Pathological Findings in COVID-19 as a Tool to Define SARS-CoV-2 Pathogenesis. A Systematic Review

Introduction: The World Health Organization declared the COVID-19 pandemic in March 2020. COVID-19 still represents a worldwide health emergency, which causesa severe disease that has led to the death of many patients. The pathophysiological mechanism of SARS-CoV-2 determining the tissue damage is not clear and autopsycan be auseful tool to improve the knowledge of this infection and, thus, it can help achieve a timely diagnosis and develop an appropriate therapy. This is an overview of the main post-mortem findings reporting data on the infection effects on several organs. Methods: A systematic literature search was conducted in the PubMed database searching for articles from 1 January to August 31, 2020. Thearticles were selected identifying words/concepts in the titles and/or abstracts that indicated the analysis of the morphological/pathological tissue injuries related to SARS-CoV-2 disease by several investigations. Results: A total of 63 articles were selected. The main investigated tissue was the lung showing a diffuse alveolar damage (DAD) frequently associated with pulmonary thrombotic microangiopathy. Inflammatory findings and vascular damage were observed in other organs such as heart, liver, kidney, brain, spleen, skin and adrenal gland. The immunohistochemical analysis showed tissue inflammatory cells infiltrates. The virus presence was detected by several investigations such as RT-PCR, immunohistochemistry and electron microscope, showing the effect ofSARS-CoV-2not exclusively in the lung. Discussion: The evidence emerging from this review highlighted the importance of autopsy to provide a fundamental base in the process of understanding the consequences ofSARS-CoV-2 infection. COVID-19 is strictly related to a hyper inflammatory state that seems to start with DAD and immuno-thrombotic microangiopathy. Massive activation of the immune system and microvascular damage might also be responsible for indirect damage to other organs, even if the direct effect of the virus on these tissues cannot be excluded.


INTRODUCTION
In December 2019, a new Coronavirus determined a cluster of infectious diseases in Wuhan, China. In a few months, it spread across the globe, leading the WHO to declare a COVID-19 pandemic on March 11, 2020. This new pathogen is SARS-CoV-2, and the related disease in humans is defined as Coronavirus disease 2019  by WHO (Weiss and Murdoch, 2020). COVID-19 is still considered a worldwide health emergency, causing pneumonia, severe acute respiratory syndrome (SARS), multiorgan dysfunction, and death among others (Weiss and Murdoch, 2020). The  Dashboard, as of September 2, 2020, reported more than 26 million confirmed cases worldwide, of which 852,758 deaths (World Health Organization, 2020).
The primary mode of human-to-human transmission of the virus is close contact, mainly by inhalation of respiratory droplets Osborn et al., 2020). The incubation period ranges between 2 and 11 days, up to a maximum of 14 days . There are various signs and symptoms of COVID-19. Patients report fever, dry cough, and diarrhea as the most common; other symptoms are represented by myalgia, fatigue, anorexia, nausea and vomiting, confusion, headache, sore throat, rhinorrhoea. In severe cases, dyspnea, ARDS, arrhythmia and acute cardiac injury, acute kidney injury, liver damage at various degrees, and septic shock (Wang D. et al., 2020;Wang et al., 2020b;Huang et al., 2020). It was reported that the median time from the first symptom to dyspnea was 5 days, to hospital admission was 7 days, and to ARDS was 8 days (Wang et al., 2020b).
Evidence highlights that the individual response against SARS-CoV-2 differs due to genetic variations in the human population that may affect the severity of the infection and a better understanding of such variations could help identify the subjects who are most at risk (Singh et al., 2020).
The main tool to identify or confirm the COVID-19 infection is the reverse transcription-PCR (RT-PCR) on respiratory tract specimens such as nose or throat swabs (Corman et al., 2020;Udugama et al., 2020). Moreover, the serological and immunological assays are used to detect the antibodies produced by subjects following virus exposure or the antigenic proteins in infected individuals (Udugama et al., 2020). The molecular test has a shorter "window" period compared to the immunological test, with a positive outcome in the earliest infection phase: a range from 0 to 5 days has been reported to ascertain RNA-virus positivity with nose or throat swabs, or combination of both (Carter et al., 2020;Sessa F. et al., 2020).
