AAV-Vectored Expression of Marburg Virus–Neutralizing Antibody MR191 Provides Complete Protection From Challenge in a Guinea Pig Model

Abstract

Although there are no approved countermeasures available to prevent or treat disease caused by Marburg virus (MARV), potently neutralizing monoclonal antibodies (mAbs) derived from B cells of human survivors have been identified. One such mAb, MR191, has been shown to provide complete protection against MARV in nonhuman primates. We previously demonstrated that prophylactic administration of an adeno-associated virus (AAV) expressing MR191 protected mice from MARV. Here, we modified the AAV-MR191 coding sequence to enhance efficacy and reevaluated protection in a guinea pig model. Remarkably, 4 different variants of AAV-MR191 provided complete protection against MARV, despite administration 90 days prior to challenge. Based on superior expression kinetics, AAV-MR191-io2, was selected for evaluation in a dose-reduction experiment. The highest dose provided 100% protection, while a lower dose provided ∼88% protection. These data confirm the efficacy of AAV-mediated expression of MR191 and support the further development of this promising MARV countermeasure.

Marburg virus (MARV) remains an ongoing threat to global public health [1]. Since its discovery in 1967, MARV—along with the highly related Ravn virus—has caused at least 18 outbreaks of Marburg virus disease (MVD), a severe illness with a fatality rate as high as 90% [1]. In 2021, MARV emerged for the first time in West Africa, and it has caused an outbreak in that region every year since, including an outbreak in Equatorial Guinea that recently ended on 8 June 2023 [2]. In March 2023, another outbreak was declared in Tanzania, marking the first time that MARV has been detected in that country [3]. To date, no therapeutics or vaccines have been licensed for the treatment or prevention of MVD, although several promising countermeasures have been developed [4].

Among the most promising experimental treatments for MARV are monoclonal antibodies (mAbs) that bind to the viral glycoprotein (GP), the only viral protein expressed on the surface of the virion and one that is critical for mediating virus attachment and entry [5]. In binding to MARV GP, mAbs neutralize the virus and/or elicit an Fc-effector response from immune cells, thereby inhibiting virus replication. mAb MR191, which was isolated from a circulating B cell from an MVD survivor [6], has proven to be particularly potent, capable of neutralizing MARV and protecting both guinea pigs and nonhuman primates (NHPs) from lethal MARV infection [6, 7]. Indeed, a single 10-mg dose of antibody delivered 4 days after challenge with MARV was sufficient to completely protect guinea pigs. Likewise, 50 mg/kg doses delivered on days 4 and 7 after MARV challenge completely protected rhesus macaques from disease, while the same dose delivered on days 5 and 8 protected 4 of 5 animals.

While mAbs have demonstrated clear therapeutic efficacy against MARV, they are not without their disadvantages. Passive administration of antibodies is typically performed through intravenous infusion over multiple treatments, which may not always be practical, particularly in resource-limited settings [8, 9]. Moreover, mAb production and purification can be time-consuming and cost-intensive, further limiting the availability of these therapeutics. Vector-mediated mAb expression—also referred to as vectored immunoprophylaxis (VIP)—presents a potential alternative to passive immunotherapy by allowing for the prophylactic production of mAbs by transduced cells within the treated individual. Adeno-associated virus (AAV) vectors are commonly used for this purpose, have an outstanding safety profile, and promote robust and sustained transgene expression [10]. Indeed, we have previously developed a novel, muscle-tropic AAV vector (known as AAV6.2FF) capable of driving high and sustained levels of protective mAbs in a variety of animal models [11–13]. We recently demonstrated that intramuscular administration of an AAV6.2FF vector encoding the MARV-specific mAb MR191 could protect mice from lethal challenge with mouse-adapted MARV [11].

In the present study, we sought to evaluate the efficacy of AAV-mediated MR191 expression in the more stringent guinea pig model. A rationally redesigned AAV vector expressing MR191 with the addition of the TLR9-inhibiting io2 sequence resulted in superior mAb expression kinetics and complete protection against guinea pig-adapted MARV. These data confirm the efficacy of AAV-mediated expression of MR191, and they will help advance the preclinical development of this important therapeutic.

