Cardiopulmonary Injury in the Syrian Hamster Model of COVID-19

The Syrian hamster has proved useful in the evaluation of therapeutics and vaccines for severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2). To advance the model for preclinical studies, we conducted serial sacrifice of lungs, large pulmonary vessels, and hearts from male and female Syrian hamsters for days 1–4, and 8 post-infection (dpi) following infection with a high dose of SARS-CoV-2. Evaluation of microscopic lung histopathology scores suggests 4 and 8 dpi as prime indicators in the evaluation of moderate pathology with bronchial hyperplasia, alveolar involvement and bronchiolization being key assessments of lung disease and recovery, respectively. In addition, neutrophil levels, red blood cell count and hematocrit showed significant increases during early infection. We present histological evidence of severe damage to the pulmonary vasculature with extensive leukocyte transmigration and the loss of endothelial cells and tunica media. Our evidence of endothelial and inflammatory cell death in the pulmonary vessels suggests endothelialitis secondary to SARS-CoV-2 epithelial cell infection as a possible determinant of the pathological findings along with the host inflammatory response. Lastly, pathological examination of the heart revealed evidence for intracardiac platelet/fibrin aggregates in male and female hamsters on 8 dpi, which might be indicative of a hypercoagulative state in these animals.


Introduction
Animal models that recapitulate the multisystemic nature of infection and disease caused by severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) in humans are critical for the interrogation of mechanisms underlying the host response, pathology and evaluation of approved or new investigational drugs that may alleviate disease [1]. The Syrian hamster has been widely accepted as a preclinical model for COVID-19 (coronavirus disease-2019) therapeutics and vaccines [2][3][4][5][6][7][8] based on the similarity of lung histopathology in human cases [9][10][11][12]. Prior to SARS-CoV-2, Syrian hamsters have been used as animal models for other human respiratory viruses [13], such as SARS-CoV [14], influenza

Cells and Virus
SARS-Related Coronavirus 2, Isolate USA-WA1/2020, NR-52281 was obtained through BEI Resources Centers for Disease Control and Prevention and, NIAID, NIH. The virus was amplified in Vero E6 TMPRSS cells using Minimum Essential Media with Earle's salts (MEM) with 1% penicillin-streptomycin (P/S). Infection of cells was conducted at an MOI = 0.1 with 2% fetal bovine serum for 1 h and then MEM containing 1% P/S, and 10% FBS was added. Passage 3 viral seed stocks were sequenced and used in animal studies. Virus seed stock titers were measured using plaque assays with Vero E6 (ATCC CRL-1586). Vero E6 were maintained in Minimum Essential Medium (MEM) with Earle's salts and L-glutamine. All reagents and cell culture reagents were purchased from Thermo Fisher Scientific unless specified.

General Animal Study Information
LVG Golden Syrian hamsters were supplied by Charles River. Hamsters were identified with ear tags and microchips (IPTT-300 by Bio Medic Data System). Hamsters were housed individually in caging under sealed, negative pressure in Allentown Biocontainment Unit (BCU-3) Rat racks. The microchips were implanted subcutaneously between the shoulder blades of the hamsters. All hamsters were anesthetized with ketamine (100 mg/kg) plus xylazine (10 mg/kg) by I.P. prior to intranasal infection. As specified, subsets of hamsters were anesthetized with ketamine (250 mg/kg) and xylazine (100 mg/kg) by intraperitoneal injection and euthanized via isoflurane anesthesia followed by a bilateral thoracotomy. Because the weights of hamsters vary, the volume of K/X injected was calculated based on the weight of the hamster. Clinical signs were monitored twice daily upon the manifestation of weight loss throughout the study. Otherwise, hamsters were monitored once per day upon recovery. The main clinical sign is weight loss; however, hamsters were evaluated for arched back and slightly rough coat (score = 1), weight loss and disinterest in food/water or changes/signs of respiratory distress (score = 2) and weight loss of greater than 10%, closed eyes or prostrate (score = 3 which requires euthanasia). If a hamster received a score of 2, a veterinarian was called. Hamsters showed minimal clinical signs and we have never observed scores above 1 for infection with SARS-CoV-2 WA1/2020. Hamsters were provided enrichment with one running wheel and one Igloo per cage both from bio-serve inc. Studies were conducted in accordance and approval by the Institutional Animal Care and Use Committee of the University of Tennessee Health Science Center (Protocol #20-0132).

