Neurovirulence Properties of Recombinant Vesicular Stomatitis Virus Vectors in Non-Human Primates

VSV shedding and lack of detectable viremia following IN inoculation with rVSV_IN

A pilot study was performed to measure viremia and shedding of infectious virus following IN inoculation of three rhesus macaques with 10⁷ PFU of rVSV_IN (Figure 1A). Blood and serum were collected pre-dose, and then, along with nasal washes and saliva, were collected every day for the first four days following inoculation and at defined intervals thereafter (see Materials and Methods). Using a standard plaque assay, infectious virus was observed in nasal wash samples of all three inoculated macaques at day 1 (Table 1). Only one macaque (number 23) had a low level (1–10 PFU/mL) of detectable infectious virus in its nasal wash at day 2. No infectious virus was detected in nasal wash samples from any macaques following day 2. The two macaques (numbers 48 and 62) that had the highest titers of virus in nasal wash samples at day 1 also shed low levels of infectious virus (10–100 PFU/mL) in their saliva at day 1. No virus was detected in saliva of the third macaque (number 23) on day 1. Infectious virus was not detected in saliva from any macaque following day 1. The unvaccinated macaques served as negative controls for all samples at all time points (Table 1).

In addition to viral shedding at mucosal surfaces, we assayed plasma and peripheral blood mononuclear cells (PBMCs) for the presence of infectious virus following IN inoculation. Infectious virus was not detected in plasma or PBMCs of any macaque post-inoculation, indicating the absence of detectable viremia in vaccinated macaques. Using a VSV neutralization assay, VSV_IN-specific serum antibodies were detected in all three vaccinated macaques by day 14 (Table 2), indicating that a productive infection occurred following IN inoculation.

No evidence of neuroinvasion following IN inoculation

Cynomolgus macaques were inoculated IN on day 0 with 10⁷ PFU of VSV_IN as shown in Figure 1B. Macaques were clinically monitored for 21 days, at which time they were subjected to euthanasia, necropsy, and tissue harvest. Group mean clinical scores of macaques in all groups were ≤ 0.1 (out of 4); the wt VSV_IN and all rVSV_IN groups were indistinguishable from the scores of the negative control group.

To assess neuroinvasiveness, brain and spinal cord tissues were harvested at necropsy and analyzed by plaque assay, real-time, quantitative RT-PCR, and histopathologic examination. No virus or viral RNA was detected in frontal cortex, occipital cortex, olfactory bulb, CSF, or nasal turbinates from any of the macaques at day 21. Furthermore, histopathological analysis revealed that no treatment related lesions or changes were present in the brains or spinal cords of macaques given wt VSV_IN or rVSV_IN vectors compared with the control macaques given PBS. These results indicate that none of the rVSV_IN vectors or even the wt VSV_IN was capable of spreading to the CNS following IN inoculation. To demonstrate that the macaques inoculated IN with each virus were productively infected, we measured anti-VSV_IN neutralizing antibodies (Table 2). Macaques in the wt VSV_IN, rVSV_IN, and rVSV_IN-HIVGag5 groups all had neutralizing antibodies in the range of 600–6000 (geometric mean titers between 1280–2167, Table 2). Macaques in the rVSV_IN-CT1-HIVGag5 and PBS groups did not have detectable neutralization titers.

Neurovirulence study clinical observations

Following IT inoculation with 10⁷ pfu of virus, all four macaques in the wt VSV_IN group and one of four in the rVSV_IN backbone group were euthanized early because of declining health status, which included head tilt, irregular gait, tremors, and ataxia, signs consistent with neurological disease (Table 3). The onset of these signs generally occurred on day 6 or 7 post-inoculation. There were no significant virus related clinical signs in the remaining macaques in the IT groups (Table 3). One of the PBS control macaques, 22290F, had a head tilt, lethargy, and bilateral pupil dilation on days 1–21 (Table 3). Because the onset of these signs occurred earlier than those in the wt or rVSV_IN groups, and because none of the other macaques in the control group displayed similar signs, it was concluded that these clinical features were procedure-related.

