SARS-CoV-2 cytopathogenesis in cultured cells and in COVID-19 autoptic lung, evidences of lipid involvement.

Background: The pathogenesis of SARS-CoV-2 remains to be dened. Elucidating SARS-CoV-2 cellular localization within cells and its cytopathic effects requires denition. We performed a comparative ultrastructural study of SARS-CoV-2 infection of Vero-6 cells and lung from COVID-19 patients. Main ndings: SARS-CoV-2 induces rapid ultrastructural changes and death in Vero cells. Ultrastructural changes in SARS-CoV-2 infection differ from those in SARS-CoV-1. Type II pneumocytes in lung tissue showed prominent altered morphological features with numerous vacuoles and swollen mitochondria with presence of abundant lipid droplets. The accumulation of lipid droplets was the most striking nding we observed in cultured cells and in infected pneumocytes. Virus particles were also found associated with lipo-lysosomes suggesting that they can play an important step in virus assembly. Interpretation: The cytopathology of SARS-CoV-2 appears to be different to that caused by SARS-CoV-1. Our ndings highlight important open topics which may represent future targets to contrast the pathogenicity of SARS-CoV-2.


Introduction
Since the rst discovery of SARS-CoV-2 as a novel human zoonotic pathogen in late December, 2019 Various mechanisms for tissue pathology in COVID-19 have been proposed, including direct cytopathic effects, ischemic injury and immune-pathology due to excessive and aberrant immune responses. SARS-CoV-2 infects the host cells using the angiotensin converting enzyme 2 (ACE2) receptor [ Ref 9], which is expressed in cells and vessels of several organs, including the lung, heart, kidney, intestine. Coronaviruses have a lipid bilayer derived from the host cell membranes and thus the intracellular membrane may play a key role in replication and cytopathic effects.
Evaluation and determination of SARS-CoV-2 virus distribution within tissues, cellular localization and its cytopathic effects is important for elucidating the pathogenetic mechanisms of SARS-CoV-2. Ultrastructural studies carried out so far, didn't show clear but rather con icting ndings concerning the presence of viral particles inside different tissues [ .
In vitro cytopathic studies of SARS-CoV-2 using cell lines may not capture in vivo pathology in affected body organs and thus performing studies in parallel is important. We thus performed a comparative ultrastructural study of SARS-CoV-2 infection of Vero-6 cells and lung tissue cells from 20 patients who died of COVID-19 disease. We also investigated the effects of SARS-CoV-2 on Vero cells, compared to effects of SARS-CoV-1.
Methods SARS-CoV-2 and SARS-CoV-1 isolates SARS-CoV-2: The rst COVID-19 cases were identi ed on January 31st at our National Institute for Infectious Diseases IRCCS "Lazzaro Spallanzani", Rome, Italy. SAR-CoV-2 was isolated and cultured from these patients [13] and was used in this study.

VERO cell lines and infection with SARS-CoV-2 and SARS-CoV-1
Mammalian cell lines Vero E6 were cultured in Dulbecco's essential medium (DMEM, Sigma-Aldrich) containing 10% fetal bovine serum (FBS), at 37°C in a 5% CO2 atmosphere. Sub-con uent cells were infected with the SARS-CoV-2 INMI1 isolate (named 2019-nCoV/Italy-INMI1, GISAID accession number: EPI_ISL_410546) obtained from sputum sample (mucus and phlegm coughed-up from the lower airways) collected at admission from the rst Covid-19 patient reported in Italy in January 2019 and hospitalized at INMI.
Speci cally, the cells were exposed to the SARS-CoV-2 (2019-nCoV/Italy-INMI1) isolate in DMEM not containing FBS for 1 hour at 37°C; at the end of the adsorption period, cells were washed, and maintained in DMEM plus 2% FBS. The same method was used for infection of Vero cells with SARS-CoV-1. Uninfected Vero cells were used as controls.

Negative Staining
Puri ed SARS-CoV-2 and SARS-CoV-2 viral suspensions were xed in 2.5% glutaraldehyde and allowed to adsorb onto a formvar carbon-coated grid for a few minutes before being stained with 1% phosphotungstic acid for 1 min. The excess uid was blotted and the grid left to dry before viewing under a transmission electron microscope JEOL JEM 2100 Plus (Japan Electron Optics Laboratory Co. Ltd. Tokyo). Images were captured digitally with a digital camera TVIPS (Tietz Video and Image Processing Systems GmbH. Gauting, Germany).

