Since the late 1980s researchers have recognized that HIV isolates (viruses sampled from infected individuals) can be divided into one of two categories based on their ability to replicate in particular cells in the laboratory. Originally, a variety of terms were used to describe this phenomenon. One of the most common was the designation of viruses as either syncytium-inducing (SI -- these viruses caused cells in the lab dish to clump together and form clusters of dying cells called syncytia) or non-syncytium-inducing (NSI). The classification of SI and NSI viruses was subsequently aided by the discovery that SI viruses could replicate in specific laboratory cell line -- known as MT-2 cells -- whereas NSI viruses could not. Confusingly, an alternative classification dubbed these viruses T cell tropic (T-tropic) or macrophage tropic (M-tropic), respectively, even though both types could replicate in T cells.

In early 1996, the underlying biological basis for this distinction became clear when researchers discovered that, in addition to latching onto the CD4 molecule in order to gain entry into T cells, HIV isolates utilized one of two different coreceptors: either CCR5 or CXCR4. It quickly became apparent that CCR5 use correlated with the NSI/M-tropic classification while viruses that used CXCR4 were SI/T-tropic. As it turns out, the MT-2 cell line only expresses CXCR4. The older nomenclature has been replaced by the simpler designations of R5-tropic or X4-tropic. Rare "dual tropic" viruses that can use either co-receptor have been reported, but this designation is more commonly (and perhaps misleadingly) applied to mixed populations of R5-and X4-tropic HIV.

One salutary result of this progress is the development of pharmaceutical compounds designed to inhibit the interaction between HIV and its various coreceptors. However, the larger questions of why HIV bifurcates into two variants with differing tropisms and how these variants relate to disease pathogenesis remain unanswered, leaving a cloud of uncertainty looming over the clinical development of both R5 and X4 inhibitors. Similarly, now that large human studies of coreceptor inhibitors are underway, outstanding questions regarding the biological functions of CCR5 and CXCR4 receptors in humans have inspired research and regulatory interest in the possibility of unpredictable toxicities. The recent termination of the development of GSK's CCR5 inhibitor, aplaviroc, due to several cases of severe liver toxicity further underscores the need for vigilance in this regard.

Co-Receptor Biology

Both CCR5 and CXCR4 belong to a family of molecules known as 7-transmembrane receptors. These receptors are snake-like structures with portions both inside and outside of the cell (think of a picture of the mythical Loch Ness monster with a trail of humps visible above the water line -- 7- transmembrane receptors have seven loops protruding from the cellular membrane). The primary function of CCR5 and CXCR4 is to interact with specific chemical messengers called chemokines; binding of the chemokine to the receptor causes the chemokine/receptor complex to submerge into the cell and initiate a cascade of signals that affect the cell's behavior. A ligand is a substance capable of binding to a receptor. For example, binding of the chemokine CCL5 (which stands for chemokine ligand 5, also known as RANTES) to CCR5 can trigger the cell's migration to specific locations within the body. The chemokine SDF-1 binds to CXCR4, and these interactions are important in many settings, including embryonic development (mice genetically lacking CXCR4 die in utero).

Notably, chemokine receptors can be rather promiscuous in their ability to bind different chemokines. CCR5 is known to interact with CCL3, CCL4, and CCL8 in addition to CCL5. The reverse is also true; certain chemokines can interact with more than one receptor. CCL5, for example, can bind CCR1 and CCR3 in addition to CCR5. The functions of all the known chemokine receptors and chemokines are not fully characterized, but broadly speaking, they seem to be involved in cell migration and/or inflammation.

In terms of which cells possess the two HIV coreceptors, CCR5 is primarily found on activated T cells but CCR5 expression has also been reported (primarily from mouse studies) on multiple other cell types including macrophages, dendritic cells, neutrophils and hepatic stellate cells. CXCR4 is broadly expressed in cells of both the immune and the central nervous systems. A critical question facing developers of co-receptor inhibitors is whether CCR5 and CXCR4 inhibitor compounds interfere with chemokine/receptor interactions, and if so, whether such interference has harmful consequences.

