Human Immunoglobulin as a Treatment for West Nile Virus Infection
- 1 Critical Care Medicine Department, Clinical Center, National Institutes of Health, Bethesda, Maryland
- 2 Division of Vector-Borne Infectious Diseases, Center for Infectious Diseases, Centers for Disease Control and Prevention, Fort Collins, Colorado
- Reprints or correspondence: Dr. Lyle R. Petersen, Div. of Vector-Borne Infectious Diseases, Centers for Disease Control and Prevention, PO Box 2087 (Foothills Campus), Fort Collins, CO 80522 (lxp2{at}cdc.gov).
During the West Nile virus outbreak in Israel in 2000, a woman with chronic lymphocytic leukemia who was comatose as a result of West Nile virus encephalitis recovered after treatment with intravenous immunoglobulin (IVIG) [1]. Antibody titers against West Nile virus were 1:1600 in Israeli IVIG; in contrast, North American IVIG preparations had no detectable West Nile virus antibody [1, 2]. A second patient, a lung transplant recipient, also recovered from West Nile virus encephalitis after treatment with Israeli IVIG [3]. Six other subsequently treated patients have had variable outcomes: 2 improved, 2 had no improvement, and 2 eventually died [4] (C. Isada, oral communication; R. Babecoff, written communication). These anecdotal reports, although inconclusive, have stimulated interest in the use of passive immunization for treating severe West Nile virus disease.
The article on the efficacy of human immunoglobulin in treating West Nile virus infection in mice by Ben-Nathan et al. [5] is provocative, in that it suggests that IVIG might ameliorate or abort established West Nile virus infection. In the series of experiments described, BALB/c mice were infected intraperitoneally with either 20 or 200 times the LD50 (100 or 1000 pfu). They received 1 of 6 treatments: nonimmune mouse serum, immune mouse serum with ELISA titers against West Nile virus of 1:3200, Omrix IVIG with ELISA titers of 1:1600, IVIG from US donors with ELISA titers of 1: 10, pooled plasma from Israeli donors (titers not described but presumed to be less than concentrated immunoglobulin), or pooled plasma from US donors. In protection experiments in which the antibody-containing treatments were given shortly before and after infection, the results were unequivocal—any treatment containing specific antibody produced 100% survival, and treatments without specific anti-West Nile virus antibody provided no protection (100% mortality; table 1).
Table 1.
Summary of published animal studies that used immunoglobulin for protection against or treatment of experimental flaviviral infection.
In subsequent experiments, the infected mice were treated with 1–5 injections of specific anti-West Nile Virus antibody, with a clear dose- and time-dependent relationship to survival (table 1). Importantly, all treatments were started during the viremic phase, before virus entered the brain. It was established that virus entry into the brain occurred on day 3 after injection with the 1000-pfu inoculum and on days 6–8 after injection with the 100-pfu inoculum in untreated West Nile virus-infected control mice. Finally, high-dose antibody treatment administered 1 and 2 days or 2 and 3 days after injection with the 100-pfu inoculum afforded complete protection; however, antibody administered 3 and 4 days after infection yielded only 50% survival, although those mice that died survived longer than controls.
Evidence suggests that West Nile virus may be more susceptible to antibody-mediated immunity than to cell-mediated immunity. Clearance of primarily neurotropic viruses is not dependent on cytolytic T cell activity, in contrast to nonneurotropic viruses [11, 12]. Neurons, as terminally differentiated cells, do not express major histocompatibility complex class I, which would subject them to lysis by CD8 T cells and nonreplacement [12]. Animal data support this concept. In the 1970s, Camenga et al. [6] challenged mice with West Nile virus, gave them cyclophosphamide 1 day later to suppress both the humoral and T cell-mediated arms of immunity, and then administered immune serum or immune syngeneic spleen cells at varying times (table 1). One injection of immune serum at day 5 or 6 resulted in 82% survival, and 1 injection at day 8 or 10 resulted in 22% survival, compared with 3% survival in controls. Notably, virus was detectable in the brain at day 6. Immune syngeneic spleen cells rescued 87% of mice at day 2 but only 13% at day 4; survival among controls was 12% [6].