Serial Sacrifice Study of Female and Male Hamsters
Within a B1 classified biosafety cabinet, twenty females at 9-weeks-old and twenty males at 7-weeks-old Golden Syrian hamsters were intranasally-infected with 2 × 10 5 pfu of SARS-CoV-2 in a total volume of 100 µL. The ages differed based on the availability of males and females by the vendor but were within the range of 7-9 weeks of age used in our studies. Prior to infection, hamsters were anesthetized as described above. On 1, 2, 3, 4 or 8 dpi, blood, lung, and heart were harvested and snap frozen in liquid nitrogen or placed into formalin. Blood was collected for hematology as described below and analyzed the same day. The whole lung was collected, the left single lobe was taken for histopathology and the right five lobes were taken for measurement of viral titers by plaque assay as described below.

Hematology Analysis
Terminal percutaneous cardiac blood samples were obtained from each hamster after the induction of deep anesthesia with ketamine/xylazine. Blood was placed into Microtainer EDTA tubes, mixed according to the manufacturer's instructions, and analyzed within 6 h of collection. A CBC (complete blood count) was performed using a DiaSys Xpedite™ HEM3 VET hematology analyzer (DiaSys Diagnostic Systems, Holzheim, Ger-Viruses 2022, 14, 1403 4 of 19 many). Tri-level quality control material (Para12Extend Controls) was run daily prior to sample analysis. The Xpedite HEM3 analyzer provides direct counting for WBC, RBC and platelets using an impedance-based system, hemoglobin is photometrically measured, and all other parameters are calculated values: MCH (mean corpuscular hemoglobin), MCHC (mean corpuscular hemoglobin concentration), MCV (mean corpuscular volume), MPV (mean platelet volume) and HCT (hematocrit). A three-part white blood cell differential analysis is performed and reported as the absolute number and percentage of lymphocytes, MID (monocytes, eosinophils, and basophils) and granulocytes (neutrophils). All reagents were purchased from DiaSys, and analysis was performed according to the manufacturer's established procedures.

Histology and Immunohistochemistry
Tissues were fixed in 10% neutral formalin for 48 h and embedded in paraffin; 5 µm serial sections were obtained and stained with H&E, NACE stain with 91C-1KT (Sigma-Aldrich, St. Louis, MO, USA), or by IHC for SARS-CoV-2 with rabbit anti-SARS-CoV-2 nucleoprotein (40143-R001, Sino Biological) or ACE-2 with rabbit anti-ACE-2 monoclonal Ab (NBP-67692, Novus Biological, CO, USA) Prosurfactant protein C (proSP-C) was detected similarly using rabbit anti-proSP-C from (Sigma-Aldrich AB3786). For immunostaining of SARS-CoV-2 nucleoprotein, slides were deparaffinized in xylene and rehydrated in decreasing concentrations of ethanol, and then treated with preheated Citrate-Based Antigen Unmasking Solution pH 6.0 (H-3300, Vector Laboratories) to reveal the nucleoprotein for 10 min at 97 • C as recommended by the manufacturer. Similarly, ACE2 was unmasked with Tris-based Antigen Unmasking Solution (H3301, Vector Laboratories, Newark, CA, USA) as recommended by the manufacturer. Unmasked tissue sections were rinsed with deionized water, endogenous peroxidase activity was quenched with 1% hydrogen peroxide. After rinsing 3× with phosphate buffer saline solution (PBS), slides were incubated overnight at 4 • C with primary rabbit antibody against nucleocapsid diluted at 1:1000 or ACE-2 diluted at 1:2000 (Cat # 40143-R001, Sino Biologicals, Beijing, China). Biotinylated secondary anti-rabbit IgG antibody was applied at 1:200 (Vector Laboratories) at room temperature for 1 h, rinsed and then incubated with Vectastain Elite using the ABC-HRP (Avidin-Biotin Complex-horseradish peroxidase) Kit from Vector Laboratories for 1 h (Vector Laboratories) at room temperature. Slides were counterstained with hematoxylin, dehydrated in ethanol and xylene and a coverslip mounted with Permont. Images were scanned with an Olympus SlideView VS200 using the Olympus OlyVIA v.3.2.1 imaging software.

Evaluation of Banked Heart Samples from Studies of Female and Male Hamsters
Four additional studies were conducted using the general parameters stated above and lung and heart tissues from sham inoculated (PBS) and SARS-CoV-2-infected were banked in paraffin blocks. In these studies, six-to nine-week-old male (n = 8) or female (n = 16) hamsters were purchased from Charles River. Hamsters were intranasally-infected with 2 × 10 5 pfu of SARS-CoV-2 in a total volume of 100 µL. Tissues were collected and placed into 10% formalin as described for the serial sacrifice above. Using these banked specimens, we examined the heart tissue from sham or SARS-CoV-2 inoculated hamsters from 8 dpi. Evaluation of the banked tissues by histology and immunohistochemistry were as described above.

