The disease COVID-19 encompasses a wide clinical spectrum, ranging from asymptomatic disease, or mild upper airway respiratory symptoms, to severe pneumonitis, respiratory failure and death (1). Before the introduction of vaccines against COVID-19, approximately 15% of symptomatic patients progressed to severe disease, which required respiratory support, and 5% of patients subsequently developed critical disease including acute respiratory distress syndrome (ARDS), septic shock or multiorgan failure (2). Advanced age, lack of SARS-CoV-2 antibody immunity, obesity, male sex, and the number of chronic comorbidities are today identified as the most important risk factors for patients to develop severe respiratory failure and death (3–6). Although a dysregulated immune response with delayed viral control and subsequent deleterious hyperinflammation is suggested to be a main feature of the pathogenesis in COVID-19, the mediators and culprit mechanisms of the hyper-inflammation are still not elucidated (7).

Previous studies and reviews of COVID-19 have suggested that damage-associated molecular patterns (DAMPs), a heterogenous group of proteins or nucleic acid molecules released from damaged cells that trigger the innate immune response, may play a key role in the development of severe COVID-19 (8, 9).

In this study we hypothesized that the extracellular form of cold-inducible RNA-binding protein (CIRP), i.e. eCIRP, a newly identified DAMP, may be a driver of COVID-19 associated respiratory failure as experimental studies have shown that it is an important mediator of sepsis induced acute lung injury (10) and bleomycin induced pulmonary fibrosis (11). Cellular stressors such as hypothermia and hypoxia are known to induce upregulation and release of eCIRP into the extracellular space and circulation (12–14). When recombinant CIRP is injected in vivo, it is shown to stimulate the release of tumor necrosis factor-α (TNF-α), interleukin 6 (IL-6), and high mobility group box 1 (HMGB1) from macrophages, and to cause tissue injury (12). eCIRP binds to the toll like receptor 4-myeloid differentiation factor 2 (TLR4-MD2) complex or directly to TLR4 on the surface on immune cells, epithelial and endothelial cells, and subsequently activates various downstream inflammatory pathways (15). Elevated serum and tissue levels of eCIRP have been reported in both systemic, organ-specific ischemic, and chronic inflammatory diseases (15). When first described serum levels of eCIRP were undetectable in healthy volunteers, while increased in patients with hemorrhagic shock admitted to the intensive care unit (ICU) (12). Later, serum eCIRP levels have been associated with severe disease and poor prognosis in patients with sepsis, community-acquired pneumonia, and acute pancreatitis (16–19). Extracellular cold-inducible RNA-binding protein knock out mice (eCIRP-/-) are protected against inflammatory disease (15, 20), and neutralizing antibodies against eCIRP are shown to attenuate inflammation and improve survival in animal models of hemorrhagic shock, sepsis, and hepatic ischemia (12, 21). Consequently, eCIRP has been proposed as a therapeutic target in several inflammatory diseases.

The role of eCIRP in the pathogenesis of COVID-19 have not yet been explored. In this study we aimed to evaluate serum concentrations of eCIRP in relation to the degree of respiratory support, the extent of lung involvement on chest computed tomography (CT), and inflammatory markers in patients with COVID-19.

Severe COVID-19 is characterized by a dysregulated, hyper-inflammatory host-response to infection. The proinflammatory mediator eCIRP shows promise as a treatment target in sepsis, but its role in COVID-19 has not yet been studied. In this study we show that serum concentration of eCIRP is associated with the severity of acute COVID-19. Patients who developed severe hypoxemia had higher eCIRP concentrations compared to patients with mild disease and the association remained after adjustment for age, sex, BMI, co-morbidities, and symptom duration. We confirmed the results in a validation cohort composed of a higher percentage of patients with mild disease, including non-hospitalized patients recruited at a separate study site. In the validation cohort, eCIRP concentrations also differed statistically between patients with mild disease and non-severe hypoxemia. Another important finding is that eCIRP correlates to parenchymal lung tissue involvement on chest CT, and to CRP, IL-6, and ferritin.

This is, to the best of our knowledge, the first study to report an association between eCIRP and respiratory failure in COVID-19. Only a small study of hospitalized patients in China (n=28), has previously studied eCIRP in COVID-19 (23). In contrast to the results of the present study, they found decreased plasma levels of eCIRP in patients with COVID-19 compared to controls (n=9), no association to severity, and a negative correlation to inflammatory cytokines (23). However, the limited sample size makes it difficult to draw firm conclusions from their study. Considering the evidence from studies on other hyper-inflammatory states and the known biology of eCIRP, it is plausible that serum concentrations would increase, not decrease, in COVID-19 patients with ongoing inflammation. Several previous preclinical and clinical studies have also demonstrated that hypoxia, both acute and chronic, induce upregulation of intracellular CIRP expression and that acute hypoxia result in release of eCIRP to the extracellular space and circulation (12, 13, 24). Also, in line with our results, pre-pandemic studies in sepsis and community acquired pneumonia, found that elevated eCIRP levels were prognostic for developing a severe disease or lethal outcome (16, 17).

In addition to an association to the level of hypoxemia, we found that eCIRP correlates to the visual extent of lung infiltration when measured by ÖCoS, a validated radiological chest CT scoring system (22). Bilateral ground-glass opacities, with or without signs of tissue consolidation, is the typical chest CT presentation of COVID-19. The extent of parenchymal infiltration, measured by chest CT on hospital admission is known to correlate to the duration of hospitalization, intensive care admission or death in COVID-19 (1, 22, 25). Patients with a high percentage of lung involvment (>75%) by the visal scoring scale by ÖCoS have a high likelihood of developing a critical disease due to respiratory failure (22). The correlation between eCIRP and ÖCoS therefore strengthens the hypothesis that eCIRP-driven inflammation contributes to development of respiratory failure of COVID-19.

