Caspase-1 activation and a predominant IL-1/IL-6 signature associated with IFNγ-induced chemokines remain highly detectable in the BALF of steroid-resistant COVID-19-ARDS, arguing for new multi-targeted therapeutics in COVID-19-ARDS. Lung biopsies from deceased COVID-19 patients show Diffuse Alveolar Damage and vasculopathy associated with activated-caspase-1. In the lungs of C-ARDS and NC-ARDS, a predominant caspase-1-induced IL-1β/IL-6 signature and IFNγ -induced chemokines persist.

Coronavirus-disease-2019 (COVID-19), caused by severe-acute-respiratory-syndrome-coronavirus-2 (SARS-CoV-2) is associated with dysregulation of host immune responses (1). In severe forms, COVID-19 causes bilateral pneumonia, that rapidly progresses to acute-respiratory-distress-syndrome (ARDS) and death (1). Immunothrombosis is one of the main complications arising in severe forms of COVID-19, due to the combined hyperactivation of the immune and coagulation systems (2–5). Multiple immune pathways have been shown to play a role in the immunopathogenesis of COVID-19. Type I interferons (IFN) promote the activation of antiviral effector mechanisms and induce the production of pro-inflammatory cytokines (6, 7). Conversely, the impaired type III-IFN response observed in severe cases of COVID-19 may promote innate immunity with hyperinflammation and immunothrombosis (2, 8–10). Increased circulating levels of pro-inflammatory cytokines have been detected in patients with severe COVID-19, as well as increased activation of nuclear-factor-kappa-light chain enhancer of activated B cells (NF-kB) (11). Inflammasomes markers such as the NACHT, LRR, pyrin domain (PYD) domains-containing protein 3 (NLRP3) and the absent-in-melanoma-2 (AIM2) are also activated during COVID-19 (12). Danger-associated molecular patterns (DAMPs) and pathogen-associated molecular patterns (PAMPS) released upon tissue injury are detected by NLRP3, a cytoplasmic protein which is in turn, activated into a macromolecular complex called NLRP3-inflammasome (13, 14). The activated NLRP3-inflammasome engages through the PYD domain of ASC (apoptosis-associated speck-like protein containing a caspase recruitment domain [CARD]) pro-caspase-1, which then undergoes autocatalytic activation to caspase-1. Activated caspase-1 (Casp1p20) is then responsible for transforming pro-IL-1β and pro-IL-18 into biologically active IL-1β and IL-18, two major pro-inflammatory and pro-coagulant cytokines of the IL-1 family. Casp1p20 also cleaves gasdermin-D, which forms cytoplasmic membrane pores and triggers pyroptosis (15). During COVID-19, strong activation of the NLRP3-inflammasome and caspase-1 is induced with subsequent release of IL-1β and IL-18, although these cytokines remain difficult to detect in patient’s blood (14, 16). Although controversial, caspase-1-induced pyroptotic cell death also appears to occur following SARS-CoV-2 infection of mononuclear cells in vitro (15, 17–19). Furthermore, in humanized mice expressing the human ACE2 receptor, SARS-CoV-2 infects macrophages by binding of anti-spike IgG to CD16, resulting in NLRP3-inflammasome activation, IL-1β and IL-18 production, and pyroptosis (19–21). Both pyroptosis and necrosis play an important role in hyperinflammatory syndromes, mainly through a significant release of DAMPs, notably IL-1α, which maintains an upward inflammatory loop (20, 22–24). IL-1α induces inflammasome activation and IL-1β production and, like IL-1β, is mediated through the IL-1 receptor. Despite their central role in amplifying the innate immune response, the role of DAMPs in COVID-19 is poorly understood to date. Moreover, the specific contribution of NLRP3 inflammasome activation and IL-1 signaling to COVID-19 immunopathogenesis remains unclear. In the present study of three prospective cohorts of COVID-19-patients, we investigated caspase-1 activation and IL-1 signature using post-mortem lung tissue, bronchoalveolar lavage fluid (BALF), and serum across the spectrum of COVID-19 severity. We also investigated the hypothetical role of other important inflammatory pathways and DAMPs in COVID-19 immunopathogenesis.

In this prospective multicenter study, we used several approaches such as histopathological and immunohistochemical analyses on post-mortem tissues from subjects with fatal C-ARDS, and immune profiling of BALF and serum from patients across the COVID-19 severity spectrum to provide information on NLRP3-inflammasome activation and the involvement of IL-1 signaling in COVID-19 patients.