Plaque Assay
Plaque assays were used to measure the SARS-CoV-2 titer from seed stocks or homogenized tissues. Twelve-well plates were seeded with Vero E6 cells overnight. Ten series of 10-fold dilutions of virus samples were made in MEM with 2% FBS, and 200 µL of each dilution was added to each well in duplicate and incubated at 37 • C, 5% CO 2 for 1 h. Following incubation, overlay media consisting of a 1:1 mix of 2% carboxymethylcellulose and 2XMEM supplemented with 10% FBS, 1% L-glutamine, 2% penicillin-streptomycin was added to each well. Plates were incubated at 37 • C, 5% CO 2 for 72 h. Wells were fixed with 10% formalin, stained with 1% crystal violet, washed with Dulbecco's phosphate buffered saline, and plaques were counted.

Data and Statistical Analyses
Body weight change, plaque forming units (viral load) and CBC (see hematology analysis above) were analyzed using GraphPad Prism 9 (GraphPad Software, Inc., La Jolla, CA, USA).

Serial Sacrifice Study Design, Body Weight, Viral Load over Time
We conducted a serial sacrifice of males (n = 2) and females (n = 2) on 1, 2, 3, 4, and 8 dpi using the high dose, 2 × 10 5 pfu, of the SARS-CoV-2 USA-WA1/2020 or PBS. Blood was taken for hematology. Lung and heart were taken for histopathology and immunohistochemistry of viral antigen as will be presented. Lungs were evaluated for microscopic pathology and scored.
At each time point, we recorded the corresponding body weight and viral load in the lung (Figure 1). Male and female hamsters dropped approximately 10% body weight by 5 dpi ( Figure 1A). In both sexes, the viral load was essentially stable in the lungs from 1-4 dpi and generally decreased markedly on 8 dpi ( Figure 1B). In one male, the virus titer remained high 8 dpi.

Plaque Assay
Plaque assays were used to measure the SARS-CoV-2 titer from seed stocks or homogenized tissues. Twelve-well plates were seeded with Vero E6 cells overnight. Ten series of 10-fold dilutions of virus samples were made in MEM with 2% FBS, and 200 μL of each dilution was added to each well in duplicate and incubated at 37 °C, 5% CO2 for 1 h. Following incubation, overlay media consisting of a 1:1 mix of 2% carboxymethylcellulose and 2XMEM supplemented with 10% FBS, 1% L-glutamine, 2% penicillin-streptomycin was added to each well. Plates were incubated at 37 °C, 5% CO2 for 72 h. Wells were fixed with 10% formalin, stained with 1% crystal violet, washed with Dulbecco's phosphate buffered saline, and plaques were counted.

Data and Statistical Analyses
Body weight change, plaque forming units (viral load) and CBC (see hematology analysis above) were analyzed using GraphPad Prism 9 (GraphPad Software, Inc).

Serial Sacrifice Study Design, Body Weight, Viral Load over Time
We conducted a serial sacrifice of males (n = 2) and females (n = 2) on 1, 2, 3, 4, and 8 dpi using the high dose, 2 × 10 5 pfu, of the SARS-CoV-2 USA-WA1/2020 or PBS. Blood was taken for hematology. Lung and heart were taken for histopathology and immunohistochemistry of viral antigen as will be presented. Lungs were evaluated for microscopic pathology and scored.
At each time point, we recorded the corresponding body weight and viral load in the lung ( Figure 1). Male and female hamsters dropped approximately 10% body weight by 5 dpi ( Figure 1A). In both sexes, the viral load was essentially stable in the lungs from 1-4 dpi and generally decreased markedly on 8 dpi ( Figure 1B). In one male, the virus titer remained high 8 dpi. Weight and viral load in male or female hamsters intranasally infected with PBS or SARS-CoV-2 at a high dose, 2 × 10 5 pfu. In the serial sacrifice study of SARS-CoV-2-infected hamsters, (A) body weight change (data are presented as mean ± SEM) was measured daily from 0 to 8 dpi, and (B) viral load was measured by plaque assay in lungs of male and female hamsters at five time points (n = 2). Data are presented as geometric mean ± geometric SD.

