Remdesivir (GS-5734) Is Efficacious in Cynomolgus Macaques Infected With Marburg Virus

Abstract

Marburg virus (MARV) is a filovirus with documented human case-fatality rates of up to 90%. Here, we evaluated the therapeutic efficacy of remdesivir (GS-5734) in nonhuman primates experimentally infected with MARV. Beginning 4 or 5 days post inoculation, cynomolgus macaques were treated once daily for 12 days with vehicle, 5 mg/kg remdesivir, or a 10-mg/kg loading dose followed by 5 mg/kg remdesivir. All vehicle-control animals died, whereas 83% of animals receiving a 10-mg/kg loading dose of remdesivir survived, as did 50% of animals receiving a 5-mg/kg remdesivir regimen. Remdesivir-treated animals exhibited improved clinical scores, lower plasma viral RNA, and improved markers of kidney function, liver function, and coagulopathy versus vehicle-control animals. The small molecule remdesivir showed therapeutic efficacy in this Marburg virus disease model with treatment initiation 5 days post inoculation, supporting further assessment of remdesivir for the treatment of Marburg virus disease in humans.

Marburg virus, filovirus, GS-5734, nonhuman primate, remdesivir

Filoviruses, including Marburg (MARV) and Ebola (EBOV) viruses, are highly virulent zoonotic pathogens that cause hemorrhagic fever disease in humans with some of the highest case-fatality ratios of any infectious diseases in the world. The African fruit bat Rousettus aegyptiacus is a natural reservoir for MARV [1, 2]. In humans, infection can occur via exposure to areas inhabited by fruit bat colonies or transmitted by infected primates or humans [1]. Marburg virus disease (MVD) was initially described in 1967 after European laboratory workers developed disease following exposure to infected African green monkeys [3]. Outbreaks of MVD occur periodically, mostly across sub-Saharan Africa, with the largest outbreaks comprising 154 cases in the Democratic Republic of the Congo (1998–2000) and 252 cases in Angola (2004–2005) [3]. The most recent outbreak occurred in Uganda in 2017 and comprised 4 cases [4]. Case-fatality rates of MVD have been as high as 90% [5]. MARV is on the 2018 World Health Organization’s Blueprint list of priority diseases with high potential to cause a public health emergency [6].

Clinical symptoms of MVD are similar to Ebola virus disease. After a 2- to 21-day incubation period, MARV-infected humans typically have rapid onset of nonspecific symptoms of severe headache, malaise, muscle aches, and high fever, followed by rash, conjunctivitis, diarrhea, and vomiting within 5 days after the onset of symptoms [5]. The disease then progresses into the early organ dysfunction phase (days 5–13 after symptom onset), where patients may develop dyspnea, abnormal vascular permeability, hemorrhagic manifestations, and neurologic manifestations, including encephalopathy, seizures, or focal neurologic abnormalities [5, 7]. The late stages of fatal disease involve severe dehydration, diffuse coagulopathy, shock, and multiorgan failure resulting in death 8–16 days after the onset of symptoms [5]. No licensed vaccines or therapeutics are currently available to counter this serious public health threat, and supportive treatment is the current standard of care.

In nonhuman primates, MARV causes pathological features similar to those experienced by humans, and exposure in experimental settings produces uniform mortality [5, 7, 8]. The cynomolgus monkey is among the most sensitive nonhuman primate species and is often used for filovirus research [7]. Monkeys infected with MARV variant Angola 2005 have more rapid and severe disease progression as well as higher fatality than infected humans [8], thus setting a high bar for the evaluation of MARV therapeutics.

