A hamster model for Marburg virus infection accurately recapitulates Marburg hemorrhagic fever

Abstract

Marburg virus (MARV), a close relative of Ebola virus, is the causative agent of a severe human disease known as Marburg hemorrhagic fever (MHF). No licensed vaccine or therapeutic exists to treat MHF, and MARV is therefore classified as a Tier 1 select agent and a category A bioterrorism agent. In order to develop countermeasures against this severe disease, animal models that accurately recapitulate human disease are required. Here we describe the development of a novel, uniformly lethal Syrian golden hamster model of MHF using a hamster-adapted MARV variant Angola. Remarkably, this model displayed almost all of the clinical features of MHF seen in humans and non-human primates, including coagulation abnormalities, hemorrhagic manifestations, petechial rash, and a severely dysregulated immune response. This MHF hamster model represents a powerful tool for further dissecting MARV pathogenesis and accelerating the development of effective medical countermeasures against human MHF.

Figure 1. Establishment of a lethal hamster model for Marburg virus.

(a,b) Seven groups of 4 hamsters each were inoculated intraperitoneally with 0.001 plaque-forming units (PFU) to 1000 PFU HA-MARV. (a) Survival of the animals is presented as a Kaplan-Meier survival curve. (b) Percentage weight loss expressed as the mean ± SEM. (c–i) Eight groups of 6 hamsters were inoculated intraperitoneally with 100 LD₅₀ (1 PFU) HA-MARV and 8 groups of 6 hamsters were inoculated with an equivalent dose of WT-MARV. Six hamsters from each group were sacrificed and samples collected each day post-infection, up to day 8. Six control hamsters were mock infected; 3 were sacrificed on day 1 and 3 on day 8. (c) Percentage weight loss over the course of the experiment of all animals sacrificed on day 8, expressed as the mean ± SEM. (d) Body temperatures over the course of the experiment of all animals sacrificed on day 8, expressed as the mean ± SEM. Two of the six animals were found dead on day 8, so n = 4 at this timepoint only. (e,f) A disseminated maculopapular/petechial rash was visible on hamsters 7 (e) and 8 (f) days post-infection. (g–i) Gross lesions revealed by necropsy included a large inguinal subcutaneous hemorrhage (arrowhead) after 7 days post-infection (g) swollen, soft, pale livers (#) and gastro-duodenal hemorrhaging (arrowhead) 7 days post-infection (h) and swollen, soft, pale spleens (#) and hemorrhaging of the adrenal cortex (arrowhead) 8 days post-infection (i). p-values ≤ 0.05 are indicated by a single asterisk (*) and p-values ≤ 0.001 are indicated by two asterisks (**), compared to WT-MARV.

Figure 2. HA-MARV induces hematologic and coagulation abnormalities.

(a–h) Blood samples were collected from sacrificed animals each day post-infection with HA- or WT-MARV and analyzed for prothrombin time (PT) (a) activated partial thromboplastin time (aPTT) (b), thrombin time (c) fibrinogen serum concentration (d) white blood cell (WBC) count (e) total neutrophil (NE) count (f) total lymphocyte (LY) count (g) and platelet count (h). All values are expressed as the mean ± SEM. n = 3–6 for the infected groups, n = 3 for the control group. p-values ≤ 0.05 are indicated by a single asterisk (*) and p-values ≤ 0.001 are indicated by two asterisks (**), compared to WT-MARV.

Figure 3. HA-MARV replicates robustly in hamsters.

(a–d) Virus titers, expressed as tissue culture infectious dose 50% (TCID₅₀) on a log10 scale, were calculated from the blood (a) the mesenteric lymph nodes (b) the livers (c) and the spleens (d). Mean values for each time point are indicated by the line graph, and individual values for each hamster are indicated by blue (WT-MARV) or red (HA-MARV) dots. n = 3–6. p-values ≤ 0.05 are indicated by a single asterisk (*) and p-values ≤ 0.001 are indicated by two asterisks (**) compared to WT-MARV. (e,f) Immunohistochemistry (IHC) to detect MARV antigen was performed on liver (e) and spleen (f) samples collected 3 or 6 days post-infection with either WT- or HA-MARV. A focus of virus antigen in the liver of a HA-MARV-infected hamster at day 3 is indicated by a white arrowhead. Note that the WT-MARV-infected hamster randomly selected for IHC analysis exhibited no detectable virus titer in the spleen and a very low titer in the liver, which correlates with the absence of antigen detected by IHC in the spleen and the small amount detected in the liver.

Figure 4. HA-MARV induces pathological changes in the liver and spleen.

(a,b) Hematoxylin and eosin stained liver (a) and spleen (b) samples collected 3 or 6 days post-infection with either WT- or HA-MARV. In the liver (a) hepatocellular necrosis is indicated by white arrowheads and neutrophilic infiltration is indicated by red arrows. In the spleen (b) tingible body macrophages are indicated with yellow arrows and necrosis and loss of lymphocytes is indicated by asterisks.

Figure 5. HA-MARV induces an early and strong innate immune response.

(a–c) Transcript levels for the indicated genes were quantified by RT-qPCR from samples derived from the livers, spleens, and blood of animals infected with WT-MARV (blue lines) or HA-MARV (red lines). Data are expressed as the mean ± SEM of the fold change over uninfected control hamsters on a log10 scale. n = 5–6 for the liver samples and 4–6 for the spleen samples. For the blood samples, n = 3–6, with the exception of HA-MARV day 7, for which n = 1. p-values ≤ 0.05 are indicated by a single asterisk (*) and p-values ≤ 0.001 are indicated by two asterisks (**) compared to WT-MARV. The grey boxes depict the range of transcript expression observed in uninfected control hamsters, n = 6.