Hematology Showed a Significant Elevation of Neutrophils, RBCs, and Hemoglobin Early in SARS-CoV-2 Infection
The complete blood counts (CBC) revealed the most significant changes on 1 dpi ( Figure 2). White blood cell measurements exhibited a significant increase in both the percentage (1 dpi, p = 0.0004; 2 dpi, p = 0.003) and absolute number (1 dpi, p = 0.03, Figure 2) of granulocytes. The WBC population was 65% neutrophils on 1 dpi followed by a decline to background levels by 4 dpi (Figure 2). Similarly, the percentage of lymphocytes in the blood was decreased at 1 dpi (p = 0.00008) and 2 dpi (p = 0.012) but returned to baseline by 3 dpi. Total RBC levels (p = 0.0004) and hematocrit (p = 0.0008) were significantly elevated from mock by 3 dpi. The hematocrit levels remained elevated from 1 to 8 dpi ( Figure 2). Hemoglobin levels were significantly elevated from mock on 3 dpi (p = 0.003) and 4 dpi, (p = 0.03). Although not significant, platelet counts were decreased at 2 dpi but subsequently were similar to mock-infected through 8 dpi.

Figure 1.
Weight and viral load in male or female hamsters intranasally infected with PBS or SARS-CoV-2 at a high dose, 2 × 10 5 pfu. In the serial sacrifice study of SARS-CoV-2-infected hamsters, (A) body weight change (data are presented as mean ± SEM) was measured daily from 0 to 8 dpi, and (B) viral load was measured by plaque assay in lungs of male and female hamsters at five time points(n = 2). Data are presented as geometric mean ± geometric SD.

Hematology Showed a Significant Elevation of Neutrophils, RBCs, and Hemoglobin Early in SARS-CoV-2 Infection
The complete blood counts (CBC) revealed the most significant changes on 1 dpi ( Figure 2). White blood cell measurements exhibited a significant increase in both the percentage (1 dpi, p = 0.0004; 2 dpi, p = 0.003) and absolute number (1 dpi, p = 0.03, Figure 2) of granulocytes. The WBC population was 65% neutrophils on 1 dpi followed by a decline to background levels by 4 dpi (Figure 2). Similarly, the percentage of lymphocytes in the blood was decreased at 1 dpi (p = 0.00008) and 2 dpi (p = 0.012) but returned to baseline by 3 dpi. Total RBC levels (p = 0.0004) and hematocrit (p = 0.0008) were significantly elevated from mock by 3 dpi. The hematocrit levels remained elevated from 1 to 8 dpi ( Figure  2). Hemoglobin levels were significantly elevated from mock on 3 dpi (p = 0.003) and 4 dpi, (p = 0.03). Although not significant, platelet counts were decreased at 2 dpi but subsequently were similar to mock-infected through 8 dpi. Blood was drawn from hamsters on 1, 2, 3, 4 or 8 dpi and a complete blood count was performed. (*: p < 0.05, **: p < 0.01;***: p < 0.001,****: p < 0.0001.)

Evaluation of Microscopic Lung Pathology in Syrian Hamsters Show Peak of Inflammation and Lesions on Day 4 Post-Infection except for Bronchiolization
Microscopic lesions within the lungs of each hamster were examined and scored at each time point (Table S1). Scores were combined to generate a heatmap ( Figure   Figure 2. Complete blood count from hamsters infected with SARS-CoV-2 or sham inoculated (PBS). Blood was drawn from hamsters on 1, 2, 3, 4 or 8 dpi and a complete blood count was performed. (*: p < 0.05, **: p < 0.01; ***: p < 0.001, ****: p < 0.0001.)