Several investigational therapies for the treatment of MVD are under development and have been evaluated in the nonhuman primate model of MVD. Favipiravir and galidesivir (BCX4430) are small molecule nucleoside analogs with broad spectrum in vitro activity that have demonstrated a survival benefit in MARV-infected nonhuman primates when treatment was initiated at the time of infection or 48 hours after virus exposure, respectively [9, 10]. MR191, a monoclonal antibody directed against the MARV glycoprotein, as well as a cocktail of MR191 with the broadly neutralizing EBOV monoclonal antibodies FVM04 and CA45, have shown increased survival in MARV-infected nonhuman primates when treatment was initiated up to 5 days after virus exposure [11, 12]. A nucleoprotein-targeting short interfering RNA (siRNA) also showed increased survival in MARV-infected nonhuman primates when treatment was initiated up to 5 days after virus exposure [13, 14].

The small-molecule antiviral compound remdesivir (GS-5734) is a monophosphoramidate prodrug of an adenosine analog that exhibits broad-spectrum in vitro antiviral activity against filoviruses including MARV, EBOV, and Sudan virus, as well as unrelated RNA viruses such as coronaviruses and paramyxoviruses [15–17]. As demonstrated for EBOV, remdesivir inhibits viral transcription and replication by targeting the viral RNA-dependent RNA polymerase and causing delayed chain termination following the incorporation of its triphosphate metabolite into the newly synthesized viral RNA [18]. Studies in nonhuman primate models of EBOV and Nipah virus infection, as well as mouse models of MERS and SARS coronaviruses, have demonstrated the therapeutic efficacy of remdesivir against viruses with the potential for high-consequence human outbreaks [15, 17, 19–21]. In this study, the therapeutic efficacy of remdesivir against MARV variant Angola 2005 (MARV-Angola) was evaluated in cynomolgus monkeys, with treatment initiation occurring 4 or 5 days after virus exposure.

METHODS

Ethics Statement

Animal research at the US Army Medical Research Institute of Infectious Diseases (USAMRIID) was conducted in Animal Biosafety Level 4 (ABSL-4) under an institutional animal care and use committee-approved protocol in compliance with the Animal Welfare Act, Public Health Service policy, and other federal statutes and regulations relating to animals and experiments involving animals. The facilities where this research was conducted are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International and strictly adhere to principles stated in the Guide for the Care and Use of Laboratory Animals [22].

Animals and Study Design

Twenty-four healthy cynomolgus macaques (Macaca fascicularis), 12 males and 12 females (age range 4–8 years), were obtained from Worldwide Primates. Animals were housed individually in stainless steel cages with enrichment toys and were provided fruit and monkey chow biscuits as well as water ad libitum. Animals were anesthetized for blood collection and treatment using ketamine (100 mg/mL)/acepromazine (10 mg/mL) at 0.1 mL/kg, administered intramuscularly (IM). Each animal was randomly assigned to 1 of 4 groups. Each group consisted of 6 animals (3 males and 3 females). On day 0, each animal was administered an IM injection of 0.5 mL USAMRIID MARV challenge stock R4410.

Three groups were administered remdesivir (lot 5734-BC-IP) in vehicle (water with 12% sulfobutylether-β-cyclodextrin [SBE-β-CD]) once daily for 12 days by slow bolus intravenous (IV) injection, beginning 4 or 5 days after virus challenge (Figure 1). The remdesivir dosing regimens were based on those that had previously demonstrated promising efficacy in studies of nonhuman primate models of Ebola virus disease [23]. Two of the groups received a single 10-mg/kg loading dose of remdesivir on day 4 or 5 followed by once-daily 5-mg/kg maintenance doses for 11 days (10/5 mg/kg d4 and 10/5 mg/kg d5, respectively). The third group received once-daily 5-mg/kg doses of remdesivir on days 5–16 (5 mg/kg d5). The control group received vehicle (12% SBE-β-CD) by IV injection on days 4–16. To maintain experimental blinding, animals not scheduled to receive remdesivir on day 4 or day 17 were administered vehicle (Figure 1). All MARV-infected animals were handled in a biosafety level 4 (BSL-4) laboratory at USAMRIID in Frederick, Maryland.

Figure 1.

