaluated in macaque challenge models using X4-tropic strains (which use CXCR4 as coreceptor) of SHIV, even though R5¬tropic viruses (which use CCR5) are thought to be responsible for the majority of sexual infections. (A lab-adapted R5-tropic SHIV has recently been developed and is being used in some animal model evaluations of newer microbicide candidates.) In a plenary address at the M2004 conference, Robin Shattock (St George’s Hospital Medical School, London) reviewed data suggesting that polyanions may be less effective against R5-tropic strains. This is because polyanion activity is mediated through an interaction with the positively charged V3 loop; this region is exposed on unbound X4-tropic viruses but only becomes accessible on R5-tropic viruses during cell binding and fusion.

Shattock suggested that the next generation of products should reflect newly-acquired scientific insights. “We now need to move into a stage of rational drug development, especially now that we know more about target cells.” Shattock also suggested that pre-clinical development incorporate more studies of tissue penetration, since products which work by blocking cell-virus interactions will have to cross the vaginal mucosa and perhaps reach the draining lymph nodes to be truly effective.

Rather than arguing against any of the trials, many leaders in the field are instead urging that the upcoming studies be carefully designed—and supplemented with coordinated pre-clinical testing using newer methods. “It’s really important to move forward with these Phase III trials; they will certainly help validate pre-clinical assays and we’re hopeful that the candidates will show some efficacy. Even a partially effective candidate could be used in combination,” says Rosenberg.

Stone too is a proponent of moving ahead with the planned trials, but says that, “...it makes a lot of sense for the planned trials to be organized so as to maximize the potential for comparative safety, effectiveness and acceptability.” Stone and others have suggested linkages or cross-membership among the Data Safety and Monitoring Boards for each of these trials, so that class-specific similarities and differences can be more readily and rapidly detected during the course of the trials.

In the weeks following the M2004 conference, a group of funders, researchers and advocates involved in microbicide research met to discuss opportunities for coordination at a meeting convened by the Bill & Melinda Gates Foundation and the Alliance for Microbicide Development (AMD). The group recommended forming a coordinating body that would help facilitate harmonization across a number of areas including protocol design, monitoring, and decision-making for next-generation candidates. At press time, an initial proposal for such a body had been reviewed by major funders in the field and broadly approved for prompt implementation. “These discussions wouldn’t have happened without the debate [over efficacy trials],” says Polly Harrison, executive director of the AMD.

Multiple pathways, multiple targets

The microbicide candidates currently entering clinical trials are broad-spectrum candidates. The next generation of products is likely to emphasize more targeted approaches and to exploit new insights into the earliest events of sexual transmission.

The current understanding of this complex process was described in detail in plenary talks at M2004 by Shattock and Melissa Pope (Population Council, New York). As both speakers described, the vaginal mucosa contains several cell types that are susceptible to HIV infection, including CD4+ T cells, dendritic cells (DCs) and macrophages. HIV uses a variety of co-receptors to enter these cells, including CD4 and CCR5 for T cells. In addition, HIV can also utilize C-type lectin receptors, including DC SIGN, mannose receptor and Langerin, to attach to or enter DCs and macrophages (see Review).

Having picked up virus, migratory DCs can ferry it to the draining lymph nodes where it can infect many more CD4+ T cells as a result of the normal DC-T cell interactions.

An effective microbicide could work by preventing the virus from reaching its target cells, perhaps by serving as a physical barrier in the mucosa or through virucidal activity in the vaginal lumen; however, it is highly likely that a truly protective candidate will also have to block some or all of these virus-cell interactions to prevent establishment of local and/or disseminated infection.

A presentation by Qinxue Hu (St George’s Hospital, London) gave some indication of the complexity of designing such a candidate. Hu reviewed a series of studies (J. Exp. Med. 199, 1065; 2004) that tested the ability of various agents to block R5-, X4- and R5X4-tropic HIV infection in cervical explant models and DC cell lines. Compounds tested included AOP RANTES (CCR5- and CCR3-inhibitor), AMD3100 (CCR5-inhibitor), TAK-779 (CCR5¬inhibitor), mannan (mannose receptor blocker), monoclonal antibody (mAb) RPA-T4 (anti¬CD4), and mAb b12 and the fusion protein CD4-IgG2, both of which target gp120.

One study in activated (phytohemagglutinin-stimulated) cervical explant tissue found that AMD3100 and TAK-779 together inhibited infection by an R5X4 tropic isolate by about 80% as measured by p24 antigen release, emphasizing that CCR5 and CXCR4 are the primary coreceptors involved in infection of human cervical tissue.