The SARS-CoV-2 outbreak has significantly changed the clinical and methodological approaches in all healthcare branches leading to specific recommended protocols that must be implemented for the safety of both patients and practitioners Wong et al., 2020). The same concerns are equally applicable in the management of corpses of the people died "from" or "with" COVID-19 or, in general, subjects died during the outbreak. The Centers for disease Control and Prevention (CDC) have published recommendations and safety strategies to adopt during confirmed and suspected COVID-19 autopsies and about the safe management of corpses at the epicenter of the outbreak (CDC, 2020). Furthermore, several scientific societies and research groups proposed documents to help healthcare professionals and morgue staff, providing data and information about possible risks related to corpse management and practical guidance of preventive measures that must be used in suspected, probable, or confirmed COVID-19 deaths (Hanley et al., 2020a;Basso et al., 2020;Fineschi et al., 2020;Yaacoub et al., 2020). However, some countries, like Italy, have chosen not to perform the clinical autopsy, except for selected cases (Pomara et al., 2020;Salerno et al., 2020). In association with epidemiological, clinical, and laboratory data, clinical autopsies may be useful to improve the knowledge on the pathophysiology of COVID-19 infection and, thus, to reach a timely diagnosis and therapy.
This review aims to provide an overview of the main microscopic findings and biomarkers described in literature to perform the diagnosis of death due to COVID-19, highlighting the importance of autoptic examination in ascertaining the cause of death. These data are fundamental in order to distinguish between death "from" and "with" COVID-19.

MATERIALS AND METHODS
The review has been conducted by employing the PubMed database, searching for articles from 1 January to August 31, 2020, written in English. The key terms used were "COVID- 19" or "SARS-CoV-2" or "nCoV" in association with the terms "AUTOPSY" or "IMMUNOCHEMISTRY" or "IMMUNOHISTOCHEMICAL" or "IMMUNOHISTOCHEMISTRY". The main rule for article selection was the identification in the titles and/or the abstracts of words/concepts indicating the analysis of the morphological/pathological tissue injuries related to SARS-CoV-2 disease. The article selection was particularly focused on those describing autopsy findings and, thus, on the macroscopic, histological, and immunohistochemical data. The articles were read entirely if the abstract indicated that the paper's content potentially met the inclusion criteria. Articles were excluded by title, abstract, or full text if not dealing with the topic. Article reviews were also excluded. The works considered relevant were analyzed in-depth, focusing on the gross examination data, histological findings, immunohistochemical results, ultrastructural morphology and molecular diagnosis of SARS-CoV-2 infection. Full-text articles were analyzed, and data were extracted by two authors and reviewed by anothertwo authors.
complexity of the pathology. Then, the review showed that researchers also investigated the heart, liver, and kidneys; few articles reported the analysis of other organs or tissues like the adrenal gland, brain, skin, spleen. The investigated papers described evidence resulting from macroscopic and histological findings, immunohistochemical investigations, post-mortem molecular diagnosis of SARS-CoV-2 infection, and electron microscopy analysis.

Risk of Bias
To reduce the risk of bias two authors independently assessed the included studies. The risk of bias of this systematic review regards data about pre-existing or concomitant pathology resulting from microscopic findings reported by each article. Additionally, it must be noted that RT-PCR data could represent another risk of bias due to their intrinsic and extrinsic limitations Carter et al., 2020).
Finally, the lung damage was investigated by electron microscopy showing viral particles predominantly located in pneumocytes (Bradley et al., 2020;Carsana et al., 2020;Fox et al., 2020;Yan et al., 2020), free in alveolar space (Yan et al., 2020), in phagosomes of alveolar macrophages (Martines et al., 2020), confirming the above reported immunohistochemical findings; moreover, the viral particles were observed either outside or inside the cells in aggregates confined within vesicles (Bradley et al., 2020). Interestingly, Ackermann et al. (Ackermann et al., 2020) described distorted lung vascularity with structurally deformed capillaries which appeared elongated, with changes in caliber and intussusceptive pillars, and endothelium ultrastructural damage.
Moreover, Tavazzi et al. (2020) examined samples of cardiac tissue using TEM, which revealed the presence of a small group of viral particles or single particles within the damaged interstitial cells of the myocardium, also showed loss of plasmalemma integrity. Lindner et al. (Lindner et al., 2020) clearly described the presence of SARS-CoV-2 RNA in interstitial cells, and macrophage infiltrates by in situ hybridization (ISH) performed on paraffin-embedded left ventricle samples.
