Abstract
The COIVD-19 global pandemic is far from ending. There is an urgent need to identify applicable biomarkers for predicting the outcome of COVID-19. Growing evidences have revealed that SARS-CoV-2 specific antibodies remain elevated with disease progression and severity in COIVD-19 patients. We assumed that antibodies may serve as biomarkers for predicting disease outcome. By taking advantage of a newly developed SARS-CoV-2 proteome microarray, we surveyed IgM/IgG responses against 20 SARS-CoV-2 proteins in 1,034 hospitalized COVID-19 patients on admission, who were followed till 66 days. The microarray results were correlated with clinical information, laboratory test results and patient outcomes. Cox proportional hazards model was used to explore the association between SARS-CoV-2 specific antibodies and COVID-19 mortality. We found that high level of IgM against ORF7b at the time of hospitalization is an independent predictor of patient survival (p trend = 0.002), while levels of IgG responses to 6 non-structural proteins and 1 accessory protein, i. e., NSP4, NSP7, NSP9, NSP10, RdRp (NSP12), NSP14, and ORF3b, possess significant predictive power for patient death, even after further adjustments for demographics, comorbidities, and common laboratory markers for disease severity (all with p trend < 0.05). Spline regression analysis indicated that the correlation between ORF7b IgM, NSP9 IgG, and NSP10 IgG and risk of COVID-19 mortality is linear (p = 0.0013, 0.0073 and 0.0003, respectively). Their AUCs for predictions, determined by computational cross-validations (validation1), were 0.74 (cut-off = 7.59), 0.66 (cut-off = 9.13), and 0.68 (cut-off = 6.29), respectively. Further validations were conducted in the second and third serial samples of these cases (validation2A, n = 633, validation2B, n = 382), with high accuracy of prediction for outcome. These findings have important implications for improving clinical management, and especially for developing medical interventions and vaccines.
Highlights
IgM/IgG responses of 1,034 patients upon admission against 20 SARS-CoV-2 proteins were analyzed.
High level of IgM against ORF7b at the time of hospitalization is an independent predictor of patient survival.
IgG responses to NSP9 and NSP10 possess significant predictive power for patient death.
Introduction
The coronavirus disease 2019 (COVID-19), the emerging infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in December 2019, has quickly become the greatest crisis of global public health and economic development in our times [1]. As of October 11, 2020, there has been 36.75 million confirmed cases and 1.06 million patients have died from SARS-CoV-2 infection worldwide [2]. The crude mortality rate of COVID-19 was approximately 2.9% as estimated by the WHO [2], and there are no highly effective therapeutics or vaccines currently available for COVID-19 [3-5], highlighting the significance to understand the role of immunity in the progression and outcome of COVID-19 patients for improving clinical management and developing effective interventions and vaccines.
Similar to SARS-CoV and MERS-CoV, SARS-CoV-2 belongs to the betacoronavirus genus and its genome encodes four major structural proteins spike (S), envelope (E), membrane (M), and nucleocapsid (N), and 15 non-structural proteins (Nsp1-10 and Nsp12-16) and 9 accessory proteins [6]. Among these, the S protein, consisted of N-terminal S1 peptide with an important receptor binding domain (RBD) region and C-terminal S2 fragment, plays an essential role in viral attachment, and fusion, entry into the target cells which express the viral receptor angiotensin-converting enzyme 2 (ACE2) [7]. There is rapidly growing serological testing evidence that IgM, IgG, and IgA antibodies against spike (S) or nucleocapsid (N) proteins of SARS-CoV-2 evolve rapidly in the serum of both asympomatic and symptomatic COVID-19 infections within one week after infection or onset of symptoms [8-11]. More importantly, these antibodies remain elevated with disease progression and severity in symptomatic COIVD-19 patients [12]. Therefore, anti-SARS-CoV-2 specific antibodies may involve in the pathogenesis and affect the prognosis and outcome of COVID-19. However, little is known about humoral immune responses to other structural and non-structural proteins of SARS-CoV-2 during disease progression and outcome.
In this study, we assumed levels of anti-SARS-CoV-2 IgM and IgG antibodies might help predict disease prognosis and outcome in patients with COVID-19. To enable the global understanding of SARS-CoV-2 specific IgM/IgG responses, we constructed a proteome microarray with 20 out of the 28 predicted proteins of SARS-CoV-2 [13]. Clinical serum specimens were further analyzed on the SARS-CoV-2 proteome microarray, which can provide a high-throughput assay for 12 samples on each microarray and a rapid turnaround time of assay results (within 5 h of sample collection).
