SARS-CoV-2 Experimental Infection of African green monkeys using the LMA MAD
We challenged six healthy, adult AGMs with a target dose of 3.0 x 106 PFU of SARS-CoV-2 (SARS-CoV-2/INMI1-Isolate/2020/Italy) via intranasal inoculation with the LMA MAD (actual delivered dose of 2.8 x 106 PFU). Three animals were euthanized at 5 days post-infection (dpi) which is thought to be the approximate time point of peak disease in AGMs [10], while the remaining three animals were euthanized at 34 dpi during early convalescence. Blood and mucosal swabs were sampled from all animals on days 0, 2, 3, 4, 5, and additionally on days 7, 9, 12, 15, 21, 28, and 34 for AGM-4, AGM-5, and AGM-6. BAL fluid collection was performed on days -8, 3, and 5 for all animals, as well as 7 dpi for AGM-4, AGM-5 and AGM-6. Consistent with our previous report describing the development of the combined intranasal and intratracheal SARS-CoV-2 challenge model in AGMs [10], we did not observe overt signs of clinical illness in any AGMs in this study, other than decreased appetite or brief (single day) anorexia (Supp Table 1). Temperature was longitudinally monitored in 15 minute increments for the entire study duration using surgically implanted temperature loggers; several animals (AGM-4, AGM-6) experienced brief (< 2 hours) periods of mildly elevated temperatures at 3 dpi, and two animals (AGM-2, AGM-3) exhibited an abnormal temperature cycling pattern at 3 dpi (Supp Figure 1).
As in our previous report, transient shifts in leukocyte populations, predominately manifested as lymphocytopenia (5/6 animals), thrombocytopenia (3/6 animals), and granulocytosis (defined by neutrophilia, eosinophilia, and/or basophilia) (6/6 animals) were observed, while markers for renal (BUN, CRE) and hepatic function (ALT, AST, ALP, GGT) remained unchanged for the most part, with the exception of mild (≤ 2-fold) increases in ALT (2/6 animals), and mild to moderate (1 to 16-fold) increases in CRP, a marker of acute systemic inflammation (5/6 animals) (Supp Table 1), although statistical significance was not reached for most parameters at most time points (Figure 1). In addition, hypercapnia (defined here as ≥ 4 mmHg increase in dissolved CO2) was observed in 3/6 animals (Supp Table 1), which as we observed previously [10], appeared to follow a biphasic pattern (Figure 1A, data shown as fold-change from baseline]).
All animals exhibited normal prothrombin times (PT) as compared to their individual baseline values; however, mild to moderate prolongation of the activated partial thromboplastin time (aPTT) was also observed in all animals through the acute phase of disease, most prominently in AGM-1 and AGM-2, indicating possible disorder of the intrinsic coagulation pathway (Figure 1H, I); this was mirrored by increased levels of circulating fibrinogen (Figure 1J). We previously showed that the pathways connected to IL-6 production are activated during SARS-CoV-2 infection of AGMs [10], indicating possible mechanisms of coagulopathy in the current study.
All animals seroconverted, with weakly neutralizing titers (as quantified by PRNT50) being detected as early as 5 dpi and gradually increasing in potency by 34 dpi, with terminal neutralizing antibody titers ranging from ~1:16-1:128 (Figure 2A-E). We next quantified SARS-CoV-2 nucleoprotein specific IgG by ELISA (Figure 2F). Seroconversion was not detected until day 15 in two animals (AGM-4 & AGM-5). Interestingly, not until 34 dpi was a modest level (1:800) of seroconversion detected in the third animal.