Experimental design for remdesivir efficacy evaluations in cynomolgus monkeys. Abbreviations: MARV, Marburg virus; No T_x, no treatment.

After viral challenge, all macaques were monitored closely for signs of clinical disease and were assigned a score of physical responsiveness ranging from 0 (active) to 4 (moderate to severe unresponsiveness, requires prodding, moderate prostration). Macaques that survived until day 41 were considered to have survived the viral challenge. Study personnel alleviated unnecessary suffering of infected animals by euthanizing clinically moribund animals according to criteria described previously [10].

Blood samples were collected from each animal under anesthesia via peripheral venipuncture on days 0, 3, 4, 5, 6, 7, 8, 10, 14, 21, 28, 35, and 41 post inoculation.

Challenge Agent

The virus used in this study was Marburg marburgvirus H. sapiens-tc/ANG/2005/Angola-200501379 (order Mononegavirales, family Filoviridae, species Marburg marburgvirus). The stock material used was USAMRIID R4410 MARV passage 3. MARV (designated 200501379) was isolated from an 8-month-old female infant during an outbreak occurring in 2005 in Angola. The patient exhibited disease, was hospitalized, and died. The first passage of virus (designated virus seed pool 810820) was conducted at the Centers for Disease Control and Prevention (CDC) using Vero E6 cells. A second passage of virus (designated WRC000041) was conducted at the University of Texas Medical Branch. WRC000041 was transferred to USAMRIID and propagated on BEI Resources Vero E6 cells to produce the USAMRIID master seed stock, Lot R4410. Each animal was exposed to a target dose of 1000 plaque-forming units (PFU) MARV diluted in 0.9% sodium chloride for injection, United States Pharmacopeia (USP). The calculated dose of virus administered, based on plaque assay of virus challenge stock, was 345 PFU.

Remdesivir

Remdesivir was synthesized at Gilead Sciences Inc.; chemical identity and sample purity were established using nuclear magnetic resonance, high-resolution mass spectrometry, and high-performance liquid chromatography analyses [15]. Small-molecule X-ray crystallographic coordinates and structure factor files have been deposited in the Cambridge Structural Database (http://www.ccdc.cam.ac.uk/); accession numbers have been supplied previously [15].

Reverse Transcription Quantitative PCR

Blood samples were subjected to quantitative and qualitative assessment of plasma viral RNA via validated reverse transcription quantitative PCR (RT-qPCR) methods. Total RNA was extracted from TRIzol LS-treated plasma samples using a MagMax 96 Blood RNA Isolation Kit (Ambion); purified RNA samples were stored at −80°C until assayed. The RT-qPCR assays were performed on samples in triplicate using the ABI 7500 Fast Dx. Quantified MARV genomic RNA was used to generate a standard curve. Quantification of genome copies per volume plasma were calculated from the cycle threshold (Ct) value obtained for each sample compared with the standard curve equation. The limits of quantitation for this assay are 8.0 × 10⁵–8.0 × 10¹¹ genomic equivalents (ge)/mL of plasma.

Qualitative assessment methods were used for all viral RNA samples for which RNA could be detected but could not be quantified. The limit of detection (LOD) Ct was defined in the context of these sample analyses as 38.28. A sample was considered positive when the Ct value was < 38.28. An animal was considered to have tested positive for detection of MARV RNA when a minimum of 2 of 3 replicates were positive.

MARV RNA values reported as less than LOD were substituted as 3 log₁₀ ge/mL; values reported as more than LOD to less than the lower limit of quantitation were imputed as 5.9 log₁₀ ge/mL; and values reported as more than the upper limit of quantitation were imputed as 11.90 log₁₀ ge/mL for computation purposes.

Clinical Pathology

Blood chemistry profiles were determined using a Vitros 350 Chemistry System (Ortho Clinical Diagnostics), hematology analysis was conducted using an Advia 120 Hematology Analyzer (Siemens), and coagulation analysis was performed using a Sysmex CA-1500 (Siemens).

