Viral antichemokines: from pathogenesis to drug discovery

In a recent JCI Commentary, Steven Kunkel reviewed the organization of the chemokine system and its important role in the development, differentiation, and deployment of mammalian leukocytes (1). The system is massive (about 50 ligands and 18 receptors are known), reflecting the complexity of its major client, the immune system. Now in this issue of the JCI, Liu and colleagues remind us that an antichemokine system has evolved in viruses, presumably as a strategy to evade the immune system, and they provide the first glimpse of how antichemokines might be exploited clinically as “ready-made” anti-inflammatory drugs (2).

Antichemokines comprise a major group of virus-encoded chemokine modulators (Table 1) and consist of three subgroups based on structure and mechanism of action: (a) chemokine homologues that act as chemokine receptor antagonists (e.g., MC148R of Molluscum contagiosum virus [MCV]; ref. 3); (b) plasma membrane–expressed chemokine receptor homologues, which function as chemokine scavengers (e.g., US28 of human cytomegalovirus; ref. 4); and (c) three subtypes of secreted chemokine-binding proteins, which have unique structures and unknown ancestry but function as extracellular chemokine scavengers (examples include M3 of γ-herpesvirus 68 [5, 6] and various poxivirus proteins, including M-T7 of myxoma, investigated by Liu et al. in this issue [2]). Three other groups of virus-encoded chemokine modulators have quite different functions. They include (a) chemokine receptor homologues, such as open reading frame 74 of Kaposi’s sarcoma–associated herpesvirus (7), which serves as a growth factor and angiogenic factor; (b) chemokine homologues that function as chemokine receptor agonists, including vMCK-1 of murine cytomegalovirus, which promotes viral dissemination via monocytes (8); and (c) nonchemokine agonists and antagonists of chemokine receptors, encoded by the HIV genome.

Table 1.

Viral chemokine modulators

Viral chemokine modulators, in turn, are part of a larger group of viral proteins with obvious homology to host proteins. These primarily include immunomodulatory, growth factor, and cell cycle control proteins (9). Interestingly, M-T7 is a hybrid. It has one domain homologous to the extracellular region of the IFN-γ receptor, as well as a COOH-terminal chemokine-binding domain. Chemokines bind M-T7 via their COOH-terminal glycosaminoglycan–binding (GAG-binding) domain (10). Fortuitously, M-T7 binds only to the rabbit form of IFN-γ, whereas for chemokines it crosses species barriers, thus permitting its use as a selective chemokine blocking agent in species other than rabbits.

Much of the rapid progress in this field can be attributed to major advances in viral genomics and to the advent of computer cloning. Examples continue to be found as additional viral genomes are sequenced and analyzed. However, in most cases it has been difficult to define relevant biological correlates, owing either to difficulty studying the virus (as in the case of MCV, which has not been grown in culture) or to the absence of animal models for the viral disease. Nevertheless, the general notion that chemokines act in antiviral host defense has solid support from studies with knockout mice lacking the chemokine macrophage inflammatory protein-1α (MIP-1α). These animals show decreased inflammatory responses to influenza A, Coxsackie B, and murine cytomegalovirus (11, 12). Thus it is tempting to speculate, for example, that the odd absence of inflammatory cells in pathologic lesions caused by MCV results from an antichemokine shield provided by secretion of its antichemokine MC148R (3). Myxoma virus, which causes a fatal systemic immunosuppressive disease known as myxomatosis in the European rabbit Oryctolagus cuniculus, is well suited to experimental analysis, and of all secreted myxoma gene products analyzed to date, deletion of M-T7 most profoundly attenuates virus virulence (ref. 13; G. McFadden, personal communication). However, whether this is due to blockade of IFN-γ, chemokines, both, or neither is still not clear.

Viral antichemokines typically have broad spectrum activity (e.g., M-T7 binds multiple CXC and CC chemokines, as well as the C chemokine lymphotactin), which suggests conversely that the host mounts a broad antiviral chemokine response. This is consistent with studies of chemokine expression in patients with immunologically mediated diseases, such as multiple sclerosis, rheumatoid arthritis, and psoriasis. Yet, counterintuitively, knockout or neutralization of just one chemokine or receptor in vivo markedly attenuates pathology in numerous and diverse animal models of inflammation (reviewed in ref. 14). While detailed studies of the spatial and temporal aspects of chemokine expression during the course of inflammatory challenge are needed to more fully interpret these results, these successes may justify developing antichemokine therapies for the clinic.

Recently several potent small molecule antagonists of specific chemokine receptors have been discovered, and more are on the way (14). Some of these were obtained “off the rack” as HIV entry blockers, after the discovery that the chemokine receptors CCR5 and CXCR4 play a major role in this process (15). Others were found by drug discovery efforts targeting specific chemokine receptors. Additional blocking strategies underway include development of humanized neutralizing antibodies to chemokines and chemokine receptors. Preclinical studies and disease indications, eagerly awaited for all of these compounds, are now available for M-T7. As Liu et al. (2) report, a single intravenous injection of just 0.017 pg/g of M-T7 protein given to rats or rabbits immediately after balloon angioplasty attenuated atherosclerosis/restenosis injury, including reduction in both neointima formation and macrophage infiltration. Since M-T7 binds the chemokine monocyte chemoattractant protein-1 (MCP-1), the results fit well with recent reports of reduced pathology in atherosclerosis-prone mice when either MCP-1 or its receptor CCR2 is genetically disrupted (16, 17). Moreover, since M-T7 does not bind rat IFN-γ, the effects in rats cannot be due to scavenging of that cytokine. Indirect arguments suggest that the effects are not mediated by IFN-γ in the rabbit model either (2).

Whether the histological effects of M-T7 in this model are really due to blockade of MCP-1 (or some other chemokine) is still unresolved. The authors show that M-T7 treatment is associated with reduced detection of MCP-1, MIP-1α, and RANTES antigen in the medial layer of the arterial wall after angioplasty injury in the rat model, but they are careful to point out that this effect could occur by several mechanisms, including antigen masking. Moreover, the maximal concentration that one might achieve in vivo from the 0.017 pg/g dose of M-T7, even assuming distribution to be restricted to the intravascular space, is much lower than the reported K_d for chemokine binding. Finally, M-T7 could potentially bind to other extracellular regulators that contain GAG-binding domains, such as basic FGF (A. Lucas, personal communication). Despite these caveats about mechanism, the efficacy of M-T7 is quite striking and no toxicity was noted. Thus, restenosis injury may now be regarded as a disease indication for further clinical development of this compound. It remains to be seen whether M-T7 will provide a useful anti-inflammatory therapy for established inflammation or systemic disease. Timing of administration, bioavailability, pharmacokinetics, and antigenicity will pose bigger problems in those clinical settings than in the present experimental study.

The M-T7 results are similar to this group’s previous reported results for SERP-1, a secreted serine proteinase inhibitor made by myxoma, in the same rat and rabbit models as well as in a rat aortic allograft model of transplant vasculopathy, and in a rabbit model of antigen-induced arthritis (18–20). These results persuaded the authors to launch a new biotechnology company, Viron Therapeutics Inc., London, Ontario, Canada, to explore the therapeutic potential of natural viral immunomodulators (21, 22). Ironically, now more than two decades after the eradication of smallpox, products of related poxviruses hold substantial promise for the treatment of human disease.