SARS-CoV-2 infection, neuropathogenesis and transmission among deer mice: Implications for spillback to New World rodents

Abstract

Coronavirus disease-19 (COVID-19) emerged in late 2019 in China and rapidly became pandemic. As with other coronaviruses, a preponderance of evidence suggests the virus originated in horseshoe bats (Rhinolophus spp.) and may have infected an intermediate host prior to spillover into humans. A significant concern is that SARS-CoV-2 could become established in secondary reservoir hosts outside of Asia. To assess this potential, we challenged deer mice (Peromyscus maniculatus) with SARS-CoV-2 and found robust virus replication in the upper respiratory tract, lungs and intestines, with detectable viral RNA for up to 21 days in oral swabs and 6 days in lungs. Virus entry into the brain also occurred, likely via gustatory-olfactory-trigeminal pathway with eventual compromise to the blood-brain barrier. Despite this, no conspicuous signs of disease were observed, and no deer mice succumbed to infection. Expression of several innate immune response genes were elevated in the lungs, including IFNα, IFNβ, Cxcl10, Oas2, Tbk1 and Pycard. Elevated CD4 and CD8β expression in the lungs was concomitant with Tbx21, IFNγ and IL-21 expression, suggesting a type I inflammatory immune response. Contact transmission occurred from infected to naive deer mice through two passages, showing sustained natural transmission and localization into the olfactory bulb, recapitulating human neuropathology. In the second deer mouse passage, an insertion of 4 amino acids occurred to fixation in the N-terminal domain of the spike protein that is predicted to form a solvent-accessible loop. Subsequent examination of the source virus from BEI Resources determined the mutation was present at very low levels, demonstrating potent purifying selection for the insert during in vivo passage. Collectively, this work has determined that deer mice are a suitable animal model for the study of SARS-CoV-2 respiratory disease and neuropathogenesis, and that they have the potential to serve as secondary reservoir hosts in North America.

Fig 1. Infection and antibody response to SARS-CoV-2.

(A) Viral RNA was detected in lungs of infected deer mice to day 6, and (B) virus was recovered only on day 3. (C) Two of three deer mice had detectable vRNA in their olfactory bulbs and (D) virus was isolated from each. (E) IgG antibodies were detected to nucleocapsid protein on day 14, (F) with neutralizing antibody detected on days 6 and 14. Each time point represents samples collected from 3 euthanized deer mice per group. Error bars represent the standard deviation of the mean and geometric mean antibody titers.

Fig 2. Immune gene expression in lungs of infected deer mice.

Gene profiling showed elevated expression of several antiviral innate immune response genes 3 and 6 dpi (A-H) and evidence of T cell infiltration and inflammatory type I immune response 6 dpi (I-S).

Fig 3. Histopathology and immunohistochemistry of SARS-CoV-2 in skull and brain of deer mice at 3- and 6- days post-infection.

(A) Acute fibrinosuppurative and ulcerative sinusitis in ethmoturbinates with degeneration and inflammation of branches of olfactory, ethmoidal, and maxillary sensory nerves (fibrinoid vascular necrosis–inset) (B) Transmural SARS-CoV-2 immunoreactivity (Fast Red staining) in MOE, 3 dpi. (C) Prominent immunoreactivity in trigeminal ganglionic neurons with mild glial reaction at 6 dpi. (D) Disruption of the BBB with associated histioneutrophilic meningoencephalitis, 6 dpi (E) Viral transmission to the glomerular layer of the MOB, 6 dpi. (F) Immunofluorescence imaging depicting entry of centrifugal afferents to the olfactory bulb at the cribriform plate in uninfected control deer mice 6 dpi showing no viral immunoreactivity in any neurons or glial cells. Neurons were identified with antibodies against microtubule associated protein (MAP2, green), microglia with anti-ionized calcium-binding adaptor molecule 1 (IBA1, cyan), anti-SARS-CoV-2 (red) and nuclei were counterstained with DAPI (blue). (G) Multifocal SARS-CoV-2 antigen was detected by 6 dpi within neurons and microglia of the afferent nerves and in the glomerular layer of the olfactory bulb (arrows; 100X high-magnification inset microglial cell). Examination of the trigeminal nerve and ganglion in (H) uninfected control with no immunoreactivity and (I) infected deer mice revealing neuroinvasion of SARS-CoV-2, with extensive co-localization of the virus within MAP2⁺ neurons proximal to activated microglia, 6 dpi. 100X high-magnification insets depict co-localization of SARS-CoV-2 with MAP2⁺ neurons (top panel, arrows), as well as the same image without MAP2 to better highlight the extent of neuronal staining with SARS-CoV-2. SARS-CoV-2 quantification by fluorescence intensity within given regions of interest (ROI) are represented in the olfactory bulb (J), nasal turbinates (K), and trigeminal nerve (L) showing the increase in viral load. Scale bars equal 100 μm (n = 2/group) *p<0.0332, **p<0.0021, ***p<0.0002, ****p<0.0001.

Fig 4. Abundance of viral RNA in oral swabs and weight loss in deer mice infected with SARS-CoV-2.

(A) Probe-based PCR was used to detect levels of E gene RNA and quantified against an E gene plasmid standard (Integrated DNA Technologies). Transient weight loss occurred in some deer mice inoculated with SARS-CoV-2. (B) Virus isolation from oral swabs of infected deer mice. *P2 contact deer mice were euthanized on day 10 to prepare P2 virus stock. Deer mice were weighed each day during the transmission study (C). All three inoculated deer mice, one passage 1 deer mouse (DM6) and both passage 2 (DM8, DM9) deer mice lost weight followed by recovery.

Fig 5. Spike protein insertion after serial passage in deer mice.

(A) Passage of SARS-CoV-2 led to fixation of a 4-residue insertion in the N-terminal domain. Oral swabs of both DM8 and DM9 passage 2 deer mice (P2-8, P2-9, respectively) were submitted to RNA-Seq and an insertion was detected in all reads spanning residues 216–219 (nt 22,208–22,209, hatches) of the spike protein, implying a strong purifying selection for the insert during passage. (B) Homology model of the spike protein-human ACE2 complex showing the location of the KLRS insertion observed in the P2 virus. Model was generated by threading the deer mouse P2 spike protein sequence into the EM structure of the open-state SARS-CoV-2 spike protein (PDB: 6VYB) and colored as follows: Core (grey), N-terminal domain (NTD, orange) with KLRS insert as magenta spheres and native loop residues as yellow spheres, interdomain linker (blue), receptor binding domain (RBD, cyan for one up-conformation copy, green for two down-conformation copies), SD1/SD2 domains (light blue). The structure of the human ACE2 receptor (brown) was then superposed on the up-confirmation RBD using the crystal structure of the RBD-ACE2 complex (PDB: 6VW1).