Vascular Inflammation in Lungs of Patients with Fatal Coronavirus Disease 2019 (COVID-19): Possible Role for the NLRP3 Inflammasome

Background: Hyperinflammation is a key event that occurs with SARS-CoV-2 infection. In the lung, hyperinflammation leads to structural damage to tissue. To date, numerous lung histological studies have shown extensive alveolar damage, but there is scarce documentation of vascular inflammation in postmortem lung tissue. Methods: Lung sections from 8 COVID-19 affected and 11 non-COVID-19 subjects [of which 8 were acute respiratory disease syndrome (ARDS) affected and 3 were from subjects with non-respiratory diseases] were stained for H & E to ascertain histopathological features including presence of thrombi/microthrombi. Inflammation along the vessel wall was also monitored by quantification of the expression of moieties of the NLRP3 inflammasome pathway (NLRP3 and caspase-1). Results: In lungs from “fatal COVID-19”, vascular changes in the form of microthrombi in small vessels, arterial thrombosis, and organization were extensive as compared to lungs from “non-COVID-19 non respiratory disease” affected subjects. The NLRP3 pathway components were significantly higher in lungs from COVID-19 subjects as compared to non-COVID-19 fatal cases without respiratory disease. No significant differences were observed between COVID-19 lungs and non-COVID-19 ARDS lungs. Conclusion: We posit that inflammasome formation along the vessel wall is a characteristic of lung inflammation that accompanies COVID-19. Thus, the NLRP3 inflammasome pathway seems to be probable candidate that drives amplification of inflammation post SARS-CoV-2 infection.


Introduction
It has been more than a year since the pandemic caused by the novel SARS-CoV-2 corona virus (Severe Acute Respiratory Syndrome Coronavirus), also known as COVID-19 has affected large populations globally [1,2]. The virus disproportionately affects the respiratory system and a major cause of fatality is the acute respiratory distress syndrome (ARDS) that accompanies the infection [3,4]. Autopsy-based lung histological studies have been an invaluable tool in understanding the pathobiology of COVID-19; indeed these have shown indications of in ammation, edema, coagulopathy and brosis [3,[5][6][7][8]. COVID-19 manifests itself under a wide spectrum of symptoms, but it can broadly be classi ed as an in ammatory disease where excessive in ammation is the main driver of poor clinical outcome [9,10]. In this direction, the vascular endothelium, a dynamically adaptable interface that is actively involved in recruitment of in ammatory cells, possibly plays a crucial role in regulation, progression, and ampli cation of in ammation. While post mortem ndings have shown alveolar damage, early or intermediate proliferative phase, and presence of thrombi and signs of in ammation in the lungs [3,6,8], histopathology in the context of vascular in ammation and altered vascular structures has been somewhat scarce [7,11].
In ammatory processes involve the participation of in ammasomes that are multimeric platforms assembled in response to pathogenic stimuli. Dysregulated in ammasome signaling has been well established as a pivotal event in hyper-in ammatory syndromes [12][13][14]. Among the in ammasomes, the NLRP3 in ammasome comprising of the NLRP3 subunit, ASC and caspase-1, is well established to be activated in response to microbial infection [15,16] and to drive cell death [17,18]. It is also involved in COVID-19, as evidenced by the detection of in ammasome subunits and products in the sera and lung tissue of COVID-19 patients [19,20]. However, there are no reports of the presence of the in ammasome in the pulmonary vasculature with COVID-19 infection. As the vasculature seems to be crucial in in ammation accompanying COVID-19, the status of NLRP3 along the vascular wall needs to be documented.
We posit that in ammasome formation is characteristic of pulmonary vascular in ammation that accompanies COVID-19. The purpose of this study is to contextualize vascular features in lung tissue in fatal cases of COVID-19 as compared to other pulmonary diseases and ascertain NLRP3 expression along the vascular wall. Here we document the major histological ndings of 8 postmortem examinations done on patients with clinically con rmed COVID-19 and compare these to lungs of non-COVID-19 subjects. This study contributes to the growing data on this topic [3,6,[21][22][23][24] .
We analyzed lung tissue samples of 8 patients that died of COVID-19 in 2020 and 11 patients that died from non-COVID complications. Written informed consent was obtained for postmortem examination from the next of kin of these patients. For the COVID-19 patients, SARS-CoV-2 infection was con rmed by real time PCR analysis at the time of hospital admission. Autopsies were done by trained personnel using personal protective equipment in accordance with the recommendations of the University of Pennsylvania School Of Medicine.