Immunotoxicities

On December 14th 2005, the Forum for Collaborative HIV Research held a roundtable discussion concerning the potential immunotoxcities of entry inhibitors (copies of the presentations are available at www.hivforum.org/projects/CCR5.htm). Mark Swain reviewed two recent studies that investigated the role of CCR5 in a mouse model of hepatitis. The model involves administering the drug Concanavalin A (Con A) to mice, which triggers an immune system attack on the liver leading to hepatitis. Swain's research group compared the severity of hepatitis in normal mice compared to mice genetically bred to lack CCR5 receptors (called CCR5-/- or CCR5 knockout mice). The study found that knockout mice who received Con A developed profound hepatitis that progressed rapidly to fulminant liver failure. Three of six knockout mice died within eight hours compared to no deaths among the normal mice. Researchers searched for an explanation and found that a specialized group of T cells called natural killer T cells (NKT cells for short) appeared more resistant to death (apoptosis) in the knockout mice. The NKT cells from knockout mice also produced more of the cytokine interleukin 4 (IL-4) compared to their CCR5 possessing counterparts. Blocking IL-4 or depleting NKT cells using antibodies reduced the extent of the liver damage in knockout mice. This study was published in June 2005 in the Journal of Immunology (J. Immunology, 174: 8027-8037, 2005).

The second study discussed by Swain was conducted by a group of Belgian researchers led by Christophe Moreno and involved the same mouse model of Con A-induced hepatitis. The researchers reported similar findings to Swain's group, namely, increased mortality in CCR5 knockout mice. They also reported that, in normal mice, serum levels of the CCR5 ligands CCL3, CCL4, and CCL5 were significantly increased following Con A injection and that CCR5- expressing liver mononuclear cells (comprising T cells, macrophages, natural killer cells and NKT cells) were recruited to the liver. The CCR5 knockout mice also exhibited increased production of interleukin 4, tumor necrosis factor, CCL3, CCL4, and CCL5, and a notable infiltration of T cells, macrophages, natural killer cells and NKT cells into the liver, among which were cells expressing the chemokine receptor CCR1 (which can also bind to CCL3 and CCL5). The researchers tried blocking CCR5 ligands with antibodies to see whether the hepatitis would improve. Blocking CCL5 significantly reduced serum ALT levels and hepatic mononuclear cell infiltration, whereas blocking CCL3 and CCL4 had no effect. Thus, it appears that the absence of CCR5-expressing cells can result in increased levels of circulating CCL5, potentially exacerbating immune-mediated liver damage. This study was published in the journal Hepatology (Hepatology 42:854-862, 2005).

Human Knockouts

While much of the mice data sounds disconcerting, it remains unclear whether the experience with mice bears any relation to what might happen to humans receiving a CCR5 inhibitor. The closest human equivalents to CCR5 knockout mice are certain rare individuals who possess a genetic mutation that prevents the expression of functional CCR5 receptors. This mutation is dubbed CCR5 delta32; when inherited from both parents, a person lacks functional CCR5 on his/her cells and is said to be "homozygous" for the mutation. If a person inherits the mutation from only one parent, he or she is displays reduced CCR5 levels on his/her cells and is said to be "heterozygous" for the mutation. People homozygous for the CCR5 delta32 are highly resistant to HIV infection, although some cases of such individuals becoming infected with X4-using HIV have been reported. While not protected from HIV infection, individuals who are heterozygous for the CCR5 delta32 mutation appear to have slower disease progression.

To date, most studies have not revealed obvious, serious health problems among delta32 homozygotes or heterozygotes, but the literature on delta32 homozygotes remains relatively sparse. There has been one report that delta32 homozygotes infected with hepatitis C experience less inflammation but more fibrosis (scarring) of the liver, compared to infected individuals who lack the delta32 mutation. Data published previously suggested that delta32 homozygotes are more susceptible to hepatitis C infection and have a diminished response to treatment, but recent studies have contradicted these assertions. The delta32 mutation has also been strongly associated with a disease called primary sclerosing cholangitis (PSC). PSC is a disease involving inflammation and scarring of the bile ducts, which can cause bile to accumulate in the liver, damaging liver cells and leading to cirrhosis.

Perhaps the most dramatic human data come from a very recent study of West Nile Virus (WNV) suggesting that delta32 homozygotes may be more susceptible to symptomatic infection with this mosquito-borne pathogen (J Exp Med 203;1:35-40, 2006). The investigation was conducted based on results of a prior study in CCR5 knockout mice demonstrating that the mice experienced exacerbated symptoms as a consequence of reduced T cell trafficking to the brain (J Exp Med 202;8:1087-1098, 2005). The human study analyzed two different cohorts of individuals with laboratory confirmed symptomatic WNV infection. The results found that delta32 homozygotes were significantly overrepresented in both cohorts relative to the expected frequency of the mutation in the population. In one of the two cohorts, delta32 homozygote genotype was also associated with an increased risk of death. The study authors went so far as to conclude: "Our results have important implications regarding the potential safety of CCR5- blocking agents now under development for the treatment of HIV/AIDS. Clinical care of individuals taking these medicines while residing in WNV-endemic areas may mandate strict measures to limit mosquito exposure and a high index of suspicion for symptoms consistent with WNV."

