SARS-CoV-2 B.1.1.7 (alpha) and B.1.351 (beta) variants induce pathogenic patterns in K18-hACE2 transgenic mice distinct from early strains

SARS-CoV-2 variants of concern (VOC) B.1.1.7 (alpha) and B.1.351 (beta) show increased transmissibility and enhanced antibody neutralization resistance. Here we demonstrate in K18-hACE2 transgenic mice that B.1.1.7 and B.1.351 are 100-fold more lethal than the original SARS-CoV-2 bearing 614D. B.1.1.7 and B.1.351 cause more severe organ lesions in K18-hACE2 mice than early SARS-CoV-2 strains bearing 614D or 614G, with B.1.1.7 and B.1.351 infection resulting in distinct tissue-specific cytokine signatures, significant D-dimer depositions in vital organs and less pulmonary hypoxia signaling before death. However, K18-hACE2 mice with prior infection of early SARS-CoV-2 strains or intramuscular immunization of viral spike or receptor binding domain are resistant to the lethal reinfection of B.1.1.7 or B.1.351, despite having reduced neutralization titers against these VOC than early strains. Our results thus distinguish pathogenic patterns in K18-hACE2 mice caused by B.1.1.7 and B.1.351 infection from those induced by early SARS-CoV-2 strains, and help inform potential medical interventions for combating COVID-19.

I n December 2019, a highly contagious novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in Wuhan, Hubei Province, China 1 . From there it has quickly spread around the world through international travelers resulting in the pandemic of coronavirus disease 2019 . Depending on age, sex and comorbidities, humans infected with SARS-CoV-2 experience variable disease severity, ranging from asymptomatic infection to pulmonary dysfunction, multi-organ failure and death 1 . As of June 2, 2021, the COVID-19 pandemic has caused over 170 million confirmed cases and >3 million deaths worldwide (https://covid19.who.int/). SARS-CoV-2 is non-segmented positive-sense, single-stranded RNA virus of bat origin belonging to Betacoronavirus family 2 . Its surface glycoproteinspike (S) contains a receptor binding domain (RBD) that has high affinity to human angiotensin converting enzyme 2 (hACE2) expressed on cell surface and thus allows virus to attach to and fuse with host cell membrane [2][3][4] . Like other RNA viruses, SARS-CoV-2 undergoes frequent recombination and evolutionary adaption 5 . Soon after the COVID-19 outbreak, a SARS-CoV-2 variant bearing D614G mutation in S outside of the RBD emerged and quickly became the dominant strain in circulation 6,7 . The D614G mutation enables S protein bind more efficiently to hACE2 and is associated with enhanced virus replication and transmissibility in humans and animal models 6,7 . As SARS-CoV-2 continues to circulate globally, more SARS-CoV-2 variants have emerged including B.1.1.7 (alpha) and B.1.351 (beta) lineages that were first detected in the UK and South Africa, respectively 8,9 . The viruses of these two lineages contain multiple mutations in S in addition to widespread D614G 10 . One of the mutations shared by both B.1.1.7 and B.1.351 variants is N501Y in RBD. N501 interacts with several residues of hACE2 including forming a hydrogen bond with Y41 of hACE2 to stabilize the virus-binding hotspot K353 at the RBD-hACE2 interface 3 . The N501Y mutation results in insertion of the aromatic ring of Y501 into a cavity between Y41 and K353 of hACE2, which increases RBD binding affinity to hACE2 11 . This ultimately leads to greater virus transmissibility, as the B.1.1.7 lineage is reportedly 43-90% more transmissible in the UK and 40-50% more transmissible in the US than early SARS-CoV-2 strains 8,12 . The B.1.351 variant is also estimated to be 50% more transmissible according to the preliminary modeling 9 . E484K is another RBD mutation detected in the B.1.351 lineage and in some of B.1.1.7 strains 10 . E484K