Evaluation of Microscopic Lung Pathology in Syrian Hamsters Show Peak of Inflammation and Lesions on Day 4 Post-Infection except for Bronchiolization
Microscopic lesions within the lungs of each hamster were examined and scored at each time point (Table S1). Scores were combined to generate a heatmap ( Figure 3). Bronchial/bronchiolar exudates and hyperplasia, endothelial hypertrophy/margination and peribronchiolar/perivascular inflammation were noted by 1 dpi with mild severity. Necrosis/apoptosis/sloughing of the bronchiolar epithelium was widespread on 2 dpi but minimal to absent at other time points. Bronchiolar epithelial hyperplasia was absent to mild on 1 dpi but increased markedly in severity and extent by 4 dpi, with some reduction in severity by 8 dpi. Minimal to mild alveolar inflammation was first detected on 2 dpi and continued to increase in extent and severity from 3 to 8 dpi. Mild to moderate pulmonary consolidation was first noted on 3 dpi and generally increased in extent and severity to 8 dpi. Endothelial hypertrophy associated with margination of inflammatory cells was widespread from 2 dpi to 4 dpi and was mostly resolved by 8 dpi. Similarly, the extent and severity of peribronchiolar/perivascular inflammation gradually increased from 1 dpi to 2-4 dpi but was declining by 8 dpi. Overall, the extent and severity of bronchiolar hyperplasia, alveolar and interstitial inflammation, and septal thickening were maximal on 4 dpi and were only slightly resolved by 8 dpi. Alveolar type 2 cell hyperplasia and/or bronchiolization of alveoli was absent until 8 dpi when this lesion was extensive in all four examined hamsters.
3)."Bronchial/bronchiolar exudates and hyperplasia, endothelial hypertrophy/margination and peribronchiolar/perivascular inflammation were noted by 1 dpi with mild severity. Necrosis/apoptosis/sloughing of the bronchiolar epithelium was widespread on 2 dpi but minimal to absent at other time points. Bronchiolar epithelial hyperplasia was absent to mild on 1 dpi but increased markedly in severity and extent by 4 dpi, with some reduction in severity by 8 dpi. Minimal to mild alveolar inflammation was first detected on 2 dpi and continued to increase in extent and severity from 3 to 8 dpi. Mild to moderate pulmonary consolidation was first noted on 3 dpi and generally increased in extent and severity to 8 dpi. Endothelial hypertrophy associated with margination of inflammatory cells was widespread from 2 dpi to 4 dpi and was mostly resolved by 8 dpi. Similarly, the extent and severity of peribronchiolar/perivascular inflammation gradually increased from 1 dpi to 2-4 dpi but was declining by 8 dpi. Overall, the extent and severity of bronchiolar hyperplasia, alveolar and interstitial inflammation, and septal thickening were maximal on 4 dpi and were only slightly resolved by 8 dpi. Alveolar type 2 cell hyperplasia and/or bronchiolization of alveoli was absent until 8 dpi when this lesion was extensive in all four examined hamsters. The severity scores for each category of microscopic pathology are illustrated using a heatmap. For each of the days examined, each box represents the average score from four hamsters for scores for these categories: tracheal/bronchiolar necrosis/apoptosis/sloughing, bronchial/bronchiolar exudates, bronchiolar hyperplasia, bronchiolar goblet cell metaplasia, endothelial hypertrophy/margination, peribronchiolar/perivascular inflammation, interstitial inflammation-septal thickening, alveolar inflammation, alveolar edema/hemorrhage, extent of alveolar involvement, consolidation and bronchiolization. Semi-quantitative scores used 0-1 = within normal limits, 2-14 = Minimal: Rare or inconspicuous lesions; 15-39 = Mild: Multifocal or small, focal, or widely separated, but conspicuous lesions; 40-79 = Moderate: Multifocal, prominent lesions; 80-99 = Marked: Extensive to coalescing lesions or areas of inflammation with some loss of structure and 100 = Severe: Diffuse lesion with effacement of normal structure. The scale to the right of the heat map reflects the scoring ranges from 0 to 70%.  The severity scores for each category of microscopic pathology are illustrated using a heatmap. For each of the days examined, each box represents the average score from four hamsters for scores for these categories: tracheal/bronchiolar necrosis/apoptosis/sloughing, bronchial/bronchiolar exudates, bronchiolar hyperplasia, bronchiolar goblet cell metaplasia, endothelial hypertrophy/margination, peribronchiolar/perivascular inflammation, interstitial inflammation-septal thickening, alveolar inflammation, alveolar edema/hemorrhage, extent of alveolar involvement, consolidation and bronchiolization. Semi-quantitative scores used 0-1 = within normal limits, 2-14 = Minimal: Rare or inconspicuous lesions; 15-39 = Mild: Multifocal or small, focal, or widely separated, but conspicuous lesions; 40-79 = Moderate: Multifocal, prominent lesions; 80-99 = Marked: Extensive to coalescing lesions or areas of inflammation with some loss of structure and 100 = Severe: Diffuse lesion with effacement of normal structure. The scale to the right of the heat map reflects the scoring ranges from 0 to 70%.

