Venezuelan equine encephalitis (VEE) is a re-emerging disease of equids and people with a long history of explosive, mosquito-borne outbreaks as well as endemic disease. Major outbreaks occur when equine-amplification-competent strains emerge from enzootic, sylvatic progenitors that circulate among rodents transmitted by arboreal mosquitoes [1
]. These epizootics/epidemics, typically involving subtype IAB and IC strains of VEE virus (VEEV) occur periodically when enzootic, subtype ID strains mutate to gain the equine-amplification-competent and equine-virulent phenotypes [2
]. These subtypes are primarily based upon similar serological differences but also upon their ability to cause epidemics. The last major outbreak occurred in Venezuela and Colombia in 1995, affecting approximately 100,000 people and large numbers of equids [3
]. However, an estimated 10,000 human cases also occur annually via endemic spillover from the rodent-mosquito cycles that occur in forests, but also periurban areas throughout much of Latin America [4
Unlike many arboviral diseases, VEE typically causes symptomatic infection, ranging from a nonspecific flu-like illness that resolves in about one week to a more severe form seen in 5–15% of cases that involves the central nervous system, especially in children. This severe form is often accompanied by permanent neurologic sequelae or death, and infection also leads to severe immunosuppression. Overall case-fatality rates are typically less than 1%, but teratogenic effects also result in many stillborn deaths during outbreaks [6
]. In addition to being a deadly re-emerging virus, VEEV is also a highly efficient biological weapon capable of disabling large populations due to its high degree of stability when delivered as an aerosol and its high degree of infectivity via the respiratory route. It was therefore highly developed by both the United States and the former Soviet Union during the cold war [7
Despite its importance as a re-emerging and biothreat virus, no licensed products exist to prevent or treat VEEV infection. Although experimental vaccines have been developed for many decades, only three have been used in clinical trials. The attenuated TC-83 strain of VEEV, developed by serial cell culture passage of the epidemic Trinidad donkey (TRD) strain [8
], has been used as an Investigational New Drug (IND) product in humans for several decades and both live forms and inactivated forms have been used for nearly 50 years in humans and equids (licensed for the latter). Although considered highly effective in equids [9
], TC-83 is reactogenic in humans and a single dose often fails to induce seroconversion, requiring a boost with an inactivated vaccine form called C-84 [11
]. Vaccines report a variety of adverse events, some severe, like rapid fever onset, headache, photophobia. TC-83 also has a number of exclusion criteria, including having egg allergies, being pregnant, or having a first relative diagnosed with diabetes. Interestingly, alternations in glucose metabolism have been modeled in TC-83 vaccinated Golden Syrian hamsters [12
] and rhesus macaques [13
]. Due to the high level of severe adverse events, TC-83 is only used for laboratory scientists at high risk of infection. The reactogenicity of TC-83 probably results from the reliance on only two point mutations for attenuation [14
]. The second vaccine to enter clinical trials, called strain V3526, was a rationally attenuated version of the TRD strain with a PE2 viral polyprotein cleavage-signal mutation, combined with a second-site suppressor mutation [15
] (Figure 1
). This vaccine, although more stably attenuated than TC-83 [16
], also proved reactogenic in a Phase 1 clinical trial, with adverse events including headache, fever, malaise, and sore throat [17
More recently, a DNA vaccine that expresses codon-optimized versions of the E3-E2-6K-E1 structural protein genes of VEEV strain TRD was developed. This vaccine induces high levels of neutralizing antibodies in rodents and nonhuman primates and protects against VEEV challenge [18
]. In a Phase 1 trial, 2–3 doses of the DNA administered intramuscularly or intradermally with an electroporation device were well-tolerated, and induced neutralizing antibodies responses in volunteers receiving low or high doses [19
Although the DNA vaccine described above could be useful in some situations, the requirement for multiple doses is not optimal for controlling an explosive arboviral outbreak or for responding to a biothreat situation. We therefore have focused mainly on live-attenuated vaccines for VEE, starting with chimeric alphaviruses [20
], and later progressing to attenuation via alterations to the expression patterns of the alphavirus open reading frames (Figure 1