confers resistance to several monoclonal antibodies and is able to afford immune escape [13][14][15] . Pseudoviruses bearing E484K alone or in combination of E484K/N501Y, K417N/ E484K or K417N/E484K/N501Y show more resistance to antibody neutralization than the parent pseudoviruses [13][14][15][16] . Convalescent or post-vaccination sera are found to less efficiently neutralize B.1.351 variant bearing K417N/E484K/N501Y than the original SARS-CoV-2 strain 15 . In addition to these RBD mutations, both B.1.1.7 and B.1.351 variants also contain several deletions in S protein, including Δ69-70 and Δ144/145 in B.1.1.7 and Δ242-244 in B.1.351 10 . These deletions are found to disrupt an immunodominant epitope in the N-terminal domain of S and alter the antigenicity of S, which not only results in resistance to monoclonal antibodies but also may accelerate SARS-CoV-2 adaptive evolution 17 . Because of increased transmissibility 8,9,12 and enhanced neutralization resistance documented [13][14][15][16][17] , both B.1.1.7 and B.1.351 are classified as variants of concern (VOC) (https://www.cdc.gov/coronavirus/2019-ncov/ variants/variant-info.html#Concern). However, the risks of increased disease severity caused by these two variants remain unclear due to many confounding factors such as different geographics, population backgrounds, public health infrastructures (https://www.who.int/publications/m/item/weekly-epidemiologicalupdate-on-covid-19-31-march-2021).
K18-hACE2 transgenic mice was originally developed for the investigations of SARS-CoV-1 pathogenesis 18 . Unlike another hACE2 knock-in mouse strain that does not exhibit mortality following SARS-CoV-2 exposure 6 , this K18-hACE2 mouse strain is susceptible to SARS-CoV-2 and produce various disease manifestations (e.g., lung inflammation and dysfunction, and death) that mimic severe SARS-CoV-2 infection in humans 19 .
We then collected various tissues for viral loads. For WA (614D) or NY (614G) infected mice, tissues were collected after animals were humanely euthanized at 3 dpi and 6 dpi. For CA (B.1.1.7) or SA (B.1.351) infection, tissues were collected at 3 dpi and 5-7 dpi, because the time of infected mice progressed to a moribund state was unpredictable and the quantities of available K18-hACE2 mice were low at the time of this study. After 3 dpi, viral RNA was readily detected in brain, lung, liver, spleen, kidney, and reproductive organ (ovary in female and seminal vesicles in male) of K18-hACE2 mice (Fig. 3). The expression of hACE2 was confirmed in these organs supporting localized SARS-CoV-2 infection ( Supplementary Fig. 2). We monitored viral loads in both testis and seminal vesicles of infected male mice and found that only seminal vesicles consistently had viral RNAs above the detection limit after SARS-CoV-2 infections. Thus, only viral loads in seminal vesicles of infected male mice were reported unless otherwise specified. Higher viral loads were detected in lung and brain homogenates than other organs across all four virus strains (Fig. 3). By 6 dpi, most mice infected with 10 4 TCID 50 /mouse of WA (614D) had tissue viral RNA levels dropped to the baseline except in lung and brain (Fig. 3). In contrast, mice inoculated with 10-fold less NY (614D) and 100fold less CA (B.   Fig. 3a-c). In contrast, nearly all WA (614D) or NY (614G) infected mice had IgM ELISA titers specific for N ( Supplementary Fig. 3a) and spike ( Supplementary  Fig. 3b) readily detected in blood as early as 3 dpi.
The results in clinical symptoms, viral burden, morbidity, and mortality are consistent, suggesting that both CA (B.1.1.7) and SA (B.1.351) variants are 100-fold more virulent in K18-hACE2 transgenic mice, followed by NY (614G) which is 10-fold more virulent, than WA (614D) that was first detected in the US in the beginning of the COVID-19 pandemic.
