Abstract

Plain language summary

Introduction

Coronavirus disease-19 (COVID-19), an infectious disease caused by a novel coronavirus called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was declared a pandemic by the World Health Organization on March 11, 2020. As of May 21st, 2021, 165,874,001 diagnosed cases and 3,438,383 deaths have been reported worldwide (https://coronavirus.jhu.edu/), with the United States of America bearing the highest disease impact in terms of morbidity and mortality (33,084,872 cases and 589,222 deaths) followed by India where is currently undergoing a massive surge of cases. COVID-19 is an established multiorgan disease in humans with the greatest involvement and derangement involving the respiratory, cardiovascular, renal, and immune systems¹.

Coronaviruses are a group of enveloped viruses with a single-stranded RNA genome. In rare instances, animal coronaviruses such as severe acute respiratory syndrome (SARS) coronavirus (SARS-CoV) in 2002, Middle East respiratory syndrome (MERS) coronavirus (MERS-CoV) in 2012, and currently SARS-CoV-2, can infect humans with variable clinical complications and impact. All three of these coronaviruses are transmitted zoonotically and spread among humans through close contact². SARS-CoV-2 is most commonly transmitted from person to person via respiratory droplets. Most patients experience mild symptoms. However, a significant subset of patients may experience severe disease outcomes including long term health sequelae and death³. The SARS-CoV-2 genome shares ~80% sequence identity with SARS-CoV and MERS-CoV, although the sequence similarity varies among genes encoding structural proteins and essential enzymes. SARS-CoV-2 is considered to be more pathogenic than the MERS-CoV and SARS-CoV^4,5. Structurally, SARS-CoV-2 is comprised of four structural proteins: spike (S), envelope (E), membrane glycoprotein (M), and nucleocapsid phosphoprotein (N) proteins⁵. SARS-CoV-2 entry into host cells depends on binding of viral spike proteins to the host receptor protein angiotensin-converting enzyme 2 (ACE2) and priming by the host serine protease, the cell surface transmembrane protease serine 2 (TMPRSS2)^6,7. In our recent study detailing transcriptional regulation of SARS-CoV-2 entry factors ACE2 and TMPRSS2, we described co-expression of the androgen receptor (AR), TMPRSS2, and ACE2 in bronchial and alveolar cells in human and murine lung tissues; demonstrated that androgens regulate the expression of ACE2, TMPRSS2, and AR in subpopulations of lung epithelial cells; and reported that transcriptional repression of the AR enhanceosome by AR antagonists inhibited SARS-CoV2 infection in vitro⁸. To further investigate transcriptional inhibition of critical host factors in the treatment or prevention of COVID-19, we set out to characterize the distribution and localization of SARS-CoV-2 viral particles, viral replication, and the host entry factor receptor ACE2, priming protease TMPRSS2 and transcriptional regulator AR by RNA in situ hybridization (RNA-ISH), immunohistochemistry (IHC) and qRT-PCR in various pulmonary and nonpulmonary tissues obtained from a series of six clinical autopsies performed on patients who succumbed to COVID-19 disease. We also correlated the presence of SARS-CoV-2 viruses and host entry machinery with clinicopathologic findings in the affected organ systems. We detect SARS-CoV-2 viral RNA and viral replication events in both pulmonary tissues and nonpulmonary tissues. We observe heterogeneity in viral load and viral cytopathic effects among various organ systems, between individuals and within the same patient. We also find the presence of host factors, ACE2, TMPRSS2, and AR, to overlap with the infection sites.

Methods

Results

Pulmonary histopathologic findings

Histologic examination of lung tissues revealed acute diffuse alveolar damage (DAD) with prominent hyaline membranes in 5/6 patients (Fig. 1a) and diffuse lymphocytic inflammatory infiltrates in all six patients. Evidence of coagulopathy (i.e., microthrombi, capillaritis etc.) was identified in a majority of patients (5/6). Additional histopathological findings attributed to viral cytopathic effects include pronounced usual hyperplastic alveolar pneumocytes (2/6 patients) and intracytoplasmic vacuolation (1/6 patient). Another histopathological feature observed in the cohort, possibly not directly attributed to COVID-19, was bronchopneumonia (3/6 patients). One patient with underlying asthma showed associated changes including mucus plugging, goblet cell metaplasia, mucus gland hyperplasia, and thickening of subepithelial basement membranes, while another patient with chronic obstructive pulmonary disease (COPD) showed large emphysematous changes and edema. A summary of pulmonary histopathologic findings is presented in Supplementary Table 2.

Open in a separate window

Fig. 1

SARS-CoV-2 virus detection in pulmonary parenchyma by SARS-CoV-2 spike RNA-ISH.

