Antibody prophylaxis and therapy for flavivirus encephalitis infections.

The outbreak of West Nile (WN) encephalitis in the United States has rekindled interest in developing direct methods for prevention and control of human flaviviral infections. Although equine WN vaccines are currently being developed, a WN vaccine for humans is years away. There is also no specific therapeutic agent for flaviviral infections. The incidence of human WN virus infection is very low, which makes it difficult to target the human populations in need of vaccination and to assess the vaccine's economic feasibility. It has been shown, however, that prophylactic application of antiflaviviral antibody can protect mice from subsequent virus challenge. This model of antibody prophylaxis using murine monoclonal antibodies (MAbs) has been used to determine the timing of antibody application and specificity of applied antibody necessary for successful prophylaxis. The major flaviviral antigen is the envelope (E) glycoprotein that binds cellular receptors, mediates cell membrane fusion, and contains an array of epitopes that elicit virus-neutralizing and nonneutralizing antibodies. The protective efficacy of an E-glycoprotein-specific MAb is directly related to its ability to neutralize virus infectivity. The window for successful application of prophylactic antibody to prevent flaviviral encephalitis closes at about 4 to 6 days postinfection concomitant with viral invasion of the brain. Using murine MAbs to modify human disease results in a human antimouse antibody (HAMA) response that eventually limits the effectiveness of subsequent murine antibody applications. To reduce the HAMA response and make these MAbs more generally useful for humans, murine MAbs can be "humanized" or human MAbs with analogous reactivities can be developed. Antiflaviviral human or humanized MAbs might be practical and cost-effective reagents for preventing or modifying flaviviral diseases.

Murine T-helper cell immune response to recombinant vaccinia-Venezuelan equine encephalitis virus.

The T-helper (Th) cell immune response following immunization of C3H (H-2k) mice with a recombinant vaccinia (VAC) virus (TC-5A) expressing the structural proteins (capsid, E1 and E2) of the attenuated vaccine strain (TC-83) of Venezuelan equine encephalitis (VEE) virus was compared with the immune response induced in mice after immunization with TC-83 virus. TC-5A virus elicited Th cells that strongly recognized both VAC and TC-83 viruses in in vitro lymphoblastogenesis tests. Th-cell activation was associated with elevated levels of interleukin-2. TC-5A virus induced long-term humoral immunity; VEE virus-binding and neutralizing antibodies were detected in mouse sera collected from mice 16 months after a single immunization.

Analysis of a yellow fever virus isolated from a fatal case of vaccine-associated human encephalitis.

The virulence of a yellow fever (YF) virus (P-16065) isolated from a fatal case of vaccine-associated viral encephalitis was investigated. P-16065 appeared identical to its parent vaccine virus (17D-204 USA, lot 6145) when examined with monoclonal antibodies except that YF wild type-specific MAb S24 recognized P-16065 but not 17D-204 USA 6145. Thus, a mutation of at least one epitope on the envelope (E) protein had occurred. Unlike 17D-204 USA 6145 and other 17D vaccine viruses, P-16065 was neuroinvasive and virulent for mice after intranasal inoculation, and neurovirulent for monkeys after intracerebral inoculation. The E protein of P-16065 differed from 17D-204 USA by two amino acids at positions 155 and 303. Changes at amino acid position 155 are found in other YF vaccine viruses that are not neurovirulent, and it is therefore postulated that the change at position 303 is involved in the alteration of the phenotype of P-16065 and may be important for virulence of YF virus.

T-helper cell epitopes on the E-glycoprotein of dengue 2 Jamaica virus.