Nevertheless, only K18-hACE2 mice infected with NY (614G) showed extensive pulmonary hypoxia signaling before reaching to a moribund state; whereas exposure to a lethal dose of CA (B.1.1.7) or SA (B.1.351) or a 10-fold higher dose of WA (614D) resulted in mild or minimum pulmonary hypoxic pathway activation.
Infection of CA (B.1.1.7) or SA (B.1.351) variant results in significant elevation of D-dimer in lung, liver, and brain of K18-hACE2 mice. Severely ill COVID-19 patients often develop complications such as pulmonary embolism and/or deep vein thrombosis 22 . Elevated D-dimer has been associated with increased disease severity and mortality in hospitalized COVID-19 patients 23 . We thus measured tissue-specific D-dimer in lung, liver, and brain of K1-hACE2 mice infected with SARS-CoV-2 and variants. Intranasal infection of CA (B.1.1.7) at 10 2 TCID 50 / mouse resulted in significant elevations of D-dimer in lungs and brains (at a lesser extent) as early as 3 dpi and the levels of D-dimer continued to rise in tissues including livers harvested at later time points (Fig. 8a). This indicated a systemic blood clotting problem developed in K18-hACE2 mice following infection of CA (B.1.1.7). Intranasal infection of SA (B.1.351) also resulted in significant increases of D-dimer in lungs and brains (but not livers) harvested at 3 dpi, which lasted through 5-7 dpi (Fig. 8a). In contrast, K1-hACE2 mice infected with early WA (614D) or NY (614G) strain showed no significant elevation of D-dimer in tissues harvested, except slight rises of hepatic D-dimer level as compared to naïve mice (Fig. 8a). Interestingly, brain D-dimer levels showed no correlation with viral loads detected in brains (Pearson correlation coefficient r = 0.022); whereas pulmonary and hepatic D-dimer levels correlated well with viral loads in individual tissuesr = 0.344 for lung (p = 0.0041) and r = 0.341 for liver (p = 0.0044), respectively (Fig. 8b). However, brain D-dimer levels strongly correlated with D-dimer levels detected in lung (r = 0.929, p < 0.00001) and liver (r = 0.417, p < 0.001) as well (Fig. 8b). The PCA on tissue-specific D-dimer responses yielded two distinct groups (PC1 + PC2 > 98%), which mice infected with early WA (614D) or NY (614G) were clustered together; whereas mice infected with CA (B.1.1.7) or SA (B.1.351) variant were together in a separate group (Fig. 8c). Moreover, VIP scores for D-dimer in lung, liver and brain were 1.17, 0.61 and 1.12, respectively, all pointing to the cluster of CA/SA infections (Fig. 8d).
These results suggest tissue-specific D-dimer levels in lung, brain, and liver (at a lesser extent) were critical to distinguish infection of CA (B.1.1.7) or SA (B.1.351) variant from those infected with early WA (614D) or NY (614G) strain in K18-hACE2 mice. That D-dimer levels in brain, lung and liver strongly correlated, suggest the blood clotting problems observed in these organs were related to the same systemic cause. K18-hACE2 mice immunized with RBD or S emulsified with AddaS03™ (an AS03-like research grade adjuvant) also developed good antigenspecific IgG ELISA titers (Fig. 9e). Immunized mice also had markedly reduced MN titers against the variants especially SA (B.1.351) as compared to the titers toward WA (614D) (Fig. 9f). Nevertheless, all mice immunized with adjuvanted S were protected from lethal challenges of NY (614 G) or SA (B.1.351) without showing substantial morbidity ( Fig. 9g-h). Immunization of adjuvanted RBD also resulted in 75% survival (3 out of 4 mice/ group/virus) from the lethal NY (614G) or SA (B.1.351) challenge ( Fig. 9g-h).