Consistent with previous reports (27–30), DAD was present in almost 60% of subjects with fatal COVID-19-ARDS, and interstitial mononuclear cell inflammation in half of them. Inflammatory vascular lesions, including endothelitis, were detected in the vast majority (>80%) of subjects died of COVID-19 disease. In the same time, pulmonary thromboembolism and thrombotic microangiopathy were also frequent (>40%), confirming the idea that vasculopathy, coagulopathy and immunothrombosis contribute centrally to the pathogenesis of refractory ARDS in patients with COVID-19 (2, 29, 31–34).

Expression and localization of Casp1p20, an indicator of NLRP3-inflammasome activation, were assessed in the lungs of patients with refractory C-ARDS. Interestingly, intense Casp1p20 expression was also detected in abundance (>90%) at sites of vascular injury and vessel thrombosis, indicating a spatial association between NLRP3-inflammasome activation and COVID-19-associated vasculopathy.

In agreement with autopsy findings, increased levels of sNLRP3 and sCasp1p20 were observed in both BALF and serum from C-ARDS, with sNLRP3 concentrations considerably higher in BALF than in serum. The lung is thus a major site of inflammasome assembly and activation during C-ARDS. In addition, serum levels of sNLRP3 and sCasp1p20 were globally increased in all COVID-19 patients, with Casp1p20 concentrations higher in severely ill compared to moderate ill patients. This suggests the involvement of caspase-1 in hyperinflammatory responses. Collectively, these findings are consistent with activation of the NLRP3-inflammasome pathway in both lungs and blood and also corroborate previous observations showing elevated Casp1p20 levels in COVID-19 patients and inflammasome activation in patient-derived PBMCs and in CD14⁺ macrophages resident in COVID-19 lung tissue (17, 29, 35, 36). NLRP3-inflammasome activation can be directly triggered by multiple SARS-CoV-2-derived proteins, including spike protein and ORF3a, which act as PAMPs (19).

Activation of the NLRP3-inflammasome and subsequent cleavage of caspase-1 are responsible for the processing of IL-1β and IL-18 precursors into biologically active cytokines. In turn, IL-1β induces IL-6 synthesis, as well as the production of IL-1Ra, a natural endogenous inhibitor of IL-1. IL-18, in combination with IL-12, is a major inducer of IFNγ production (37). Due to its extremely short half-life, IL-β is usually detected at very low concentrations in the bloodstream, even in disorders typically driven by IL-1β. As expected, we observed low serum IL-β concentrations in COVID-19 patients, which were, however, significantly higher than controls. Interestingly, we found high concentrations of IL-1β (reaching >10,000pg/ml in some individuals) in BALF, indicating intense activation of the pulmonary NLRP3-inflammasome and IL-1β production during COVID-19 disease. These observations corroborate and extend data from single-cell RNA transcriptomic analyses suggesting increased IL-1β in BALF of smaller cohorts of COVID-19 patients (38–40). In contrast to IL-1β, IL-1Ra and IL-6 are generally secreted at higher concentrations and appeared elevated in both BALF and serum of COVID-19 patients. In our study, serum IL-6 correlates with the severity of COVID-19, as previously shown by others (41, 42). Our results, showing abundant expression of activated caspase-1 associated with pulmonary vascular injury and thrombosed vessels, as well as elevated levels of NLRP3, caspase-1, IL-1β and IL-6 in the pulmonary microenvironment, confirm the role of the NLRP3-inflammasome pathway in COVID-19-associated immunothrombosis, as we previously suggested (14).

IL-18 was increased in BALF but especially in the serum of C-ARDS patients compared to controls, showing a positive association with disease severity as previously described (17, 43). However, serum IL-18 levels remained lower than those observed in other hyperinflammatory syndromes (44). We found low but detectable concentrations of IFNγ in C-ARDS patients, which in BALF appeared higher in C-ARDS and NC-ARDS than in controls. However, an IFNγ-induced signature is probably present in COVID-19, since CXCL10 (in common with CXCL9), an important IFNγ-induced chemokine, is highly detectable, as previously observed in several IFNγ-mediated syndromes (45). CXCL10 levels were elevated in BALF and serum of C-ARDS patients and correlated with disease severity. Several other reports have shown an elevated CXCL10 signature during COVID-19, including in BALF (39, 40, 46) and CXCL10 is a good circulating marker of disease severity (43, 47), consistent with the fact that a Th-1 signature can develop over time during severe COVID-19 (43, 48).

