Pathology and Pathogenesis of SARS-CoV-2 Associated with Fatal Coronavirus Disease, United States

An ongoing pandemic of coronavirus disease (COVID-19) is caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Characterization of the histopathology and cellular localization of SARS-CoV-2 in the tissues of patients with fatal COVID-19 is critical to further understand its pathogenesis and transmission and for public health prevention measures. We report clinicopathologic, immunohistochemical, and electron microscopic findings in tissues from 8 fatal laboratory-confirmed cases of SARS-CoV-2 infection in the United States. All cases except 1 were in residents of long-term care facilities. In these patients, SARS-CoV-2 infected epithelium of the upper and lower airways with diffuse alveolar damage as the predominant pulmonary pathology. SARS-CoV-2 was detectable by immunohistochemistry and electron microscopy in conducting airways, pneumocytes, alveolar macrophages, and a hilar lymph node but was not identified in other extrapulmonary tissues. Respiratory viral co-infections were identified in 3 cases; 3 cases had evidence of bacterial co-infection.


Study Patients and Data Collection
As part of the public health response to COVID-19, the CDC Infectious Diseases Pathology Branch (Division of High-Consequence Pathogens and Pathology, National Center for Emerging and Zoonotic Infectious Diseases) was consulted on autopsies of 8 patients with laboratory evidence of SARS-CoV-2 by reverse transcription PCR (RT-PCR) on respiratory swab specimens collected either before or after death. We reviewed available medical records and preliminary autopsy reports for information regarding demographics, symptom history, underlying conditions, infectious disease testing, imaging study findings, treatment and advanced supportive care received, and date of death. This investigation was reviewed in accordance with CDC's human subjects review procedures and was determined to not meet the definition of research.

Histopathology and Immunohistochemistry
We performed routine hematoxylin-eosin stains for histopathologic evaluation. We conducted an IHC assay for SARS-CoV-2 using a rabbit polyclonal antibody raised against SARS-CoV nucleocapsid (Novus Biologicals, https://www.novusbio.com) (19) at 1:100 dilution and a Mach 4 Universal AP Polymer Kit (Biocare Medical, https://biocare.net) with Permanent Red Chromogen (Cell Marque/Millipore Sigma, https://www.cellmarque.com). We pretreated the slides with heat-induced epitope retrieval with a citrate-based buffer (Biocare Medical). We ran appropriate negative controls in parallel, using normal rabbit serum in place of the primary antibody. We validated cross reactivity of the anti-SARS-CoV antibody with SARS-CoV-2 by testing controls created from SARS-CoV-2-infected Vero cells embedded with normal human tissues; we used this control as the positive control for subsequent IHC assays. The SARS-CoV nucleocapsid antibody did not cross-react with influenza A(H1N1) virus, influenza B virus, respiratory syncytial virus, parainfluenza virus type 3, human coronavirus (HCoV) 229E, or MERS-CoV in PCR-confirmed tissue samples. For cases with bronchopneumonia, we performed IHC testing for bacterial agents using a mouse monoclonal antibody raised against Streptococcus pneumoniae but known to also detect other Streptococcus spp. and a rabbit polyclonal anti-Klebsiella pneumoniae antibody (both from Thermo Fisher, https://www.thermofisher.com) known to also detect other gram-negative bacteria (Escherichia coli, Haemophilus influenzae, and Pseudomonas spp.).
For double-stained assays, we used Envision G/2 Double Stain System, Rabbit/Mouse (DAB Permanent Red) from Agilent Technologies (https://www. agilent.com). We used antibodies against CD163 (Leica Biosystems, https://www.leicabiosystems.com) or surfactant apoprotein A (Dako, https://www. agilent.com), followed by the anti-SARS-CoV nucleocapsid antibody (Novus Biologicals). We performed all assays according to the manufacturer's guidelines. We used SARS-CoV-2-infected Vero cells as a positive control and used non-COVID-19 cases and normal rabbit serum in place of primary antibody as negative controls.

Electron Microscopy
We obtained upper airway and lung tissue specimens from formalin-fixed samples, cut them into cubes, rinsed them with 0.1 mmol/L phosphate buffer, postfixed them with 2.5% glutaraldehyde, and rinsed them in phosphate buffer. In addition, we removed tissue samples from areas corresponding to positive SARS-CoV-2 immunostaining from paraffin blocks with a 2-mm punch or from 4-µm sections on glass slides; we deparaffinized the samples in xylene and rehydrated. We processed tissues for transmission EM as described previously (20). We immersed the sections embedded in epoxy resin in boiling water, removed them from the slides with a razor blade, and cut out the areas of interest and glued them onto a blank EM block. We stained EM sections with uranyl acetate and lead citrate and examined them on a Thermo Fisher/FEI Tecnai Spirit or Tecnai BioTwin electron microscope.