These results together suggest the SARS-CoV-2-specific immunity developed from prior infection or immunization is effective in protecting K18-hACE2 mice from later exposure to CA (B.  10 . Both N501Y and D614G have been shown to increase RBD binding to hACE2 and promote virus entry and replication in humans and animal models 6,7 . The N501Y alone or in combination with K417N, E484K or K417N/E484K of S mutations identified in B.1.351 variant also greatly enhances viral infectivity in HEK293T with hACE2 overexpression 16 . In addition, the deletion of H69/V70 in S1 subunit is reported to increase the infectivity of B.1.1.7 variant and benefit more cleaved S incorporated into virions 24 . In this study, K18-hACE2 mice infected with CA (B.1.1.7) or SA (B.1.351) variant had comparable or higher viral loads in various organs than those inoculated with a 10-fold higher dose of NY (614G) or a 100-fold higher dose of WA (614D). The systemic dissemination of virus in K18-hACE2 transgenic mice was in consistence with tissue tropism of hACE2 expression. Some infected mouse organs had hACE2 expression lower than naïve controls, which may be due to virus-induced hACE2 shedding/ downregulation or cell death in these individual tissues 19,25 . Nevertheless, the overall expression of hACE2 in most tissues was enhanced after SARS-CoV-2 infection.
We observed that male K18-hACE2 mice were more susceptible than female mice to early SARS-CoV-2 strains (Supplementary Fig. 6). This is consistent with the meta-analysis of global cases that shows male sex is a risk factor for COVID-19 and male patients have higher odds for intensive care admission or death during the early pandemic 26 . Sex-specific differences in hormones, immunity and social behaviors may account for the relative resistance of females to COVID-19 [27][28][29] . In addition, human reproductive organs have high levels of ACE2 expression and SARS-CoV-2 RNA has been detected in testis and semen from male patients, and placenta and vaginal mucous membrane from pregnant women, respectively [30][31][32] . However, the long-term effects of COVID-19 on both male and female fertility remain unknown. In K18-hACE2 mice, viral RNA was also detected in ovary of female mice or seminal vesicles/testis of male mice following infection of SARS-CoV-2 or variants. Thus, K18-hACE2 mice would be an excellent model to assess the impact of SARS-CoV-2 infection on the fertility of both sexes.
Unlike those early COVID-19 cases showing sex differences, infections with B.1.1.7 or B.1.351 variant were found to have men and women almost in the equal ratio 33 . Human infections of B.1.1.7 or B.1.351 variants also had significantly higher adjusted odds ratios for hospitalization and intensive care admission than non-VOC cases based on the data collected from seven EU/EEA countries 33   mice not only indicate the severity of infection but also provide insights into potential immunomodulatory interventions for combating COVID-19.
Low blood oxygen is one of the early clinical signs present in COVID-19 patients. Some patients may experience dangerously low oxygen levels without developing dyspnea or showing other signs of respiratory distress-the so-called "silent hypoxia" 47 . Hypoxia has been reported to be independently associated with mortality in hospitalized COVID-19 patients and has been suggested as an indicator for COVID-19-related intensive care admission 48 . In this study, K18-hACE2 mice exhibited extensive hypoxia signaling in lungs at 3 dpi of NY (614G), but had minimum hypoxic genes upregulated or downregulated after WA (614D) infection. The latter may be because a sublethal dose of WA (614D) was used to inoculate K18-hACE2 mice. However, K18-hACE2 mice infected with CA (B.1.1.7) or SA (B.1.351) variant had fewer hypoxic genes altered than those infected with NY (614G), despite they all were administered a lethal infectious dose. Furthermore, infections of CA (B.1.1.7) and SA (B.1.351) variants resulted in significant D-dimer depositions in lungs and brains (livers at less extent) of K18-hACE2 mice as compared to those exposed to early WA (614D) or NY (614G) strain. Pulmonary embolism and venous/arterial thromboembolism are frequently diagnosed in critically ill COVID-19 patients and are significantly associated with disease severity  (intensive care admission or death) 22,49,50 . Elevated D-dimer indicating blood clot formation has been used as a diagnostic marker for pulmonary embolism in COVID-19 patients, and hospitalized patients with superhigh D-dimers often have poor prognosis and low chance of survival 1,22,23,42,51 . The