The viral signals were observed within the intra-alveolar hyaline membranes (a H&E and b SARS-CoV-2 spike RNA-ISH) and within the intact lining alveolar epithelial cells as well as in the desquamated alveolar epithelial cells (c H&E and d SARS-CoV-2 spike RNA-ISH). Viral particles were also noted in the nonalveolar region like bronchus as a cluster of viral signals within peri-bronchial sero-mucinous glands, as well as within the lining pseudostratified respiratory bronchial epithelium. (e H&E and f SARS-SoV-2 spike RNA-ISH). Viral signals were also detected in the necrotic and fibrinous material within damaged pulmonary tissue (g H&E and h SARS-CoV-2 spike RNA-ISH). Inset: Viral signals shown as individual punctate brown dots and clusters. Scale bars=100µm. Inset scale bar=10µm.

Detection of SARS-CoV-2 virus in pulmonary tissues by RNA-ISH

We performed RNA-ISH using a target probe against the SARS-CoV-2-S gene for detection of viral particles in formalin fixed paraffin embedded (FFPE) tissue sections. The SARS-CoV-2 RNA-ISH signals were identified as individual brown punctate dots or clusters. As expected, SARS-CoV-2 viral particles were detected in lung tissues of all six patients and no signal was observed in normal lung tissue or H1N1 influenza patient’s lung tissue. The viral loads, however, varied drastically among patients and different sampling sites from the same organ of a single patient. RNA-ISH signals were observed to be patchy. At anatomic level, the SARS-CoV-2 viral particles were observed in the intact as well as desquamated alveolar epithelial cells lying in the edema fluid of the intra-alveolar space and within the hyaline membranes lining the alveoli. (Fig. 1a–d). SARS-CoV-2 virus was also detected in the nonalveolar region such as peri-bronchial sero-mucinous glands in 2/4 patients (Fig. 1e, f) and peri-tracheal region in 1/4 patients, but with lower viral counts compared to the alveolar region as indicated by RNA-ISH staining. Viral signals in lung tissue were also noted to be prominent in areas exhibiting necrosis and sero-fibrinous deposits. (Fig. 1g, h).

As described in the methods, we also employed a sense probe against the minus strand of SARS-CoV-2-S gene to detect viral replication. Utilizing this approach, we observed active viral replication events as brown signal dots and clusters within the intra-alveolar fibrinous deposits, the cellular proliferation seen in the alveolar septa and bronchiolar region of lung parenchyma in all six patients. Signals were also identified in primary/parent bronchus tissues in 2/4 patients, and tracheal tissues in 1/4 patients (Fig. 2).

Open in a separate window

Fig. 2

Detection of SARS-CoV-2 viral replication in pulmonary parenchyma by SARS-CoV-2-S-sense RNA-ISH.

SARS-CoV-2 viral replication events were observed in the intra-alveolar septal region (a), hyaline membranes lining the alveolar space (b), ciliated columnar respiratory epithelium of the bronchus (c), and the subepithelial sero-mucinous tracheal glands (d). Inset: a signal cluster representing viral replication. Scale bars=200µm in a, 100 microns in b–d. Inset scale bar=10µm. Arrows point to RNA-ISH signals.

To demonstrate and confirm the specificity of the virus signals detected by RNA-ISH described above, we performed duplex RNA-ISH assay for co-detection of SARS-CoV-2-N gene and S gene in the lung autopsy tissue sections. We observed co-presence of both N gene and S gene signals within the same topographical locations of intra-alveolar region, hyaline membranes and peri-tracheal sero-mucinous glands (Fig. 3a, b). Similar strategy was utilized by combining the target probes against N gene and the minus strand of SARS-CoV-2-S gene, to show case SARS-CoV-2 viral replication event co-existing within the same cell types harboring the virus genome (Fig. 3c, d).

Open in a separate window

Fig. 3

Co-detection of SARS-CoV-2 virus spike (S), nucleocapsid (N) genes, and viral replication in pulmonary parenchyma.

Co-expression of the SARS-CoV-2-S (green) and N (red) genes in the intra-alveolar (a) and hyaline membranes in lung parenchyma (b) utilizing the duplex RNA-ISH with probes against the SARS-CoV-2-N gene and -S gene. Co-detection of the SARS-CoV-2 virus (red) and viral replication (green) in the intra-alveolar and hyaline membranes in lung parenchyma (c, d) with probes against the SARS-CoV-2-N gene and minus strand of S gene. (a, c from patient 1, b, d from patient 3) Scale bars=50µm.