The myocardial data described in the selected articles are summarized in Table 2.
Other less frequent changes were disseminated intravascular coagulation (Menter et al., 2020), hemosiderin in renal tubules (Remmelink et al., 2020), chronic interstitial inflammation with sporadic prominent perivascular lymphocytic inflammation (Bradley et al., 2020), hypertensive and diabetic nephropathy (Menter et al., 2020), and unspecific nephrosclerosis . Su et al. (2020) evaluated six kidney samples using immunofluorescent stain with positive results for SARS-CoV-2 NP in 50% of the cases and, using TEM, observed coronaviruslike particles in 7 out of 9 cases together with dense deposit and subendothelial expansion. Other researchers also performed TEM, which revealed prominent activation of podocytes with multiple cytoplasmic vesicles containing virus-like particles, also detected in endothelial cells and proximal tubular epithelial cells (Bradley et al., 2020;Menter et al., 2020). The virions were also detected in proximal convoluted tubules (Rapkiewicz et al., 2020).
All the main kidney findings are reported in Table 4.

Other Organs
Data on brain involvement in COVID-19 are controversial. In particular, in a study conducted on six autopsy cases, von Weyhern et al. (2020) observed massive intracranial hemorrhage and diffuse petechial hemorrhages along with microscopic findings of localized perivascular and interstitial encephalitis, neuronal cell loss, and axon degeneration of multiple neuronal areas. Remmelink et al. (2020) described cerebral focal necrosis and cerebral hemorrhage. Similarly, another study (Reichard et al., 2020) on a single case described destructive hemorrhagic white matter lesions, focal microscopic necrosis, perivascular cellular infiltrates, and axonal injury, then confirmed by the immunohistochemical positivity to different markers such as CD68, CD3, CD20, GFAP (glial fibrillary acidic protein), APP (amyloid precursor protein) and PLP (myelin proteolipid protein). Conklin et al. (2020) reported microscopic ischemic lesions associated with widespread microvascular injuries as perivascular and parenchymal petechial hemorrhages. The ischemic damage was also found in another report by the BAPP (β amyloid precursor protein) immunohistochemical stain and T-cell infiltration around blood vessels and capillaries (Hanley et al., 2020b). Similar findings were described by Kantonen et al. (2020), reporting enlarged perivascular spaces, microhemorrhages, scattered T-lymphocytes, and minor intravascular fibrinoid deposits in some cerebral and subarachnoidal vessels. Moreover, in the nine autopsy cases examined by Nunes Duarte-Neto et al. (2020), reactive gliosis, neuronal satellitosis, small vessel disease, and perivascular hemorrhages were reported. On the contrary, Solomon et al. (2020) described only acute hypoxic-ischemic damage in the absence of microscopic specific elements; however, the same research group highlighted negative SARS-Cov-2 immunohistochemistry and positive molecular diagnosis by RT-PCR in few samples of the medulla, olfactory nerves, and frontal lobe. Some reports described macroscopic and histological evidence of the spleen. Oprinca and Muja (2020) reported histologically marked congestion and white pulp atrophy associated with the absence of lymphoid follicles. Prilutskiy et al. (2020) observed an enlarged, soft, and friable organ just in one of the four analyzed cases. Microscopically, they described white pulp depletion with red pulp hemorrhage or infarction and histiocytic hyperplasia with hemosiderin-laden macrophages, suggestive of a prior red blood cells phagocytosis, or hyperplastic white pulp with red pulp congestion but lacking hemophagocytosis. White pulp depletion and red pulp hemorrhage were reported also by Rapkiewicz et al. (2020). Even Nunes Duarte-Neto et al. (2020) studied the spleen in five cases reporting lymphoid hypoplasia, red pulp hemorrhages, and splenitis.
Additionally, they described follicular arterioles endothelial changes, vasculitis, and arterial thrombus. Likewise, an acute splenitis was observed by Menter et al. (2020) in six out of 21 cases, while Lax et al. (2020) found lymphocyte depletion affecting both the spleen and lymph nodes. Furthermore, in two cases necrotizing granulomata was reported (Sekulic et al., 2020) in the spleen.