1,034 patients hospitalized for confirmed COVID-19 at Tongji hospital from the day of hospitalization to the day of discharge or death were enrolled in this study and were classified into two groups, namely survivors and nonsurvivors based on the known clinical outcome. Serum IgM and IgG profiles for 1,034 patients with COVID-19 on admission were probed using the SARS-CoV-2 proteome microarray, which were further correlated with laboratory biomarkers of disease severity and comorbidities, and with death. We found that elevated ORF7b specific IgM serum levels at presentation is a useful predictor of survival, while high levels of IgG responses to most of non-structural proteins, especially NSP9 and NSP10 are powerful predictions of COVID-19 death. Our results indicate that the set of anti-SARS-CoV-2 antibody signatures are independent from other biomarkers of laboratory and clinical severity factors, which could be used to guide clinical management, vaccine developments, and interventional studies.
Materials
Patient information and data source
1,056 confirmed COVID-19 patients were recruited from Tongji Hospital, Wuhan, China, between 17 February 2020 and 28 April 2020. COVID-19 was diagnosed based on positive SARS-CoV-2 nucleic acid test from respiratory tract specimens or based on clinical diagnosis with clinical symptoms and imaging features of pneumonia on chest computed tomographic (CT) according to the fifth version of COVID-19 diagnostic and treatment guideline published by the National Health Commission of China (NHCC) [14]. Demographic information, medical history, comorbidities, signs and symptoms, chest CT, laboratory findings on admission, and clinical outcomes were collected from electronic medical records. Among these, laboratory biomarkers related with disease severity factors such as the blood routine (leucocytes, lymphocytes, platelets, and neutrophils), liver and kidney functions (aspartate aminotransferase, alanine aminotransferase, lactate dehydrogenase, and creatinine), coagulation function (D-dimer) and infection markers (C-reactive protein, procalcitonin) were performed by automated analyzers according to the manufacturers’ instructions. The level of IL-2R in serum was measured by automatic procedure of a solid-phase two-site chemiluminescent immunometric assay via IMMULITE 1000 Analyzer (Siemens). Serum IL-6 was measured by electro-chemiluminescence method (Roche Diagnostics). Serum specimens were collected from each patient on admission and were stored at -80 °C until use. Serum detection based on proteome microarray and data analysis were performed during April 2020 to July 2020. After excluding individuals whose 23 anti-SARS-CoV-2 antibody indicators were missing more than three, a total of 1,034 eligible participants (524 females and 510 males) with available data from serum proteome microarray and clinical outcomes were enrolled for the final analysis. Among 1,034 eligible participants, some of whom had serial serum samples and collected for a total of 2,973 samples.
Ethical approval
The study was approved by the Ethical Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China (IRB ID:TJ-C20200128).
Protein microarray fabrication
The microarray used for serum IgM and IgG profiling was prepared as described previously [11, 13]. 20 proteins of SARS-CoV-2 with indicated concentrations, along with the negative (GST, Biotin-control, and eGFP) and positive controls (Human IgG, Human IgM, and ACE2-Fc), were printed in quadruplicate on PATH substrate slide (Grace Bio-Labs, Oregon, USA) to generate identical arrays in a 2×7 subarray format using Super Marathon printer (Arrayjet, UK). The prepared protein microarrays were incubated in blocking buffer (3% BSA in 1×PBS buffer with 0.1% Tween 20) for 3 h, and then stored at -80 °C until use.
Microarray-based serum analysis
The protein microarrays stored at -80 °C were warmed to room temperature before detection and were performed to probe all available seral samples. A 14-chamber rubber gasket was mounted onto each slide to create individual chambers for the 14 identical subarrays. Serum samples were diluted 1:200 in PBS containing 0.1% Tween 20 and a total of 200 µL of diluted serum or buffer only (negative controls) was incubated with each subarray for 2h at 4o C. The arrays were washed with 1×PBST and bound antibodies were detected by incubating with Cy3-conjugated goat anti-human IgG and Alexa Fluor 647-conjugated donkey anti-human IgM (Jackson ImmunoResearch, PA, USA), which were diluted 1: 1,000 in 1×PBST, and incubated at room temperature for 1 h. The microarrays were then washed with 1×PBST and dried by centrifugation at room temperature and scanned by LuxScan 10K-A (CapitalBio Corporation, Beijing, China) with the parameters set as 95% laser power/ PMT 550 and 95% laser power/ PMT 480 for IgM and IgG, respectively. Data of fluorescent intensity (FI) from each microarray was extracted by GenePix Pro 6.0 software (Molecular Devices, CA, USA). The result of FI for each serum response to each protein was defined as the median of the foreground subtracted by the median of background for each spot and then averaged the triplicate spots for each protein. The result of the protein-specific antibody in the serum was expressed as log2(FI). IgG and IgM data were analyzed separately.