Quantification of viral load in blood, mucosal swabs, and lungs
Viral RNA (vRNA) was purified from whole blood, oral, nasal and rectal mucosa, and BAL fluid from all collection days, as well as from lung tissue harvested at necropsy. As we previously reported [10], we were unable to detect SARS-CoV-2 vRNA in whole blood by RT-qPCR, nor were we able to recover infectious virus in the plasma fraction by plaque assay, confirming a lack of either cell-associated or freely-circulating virus in the peripheral blood. SARS-CoV-2 vRNA and infectious virus was detected in the nasal mucosa from all animals as early as 2 dpi, with vRNA persisting in a single animal up to 15 dpi (Figure 3A, B). Likewise, vRNA was detected in oral swabs from all animals beginning 2-3 dpi before falling below the limit of detection by 7 dpi, while low quantities of infectious virus (1-2 log10 PFU/mL) were only isolated from three animals (AGM-4, AGM-5, and AGM-6) (Figure 3C, D). Remarkably, vRNA was transiently shed from the lower gastrointestinal tract up to 28 dpi (AGM-4 and AGM-6), although infectious virus could only be recovered from the rectal swab of a single animal (AGM-3) 4-5 dpi (Figure 3E, F). vRNA was detected in BAL fluid from 4/6 animals 3 dpi and up to 7 dpi in all three animals held past 5 dpi, while infectious virus was recovered from 3/6 animals (Figure 3G, H). Detectable quantities of vRNA were absent from lungs harvested during necropsy of AGMs euthanized 34 dpi, while 6-9 log10 GEq/g were detected from all three animals euthanized at 5 dpi (Figure 3I).
Gross pathology, histopathology, and immunohistochemistry
Necropsy was performed on all animals following euthanasia, and lungs were collected for gross examination and histopathological analysis. Consistent with our previous study utilizing a combined i.n. and i.t. inoculation route [10], all AGMs displayed varying degrees of pulmonary consolidation with hyperemia and hemorrhage, characterized by depressed and patchy dark red to light pink regions (Fig. 4, arrows). In all AGMs, the most severe lesions were located in the dorsal aspects of the lower lung lobes. A board-certified veterinary pathologist approximated lesion severity for each lung lobe (Supp Table 2). All AGMs at 5 dpi also had segmentally flaccid and gas distention of small intestines. There were no other significant gross lesions.
Histologically, all three AGMs euthanized at 5 dpi developed mild multifocal neutrophilic bronchointerstitial pneumonia (Figure 5A-E, O). Histologic features include acute inflammation centered within the airways of terminal bronchioles with occasional flooding of adjacent alveolar spaces with neutrophils, macrophages, fibrin, edema, hemorrhage, mucous and rarely multinucleated giant cells (5A, B). In lesser-affected regions alveolar septate were expanded with mixed inflammatory cells and alveolar spaces contain increased numbers of alveolar macrophages with scattered red blood cells. Ulcerative tracheobronchitis was also present in all three AGMs and characterized by multifocal epithelial erosion associated with underlying hemorrhage, fibrin accumulation and infiltrating acute inflammation. Polymerized fibrin, highlighted by IHC, colocalized with acute inflammation within the bronchial lumen, alveolar spaces, alveolar walls and ulcerated regions of the trachea and bronchus (Figure 5C). Fibrin was also present within medium and small caliber vessels but was not associated with an obvious adherent thrombus. Trichrome stain of representative lung sections identified modest collagen deposition within multifocal regions of alveolar septae (Figure 5D). IHC for SARS-CoV-2 antigen was positive in all three AGMs associated with pulmonary lesions. Positive IHC labeling was noted diffusely within the cytoplasm of respiratory epithelium of the bronchus (Figure 5O) and less in type I and type II pneumocytes (Figure 5E).
Histologically, all three AGMs euthanized at 34 dpi developed moderate multifocal chronic interstitial pneumonia Figure (5F-J). Histologic features include expansion of alveolar septae with macrophages, lymphocytes, and very rarely neutrophils (Figure 5F, G). Wispy, pale eosinophilic, acellular material also multifocally expanded the alveolar walls and stained as immature collagen with trichrome staining (Figure 5I). Polymerized fibrin was present within medium and small caliber vessels but was not associated with an obvious adherent thrombus (Figure 5H). No immunolabeling for SARS-CoV-2 was noted with IHC in any of the examined tissue sections from this 34 dpi cohort (Figure 5J).