Serum-soluble ST2 as a novel biomarker reflecting inflammatory status and illness severity in patients with COVID-19

The outbreak of SARS-CoV-2, that first emerged in Wuhan in December 2019, spread rapidly throughout China in the first three months of 2020, leading to a serious global epidemic of COVID-19 [1–3]. The coronavirus infection can induce not only mild-to-severe respiratory diseases, but also inflammation and dysfunction of internal organs that may cause death [4]. Initial diagnosis of COVID-19 was based on positive viral nucleic acid test combined with chest CT scan [5]. These two indexes can reflect viral load and pneumonia severity, respectively. However, an index of monitoring disease progress, inflammatory state and pathological changes in other organs is lacking. For clinical practitioners, laboratory parameters that could be used as a fast and convenient alert to COVID-19 severity and inflammation were urgently required.

Suppression of tumorigenicity-2 (ST2), a member of the IL-1R/Toll-like receptor (TLR) superfamily, exists in three isoforms, full-length transmembrane form (ST2L), membrane bound variant (ST2V) and secreted soluble form (soluble ST2, sST2) [6,7]. The only known ligand of ST2 is interleukin-33 (IL-33), a member of the interleukin-1 family widely expressed in epithelial cells, endothelial cells and fibroblasts [8,9]. Previous researches have reported that IL-33/ST2 axis is involved in the inflammatory responses to several viral infections, including bronchiolitis, influenza, herpes simplex virus, coxsackie B virus, lymphocytic choriomeningitis virus and metapneumovirus [10–13]. During different infections, the deleterious or protective role of IL-33/ST2 axis depends on the type and invasiveness of infectious agent, the involved organs, whether the infection is acute or chronic, the host immune compartment and microenvironments [14]. Clinical studies have found higher sST2 levels with no difference in IL-33 levels in patients with hepatitis B virus infection, and the increased sST2 levels were found to correlate with disease severity and predicted poor survival [15]. In sera of patients with HIV infection, high levels of sST2 and low levels of IL-33 have been detected [16]. Moreover, it has been reported that persistently high sST2 concentrations were associated with severity of viral acute lower respiratory infection [17]. As an important biomarker of severe forms of dengue virus (DENV) infection, serial changes of sST2 levels may be a more reliable predictor for dengue fatality [18]. For the novel SARS-CoV-2 infection in COVID-19 patients, the role and change of sST2 and IL-33 is still unknown.

The study aimed to investigate the levels of serum sST2 and IL-33 in patients with COVID-19, and evaluate the association between the IL-33/ST2 axis, along with other common serological markers in clinic (C-reactive protein, serum amyloid A, interleukin-6, procalcitonin), and the illness severity.

Patients & methods

Study population

Eighty patients were enrolled with COVID-19, which were confirmed by detection of SARS-CoV-2 nucleic acid in throat swab samples using a novel coronavirus PCR fluorescence diagnostic kit (Daan Gene Co., Ltd of Sun Yat-Sen University, Guangzhou, China). All patients were initially admitted to Zhongnan Hospital of Wuhan University from 15 January to 25 February 2020. Meanwhile, we tested 38 healthy blood donors with exclusion of SARS-CoV-2 and other potential viral infections as the control subjects in this study. The study was approved by the ethics committee of Zhongnan Hospital of Wuhan University (no. 2019125). Written informed consent was obtained from all.

Data collection

A general medical history was collected, including age, sex, medical history, symptoms, severity assessment on admission, standard laboratory tests and chest CT findings. Among all COVID-19 patients, 36 patients were classified as mild cases (i.e., mild clinical symptoms with imaging feature of pneumonia), 41 patients as severe cases (i.e., dyspnea, respiratory frequency ≥30/min, blood oxygen saturation ≤93%, partial pressure of arterial oxygen to fraction of inspired oxygen ratio <300 and/or lung infiltrates >50% within 24 to 48 h) consisting of 16 nonsurviving and 25 surviving patients, and three patients as asymptomatic cases (diagnosis by positive viral nucleic acid test result but lacking typical symptoms including fever, dry cough and fatigue). Patients were discharged, if they met all of the following three criteria from Chinese Center for Disease Control guidelines: with normal temperatures for more than 3 consecutive days; remission of clinical symptoms; and twice negative SARS-CoV-2 nucleic acid results performed in every other day.