Autopsy Lung tissues
Post-mortem examination of 20 consecutive patients, who died of laboratory con rmed COVID-19 disease, were performed at the National Institute for Infectious Diseases Lazzaro Spallanzani-IRCCS Hospital (Rome, Italy). Autopsies were performed according to guidance for post-mortem collection and submission of specimens and biosafety practices (CDC March 2020, Interim Guidance; Hanley B, Lucas SB, Youd E, Swift B, Osborn M. Autopsy in suspected COVID-19 cases. J Clin Pathol. 2020;73(5):239-242. doi:10.1136/jclinpath-2020-206522) to reduce the risk of transmission of infectious pathogens during and after the post-mortem examination. The study was approved by the local Clinical Research Ethics Committee (approval number: n° 9/2020).
Specimens from lungs were xed in 10% neutral-buffered formalin, and routinely processed to para n blocks. Sections of tissues (4 μm) were stained with hematoxylin and eosin (H&E). For immunohistochemistry depara nized and rehydrated sections were used. Immunostaining was performed on BenchMark ULTRA system fully automated instrument (Roche) with antibody directed against CD68 Ventana (KP-1), and anti-Coronavirus (FIPV3, Santa Cruz).

Light and Transmission electron microscopy
Light and transmission electron microscopy (TEM) was performed on Vero cells and autopsy lung tissue specimens using standard procedures. Cultured cells and small pieces of tissues were xed with 2.5% glutaraldehyde in 0.1M cacodylate buffer, for 4h at 4 °C. Post-xation was performed with 1% OsO 4 .
Samples were then dehydrated in graded ethanol and embedded in Epon resin, as previously described [14,15]. Ultrathin sections were stained with 2% uranyl acetate and observed under a transmission electron microscope JEOL JEM 2100 Plus (Japan Electron Optics Laboratory Co. Ltd. Tokyo, Japan).
Images were captured digitally with a digital camera TVIPS (Tietz Video and Image Processing Systems GmbH. Gauting, Germany).

Negative Staining
Negative staining electron microscopy of puri ed SARS-CoV-2 particles revealed a spherical (Fig. 1A) or slightly pleomorphic shapes (Fig. S1A,C,D).. On the surface of the virions the typical rim of cone-shaped spikes were identi ed, but their distribution was not as regular as usually reported for other Coronavirus, in fact they appeared in multi-aggregated fashion (Fig. 1A, Fig. S1).. The diameter of the viruses ranged from 80 to 102 nm (average size 93.61 nm), while the length of the spikes ranged from 9 to 12.5 nm (average length 10.99 nm). Some of the viral particles showed part of the ribonucleic-protein material extruding from rupture of the envelope (Fig. S1C)..

EM ndings -SARS-CoV-2 infected Vero cells.
Light microscopy of thin-sections from resin embedded cell samples, showed cytopathic effects of the virus. At 24 hours post-infection, many cells lose their typical elongated shape of uninfected cells (Fig.  S2A) and become roundish and rich in plasma membrane extro essions (Fig. S2B).. After 48 hours from the infection cell morphology further changed dramatically. Most cells appeared swollen and numerous cytoplasmatic vacuoles were visible inside them; in contrast, other cells appeared dark colored suggesting that cell shrinkage occurred (Fig. S2C,D).. Transmission electron microscopy analysis at 24 hours post-infection showed several round shaped cells, with prominent presence of lopodia at the plasma membrane (Fig. 1B).. Many mature viral particles were visible budding at the cell surface ( Fig.  1B,C).. Inside the cells, SARS-CoV-2 particles were detected in virus containing compartments (VCC) that were with different size and shape (Fig. 1E,F).. Group of virions were enclosed in single membrane vacuoles, similar to endosomes (Fig. 1E).. Other vesicles, small in size, containing single viral particles, resembled the "spherules" described for other coronaviruses (Fig. 1F) [ Refs 16,17]. At 48h post-infection most cells showed strong signs of degeneration and many were clearly dying (Fig. 2).. Some cells showed extensive vacuolization of the cytoplasm and depletion of all organelles ( Fig. 2A).. Vacuoles containing viruses were still present in necrotic cells (Fig. 2B).. Some cells seemed to die with morphological features of both apoptosis and necrosis, in which condensed cellular contents was dispersed by means of plasma membrane leakage. Free released viruses were also observed associated with cell remnants (Fig. 2C,D)..
The most striking nding we observed in infected cells was the presence of numerous lipid droplets (LDs), with variable size and morphology (Fig 2E,F).. Some LDs showed homogeneous content, the typical feature of lipid storage without encompassing membrane (Fig. 2E).. Other droplets presented an external dark membrane (Fig. 2E,F) and were identi ed as lipolysomes described in humans with abnormalities in lipid metabolism [17]. Mitochondria in contact with the lipid droplets were often found ( Fig. 2E).. In most cells mitochondria appeared altered and display swollen cristae (Fig. 2F).. Of note, virus particles were also found associated with lipolysosomes suggesting that they can play an important step in virus assembly (Fig. 2F)..