The extent to which any or all of these problems might occur in the setting of pharmacological CCR5 inhibition cannot be known until a larger amount of safety data accumulates on CCR5 inhibitors. It is possible that the redundancy present in the chemokine/chemokine receptor system may allow other receptors to assume the function of CCR5 in delta32 homozygotes, and that something similar may occur in people receiving CCR5 inhibitors. But the safety and toxicity issues associated with CCR5 receptor blocking are significant enough that regulatory authorities are requiring extensive long term follow up -- up to 5 years -- of individuals participating in clinical trials of CCR5 inhibitors.

Coreceptors and Pathogenesis

Even back when X4-using viruses were characterized as syncytium-inducing, researchers recognized that these viruses became detectable almost exclusively during late stages of disease. Large cohort studies now indicate that around 50% of people progressing to AIDS will show evidence of a shift from R5- to X4-using HIV. Initially, researchers assumed that the emergence of X4 viruses was causing accelerated disease progression in these individuals. This assumption remains one of the focal concerns regarding current trials of CCR5 inhibitors: that these drugs might select for X4 viruses and that X4 viruses might accelerate disease progression.

Recently, however, the supposition that X4 virus causes rapid progression has been questioned. The countervailing hypothesis is that X4 virus emerges as a consequence of the severe T cell depletion seen in advanced disease, perhaps due to the loss of appropriate cellular targets for R5-using HIV (this argument is rehearsed in excruciating mathematical detail in a new paper in the Journal of Virology, see J. Virol 80;2:802-9, 2006). This line of reasoning is further based on evidence that R5 HIV has a competitive advantage over X4 HIV. The precise nature of this advantage remains uncertain, but a number of possible explanations have emerged:

• R5 HIV makes more virus per infected cell than X4 HIV

• R5 HIV is preferentially taken up by dendritic cells and transferred to CD4 T cells

• X4 HIV is preferentially suppressed by CD8 T cells

There are data that support each of these notions, but as of yet, nothing conclusive. It is fair to say, however, that some of the heightened early concern about the potential danger of tropism shifts has waned. A shift to X4 virus has been reported in several recipients of CCR5 inhibitors, but it does not appear to have harmful clinical consequences. CCR5 inhibitor studies are currently using an assay that attempts to quantify proportions of R5- and X4-using HIV as a screening tool. When applied to recipients of normal HAART regimens, the assay has found that roughly 40-50% of HIV+ people show evidence of X4 virus, though there is no association between X4 presence and disease stage.

CXCR4 Inhibitors

One way to address the concern regarding emergence of X4 virus is to focus pharmacological inhibitor development on this coreceptor. Since CXCR4 knockout mice cannot be bred, and there are no humans lacking CXCR4 expression (as there are with CCR5), trials of CXCR4 inhibitors are subject to equally rigorous scrutiny for signs of toxicity.

At least one such drug (AMD070) is undergoing clinical evaluation. AMD070 has been tested in a very small group of HIV-infected individuals, and the ACTG is currently sponsoring a larger Phase II trial that has enrolled just four people after more than a year of accruing. Recruitment was recently temporarily stopped due to hepatotoxicity seen in a parallel dog study. Upon announcement of the halting of the ACTG study, Anormed opened a similar study. The FDA is likely to keep a close eye on these studies should they move forward.

Conclusion

The targeting of HIV co-receptors represents an exciting new frontier for antiretroviral therapeutics even as it signals a journey into deep, uncharted waters. There are no therapeutic precedents for inhibiting human chemokine receptors. Until more data become available, it is difficult to predict just how rocky this sail may become. In the meantime, FDA is paying very close attention to safety concerns, including convening a consultative meeting in conjunction with the Forum for Collaborative HIV Research to solicit input from the wider community. This meeting was originally slated for January 18th 2006 but has been rescheduled for May 30, 2006.

For more information about coreceptor inhibitors and other drugs in development, see Rob Camp's 2006 Clinical Pipeline report with links to data.