Lymph nodes and bone marrow histological changes have been observed in two of the above-mentioned studies. Indeed, one of these (Prilutskiy et al., 2020) described enlarged mediastinal and pulmonary lymph nodes that showed a hemophagocytic histiocytes CD163+, while the other (Menter et al., 2020) reported lymph nodes congestion and increased presence of plasmablasts. As for the bone marrow, both research groups reported leftshifted myeloid hyperplasia; in addition, Prilutskiy et al., 2020) also observed histiocytic cells CD163+.  reported data on lymphoid tissue describing primary lymphoid follicle, scattered T lymphocytes, and focal necrosis. The virions Frontiers in Pharmacology | www.frontiersin.org April 2021 | Volume 12 | Article 614586 detection in bone marrow was reported only in one study using TEM, which detected megakaryocytes (Rapkiewicz et al., 2020).
Adrenal gland findings were reported by Iuga et al. (2020), who described small vessels with acute fibrinoid necrosis, subendothelial vacuolization, and apoptotic debris. In a further study , adrenal cortical hyperplasia was described. Interestingly, Hanley et al. (2020b) described patchy areas of infarct-type adrenocortical necrosis and organizing microthrombi in adrenal vessels.
Furthermore, Yang et al. (2020) studied the tests in 12 cases using post-mortem biopsy, detecting Sertoli cells swelling, reduced Leydig cells, mild lymphocytic inflammation, detachment from tubular basement membranes and lumen intratubular cell mass loss and sloughing; in the same study, the immunohistochemical positivity to different markers such as CD3, CD20, CD68, CD138, and ACE-2 was observed, but the RT-PCR confirmed the presence of the virus only in a biopsy sample. Nevertheless, Nunes Duarte-Neto et al. (2020), in two cases out of two, observed an orchitis condition.
Finally, the involvement of the skin was included in the study performed by Magro et al. (2020). Five cases with purpuric lesions were described, microscopically characterized by thrombogenic vasculopathy, epidermis, and adnexal structures necrosis, interstitial and perivascular neutrophilia with prominent leukocytoclasia or perivascular lymphocytic infiltrate in the superficial dermis with small thrombi within rare venules of the deep dermis. The same study highlighted the immunohistochemical positivity to different markers like C4d, C3d, C5b-9, MASP2, and SARS-CoV-2 spike and envelope proteins.
Purpuric lesions, superficial perivascular mononuclear infiltrate, and endothelial changes were also described by another work in which interesting findings in the skeletal muscle were evaluated, consisting of myositis and necrotic fibers (Nunes .
The findings about the mentioned organs are summarized in Table 5.

DISCUSSION
In this work, we reported the main findings related to COVID-19 tissues. The authors reviewed studies describing the macroscopic and microscopic (histological and immunohistochemical) data and the ultrastructural and molecular evidence observed in different organs by both biopsies and autopsies. Only the histological samples and, in general, the autopsy findings enable us to ascertain the exact cause of death, especially in suspect or probable cases ofSARS-CoV-2 infection; moreover, the post-mortem evidence provides fundamental data which are useful to better understand the pathophysiology of this infection; this can help clinicians identify the most appropriate and effective treatment to reduce mortality.
Although documented by clinical and imaging data of patients affected by this infection (Lake, 2020), the revision revealed that the lung is the most affected organ, showing injuries similar to those observed in diseases related to other coronaviruses, the Severe Acute Respiratory Syndrome (SARS) and the Middle East Respiratory Syndrome (MERS) (Ding et al., 2003;Ng et al., 2016), even if the alterations related to COVID-19 are more extensive and severe, and collectively constitute a distinctive feature (Buja et al., 2020).