). These include: (1) inactivation of the subgenomic promoter to eliminate the molar excess of subgenomic RNA that results in high levels of structural protein expression, combined with the introduction of a picornavirus (encephalomyocarditis virus) internal ribosome entry site (IRES) to allow for translation of the structural polyprotein from genomic RNA (Version 1), and; (2) translocation of the capsid protein gene to a separate open reading frame downstream of the envelope glycoprotein genes and behind the IRES to downregulate expression of capsid only (Version 2). This approach was first used with the TC-83 live-attenuated VEE vaccine strain [21
]. The subgenomic RNA was eliminated as expected, and mosquito infectivity was also eliminated because the IRES is not efficiently recognized by insect ribosomes. This Version 2 of VEEV/IRES was slightly more immunogenic [22
Although the original IRES-based attenuation was not successful in terms of immunogenicity using the attenuated TC-83 backbone, the application of this approach to wild-type VEEV strains demonstrated a high degree of attenuation, immunogenicity, and efficacy. Starting with an enzootic subtype IE VEEV strain that represents the cause of extensive endemic human VEE in Mexico and Central America [4
], IRES-based, live-attenuated vaccine candidates proved highly attenuated and immunogenic after a single immunization of mice, and provided full protection against lethal challenge [24
]. Version 1 of this vaccine produced no detectable disease in cynomolgus macaques (Macaca fascicularis
), the preferred nonhuman primate model for VEE, and protected against aerosol challenge including fever and viremia (this nonhuman primate model, like most human infections, is not lethal) [25
Although VEEV subtype IE is an important human and sometimes also an equine pathogen [26
], its southern distribution ends in western Panama, where enzootic subtype ID occurs southward through eastern Panama, Venezuela, Colombia, Ecuador, Peru, and Bolivia [4
]. Furthermore, subtype ID strains are the progenitors of subtype IAB and IC strains, which are responsible for all major epizootics/epidemics ever documented [27
]. There is also significant antigenic divergence between subtypes ID and IE, as reflected in weak cross-neutralizing antibody titers after infection or vaccination [28
]. Furthermore, although the subtype IAB-based V3526 vaccine strain protects mice against aerosol challenge with a subtype IE VEEV strain, it did not significantly limit challenge virus replication [29
Based on the potential limitations of a subtype IE-based vaccine to protect against subtypes IAB, IC, and ID, we developed additional VEE vaccines based on these same IRES-based attenuation approaches but using ID strain ZPC738 [30
], a close relative of subtypes IAB and IC [31
]. Here, we describe the safety, immunogenicity and efficacy of these vaccines as determined in mice and cynomolgus macaques.
2. Materials and Methods
2.1. Cell Lines and Viruses
Vero-76 cells were maintained in Dulbecco’s Minimal Essential Media (DMEM) supplemented with 5% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) within a 37 °C incubator containing 5% CO2
. All viruses used in this study were rescued from infectious cDNA clones. ID ZPC-738 VEEV [32
] and IC 3908 VEEV [33
] were rescued from cDNA clones as described below.
2.2. Creating the IRES Vaccines and Rescue
Creating the ZPC/IRES vaccines was done by modifying the original cDNA clone of the ZPC full-length infectious clone pMI-738 using standard molecular techniques as previously descried [24
]. For ZPC/IRESv1, site-directed mutagenesis of the nsP4-subgenomic promoter region was done using a fusion PCR reaction conducted with the following primer set: 5′gATtACgtTgTAtGGaTgAtaaAACGTTACTGGCCGAAGCCGCTTGGAA3′ and 5′TTttaTcAtCCaTAcAacGTaATcGGGGCCCCTCTCAGGTAGCTGAAT3′. The lower-case letters denote mutations. The EMCV IRES sequence was obtained from the VEEV TC-83/IRESv1 clone previously made [21
]. Naturally occurring SwaI and AflII sites were used to insert the mutated subgenomic promoter and IRES into the pMI-738 cDNA clone. ZPC/IRESv2 was created by fusing the EMCV IRES between the E1 and capsid genes. A TAA stop codon was inserted immediately after E1 followed by a short 3′ UTR sequence upstream of the start of the IRES sequence. A start codon was inserted immediately before the capsid gene. Naturally occurring AvrII, NdeI, and NotI restriction sites in the pMI-738 cDNA were used to assemble the fusion PCR fragments. ZPC/IRES vaccines with individual mutations were also engineered using similar methods using plasmids encoding the desired mutation.
All ZPC/IRES vaccine cDNA clones were fully sequenced by Sanger sequencing and the contigs were assembled into a consensus sequence to ensure only the inserted sequences were present. Viral RNA was obtained by in vitro transcription using an SP6 promoter (mMessage mMachine SP6 transcription kit, Invitrogen, Carlsbad, CA, USA) on a NotI-linearized cDNA template. Vero cells were electroporated with viral RNA as previously described [24
]. Supernatant was collected when cytopathic effects were observed and was clarified by centrifugation prior to aliquoting samples and storing at −80 °C.