EM ndings -SARS-CoV-1 infected Vero cells.
During the rst 24 hours post SARS-CoV-1 infection, Vero cells also showed modi cation of plasma membrane, which became enriched in lopodia and extro essions associated to the presence of numerous virus particles (Fig. 3A).. At 48h post-infection appearance of vacuoles and roundish of cells was displayed. Some infected cells showed the formation of large septa, resulting in a more dramatic compartmentalization of cytoplasm compared to SARS-CoV-2 (Fig. 3B).. Cytopathic effects in SARS-CoV-1 infected cells resulted in both apoptotic (Fig. 3B) and necrotic cell death (Fig. 3C).. Large vacuoles containing virus particles, resembling dilated spaces of endoplasmic reticulum were also detected (Fig.  3D, S3A).. Mitochondria displayed loss of their typical morphology, they appeared swollen with progressive cristae disappearance, resulting in the formation of vesicles, which occasionally still maintain mitochondrial matrix into the lumen. Virus particles were observed inside these vesicles, lining the membrane or in the process of pinching off (Fig. 3E).. Virus particles were also found in deep association with particular multilamellar structures (Fig. 3F).. These structures were thought to derive from the rER, since ribosome-carrying membrane were observed in close continuity (Fig. S3B).. We didn't observe the presence of lipid droplets at any time of infection.
A comparative of the main features of the SARS-CoV-2 vs the SARS-Cov ultrastructural feature is shown in (Fig 6).
Histopathological examination of Lung tissue from COVID-19 patients.
In order to evaluate whether SARS-CoV-2 exerts similar cytopathic effects in vivo, we analyzed lung tissues obtained at autopsy from 20 COVID-19 patients. Histopathological analysis showed diffuse alveolar damage with hyaline membranes, brinous exudate, and in ammatory in ltrate (Fig. 4A).. Damage of alveolar epithelium was associated with the presence of reactive type II pneumocyte, characterized by hyperplasia, amphophilic cytoplasm, large nuclei and prominent nucleoli (Fig. 4B).. Type II pneumocytes showed increased detachment from the alveolar walls and displayed signs of degeneration consisting in highly vacuolated cytoplasm and nuclear changes, making the nucleus di cult to distinguish (Fig. 4C,D).. The immunohistochemistry anti-coronavirus revealed a focal distribution of the positivity, restricted to the activated type II pneumocytes (Fig. 4E,F).. It is important to note that absence of immunoreactivity to CD68 of these altered cells excluded that they were macrophages (Fig. 4G,H)..