METHODS

Animal Ethics and Biosafety Statement

All animal experiments were reviewed and approved by the animal care committees at the University of Guelph, Guelph, Ontario, or the Canadian Science Centre for Human and Animal Health (CSCHAH), Winnipeg, Manitoba in accordance with the guidelines of the Canadian Council on Animal Care. All work with infectious MARV was performed in the containment level 4 laboratory at the CSCHAH following standard operating protocols.

AAV Vector Construction and Production

The heavy and light chain variable region sequences of the MR191 mAb were used as reported [6]. Four different versions of the MR191 mAb sequence were synthesized as immunoglobulin G1 (IgG1) molecules (Figure 1): unmodified MR191possessing a human IgG1 constant domain (MR191); MR191 possessing a human IgG1 constant domain and containing 2 TLR9-inhibitory io2 sequences [14] in the genome located immediately adjacent to the 5′ and 3′ inverted terminal repeats (ITRs) (MR191-io2); guinea pig codon optimized MR191 possessing a human IgG1 constant domain (MR191-GP_op); and MR191 containing guinea pig heavy and light chain constant domains (MR191-GP_HL). mAb genes were expressed from a single gene expression cassette under the control of a CASI promoter to generate heavy chain and lambda light chain polypeptides separated by a furin-2A (F2A) self-cleaving peptide. A woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) and SV40 polyadenylation signal were added downstream of the mAb coding sequence, and the entire expression cassette was flanked by AAV2 ITRs. AAV genomes were packaged into the AAV6.2FF capsid [13], and vector production was carried out as previously described using adherent HEK293 cells and heparin affinity chromatography [15]. AAV vector genome titers were determined by quantitative polymerase chain reaction (qPCR) using a Taqman probe specific for the SV40 sequence as described [15].

Figure 1.

Design of AAV genomes expressing MR191. All adeno-associated virus (AAV) genomes are flanked by AAV2 inverted terminal repeats (ITR) and contain the ubiquitous CASI promoter, a WPRE (woodchuck hepatitis virus posttranscriptional regulatory element), and pA (SV40 poly-adenylation signal). The variable heavy and light chain domains are separated by an F2A (furin 2A self-cleaving peptide). The AAV6.2FF-MR191 vector encodes the variable heavy and light chain domains of MR191 with human IgG1 heavy and human lambda light chain constant domains. AAV6.2FF-MR191-GP_op contains the same sequences as MR191 but codon optimized for guinea pigs. AAV6.2FF-MR191-io2 contains the same MR191 sequences with the addition of 2 TLR9 inhibitory sequences (io2) adjacent to the 5′ and 3′ ITRs. Lastly, the transgene cassette for AAV6.2FF-MR191-GP_HL contains the guinea pig heavy and light chain constant domains.

Mouse Experiments

Five-week-old BALB/c mice were purchased from Charles River Laboratories (Saint Constant, QC, Canada) and allowed to acclimate for 1 week prior to experimentation. BALB/c mice were used for monitoring the kinetics of mAb expression in vivo. AAV vectors were administered at a dose of 1 × 10¹¹ vector genomes (vg) per mouse to the gastrocnemius muscle using a 29-guage needle in a total volume of 40 μL. Weekly blood collection was completed by saphenous bleed using EDTA microvettes (16.444.100; Sarstedt) and plasma samples obtained.

Guinea Pig Experiments

Female Hartley guinea pigs (200–250 g) were purchased from Charles River Laboratories. For the pilot study, 4 groups of 3 guinea pigs were treated with 5 × 10¹² vg of AAV6.2FF-MR191, AAV6.2FF-MR191-io2, AAV6.2FF-MR191-GP_op, or AAV6.2FF-MR191-GP_HL via intramuscular (IM) administration. For the dose-response study, 3 groups of 6 guinea pigs were treated with AAV6.2FF-MR191-io2 via IM administration at doses of 5 × 10¹² vg (high), 1 × 10¹² vg (medium), or 5 × 10¹¹ vg (low). A control group of 6 animals was also treated with 5 × 10¹² vg of AAV6.2FF-FluA20, which expresses an irrelevant mAb. All AAV vectors were diluted in phosphate-buffered saline (PBS) to a final volume of 300 μL. Blood samples were collected in serum separator tubes (BD Microtainer SST; BD) at predetermined time points leading up to challenge: 0, 14, 28, 56, and 90 days after AAV administration for the pilot study and 0, 7, 14, 21 and 28 days for the dose-response study. At 92 days (pilot study) or 28 days (dose-response study) after AAV administration guinea pigs were challenged with 1000 median lethal dose (LD₅₀) of guinea pig-adapted Marburg virus (GPA-MARV; Marburg virus NML/C.porcellus-lab/AGO/2005/Ang-GA-P2, GenBank accession No. MF939097) via intraperitoneal injection. Animals were monitored for up to 28 days after challenge for survival and signs of disease, including weight loss.