VEEV virus and vaccine titrations were done as previously described [24
]. Briefly, Vero cell monolayers in 6- or 12-well plates were infected with serial dilutions of virus supernatant or serum in DMEM supplemented with 2% FBS and 1% P/S. After 1-hour adsorption, monolayers were overlaid with dilution media supplemented with 0.4% agarose and incubated in a 37 °C incubator containing 5% CO2
. Monolayers were fixed with formalin and plaques were visualized using a crystal violet stain after removing the agarose overlay. Plaques were counted in each well and titers are reported as plaque forming units (pfu)/milliliter (mL).
2.4. Plaque Reduction Neutralization Tests
Plaque reduction neutralization tests were performed using sera from vaccinated mice as described in Rossi et al. [24
]. Briefly, samples were diluted 1/10 in PBS prior to heat inactivation at 56 °C for 1 h. Serial two-fold dilutions of sera were performed in media prior to adding 800 pfu of virus. After incubating the sera and viruses for an hour at 37 °C, the mixture was added to a monolayer of Vero cells for hour. An overlay containing 0.4% agarose in media was added to each well and incubated for 48 h. Plaques were visualized by crystal violet staining. Average un-neutralized plaque counts varied between 24 and 35, from which the 50% (PRNT50
) and 80% (PRNT80
) neutralization levels were calculated.
2.5. Replication Curves
Semiconfluent Vero monolayers in T25 flasks were infected with either TC-83, ZPC-738, ZPC/IRESv1, or ZPC/IRESv2 at a multiplicity of infection (MOI) equal to 0.1 in triplicate. After 1 h, the infection was washed thrice with cell culture media and a sample was taken to determine the level of residual virus after infection. At 6, 12, 24, and 48 h post infection, a small sample was removed from the flask and frozen at −80 °C. To create the growth curve, the initial inoculum and all samples were titrated on Vero monolayers in 12-well plates [24
2.6. Serial Passage in Vero Cells
ZPC/IRESv1 was serially passaged in triplicate in T25 flasks of Vero cells at an MOI of 0.1 Forty-eight hours after infection, each replicate was titrated prior to reseeding a new set of Vero cells into a T25 at an MOI of 0.1 to initiate the next passage. At the end of 10 passages, viral RNA was isolated following TRIzol (Invitrogen, Carlsbad, CA, USA) extraction of supernatant. The entire coding region was Sanger sequenced to identify consensus mutations. Mutations observed in the E2 protein, either alone or together, were incorporated into the ZPC/IRESv1 and ZPC/IRESv2 cDNAs.
2.7. Mouse Studies
All animal handling was conducted at UTMB in accordance with the UTMB Institutional Animal Care and Use Committee approval (IACUC #0209068). Female CD1 mice were purchased from Charles River Laboratories (Wilmington, MA, USA). Adult mice (aged 6–8 weeks) were used for vaccination and immunogenicity studies similar to previous studies [24
]. For vaccination, mice were anesthetized with inhaled isoflurane, then injected subcutaneously in the scruff of the back with 100 μL of virus diluted in DMEM supplemented with 2% FBS and 1% P/S. Vaccination dose was confirmed by back titration of inocula to be 1 × 105
pfu/mouse. Challenge was achieved by delivering VEEV 3908 strain diluted in PBS either subcutaneously (1 × 105
pfu/mouse) or intranasally (1 × 104
pfu, anesthetized mice aspirated 20 μL of virus through the nares). Following vaccination and challenge, mice were observed daily for signs of illness, which included lethargy, ruffled fur, hunched posture, and anorexia. Individuals’ weights were also measured to quantify illness. Any mouse that exhibited signs of neurological infection, including tremors, failure to right and partial hindlimb paralysis, or were unable to reach food or water were humanely euthanized. Euthanized animals were recorded as having died the following day.
To acquire serum to determine viremia or neutralizing antibody levels, blood was collected from anesthetized mice by retro-orbital puncture with a heparinized micro-hematocrit capillary tube (Fisher Scientific, Pittsburgh, PA, USA). In some cases, blood was collected into microfuge tubes containing 225 μL of PBS to achieve a 1:10 dilution of plasma. Blood was centrifuged for 5 min at 3380× g, then diluted plasma was removed and frozen at −80 °C until needed.
Neurovirulence safety studies were done in six-day-old CD1 pups from litters randomized among dams. These pups were inoculated intracranially with 20 μL of virus at a titer of 1 × 104 pfu/mouse, and monitored daily for survival.