Laboratory tests

All samples were processed at the Department of Laboratory Medicine of Zhongnan Hospital of Wuhan University. All patients were tested for SARS-CoV-2 nucleic acid by use of quantitative real-time PCR on samples from the respiratory tract.

Blood samples were collected from each of the 80 fasting patients with COVID-19 on admission before initial treatment as well as from the 38 fasting control subjects. After standing for 30 min, the blood samples were centrifuged for 10 min at 3800 r.p.m. The sera were separated and preserved at -80°C for further analysis. Serum levels of human sST2 and IL-33 were measured according to the ELISA kits (Westang Biotech Co., Shanghai, China). In a previous study, demonstrated the basic analytical performances of human sST2 and IL-33 assays [19].

The common inflammatory parameters in clinic, C-reactive protein (CRP; Sekisui Medical Co., Tokyo, Japan), serum amyloid A (SAA; Purebio Biotech Co., Ningbo, China), serum urea and creatinine were tested on an Olympus 5800 analyzer (Beckman Coulter, CA, USA) using the latex-enhanced immunoturbidimetric assay. Interleukin-6 (IL-6) was detected using the automatic electrochemiluminescence immunoassay system (Cobas e601, Roche, Basel, Switzerland), and procalcitonin (PCT) was tested on an automatic immunoassay analyzer (VIDAS, Biomerieux, Marcy L'Etoile, France) using the enzyme-linked fluorescence analysis. Hypersensitive troponin I (hs-cTnI) was tested on an automatic chemiluminescence immunoassay analyzer (Abbott ARCHITECT i4000, Abbott Laboratories, IL, USA), and N-terminal pro-B-type natriuretic peptide (NT-proBNP) was analyzed on a VITROS 5600 analyzer (Johnson & Johnson, NJ, USA) using chemiluminescence immunoassay.

We also measured lymphocyte subsets in samples of EDTA anticoagulated peripheral blood from patient with COVID-19 on admission using multiple-color flow cytometry. The cells were analyzed on a BD FACS Canto II flow cytometry system (BD Biosciences).

Statistical analysis

All data were represented as mean ± SEM in normally distributed variables or medians (interquartile interval) for non-normally distributed variables. A nonparametric comparative test for continuous data was used to compare variables between groups. Chi-square test was used to compare the frequencies. Pearson analysis was used for the correlation analysis. Multivariate analysis was used to identify independent predictors of inflammation in COVID-19. The receiver operating characteristic (ROC) curve was carried out for measurement of the sensitivity and specificity of serum sST2 at different cut-off values. Bootstrap test was used to compare two correlated ROC curves. Two-way analysis of variance followed by a Dunnett or Tukey post hoc test was used to determine the differences between more than two groups. All statistical analyses were performed by GraphPad Prism 8.0 (GraphPad Software). p < 0.05 was considered statistically significant.

Results

Clinical Characteristics of mild & severe COVID-19 patients

The baseline characteristics of 36 mild cases and 41 severe cases with COVID-19 are shown in Table 1. The gender distribution did not differ significantly among the control group, mild and severe COVID-19 cases. However, the median age in severe COVID-19 cases was older than that in the mild cases and control group. Compared with the control subjects and mild cases, hypertension was the more prevalent comorbidity in the severe COVID-19 patients. Total lymphocytes, CD3⁺CD4⁺ T cells and CD3⁺CD8⁺ T cells decreased in COVID-19 patients, and severe cases had a lower level than mild cases. Myocardial injury biomarkers hs-cTnI and NT-proBNP were significantly higher in severe COVID-19 patients than those in the control subjects and mild cases. For the common inflammatory indicators, elevated levels of serum CRP and SAA were found in COVID-19 patients, and serum CRP level in severe cases was higher than that in mild cases. Compared with the control group and the mild cases, the IL-6 and PCT levels were increased in severe cases. Nevertheless, no obvious significance in the levels of IL-6 and PCT was observed between the control group and the mild COVID-19 patients.