Enzyme-Linked Immunosorbent Assays

Human IgG and guinea pig IgG serum levels were quantified using commercially available kits (ab195215 and ab190813, respectively; Abcam). Binding of AAV-expressed MR191 variants to their MARV GP target was determined using half-area 96-well plates (Corning) coated with 100 ng/well in 30 μL of MARV GP (IBT Bioservices) in PBS overnight at 4°C. Plates were washed 3 times with 0.2% PBS-Tween (PBS-T) and blocked using SuperBlock buffer (Thermo Fisher Scientific) for 30 minutes. Serum samples diluted 1:100 were plated and incubated at 37°C for 1 hour and then washed 3 times with 0.1% PBS-T. Donkey anti-human HRP-conjugated secondary antibody (ab102410; Abcam) or goat anti-guinea pig HRP-conjugated secondary antibody (ab6908; Abcam) diluted 1:50 000, were added to each well, and incubated at 37°C for 1 hour followed by 3 washes with 0.2% PBS-T. Plate was developed with 3,3',5,5'-tetramethylbenzidine (TMB) substrate (Thermo Scientific Pierce) for 15 minutes at room temperature, after which 2 M sulfuric acid was added. Optical density (OD) values (450 nm) were graphed after subtraction of the mean OD of the negative control.

Statistical Analyses

All graphs were generated using GraphPad Prism 9, and all statistical tests were performed using the same software. Survival of guinea pigs treated with AAV-MR191-io2 was compared to the control group using the log-rank (Mantel-Cox) test with the Bonferroni correction for multiple comparisons. A 1-way ANOVA and Tukey multiple comparison test were used to compare prechallenge serum human IgG concentrations between groups. P values less than ≤ 0.05 were marked with an asterisk *, ≤ .01 **, ≤ .001 ***, ≤ .0001 ****.

RESULTS

AAV-Mediated Delivery of MR191 Variants to Mice Provides Sustained mAb Expression

We previously demonstrated that 1 × 10¹¹ and 1 × 10¹⁰ vg doses of AAV6.2FF expressing full-length human IgG MR191 given IM provided 100% survival in BALB/c mice when administered 28 days prior to challenge [11]. To advance our MARV-specific AAV platform in a larger animal model, we sought to investigate the efficacy of AAV6.2FF-MR191 in Hartley guinea pigs. In an effort to increase transgene expression in vivo, we first synthesized 3 additional gene variants of MR191: MR191 codon optimized for guinea pigs (MR191-GP_op); MR191 containing two inflammation-inhibiting oligonucleotide 2 (io2) [14] sequences (MR191-io2); and MR191 containing constant domains of guinea pig origin (MR191-GP_HL) (Figure 1).

To evaluate AAV6.2FF-mediated expression of the MR191 variants in vivo, groups of 6-week-old female BALB/c mice (n = 4) were IM administered 1 × 10¹¹ vg of AAV6.2FF-MR191, AAV6.2FF- MR191-GP_op, AAV6.2FF- MR191-io2, or AAV6.2FF- MR191-GP_HL. Blood collection occurred weekly on days 0, 7, 14, 21, and 28 after AAV administration, followed by quantification of vectorized mAbs via enzyme-linked immunosorbent assay (ELISA; Figure 2). Levels of MR191 expression increased over the course of the experiment, with peak levels of transgene expression reached at day 28 in most animals. At 28 days postadministration, AAV6.2FF-MR191 resulted in plasma human IgG concentrations of 9.6–15.1 μg/mL, AAV6.2FF-MR191-io2 resulted in concentrations of 11.6–29.3 μg/mL, and AAV6.2FF-MR191-GP_op resulted in concentrations of 22.6–30.2 μg/mL. Interestingly, AAV6.2FF-MR191-GP_HL produced plasma guinea pig IgG concentrations of 97.6–131.2 μg/mL, which was 5.6 to 9.2 times greater than what was observed from the other vectors.

Figure 2.