Respiratory failure and severe hypoxemia in COVID-19 is known to be linked to a hyper-inflammatory, pro-coagulative state in the lungs, which causes a mismatch between ventilation and perfusion due to edema, fibrosis, and to activated lung endothelia with disseminated micro vasculopathy with thrombosis and bleeding. The injection of exogenous eCIRP into healthy mice has been shown to cause a similar picture with lung injury characterized by vascular leakage, edema, neutrophil infiltration, NLRP3 inflammasome activation, and local production of proinflammatory cytokines (TNF-α, IL-1β) (26). Moreover, eCIRP is shown to induce pulmonary fibrosis in mice, and to induce release of neutrophil extracellular traps (NETs) during sepsis, resulting in lung injury (27). In COVID-19, NET concentration has shown to be increased in both human serum and tissue, suggesting that it contributes to the development of severe disease. Although this study cannot address the mechanisms of eCIRP induced respiratory failure in COVID-19, the results from preclinical studies and pre-pandemic clinical studies demonstrate that CIRP itself can induce the same pathology that is characteristic for respiratory failure in COVID-19. Taken together, it is likely that eCIRP may aggravate inflammation and tissue injury leading to the development of severe respiratory failure in COVID-19. Future studies are required to define the pathways involved and if active or passive secretion is the leading mechanism of eCIRP secretion in COVID-19. In sepsis, active eCIRP secretion is shown to be linked to inflammasome activation, regulated by the gasdermin D (GSDMD) pathway (28). It is possible that the same pathway regulates eCIRP secretion in COVID-19 given that inflammasome activation with engagement of the GSDMD-pathway is shown to contribute to the pathology of COVID-19 (29). However, the lytical viral replication stage and necrosis may also lead to passive eCIRP release, further aggravating inflammation. Thus, it can be proposed that eCIRP is released by a combination of active and passive processes in COVID-19.

Extracellular CIRP was positively correlated to the pro-inflammatory biomarkers CRP, IL-6, and ferritin. Similar correlations between eCIRP and biomarkers for severity, i.e., CRP, IL-6, and procalcitonin (PCT) has been observed in sepsis, community acquired pneumonia, and acute pancreatitis (16–18). An association between eCIRP and non-COVID-19 induced inflammation has previously been established in a number of cells resident in pulmonary tissue, e.g., macrophages, alveolar epithelial cells, endothelial cells, and pulmonary fibroblasts (10, 12, 26, 30, 31).

This study has several limitations. Chest CT was not included in the prospective study design and therefore only possible to be assessed in hospitalized patients from site 1, where it was part of the clinical routine of hypoxemic patients (22). Moreover, inflammatory biomarkers sampled simultaneously as eCIRP were only possible to assess in patients from the cohort at site 1. The relatively small sample size of patients with mild disease at site 1 may therefore affect the generalizability of these correlations in a cohort with non-hypoxemic patients. Second, in contrast to previous studies on COVID-19, we had a greater proportion of older patients in the mild disease group, which may also affect the generalizability. However, age did not affect the concentrations of eCIRP in the present study. Third, we focused on samples with the highest concentrations on day 0-4, not from study inclusion only. This was possible due to the variation of symptom duration on study inclusion. To account for the effect of symptom duration on eCIRP expression, we include this as a covariate in the regression analysis.

The strength of the study is the prospective design with data on the levels of respiratory support and that the results are validated in a different cohort including a different case-mix, recruited at a separate study site.

In this two-center study, we conclude that high serum levels of eCIRP are associated with acute respiratory failure in COVID-19. Due to the observational design, we cannot draw firm conclusions about the role of eCIRP in the pathogenesis of COVID-19. Nevertheless, with the evidence from previous studies on sepsis, demonstrating that eCIRP is involved in the same inflammatory pathways that have been identified as cornerstones in the pathogenesis of COVID-19 induced acute respiratory failure, it is biologically plausible that it has a key function in the pathogenesis of severe COVID-19.

Future clinical, and experimental studies are required to define the mechanistic role of eCIRP in COVID-19. An important area of investigation would be to study how eCIRP affects NETosis and to determine if eCIRP antagonists protects from progression to severe respiratory failure in COVID-19.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The study involving human participants was reviewed and approved by the Swedish Ethical Review Authority, file number 2020-01557. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AL and SC were involved in overall planning and supervised the work. SC, CA, JN, and AE initiated and supported the prospective study and SA, SC, AL, JN, AE participated in patient enrollment. SC, AL, FS and KD processed the laboratory data, performed the statistical analysis, and designed the figures. ML performed ÖCoS grading of Chest CTs where the initial grading was missing. FS and KD drafted the manuscript. AE edited the figures. ML, JS, OH, SA and AE aided in interpreting the results. All authors critically revised the manuscript. All authors approved the submitted version of the manuscript.

This work was supported by grants from Nyckelfonden Region Örebro County (DNR OLL-938628, OLL-961416) to OH and SciLifeLab National COVID-19 Research Program (VC-2020-0015), financed by the Knut and Alice Wallenberg Foundation, Umeå University & Region Västerbotten, Swedish Heart-Lung Foundation to CA and a consolidator grant from the Swedish Research Council (DNR 2020-06235) to MF and Centrum för klinisk forskning (CKF) Region Västmanland.

We would like to thank patients and controls for their participation, and assistance from research and clinically working nurses, assistant nurses and other personnel in their help with the study during pressured times.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.