Serum sST2 expression was elevated in patients with COVID-19

We first tested the serum levels of sST2 and IL-33 in mild and severe COVID-19 patients. Our results showed that the median level of serum sST2 in control subjects was 147.1 pg/ml (interquartile interval, 51.4–321.6). The level of serum sST2 was upregulated in mild and severe cases compared with the control subjects, and the serum sST2 value in severe cases was higher than that in mild cases (Figure 1A). Previous studies have reported the possible prognostic value of sST2 in patients with hypertension, Type 2 diabetes mellitus or heart diseases including coronary heart disease, myocardial infarction and heart failure [20–23]. We divided the mild and severe patients into cases with or without the above comorbidities to investigate their effects on serum sST2 levels. Nevertheless, there is no obvious difference in the serum sST2 level between mild or severe cases with or without the above comorbidities (Figure 1B). Moreover, no significant difference in the level of serum IL-33 was found among control subjects, mild and severe cases (Figure 1C).

We also analyzed the serum samples from three asymptomatic cases infected by SARS-CoV-2 without underlying diseases, and their laboratory findings are shown in Table 2. The levels of serum sST2 were significantly elevated in all three asymptomatic cases. Case 3 had a high level of serum IL-33, but low levels represented in Case 1 and Case 2. The levels of CRP, SAA, IL-6 and PCT did not change obviously.

Serum sST2 positively correlated with CRP concentration in patients with COVID-19

Correlation coefficients were calculated with the purpose of elucidating the association between the inflammatory status and the serum sST2 level in COVID-19 patients. In clinic, CRP, SAA, IL-6 and PCT were the common inflammatory parameters. Interestingly, we found that there was a significant positive correlation between the serum sST2 and CRP level in patients with COVID-19, but no significant correlation with SAA and IL-6 (Figure 2A–C). In COVID-19 patients with elevated PCT (>0.05 ng/ml), there was also a positive correlation between serum sST2 and PCT level (Figure 2D). However, for patients with normal PCT (<0.05 ng/ml), the correlation between sST2 and PCT cannot be compared.

ROC curves were constructed to evaluate the predictive value of inflammatory biomarkers in patients with COVID-19 on admission (Figure 3). Table 3 showed that the area under curve (AUC) of sST2, CRP, SAA and IL-6 were 0.9896, 0.7963, 0.8104 and 0.7387, respectively. The cut-off values were 10 mg/l, 10 mg/l and 7 pg/ml for CRP, SAA and IL-6 based on the reference range used in clinic. For serum sST2, the optimal cut-off value of 356.0 pg/ml yielded good sensitivity at 100.0% and specificity at 97.4%. The AUC for the integrated indicator was 0.8067, and bootstrap testing indicated a higher predictive accuracy of serum sST2 than the integrated indicator. Therefore, higher expression of serum sST2 may be a potential marker to reflect the inflammatory status in patients with COVID-19.

Serum sST2 negatively correlated with CD4⁺ & CD8⁺ T cells in patients with COVID-19

Recent research has found that total lymphocytes are decreased in COVID-19 patients, and lymphocyte subset (CD4⁺ and CD8⁺ T cell) counts may reflect disease severity and associate with inflammatory status [24–26]. Here, we also analyze the relationship of serum sST2 to the lymphocyte subsets. Our results showed that CD3⁺CD4⁺ and CD3⁺CD8⁺ lymphocyte absolute counts were negatively correlated with the levels of serum sST2 in patients with COVID-19, suggesting that the elevated serum sST2 may contribute to the dysfunction of T cells in COVID-19 progression (Figure 4A & B).