Clinical Data
Of the 8 case-patients, 7 were residents of a longterm care facility (LTCF) in Washington state (Table  1) (23,24). Seven (87.5%) were White, non-Hispanic. The median age of the 8 case-patients was 73.5 years; 2 were <65 years of age. The median number of days from illness onset until death was 12.5 (range 6-15). Common signs and symptoms reported included fever (75%), cough (62.5%), and shortness of breath (62.5%). All patients had abnormal findings on chest radiographs. Case-patients were all hospitalized for a median of 3.5 days. Six (75%) patients received mechanical ventilation; 2 received comfort care. Underlying medical conditions were identified in all case-patients; hypertension (75%), chronic kidney disease (75%), cardiovascular disease (75%), obesity (62.5%), and diabetes (50%) were the most frequent conditions reported.

Histopathology, Immunohistochemistry, and Electron Microscopy
Histopathologic findings and results of testing performed on FFPE tissues showed mild to moderate tracheobronchitis was consistently present and characterized by mononuclear inflammation, with epithelial denudation and submucosal congestion (Table 2; Figure 1, panels A, B). The predominant lung pathology was diffuse alveolar damage (DAD); acute phases, organizing phases, or both were present in 7 (87.5%) of 8 patients. Desquamation of pneumocytes and the presence of hyaline membranes, alveolar edema and fibrin deposits, type II pneumocyte hyperplasia, and alveolar infiltrates, including increased alveolar macrophages, were seen ( Figure 1, panels C, D). Squamous metaplasia and atypical pneumocytes were present in 3 case-patients, and rare multinucleated cells were present in 1 case-patient (patient no. 1) (Figure 1, panel E); no definitive viral inclusions were seen. One case-patient without DAD (patient no. 7) had diffuse bronchopneumonia with filling of alveolar spaces by mixed inflammation with abundant neutrophils (Figure 1, panel F). Three additional casepatients had focal bronchopneumonia and increased pulmonary intravascular leukocytes. Hemosiderinladen macrophages (4/8), hemorrhage (4/8), mucus aspiration (3/8), emphysema (2/8), and microthrombi (1/8) were seen ( Figure 2). Anthracosis, common in elderly persons as a result of chronic carbon accumulation, was present in the lungs and pulmonary hilar lymph nodes in all cases. In 6 cases, lymph nodes also showed sinus histiocytosis and hemophagocytosis in subcapsular sinuses. Notable pathologic findings in extrapulmonary tissues included evidence of chronic renal disease (5/8), acute renal tubular injury (3/8), hepatic steatosis (4/8) and cirrhosis (1/8), and focal myocardial fibrosis (3/8) (Figure 2). No myocarditis or myocardial necrosis and no notable histopathologic changes in the intestine were seen in any case. Brain tissues were not available for histopathologic evaluation or testing.
We detected SARS-CoV-2 by IHC in the upper airways in 4/8 (50%) case-patients and in the lungs in 7/8 (92%) case-patients. We observed immunostaining of viral antigens in upper airway and bronchiolar epithelium, submucosal gland epithelium, and in type I and type II pneumocytes, alveolar macrophages, and hyaline membranes in the lung ( Figure  3, panels A-C, F). Upper airways and lung tissues *Chronic kidney disease requiring treatment with hemodialysis was reported in 2 case-patients. †Number of cases includes patients with history of coronary artery disease, congestive heart failure, valvular heart disease, and/or cerebrovascular accident. ‡Obesity includes reported BMI >30 or mention of obesity or morbid obesity in the submitted medical records. Two patients had morbid obesity listed in the submitted medical records.
§Chronic obstructive pulmonary disease was the only reported underlying condition with chronic lung disease. ¶This includes persons immunosuppressed by solid organ transplant, corticosteroid therapy, or immunosuppressive medication. #Includes any of the following: hyperlipidemia, thyroid disease, rheumatologic disorder, and obstructive sleep apnea. **Two patients were transitioned to comfort care and were not intubated.  Figure 3, panel E). We also found viral immunostaining in scattered macrophages in the hilar lymph node from 1 severely immunosuppressed patient with a history of solid-organ transplant (Figure 3, panel G). We did not detect SARS-CoV-2 by IHC in heart, liver, kidney, spleen, or intestine from any patient. Six (75%) of 8 case-patients had either viral or bacterial co-infections, but not both, identified by IHC, PCR, or both in addition to SARS-CoV-2. Respiratory viral PCR testing detected parainfluenza virus type 3 coinfection in upper airway and lung tissue in 2/8 (25%) case-patients and influenza B virus coinfection in upper airway in 1 case-patient. Three (75%) of the 4 case-patients with SARS-CoV-2 and bronchopneumonia had immunostaining for Streptococcus spp. Two of these patients had nonpneumococcal Streptococcus spp. confirmed by PCR testing.
EM examination of respiratory tissues showed virions with prominent surface projections (spikes) characteristic of the family Coronaviridae. In the lung, extracellular virions free in the alveolar space were, on average, 105 nm in diameter, including surface projections (Figure 4, panel A). In upper airways, virions were seen extracellularly among the cilia and within the cytoplasm of respiratory epithelial cells