Anatomic Pathology, Histopathology, and Immunohistochemistry

Animals were euthanized via pentobarbital, administered IV. A full necropsy was conducted on the carcass of each animal, and a complete set of tissue samples was also collected from each animal for analysis by routine histology and immunohistochemistry (IHC).

Tissues collected for histopathology and IHC were immersion-fixed in 10% neutral buffered formalin for a minimum of 28 days before removal from BSL-4 containment. The tissue samples were trimmed, routinely processed, and embedded in paraffin. Sections of the paraffin-embedded tissues 5 µm thick were cut for histology. The histology slides were deparaffinized and stained with hematoxylin and eosin.

Replicate sets of the slides produced for routine histology were made for IHC. An immunoperoxidase assay using a cocktail of 2 mouse monoclonal antibodies against MARV as the primary antibodies was completed on unstained slides of all tissue sections. These slides were then counterstained with hematoxylin.

Statistical Analyses

Animals were randomly assigned to experimental treatment groups, stratified by sex (with equal number of males and females per group) and balanced by body weight, using SAS statistical software. Statistical power analysis was used to predetermine sample size. Study personnel responsible for assessing animal health (including euthanasia assessment) and administering treatments were experimentally blinded to the group assignments of all animals in the study. Study data were analyzed using SAS version 9.4. The proportion of animals surviving at day 41 was evaluated using a Fisher exact test, adjusted for multiple comparisons using a Hochberg test. Kaplan-Meier survival analysis was used to calculate survival curves as well as median and mean survival times. Resulting survival curves were compared by log-rank (Mantel-Cox) with the Dunnett-Hsu procedure adjustment for multiple comparisons. Plasma viral RNA was compared across groups using analysis of variance (ANOVA), with Dunnett test to adjust for multiple comparisons. When sample sizes permitted, clinical pathology parameters were compared between groups using a Wilcoxon rank-sum test.

Data Availability

The authors declare that the data supporting the findings of this study are available within the article or are available from the corresponding author upon request.

RESULTS

All vehicle-control animals developed acute signs characteristic of MVD infection, such as fever and rash, and all succumbed or were euthanized between days 7 and 9 post inoculation after showing behavioral depression and deteriorating physical responsiveness (Figure 2A). Treatment with remdesivir was associated with a statistically significant increased survival of 83% in both of the 10/5 mg/kg groups, regardless of whether treatment was initiated on day 4 or 5 post virus exposure (P < .05 vs vehicle), and with a trend toward increased survival of 50% (not significant vs vehicle) in the 5 mg/kg group with treatment initiated on day 5 (Figure 2A).

Figure 2.

Remdesivir postexposure protection against Marburg virus in cynomolgus monkeys. A, Kaplan-Meier survival curves. *P < .05 for treatment vs vehicle groups assessed by Fishers exact test using the Hochberg test to adjust for multiple comparisons. B, Group mean of plasma viral RNA concentrations. *P < .05, **P < .01 for treatment vs vehicle groups assessed by analysis of variance using Dunnett test to adjust for multiple comparisons. C, Group mean clinical disease score. Remdesivir treatment groups and day initiated: black, vehicle; red, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 4; green, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 5; blue, 5 mg/kg maintenance dose, day 5. B and C, Error bars are omitted for clarity. Abbreviations: LLOQ, lower limit of quantitation; LOD, limit of detection.

Systemic (plasma) MARV RNA was detected in multiple animals on day 3 post inoculation, the first time point at which samples were collected for viral load analysis. In vehicle-control animals, plasma viral load increased rapidly, with mean values of 10.77 log₁₀ ge/mL obtained on day 6 and 11.45 log₁₀ ge/mL on day 8. Notably, all animals in which remdesivir was initiated on day 5 had detectable viral RNA levels at the time of treatment initiation, with values > 8 log₁₀ ge/mL in 6/12 animals (Supplementary Figure 1B and 1C, and Supplementary Table 1). Consistent with increased survival, treatment with remdesivir reduced MARV viral load, with significantly lower (P < .05 vs vehicle) plasma viral RNA levels in both 10/5 mg/kg remdesivir treatment groups compared with vehicle, beginning 1 day after initiation of treatment with remdesivir (Figure 2B, Supplementary Figure 1, and Supplementary Table 2). In surviving animals, plasma viral RNA levels were below the LOD by day 21 (Figure 2B, Supplementary Figure 1, and Supplementary Table 1).