Course of serum sST2 concentrations & duration of hospital stay

To further explore the correlation between serum sST2 levels and the disease severity of SARS-CoV-2 infection, serial sST2 levels were tested in mild and severe patients with COVID-19 from on admission to the time of discharge or death. The average timeline of disease course was shown in Figure 5. On admission, the mean levels of serum sST2 in severe cases were increased, then elevated rapidly to a peak during disease progression. For nonsurviving severe cases (n = 16), the serum sST2 persisted at a high level until death. Compared with the severe cases, the mean levels of serum sST2 in the mild cases persisted at a relatively high level during the progression. To the time of discharge, the levels of serum sST2 tended to be decreased in mild and surviving severe cases (Figure 5). The results showed that persistent high level of serum sST2 may indicate more severe disease status of SARS-CoV-2 infection and worse outcomes.

A total of 16 cases (39%) with severe COVID-19 did not survive. The median age in severe COVID-19 cases was older with a more prevalent complication of hypertension, therefore, we further analyzed whether elevated serum sST2, age and hypertension were independent risk factors for death in COVID-19 patients. Multivariate logistic analysis showed that increased sST2 (OR: 5.876, 95% CI: 2.737–9.211; p = 0.003), and age (OR: 3.124, 95% CI: 1.873–6.598; p < 0.001) were the risk factors of death in COVID-19 patients.

Discussion

To our knowledge, this is the first study to evaluate serum sST2 levels in patients with COVID-19 during hospitalization along with other inflammatory cytokines and the association with illness severity. In this study, we found that among the traditional inflammatory markers used in clinic, serum CRP seems to be superior to SAA, IL-6 and PCT in reflecting the inflammatory status in the early stage of COVID-19 patients, which is consistent with recent COVID-19 reports [27,28]. Several studies have shown changes in SAA rather than significant CRP changes in some cases of SARS-CoV-2 infection, however, it should be noted that SAA is responsive to both viral and bacterial infections compared with CRP [29–31]. Additionally, clinicians may use IL-6 to identify severity earlier, but it is somewhat difficult to ascertain what level of IL-6 corresponds to what negative outcome due to the varying outcomes [32].

ST2 receptor and its ligand IL-33 has been reported to be involved in the inflammatory responses to several viral infections, but their role in coronavirus infections including SARS, MERS-CoV and the pandemic SARS-CoV-2 has been unclear [14]. Our results showed that the high level of serum sST2 occurred in COVID-19 patients and its level was positively correlated with the level of serum CRP, with exclusion of the influences from the cardiovascular complications, which means that the elevated serum sST2 was caused by the SARS-CoV-2 infection in COVID-19. According to the ROC curve analysis, the AUC in serum sST2 was 0.9896, which was fairly close to 1, indicating a better diagnostic value than serum CRP in reflecting the inflammatory status after SARS-CoV-2 infection. Moreover, in this pandemic of COVID-19, early detection and monitoring of asymptomatic cases has proved difficult [33]. Although few asymptomatic cases enrolled in the study, high levels of serum sST2 can be detected in all three cases without obviously elevated CRP, SAA, IL-6 and PCT levels. Under the condition of SARS-CoV-2 infection, serum sST2 as a predictive marker could be used in combination with clinical algorithms. Thus, the elevated serum sST2 may be a promising sensitive biomarker for reflecting the inflammatory status during early periods in patients with COVID-19.

IL-33 is a multifunctional immunomodulatory cytokine, which may regulate T-cell immune responses in different disease states or experimental conditions via its membrane receptor ST2L [8,34]. Conversely, sST2 acts as a decoy receptor by binding IL-33 to prevent its signaling through ST2L [35]. Here, we also tested the level of serum IL-33 in patients with COVID-19, and found that the median levels of serum IL-33 were at a very low level in control subjects and COVID-19 patients without significant difference, suggesting that IL-33 function may be inhibited or may not be activated after SARS-CoV-2 infection.