Detection of VSV in tissues from IT groups

At the time of necropsy, samples were taken from the frontal and occipital cortices, as well as from CSF, and were assayed for infectious VSV by plaque assay and for viral RNA by real time, quantitative RT-PCR. All five macaques euthanized by day 8 had infectious virus in the frontal cortex, which was the brain section most proximal to the injection site. Viral titers were within the range of 10² – 10⁷ PFU/g (Table 4). Similarly, viral genomic RNA within the range of 10⁶ – 10¹⁰ copies/g was detected in the frontal cortex of all five of these macaques. None of the five macaques euthanized by day 8 had infectious virus in the occipital cortex, which was distal to the injection site. However, three of the four macaques in the wt group did have low levels (10⁵–10⁶ copies/g) of detectable viral RNA in their occipital lobes. In addition, low levels of infectious virus (56 PFU/mL) and viral genomes (5.8 × 10⁵ copies/mL) were detected in the CSF of one macaque in the wt group, F22190F (Table 4). All other tissues and CSF from this group were negative for infectious virus and viral genomes. Viruses recovered from brain samples of macaques subjected to euthanasia by day 8 were sequenced across their entire genomes. Virus isolated from two out of four macaques in the wt group had no mutations relative to the inoculum. Of the viruses isolated from the remaining two wt-inoculated macaques, one had a single amino acid substitution in G and the other had two substitutions in L (Table 5). The virus isolated from the one rVSV_IN macaque euthanized by day 8 was found to contain identical sequence relative to input virus (Table 5). Taken together, these results suggest that the clinical outcomes in these macaques were caused by the input viruses, rather than by selection mutants with increased NV potential.

For the remaining macaques in the rVSV_IN group and all the macaques in the rVSV_IN-HIVGag5 and rVSV_IN-CT1-HIVGag5 groups, no infectious virus or viral RNA was detected in any tissue sample collected at necropsy (day 21). Similarly, the tissues from the PBS group of macaques were negative in both assays.

To demonstrate that productive infections occurred in all groups, including those with no clinical scores or detectable virus, neutralizing antibodies were measured in the IT groups of macaques (Table 2). All macaques that were euthanized by day 8 had low levels of neutralizing antibodies (geometric mean titers between 40–230). All rVSV_IN and rVSV_IN-HIVGag5 macaques that survived to day 21 had neutralizing antibodies between 600 and 6000 (geometric mean titers between 1280–4305). With the exception of one macaque that had an undetectable neutralizing antibody titer, those in the rVSV_IN-CT1-HIVGag5 group had neutralizing antibodies titers of 80–160 (geometric mean = 80).

Histopathological analysis of brain and spinal cord tissues from macaques in IT groups

Slides of fixed brain and spinal cord tissues were analyzed for histological changes. For each tissue section, observed lesions were noted according to relative size, number, and type (e.g., inflammatory, necrotic, perivascular, periventricular). To summarize the histological findings, the following scoring system was used to quantify the observed lesion severities: 0= no lesions, 1= minimal, 2= moderate, 3= marked, 4= severe (Table 6). Macaques inoculated with PBS had only minimal to moderate areas of malacia and inflammation localized to the injection site, which indicated tissue damage resulting from the needle. Macaque 22290F (PBS), which had an average clinical score of 3 for the length of the study (Table 3), displayed only needle track-associated tissue damage (Table 6), which supported the notion that the clinical features were procedure related. There were no notable lesions in other tissue sections in the PBS group (Figure 2E, Table 6). Brain sections from macaques inoculated with wt VSV_IN showed moderate to severe necrotizing meningoencephalitis characterized by malacia, hemorrhage, choroiditis, loss of ependymal cells, and moderate numbers of perivascular cuffs of mononuclear cells (Figure 2A). Macaques inoculated with rVSV_IN had lesions that were of similar severity to those in the wt VSV_IN group; however, the lesions in the rVSV_IN macaques were characterized by a predominant cellular inflammatory response with fewer areas of malacia than in the wt VSV_IN inoculated macaques (Figure 2B). The macaques in the rVSV_IN-HIVGag5 and rVSV_IN-CT1-HIVGag5 groups had lesions that were primarily inflammatory with minimal to moderate severity (Figure 2C, D, F, and Table 6). Statistical analysis of the brain lesion scores (Table 6) revealed that the wt VSV_IN and rVSV_IN groups were significantly greater than those of the PBS control goup (p <0.01) whereas the rVSV_IN-HIVGag5 and rVSV_IN-CT1-HIVGag5 group scores did not differ significantly from those of the PBS control group.

It is noteworthy that the wt VSV_IN inoculated macaques generally had fewer, less severe lesions in their spinal cords than the macaques in the rVSV_IN group (Figure 3 and Table 6). The mean severity score for spinal cord lesions in the wt VSV_IN group (Figure 3A) was 0.5 (minimal), while the mean score for the rVSV_IN group (Figure 3B) was 2.7 (marked). Macaques in both HIVGag5 vector groups had spinal cord lesions that were less severe than those in the rVSV_IN group, with mean severity scores of 1.0 (minimal) and 0.5 (minimal) for rVSV_IN-HIVGag5 and rVSV_IN-CT1-HIVGag5, respectively (Figure 3C, D, F, and Table 6). These observations may have resulted from the differences in days of necropsy; that is, viruses in macaques that survived to day 21 may have had more opportunity to spread to the spinal cord than those in macaques euthanized by day 8. Statistical analysis of the mean spinal cord lesion scores (Table 6) revealed that no significant differences existed between the PBS group, the wt group, or the two rVSV-HIV groups (p>0.05); however, all four of these groups had significantly lower lesion scores than the rVSV_IN group (p<0.05).