In critical cases, the acute COVID-19 respiratory picture revealed a severe acute respiratory distress syndrome (ARDS) associated with the histological detection of DAD in different stages, mainly in exudative and proliferative phases. In particular, this condition was characterized by hyaline membranes, intraalveolar and/or septal and/or interstitial edema, intra-alveolar fibrinous exudate and inflammatory cells infiltrates, type 2 pneumocyte hyperplasia/activation, and squamous metaplasia (Okudela et al., 2020;Wichmann et al., 2020;Heinrich et al., 2020;Bradley et al., 2020;Aguiar et al., 2020;Sekulic et al., 2020;Cipolloni et al., 2020;Edler et al., 2020;Suess and Hausmann, 2020 Immunohistochemical analysis was fundamental to better define the immune cells infiltrating the lung tissue Aguiar et al., 2020;Cipolloni et al., 2020;Suess and Hausmann, 2020;Navarro Conde et al., 2020;Carsana et al., 2020;Menter et al., 2020;Fox et al., 2020;Buja et al., 2020;Oprinca and Muja, 2020;Nunes Duarte-Neto et al., 2020;Grimes et al., 2020;Hanley et al., 2020b;Aiolfi et al., 2020;von Weyhern et al., 2020;Ackermann et al., 2020;Barton et al., 2020;Rapkiewicz et al., 2020;Varga et al., 2020), it showed all aspects of innate immune response and T and B cells immunity once the virus has entered the tissue cells. In fact, it was reported that, in COVID-19, white cells immunity acts as in a classical respiratory virus-like infection (Jansen et al., 2019;Azkur et al., 2020): recognition of the whole virus and viral particles by professional antigen-presenting cells, which are mainly dendritic cells and macrophages, that present viral peptides to CD4 + T cells; activation of CD8 + cytotoxic T cells to lyse the virus-infected cells; activation of B cells that can directly recognize the viruses and also interact with CD4 + T cells. Additionally, other two investigated markers were ACE-2 and TTF-1 on which some interesting consideration can be carried out. ACE-2 was recognized as a receptor for host cell entry of Frontiers in Pharmacology | www.frontiersin.org April 2021 | Volume 12 | Article 614586 SARS-CoV-2, being positive in alveolar epithelial, endothelial cells, alveolar macrophages, and lymphocytes in lung tissue samples (Ackermann et al., 2020;Cipolloni et al., 2020;Grimes et al., 2020). It was also reported that both the tissue activity and membrane expression of ACE-2 increase the susceptibility to COVID-19. On the contrary, several studies indicated that downregulation of ACE-2 is associated with the aggravation of inflammatory events due to both antiinflammatory and anti-fibrotic actions of ACE-2 activators and the modifications in renin-angiotensin-aldosterone system function (Behl et al., 2020). TTF-1 (thyroid transcription factor-1) was used to evaluate the pneumocytes, resulting positive in those hyperplastic and dysmorphic (Aiolfi et al., 2020;Buja et al., 2020;Carsana et al., 2020;Fox et al., 2020;Menter et al., 2020;Navarro Conde et al., 2020;Nunes Duarte-Neto et al., 2020;Schaefer et al., 2020;Shao et al., 2020;Suess and Hausmann, 2020). Notably, this marker was detected in dysmorphic cells showing features of syncytia, characterized by the presence of multi-nucleation and an ample cytoplasm surrounded by a single plasmalemma; these syncytial elements of pneumocytic origin are considered characteristic of COVID-19 (Bussani et al., 2020). The pulmonary thrombotic microangiopathy is another observed microscopic pattern, together with DAD, in several cases resulting often in fibrin-platelet thrombi in alveolar capillaries and/or in small vessels (Konopka et al., 2020a;Barton et al., 2020;Konopka et al., 2020b;Bradley et al., 2020;Buja et al., 2020;Cai et al., 2020;Carsana et al., 2020;Dolhnikoff et al., 2020;Fox et al., 2020;Grimes et al., 2020;Kantonen et al., 2020;Menter et al., 2020;Navarro Conde et al., 2020;Nunes Duarte-Neto et al., 2020;Oprinca and Muja, 2020;Rapkiewicz et al., 2020;Remmelink et al., 2020;Santana et al., 2020;Schaefer et al., 2020;Shao et al., 2020;Suess and Hausmann, 2020;Wichmann et al., 2020) associated with endothelial damage showing different features (i.e. thrombotic necrotizing capillary injury, perivascular inflammation/endothelialitis) (Ackermann et al., 2020;Cai et al., 2020;Dolhnikoff et al., 2020;Edler et al., 2020;Wang et al., 2020a;Magro et al., 2020;Menter et al., 2020;Remmelink et al., 2020;Varga et al., 2020). In particular, Achermann et al. (Ackermann et al., 2020) reported distinctive angiocentric features in some cases of COVID-19 represented by severe endothelial injury associated with intracellular SARS-CoV-2 virus and endothelial cell membranes destruction, widespread vascular thrombosis with microangiopathy and alveolar capillaries occlusion, a significant new vessel growth by intussusceptive angiogenesis. The latter morphological/pathological aspect suggests a severe degree of endothelialitis and thrombosis in the lung. Very interesting evidence was described by Magro et al. (2020) because the authors analyzed pulmonary abnormalities largely restricted to the alveolar capillary with immunohistochemical detection of complement components and MAC in the lung septal microvasculature, this suggested an "atypical ARDS" in which a catastrophic microvascular injury syndrome occurs activating the complement pathway. Moreover, this would be supported by C5b-9 involvement in pathologies as antiphospholipid antibody syndrome and purpura fulminans (Perera and Murphy-Lavoie, 2020;Ruffatti et al., 2020). MAC contributes to endothelial activation, resulting in cell adhesion molecules and releasing chemotactic and activating factors related to migration and activation of white cells (Kilgore et al., 1995;Mondello et al., 2017;Mondello et al., 2018;Mondello et al., 2020). The role of the complement system in lung injuries due to SARS-CoV-2 was also supported by the immunohistochemical detection of MASP-2 (mannan-binding lectin-associated serine protease-2) (Magro et al., 2020), an effector enzyme of the lectin pathways capable of binding the SARS-CoV-2 nucleocapsid protein, resulting in complement activation and endothelial/tissue injury (Rambaldi et al., 2020).