As an additional confirmatory validation for RNA-ISH signal specificity, we treated the autopsy tissue sections with RNase. Upon RNase treatment just prior to SARS-CoV-2 target probe hybridization, RNA-ISH signals were no longer detectable within the tissue samples, which were earlier reported to be SARS-CoV-2 positive by conventional RNA-ISH (Supplementary Fig. 1).

Detection of SARS-CoV-2 virus in pulmonary tissues by qRT-PCR

We also performed qRT-PCR on eight SARS-CoV-2 RNA-ISH-positive pulmonary tissue samples from four patients using primer sets for E gene, N gene, and RdRP gene. All the selected pulmonary tissues were also qRT-PCR positive (Fig. 4). On the other hand, none of the genes were detected in normal lung tissue or H1N1 influenza patient’s lung tissue. The SARS-CoV-2 RNA-ISH results were highly consistent with the qRT-PCR results with an overall good correlation of signal intensities among pulmonary tissues (Supplementary Table 3).

Open in a separate window

Fig. 4

qRT-PCR of SARS-CoV-2 infection in COVID-19 autopsy tissues.

2^(Ct) for SARS-CoV-2 E gene, N gene, and RdRP gene using RNaseP housekeeping gene as reference. The undetermined Ct was set to 40. Normal lung, H1N1 influenza lung, and normal prostate tissues were used as negative controls. The 2^(Ct) of normal lung was set as a cutoff point (0.000224) and the gray area below the cutoff line is considered as SARS-CoV-2 negative. Minimum 1 out of the 3 genes should have Ct value<40 (or 2^(Ct)>0.000224) to be considered as SARS-CoV-2 positive.

To further confirm the RNA-ISH signals of viral replication, we performed minus strand-specific reverse transcription followed by E gene real-time PCR for lung samples positive for viral replication as detected by RNA-ISH. The PCR results are consistent with the RNA-ISH results for both S-sense RNA-ISH-positive and -negative samples. (Supplementary Table 4)

Detection of SARS-CoV-2 virus in nonpulmonary tissues

To further investigate the SARS-CoV-2 tissue distribution, we selected 14 representative nonpulmonary tissue samples, including kidney, heart, liver, lymph node, spleen, thyroid, prostate, uterus, colon, adrenal, pancreas, stomach, esophagus, and small intestine, using RNA-ISH and qRT-PCR methodologies. Low level signals with high cycle threshold (Ct) values were detected in kidney, heart, liver, lymph node, thyroid, spleen, prostate, uterus, colon, and adrenal by qRT-PCR. (Fig. 4 and Supplementary Table 3). Rare RNA-ISH signal dots were observed within the germinal center of the cortical lymphoid follicle of a lymph node sample from Patient 1 (Fig. 5) and within the distal tubules of a kidney sample form Patient 7 (Fig. 6c, d). Both tissues were qRT-PCR positive (Fig. 4, Supplementary Table 3, Patient 1 Lymph node with average Ct 30.8 and Patient 3 Kidney with average Ct 36.5). Very rare signal clusters were observed in kidney, heart, liver, thyroid, spleen, prostate, uterus, and colon tissue sections, but no individual signal dots representing viral particles were detected in any of those tissues. Therefore, we did not consider these tissues as SARS-CoV-2 positive by RNA-ISH.

Open in a separate window

Fig. 5

SARS-CoV-2 virus detection within lymph node.

Viral signal clusters were observed within lymph node germinal center as individual brown dots (a H&E and b SARS-CoV-2 RNA-ISH, in arrow pointed area). Scale bars=50µm.

Open in a separate window

Fig. 6

SARS-CoV-2 infection in kidney transplant patient (patient 3) on immunosuppressant therapy.

Highly abundant viral signals were observed within the lung parenchyma exhibiting diffuse alveolar damage, reflecting the high SARS-CoV-2 virus infection load in Patient 3. The viral signals were detected in intra-alveolar hyaline membrane and interstitial fibroblastic proliferation region (a, b. SARS-CoV-2 RNA-ISH). Scattered viral signals were identified in the renal tubules (c H&E and d SARS-CoV-2 RNA-ISH). Scale bar= 500µm in a, 100µm in b, 50µm in c, d.

Discussion

The highly contagious and rapidly spreading COVID-19 has evolved into a global health threat in 2020. This disease is now known to manifest with a wide spectrum of severity with severe disease resulting in intense and multiorgan dysfunction^13,14. Currently, there is a limited understanding of the pathophysiology of COVID-19, and no targeted therapy is yet available for treatment of established SARS-CoV-2 infection. As a follow-up to our recent study on transcriptional regulation of SARS-CoV-2 entry factors in lung⁸, herein, we studied the tissue distribution and localization of SARS-CoV-2 virus, viral replication, the transcripts of the host cell entry factors, ACE2 and TMPRSS2, and the transcriptional regulator of TMPRSS2, AR, in postmortem human organ systems using in situ and qRT-PCR approaches, and correlate those results to the histopathologic findings in COVID-19 patient autopsy cases.