2.8. Nonhuman Primate Studies
Nonhuman primate (NHP) cynomologous macaques (Macaca fascicularis) were housed at Tulane National Primate Research Center (TNPRC), which is an AAALAC-accredited facility. All work was approved by the Tulane Institutional Animal Care and Use Committee (Protocol P0171).
Healthy NHPs weighing between 3–6 kg were screened and free of prior infection against alphaviruses (Venezuelan and eastern equine encephalitis viruses, Sindbis, Semliki Forest, and chikungunya viruses) as well as simian immunodeficiency virus, simian type D retrovirus and simian T-lymphotropic virus. NHPs were housed in open metal caging allowing visual contact with others in the room. A standard primate chow and fresh fruits and vegetables were provided daily. An intramuscular injection of ketamine hydrochloride for anesthetization was given prior to collecting blood samples or vaccination.
Prior to the start of the study, some NHPs were implanted with a telemetric device as previously described (Konigsberg Instruments, Pasadena, CA, USA) [25
]. Buprenorphine was given during and after the procedure for analgesia. Post-surgery, NHPs were monitored for signs of infection, lethargy, anorexia, dehydration, and device rejection. Core body temperature was recorded wirelessly and reported in one-hour observation intervals. Baseline temperatures were recorded for 6 days prior to the aerosol challenge. Baseline data over a 24-h period were averaged to determine the temperature threshold and deviations were recorded as 1.5 times higher or lower than the standard deviation for each time. Each NHP served as its own temperature control through the use of collected pre-exposure baseline data. NHPs that were not monitored telemetrically had rectal temperatures taken only daily when anesthetized during a procedure.
NHPs assigned to the control group (N = 2) were given saline and those assigned the vaccine group (N = 5) were given 1 × 105 pfu of ZPC/IRESv1, both in the upper deltoid with a single subcutaneous injection of 100 μL. No signs of disease and distress were noted following vaccination. Blood samples were taken by venipuncture on days 1, 10, 23, and 35 to determine neutralizing antibody levels. Sera were clarified from this blood by centrifugation to determine PRNT titer.
Aerosol exposure to VEEV 3908 was performed on day 35 post vaccination. A head-only 16-liter dynamic inhalation aerosol exposure was used to challenge each NHP as previously reported [34
]. Collision and AGI samples were used to determine the infectious dose received by each NHP, which fell in the range between 8 × 105
and 9 × 106
pfu. Animals were observed for signs of illness and telemetrically monitored for temperature changes. Blood was taken for the first 3 days following challenge to determine viremia. On day 24 post challenge, NHPs were euthanized by a pre-dose of ketamine hydrochloride followed by an overdose of sodium pentobarbital. Necropsies were performed to determine any gross pathological changes.
2.9. Statistical Analyses
All graphs and statistical analyses were created and performed using GraphPad software v8.0. Significance in growth curve titers was determined by a two-way ANOVA with Tukey posthoc tests to compare multiple groups. Vaccine viremia and NHP fever-hours were compared individually using Student’s t-test. For values below or above the LOD, one-half of the limit of detection (LOD) was recorded and statistics were conducted. Survival significance was determined by Log-rank (Mantel-Cox) test comparing two groups.
There is a continued need and demand for safe and efficacious vaccines against the diseases caused by alphaviruses, including chikungunya, Mayaro, eastern and Venezuelan equine encephalitis viruses. We have shown for many of these viruses that the IRES platform addresses this need while also limiting their spread through mosquito vectors [22
]. Here, we continue to improve upon and broaden the scope of our IRES-vaccines by including the ID serotype of VEEV.
In a previous study, we reported on the similar construction and evaluation of the IE subtype of VEEV IRES vaccines [24
]. The 68U201/IRESv1 (VEEV/mutSG/IRES/1) and 68U201/IRESv2 (VEEV/IRES/C) both were able to elicit protective immunity in CD1 mice. Both vaccines also showed a marked decrease in neurovirulence in CD1 pups compared to wild type parent viruses, with little difference between them. The differentiation between these vaccines was apparent with the neutralizing antibody titers, as VEEV/IRES/C elicited two-fold greater PRNT50
values in CD1 mice than VEEV/mutSG/IRES/1. A very similar trend is observed here with ZPC/IRESv1 and ZPC/IRESv2, where the latter produced higher PRNT80
titers, especially at 20-weeks-post vaccination (Table 3
). More importantly, all ZPC/IRESv2-vaccinated mice produced a neutralizing antibody response after a single vaccination.