Diamond et al. [7] have described the effects of immunoglobulin in several different mouse models of West Nile virus infection. In a B cell knockout model, a single plaque-forming unit of West Nile virus given by footpad inoculation constitutes the LD50. All mice given specific immunoglobulin 1 day before and 1 day after infectious challenge with a lethal inoculum of West Nile virus (100 pfu) survived, compared with none of the controls (table 1). In wild-type mice, immune serum produced results similar to those reported by Ben-Nathan et al. [5]; after inoculation with 100 pfu of West Nile virus, immune serum produced 80% survival if administered at 24 h and 50% survival if administered at 4 days, versus 20% survival in control animals (M. S. Diamond, unpublished data).
Xiao et al. [13] established a model of West Nile virus encephalitis in Syrian golden hamsters. After intraperitoneal inoculation, moderate viremia was detected for the first 5 days after infection. Natural antibody levels increased by day 5, and viremia became undetectable by day 7. By day 5, neuronal degeneration was evident, and by days 7–8, degeneration had localized to the deeper layers of the cerebral cortex, Purkinje cells in the cerebellum, and the brain stem. Death occurred in ∼50%–70% of the infected animals between 7 and 14 days after infection. In another study using the same model, hamsters (>8 weeks old) treated with exogenous immune serum 24 h before infection were completely protected from viremia and death (table 1) [8]. However, if antibody administration was delayed 48 h after infection, no survival effect was seen (R. B. Tesh, unpublished data).
Finally, in recent experiments, golden hamsters and BALB/c mice were given Israeli IVIG, IVIG from US donors, or saline on day 1 and 2 after intraperitoneal injection with West Nile virus. Older hamsters (>8 weeks old) showed no survival benefit with IVIG, but younger hamsters (4–5 weeks old) had 100% survival when Israeli IVIG was used and 20% survival when US IVIG was used. Among BALB/ c mice, 90% of those given Israeli IVIG survived, compared with 10% of those given US IVIG (J. D. Morrey, unpublished data).
The results of Ben-Nathan et al. [5] and other groups showing clear-cut protection by IVIG when animals are treated before or shortly after infectious challenge with West Nile virus provide support for the potential usefulness of IVIG in the treatment of acute West Nile virus infection in humans. Passive immunization of humans for treatment of other viral infections has precedent. Examples include parvovirus in patients with human immunodeficiency virus [14]; chronic echovirus meningoencephalitis in children [15–17]; disseminated vaccinia infection [18]; Junin virus, the agent of Argentine hemorrhagic fever [19]; and cytomegalovirus pneumonitis in bone marrow transplant recipients [20].
A critical question is whether IVIG is effective when cerebral infection has been established. Successful treatment of animals challenged with St. Louis encephalitis and tickborne encephalitis viruses, flaviviruses related to West Nile virus, was possible several days after infection; however, the therapeutic effect decreased rapidly once virus was found in brain (table 1) [9, 10]. Other neurotropic viruses are amenable to treatment with antibody [21–23]. Studies by Griffin and coworkers [12, 21] indicated an important role for antibody in the clearance of neuroadapted Sindbis virus after well-established brain infection. Virgin and Tyler and their associates [22, 23] successfully treated reovirus brain infection with intraperitoneally administered antibody up to 7 days after infectious challenge. In the preantibiotic era, horse serum treatment for meningitis caused by Neisseria meningitidis significantly reduced mortality [24]. The efficacy of antibody therapy for cerebral infection depends on passage of IgG through the blood-brain barrier. Studies in humans without inflamed meninges show that IgG enters at very low levels [25]. Entry is enhanced, however, once inflammation is present. In a study of patients who developed aseptic meningitis as a complication of highdose IVIG therapy for autoimmune disease, IgG levels of 1.5–7 times the upper limit of normal were seen in cerebrospinal fluid [26].
