Modeling pathogenesis of emergent and pre-emergent human coronaviruses in mice

Abstract

The emergence of highly pathogenic human coronaviruses (hCoVs) in the last two decades has illuminated their potential to cause high morbidity and mortality in human populations and disrupt global economies. Global pandemic concerns stem from their high mortality rates, capacity for human-to-human spread by respiratory transmission, and complete lack of approved therapeutic countermeasures. Limiting disease may require the development of virus-directed and host-directed therapeutic strategies due to the acute etiology of hCoV infections. Therefore, understanding how hCoV-host interactions cause pathogenic outcomes relies upon mammalian models that closely recapitulate the pathogenesis of hCoVs in humans. Pragmatism has largely been the driving force underpinning mice as highly effective mammalian models for elucidating hCoV-host interactions that govern pathogenesis. Notably, tractable mouse genetics combined with hCoV reverse genetic systems has afforded the concomitant manipulation of virus and host genetics to evaluate virus-host interaction networks in disease. In addition to assessing etiologies of known hCoVs, mouse models have clinically predictive value as tools to appraise potential disease phenotypes associated with pre-emergent CoVs. Knowledge of CoV pathogenic potential before it crosses the species barrier into the human population provides a highly desirable preclinical platform for addressing global pathogen preparedness, an overarching directive of the World Health Organization. Although we recognize that results obtained in robust mouse models require evaluation in non-human primates, we focus this review on the current state of hCoV mouse models, their use as tractable complex genetic organisms for untangling complex hCoV-host interactions, and as pathogenesis models for preclinical evaluation of novel therapeutic interventions.

Fig. 1

Emergent and pre-emergent coronavirus genome organization. The Orf1a and Orf1b genes (green) encode 16 non-structural proteins (nsp1–nsp16) that are highly conserved throughout coronaviruses. The structural genes (red) encode the structural proteins spike (S), envelope (E), membrane (M), and nucleocapsid (N), which are common to all coronaviruses. The accessory genes (dark shade) are unique to different coronaviruses with regard to number, genomic organization, sequence, and function

Fig. 2

Mouse model development should be tailored to the coronavirus of interest. Mouse-adapted SARS-CoV column (blue). SARS-CoV infects humans via an interaction of the spike protein [RCSB PDB ID: 5X58 (Yuan et al. 2017)] with its cognate receptor, ACE2 [RCBS PDB ID: 1R42 (Towler et al. 2004)]. In order to establish lethal mouse models, the spike protein, and other genomic determinants, have been modified through adaptive evolution in mouse lungs (Day et al. ; Frieman et al. ; Roberts et al. 2007a). Thereby, any mouse subspecies that exhibits an unaltered mouse Ace2 can be infected with the mouse-adapted SARS-CoV. Wild-type MERS-CoV column (green). MERS-CoV infects humans via an interaction of the spike protein [RCSB PDB ID: 5X5F (Yuan et al. 2017)] with its cognate receptor, DPP4 [RCSB PDB ID: 2ONC (Feng et al. 2007)]. The MERS-CoV spike protein is not able to interact with the mouse orthologue of human DPP4 (Cockrell et al. ; Coleman et al. 2014). Therefore, the mouse Dpp4 gene had to be genetically modified in order to allow for infection with MERS-CoV (Cockrell et al. ; Li et al. ; Pascal et al. 2015). Pre-emergent CoVs column (grey). Pre-emergent CoVs can use either known human receptors for CoVs (ACE2 or DPP4) (Ge et al. ; Menachery et al. ; Wang et al. ; Yang et al. 2016, 2014), or novel, unknown receptors to infect their host. The genetically highly diverse panel of mouse strains from the Collaborative Cross has the potential to provide mouse models for studies of newly emerging CoVs due to the high genetic variability present in this resource. For all virus images: yellow represents the envelope protein; light/dark blue represents the membrane protein; and red represents the nucleocapsid protein

Fig. 3

Comparison of mouse models for MERS-CoV. a Wild-type mouse DPP4 (mDPP4) with 26 exons (NM_010074.3). b Wild-type human DPP4 (hDPP4) with 26 exons (NM_001935.3). c DPP4 sequence used in VelociGene hDPP4 knock-in mouse model (Pascal et al. 2015). Mouse genomic sequence, from exon 2 through the stop codon in exon 26, was deleted and replaced with exon 2 through exon 26 and a portion of 3′ untranslated sequence of human genomic sequence. d DPP4 sequence used in hDPP4 knock-in model (Li et al. 2017). Mouse genomic sequence from codon I264 in exon 10 to codon V340 in exon 12 was replaced with the human equivalent. e DPP4 sequence used in 288–330^+/+ mouse model (Cockrell et al. 2016). CRISPR/Cas9 technology used to make A288L substitution in exon 10 and T330R substitution in exon 11 of mouse DPP4. f DPP4 construct used in hDPP4 overexpression mouse models. Constructs included cDNA from human DPP4 flanked by constitutive promoter, polyadenylation signals, and other regulatory elements (Agrawal et al. 2015); or by 5′ and 3′ genomic regions of human cytokeratin 18 or human surfactant protein C (Li et al. 2016)