Tissue blocks taken from the most representative areas of the lung (as identi ed by macroscopic examination) were xed in formalin. Para n embedded sections of 3 to 5 µm thickness were stained with hematoxylin and eosin (H & E). Images were captured on the Aperio Pathology System and visualized by ImageScope (Leica Biosystems, Buffalo Grove, IL). High and low powered elds were selected for evaluation. In ammation and in ammation induced cell death (pyroptosis) were characterized by immunostaining for NLRP3 in ammasome and caspase-1 respectively. Sections were depara nized; after antibody retrieval, were stained using anti-human NLRP3 monoclonal antibody at 1:200 or anti-human caspase antibody at 1:100 (both from R&D Systems, Minneapolis, MN). Secondary antibody used was conjugated to Alexa 488 at 1:200 (Life Technologies, Eugene, OR). Appropriate IgG controls were used to x exposure settings. Vectashield antifade mounting medium used was from Vector Labs (Burlingame, CA). Images were acquired by epi uorescence microscopy using a Nikon TMD epi uorescence microscope, equipped a Hamamatsu ORCA-100 digital camera, and MetaMorph imaging software (Universal Imaging, West Chester, PA, USA). Fluorescence images were acquired at excitation = 488 nm; all images were acquired with the same exposure and acquisition settings as reported previously [25][26][27]. Quantitation of the uorescence signal was carried out using the MetaMorph Imaging Software. Integrated Intensities were normalized to the eld area as reported by us elsewhere 40 . Tables 1 and 2, histological characteristics in Tables 3 and 4. COVID-19 patients were 4 men and 4 women, with a mean age of 71.8 years (SD 13.9); non-COVID-19 patients were 7 men and 4 women, with a mean age of 64 (SD 10.7). Lung sections from all patients showed diffuse alveolar damage including hyaline membranes, intra-alveolar brin deposition, and thickening of the alveolar-capillary membrane. All sections from lungs also stained positively for the NLRP3 in ammasome associated markers that were assessed and quantitated by uorescence imaging.    Upon light microscopic examination, the lungs of all COVID-19 patients showed extensive alteration of lung microstructure (Fig. 1A, B). A closer inspection of COVID-19 lungs revealed brin exudation into alveolar space, extensive thrombi and broblastic proliferation, hyaline membrane, brin deposition and early and advanced proliferative phase of diffuse alveolar damage (Fig. 1B). Thrombi and microthrombi were identi ed in 7 of the 8 patients (Fig. 1C). Vascular changes were extensive, with microthrombi in small vessels and arterial thrombosis and organization. Microthrombi were also observed in alveolar septa. Thrombi and microthrombi were found in > 75-80% of the elds imaged. Histological ndings are detailed in the legends of Fig. 1 and in Table 2.

Patient demographics and clinical information are summarized in
In contrast, the lungs from non-COVID fatal cases, showed less thrombi and brin exudation ( Fig. 2A, B). While higher magni cation showed certain key features of lung injury such as diffuse alveolar damage, thickening of the alveolar-capillary membrane, broblastic proliferation, the presence of hyaline membranes, edema and proliferative phase of diffuse alveolar damage, the non-ARDS lungs (nc 1, 8 and 11) have intact structure and did not show alveolar in ltration or hemorrhage (Fig. 2B). Furthermore, in non-COVID-19 lungs, vessels showed thrombus in about < 40% of the elds (Fig. 2C).
We next assessed the expression of the NLRP3 subunit and its downstream effector caspase-1 in all samples. In lungs from COVID-19 subjects, intense expression of the NLRP3 and caspase-1 as observed from the green-uorescent signal, is shown in Fig. 3. Fluorescence around the vessel walls implied NLRP3 expression along the endothelial layer (Fig. 3A, upper panels). The effector enzyme, caspase-1 was widely distributed throughout the lungs and was not limited to the vascular structures (Fig. 3B, upper panels). In lungs from non-COVID subjects that were not affected by respiratory disease (nc1, 8 and 11), NLRP3 (Fig. 3A, lower panels) and caspase-1 expressions were signi cantly lower (Figs. 3A, B: lower panels and Fig. 3C). However, in lungs of subjects, that were affected by ARDS, NLRP3 and caspase1 expression was not signi cantly different from COVID-19 lungs (Fig. 3D).