There are several limitations to the study by Ben-Nathan et al. [5]. One concern is that success was shown in animal models when antibody was given during the viremic phase; however, nearly all patients with West Nile virus infection are no longer viremic when they present, and most have already developed IgM antibody [27, 28]. Another limitation was that even the lower infecting dose used by Ben-Nathan et al. [5] produced 100% mortality in untreated mice. But the infectious inoculum from a mosquito bite is far less, and most persons infected remain asymptomatic. It would be of interest to see how long after an infectious inoculum is injected IVIG can produce a survival benefit, particularly when the inoculum is low. Another potential limitation was the choice of infecting strain. Whereas other recent work investigating the efficacy of antibody against West Nile virus in animal models has used contemporary strains from the US outbreak [7, 8], Ben-Nathan et al. [5] used a 1954 strain. Although the establishment of an LD50 still ensures adequate virulence, differences in strain may produce different responses to treatment. In work screening other compounds in vitro, Morrey et al. [29] found an up to 30-fold difference in EC50 (the amount of a compound required to abrogate 50% of the cytopathic effect produced by a virus) for some compounds when they used a recent New York strain, compared with the 1937 Uganda strain of West Nile virus. Finally, the efficacy of antibody therapy demonstrated in animal models may not translate directly to humans. As an example from the preantibiotic era, the mouse model of pneumococcal infection showed that treatment with horse serum was only effective when given within 12 h, but when this therapy was transferred to humans for this otherwise untreatable disease, it was very effective when given within 4 days [24].
Current candidate agents for treatment of severe West Nile virus disease include ribavirin, interferon (IFN)-α2b, and hightiter anti-West Nile virus immunoglobulin. Ribavirin has some activity at very high doses in vitro [29, 30]. Patients who received ribavirin during the 2000 West Nile virus outbreak in Israel had increased mortality; however, this was likely due to the fact that ribavirin was administered to sicker patients [31]. If the high levels required for any antiviral activity in vitro are extrapolated to levels achievable in human cerebrospinal fluid, extraordinarily high intravenous doses (on the order of 4 g/day) would be required achieve levels between the in vitro EC50 and the EC90 for oligodendroglial cells [30]. Such doses are associated with significant, although reversible, hemolytic anemia.
IFN-α2b is active in vitro against West Nile virus [32] but has not been tested in an animal model of West Nile virus. Its in vitro activity and broad immunostimulatory activity [33] have prompted several clinical trials involving flaviviral diseases. In an outbreak of St. Louis encephalitis in Louisiana, IFN-α2b was used expediently in a nonrandomized open-label fashion. The first 17 patients in the outbreak were treated with supportive care only; the next 15 patients were treated with IFN-α2b [34]. Neither group experienced mortality. Although the mean neurological score improved in the treated group and worsened in the untreated group by week 3, these results must be interpreted with considerable caution, because the study was not randomized, there was no placebo group, and the scoring scale was applied in an unblinded fashion [34]. An open-label study examining the effect of IFN-α2b versus supportive care for West Nile virus infection is ongoing (J. J. Rahal, written communication). A large, double-blind, placebo-controlled study failed to find a difference in mortality or functional outcome resulting from the use of IFN-α2b for treatment of Japanese encephalitis, a related flavivirus, in Vietnam [35].
Although studies such as that by Ben-Nathan et al. [5] suggest that flaviviral infections are at least partially treatable with passive immunity, if antibody is going to work at all, it must be administered early. The side-effect profile of IVIG compares favorably with that of the other 2 main candidate therapies (ribavirin and IFN), and, without a human trial, there will be no way to know whether the results in animals will translate to a significant benefit in humans. The high mortality seen among persons with West Nile virus encephalitis and the long-term morbidity among survivors [36, 37] provide impetus for such a trial. Patients with established neurological disease or who are at high risk for progression to severe disease but who have not yet developed encephalomyelitis would be candidates for any effective therapy. Risk factors for progression of disease are advanced age or immunosuppression, although the type of immunosuppression is not yet well defined [31, 36, 38]. If IVIG or any agent is found to be effective, rapid diagnosis will become far more important to identify potentially treatable patients early in the course of disease. The unpredictability of flaviviral outbreaks other than Japanese encephalitis complicates planning of human clinical trials; if the high incidence of severe West Nile virus illness observed in 2002 in North America is repeated in coming years, there may exist a unique opportunity to determine whether early treatment of West Nile virus infection with specific antibody is beneficial.
- Received May 12, 2003.
- Accepted May 14, 2003.
References
This Article
- ExtractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 6.288
5-Yr impact factor: 6.103
For The Media
For Authors
- Services for authors
- Self-archiving policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
