Smallpox: An ancient disease enters the modern era of virogenomics

When the World Health Organization certified in 1980 that the world was finally free of smallpox as an extant human disease, few suspected that the viral agent that caused it (variola virus) would be a subject of intense debate a quarter of a century later (1). Today, with the threat of bioterrorism, variola virus is once again subject to investigative scrutiny, now with particular focus on antiviral drug discovery and the development of new generations of safer vaccines (2, 3). However, serious researchers have had to come to grips with the fact that almost all our knowledge of variola virus and of the pathogenesis of the disease it once so routinely caused in humans is derived from an era before many of the recent advances of molecular virology and immunology (4). For example, we still do not understand precisely the pathological dysfunction that smallpox patients actually died of, nor do we have exacting surrogate animal model systems with which to address such fundamental issues as the mechanisms of virus pathogenesis and the basis for the immune responses to either the disease or its vaccine. In this issue of PNAS appear two recent entries into this new arena (5, 6). Both papers provide provocative insights into these issues, as well as serious food for thought about the implications and complications of trying to exploit variola virus as a tool for medical research.

In the first of the two companion papers, Jahrling et al. (5) investigated the pathogenesis of variola virus in cynomolgus macaques to ascertain whether a nonhuman primate model could recapitulate at least some of the pathologic features of human smallpox. Many closely related orthopoxviruses have been investigated over the years as surrogate models with which to dissect the pathogenesis of smallpox, for example monkeypox in monkeys, ectromelia virus in mice, and rabbitpox in rabbits. However, these models do not precisely recapitulate the pathogenic spectrum and kinetics of true smallpox, and none of these models is currently recognized by the Food and Drug Administration as fulfilling the animal efficacy requirement for the licensing of any new-generation antiviral drugs against smallpox. Despite many advances in the area of molecular poxvirology, it still has not been worked out why individual poxviruses frequently have a predilection to cause disease in only a certain narrow range of host species. Nor is it understood why poxviruses occasionally leap species to cause zoonotic infections in nonevolutionary hosts. However, the subject of poxvirus tropism and host range determinants in general is a particularly fertile ground for future research.

In the case of smallpox, one of the major reasons the World Health Organization worldwide vaccination campaign was so successful relates to the fact that variola virus has no known animal reservoir and successfully transmits only among humans (1, 7, 8). Despite this, some attempts to induce clinical disease in monkey models with variola virus have been made in the past, but such efforts never managed to recapitulate the fulminant disease course of smallpox. Jahrling et al. (5) investigated this issue by examining several routes of exposure of cynomolgus monkeys to variola virus and confirmed that monkeys are relatively refractory to infection after moderate inoculation dosages. In fact, by the standard respiratory inoculation route that is so infectious for humans, it was not possible to establish clinical disease in monkeys, suggesting that some aspect of primary viral replication in pulmonary tissues or infected cell trafficking into the draining lymph nodes is suboptimal. However, when variola virus was injected at high doses (10⁹ infectious units per animal) by the i.v. route, the direct access to the blood-stream appeared to mimic the consequences of secondary viremia, and this protocol was able to induce the classical centrifugal lesion progression plus the coagulopathy and toxemia reminiscent of full-blown smallpox. The exact details of the tissue dysfunctions, coagulation abnormalities, and induced cytokine storm observed in the model will require further analysis, but this represents the closest surrogate described to date for end-stage hemorrhagic smallpox. Although this model may accurately mirror only the secondary viremic phase of smallpox in humans, it does at least allow the potential for fulfilling one of the key Food and Drug Administration criteria for new antiviral drug development and licensure.

Rubins et al. (6) have exploited this monkey/variola virus model to ask more fundamental questions relating to pathogenic mechanisms that could have been addressed only in the modern era. Virogenomics exploits the ever-expanding “-omics” technologies, particularly genomics, proteomics, and bioinformatics, to investigate the complex interplay between a viral pathogen and the infected host (9–12). In the current work, the authors isolated peripheral blood samples during disease progression in variola-infected monkeys and subjected the extracted mRNA to exhaustive microarray analysis by using chips that can query 18,000 human cDNA sequences and can detect the expression of a large majority of the orthologous cynomolgus genes. In the blood samples themselves, there was a mixture of variola-infected and uninfected leukocytes, but the virus replication was mostly confined to monocyte/macrophages, whereas the lymphocytes remained largely uninfected, although significant lymphopenia was detected. In terms of the innate immune responses, perhaps the most striking result was the consistent early up-regulation of IFN responsive genes, coupled with a noted absence of the expected proinflammatory responses mediated by NF-κB activation and tumor necrosis factor. The absence of these latter inflammatory markers was not a simple case of an inability to detect the mRNA for human-related monkey genes, because cynomolgus monkeys infected with another acute viral pathogen, Ebola virus, clearly caused their robust activation (6).