Viruses and vectors tested in NHP studies

The viruses tested in these studies included tissue culture-adapted, wt VSV_IN; the prototypic, attenuated vector rVSV_IN; the rVSV_IN vector encoding HIV Gag protein; and a further attenuated vector encoding HIV Gag, rVSV_IN-CT1-HIVGag5 (Figure 1). The tissue culture-adapted San Juan isolate of VSV Indiana serotype (wt VSV_IN) was least attenuated in vitro and was highly pathogenic in mice (Roberts et al., 1998). The rVSV_IN was derived from a cDNA encoding the entire genome of VSV_IN (Lawson et al., 1995), which was constructed using RT-PCR products combined from the genomes of VSV_IN, San Juan and Mudd-Summers isolates. The resulting virus contained multiple nucleotide sequence differences from the parental San Juan isolate that presumably accounted for the in vivo attenuation of rVSV_IN relative to wt VSV_IN (Roberts et al., 1998). This cDNA served as the “backbone” vector from which the two rVSV-HIV vectors were derived. The prototypic HIV vaccine vector, rVSV_IN-HIVGag5, was similar to the vectors tested previously in SHIV challenge studies in macaques (Rose et al., 2001), except that this vector encoded an HIV-1 Gag protein rather than an SIV Gag. Finally, the rVSV_IN-CT1-HIVGag5 contained a mutant G gene that encodes a protein in which the cytoplasmic tail domain is truncated from 29 amino acids to a single amino acid (CT1); this mutation has been shown to greatly reduce vector pathogenicity in mice following IN inoculation (Roberts et al., 1999; Schnell et al., 1998). Since PBS was the diluent for each virus, it was used as the negative control inoculum.

The wild type, rVSV_IN, and rVSV_IN-CT1-HIVGag5 viruses were provided by Dr. John K. Rose (Yale University, New Haven, CT). The rVSV_IN-HIVGag5 virus was rescued from cDNA pVSV_IN-HIVGag5 using standard reverse genetics procedures (Lawson et al., 1995; Schnell et al., 1996b). The cDNA pVSV_IN-HIVGag5 was generated by PCR amplification of the ORF of the HIV-1 gag gene (HXB2 strain) and insertion into the VSV vector backbone pVSV-XN1 (Schnell et al., 1996b) via Xho I and Nhe I restriction sites as the fifth gene in the genome. All viruses were amplified on baby hamster kidney (BHK) cells at a multiplicity of infection (MOI) between 0.1 and 1, concentrated by ultracentrifugation through 10% sucrose (28,000 rpm in a Beckman SW28 rotor at 4°C for 75–90 min), and resuspended in Phosphate Buffered Saline (PBS).

Viral shedding study in rhesus macaques

The rhesus macaque study was performed at the University of Louisiana at Lafayette-New Iberia Research Center, New Iberia, LA, with adherence to the regulations outlined in the USDA Animal Welfare Act (9 CFR, Parts 1, 2 and 3) and the conditions specified in the Guide for the Care and Use of Laboratory Animals (ILAR publication, 1996, National Academy Press). Any clinical signs of illness or distress were promptly reported to the responsible veterinarian, who recommended either treatment for minor ailments and injuries or euthanasia for profound effects such as severe emesis or convulsions.

Because of limited availability of macaques, three male rhesus macaques (3–4 years old) were vaccinated, and two un-inoculated macaques served as negative controls (Figure 1A). Following light sedation with ketamine, macaques were inoculated IN with a total of 10⁷ plaque-forming units (PFU) of rVSV_IN in PBS, delivered to both nostrils in a total of 0.8 mL by micropipettor. Samples were collected as follows: saliva and nasal washes (in that order, ≥ 1.0 mL each) on days 1–4, 6, 8, 10, 13, 16, 19, 22; serum (1.0 mL) pre-dose and on days 1–4, 8, 13, 22; whole blood (including sodium heparin as an anticoagulant, 10mL) pre-dose and on days 8, 14, and 22.