Kinetics of vectorized MR191 variants in mice. Six-week-old female BALB/c mice (n = 4) were administered 1 × 10¹¹ vector genomes of (A) AAV6.2FF-MR191, (B) AAV6.2FF-MR191-GP_op, (C) AAV6.2FF-MR191-io2, or (D) AAV6.2FF-MR191-GP_HL. Human (A–C) or guinea pig (D) IgG levels in the plasma were monitored weekly via ELISA. Data from individual mice are plotted. Abbreviations: AAV, adeno-associated virus; ELISA, enzyme-linked immunosorbent assay; IgG, immunoglobulin G.

AAV6.2FF-MR191 Variants Provide Complete Protection from MARV in a Pilot Study

To assess the long-term expression kinetics and GP binding activity of the 4 MR191 variants in vivo, we immunized 4 groups of 3 guinea pigs with 5 × 10¹² vg of AAV6.2FF-MR191, AAV6.2FF-MR191-io2, AAV6.2FF-MR191-GP_op, or AAV6.2FF-MR191-GP_HL IM. Blood was collected on days 0, 14, 28, 56, and 90 after vector administration for expression analysis. The binding kinetics of MR191 variants were evaluated using a MARV GP ELISA. As shown in Figure 3A and 3B, all 4 MR191 variants retained their MARV GP binding activity as demonstrated by the increasing OD values of the collected serum samples over time, concomitantly aligning with the escalating levels of serum mAb expression (Figure 3C). Variability in the efficiency of the IM injection may have played a role in the reduced OD values observed in the serum samples collected from 1 of the guinea pigs treated with MR191-GPop. In instances where a guinea pig did not receive the complete AAV dose, it could potentially result in slower mAb expression kinetics and subsequently lower levels of mAb in the serum. This reduction in serum mAb levels could, upon dilution to 1:100, fall below the detectable threshold in the assay. Serum concentrations for the 3 MR191 variants containing a human IgG1 constant domain were also quantified using a human IgG ELISA. As was observed in the MARV GP binding ELISA, these MR191 variants displayed a similar pattern of mAb expression, with MR191-io2 being detected at consistently higher concentrations than MR191 and MR191-GP_op, plateauing at day 56 post-AAV with peak serum concentrations >400 μg/mL (Figure 3C).

Figure 3.

Kinetics of vectorized MR191 variant expression in guinea pigs. Guinea pigs (n = 3/group) were intramuscularly administered 5 × 10¹² vector genomes of AAV6.2FF-MR191, AAV6.2FF-MR191-io2, AAV6.2FF-MR191-GP_op, or AAV6.2FF-MR191-GP_HL and monitored for 90 days. Serum samples were obtained from all animals on 14, 28, 56, and 90 days after AAV administration and used to quantify Marburg virus glycoprotein-specific human (A) or guinea pig (B) IgG levels. Total human IgG levels for 3 of the AAV variants encoding a human IgG1 constant domain were also quantified in the same serum samples (C). Data from individual guinea pigs are plotted. Abbreviations: AAV, adeno-associated virus; IgG, immunoglobulin G; OD, optical density.

To demonstrate the protective efficacy of the vectorized MR191 variants against MARV infection, all the animals were challenged with a lethal dose of guinea pig-adapted MARV (GPA-MARV) 92 days after vector administration. Remarkably, all animals in each of the 4 groups survived infection (Figure 4A). Although some animals experienced limited weight loss (Figure 4B), no serious signs of disease were observed in any of the animals. Not only do these results demonstrate long-term AAV-mediated expression of the MR191 variants at relatively high levels, but they also confirm the ability of all 4 of these MR191 variants to protect against MARV infection in guinea pigs.

Figure 4.

AAV6.2FF-mediated expression of MR191 variants provides complete protection against MARV. Guinea pigs (n = 3/group) were administered 5 × 10¹² vector genomes of AAV6.2FF-MR191, AAV6.2FF-MR191-io2, AAV-MR191-GP_op, or AAV-MR191-GP_HL. All animals were challenged with 1000 LD₅₀ GPA-MARV on day 92 after AAV administration and monitored for survival (A) and weight loss (B). Data from individual animals are plotted in (B). Abbreviations: AAV, adeno-associated virus; GPA-MARV, guinea pig-adapted Marburg virus; LD₅₀, median lethal dose.