Discussion
COVID-19 has been described largely as a respiratory disease; indeed, the respiratory tract and alveolus are amongst the primary sites of infection. However, it is also an in ammatory disease where release of in ammatory cytokines is the cause of organ injury and damage. The endothelium is the converging site of the in ammation as its activation (expression of adhesion molecules and cytokines) leads to immune cell recruitment; thus it is reasonable to conclude that COVID-19 is potentially a vascular disease [11,28,29]. While this would be an indirect impact of the virus, more recent studies also provide evidence of a direct effect i.e. infection by SARS-CoV-2 virus of endothelial cells [30]. Our inspection of autopsies of the 8 COVID-19 patients showed macro and microthrombi in almost all elds imaged, indicating coagulation pathology. This was not observed in autopsies of non-COVID-19 lung sections. As is well established, coagulation is closely linked to endothelial in ammation signaling; in ammatory moieties on the endothelium increase leukocyte in ltration and alter coagulation control driving a procoagulant direction [31]. Thus, COVID-19 which is increasingly being described as a vascular disease should perhaps be more accurately de ned as a pathology which has its origins in "endothelial in ammation" signaling.
In ammasome activation on the endothelium plays a major part in cell death and injury with in ammation. The NLRP3 in ammasome is a multiprotein complex comprised of three basic components: (1) A sensor such as a NOD-like receptor (NLR) (2) the adaptor protein apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) and (3) the in ammatory cysteine aspartase caspase-1. The assembly of this complex leads to release of caspase-1 which then exerts its catalytic activity on the pro-in ammatory cytokines (IL-1β) that after their release perpetuate cell death, speci cally in ammation induced cell death or pyroptosis [17,18].
A recent report showed high levels of NLRP3 in ammasome and caspase-1 in patients with fatal COVID [20]. This is not surprising as increased NLRP3 is associated with various in ammatory lung pathologies including acute lung injury and ARDS [32,33]. The COVID-19 lung autopsies in this study, showed NLRP3 expression throughout the lung, but intense expression was seen along the lung vessel walls implying in ammasome expression on the endothelium. The downstream effector of NLRP3 in ammasome activation, Caspase-1 was found to be expressed throughout the lungs including in the vascular structures.
Caspase-1 is considered as a key pyroptotic mediator; it reportedly drives pulmonary vascular endothelial cell death [17]. Elsewhere too, high caspase-1 expression has been reported with both COVID-19 [20] and with other lung in ammatory pathologies [34]; however its expression on the endothelium or vascular wall with COVID-19 has not been documented. Possibly the NLRP3-caspase-1 axis can directly (via caspase-1 driven pyroptosis) or indirectly (via NLRP3 driven chemotactic immune cell recruitment [35]) injure the endothelial layer. This con uence of vascular injury, thrombosis and dysregulated in ammation seems to propagate lung damage with COVID-19 and supports a pivotal role for the pulmonary endothelium in severe and fatal COVID-19. In contrast, non-COVID-19 lungs of subjects that did not have respiratory disease, had signi cantly lower expression of NLRP3 and caspase-1, indicating that an engagement of the NLRP3 pathway in COVID-19 and in ARDS.
As NLRP3 in ammasome driven pyroptosis is being considered to play a leading role in the pathogenesis of multi-organ failure with COVID- 19 [36], there is some speculation on the mechanisms by which in ammasome activation occurs upon SARS-CoV-2 infection. One possibility is that the SARS-CoV-2 spike protein's binding to cell surface-expressed angiotensin-converting enzyme 2 (ACE2) directly triggers its enzymatic activation and alters membrane polarity that can result in activation of NLPR3 in ammasome [37]. Or NLRP3 could be activated via Angiotensin II which is reported to facilitate the assembly of the in ammasome. A third possibility could be via interaction of damage associated molecular patterns (DAMPs that are released post infection) and members of the complement cascade with the SARS-CoV-2 virus. Potent cleavage fragments of DAMPs and complement cascade can potentially activate the in ammasome [38]. Yet another possibility is that the stretch from ventilation activates the in ammasome [39]. Once activated around the vascular wall (endothelial layer), the NLRP3 in ammasome would lead to release of caspase-1 and interleukin-1β that would facilitate pyroptosis (cell death) of the endothelium (Schema 1).