Viral Load, Neutrophil Presence and Lung Pathology in Syrian Hamsters Peaked at 4 dpi and Showed Resolution by Eight Days Post-Infection except for and Bronchiolization
Viral antigen (Figures 4-8; see B, D, F, H, J) and immune cell infiltrates (Figures 4-8; see B, D, F, H, J) were widespread in the bronchial epithelium and lower airways starting on 1 dpi and continuing through 2 dpi (Figures 4 and 5; see E, C, G, I). On 2 dpi, a notable increase in debris and virus was noted in the lower airways (Figure 5C,E,G,I). By 3 dpi the numbers of ciliated cells in the large airways were reduced and disorganized ( Figure 6F). We noted hyperplasia in the bronchial epithelia on 3 dpi. Consolidation was apparent around blood vessels ( Figure 6I) with colocalization of the virus in consolidated areas of alveolar epithelial cells ( Figure 6J). Hyperplasia was observed at 4 dpi with bronchiolization as well as the proliferation of the bronchial epithelial and pneumocytes ( Figure 6C,I) and vasculitis ( Figure 7J). Neutrophils were observed in airways over the first 4 dpi (Figures 4-7; see E). By 8 dpi, virus antigen was absent ( Figure 8B,D,F,H) and there was decreased immune cell infiltration and neutrophils in the bronchial epithelium ( Figure 8E, J). The bronchial epithelium continued to show hyperplasia ( Figure 8C,D). Eosinophils were also present in the alveolar spaces ( Figure 8F).
We noted hyperplasia in the bronchial epithelia on 3 dpi. Consolidation was apparent around blood vessels ( Figure 6I) with colocalization of the virus in consolidated areas of alveolar epithelial cells ( Figure 6J). Hyperplasia was observed at 4 dpi with bronchiolization as well as the proliferation of the bronchial epithelial and pneumocytes ( Figure 6C,I) and vasculitis ( Figure 7J). Neutrophils were observed in airways over the first 4 dpi (Figures 4-7; see E). By 8 dpi, virus antigen was absent ( Figure 8B,D,F,H) and there was decreased immune cell infiltration and neutrophils in the bronchial epithelium ( Figure 8E, J). The bronchial epithelium continued to show hyperplasia ( Figure 8C,D). Eosinophils were also present in the alveolar spaces ( Figure 8F).

Pathology of Pulmonary Arteries and Veins in Syrian Hamsters Were Characterized by Extensive Leukocyte Transmigration
The large pulmonary arteries and veins were evaluated on 1, 2, 3, 4 and 8 dpi by H&E (Figure 9 and Figure S3). By 1 dpi, leukocytes were attached to the endothelium of pulmonary veins in all cases and the artery of one female hamster. An increased number of leukocytes penetrated through endothelial cells and the tunica intima by 2 dpi. By 3 dpi, neutrophils were abundant in the tunica media of the pulmonary arteries and veins. Inflammatory cells continued to infiltrate the endothelial cell layer of the arteries on 4 dpi. In the pulmonary veins on 4 dpi, we noted continued infiltration of inflammatory cells and active mitosis in endothelial cells. We also noted clusters of cells with single nuclei in the tunica media (smooth muscle cells). By 8 dpi, there was decreased inflammatory cell infiltration in blood vessels and a higher density of endothelial cells in the tunica intima.
hemorrhage. (B) A representative section of the lung was subjected to IHC for the discovery of viral antigens. No staining for viral antigen was observed. (C) A higher magnification image of bronchial epithelial cells shows hyperplasia and few inflammatory cells (D). No viral staining was detected in adjacent serial section to (C). (E) A representative image stained with NACE stain shows no neutrophils in mucosa or submucosa. (F) H&E imaging shows increased eosinophils and neutrophils in the interstitial tissues (arrow) of the lung. Alveolar epithelial are hypertrophic and hyperplastic. (G) Representative section of the lung showing alveolar bronchiolization and inflammatory cell infiltration. (H) An adjacent section of (G) probed by IHC for viral antigen. No viral antigen was detected.

Pathology of Pulmonary Arteries and Veins in Syrian Hamsters Were Characterized by Extensive Leukocyte Transmigration
The large pulmonary arteries and veins were evaluated on 1, 2, 3, 4 and 8 dpi by H&E (Figures 9 and S3). By 1 dpi, leukocytes were attached to the endothelium of pulmonary veins in all cases and the artery of one female hamster. An increased number of leukocytes penetrated through endothelial cells and the tunica intima by 2 dpi. By 3 dpi, neutrophils were abundant in the tunica media of the pulmonary arteries and veins. Inflammatory cells continued to infiltrate the endothelial cell layer of the arteries on 4 dpi. In the pulmonary veins on 4 dpi, we noted continued infiltration of inflammatory cells and active mitosis in endothelial cells. We also noted clusters of cells with single nuclei in the tunica media (smooth muscle cells). By 8 dpi, there was decreased inflammatory cell infiltration in blood vessels and a higher density of endothelial cells in the tunica intima. Figure 9. Selected H&E images of the large pulmonary blood vessels from Syrian hamster following infection with SARS-CoV-2 on 1, 2, 3, 4, and 8 dpi. In the sham inoculated group (PBS), the main pulmonary artery and veins display a normal character in each layer of the blood vessel. In subsequent rows, representative images of pulmonary artery and veins are presented for 1, 2, 3, 4, and 8 dpi. The pulmonary artery at 1 dpi is normal, although in one female hamster, a few leucocytes were attached to the tunica intima. At 1 dpi, the image of the pulmonary vein shows leucocytes attached to the wall of pulmonary veins. At 2 dpi, blood vessels show an increased level of inflammatory cells or leucocytes penetrating through endothelial cells, and tunica intima, and appeared to be greater in the vein. The image on 3 dpi shows neutrophils that have penetrated through tunica media of pulmonary artery (male) or less severe (female). In the image of the pulmonary veins, the inflammatory cells have clotted in the tunica media. The image on 4 dpi shows inflammatory cells passing through an intact endothelial cell layer of the artery (male) although some vacuoles are present in endothelial cells and in muscle cells (female). The image of the vein shows the infiltration and proliferation of mononuclear cells. Mitosis was also noted in the endothelial cells (arrow in female vein implies proliferation of endothelial cells). We also noted groups of increased cells with single nuclei in the tunica media (smooth muscle cells). By 8 dpi, there was a decrease in the level of inflammatory cells infiltrating the blood vessels (recovered), and a higher density of endothelial cells in tunica intima. Scalebar = 20 µm.