Electron Microscopic examination of Lung tissue from COVID-19 patients.
Lung tissues showed the presence of SARS-CoV-2 virus inside type II pneumocytes (Fig. 5A).. As found in cultured infected cells, virions were observed enclosed in single layered cytoplasmic compartments of variable size, containing numerous viral particles, or as sole particles into the "Spherules" (Fig. 5A,B).. The pneumocytes showed altered morphological features, for example the nucleus appeared with nely and uniformly dispersed chromatin (Fig. 5A) or with convoluted pro le and marginated chromatin alternated to cleared regions (Fig. 5E).. Those cells displayed organelles injury comparable with those observed in SARS-CoV-2 infected cells. The pneumocytes showed the presence of numerous vacuoles and swollen mitochondria (Fig. 5A-D).. The rough endoplasmic reticulum and free ribosomes, (which are typically abundant in type II pneumocytes, due to the production of surfactant), in the infected cells were respectively enlarged and compartmentalized (Fig. 5C,D,F).. Of note, in agreement to what observed in cultured cells, the infected type II pneumocytes showed unusual presence of abundant lipid droplets (LDs) (Fig. 5E,F).. All SARS-CoV-2 infected cells ultimately died. Dying pneumocytes had morphological features which did not resemble neither necrosis or classical apoptosis. Some cells displayed partially condensation together with plasma membrane leakage and release of the cellular contents (Fig. 5E).. These membranous structures not only harbor viral proteins but also contain a speci c array of hijacked host factors, which collectively orchestrate a unique lipid micro-environment optimal for coronavirus replication [ . Ultrastructural studies on Coronavirus genera have revealed that alpha-and betacoronaviruses formed clusters of the double-membrane vesicle (DMV), sometimes linked by a convoluted membrane [ Ref 25], whereas the gamma-coronavirus IBV induced extensive paired membranes and smaller 60-80 nm spherules in addition to the DMVs [ . In a recent study, it has been demonstrated that during human coronavirus infection the cell lipid pro le is signi cantly altered [ Ref 28].
An observation in our study regarding SARS-CoV-2 is that there is the striking difference compared to SARS-CoV-1, consisting in the presence of numerous lipid droplets. In SARS-CoV and MERS-CoV, host lipids have been reported to be linked most to pathogenesis. In particular, the lipid rafts are required for cellular entry [ Refs 29,30]. Some (+)RNA viruses exploit lipid droplets (LDs) to acquire lipids for membrane or energy production to support their replication [ Ref 31]. Lipid droplets found in SARS-Cov-2 infection, both in vitro and in type II pneumocytes, appear similar to those known to occur in hepatocytes as a consequence of HCV infection. Lipid droplets with typical features of lipid storage, without encompassing membrane and translucent omogeneous appearance, were often observed. Other vesicles with the characteristics similar to lipolysosomes were also present, with external membrane and whorls. In Vero cells viral particles were also found associated with lipolysosomes suggesting that they can play a role in virus assembly. Another important observation concerns mitochondria. A number of mitochondria were in close contact with lipids droplets. These contacts site have recognized as a key feature of lipid dynamics [ Ref 32]. The proximity of mitochondria and lipid droplets is necessary for the ATP production, via β-oxidation. Recent ndings described host lipid metabolic remodelling associated with human-pathogenic propagation of HCoV-229E, suggesting that lipid metabolism regulation could be a common event for coronavirus infections [ Ref 28].
Modulation of host lipid metabolism has been reported to be necessary for replication of virus, such as hepatitis C virus (HCV), and picornaviruses [ Refs 33,34]. Several studies demonstrated that targeting host lipid metabolism by statins, allow to suppress viral replication of many positive-strand RNA viruses, such as Hepatitis C virus, Dengueviruses, Japaneseen-cephalitis virus, West Nile virus and in uenza A virus. Statins, are able to destabilize lipid rafts involved in the viral replication phases, as they constitute packets of vesicles capable of concentrating virus replication factors [ Refs 35,36].
We showed that type II alveolar epithelial cells appear to be the main target of the SARS-CoV-2. Whilst SARS-CoV-2, induces rapid dramatic ultrastructural changes and death in the host cells, the morphological features do not correspond to the activation of apoptosis, which is, by contrast, the main mechanism of cell death induced by SARS-CoV infection [ Refs 37,38]. Our results on SARS-CoV-2 infected cells suggested that a distinct type of cell death, with morphological features of both apoptosis and necrosis, namely Pyroptosis     Main differences in the ultrastructural cytopathogenesis induced by SARS-Cov-2 and SARS-Cov infection in cultured Vero cells, and analogye between in vitro and in vivo.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.