Current collected evidence integrates and completes the clinical studies on understanding the pathological mechanism underlying the SARS-CoV-2 infection in which both the innate and the acquired immune system (i.e. tissue barriers, innate and adaptive cells, and mediators) are involved (Azkur et al., 2020). Currently, it is postulated that the virus determines a state of hyper-inflammation in pulmonary tissue variously designated as macrophage activation syndrome (MAS), cytokine storm, and secondary hemophagocytic lymphohistiocytosis (sHLH) (Barth et al., 2020a;McGonagle et al., 2020a;Ruan et al., 2020). In particular, the cytokine storm induced by the activation of a large number of white cells (i.e. B cells, T cells, NK cells, macrophages) and resident tissue cells (i.e. epithelial and endothelial cells) is a common feature of severe COVID-19 cases, as it is responsible for ARDS and multiorgan failure (Azkur et al., 2020).
Notably, the post-mortem evidence regarding both the lung tissue damage supporting ARDS and the pulmonary vascular thrombotic findings might be consistent with the conclusion of a recent clinical prospective observational study. Such study reports that patients with a combination of very high D-dimer concentration and low lung static compliance due to COVID-19 have a dramatic increase in mortality rate (Grasselli et al., 2020). Thus, a critical prognosis could be considered for patients with pulmonary cells and vascular system affected by SARS-CoV-2.
Moreover, the data emerging from the literature review lead to consideration of the SARS-CoV-2 infection no longer as a disease almost exclusively affecting the lungs, but as a disease that affects multiple organ systems.
The heart findings in subjects who have died from COVID-19 are acute inflammatory signs such as focal lymphocytic infiltrates, individual cardiomyocyte injury, acute myocardial infarction, Frontiers in Pharmacology | www.frontiersin.org April 2021 | Volume 12 | Article 614586 lymphocytic myocarditis/epicarditis/pericarditis (Hanley et al., 2020b;Buja et al., 2020;Lax et al., 2020;Menter et al., 2020;Rapkiewicz et al., 2020;Remmelink et al., 2020;Schaller et al., 2020). These tissue alterations could cause both cardiomyopathy and heart failure characterizing the potential natural evolution of COVID-19 as suggested by some authors . Previous studies revealed that the dysregulated inflammatory response related to the failure of infection control process can affect the heart determining septic cardiomyopathy characterized by myocardial depression . Coronary artery atheroma and/or atherosclerosis were other pathological findings that might cause an increased susceptibility to the negative consequences of the SARS-Cov-2 infection, including death . Additionally, the endothelial small vessels damage and thrombosis were also observed in this organ, even if these were found in just a few cases (Hanley et al., 2020b;Nunes Duarte-Neto et al., 2020;Oprinca and Muja, 2020;Rapkiewicz et al., 2020). However, more interesting data emerged from myocardial ultrastructural analysis, revealing the presence of a small group of viral particles or single particles within the damaged interstitial cells of the myocardium, demonstrating loss of plasmalemma integrity (Tavazzi et al., 2020). ISH also showed the presence of the virus in interstitial cells and macrophage infiltrates (Lindner et al., 2020). This evidence supports the thesis that the direct viral infection of the heart could be another possible cause of myocardial damage (Bonow et al., 2020). The microscopic liver analysis revealed in some cases lymphocytic infiltrates, especially in the portal/periportal tract and focal lobular inflammation with infiltration of predominant lymphocytes and neutrophils along with Kupffer cell hyperplasia and hypertrophy (Wang et al., 2020a;Lax et al., 2020;Nunes Duarte-Neto et al., 2020;Oprinca and Muja, 2020;Prilutskiy et al., 2020;Schaller et al., 2020;Varga et al., 2020). The described hepatocyte necrosis can be more likely attributed to hypoxia and systemic inflammation in severe cases (Horvatits et al., 2019). Microthrombi were observed by Rapkiewicz et al. (2020) in hepatic sinusoids. Interesting vascular findings were also reported by Sonzogni et al. (2020), such as portal vein endotheliitis and diffuse alterations of intrahepatic vascular structures together with luminal thrombosis. Also, besides the heart, the TEM analysis revealed the presence of typical coronavirus particles in the cytoplasm of hepatocytes with most viral particles without membrane-bound vesicles (Wang et al., 2020a).