Intracardiac Platelet and Fibrin Aggregates in the Heart
In addition to the samples herein, we examined banked paraffin-embedded heart tissues from four studies using 6-9-week-old male (n = 8 SARS-CoV-2, n = 2 PBS-inoculated) or female (n = 18 SARS-CoV-2, n = 6 PBS-infected) hamsters. The body weight and viral load were measured for the four studies ( Figure S4). Across these studies, sagittal sections of the heart were evaluated by H&E and probed by IHC for the nucleocapsid antigen. No viral antigen was detected in over 80 different sections in several studies ( Figure 10C). Mild degeneration of the heart muscle was noted as soon as 1 dpi with the presence of vacuoles within cardiomyocytes in the right ventricle ( Figure 10A) and abnormal appearance of the coronary artery muscle layer ( Figure 10B). However, no other findings were noted with the SARS-CoV-2 WA20 strain in the examination of tissues across these studies until 8 dpi.
In one study of female hamsters, two of ten specimens of heart on 8 dpi showed platelet and fibrin aggregates in the right ventricle and atrium. In one animal, this is shown for the right ventricle ( Figure 10E) and both WBCs and RBCs were noted within the fibrin close to the tricuspid valve (higher magnification) ( Figure 10F). In the second hamster, we show the right atrium, but similar findings were noted in the right ventricle ( Figure 10G). Lastly, in a third, female hamster (8 dpi), we noted an abnormality in the endothelial cells with hyperplasia and WBCs near the damaged area (mononuclear cells and eosinophils) at the mitral valve ( Figure 10H). Examination of heart specimens from female animals on 8 dpi from three additional studies suggested one additional animal with platelet and fibrin aggregates out of a total of eight. Lastly, in two studies of male hamsters infected with SARS-CoV-2, one of eight hearts was noted to have platelet and fibrin aggregates. In summary, evidence for intracardiac platelet/fibrin aggregates was noted in male (13%) and female (17%) hamsters on 8 dpi, which might be indicative of a hypercoagulative state in these animals.