The decision to pursue ZPC/IRESv1 for characterization in the NHP model was made based upon the results from not only ZPC/IRES vaccine data, but also from other IRESv1 vaccines that have been heavily tested in mice and NHPs with positive results [25
]. ZPC/IRESv1 was chosen as the first candidate because it provided protective murine immunity and had a similar safety profile to TC-83 in intracranially injected neonatal mice. The Version 2 vaccine based either the 68U201 [24
] or TC-83 [22
] VEEV backbones were also more virulent in previous studies. Five NHPs were vaccinated with ZPC/IRESv1 and 4 showed positive neutralizing antibody responses immediately prior to 3908 aerosol challenge. Interestingly, none of the vaccinated NHPs had a detectable viremia on days 1–3, even in the non-responder. However, the non-responder and two NHPs with a PRNT80
titer of 20 and 40, had 107, 110, or 10 fever-hours, respectively, suggesting that protection from disease was incomplete. The reason for this discordance between protection against viremia and fever is puzzling. The challenge virus used as the IC strain 3908 while the vaccine backbone was the ID strain ZPC-738. While the majority of vaccinated NHPs (and mice) did have heterologous cross-protection, there may be some incomplete protection at the lowest measurable PRNT80
titers. Further, viremia is not routinely measured as a correlate of protection, and human IND trials with the TC-83 vaccine have shown disconnects between adverse, flu-like symptoms, and immunogenicity [11
]. Regardless, immunogenicity elicited by the ZPC/IRESv2 vaccine was superior to that of ZPC/IRESv1 in both mice and NHPs. CHIKV/IRESv2 also provided a better neutralizing antibody immune response than CHIKV/IRESv1 in cynomolgus macaques [40
]. Furthermore, our earlier studies of the VEEV 68U201/IRESv2-vaccinated mice also had a slightly higher PRNT80
titers than those induced by 68U201/IRESv1, which lasted for more than a year after vaccination [24
]. Unfortunately, due to the study design, fever-hour reductions could not be compared between ZPC/IRESv1 and ZPC/IRESv2.
One confounding issue in our challenge of the ZPC/IRESv2 NHPs was the inconsistent viremia in the sham-vaccinated group. This lack of consistent viremia makes it difficult to determine if the ZPC/IRESv2-vaccinated NHPs were truly protected from viremia or not adequately exposed to an appropriate dose of VEEV. Furthermore, fever-hours were not reported as part of this study so protection from disease could not be assessed either. It is likely that these NHPs would be protected from both viremia and fever based on their high PRNT80
titers (Table 4
) but that conclusion cannot be reached from the data gathered in this experiment. However, experience from a similar study evaluating the efficacy of 68U201/IRESv1 showed complete protection from IE subtype 68U201-challenge in NHPs with a similar PRNT80
Serial passaging of ZPC/IRESv1 in Vero cells resulted in several consensus mutations, 3 of which were nonsynonymous changes to the E2 and E1 glycoproteins. When these E2 mutations, either in combination or alone, were genetically engineered back into the ZPC/IRESv1 vaccine, some constructs lost their ability to elicit a protective immune response. Surprisingly, ZPC/IRESv1 A76, R85 generated no vaccine-related viremia, and vaccinated mice were protected or had neutralizing antibody titers. When these mutations were investigated alone, vaccine-related viremia and PRNT80
titer were detected in only a portion of the mice but all were protected from lethal 3908 challenge. One explanation for this phenotype is the accumulation of positively charged amino acids on the E2 and E1 proteins that contribute to commonly seen cell culture adaptive mutations that facilitate binding to heparin sulfate (HS). Studies in Sindbis virus have long shown a negative correlation between HS binding and virulence [41
], but this phenotype is more complex for VEEV [43
]. Sindbis virus passaged on BHK-21 cells showed nonsynonymous mutations favoring positively charged amino acids like arginine that increased the ability of virions to bind to cells and a reduction in murine neonatal fatalities [42
]. Likewise, Griffin and colleagues showed that the loss of positively charged amino acid changes in the E2 protein resulted in increased Sindbis virulence in sucking mice compared to those with more positive E2 proteins [41
]. This is a common feature for serially passaged alphaviruses and underlines the importance of using low-passage stocks for studies. Additional experiments are required to determine when in the 10-passage series these mutations arose and whether genetically stabilizing this region using synonymous mutations can reduce the frequency of these mal-adaptive mutations. As it is highly unlikely that these mutations would arise naturally in vivo after vaccination, sera at peak viremia from these mutations were not sequenced. If a mutation did occur, it would likely be associated with a high viremia titer, which was not observed in these studies. Therefore, care would need to be taken to limit tissue culture amplification when creating vaccination stocks.