Not many research groups studied the brain in COVID-19 deceased patients. In some cases, researchers found that the encephalic small vessels were affected, and they described inflammatory lymphocytic infiltrates and hemorrhagic foci in the perivascular setting (Nunes Hanley et al., 2020b;von Weyhern et al., 2020;Reichard et al., 2020;Conklin et al., 2020). Although they had some histological aspects in common, none of the authors found specific elements to establish a direct correlation with the viral infection. Nevertheless, von Weyhern et al. (2020) documented lymphocytic pan-encephalitis and meningitis with diffuse petechial hemorrhage in the entire brain even if they did not observe significant endotheliitis on the TEM and reported that the hypoxic morphological changes were localized in brain areas most commonly susceptible to developing damage (i.e., neocortex, cerebellum, and brainstem). Similarly, the hypoxic changes observed by Solomon et al. (2020) have no specific characteristics correlated to the SARS-Cov-2 infection. Finally, Reichard et al. (2020) excluded an association between the larger hemorrhagic lesions and both the perivascular inflammation and small vessels necrosis, suggesting that these changes occurred in close chronological proximity to the patient's death, hypothesizing micro-thromboembolic pathogenesis, possibly related to COVID-19, based on the observation of microscopic neocortical infarcts.
The digestive tract involvement, except for the liver, was little discussed, resulting in only mesenteric ischemia, small bowel sub-mucosal vessels endotheliitis, and pancreatitis (Hanley et al., 2020b;Lax et al., 2020;Remmelink et al., 2020;Varga et al., 2020). A research group made a very interesting consideration, evaluating only bowel ischemic damage and lung injuries and pointing out that the intestine lesions frequently do not correlate with tissue virus detection: RT-PCR positivity in cases without ischemia, and RT-PCR negativity in cases with ischemia . However, clinical reports described the recurrence of intestinal ischemia in COVID-19 patients with gastrointestinal symptoms noting an increased mortality rate with respect to patients with SARS-CoV-2-negativity (Norsa et al., 2020).
Two different studies (Nunes Duarte-Neto et al., 2020;Yang et al., 2020) described macroscopic and histological testis findings with results compatible with viral orchitis, characterized by lymphocyte inflammation with a prevalence of T-cells. Moreover, Yang et al. (2020) also observed the swelling of Sertoli cells and reduced Leydig cells, two cell populations with a strong expression of ACE-2 receptors.
As for the skin involvement, purpuric lesions were observed in association with other microscopic findings such as a microvascular endothelial and thrombotic damage and perivascular inflammatory infiltrate (Magro et al., 2020;Nunes Duarte-Neto et al., 2020). It must also be highlighted that the complement components analyzed by Magro et al. (2020) in lung tissue were detected in skin samples as extensive endothelial and subendothelial deposits of C5b-9.