To identify T-helper (Th)-cell epitopes, we analyzed 25 synthetic peptides, which included most of the 495-amino-acid sequence of the envelope (E)-glycoprotein of dengue 2 virus. The peptides were analyzed in three mouse strains, BALB/c (H-2d), C57BL/6 (H-2b), and outbred NIH-Swiss, for their ability to elicit antibody or prime the Th-cell compartment following two inoculations in Freund's incomplete adjuvant. Sixteen peptides were able to elicit an antipeptide antibody response in one or more mouse strain. Eleven antipeptide serum pools were able to bind to virus in ELISA. Fifteen peptides primed one or more haplotype for an in vitro antipeptide Th-cell response as measured by blastogenesis. Th-cell activation was generally confirmed by measurable in vitro production of interleukin (IL)-2/IL-4. Nine peptides that were positive for in vitro blastogenesis, 1-2, 35, 4-6, 79, 142, 208, 06, 16, and 17, elicited virus-reactive Th-cells in vitro in H-2d mice. Two of these peptides (4-6 and 17) were able to prime virus-reactive Th-cells in H-2b mice. Nine peptides primed outbred mice in vitro for an antiviral antibody response significantly greater than that seen in animals primed with an irrelevant peptide. These results correlate with, and expand on, our previous observations based on a smaller set of synthetic peptides derived from the E-glycoprotein of Murray Valley encephalitis virus and suggest that synthetic peptides can function as E-glycoprotein Th-cell epitopes. The similarity of results between two distantly related flaviviruses suggests that E-glycoprotein Th-cell epitopes are consistent in location and activity.

Molecular and biological characterization of a non-glycosylated isolate of St Louis encephalitis virus.

The glycosylation patterns of the envelope (E) glycoprotein of several naturally occurring strains of St Louis encephalitis (SLE) virus were investigated. SLE viruses were found that contained both glycosylated and non-glycosylated E proteins, and one isolate (Tr 9464) that lacks N-linked glycosylation sites on its E protein was identified. SLE virus monoclonal antibodies that define E protein B cell epitopes and demonstrate biological activities reacted essentially to the same extent with glycosylated and non-glycosylated virions. These results indicate that glycosylation is not essential for epitope conformation or recognition. However, failure to glycosylate the E protein was associated with possible morphogenetic differences as manifested by reduced virus yields and differences in specific infectivity.

A synthetic peptide to the E glycoprotein of Murray Valley encephalitis virus defines multiple virus-reactive T- and B-cell epitopes.

Synthetic peptides from the envelope glycoprotein sequence of Murray Valley encephalitis (MVE) virus were previously evaluated in various strains of mice for both the induction of antibody and the in vitro proliferation of peptide-primed T-helper (Th) cells. MVE peptide 6 (amino acids 230 to 251) elicited reciprocal Th- and B-cell reactivity with native MVE virus after primary inoculation of C57BL/6 mice. In this study, we prepared overlapping subunit peptides of MVE peptide 6 and evaluated their immunogenicity. Analysis of these peptides delineated at least two B-cell epitopes that induced antibody reactive with MVE and other Japanese encephalitis serocomplex viruses. This antibody at low titer neutralized MVE virus. Genetic restriction of the antibody response to various T-cell elements within peptide 6 was observed in C3H, BALB/c, C57BL/6, and B10 congenic mice. One element demonstrable after primary immunization, located in the carboxy terminus, associated only with major histocompatibility complex class II IAb and IAbiEk glycoproteins. Functional stimulation with the peptides in association with IAkIEk and IAdIEd molecules was observed only after in vivo secondary stimulation. Peptide 6-1 (amino acids 230 to 241) was nonimmunogenic but could be recognized by Th cells from peptide 6-immunized mice. Further association of peptide 6 with the IAkIEk and IAdIEd subregions was demonstrated by the finding that T cells from MVE peptide 6-inoculated C3H and BALB/c mice primed for an antibody response to MVE virus. These results suggest that the peptide 6 sequence, which is relatively conserved among a number of flaviviruses, should be given consideration when synthetic immunogens for vaccine purposes are designed.

Enhancement of the antibody response to flavivirus B-cell epitopes by using homologous or heterologous T-cell epitopes.