deposition of D-dimers in the vital organs along with a lack of extensive pulmonary hypoxia signaling before death indicates that the pathogenic patterns of B.1.1.7 and B.1.351 variants are different from those of early SARS-CoV-2 strains in K18-hACE2 mice. Both pulmonary and hepatic D-dimer levels were also found to correspond with local viral loads, indicating virus replication induced tissues damages in these mice. Interestingly, brain D-dimer levels of infected mice showed no correlation with local viral loads but were strongly associated with pulmonary D-dimer levels. These are likely due to SARS-CoV-2 caused systemic endothelial injury and microthrombosis 50,52 . The postmortem histopathological examination reveals that many COVID-19 patients suffered multifocal microvascular injury in brains, e.g., congested blood vessels, fibrinogen leakage, and microhemorrhages in addition to widespread thrombosis in pulmonary vessels 50,52 . K18-hACE2 mice infected with CA (B.1.1.7) or SA (B.1.351) variant also had severe hemorrhages in brain before they succumbed to sudden death. These differences in pulmonary hypoxia signaling, tissue-specific D-dimer deposition and cytokine file confirm that the B.1.1.7 and B.1.351 variants have distinct pathogenic characteristics from early SARS-CoV-2 strains.
Sera collected from K18-hACE2 with prior exposure to early SARS-CoV-2 strains or immunized with adjuvanted S or RBD had significantly reduced neutralization titers against CA (B.1.1.7) and SA (B.1.351). This is consistent with the reports that viruses bearing escape mutations such as Δ69-70 and Δ144/145 in B.1.1.7 or K417N/E484K/N501Y in B.1.351 can avoid recognition by monoclonal antibodies or show resistance to neuralization by convalescent/post-vaccination sera [13][14][15][16][17] . Despite having reduced neutralizing antibodies, K18-hACE2 mice with prior SARS-CoV-2 exposure or immunized with AS03-like adjuvanted S all survived the lethal challenge of CA (B.1.1.7) or SA (B.1.351) variant with minimum BW loss. Immunization of AS03-like adjuvanted RBD also protected most K18-hACE2 mice from the lethality of SA (B.1.351) or NY (614G). Infection or immunization induced cellular immunity may contribute to the protection observed in these K18-hACE2 mice with low neutralization titers against variants. It has been reported that SARS-CoV-2 infection induces long-lasting memory B cells and T cells in humans 53,54 . Individuals with prior infection are able to develop robust T cell and memory B cell responses against B.1.1.7 and B.1.351 variants after receiving first mRNA vaccine dose 55 . Investigations on characterizing T cell and B cell responses in K18-hACE2 mice following exposure to different SARS-CoV-2 variants or immunization are currently underway. Although K18-hACE2 mice are highly susceptible to SARS-CoV-1 and SARS-CoV-2 infections as demonstrated in this study and by others 18,19 , this mouse strain like other mouse models has the limitations. The hACE2 expression in K18-hACE2 strain is driven by the human cytokeratin-18 (K18) gene promoter, which results in a different hACE2 physiological distribution pattern from that of human 18 . In addition, mice are unsuitable for SARS-CoV-2 transmission investigations and the genetic disparities between mouse and human immune systems also make them less appealing than non-human primates for studying vaccine-induced immune responses 56,57 .
In summary, the current study in K18-hACE2 mice reveals that B.1.1.7 and B.1.351 are at least 100 times more lethal than the original SARS-CoV-2 WA (614D) strain detected in the beginning of the pandemic. K18-hACE2 mice infected by B.1.1.7 and B.1.351 variants develop BW loss, hypothermia, and severe internal organ damages. B.1.1.7 and B.1.351 infection also produces a different pathogenic profile from that induced by early SARS-CoV2 strains (614D or 614G), which is characterized as distinct tissue-specific proinflammatory cytokine signatures, significant D-dimer depositions in vital organs and a lack of extensive pulmonary hypoxia signaling before death. Moreover, K18-hACE2 mice with the preexisting immunity from prior infection of early SARS-CoV-2 strains or immunization of S or RBD are protected from lethal B.1.1.7 or B.1.351 reinfection, despite exhibiting lower neutralizing titers against these VOC than early strains. The current study not only identifies the in vivo pathogenic characteristics of B.1.1.7 and B.1.351 variants but also provide insights into potential medical interventions to control inflammation, mitigate tissue damages and improve survival outcome of COVID-19 patients.