Recently, it has been reported that COVID-19 may damage lymphocytes, especially T lymphocytes, and that the immune system is impaired during the period of disease [24]. Cytotoxic immunity (particularly CD8⁺ T lymphocytes) was involved in antiviral processes and may indicate the infection severity [24]. Similar to the findings on SARS, there was a decreased CD4⁺ and CD8⁺ T lymphocytes after SARS-CoV-2 infection [25,26,36]. The results showed that there was a negative correlation between the high levels of serum sST2 and the decreased counts of CD4⁺ and CD8⁺ T cells in COVID-19 patients. Thus, it was speculated that the upregulated sST2 may weaken the activity of IL-33, which may contribute to the dysfunction of T-cell immune response and rapid disease progression. High sST2 levels, a decrease of CD4⁺ and CD8⁺ T cells, and low IL-33 levels could represent surrogate markers of T-cell immunosuppression during SARS-CoV-2 infection in human.

Similar to CRP, the level of serum sST2 in severe COVID-19 patients was higher than that in mild cases. To further verify whether the serum sST2 may reflect disease severity, we monitored the serum sST2 from on admission to discharge. The average timeline of disease course was longer in severe cases than that in mild cases. In mild cases, serum sST2 was maintained at a high level alone with disease progression, then decreased by the time of discharge. Conversely, the serum sST2 in severe cases showed a higher level during the progression period compared with the level on admission. In particular, the serum sST2 levels in severe COVID-19 patients who did not survive elevated continuously until death. Multivariate analysis also showed serum sST2 may be an independent risk factor of death in COVID-19 patients. These results revealed that high serum sST2 level may be a helpful predictor for the disease severity. One of the most dreaded thoracic complications in patients with COVID-19 is pulmonary fibrosis, which is associated with an increased risk of morbidity and mortality [37]. sST2 has been suggested to be a potential biomarker of fibrosis [38], so it was speculated that patients with the persistently elevated sST2 may have a higher risk of developing pulmonary fibrosis, which merits further elucidation.

There were several limitations in the study. First, it was a retrospective, single-center and small sample study of COVID-19 patients admitted to hospital. Additional larger sizes of COVID-19 patients from multiple centers are needed to validate our findings. A high level of serum IL-33 was found in one asymptomatic case, which merits further elucidation. However, not enough asymptomatic cases were enrolled to this COVID-19 study. Second, the sST2 levels in COVID-19 cases were not tested prior to their viral infections and during the incubation period and symptom onset, which may contribute to recording the dynamic changes of serum sST2 over the entire disease course. Although the baseline level in serum sST2 in COVID-19 patients with cardiovascular complications was unclear, the decreased tendency by the time of discharge in these cases suggested that the elevated serum sST2 level was induced by the SARS-CoV-2 infection, but not related to the corresponding complications. Furthermore, it should be noted that the cut-off value of serum sST2 works for fasting samples using ELISA method, and other factors that may impact on the cut-off values of sST2 need further investigation.

Conclusion

Taken together, serum sST2 levels were higher in patients with COVID-19 than the control subjects, which has not been previously reported. Serum sST2 was positively correlated with CRP, and negatively correlated with CD4⁺ and CD8⁺ T lymphocytes. Combined with the decreased T-cell subsets, high sST2 level may reflect the dysfunction of T-cell immune response and infection severity in COVID-19 patients.

Future perspective

As a sensitive inflammatory and prospective marker, high sST2 level combined with the decreased T lymphocytes may reflect the dysregulated T-cell immune response and infection severity in patients with SARS-CoV-2 infection. Serum sST2 test is helpful in the early screening of inflammatory status and critical illness of COVID-19.