Clinical autopsy examination revealed ARDS due to SARS-CoV-2 infection to be implicated in the demise of all six patients. The medical history in this series was enriched for at least one underlying or predisposing medical condition in each of the six patients including diabetes, hypertension, coronary artery disease, asthma, and obesity. These findings are consistent with previous reports in the literature suggesting older age (>60 years), serious co-morbidities, and tobacco exposure to be associated with a higher risk of developing ARDS and mortality in COVID-19 patients, especially in males¹³.

Histologic examination of pulmonary tissues revealed acute diffuse alveolar damage with prominent hyaline membranes as a common COVID-19 pathological feature, in concordance with the literature^1,15–17. In our study, the most common pathological changes encountered in nonpulmonary tissues included myofibrillary hypertrophy in the heart, tubular injury in the kidney, and steatosis in the liver. These observations are in line with published reports that SARS-CoV-2 infection may cause acute injury in heart, kidney, and other tissues through a direct cytopathic effect^14,18 In addition, catecholamine surge leads to cytokine storm imploding as systemic inflammatory response syndrome (SIRS), leading to multiorgan failure, which has been documented as a major cause of death in COVID-19 patients^1,14.

We further employed a single-molecule RNA-ISH technology for sensitive and specific detection of SARS-CoV-2 in FFPE tissue sections. This technology has been adapted to facilitate SARS-CoV-2 diagnosis and research, and has demonstrated SARS-CoV2-2 virus detection in cultured cells^19–21, rodent²², non-human primates^23,24, and human tissues autopsies^25–29. Employing RNA-ISH, we were able to concurrently document both the tissue distribution of SARS-CoV-2 viral particles and the host entry factors in pulmonary and extrapulmonary tissues. The SARS-CoV-2 RNA-ISH staining pattern and signal localization resembles those of the SARS-CoV-2 IHC signals. Our observations of RNA-ISH staining patterns in pulmonary tissues are consistent with published studies^28,30 and the viral localization matches the pulmonary histopathological findings in our study and other postmortem and retrospective studies^16,25,26 Among nonpulmonary tissues, very rare signal clusters were observed in kidney, heart, liver, thyroid, spleen, prostate, uterus, and colon tissue sections, but no individual signal dots were detected. Low viral signals were detected in these specimens by qRT-PCR as indicated by the high Ct values in the range of 27.4–38.8. The failure to detect SARS-CoV-2 viral particles by RNA-ISH might be due to the low viral count and/or RNA degradation prior to tissue fixation.

It is noteworthy that apart from pulmonary infection, there has been limited experimental evidence of SARS-CoV-2 infection in nonpulmonary tissues previously^1,31. A few research groups have attempted by means of RT-PCR, electron microscopy or IHC, to detect SARS-CoV-2 viruses in nonpulmonary tissues and have recorded the presence of viral particles in heart^27,32, kidney^16,32, liver^33,34, placenta^35–37, gastrointestinal tract³⁸, and skin³⁹. SARS-CoV-2 viruses were also detected in kidney^16,29 and placenta^36,37,40 by RNA-ISH. The current study is by far the most comprehensive report of SARS-CoV-2 infection in multiorgan systems and the first report of SARS-CoV-2 viral detection in lymph node, spleen, thyroid, colon, prostate, and uterus. The identification of SARS-CoV-2 viral infection in distant, nonpulmonary tissues, like prostate and uterus, suggest the wide spread of SARS-CoV-2 in the diverse human organ systems.

The SARS-CoV-2 host entry mediator, TMPRSS2, was first found to be highly enriched in prostate cancer⁴¹ were reported in the prostate of six male COVID-19 patients³³. Recent single-cell RNA expression studies mapped SARS-CoV-2 entry factors in a broad range of human tissues and predicted that testicular spermatogonial cells and prostatic endocrine cells are susceptible to SARS-CoV-2 infection^8,42, and we presented evidence of SARS-CoV-2 infection in prostatic tissue. Overall, our examination of SARS-Cov-2 tissue localization by in situ and qRT-PCR approaches revealed a systemic fashion of viral infection, which may explain the multisystemic involvement observed in COVID-19 disease.