Methods
Cells and viruses. Vero E6 (CRL-1586; American Type Culture Collection or ATCC) was maintained in high-glucose Dulbecco's modified Eagle's medium . Infectious stocks were prepared by inoculating 90% confluent Vero E6 cells in DMEM supplemented with 3% FBS at 37°C, 5% CO2 until observation of cytopathic effect. Supernatants were centrifuged and passaged through a 0.45-μm filter. Aliquots were stored in a secured −80°C freezer until use. All work with live SARS-CoV-2 isolates was performed in an animal biosafety level (ABSL) 3 laboratory equipped with advanced access control devices and by personnel equipped with powered air-purifying respirators.
Determination of 50% tissue culture infectious dose (TCID50). Virus infectivity was determined using an ELISA-based TCID 50 method adopted from a protocol used for titrating influenza virus 58   (d) The Variable Importance in Projection (VIP) scores for the D-dimer levels in lung, liver and brain that drive the differences between WA/NY and CA/ SA infected mice. Bar length corresponds to relative importance. (e) Serum IgG ELISA titers specific for RBD or S or (f) MN titers for SARS-CoV-2 and variants before challenge (n = 8 mice/group). BW and survival of immunized mice (n = 4 mice/group) and naïve K18-hACE2 mice (n = 3 mice/group) were monitored for two weeks after challenge of (g) NY (614G) or (h) SA (B.1.351) variant. Short lines represent geometric means and dotted horizontal lines indicate the limit of detection. BW are expressed as mean ± s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001 and **** p < 0.0001 by one-way nonparametric ANOVA with Dunn's multiple comparisons test after log-transformation of MN titers. (d) and (h) were conducted at the same time with the same naïve mice as the control.
trimerization domain and a 6xHis tag at the C-terminus were codon-optimized and were subcloned into pcDNA™5/FRT mammalian expression vector (Invitrogen #V601020). Expi293™ Expression system (ThermoFisher #A14635) was used to produce recombinant SARS-CoV-2 S and RBD proteins according to the manufacturer's protocol. The cultured media harvested on day 6 post-transfection were centrifuged at 27,000 × g, 4°C for 60 min. The supernatants were concentrated and diafiltrated through Amicon Ultra-15 10 K (EMD Millipore #UFC901008) in the equilibration buffer (50 mM Tris-HCl, 150 mM NaCl and EDTA-free protease inhibitor cocktail, pH 8.0). The diafiltrated supernatants were then loaded to equilibrated HisTrap™ HP His tag protein purification columns (Cytiva #17524802) and bound proteins were eluted in the presence of 500 mM imidazole. Expressed S and RBD proteins were purified using an AKTA Pure system (Cytiva) with >90% purity as confirmed by SDS-PAGE. Expression of the full-length E, M and N proteins 59 was carried out in propriety Xpress CF™ system with reactions in thinfilm Petri dish incubated at 25°C overnight without shaking. For recombinant E and M, 0.1% Brij35 was added to the cellfree reactions and purification buffers to aid solubility and slow aggregation. Expressed E, M and N were purified by Ni-NTA affinity chromatography with >85% purity by SDS-PAGE.