Discussion
Given the global impact of COVID-19 pneumonia, a preclinical animal model that recapitulates the disease is critical for the evaluation of therapeutics and vaccines [1,[21][22][23]. The results of this study of SARS-CoV-2 infection in male and female hamsters provide several parameters that can enable the evaluation of therapeutics and vaccines or mechanistic interrogations of disease pathways. In the context of a serial sacrifice study, we present evidence of hematological abnormalities, such as increased RBC count and hematocrit. Future studies will assess if this was driven due to dehydration and decreased oxygen levels. Neutrophil levels were pronounced in whole blood on 1 and 2 dpi (Figure 2), around blood vessels on 2 dpi ( Figure 5I), penetrating the tunica media on 3 dpi (Figure 9) in male and female animals and were present in lung debris, the epithelium and submucosal layer by 1 dpi ( Figure 4E) and 2-4 dpi ( Figures 5, 6 and 7E). By 8 dpi, neutrophils were no longer present in the mucosa or submucosa ( Figure 8E) but were present in the interstitial tissues of the lung along with eosinophils ( Figure 8F). Monocytes and eosinophils were noted at 4 dpi around blood vessels with proliferating endothelial cells ( Figure 7J).
We observed blood vessel injury with the loss of endothelial cells and tunica media, and diffuse damage in surrounding alveoli, pulmonary hemorrhage, and destruction of alveolar septa. This corroborates with the lung pathophysiology observed in human COVID-19 cases with significant multi-focal lesions indicative of moderate disease [9][10][11]. Endothelial injury of the pulmonary vessels is a principal determinant of microvascular dysfunction [24]. It results in shifting the vascular equilibrium towards vasoconstriction with subsequent organ ischemia, and inflammation, with associated tissue edema [24,25]. Our findings show the accumulation of inflammatory cells, with evidence of endothelial and inflammatory cell death, and suggest endothelialitis secondary to SARS-CoV-2 epithelial cell infection as a possible determinant of the pathological findings along with the host inflammatory response [24,25]. Immunothrombosis (i.e., upregulation of the immune and coagulation systems to block pathogens) driven by endothelialitis may be a major driving factor in lung injury observed in SARS-CoV-2 infection [24,25]. In the review by Bonaventura et al. [24], they propose that clinical manifestations of COVID-19 are driven by exaggerated immunothrombosis within the lung microvessels. In addition to regulating vascular tone, endothelial cells are critical for regulating thrombosis and maintaining an anti-thrombotic state. Currently, there are several proposed mechanisms of lung vascular disease and thrombosis formation during SARS-CoV-2 infection that include inflammation and cellular injury, NETosis (neutrophil extracellular traps (NETs) formation), complement activation, and injury mediated by platelet activation [26,27]. Our findings show an accumulation of inflammatory cells, with evidence of endothelial cell death. We did not observe direct infection of endothelial cells in areas where there was an accumulation of inflammatory cells and injury. Endothelialitis has been reported in one other study at 5 dpi in the COVID-19 hamster model with no detectable viral RNA [12]. Exposure of endothelial cells to lytic cell debris and proinflammatory cytokines can upregulate adhesion molecules to allow for the recruitment of leukocytes and platelets [26]. A lack of tropism for endothelial cells and the requirement of exposure to serum proteins may limit our ability to elucidate mechanisms of disease using in vitro assays.
Initial low platelet levels observed in our study on 2 dpi, may be due to platelet apoptosis [28]. Platelet count in human cases of COVID-19 may be dependent on the severity of infection [28,29]. In the serial sacrifice study, mild degeneration of the heart muscle was observed by 1 dpi, however, no other pathologies were detected. We expanded our analyses of the heart to include additional PBS or SARS-CoV-2-inoculated samples that we have banked as paraffin blocks from the start of the pandemic. In these four studies, male and female hamsters were intranasally infected with the 2 × 10 5 PFU of SARS-CoV-2 in the 6-9-week-old age range. Intracardiac platelet and fibrin aggregates were detected in 13-17% of the male and female hamsters ( Figure 10). In these specimens, we observed the platelet and fibrin in the right atrium and right ventricles. The PBS-inoculated hamsters from the studies reported herein, or from these studies, had no platelet and fibrin or any pathology. Viral antigen was not detected in over 80 sagittal sections from these studies. These findings suggest a hypercoagulative state in these animals. Moreover, our findings provide a premise to expand our examination of inflammatory markers, and coagulation pathways, to define the mechanisms driving the progression of heart failure to the chronic stages and the potential involvement of the pulmonary right ventricular axis.
RBC alterations and thrombosis in COVID-19 disease need to be further investigated in animal models. While the incidence of cardiac abnormalities was low in the 6-9-weekold COVID-19 hamster model, it is possible that this may increase in older hamsters or hamsters with comorbidities, such as diabetes. As our studies used 6-to 9-week-old hamsters, our future studies will evaluate incidence in older age groups with the approaches and markers identified herein [18]. Further development of COVID-19 animal models is important for the elucidation of mechanisms and to provide a strong premise for the development of therapies and interventions targeting cellular and molecular mechanisms involved in the development of cardiac dysfunction which is key to the progression and severity of the COVID-19 disease process [30,31]. Our findings in the hamster model (with mild/moderate infection) are reflective of the pathology observed in SARS-CoV-2 infection of human subjects in the acute stages. However, patients with mild/moderate infection are not severely ill, thereby limiting access to the availability of specimens for histopathological analysis. We observed in our model early evidence of cardiopulmonary failure, and early thrombus formation. Further work to elucidate the beneficial effects of the early institution of anti-platelet, anti-coagulant, fibrinolytic and/or thrombolytic therapy to ameliorate the prothrombotic and procoagulant state would be of significant benefit and lead to further evaluation in afflicted human subjects and have significant translational potential. Finally, understanding the complexity and progression of cardiac involvement using additional approaches, such as echocardiography of the hamster-COVID-19 model, will advance translation [30,31].
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/v14071403/s1, Figure S1. Weight and viral load in hamsters following intranasal infection with SARS-CoV-2 at a low and high dose.; Figure S2. Representative images of male Syrian hamster lungs from mock or SARS-CoV-2 infection.; Figure S3. Additional pathology and immunohistochemistry of the large pulmonary blood vessels from Syrian hamster following infection with SARS-CoV-2. Table S1