Remdesivir generally ameliorated clinical signs (Figure 2C and Supplementary Figure 2) associated with MVD in cynomolgus monkeys compared with vehicle treatment. Vehicle-control animals exhibited a mean 1.3°C elevation in body temperature from baseline by day 6, and overt signs of disease were noted beginning on day 7 post virus inoculation, at which time all animals in this group exhibited behavioral depression and/or changes in posture. In comparison, all remdesivir treatment groups had statistically significantly reduced body temperature compared with vehicle-control animals at day 6 (mean 38.4°C for each remdesivir treatment group vs 39.2°C for vehicle-control animals; all P < .05). Furthermore, 7 of 18 of the remdesivir-treated animals exhibited either no clinical disease signs or only mild and transient signs, as indicated by assignment of responsiveness scores ≤ 1 at all observations for the study duration (Supplementary Figure 2).

Consistent with MVD disease course, vehicle-control animals had profound changes from baseline in hematology, coagulation, and serum chemistry parameters. Early changes from baseline indicative of an inflammatory response, including elevated C-reactive protein, fibrinogen, neutrophil counts, and reduced lymphocyte count, were apparent on or before day 5 in vehicle-control animals. These changes were ameliorated in most remdesivir-treated animals by day 8 (3–4 days after initiation of treatment with remdesivir). On day 8, antithrombin and d-dimer levels in all remdesivir-treated groups were statistically significantly improved versus the vehicle group (P < .05; Figure 3A and 3B, and Supplementary Table 3). Consistent with the highest rate of survival, the remdesivir regimens that included the 10-mg/kg loading dose initiated on day 4 or day 5 (10/5 mg/kg d4 and 10/5 mg/kg d5, respectively) were associated with the lowest levels of serum markers predictive of kidney and liver injury. On day 8, blood urea nitrogen (BUN) and creatinine levels were lower in the remdesivir-treated animals (all P < .05 vs vehicle except BUN in the 5 mg/kg d5 group; Figure 4A and 4B, and Supplementary Table 3). Alanine aminotransferase and aspartate aminotransferase levels were lower in animals that received a loading dose of 10 mg/kg remdesivir (P < .05 vs vehicle; Figure 4C and 4D, and Supplementary Table 3).

Figure 3.

Improvement in clinical pathology after remdesivir treatment in cynomolgus monkeys. Group mean (n = 6 per group) values of prothrombin time (A) and antithrombin (B). Remdesivir treatment groups and day initiated: black, vehicle; red, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 4; green, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 5; blue, 5 mg/kg maintenance dose, day 5. Error bars are omitted for clarity. *P < .05 for remdesivir vs vehicle using Wilcoxon rank-sum test without adjustment for multiple comparisons.

Figure 4.

Improvement in clinical pathology after remdesivir treatment in cynomolgus monkeys. Group mean (n = 6 per group) values of blood urea nitrogen (BUN, A), creatinine (B), alanine aminotransferase (ALT, C), and aspartate aminotransferase (AST, D). Remdesivir treatment groups and day initiated: black, vehicle; red, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 4; green, 10 mg/kg loading dose/5 mg/kg maintenance dose, day 5; blue, 5 mg/kg maintenance dose, day 5. Error bars are omitted for clarity. *P < .05 for remdesivir vs vehicle using Wilcoxon rank-sum test without adjustment for multiple comparisons.