Statistical analysis
Shapiro-Wilk test was used to test data normality. Two-tailed t-test was conducted to test difference in means between survivor and nonsurvivor groups, Mann-Whitney U test was performed to test difference in skewed parameters. Chi-square tests or Fisher’s exact test, when appropriate, was used for categorical variables. Cox proportional-hazards model was performed to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) of COVID-19 mortality for individual levels of virus-specific IgM and IgG responses categorized into tertiles according to distributions. The lowest tertiles were assigned to be the reference groups. Age and sex were included in Model 1. In Model 2, we further adjusted hypertension (yes/no), diabetes (yes/no), lymphopenia (<1.1, ≥1.1, ×10^9/L), increased alanine aminotransferase (<40, ≥41, U/L), and increased lactate dehydrogenase (<214, ≥214, U/L). Linear trend p-values were calculated by modeling the median value of each metal tertiles as a continuous variable in the adjusted models. Spearman’s rank correlation analysis was performed to explore the correlations between virus-specific IgM/IgG responses and laboratory results in the study population. The principal component analysis (PCA) based on the 20 proteins of SARS-CoV-2 specific IgM/IgG responses was used to optimize the type of data and extract principal components (PCs). Proteins of SARS-CoV-2 specific IgM /G responses with factor loadings over 0.7 on a particular PC were regarded as main contributors of it. Each PC was modeled into the Cox proportional-hazards models as tertiles to evaluate the association with SARS-CoV-2 specific IgG responses and COVID-19 mortality.
In addition, we also conducted sensitivity analyses by exclusion of patients who died within the initial 7 days after hospitalization to avoid reverse causality. The associations of viral specific IgG and IgM responses with the risk of COVID-19 mortality were also evaluated using restricted cubic splines, with 3 konts defined at the 5th, 50th, and 90th percentiles of its distribution; the reference value (HR = 1.00) set at the 10th percentile. The measured level replaced with the mean level ± 3SD for all observations with measured concentrations above this value. The results of antibodies were classified as two groups of the high levels (≥ median) and low levels (< median) based on the medians of IgM and IgG responses to each protein of all involved patients and further correlated these results with on day 66 mortality of COVID-19 by Kaplan-Meier survival curve and log-rank test. The receiver operating characteristic curve was conducted for the prediction of COVID-19 survival and death, and 1,000 times computational cross-validations were conducted. For each cross-validation procedure, 477 survivors and 39 nonsurvivors were randomly selected as the training set. The rest of the samples were treated as the testing set (478 survivors and 40 nonsurvivors). Further validation was conducted using the second and third serial samples after hospitalization (validation2A, n = 633, validation2B, n = 382). Loess regression was used to establish the kinetics of SARS-CoV-2 specific antibodies. Cluster analysis was performed with pheatmap package of R. SAS (version 9.4), R (version 4.0.0), and SPSS (version 23.0) were used to conduct statistical analyses when applicably used. Two-sided statistical tests were considered to be significant at p values below 0.05.
Results
Characteristics of the study population
1,034 participants, having available serum microarray results and consisting of 955 survivors and 79 nonsurvivors, were enrolled in this study. Baseline characteristics of participated patients based on electronic medical records were analyzed as Table 1. The median age of all enrolled patients was 63 years old (IQR, 51-71). The median intervals from onset of symptoms to hospital admission, from onset of symptoms to recovery, and from onset of symptoms to death were 13 days (IQR, 8-21), 41 days (IQR, 33-52), and 32 days (IQR, 25-39), respectively. The median length of all COVID-19 patients’ hospital stay was 24 days (IQR, 15-35). 37% patients with COVID-19 had hypertension and 18.5% with diabetes. 30.7% patients had lymphopenia, and increased levels of lactate dehydrogenase and alanine aminotransferase were detected in 43% and 25.4% patients, respectively. Consistent with previous reports [15, 16], nonsurvivors were more likely to be male and older than survivors (p < 0.001). Our study also demonstrated that higher proportion of abnormal laboratory results and shorter hospitalization time were obtained in nonsurvivors than those of survivors (p < 0.001).