We have been investigating the T-helper (Th)-cell response to the flavivirus envelope (E) glycoprotein. In our studies with Murray Valley encephalitis (MVE) virus, we previously identified synthetic peptides capable of priming Th lymphocytes for an in vitro antivirus proliferative response (J. H. Mathews, J. E. Allan, J. T. Roehrig, J. R. Brubaker, and A. R. Hunt, J. Virol. 65:5141-5148, 1991). We have now characterized in vivo Th-cell priming activity of one of these peptides (MVE 17, amino acids 356 to 376) and an analogous peptide derived from the E-glycoprotein sequence of the dengue (DEN) 2, Jamaica strain (DEN 17, amino acids 352 to 368). This DEN peptide also primed the Th-cell compartment in BALB/c mice, as measured by in vitro proliferation and interleukin production. The failure of some MVE and DEN virus synthetic peptides to elicit an antibody response in BALB/c mice could be overcome if a Th-cell epitope-containing peptide was included in the immunization mixture. A more detailed analysis of the structural interactions between Th-cell and B-cell epitope donor peptides revealed that the peptides must be linked to observe the enhanced antibody response. Blockage or deletion of the free cysteine residue on either peptide abrogated the antibody response. The most efficient T-B-cell epitope interaction occurred when the peptides were colinearly synthesized. These Th-cell-stimulating peptides were also functional with the heterologous B-cell epitope-containing peptides. The Th-cell epitope on DEN 17 was more potent than the Th-cell epitope on MVE 17.

T-helper cell and associated antibody response to synthetic peptides of the E glycoprotein of Murray Valley encephalitis virus.

A battery of 16 synthetic peptides, selected primarily by computer analysis for predicted B- and T-cell epitopes, was prepared from the deduced amino acid sequence of the envelope (E) glycoprotein of Murray Valley encephalitis (MVE) virus. We examined all of the peptides for T-helper (Th)-cell recognition and antibody induction in three strains of mice: C57BL/6, BALB/c, and C3H. Lymphoproliferative and interleukin-2 assays were performed on splenic T cells from mice inoculated with peptides in Freund's incomplete adjuvant or with MVE virus. Several peptides found to contain predicted T-cell epitopes elicited a Th-cell response in at least one strain of mice, usually with a concomitant antibody response. Peptides 145 (amino acids 145 to 169) and 17 (amino acids 356 to 376) were strongly recognized by T cells from all three inbred strains of mice. Peptide 06 (amino acids 230 to 251) primed C57BL/6 mice for Th- and B-cell reactivity with native MVE virus, and T cells from virus-immune mice were stimulated by this peptide. Peptide 06 was recognized by several Th-cell clones prepared from mice immunized with MVE, West Nile, or Kunjin virus. These results indicate that it may be feasible to design synthetic flavivirus peptides that define T-cell epitopes capable of generating a helper cell response for B-cell epitopes involved in protective immunity.

Identification of monoclonal antibodies that distinguish between 17D-204 and other strains of yellow fever virus.

Eight monoclonal antibodies (MAbs) prepared against the flaviviruses Saint Louis encephalitis, dengue 2 and dengue 3 viruses all recognized epitopes on the envelope protein of the prototype flavivirus, yellow fever (YF) virus. Three of these MAbs with flavivirus group-common specificity and two MAbs with a flavivirus-subgroup specificity were found to distinguish wild-type YF viruses from YF 17D-204 vaccine virus, but not from the closely related 17DD vaccine virus, nor from the French neurotropic vaccine virus. This pattern of reactivity was seen only with viruses grown in Aedes albopictus C6/36 cells and not with viruses grown in vertebrate cells (SW13 and Vero cells), where all five MAbs recognized epitopes on both wild-type and 17D-204 viruses. Examination of adult A. aegypti mosquitoes infected with the same YF viruses as above gave a different pattern of results to those in C6/36 cells. Thus, epitope expression differs between mammalian and arthropod cells and between arthropod cells in vitro and in vivo. Neutralization tests showed that all five MAbs would neutralize wild-type Asibi virus grown in SW13 cells, but not Asibi virus grown in C6/36 cells, nor 17D-204 vaccine virus grown in either cell type. Therefore, it is concluded that when YF virus is grown in mosquito cells, wild-type virus is antigenically and biologically distinct from the 17D-204 vaccine virus.

Specificity of the murine T helper cell immune response to various alphaviruses.