The researchers then asked whether the panel of inhibitors could prevent DCs from internalizing whole HIV. They collected migratory cells emigrating from activated cervical explants 48 hours after HIV inoculation in the presence or absence of inhibitors, and then co-cultured them with indicator cells (PM1) to determine whether they contained infectious virus. Migratory cells containing infectious virus could be found in samples that were exposed to TAK-779 and AMD3100, suggesting that compounds which only block coreceptors may provide incomplete protection and that infection via migratory cells might still proceed. Substantial (80%) inhibition of infection of migratory cells was only achieved by a combination of mAb RPA-T4 and mannan.

Is it possible to simultaneously block the pathways that lead to localized infection and viral dissemination? As Hu reported, the only substances capable of blocking both pathways were ones that targeted gp120, rather than cellular co-receptors. MAb b12 and CD4-IgG2 reduced infection of T cells and migratory DCs by more than 95% in activated cervical explant tissue.

Extrapolating to vaccines, study co-author Shattock says, “It’s a really striking thing that if you do have neutralizing antibody, it just halts infection—even neutralizing virus that is taken up by DCs.”

Assessing the impact of local immune activation and inflammation

Heterosexual transmission accounts for the majority of new HIV infections worldwide, yet relatively little is known about the determinants of male-to-female or female-to-male viral transmission. This is due to the difficulty of identifying individuals during the acute phase of infection and to the limitations of current methods for measuring mucosal immune responses. The available data strongly suggest that sexually transmitted infections (STIs) and bacterial infections such as bacterial vaginosis in women and urethritis in men can all increase viral shedding in the genital tract of HIV-infected individuals. These infections—particularly herpes simplex virus type 2—may also increase susceptibility to HIV infection in men and women.

Several mechanisms may explain these enhancing effects, including STI-related ulcers which can serve as portals for viral entry across the mucosa. Non-ulcerative infections have also been linked to increased susceptibility to HIV infection and it’s thought that local immune activation may play a role, perhaps by triggering proinflammatory responses that enhance viral replication or by increasing the number of activated CD4+ T cells and DCs trafficking from the genital mucosa to the lymph nodes.

STIs and bacterial infections that enhance viral replication might raise the bar for microbicide or vaccine efficacy. Many planned efficacy trials (see Table 1) and vaccine preparedness efforts are collecting data on STI prevalence and incidence in order to tease out the impact of STI coinfection on other interventions; diagnosis and treatment of STIs is considered an important HIV prevention strategy. By the same token, a microbicide that protected against STIs other than HIV might still have a protective effect against HIV simply by reducing inflammation.

Other agents may actually cause inflammation. Like all mucosal surfaces, the vaginal epithelium can respond to any foreign substances, including microbicides and lubricants. At M2004, a presentation by Gustavo Doncel (CONRAD, US) showed that changes in the local vaginal environment can be caused by vaginal products such as lubricants, spermicides, placebo gels and candidate microbicides. Doncel measured the expression of proinflammatory genes and cytokines, including interleukin (IL)-1, -6, and -8 in immortalized human vaginal keratinocytes (VK-2/E6E7) that were pretreated with candidate microbicides, placebo gels and over-the-counter vaginal lubricants. Some of these products, including N-9 and a lubricant, caused increased production of these cytokines, as well as NF-κB.

Doncel is working to improve the field’s ability to identify candidates with potential safety problems at very early stages in the development process, a consequence of the adverse effects reported for N-9 during the Phase III efficacy trial. The standard preclinical safety analyses for that trial included in vitro cytotoxicity studies and a standard panel of animal safety tests that failed to predict the observed effects. But later analyses found that N-9 use increased proinflammatory responses, including cytokine and chemokine induction and an influx of activated macrophages in the cervicovaginal lavage (CVL) (J. Infect. Dis. 184, 418 2001). An assay such as Doncel’s that measured pro-inflammatory responses could be used as an early indicator of product safety, and several presenters suggested that these data should be collected for the candidates currently entering Phase IIb and Phase III trials. Efficacy data could then later be compared with these results to help validate predictive assays.

The field is also seeking early indications of product efficacy that could help guide decisions about launching large-scale efficacy trials. At M2004, Marla Keller (Mount Sinai School of Medicine, New York) described a strategy for gathering CVL fluid from HIV-infected women before and after treatment with a candidate microbicide called PRO 2000. Diluted CVL was spiked with HIV and inoculated on to susceptible cells to measure viral replication. Preliminary results indicate that post-treatment CVL inhibited replication of both HSV-2 and HIV. Since women already infected with HIV but unaware of their status will inevitably use any effective microbicide developed, safety and efficacy—including effects on viral activity and shedding—must be evaluated in this group. The same type of study could be used to gather preliminary data in HIV-uninfected women.

Timing of early events in sexual transmission

As the picture of sexual transmission becomes clearer and more detailed, both vaccine and microbicide researchers are focusing on the timing and sequence of early events—specifically the interval between initial infection of target cells in the vaginal mucosa and dissemination to the draining lymph nodes. Rapid dissemination would pose challenges for both of these prevention measures, since rapid transport to the draining lymph nodes could remove HIV from either sufficiently inhibitory concentrations of microbicide or potentially protective vaccine-induced immune responses.

