Main

In the best of circumstances, clinical investigation is difficult. To draw definitive conclusions, well-controlled studies must be done in defined populations, but this is secondary to the obligation to provide care that benefits individual patients. In addition to these limitations, access to patients and the collection of specimens during the relevant time in a disease is often problematic. Such has been the case with studies of Ebola virus infection, which have proven particularly difficult because outbreaks occur in remote locations and because this highly infectious, lethal virus poses substantial risks to medical personnel and scientific investigators. Ebola virus infection causes hemorrhagic fever, a syndrome that begins with flu-like symptoms and progresses rapidly to fever, diffuse bleeding, shock and death1,2,3. Rigorous study of this disease in humans has been nearly impossible, and its immunopathogenesis is poorly understood. For example, determinants of host resistance, mechanisms of viral clearance and prognostic indicators of survival have not been defined. Such information could provide important insights into the natural course of this disease in humans that would facilitate the development of vaccines and antiviral therapies. In this issue of Nature Medicine, Baize, Leroy and colleagues have analyzed immune responses in cohorts of patients from two outbreaks in northern Gabon in 1996 ( ref. 4). The mortality rates of the epidemics were similar (66% and 75%), and the investigators collected serum and blood samples during these outbreaks from both survivors and non-survivors. Their analyses of immune response and serum biochemical parameters suggest possible mechanisms of protection against this highly pathogenic virus.

Viral load and immune responses during acute Ebola virus infection.

Patients who survived showed several features that distinguished them from non-survivors. No difference in the viral antigen load, measured in the serum, was detected in survivors and non-survivors, indicating that host defense rather than inoculum size determines survival. Although non-survivors generated a higher level of gamma interferon (IFN-γ) in the serum early in infection, T-cell cytokine responses, measured by T-cell cytokine RNA levels, were not maintained. In contrast, surviving patients developed more effective antibody responses to the viral nucleoprotein and also generated more sustained T-cell immune responses that persisted into the convalescent phase. Lethal infection was also accompanied by a considerable increase in apoptosis in peripheral blood cells, and cell death increased as disease progressed in fatal cases. This increase in cell death markers correlated with mortality.

These observations are among the most comprehensive analyses of the immune response to Ebola virus infection in humans, and they provide several clues about its pathogenesis. Several similarities between human disease and relevant animal models were found. For example, in guinea pigs immunized with the Ebola virus gene products, successful protection against lethal challenge involves T cell-dependent immune responses5. As with immunization in the guinea pig, the nucleoprotein response in humans is mainly humoral, and immunoglobulin production to the viral proteins GP and soluble GP is minimal, although T-cell immunity to these antigens is generated. The findings by Baize et al. in infected humans4 are thus consistent with those in the vaccine model in rodents5. The generation of antibody responses correlated with survival in the guinea pig and probably reflected successful T-cell responses that stimulate production of antibody5, as is suggested now for humans4. It is not certain whether the ability to counteract early infection is due to this T-cell immune response in natural infection, and it is possible that an inflammatory response or innate immunity is involved. In addition to these immune parameters, the authors also analyzed levels of apoptosis in peripheral blood cells. These markers reflect the death of immune and non-immune cells. Although the death of immune cells probably contributes to the demise of patients, the significance of this finding is not entirely evident. Apoptosis often accompanies antigen stimulation, particularly when proliferating lymphocytes are deprived of growth factors. As noted by the investigators, these markers of apoptosis could be derived from inflammatory cells, such as monocytes, or endothelial cells. Both cell types are productively infected in primate models of this disease6,7,8. Despite these valuable observations regarding the immune response and its relationship to the progression of the disease, many questions remain. Enhanced apoptosis of lymphocytes has not been demonstrated directly. If this occurs, it will be important to define which subsets are affected and by what mechanism, as the virus does not efficiently infect lymphoid cells9,10. The mechanism by which innate immune or inflammatory responses control viral replication in early stages of this disease requires further definition, and the point at which acquired immunity becomes effective has not yet been determined.

The remarkable inverse correlation between the serum IFN-γ levels and mortality indicates that this marker may be useful in predicting the outcome of infection. Semi-quantitative PCR analysis may also be helpful. In the absence of effective therapy, the utility of this information is limited; however, antiviral compounds are being investigated11 and may prove useful in the future. In the interim, Baize, Leroy et al. have provided insight into the natural history of the immune response to infection by this highly pathogenic virus. Successful immune protection and vaccination against viral infections other than Ebola have been demonstrated in humans. Despite these successes, the specific contribution of innate and acquired immune responses to viral infections in general remain poorly understood12,13. Ebola virus infection can be regarded as a paradigm for other viruses and for other infectious causes of hemorrhagic fever. Such information will help to unravel the pathophysiology of this aggressive infection and will contribute to strategies for vaccine development and treatment.