We investigated the specificity of the T helper (Th) cell immune response to three alphaviruses: Venezuelan equine encephalomyelitis (VEE), eastern equine encephalitis (EEE) and western equine encephalitis (WEE). Single cell suspensions were prepared from spleens of virus-primed C3H mice, and T lymphocyte populations were enriched by nylon wool chromatography. T cells were incubated in vitro with irradiated, syngeneic splenic stimulator cells previously exposed to purified virus. Cellular proliferation was measured by [3H]thymidine uptake 5 days post-stimulation. The predominant proliferating cell type secreted interleukin-2 and was of the Th cell phenotype Thy-1+, Lyt-1+,2-, L3T4+. Stimulation of VEE, EEE and WEE virus-primed Th cells with homologous and heterologous virus resulted primarily in a proliferative response specific for the immunizing virus. The corresponding antibody response, as measured by ELISA using purified virus as antigen, was also specific for the immunizing virus. The magnitude of the blastogenic response of VEE TC-83 virus-primed lymphocytes to a battery of VEE subtype viruses was remarkably similar to schemes of antigenic classification. The results indicate that the dominant Th cell epitopes on these alphaviruses represent regions largely virus-specific and lead to a virus-specific B cell response which does not change over time after primary inoculations of mice with VEE and WEE viruses and multiple inoculations of mice with EEE virus.

Recombinant vaccinia virus/Venezuelan equine encephalitis (VEE) virus protects mice from peripheral VEE virus challenge.

Mice immunized with recombinant vaccinia virus (VACC) expressing Venezuelan equine encephalitis (VEE) virus capsid protein and glycoproteins E1 and E2 or with attenuated VEE TC-83 virus vaccine developed VEE-specific neutralizing antibody and survived intraperitoneal challenge with virulent VEE virus strains including Trinidad donkey (subtype 1AB), P676 (subtype 1C), 3880 (subtype 1D), and Everglades (subtype 2). However, unlike immunization with TC-83 virus, immunization with the recombinant VACC/VEE virus did not protect mice from intranasal challenge with VEE Trinidad donkey virus. These results suggest that recombinant VACC/VEE virus is a vaccine candidate for equines and humans at risk of mosquito-transmitted VEE disease but not for laboratory workers at risk of accidental exposure to aerosol infection with VEE virus.

Recombinant vaccinia/Venezuelan equine encephalitis (VEE) virus expresses VEE structural proteins.

cDNA molecules encoding the structural proteins of the virulent Trinidad donkey and the TC-83 vaccine strains of Venezuelan equine encephalitis (VEE) virus were inserted under control of the vaccinia virus 7.5K promoter into the thymidine kinase gene of vaccinia virus. Synthesis of the capsid protein and glycoproteins E2 and E1 of VEE virus was demonstrated by immunoblotting of lysates of CV-1 cells infected with recombinant vaccinia/VEE viruses. VEE glycoproteins were detected in recombinant virus-infected cells by fluorescent antibody (FA) analysis performed with a panel of VEE-specific monoclonal antibodies. Seven E2-specific epitopes and two of four E1-specific epitopes were demonstrated by FA.

In vitro mechanisms of monoclonal antibody neutralization of alphaviruses.

We have previously identified at least eight epitopes on the E2 glycoprotein of Venezuelan equine encephalomyelitis (VEE) virus vaccine strain TC-83 by using monoclonal antibodies (MAbs). Several of these antibodies identified a critical neutralization (N) domain in competitive binding assays. Passive transfer of these MAbs protected animals from a lethal virus challenge. Using radioactive, purified virus as a marker, we have demonstrated that antibody-mediated virus N, preattachment, can be effected by one of three mechanisms. Interaction of antibody can block virus attachment to susceptible Vero or human embryonic lung cells. The MAbs that were most efficient at blocking attachment were those that defined epitopes spatially proximal to the E2c epitope. The E2c MAbs were, however, the most efficient antibodies for neutralizing virus postattachment. Other E2 MAbs were unable to efficiently block virus attachment to cells; however, resulting replication as monitored by plaque assay or intracellular viral RNA synthesis could not be detected. One novel MAb that defined the E2f epitope appeared to enhance virus attachment to Vero cells, but not BHK-21 or LLC-MK2 cells, by stabilizing virus-cell interaction. This antibody did, however, efficiently neutralize virus infectivity. Once virus had attached to cells, the ability of most MAbs to neutralize infectivity was diminished, except for E2c MAbs. On a molar basis antibody Fab fragments were less efficient than intact antibody at blocking virus attachment.