Inflammasome-dependent caspase-1 activation can also trigger pyroptosis with subsequent release of intracellular contents (49, 50). Increased levels of intracellular proteins such as NLRP3 and Casp1p20 observed in BALF and serum from C-ARDS patients suggest widespread cell death resulting from direct viral cytotoxicity or inflammatory damage. Consequently, we also found detectable levels of DAMPs such as IL-1α and HMGB1 in BAL of C-ARDS patients. These intranuclear proteins are released after cell death. They probably originate from injured cells of respiratory tractus, although the nature of these cells (bronchial epithelial, alveolar or endothelial…) remains uncertain. The specific contribution of DAMPs to COVID-19 immunopathogenesis remains incompletely understood, but the role of DAMPS is investigated in the review by Land et al. When viral load is too high, for example, in the respiratory tract, authors suppose a “forcing” of many virus-infected host cells to decide to commit “suicidal” regulated cell death (e.g., necroptosis, pyroptosis) associated with release of large amounts of DAMPs. Ironically, although the aim of this “suicidal” cell death is to save and restore organismal homeostasis, the intrinsic release of excessive amounts of DAMPs leads to those dysregulated hyperinflammatory responses—as typically involved in the pathogenesis of acute respiratory distress syndrome and systemic inflammatory response syndrome in respiratory viral infections (21). DAMPS could be released as a direct consequence of SARS-CoV-2 cytotoxicity, triggering cell necrosis, or indirectly because of the excessive inflammatory response, leading to necroptosis or pyroptosis associated with the release of large amounts of DAMPs. The excessive emission of DAMPS thus released into the extracellular environment could promote dysregulated hyperinflammatory responses, and pro-inflammatory programs, such as the activation of inflammasomes, notably the NLRP3 inflammasome (21, 51–53). In response to NLRP3 inflammasome activation, Caspase-1 is cleaved, resulting in pyroptosis. Moreover, IL-1ß and IL-18 released after cleavage by activated Caspase-1 belong to DAMPS. DAMPS could therefore participate in a positive feedback loop of the innate immune response induced by SARS-CoV-2 infection. The authors defined different classes of activating DAMPS in the event of cell lysis, and suppressor DAMPS or compensatory SAMPS in the initiation of controlled defense responses favoring, depending on their balance, either inflammation or resolution of inflammation in the event of infection (21). Two studies have reported elevated circulating concentrations of HMGB1 and calprotectin (S100A8/A9, a marker of neutrophil activation) in serum from COVID-19 patients (54, 55). In a study by Renieris et al., mice injected with serum from COVID-19 patients showed elevated pulmonary expression of pro-inflammatory molecules, including TNFα, IL-6 and CXCL10 resembling human COVID-19. Treatment of these mice with anakinra (which blocks both IL-1α and IL-1β signaling) or with antibodies selectively targeting IL-1α significantly attenuated COVID-19-like pulmonary immunopathology, potentially identifying IL-1α as a mediator of inflammation and tissue-specific injury (54, 56). Taken together, these observations may help to explain the limited efficacy of the anti-IL-1β monoclonal antibody, canakinumab, compared to anakinra in COVID-19 (57–61). Aware of the limitations of these observations, we plan to pursue this work by characterizing and specifically targeting DAMPS and eventually SAMPS in BALF and serum from ARDS patients.

We found that NLRP3-inflammasome activation and concentrations of IL-1α, IL-β, IL-6, IL-1Ra or CXCL10 were broadly similar in C-ARDS and NC-ARDS. However, concentrations of TNFα, sTNF-R1 and CXCL8 both in BALF and serum, were significantly higher in NC-ARDS than in C-ARDS, suggesting that inflammatory pathways involving NF-kB could be probably more up-regulated in NC-ARDS than in C-ARDS. A more likely explanation of these results may be due to the different use of steroids, which are well-known inhibitors of NF-kB-induced gene transcription. Glucocorticoids act by modulating the gene expression of several proteins involved in the inflammatory response. The reduced transcriptional activity of NF-κB is key to the anti-inflammatory effects of glucocorticoids. As expression of many of the pro-inflammatory cytokines implicated in COVID-19 is NF-κB-dependent, including IL-6, TNFα, and CXCL10 among others (62). Indeed, all C-ARDS subjects included in this study received steroids as a part of standard and recommended treatment, whereas NC-ARDS subjects were steroids free. It should be noted that all C-ARDS patients included in this study were in fact refractory to this steroid therapy, experiencing severe clinical deterioration despite well-conducted treatment, and that in these patients, steroids demonstrated almost a poor biological inhibitory effect on IL-1α, IL-β, IL-6, IL-1Ra or CXCL10 levels (62). The deterioration of COVID-19 patients is inextricably linked to immunopathological phenomena rather than viral load. The aggressive inflammatory responses in COVID-19 result in damage to the airways, termed ARDS, which may lead to respiratory failure and death. Corticosteroid treatment reduced mortality, but subsequent meta-analyses have again highlighted the importance of timing when considering corticosteroid administration, as corticosteroid use was associated with increased viral load. NLRP3 inflammasome activation is strongly correlated with COVID-19 severity and part of dexamethasone’s clinical effect in COVID-19 may be via NLRP3 inhibition (63). But our results suggest that these inflammatory cytokines,IL-1α, IL-β, IL-6, IL-1Ra or CXCL10, may be poorly targeted by steroids and may require to be specifically targeted in C-ARDS patients with poor response to steroids (64–66). Thus, in pursuit of this work, we aim to assess and analyze some of the anti-inflammatory cytokines, dependent from NLRP3 inflammasome pathway or from NF-κB pathway, to show how the balance between pro- and anti-inflammatory cytokines is affected in C-ARDS, NC-ARDS and control pulmonary vasculopathies, before and after steroid treatment. Furthermore, data showing similar IL-1/IL-6 cytokine profiles in C-ARDS and NC-ARDS patients suggest that these cytokine pathways may be highly activated in all types of ARDS, leading to the conclusion that ARDS should be considered as an hyperinflammatory state with DAD and severe prognosis in which novel anti-inflammatory strategies targeting these cytokines merit further mechanistic investigation (30). Specific NLRP3 inhibitors are currently undergoing clinical trials for the treatment of COVID-19 (63).