Under some conditions viral dissemination may occur within a matter of hours. Chris Miller (University of California, Davis) and colleagues found SIV-infected cells in the draining lymph nodes within 18 hours of intravaginal exposure to highly-pathogenic SIVmac251 (J. Virol. 74;6087, 2000). However there may also be conditions under which systemic infection does not emerge for days or even weeks after initial exposure. At M2004 Miller raised this possibility with data from a new study (in press, J. Virol.) in which eight macaques were challenged intravaginally with multiple low-dose inocula (103 TCID50 of SIVmac251) over the course of several weeks; the animals became infected after an average of 8 challenges. In six infected animals, systemic infection was preceded by a period of transient viremia followed by an interval of “occult infection,” during which no virus was detected in PBMCs. Some of these animals also had SIV-specific T-cell responses during the period of occult infection.

Miller noted that these results differ from the kinetics of viremia and immune responses in animals infected with a high dose intravaginal challenge, after which systemic infection develops immediately with subsequent emergence of SIV-specific immune responses. He suggested that timing of viral dissemination may vary depending on the size of the initial challenge and that, under certain conditions, multiple exposures (such as might occur during intercourse with an infected partner) might give rise to localized infection that is not detectable in the blood. Miller also referenced the highly exposed, persistently seronegative sex workers in Nairobi, some of whom have HIV-specific immune responses in the absence of detectable HIV infection. He suggested that some of these women might actually have occult infection and that this phenomenon might be more common than is appreciated.

The establishment of infection at the portal of entry and timing of dissemination might also be affected by the numbers and types of cells that are initially infected. In a paper published shortly after M2004 (Proc. Natl. Acad. Sci. USA 101, 5640; 2004), Ashley Haase (University of Minnesota) proposed that HIV exploits whichever cell types are most common in the vaginal epithelium, draining lymph nodes and other lymphatic tissues to which virus spreads. These are principally resting T cells, which have not been considered permissive targets for HIV infection, but Haase found that small Ki67 negative T cells—which he thinks might actually be activated cells returning to a resting state—can indeed be infected. (Ki67 is a marker of cell proliferation.) In the healthy mucosa, these cells greatly outnumber activated T cells, DCs and macrophages, and are thus more like to be infected because of their availability. While they only maintain a low level of viral replication, Haase thinks they may play a critical role in sustaining infection in its earliest stages and that the role that the infected activated CD4+ T cell plays is to more efficiently disseminate virus because of higher levels of replication. “The kinetics of infection will depend on the sizes of these founder populations of infected cells, and in turn on factors that influence the integrity of the mucosal barrier and thus the access of virus to cellular targets” Haase says.

A novel ex vivo model presented by Julie McElrath (University of Washington) could shed additional light on the types of cells first infected. McElrath described a model developed with colleague Florian Hladik that uses a suction blister method to separate the outermost epithelial layer from the underlying stroma of sheets of vaginal epithelia obtained from healthy women undergoing tissue repair surgery. She described how the basal side of the epithelium remains fully intact during the procedure, allowing for in situ analysis of virus-cell interactions with minimal damage to or changes in cell function. The basal side of this epithelial sheet displays an organized pattern of deep depressions which nearly reach the vaginal surface. Most of the mucosa’s potential target cells for HIV (infiltrating lymphocytes and Langerhans cells) accumulate around these basal depressions, which are also among the thinnest regions of the epithelial barrier and therefore most easily breached.

McElrath presented unpublished data from a study in which sheets of the outer epithelial layer were co-cultured with fluorescently-tagged lab-adapted strains of HIV. Confocal and electron microscopy of fixed sheets revealed HIV on the surface and inside of both CD4+ T cells and Langerhans cells, suggesting that the two cell types might be infected in parallel. (One model of HIV infection holds that DCs are the first cells to be infected and that they then transmit the virus to CD4+ T cells). In discussion, McElrath noted that since her group had not yet seen viral budding or fusion from epithelial cells, it was not possible to say with certainty whether HIV productively infects these cells or merely sequesters virus within these cells. This new model remains to be fully validated, but it could prove to be an additional tool for understanding early events in a physiologically relevant system.

Next steps

The next microbicide biannual meeting will take place in 2006, as will the next International AIDS Conference. By then early data could be available on safety or perhaps even efficacy from the microbicide candidates currently advancing into clinical trials. These data will give the first hint of whether these current approaches provide any protection against HIV infection. As new tools and techniques are developed, the next two years will likely bring further insights into the early events of sexual transmission of HIV that will help guide the design of next generation microbicide and vaccine candidates. The challenges are formidable, but as Zeda Rosenberg said in Bangkok, “The science is there, the technology is there, and most of all the passion and dedication of those in the field is palpable. And failure is not an option.”

*Emily Bass is senior writer of the IAVI Report