Antigenic variants of Colorado tick fever virus.

Twenty strains of Colorado tick fever (CTF) virus, isolated from ticks, mammals and humans, and two antigenic relatives of CTF virus were compared in cross-neutralization tests. Viruses were tested using single-inoculation sera prepared in hamsters. Antigenic variation, as measured by differences detected in the neutralization test, was noted among the virus isolates identified as strains of CTF virus. The virus strains isolated from humans appeared to vary the most in serological reactions. The two antigenic relatives of CTF virus are clearly distinct from strains of CTF and are different from each other. Antigenic relationships between these two viruses were established using two sets of single-inoculation antisera and both complement fixation and neutralization tests. Six distinct antigenic variants of CTF virus isolated from humans and the virus strain from ticks (75V1906) that showed the least antigenic variation, were tested against 49 coded serum pairs from clinically diagnosed cases of CTF. Significant differences were noted in the number of convalescent-phase sera that reacted with each virus strain and in the number of seroconversions observed with each test virus strain. Convalescent phase sera that reacted with multiple virus strains often varied significantly in antibody titre from one virus strain to another. This indicates that, in some instances, antibody was probably produced in response to infection by different antigenic variants of CTF virus.

Role of complement and the Fc portion of immunoglobulin G in immunity to Venezuelan equine encephalomyelitis virus infection with glycoprotein-specific monoclonal antibodies.

We have previously characterized with monoclonal antibodies (MAbs) seven unique epitopes on the two envelope glycoproteins of Venezuelan equine encephalomyelitis (VEE) virus vaccine strain TC-83. The epitopes important in protection from VEE virus infection were determined in passive antibody transfer studies, with virulent VEE (Trinidad donkey) virus as the challenge virus. Selected high-avidity MAbs to the three major protective epitopes (E2c, E1b, and E1d) were assayed for in vitro complement activity. All three fixed murine complement to high titer. Limited pepsin digestion of the anti-E2c in the presence of cysteine resulted in a rapid decrease and complete loss of complement-fixing ability by 2 h, but the majority of mice, except at the lowest dilution of MAb, were protected until the Fc termini were cleaved at 3 h. Anti-E2c F(ab')2 would neutralize VEE (Trinidad donkey) virus more efficiently than either Fab' or Fab; none of the fragments would fix complement or was effective in passive protection. C5-deficient mice and mice depleted of C3 with cobra venom factor were still protected from VEE (Trinidad donkey) virus challenge after passive transfer of either anti-E2c or anti-E1b MAb. The results show that the anti-E2c MAb mediates neutralization through bivalent binding at a critical site on the virion and that Fc effector functions, other than complement, are necessary for protection. Although the ability of the anti-E2c MAb to fix complement was associated with its ability to protect in vivo, no direct cause-and-effect relationship was found. Since the epitope defined by the anti-E1d antibody is found on the cell membrane, but is not expressed on the infectious virion, protection in mice was most likely mediated at the cellular level, possibly by inhibition of the final stages of virion maturation.

The neutralization site on the E2 glycoprotein of Venezuelan equine encephalomyelitis (TC-83) virus is composed of multiple conformationally stable epitopes.