This study is limited in terms of power and clinical significance. Our control patients represented an imperfect control, since they were patients who might have incipient bronchopulmonary cancer. Bronchopulmonary cancer develops in inflammatory tumor microenvironment, which may involve the NLRP3 pathways. Tumor cell-intrinsic mechanisms of NLRP3 activation in the tumor microenvironment has been described by Tengesdal. et al (67). Aberrant activation of NLRP3 within the tumor microenvironment results in increased IL-1β and IL-18 secretion. Dysregulation of these cytokines induce tumor promoting mechanisms, such as angiogenesis, immunosuppression and metastasis. NLRP3 activation therefore represents a key immune checkpoint within the tumor microenvironnement, acting as a master switch for inflammation-mediated tumor progression. Nevertheless, according to our results from controls, ELISA protein expression levels of the inflammatory cytokines from the IL-1 family, and dependent from NLRP3 pathway, were significantly lower in BAL and in serum than in the other groups studied. Moreover, NLRP3 activation in cancer remains controversial. The role of the NLRP3 inflammasome therefore appears to be secondary in this inflammatory tumor microenvironment, after Transforming growth factor β (TGF-β). The article by Zhao et al. characterizes the tumor microenvironment in relation to carcinogenesis, and the predominant role of TGF-ß/SMAD4 signaling in cancer. TGF-β signaling pathway plays important roles in many biological processes, including cancer initiation and progression (68). Another predominantly inflammatory pathway, is interferon persistent signaling. Mathew et al. described the persistent IFN signaling as a potent immunoregulatory effect, promoting carcinogenesis. In cancer cells and immune cells, chronic IFN-I signaling is linked with cancer resistance in humans. IFN signaling thus appears to be involved in the carcinogenesis process, and argue to using combined JAK inhibition and PD-1 immunotherapy for non-small cell lung cancer patients (69). It seemed ethically difficult to propose another group controls totally asymptomatic.

Limitations of our study lie in the small number of patients in the study groups, and the missing data, resulting from mainly technical reasons, like insufficient quantity of samples or poor preservation of samples.

Finally, these results are therefore not generalizable but raise the hypothesis of an IL-1/IL-6 and IFNγ/IL-18 cytokine signature in ARDS BAL, in response to activation of the NLRP3 inflammasome in lung cells. These results are preliminary data to support clinical trials for new NLRP3 inhibitors, anti-IL-1 and anti-IL-6 therapies or JAK ½ inhibitors in COVID-ARDS and in ARDS.

As a conclusion, the results of this study support a role for the NLRP3-inflammasome activation and IL-1 in the immunopathogenesis of COVID-19. The presence of increased plasma levels of sNLRP3, sCasp1p20 and IL-1 cytokines (IL-1β, IL-18) was demonstrated in hospitalized patients across the spectrum of COVID-19 severity. Upregulation of this pro-inflammatory IL-1β/IL-6 pathway was also observed in both BALF and plasma obtained from critically ill patients with steroid-refractory-C-ARDS, but predominantly in BALF, better reflecting the inflammatory alveolar microevironment. Consistent with NLRP3-inflammasome activation, caspase-1 activation was detected on post-mortem lung tissues and mainly registered with lung-residing macrophages, localizing predominantly at the areas of alveolar injury and foremost-injured vasculature and thrombosis, hence supporting NLRP3-inflammasome activation as a putative mechanism contributing to immunothrombosis in COVID-19. IL-18/IFNγ-induced pathways also appeared to play a role in patients with steroid-resistant C-ARDS and may be insufficiently targeted by steroid therapy alone, arguing for new multi-target strategies in COVID-ARDS.