Discussion
The clinical distinction between SARS-CoV-2 and other respiratory viral infections is difficult because there are overlapping clinical features characterized by febrile illness with cough that lasts for several days before progressing to acute pneumonia. In addition, persons with SARS-CoV-2 and other respiratory viral infections may have atypical or minimal symptoms (25)(26)(27). Besides respiratory failure, particularly in patients with severe disease, fatigue, myalgia or arthralgia, chills, hepatic and renal dysfunction, lymphocytopenia, leukopenia, thrombocytopenia, and elevated inflammatory biomarkers have been described (5,28,29).
Histopathologic lesions attributed directly to the virus in these cases were limited to respiratory tissues; the predominant finding was DAD, with various levels of progression and severity. We saw no clear correlation of the pathologic phase of DAD to known symptom duration, which could be the result of underrecognition of early symptoms in elderly residents of LTCFs and underestimation of illness duration. Together, the histopathologic, IHC, and  Tracheobronchitis  (15,20). Viral antigen was also seen in respiratory epithelium of conducting airways (trachea, bronchi, and bronchioles) and occasionally in alveolar macrophages; infection of these cell types may be key in viral replication and trafficking. The respiratory epithelium is one of the first cell types encountered by inhaled virus; SARS-CoV-2 antigens were detected by IHC in ciliated epithelial cells from 50% of these case-patients and up to 16 days after known symptom onset. Ultrastructural analysis showed numerous extracellular viral particles along the ciliated surface and within ciliated columnar epithelial cells. These findings corroborate reports of high viral loads in the upper respiratory tract and support the potential for persons infected with SARS-CoV-2 to readily transmit the virus, with prolonged and continued viral shedding in severe cases (30,31). Overall pathologic features in these 8 COVID-19 deaths were similar to those seen in SARS-CoV and MERS-CoV infections, and in available COVID-19 reports (10)(11)(12)(13)(14)(15)18,20,32). However, the amount of viral antigen detected by IHC in lung tissue from these cases is more than what we have seen in SARS and MERS (16,18) cases submitted to our laboratory, and its extensive detection in epithelial cells of the upper respiratory tract is unique among these highly pathogenic coronaviruses (33). In addition to direct viral effects on tissues, the immune response to viral infection likely plays a major role in determining clinical outcome, and acute decline in COVID-19 patients has been linked to an immune-mediated cytokine storm (34). Preliminary evaluation of immune cell populations in the respiratory tissues from these 8 cases revealed abundant T lymphocytes in the upper airways and lung parenchyma, with B lymphocytes in smaller numbers and predominating in areas of lymphoid aggregates (data not shown). Further investigation into the roles of these cell populations in COVID-19 is needed. In 6 of the 8 case-patients, sinus histiocytosis and hemophagocytosis were seen in hilar lymph nodes. However, SARS-CoV-2 anti-gens were detected by IHC in hilar lymph node macrophages from only 1 immunosuppressed patient. Lymph nodes are key sites for immune recognition and elimination of respiratory pathogens. Elucidating the immune response to, and the effects of immunosuppression on, SARS-CoV-2 infection is therefore of fundamental importance. SARS-CoV-2 uses the angiotensin-converting enzyme 2 (ACE2) receptor to facilitate viral entry into target cells. ACE2 is expressed in multiple tissues throughout the body, including type II pneumocytes, myocardial cells, cholangiocytes, enterocytes, and oral mucosal epithelium (5,35,36). However, among these patients, SARS-CoV-2 antigens were not detected in extrapulmonary tissues besides hilar lymph node, and pathologic findings in other tissues were 2010 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 26, No. 9, September 2020 attributable to other underlying concurrent conditions. Some of the underlying conditions in these case-patients (e.g., hypertension, COPD) are associated with upregulation of ACE2 receptors; possible correlation of these conditions with COVID-19 severity warrants further exploration (5,28,37). COVID-19 cardiomyopathy and acute cardiac death during clinical resolution of pulmonary disease have been described (24,38). However, we did not observe any