The neutralization (N) site on the gp56 (E2) surface glycoprotein of the TC-83 vaccine strain of Venezuelan equine encephalomyelitis (VEE) virus has been characterized using monoclonal antibodies. Five new epitopes (E2d-h) were identified three of which could be mapped into the critical N site by using a competitive binding assay (CBA). Antibodies reactive with these three epitopes had either N or N and hemagglutination-inhibition activity. All epitopes contained within this N site elicited monoclonal antibodies that could protect mice from peripheral virus challenge. Antibodies reactive with the N site on other subtypes of VEE virus (IC and II) bound to, but failed to neutralize, TC-83 virus. Epitopes defined by these antibodies could be located outside of the N site on TC-83 virus by CBA. Antigenic activity of all epitopes except E2d was resistant to treatment with 2% SDS, 3% beta-mercaptoethanol, or cleavage with Staphylococcus aureus V8 protease. Those antibodies which defined epitopes located within the N site of TC-83 with CBA bound the same V8 fragments in immunoblots. Those antibodies which defined epitopes not located within the N site bound a different set of fragments than neutralizing antibodies. These results indicate that there is a specific N site on the E2 of VEE virus which undergoes significant antigenic drift while maintaining structural and functional integrity.

Elucidation of the topography and determination of the protective epitopes on the E glycoprotein of Saint Louis encephalitis virus by passive transfer with monoclonal antibodies.

We have identified and characterized eight antigenic epitopes on the 53,000 dalton envelope (E) glycoprotein of Saint Louis encephalitis (SLE) virus by using monoclonal antibodies. One of these epitopes (E-1c) encoded for the type-specific biologic functions of hemagglutination (HA) and neutralization (N). Injection of 50 ng of anti-E-1c antibody protected the majority of mice from peripheral challenge with 100 i.p. LD50 of SLE virus. Similar levels of protection with antibodies specific for other epitopes usually required greater than or equal to 1000-fold additional antibody. Attempts to block N or protection at the E-1c antigenic domain by using antibody to several other SLE epitopes that strongly competed for the E-1c site were unsuccessful. Enhancement of protection was observed with mixtures of the more cross-reactive antibodies. The E-1c antibody was also effective in abrogating SLE virus replication until neural invasion occurred. On the basis of these findings, the topologic arrangement and function of the eight SLE E glycoprotein epitopes on the virion spike is proposed.

Identification of epitopes on the E glycoprotein of Saint Louis encephalitis virus using monoclonal antibodies.

Twenty-one hybridomas producing monoclonal antibodies specific for the E glycoprotein of St. Louis encephalitis (SLE) virus, strain MSI-7, have been isolated. Serologic reactivities were initially determined by cross-reactivity indirect immunofluorescence assays using 22 strains of SLE virus and 8 other related flaviviruses. Four groups demonstrating type-, subcomplex-, supercomplex-, and group-specific reactivity patterns were identified. Analysis of hemagglutination-inhibition (HI) and virus neutralization (N) subdivided the cross-reactivity groups into eight epitopes (E-1a,b,c,d, E-2, E-3, and E-4a,b). The antibodies could detect strain differences between SLE viruses isolated from various geographic areas. Analysis of the spatial arrangements of these epitopes using competitive binding assays with representative antibodies possessing similar binding avidities, indicated that the protein was a continuum of six overlapping domains.

Determination of the protective epitopes on the glycoproteins of Venezuelan equine encephalomyelitis virus by passive transfer of monoclonal antibodies.

We have previously characterized seven unique antigenic epitopes on the two envelope glycoproteins of the Venezuelan equine encephalomyelitis (VEE) virus vaccine strain, TC-83, by using monoclonal antibodies. The in vitro function of virus neutralization was primarily associated with one epitope on the gp56 (gp56c). To determine which epitopes were important in protecting animals from VEE infection, purified monoclonal antibodies were inoculated i.v. into 3-wk-old Swiss mice. Twenty-four hours later these animals were challenged i.p. with 100 IPLD50 of virulent VEE virus (Trinidad donkey). High-avidity anti-gp56c, anti-gp50b, anti-gp50c, and anti-gp50d monoclonal antibodies protected animals from virus challenge. Rabbit antisera to the gp56 and the gp50 glycoproteins were also effective in protecting mice from challenge with virulent VEE virus. Less antibody was needed to protect animals if the antibody was directed against the critical neutralization site. Less avid antibodies to the gp56c and gp50b epitopes demonstrated little or no protection in vivo. Protection, therefore, appeared to be a function of the passive antibody's specificity, avidity, and ability to bind to virion antigenic determinants topologically proximal to the critical neutralization site.