To the best of our knowledge, this is the rst study on NLRP3 expression in the vascular structures in lungs of fatal cases of COVID-19. The origin of several events that exacerbate in ammation and injury with COVID-19 (such as immune cell aggregation and extravasation, edema, formation of thrombi and leukopenia) possibly lies in pulmonary endothelial in ammasome activation and pyroptotic cell death. Therefore, NLRP3 inhibitors have been suggested for as a potential treatment strategy and are currently being explored for management of moderate COVID-19 symptoms (NCT04540120) [19,40].
A major drawback of this study is that our sample size is small. Moreover, para n based post-mortem samples offer a snapshot of the disease and cannot recreate the evolving disease process. Histology is also impacted with the effects of clinical care and treatment as comorbidities, ventilation and medication pose as challenges in interpretation of results. Nevertheless, this study identi es endothelial NLRP3 in ammation, and documents thrombi and altered vascular structures in the lungs of fatal COVID-19 patients. Hematoxylin and Eosin-stained sections staining from representative regions of the lung parenchyma of post-mortem lung tissue of 8 COVID-19 patients. A. All patients show extensive alteration of lung microstructure in the form of alveolar damage, brin exudation into alveolar space, thrombi and broblastic proliferation. The septa are thickened and there is presence of hyaline membranes and dense in ltrates. Scale bar is 3 mm. 1: Alveolar damage with collagen deposition and exudative pattern of damage 2. Large thrombi and smaller caliber arteries showing brin thrombi (arrows) 3. Alveolar damage pattern arising from broblastic proliferations 4 and 5. Exudate in the entire lung 6. Necrosis with blood and exudate in the lung parenchyma 7. Hemorrhagic infarction of lung tissue adjacent to a pulmonary artery with thrombotic material 8. Pulmonary hemorrhage with blood and brin exudation into the parenchyma B. H and E staining at higher magni cation: All patients had extensive diffuse alveolar damage, microthrombi and edema in regions of the lung. A. Fibroblastic proliferation B. Plugged airway due to remodeling C. Coagulation necrosis with blood in the lung tissue D. Proliferative phase of diffuse alveolar damage E. Patchy distribution of damage F. Proteinaceous exudates in alveolar spaces G. Blood and brin exudation into parenchyma H: Fibroblastic proliferation I: Endotheliitis of small vessel <100 μm with in ltration of the vessel wall by lymphocytes (arrow shows in ltrated cells) C. (unavailable with this version):Thrombi and microthrombi were identi ed in 7 of the 8 patients. Images of vessels were chosen to emphasize the microthombi. Box is magni ed in the right panel. Arrow shows microthombi on alveolar septa.

Figure 2
A. Hematoxylin and Eosin-stained sections staining from representative regions of the lung parenchyma of post-mortem lung tissue of 11 non COVID-19 patients. Scale bar is 3 mm. B. H and E staining at higher magni cation: All patients had diffuse alveolar damage, microthrombi and edema in regions of the lung. Arrows show proteinaceous exudate in the airspaces. Scale bar is 200 microns C. Vascular structures in lungs from non-COVID-19 sources. Arrows show thrombi in vessels. About 40% of the elds showed thrombi. Scale bar is 100 microns.

Figure 3
In ammasome in the lungs of patients with COVID-19 infection. Representative images of the immuno uorescence in lung sections stained with anti-NLRP3 and Caspase-1. A. The NLRP3 subunit (green) was visualized along the walls of arterioles (arrow). Upper panels: COVID-19 lungs. Lower Panels: Lungs from non-COVID 19 subjects, without respiratory disease. B. Caspase staining (green). Upper panels: COVID-19 lungs. Lower Panels: Lungs from non-COVID 19 subjects, without respiratory disease. C and D. Quantitation of the uorescence Intensity of the images using MetaMorph Imaging Program. *p<0.01 as compared to non-COVID lungs. Overview of SARS-CoV-2 entry, infection and endothelial in ammation and cell death. As is well established, oral nasopharyngeal entry of SARS-CoV-2 is followed by its binding to the alveolar epithelium. The infected pneumocytes secrete cytokine and chemokines, which attract neutrophils to the alveolar space, leading to a possible breach of the alveolar wall. Meanwhile, endothelial cells overexpress NLRP3 as we observed in the autopsies (either by infection, or via increased amounts of chemokines and cytokines). The NLRP3 pathway drives endothelial pyroptosis. The leads to breakdown of the endothelialalveolar barrier and causes interstitial and alveolar space ooding. Endothelial cell death and debris activates coagulation cascades that promotes thrombi formation.