Mice. Specific pathogen-free hemizygous B6.Cg-Tg(K18-ACE2)2Prlmn/J (K18-hACE) female mice (JAX Stock No. 034860) and noncarrier c57BL/6 J male mice (JAX Stock No. 000664) were mated to maintain live colonies at the ABSL2 facility of AAALAC International-accredited FDA White Oak Vivarium. Only hACE2 (+) offspring as confirmed by genotyping (Transnetyx) were retained and were used for experiments. At the age of 8-10 weeks, K18-hACE mice of both sexes were randomly assigned to experimental groups at~1:1 ratio and were transferred to the ABSL3 facility for infection. The ABSL3 suite was maintained at 65-75°F with 40-60% humidity and 12 h light:dark cycles. All K18-hACE mice were ear tagged and were housed at two per microcage in the ABSL3 facility with ad libitum access to food and water. Experimental mice and naïve controls were separated housed in different microcages. Under isoflurane anesthesia, mice were inoculated with SARS-CoV-2 or variants via intranasal route. Body weight (BW) and mortality were monitored daily for up to 14 days post infection (dpi). Cutaneous temperature was also measured on mouse tail surface using a rodent infrared thermometer (Braintree Scientific) following infection. BW, mortality, and cutaneous temperatures were tableted and processed in Microsoft Excel. Mice becoming moribund or reaching humane endpoints (e.g., 30% BW loss) were immediately euthanized with CO2 inhalation in a euthanasia chamber. Mice survived from prior infection of WA (614D) or NY (614G) were reinfected at 6-8 weeks later with 10 2 TCID 50 / mouse of CA (B.1.1.7) or SA (B.1.351). In separated experiments, K18-hACE2 mice of both sexes were primed and boosted at 3 weeks intervals with S (10 μg/dose) or RBD (20 μg/dose) emulsified with AddaS03™ (InvivoGen #vac-as03-10) via intramuscular route. Sera were collected right before the challenges for antibody determination. All procedures were performed according to the animal study protocols approved by the FDA White Oak Animal Program Animal Care and Use Committee.
Viral burden and hACE2 expression. Tissues were weighed and homogenized in PBS (pH7.2) (10%, v/w) with ceramic beads in a Fisherbrand™ Bead Mill 24 Homogenizer (FisherScientific). Total RNA was extracted from tissue homogenates using the RNeasy Plus Mini Kit (Qiagen #74136) and was immediately reverse transcribed and amplified using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) following the manufacturer's protocol. Copies of SARS-CoV-2 nucleocapsid (N) gene in homogenized tissues were determined using QuantiNova SYBR Green PCR kit (Qiagen #208052) along with 2019-nCoV RUO Kit (Integrated DNA Technologies #10006713) as gene specific primers at the final concentration of 500 nM. Quantitative amplification in Stratagene MX3000p qPCR system (Agilent) was conducted according to the following program: 95°C for 120 s, 95°C for 5 s (50 cycles) and 60°C for 18 s. Threshold cycle (Ct) values were calculated using MxPro qPCR software (Agilent) and the N gene copies in individual tissues were interpolated from a standard curve constructed by serial dilutions of a pCC1-CoV2-F7 plasmid expressing SARS-CoV-2 N 60 . hACE2 expression in different mouse organs was determined similarly using hACE2 gene specific primer set (Integrated DNA Technologies, Assay ID: Hs.PT.58.27645939) at the final concentration of 500 nM. The copies of expressed hACE2 were interpolated from a standard curve created by serial dilutions of a pSBbi-bla ACE2 plasmid expressing hACE2 (kindly provided by J Yewdell, NIAID/NIH). A value of 1 was assigned if gene copies were below the detection limits.
Hypoxia signaling pathway PCR array. Total RNA extracted from lungs on dpi 3 was reverse transcribed as described above and was mixed with RT 2 SYBR Green ROX qPCR Mastermix (Qiagen #330523) to perform RT² Profiler™ PCR Array Mouse Hypoxia Signaling Pathway (Qiagen #PAMM-032ZA-24/330231) real-time PCR in Stratagene MX3000p qPCR system under these cycling conditions: hold for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 60 s at 60°C. Ct values were determined using MxPro qPCR software (Version 3.0). Individual gene expressions based on Ct values were normalized using Gusb as the internal housekeeping gene and were calculated for fold changes using Qiagen web-based GeneGlobe data analysis tool. Heatmap was generated using Prism 8.4.3 (GraphPad).