The 2003 SARS coronavirus and 2019 SARS-CoV-2 virus employ host proteins ACE2 and TMPRSS2 to gain cell entry. The viral spike (S) glycoprotein recognizes and binds to ACE2 receptor on host cells and is cleaved and activated by cell surface transmembrane protease TMPRSS2^6,43. Targeting the transcriptional regulation or activity of these host factors could inhibit SARS-CoV-2 infection and replication. TMPRSS2 has been widely studied in prostate cancer where it is highly expressed in an androgen-dependent manner⁴⁴. The recurrent oncogenic TMPRSS2-ETS gene fusions are also found in more than 50% of prostate cancers⁴¹. AR inhibitors have been developed for treatment of prostate cancer and could be repurposed for COVID-19⁴⁵. In our recent study, we demonstrated that ACE2, TMPRSS2 and AR are co-expressed in a subset of lung epithelial cells under the transcriptional regulation of androgen; and transcriptional repression of AR inhibited SARS-CoV-2 infection in vitro⁸. However, the distribution of ACE2 and TMPRSS2 expression in human organ systems associated with viral infection has not been comprehensively characterized in the literature. Herein, we further characterized the tissue distribution of ACE2, TMPRSS2, and AR transcripts in multiorgan systems and found that the distribution of host entry factors, ACE2 and TMPRSS2, and regulator, AR, highly overlaps with the viral infection sites in pulmonary tissues and numerous nonpulmonary tissues, supporting the notion that targeting ACE2 and/or TMPRSS2 could be an effective treatment strategy to counter SARS-CoV-2 infection from a multiorgan perspective and inhibition of AR would be effective against SARS-CoV-2 infection⁸. The expression and tissue distribution of entry and attachment receptors can influence viral tropism and pathogenicity. Our study provides a panoramic view of such factors at a systemic level.

We also investigated viral replication activity by targeting the minus strand of SARS-CoV-2. Active viral replication events were detected in lung, bronchus, and trachea. The detection frequency of viral replication is much less than that of viral particle signals in the tissues mentioned above. Similar observation has been reported in lung tissue²⁸. Replication events and kinetics may influence viral pathogenicity, tissue tropism, and accelerated clinical organ system involvement and decline. Our study documents such events at a systemic level in COVID-19 disease. Such findings also provide valuable resources for the development of SARS-CoV-2 replication inhibitors for disease management.

Based on our observations, COVID-19 disease demonstrates heterogeneity with respect to viral load and viral cytopathic effects among various organ systems, even within the same patient. The highest viral load was observed in lung tissues and concentrated within the lung alveolar hyaline membrane and intra-alveolar spaces, followed by bronchus and trachea in the respiratory system. It is currently unclear whether heterogeneous viral abundance and distribution is associated with disease severity.

Our study cohort included a COVID-19 patient with prior kidney transplant receiving immunosuppression therapy. The SARS-CoV-2 RNA-ISH staining revealed extremely high viral infection in the lung tissue of this decedent along with SARS-CoV-2 viral infection and viral inflammatory changes in the transplant kidney, liver, and uterus. Recently studies linked high viral load with disease severity and transmissibility. A hospital-based study in China revealed that the severe COVID-19 cases tend to have higher viral load and longer virus-shedding period than mild cases⁴⁶. An observational study conducted in India showed that the index cases with high viral load detected by RT-PCR transmitted 6.25 secondary cases on average whereas the cases with low viral load transmitted an average of 0.8 case⁴⁷. Immunocompromised persons may experience persistent SARS-CoV-2 infection and bear accelerated viral evolution under immunocompromised state⁴⁸. Therefore, the high viral load observed in the COVID-19 patient under immunosuppression therapy could be related to the immune suppression status and SARS-CoV-2 infected, immunosuppressed patients may act as super-spreaders under certain circumstances.

Overall, our findings show co-existence of SARS-CoV-2 infection and host entry factors in multiple pulmonary and nonpulmonary tissues. We believe that a detailed characterization of such biomarkers in conjunction with histopathological assessment to assess disease severity and progression will guide future coronavirus biology studies on patients with advanced disease, as well as provide a framework for identifying better, novel or specific antiviral therapeutics.

Supplementary information

Acknowledgements

Author contributions

X.M.W., R. Mannan, C.F., J.L.M., L.P., A.M.C. and R. Mehra designed research; X.M.W., R. Mannan, L.X., L.M., F.S., R.W., S.Z.W., Y.Z. and X.C. performed research; J.J. and A.W. performed autopsy and generated clinical report; X.M.W., R. Mannan, L.X., Y.Q., Y.Z. and R. Mehra analyzed data; X.M.W., R. Mannan, L.X., E.A. and R. Mehra wrote the paper.

Data availability

The RT-PCR data are available in the supplementary information files. All other data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Footnotes

References