All animals that died during the critical phase had similar microscopic findings that are consistent with MVD, including liver hepatocyte degeneration and necrosis; lymphoid depletion, lymphocytolysis, and fibrin deposition within the spleen; necrosis within the gastrointestinal tract; adrenal gland necrosis; lymphocytolysis in various lymph nodes; skeletal muscle degeneration; and necrosis at the inoculation site. With few exceptions, all assessed tissues were positive for MARV antigen by IHC. In surviving animals following terminal euthanasia on days 41–44, the most common microscopic findings were lymphoid hyperplasia in various lymph nodes and inflammation in various organs and tissues. Six of 13 surviving animals exhibited viral persistence, with at least 1 tissue sample positive for MARV antigen by IHC. While tissues were not assessed for presence of viral RNA or infectious virus, this is of considerable interest for future studies.

Discussion

These data extend the previously shown antiviral activity of remdesivir against MARV in cell-based assays [15] to demonstrate efficacy of remdesivir in an animal model that reproduces many of the hallmarks of MVD observed in infected humans. In this study, therapeutic administration of remdesivir to MARV-infected cynomolgus macaques up to 5 days after virus exposure resulted in a survival benefit, significant reductions in plasma viral load, and amelioration of disease signs. These results show that a small-molecule antiviral can be successfully initiated as late as 5 days after virus exposure and still elicit robust efficacy against lethal infection in this model.

Further nonhuman primate studies evaluating remdesivir efficacy when initiated later in disease, such as on day 6 post inoculation, may be helpful to inform the potential efficacy of remdesivir in infected humans, who are likely to seek to treatments only after onset of signs or symptoms. Recent results from a clinical trial conducted during an outbreak of Ebola virus disease in the Democratic Republic of the Congo demonstrated that many patients presented for treatment late in disease, and this may have resulted in the reduced efficacy of experimental therapies relative to nonhuman primate studies where treatment was initiated earlier in disease [24]. Furthermore, the utility of combining remdesivir treatment with a therapeutic that inhibits MARV via a different mechanism of action, such as the monoclonal antibody MR191, should be explored to see whether efficacy could be further improved, particularly when treatment is initiated later in the disease course. Given the limited opportunities to study the safety and efficacy of MARV therapeutics in humans, nonclinical studies are critical to be well positioned for the next outbreak of MVD and to inform the potential clinical testing of promising candidate therapeutics.

Remdesivir has broad-spectrum antiviral activity against a range of pathogenic RNA viruses, its large-scale manufacturing has been demonstrated, and phase 1 clinical studies in healthy human subjects have been completed. Remdesivir is also currently undergoing clinical evaluation for the treatment of COVID-19. The results of this study support continued assessment of remdesivir as a potential therapy for the treatment of MVD.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Presented in part: 9th International Symposium on Filoviruses, Marburg, Germany, 13–16 September 2017.

Acknowledgment. B. Norquist provided medical writing assistance on behalf of Gilead.

Author contributions. T. K. W. designed and supervised activities associated with efficacy evaluations, and interpreted study results. J. M. W. and L. G. coordinated study activities. J. W., K. W., N. G., S. V. T., G. D., and J. S. conducted the study and performed associated sample analyses. A. M. performed interpretations of clinical pathology results. J. B. and E. L. performed anatomic pathology examinations and histological analyses. C. B. conducted statistical analyses. D. P. P., R. B., R. J., T. C., and S. B. evaluated results and provided project oversight. T. K. W. and D. P. P. outlined and wrote the manuscript.

Disclaimer. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the US Army.

Financial support. This work was supported by the US Army Joint Program Executive Office for Chemical, Biological, Radiological, and Nuclear Defense Joint Project Manager for Chemical, Biological, Radiological, and Nuclear Defense Medical under (agreement number W911QY-16-9-0001).

Conflict of Interest Statement. D. P. P., R. B., C. B., R. J., and T. C. are current or former employees, and may be shareholders, of Gilead Sciences. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.