D-dimer measurement. Detection of D-dimer in tissue homogenates was carried out using mouse D-Dimer ELISA Kit (MyBioSource #MBS764560) according to the manufacturer's instructions.
Antibody ELISA. All mouse sera were heat-inactivated before antibody assessment. ELISA was performed in 96-well microtiter plates pre-coated with 1 μg/ml of recombinant S, RBD or N protein. Bound antibodies were detected using peroxidase-conjugated goat anti-mouse IgM heavy chain (Southern Biotech #1021-05, 1:4000) or IgG (H + L) highly cross-adsorbed antibodies (Invitrogen #A24518, 1:2000) followed by 1-step™ Ultra TMB-ELISA substrate (ThermoFisher #34029). ELISA titers were interpolated based on a standard curve constructed using inhouse developed rabbit polyclonal antibodies specific for S, RBD or N, respectively.
Microneutralization (MN) assay. Neutralizing antibody titers against live SARS-CoV-2 and variants were determined using a cell-based MN assay. Briefly, heatinactivated mouse sera were serially diluted and were incubated with 100 TCID 50 / well of live virus at room temperature for 1 h. The virus-serum mixtures were then added to Vero E6 cells (2 × 10 4 cells/well) pre-seeded in 96-well tissue culture plates and were incubated at 37°C, 5% CO2 for 2 days. Cells were then fixed with 4% paraformaldehyde for 30 min followed by permeabilization in 0.1% NP-40 detergent for 15 min. Virus-infected cells were detected using in-house raised rabbit polyclonal antibody specific for SARS-CoV-2 E/M/N (1:1000) along with peroxidase-conjugated goat anti-rabbit IgG (H + L) secondary antibody (Invitrogen #31460, 1:2000). MN titers represented the reciprocal of the highest serum dilution that yielded >50% reduction in OD values as compared to wells containing virus only. A MN titer of 20 was assigned if no neutralization was observed at the initial serum dilution of 1:40.
Histology. The whole brains and lungs harvested from infected mice were fixed in 10% neutral buffered formaldehyde for at least two weeks before being embedded in paraffin for histology (Histoserv). Tissues embedded in paraffin were sectioned and stained with hematoxylin and eosin by the standard protocols (Histoserv). Uninfected mouse brains and lungs were used as negative controls and stained in parallel. Tissue sections were visualized using a Leica Aperio AT2 slide scanner with Aperio ImageScope DX clinical viewing software version 12.4.3.5008 (Histoserv).
Statistical analysis. Multivariate Principal Component Analysis (PCA) and Partial Least Squares Discriminant Analysis (PLSDA) were used to analyze cytokine profile and D-dimer among different infection groups. PCA plots were used to visualize if samples from different virus exposures were separated into discrete groups within the two-dimensional principal component (PC) space. The variable importance in projection (VIP) scores, which measures the variable's importance, were reported from the PLSDA model. Linear models were also performed to associate tissue-specific D-dimer with viral loads in individual organs. R version 4.0.3 (https://www.r-project.org/) was used to perform all the above analyses. Prism 8.4.3 (GraphPad) was used to conduct one-way nonparametric ANOVA or twoway mixed ANOVA after the log transformation of antibody titers, or Log-rank (Mantel-Cox) survival test. A p < 0.05 was considered statistically significant.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
PCR array data that support the findings of this study have also been deposited in NCBI GEO with the accession number GSE183621 and are provided in Supplementary Table 1. The DNA sequences encoding the full-length S (residues 1-1213) and RBD (residues 319-541) of WA are obtained from GenBank with accession number MN985325.1. The unique materials/reagents used in this study are available from the corresponding author upon reasonable request and by Material Transfer Agreement. Source data are provided with this paper.