At this point, the leading hypothesis to explain the marked host response differences after infection with Ebola and variola virus resides in the large repertoire of immunomodulatory proteins that are encoded by poxviruses (13). Variola encodes several putative inhibitors of IFN and tumor necrosis factor, as well as many others (14–16), and so at least some of these results can be rationalized (Table 1).

Table 1. Portrait of a serial killer: Immunomodulator genes of smallpox virus.

Immunomodulator	Variola major*	Variola minor†	Comments‡
IFN-α/β inhibitor	B20R	D9R	Untested
IFN-γ inhibitor	B9R	H9R	Inhibits human IFNγ (24)
eIF2α mimic	C3L	D3L	Untested
dsRNA-BP	E3L	C3L	Untested
Tumor necrosis factor inhibitor	G2R	G2R	Untested
NF-κB inhibitor	P1L	R1L	Untested
TLR signaling inhibitor	A52R	A56R	Untested
IL-18-binding protein	D5L	B6L	Inhibits human IL-18 (25)
Chemokine-binding proteins	G3R, A46L	G3R, A49L	G3R inhibits human CC-chemokines (26)
Viral growth factor	D2R	B2R	Binds and activates human EGF-R (erbB1) (27)
Complement inhibitor/SPICE	D12L	B18L	Inhibitor of human complement cascade (28)
GM-CSF/IL-2 inhibitor	A46R	A49L	Untested
Semaphorin	A42R	A45R	Untested
Serpins	B13R, B25R, C2L	D14R, D2R	Untested

GM-CSF, granulocyte—macrophage colony stimulating factor; dsRNA, double-stranded RNA; TLR, Toll-like receptor; SPICE, smallpox inhibitor of complement enzymes; EGF-R, epidermal growth factor receptor.

^*Strain India, 1967.

^†Strain Garcia, 1966 (alastrim).

^‡Untested variola immunomodulators are presumed to share some of the properties of orthologous versions from other poxviruses.

However, it should be noted that the less virulent variola minor virus (alastrim) also encodes very closely related inhibitors directed against these kingpins of the innate immune responses, so it is clear that our understanding of how this virus actually subverts the host responses and causes clinical disease remains rudimentary. For other poxviruses, the use of directed gene knockouts in specific viral immunomodulatory genes has provided a wealth of information about poxvirus pathogenesis (17), but targeted recombinant manipulations are currently proscribed for variola virus. Monkeypox virus, on the other hand, at least offers a more tractable model with which to address this issue in the monkey model (18).

The results from these two papers (5, 6) also have implications that extend above and beyond the study of a human disease that currently exists only in history books. Many poxviruses can zoonotically infect humans (19), and one in particular, monkeypox, induces a disease that is clinically almost indistinguishable from smallpox (20, 21). What differentiates cases of human monkeypox from true smallpox is the relative inefficiency of the former to serially propagate in successive human-to-human transmissions. We know almost nothing about the basis for this apparent deficiency of human transmission by monkeypox, and our ignorance is certainly not comforting. In 2003, an imported monkeypox-infected rodent, possibly a Gambian rat, managed to play the role of bioterrorist in the Midwest of the U.S. and initiated a miniepidemic of human monkeypox that was quickly contained, but by virtue more of good luck than of public health responses (22, 23). In this case, the isolate of monkeypox virus introduced by this one terrorist rat was likely a naturally attenuated virus clone from West Africa that is less pathogenic in humans than versions from central Africa, but next time we may not be so lucky.

It is imperative that we better understand the parameters that distinguish pathogenic poxviruses from their benign cousins and prepare for the possibility that a virulent transmissible poxvirus pathogen might someday again appear in the human population. From the standpoint of preparedness, it matters little whether the next event derives from smallpox release from a covert source, an engineered poxvirus, or simply a zoonotic infection by a poxvirus that self-educates how to adapt to the human immune system. Although the immediate agenda in current counter-strategies against pathogenic poxviruses is the development of antiviral drugs and safer vaccines, there is still a genuine need to probe the fundamental nature of virus/host interactions with lethal human pathogens like variola and monkeypox. There will continue to be debate about what limits should be placed on working with live variola virus and the kinds of recombinant constructs that should be permitted with poxviruses capable of infecting humans, but there is no arguing that we live in a world where ignorance is more dangerous than knowledge. As we know so well, emerging viruses care little of their origins once they are afoot in a susceptible host population. The task of the medical research community is to anticipate future catastrophic scenarios by continuing to learn from our past adversaries. It remains true that our worst enemies know us the best, and they still have much to teach us about our future enemies as well.