evidence of myocarditis or myocardial necrosis in the tissues of the 8 case-patients we examined. Reports have been made of coagulation abnormalities and pulmonary vascular perfusion issues without DAD in some CO-VID-19 patients (39,40), and we saw microthrombi in the lung from 1 case-patient who lacked DAD but had severe bacterial bronchopneumonia. These various and potentially severe cardiovascular complications of COVID-19 warrant further investigation into the specific mechanisms of SARS-CoV-2-induced cardiovascular injury, homeostatic derangement, or both. Clinical studies have reported elevated liver enzymes in patients with COVID-19 (28,29). The lack of viral detection by IHC in the liver in this investigation suggests that for these case-patients, abnormal biomarkers of hepatic injury may not be the result of direct viral infection of hepatocytes. Gastrointestinal symptoms are not typically a prominent feature of COVID-19 but have been reported, and SARS-CoV-2 has been detected in fecal samples (41)(42)(43). However, no histopathologic findings or SARS-CoV-2 antigens were detected in gastrointestinal tissues, and diarrhea was reported for only 1 of these case-patients. We identified viral co-infections in upper respiratory tract tissues from 3 case-patients, including 2 with parainfluenza virus 3 and 1 with influenza B virus, but the contribution of these co-infections to pulmonary disease and fatal outcomes is unknown. Although we identified streptococcal lower respiratory infections in 3 case-patients, none were caused by Streptococcus pneumoniae, and there was no strict correlation of these infections with mechanical ventilation among these case-patients. Because 7 of 8 case-patients discussed in this report were residents of a LTCF, their exposures and risks for viral and bacterial co-infections may be different from those for other patients. Few community-acquired bacterial infections have been reported in critically ill patients with COVID-19, but co-infections are not frequently reported with SARS and MERS (17,32,44). Co-infections may play a key role in increasing susceptibility to, and illness from, SARS-CoV-2 infection in a LTCF setting. Further investigation into this association, and characterization of the etiologic agents most commonly involved, is warranted and may contribute to improved overall management of COVID-19 disease. This report describes the specific cellular and extracellular localization of SARS-CoV-2 in respiratory tissues, without any IHC evidence of the virus in other tissues. Although detection of SARS-CoV-2 RNA in blood or serum has been reported (34,41), we did not find evidence of systemic virus dissemination in these case-patients. Our findings highlight the importance of underlying conditions and pulmonary co-infections in COVID-19; these factors may potentially delay or confound diagnosis and contribute to adverse outcomes.
A limitation of this study is that 7 of 8 cases were from a single skilled nursing facility; findings may therefore not be representative of community-acquired SARS-CoV-2 infections. However, nosocomial transmission of viruses often parallels community outbreaks, and understanding disease transmission in healthcare settings is crucial (25,45). None of these case-patients had diagnoses of acute cardiac injury, myocarditis, or cardiomyopathy, so understanding the pathogenesis of cardiac injury with SARS-CoV-2 infection requires additional investigations in fatal cases with evidence of cardiac injury.
No clinical or histopathologic features are specific to SARS-CoV-2 infection. Demonstrating SARS-CoV-2 directly in lung tissue, when taken in context with any other pathology present, is critical to assessing its contribution to mortality. Herein, we establish the utility of IHC as a diagnostic modality for SARS-CoV-2 in FFPE tissues by localizing viral antigens in respiratory tissues from RT-PCR confirmed cases. This diagnostic method is particularly valuable for FFPE specimens from cases in which antemortem or postmortem respiratory swab testing for SARS-CoV-2 was not performed. We also demonstrate virus identification in tissues by EM using various tissue sources (formalin-fixed wet tissue, FFPE blocks, and stained slides). Identification of SARS-CoV-2 cellular tropisms in the respiratory tract represents a crucial step forward in understanding the pathogenesis of SARS-CoV-2 infection and provides some insights relevant to the development of targeted therapeutic and preventive measures to combat COVID-19.