Neuroinvasiveness and neurovirulence (NV) studies in cynomolgus macaques

Both the NV and neuroinvasivenss studies were performed at Charles River Laboratories, Inc.-Preclinical Services (Sparks, NV). The cynomolgus macaque studies adhered to the regulations outlined in the USDA Animal Welfare Act (9 CFR, Parts 1, 2 and 3) and the conditions specified in the Guide for the Care and Use of Laboratory Animals (ILAR publication, 1996, National Academy Press). Any clinical signs of illness or distress were promptly reported to the responsible veterinarian, who recommended either treatment for minor ailments and injuries or euthanasia for profound effects such as severe emesis or convulsions.

Cynomolgus macaques (2–3.5 years old) were inoculated either intranasally (IN) or intrathalamically (IT) with a total of 10⁷ PFU of virus per macaque (Figure 1B). Each group contained two females and two males. For IN inoculations, macaques were lightly sedated with ketamine. Each macaque received a total of 1.0 mL of inoculum, injected via syringe into both nostrils over the course of approximately one minute. Non-anti-inflammatory analgesics were given at the discretion of the veterinarian to prevent post-inoculation discomfort.

For IT inoculations, macaques were anesthetized with a mixture of ketamine and valium; each macaque’s head was shaved, and incisions were made through the skin at approximately 1.5 cm on either side of the sagittal suture and at approximately 0.5 cm from the coronal suture of the head of the macaque. A high-speed, general twist drill was used to make two 0.16 cm holes through the skull, corresponding to the positioned incisions described above. Subsequently, 0.1 mL of the virus or control inoculum was injected into the right and left thalamic regions for a total dose volume of 0.2 mL/macaque, using a 25 gauge × 1½ inch needle (to a depth of approximately 1 inch) on a 1.0 mL tuberculin syringe, introduced in a medial-oral direction. After inoculation, the surgical area was closed with a subcuticular suture to minimize self-inflicted trauma by the macaque. Each macaque was observed until it recovered from anesthesia and was then returned to its cage. Non-anti-inflammatory analgesics were given at the discretion of the veterinarian to prevent post-surgical discomfort.

Macaques were monitored for 21 days after challenge for the following parameters: food consumption, body weight and temperature, serum chemistry, hematology, and general clinical parameters, the latter of which were scored using a system based on that used for yellow fever vaccine (Levenbook, Pelleu, and Elisberg, 1987). Macaques were examined twice daily for possible clinical signs of encephalitis and were assigned numerical scores as follows: 0, no clinical signs of encephalitis; 1, rough coat, not eating; 2, high pitched voice, inactive, slow moving; 3, shaky movements, tremors, incoordination, limb weakness; 4, inability to stand, limb paralysis, moribundity, or death. The higher of the two daily individual macaque scores was assigned as the daily score. The mean daily clinical score for each macaque was calculated by averaging the 22 individual macaque daily scores, and the group mean daily scores were calculated for each study day by averaging the individual macaque scores for each group. At the time of euthanasia (scheduled or unscheduled), the macaques were exsanguinated while under deep anesthesia induced with ketamine and Beuthanasia®-D.

Tissue Preparation

All tissues and fluids were collected at the time of euthanasia. For quantification of virus, frozen samples of occipital and frontal cortex were thawed, weighed and suspended in PBS supplemented with 1x sucrose, phosphate, glutamate buffer (0.2M sucrose, 7.0mM K₂HPO₄, 3.8mM KH₂PO₄, 5.0mM glutamic acid) to generate 10% (w:v) tissue suspensions. Samples were homogenized using an Omni Mixer ES with aerosol-sealed stainless steel chambers (Omni International). Tissue homogenates were centrifuged at 3000 rpm for 15 min at 4°C and clarified supernatants used immediately for plaque assay or RT-PCR, as described below. Cerebral spinal fluid (CSF) samples were thawed and used directly in assays. PBMCs were assayed for the presence of infectious VSV by two methods: direct co-culture of the PBMCs with BHKs or following a freeze-thaw and clarification of the lysates.

Histopathological Evaluation

Brain and spinal cord tissue sections were collected from each macaque (following gross necropsy) and fixed in neutral buffered 10% formalin. Samples were taken from five brain regions (Figure 1C) including frontal cortex, occipital cortex, cerebellum, thalamus, basal ganglia, and from three spinal cord regions (cervical, thoracic, and lumbar). Fixed tissues were mounted onto slides and stained with eosin and hematoxylin for histological examinations, which were performed by Roger Price, D.V.M., Ph.D., D.A.C.V.P. (Veterinary Pathology Services, Houston, TX), and peer-reviewed by Andrew Lackner, D.V.M., D.A.C.V.P. (Tulane National Primate Research Center, Covington, LA) and Mary E. P. Goad, D.V.M., Ph.D. (Exploratory Drug Safety, Wyeth Research, Andover, MA).