The molecular analysis performed by several researchers using RT-PCR on post-mortem samples showed valuable data leading to different considerations. This investigation is fundamental to assess the SARS-CoV-2 infection and, as suggested by other authors , it should always be performed, especially in probable and suspected cases. The present literature review confirmed that the virus is located mainly in lungs and upper airway tissue (Aguiar et al., 2020;Konopka et al., 2020a;Bradley et al., 2020; Tian et al., 2020b;Buja et al., 2020;Cipolloni et al., 2020;Craver et al., 2020;Edler et al., 2020;Fitzek et al., 2020;Fox et al., 2020;Kantonen et al., 2020;Karami et al., 2020;Lacy et al., 2020;Martines et al., 2020;Menter et al., 2020;Nunes Duarte-Neto et al., 2020;Oprinca and Muja, 2020;Schaller et al., 2020;Sekulic et al., 2020;Wichmann et al., 2020;Youd and Moore, 2020). The negativity reported in some cases could be explained by: 1) the early "window" period in which RT-PCR cannot detect the RNA; 2) the different clinical and biological evolution of the infection in the late phase resulting in subjects both negative or continuously positive; 3) the immune interindividual variability and the different virus haplotype Carter et al., 2020;Kucirka et al., 2020;Lescure et al., 2020;Norsa et al., 2020;Shen et al., 2020;Yi, 2020). Other factors affecting the RT-PCR results can be the sampling technique and post-mortem tissue degradation (Ventura Spagnolo et al., 2019b;Stassi et al., 2020;Youd and Moore, 2020). Moreover, this investigation was performed on other organs, resulting in positivity in the heart, liver, spleen, kidney, and testis of few cases (Wichmann et al., 2020;Bradley et al., 2020;Sekulic et al., 2020;Menter et al., 2020;Hanley et al., 2020b;Tian et al., 2020b;Lindner et al., 2020;Iuga et al., 2020], supporting that the SARS-CoV-2 might determine a direct effect not only on respiratory tissue. Moreover, Remmelink et al. (2020) reported virus detection in the heart, spleen, liver, kidney, bowel, and brain even if the histological findings were non-specific. Similar considerations regarded the study of Skok et al. (2020), which described bowel histological damage in both cases with and without virus RNA tissue detection. The summarized evidence collected from the articlesincluded in this review show that COVID-19 is strictly related to a hyperinflammatory state that seems to start with DAD and immuno-thrombotic microangiopathy. Massive activation of the immune system and microvascular damage might also be responsible for indirect damage caused to other organs, even if the direct effect of the virus on these tissues cannot be excluded.
Research is still at an early stage in understandinghow SARS-CoV-2 both affects the immune system and determines organ damage, especially lung, but a better understandingisextremely importantin order to develop better(s)elective treatment strategies . To date, in fact, there is yet no specific and standardized treatment even if in vitro studies and/or clinical trial of some antiviral drugs (such as inhibitors of RNA-dependent RNA polymerase, viral entry inhibitors, and viral fusion inhibitors) showed encouraging results (Kabir et al., 2020). Moreover, in this context, studies aimed at evaluating the genetic diversity of SARS-CoV-2 appear to be crucial for the potential association with virus conformational structural changes and, thus, an altered antigenicity supporting different immune responses (Kabir et al., 2020).
Frontiers in Pharmacology | www.frontiersin.org April 2021 | Volume 12 | Article 614586 various reasons. This way, it is possible to assess if the subject died "from" or "with" COVID-19, ensuring reliable epidemiological, pathological, molecular, and global health data. Our study has limitations, mainly related to the unknown validity of pathological assessment of macroscopic and microscopic findings; actually, there is no published risk of bias assessment tool that could be applied to the material and methods analyzed in this review. Another limitation was missing details regarding the prevalence of the detected features stratified through the characteristics of the subjects, such as age, gender, race, and comorbidities and their pharmacological treatment. Additionally, only data available on PubMed database were taken into account. Finally, limitations secondary to paucity of performed autopsies and data available in literature did not allow a reliable estimation to be conducted of the overall quality of the included studies through the statistical evaluation of the pathological findings rate.
All post-mortem findings, along with clinical pictures, are useful to better understand pathogenesis and pathophysiology, and, consequently, the best therapeutic management of patients. Therefore, this is the right time "to reverse the decline of autopsy rates", as affirmed by De Cock et al. (2019) a few months before the COVID-19 outbreak, in order to benefit global public health. As observed by Barth et al. (2020b), a "call to action" for more autopsy reports containing detailed findings is necessary.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
Conceptualization was carried out by AA and CP. Literature search was performed by SR, OM, NDN, MS, CM, DS, PG, and EVS with guidance from AA. Draft writing was provided by CM, SR, OM, prepared the figure and tables. Final manuscript editing was performed by CM, SR, OM, and AA. All authors contributed to the article and approved the submitted version.