AAV6.2FF-MR191-io2 Vector Protects Guinea Pigs in a Dose-Dependent Manner

We next evaluated the protective efficacy of AAV6.2FF-MR191-io2 against MARV in a dose-response study in guinea pigs. We chose to move forward with this particular AAV construct for 3 reasons: (1) AAV6.2FF-MR191-io2 resulted in the highest levels of serum IgG at almost every time point postadministration in our guinea pig pilot experiment (Figure 3C), albeit with somewhat lower GP binding activity (Figure 3A); (2) this construct preserved the human IgG coding sequence of MR191, making it feasible to quantify the amount of MR191 expressed in guinea pig serum samples; and finally, (3) MR191-io2 is the optimal candidate to be advanced through the clinical translation pipeline and evaluated as a prophylactic in both NHPs and humans because the reduced inflammation mediated by the io2 sequence [14] may lower the potential for antidrug antibody responses, as has been observed in a majority of NHP and human clinical trials for AAV VIP [16]. Four groups of 6 animals were IM administered AAV6.2FF-MR191-io2 at a dose of 5 × 10¹² vg (high dose), 1 × 10¹² vg (medium dose), or 5 × 10¹¹ vg (low dose). A group of control animals were administered a vector expressing an irrelevant antibody (FluA20) at 5 × 10¹² vg. At day 28 postadministration, all animals were challenged with 1000 LD₅₀ GPA-MARV. As expected, all control animals succumbed to MARV infection by 12 days postchallenge following dramatic weight loss (Figure 5A and 5B). In contrast, the animals that received the high dose of AAV6.2FF-MR191-io2 were completely protected against MARV infection (Figure 5A), remaining healthy and gaining weight throughout the study (Figure 5B). Of the 6 animals that received the medium dose, 5 survived (approximately 83%) and exhibited no weight loss or clinical signs of disease (Figure 5A and 5B). Only 2 of the 6 (approximately 33%) animals that received the low dose survived, with all but 1 animal showing pronounced weight loss and signs of disease (Figure 5A and 5B). These results confirm the efficacy of AAV6.2FF-MR191-io2 in guinea pigs and demonstrate a clear dose-dependent effect.

Figure 5.

AAV6.2FF-MR191-io2 protects against MARV in a dose-dependent manner. Guinea pigs (n = 6/group) were administered AAV-MR191-io2 at a dose of 5 × 10¹² vg (high), 1 × 10¹² vg (medium), or 5 × 10¹¹ vg (low). A control group (n = 6) was administered 5 × 10¹² vg of an AAV vector expressing an irrelevant monoclonal antibody (AAV6.2FF-FluA20). At 28 days after AAV administration, all animals were challenged with 1000 LD₅₀ GPA-MARV and monitored for survival (A) and weight change (B). Data from individual animals are plotted in (B). Serum concentrations of human IgG (MR191 or FluA20) were quantified immediately prior to challenge (C). Data represented as the mean ± standard deviation. ** P ≤ .01, *** P ≤ .001. Nonsignificant statistical comparisons were not labelled. Abbreviations: AAV, adeno-associated virus; GPA-MARV, guinea pig-adapted Marburg virus; IgG, immunoglobulin G; LD_50, median lethal dose; vg, vector genome.

The dose-dependent efficacy of AAV6.2FF-MR191-io2 was mirrored by the expression levels of MR191 in the guinea pig serum obtained just prior to MARV challenge (Figure 5C). With a mean MR191 level of 147 µg/mL, the high dose of AAV6.2FF-MR191-io2 resulted in antibody expression that was significantly greater than both the medium-dose group (51 µg/mL) and the low-dose group (45 µg/mL). The level of FluA20 detected in the control animals was also significantly lower than the high dose of MR191 despite the fact that both AAVs were administered at the same dose. This result may be linked to the absence of io2 sequence in AAV6.2FF-FluA20, which has been shown to enhance transgene expression [14].