Plaque Assay

Ten-fold serial dilutions of tissue homogenates, PBMCs, PBMC lysates, or CSF were prepared in Dulbecco’s Modified Eagle Medium (DMEM, Mediatech, Inc.) with 10% fetal bovine serum (FBS). Aliquots (100 μL) of each dilution were placed on confluent BHK or Vero cell monolayers in 6 well plates and allowed to absorb for 15 min at room temperature, followed by incubation at 37°C for 30 min. Inoculum was removed from each well, which was then overlaid with 0.8% agarose in DMEM + 10% FBS. Plates were incubated at 37°C overnight. Agar overlays were removed and plates were stained for 10–20 min at room temperature with 2 mL/well of crystal violet stain solution (1% crystal violet in 70% methanol).

Real-time quantitative RT-PCR

Aliquots of tissue homogenates (500 μL) were centrifuged at 14,000 rpm for 5 min. Viral RNA was extracted from clarified supernatants or CSF samples (140 μL) using QIAamp Viral Mini Kit (Qiagen, Inc.). An equivalent volume of buffer was extracted as a negative control, and an aliquot of the viral inoculum was extracted as a positive control.

Using Primer Express software (PE Applied Biosystems), primers (Qiagen) and probe (AB Applied Biosystems) for the VSV N gene were designed to amplify a 70-nucleotide fragment. The sequence of the forward primer was as follows: 5′-GATAGTACCGGAGGATTGACGACTA -3′ corresponding to nucleotide positions 1075–1099 in the N open reading frame. The reverse primer sequence was as follows: 5′-AACCATCCGAGCCATTCGA -3′ corresponding to nucleotide positions 1127–1145 in the N open reading frame. The fluorogenic probe was labeled at the 5′ end with the reporter dye, 6-carboxyfluorescein (FAM), and at the 3′ end with the quencher dye, 6-carboxytetramethylrhodamine (TAMRA), and its sequence was as follows: 5′-TGCACCGCCACAAGGCAGAGA –3′ corresponding to nucleotide positions 1101–1121 in the N open reading frame.

To determine the number of genome copies of VSV, a synthetic RNA oligonucleotide corresponding to the entire amplified region was synthesized and PAGE-purified (TriLink BioTechnologies). Following precipitation and centrifugation, the number of oligonucledotide copies per μL of solution was calculated using Avogadro’s number, the molecular weight, and the extinction coefficient of the oligonucleotide. Serial 10-fold dilutions were made of this synthetic oligonucleotide and used to generate a standard curve.

The quantitative RT-PCR assay was developed as a two-step procedure to detect the genomic (negative sense) RNA of VSV. The initial reaction mixture for reverse transcription (25 μL) contained 7.5 μL purified RNA, 900 nM forward primer, reagents from the TaqMan® Reverse Transcription Reagents kit (Applied Biosystems), and 15 units/μL of ThermoScript™ Reverse Transcriptase (InVitrogen^™ Life Technologies). The reaction was incubated at 60°C for 30 min., which was followed by an inactivation step of heating at 95°C for 15 min. Then, 30 μL of 2x Master mix buffer (TaqMan® One Step PCR Master Mix Reagent Kit; Applied Biosystems), 900 nM of reverse primer, and 200 nM of probe were added to the cDNA reaction mix to bring the final volume to 60 μL. Amplification and detection were performed with an ABI Prism 7700 Sequence Detection System under the following conditions: 10 min at 95°C, and 40 cycles of 15 s at 95°C and 1 min at 60°C. All samples were tested in duplicate.

VSV Serum Neutralization Assay

Sera were diluted with PBS in a volume of 50 μL in serial two-fold dilutions in 96-well plates. One hundred PFU of virus (rVSV_IN) in 50 μL of serum-free DMEM was added to each well. Plates containing serum and virus were incubated at 37°C for 1 h. Approximately 1,500 BHK cells in 100 μL of DMEM-10% FBS were then added to each well. The plates were incubated at 37°C and 5% CO₂ for 2 to 3 days and assayed by microscopy for the presence of VSV-induced cytopathology. Each assay was performed in duplicate. Neutralizing titers were assigned as the highest dilutions that completely inhibited VSV-induced cytopathology.

Statistical Analysis

Differences in histopathological scores among groups of macaques were analyzed by one-way analysis of variance using the Dunnett multiple comparisons post test (InStat version 2.02, GraphPad Software, San Diego, CA).