DISCUSSION

AAV-based gene therapy has achieved a number of important milestones over the last several years, with at least 4 products reaching clinical licensure and many more currently awaiting regulatory approval [17–19]. VIP using AAVs to drive expression of monoclonal antibodies has also made encouraging progress [20], particularly in the context of human immunodeficiency virus (HIV) infection [21, 22]. Thus, given the safety and efficacy of AAV-mediated gene expression, we sought to advance the development of AAV VIP against filoviruses, for which vaccines and therapeutics are still acutely needed. We have previously demonstrated the efficacy of several different AAV-vectored antibodies against Ebola virus in mouse models [12, 23], and, more recently, we demonstrated the efficacy of AAV-vectored MR191 against MARV in mice [11]. Furthermore, we have also shown that AAV-mediated antibody expression can persist over long periods of time in mice and sheep, and we demonstrated that this antibody expression does not impact the endogenous humoral response to other viral pathogens [11, 12, 16]. In the present study, we have continued to advance the development of AAV VIP against filoviruses by demonstrating the remarkable efficacy of AAV-mediated MR191 expression against MARV in guinea pigs. Notably, this represents the first such demonstration of AAV VIP efficacy against a filovirus in the guinea pig model, which is typically considered to be highly predictive of efficacy in humans and NHPs [24].

We evaluated 4 different MR191 variants in this study, including 2 that were modified to accommodate expression in guinea pigs (AAV6.2FF-MR191-GP_op and AAV6.2FF-MR191-GP_HL) and 1 that was modified to inhibit the immunogenicity of AAV itself (AAV6.2FF-MR191-io2). While all 4 variants performed equally well in a head-to-head pilot evaluation, demonstrating complete protection against MARV approximately 3 months after administration, AAV6.2FF-MR191-io2 possessed 2 significant advantages that made it the obvious choice for further evaluation. Namely, this construct not only preserved the human IgG sequences of the original MR191, making comparisons to other studies more feasible and strengthening the predictive efficacy of our study, but it also demonstrated uniformly enhanced expression kinetics thanks to the inclusion of an immune-modulating factor.

The AAV DNA genome has been shown to activate TLR9, thus eliciting an immune response that can reduce transduction efficiency [25, 26]. To counter this effect, we added to the vector genome an inflammation-inhibiting oligonucleotide 2 (io2) sequence that has been previously shown to prevent TLR9 activation, reduce the innate immune and T-cell responses to AAV administration, and thereby enhance transgene expression [14]. Indeed, the io2 sequence did appear to enhance antibody expression kinetics over the unmodified vectors, as assessed through the quantification of human IgG in transduced guinea pigs (Figure 3C). Interestingly, the levels of human IgG did not closely parallel the levels of MARV GP-specific IgG (Figure 3A). While the reasons for this discrepancy are not known, it is conceivable that this could be an artifact of the mouse model. Because MR191, MR191-io2, and MR191-GP_op are all composed of a human IgG constant domain, it is possible the mice mounted an immune response to the human IgG domain, which may have led to a reduction in mAb expression levels. Conversely, MR191-GP_HL contains a guinea pig IgG constant domain that, being of rodent origin, may have been less immunogenic in mice. In any case, AAV6.2FF-MR191-io2 exhibited remarkable therapeutic efficacy in our dose-response study, with the high dose offering complete protection against MARV and the medium dose offering near-complete protection. These data clearly demonstrate the efficacy of AAV VIP against a filovirus in a stringent animal model, and they position AAV6.2FF-MR191-io2 as a therapeutic worthy of continued preclinical development.

The frequency with which MARV has emerged to cause outbreaks over the past 3 years is alarming. The fact that we still lack a clinically approved vaccine or therapeutic with which to address these outbreaks is doubly alarming. AAV VIP may offer a tangible solution for quickly conferring a high-quality and protective immune response against filoviruses such as MARV. Unlike passive administration of monoclonal antibodies, AAV VIP can offer longer lasting immunity at a lower cost, and, unlike traditional vaccination, AAV VIP can offer prophylactic protection even in those individuals who are immunocompromised. Thus, AAV VIP in general—and AAV6.2FF-MR191-io2, in particular—warrants further development.

Notes

Author contributions. L. B. and S. K. W. conceived the study. A. D. R., W. C., S. H., J. A. L., N. Z., Y. P., B. T., L. B., and S. K. W. performed the experiments. A. D. R., W. C., L. B., and S. K. W. analyzed the data. W. C., L. B., A. D. R., and S. K. W. wrote the manuscript, which all authors have read and support.

Financial support. This work was supported by the Public Health Agency of Canada; and the Canadian Institutes of Health Research (grant number PJ4 179807).

Supplement sponsorship. This article appears as part of the supplement “10^th International Symposium on Filoviruses.”