The magnitude of antibody responses correlates with disease outcome
To establish the association of anti-SARS-CoV-2 IgM and IgG antibodies with COVID-19 survival and death, serum collected from each involved patients on admission was used for microarray-based serum analysis. Based on the FI extracted from proteome microarray for each serum of 1034 patients, we first presented overall visualizations and quantitative data of IgM (Figure S1 and Table S1) and IgG (Figure S2 and Table S2) profiles against 20 proteins of SARS-CoV-2, respectively. We demonstrated that higher levels of both IgM and IgG responses against N, ORF3a, and ORF7b proteins were induced in survivors than those of nonsurvivors, apart from ORF6 specific IgM antibody (p < 0.05, Figure 1a, Table S1 and Table S2). On the contrary, nonsurvivors elicited higher levels of NSP10 specific IgM antibody (p < 0.05, Table S1) and IgG responses against E, NSP1, NSP2, NSP4, NSP5, NSP7, NSP8, NSP9, NSP10, RNA-dependent RNA polymerase (RdRp or NSP12), NSP14, NSP15, NSP16, ORF3b, and ORF9b proteins than survivors (p < 0.05, Figure 2a and Table S1). The levels of ORF7b IgM, NSP9 IgG, and NSP10 IgG fluctuated with the days after symptoms onset, but they were not obvious (Figure S4). Our results strongly indicate that the magnitude of IgM or IgG responses against most of structural and non-structural proteins of SARS-CoV-2 might involve in the prognosis and outcome of COVID-19.
Comparison of SARS-CoV-2 specific IgM responses ([log2(FI)]) between survivors and nonsurivivors
Comparison of SARS-CoV-2 specific IgG responses ([log2(FI)]) between survivors and nonsurivivors
Heatmap of IgM antibody responses of 955 survivors and 79 nonsurvivors. Each square indicates the IgM antibody response against the protein (column) in the serum (row). The proteins were arranged with clustering. FI: Fluorescence Intensity.
Heatmap of IgG antibody responses of 955 survivors and 79 nonsurvivors. Each square indicates the IgM antibody response against the protein (column) in the serum (row). The proteins were arranged with clustering. FI: Fluorescence Intensity.
(a) Comparison of the levels of IgM response to ORF7b between 955 survivors and 79 nonsurvivors. The boxplots show medians (middle line) and the third and the first quartiles (boxes), while the tentacles show 97.5 and 2.5 percentiles of the upper and lower parts of the box. (b) Kaplan-Meier survival curves of patients with different levels of IgM antibody against ORF7b. Based on the median level of ORF7b specific IgM responses, patients were classified as both high and low level groups. (c) The restricted cubic spline for the association between ORF7b IgM and risk of COVID-19 mortality. The lines represent adjusted hazard ratios (HRs) based on restricted cubic splines for the levels of ORF7b IgM in Cox regression model. Knots were placed at the 5th, 50th, and 95th percentiles of the distribution of ORF7b specific IgM levels, and the reference value was set at the 10th percentile. Age, sex, diabetes, hypertension, lymphopenia, increased alanine aminotransferase, and increased lactate dehydrogenase were used as adjustment factors.
(a) Comparison of the levels of IgG response to non-structural proteins from the proteome of SARS-CoV-2 between 955 survivors and 79 nonsurvivors. The boxplots show medians (middle line) and the third and the first quartiles (boxes), while the tentacles show 97.5 and 2.5 percentiles of the upper and lower parts of the box. (b) Kaplan-Meier survival curves of patients with high and low levels of IgG antibodies to each protein. Based on the median level of IgG responses to each protein, patients were classified as both high and low level groups.
High levels of ORF7b IgM antibody predict disease survival
To assess the relationship of the magnitude of IgM antibodies with the mortality risk of COVID-19, the HRs (95% CIs) for the mortality risk of COVID-19 associated with the levels of IgM responses against different proteins of SARS-CoV-2 were categorized into tertiles (Table 2 and Table S3). We first analyzed the effects of age and gender on disease death as model 1. After adjusting for age and gender, we found that IgM responses to N, ORF3a, or ORF7b were significantly associated with COVID-19 mortality (all p trend < 0.05), while no significant association was observed among other protein-specific IgM responses and the death, respectively. Previous studies reported that comorbidities and laboratory biomarkers related with the function of important organs are also the risk factors resulting in the COVID-19 death [16, 17]. We further adjusted the association for hypertension, diabetes, lymphopenia, increased alanine aminotransferase and lactate dehydrogenase as shown in model 2. Interestingly, only the IgM response to ORF7b was significantly associated with the mortality risk of COVID-19 (T2 vs T1: HR = 0.86, 95% CI: 0.51-1.44; T3 vs T1: HR = 0.19, 95% CI: 0.07-0.55; p trend = 0.002, Table 2), even independently of the factor excluding patients who died within 7 days after admission (Table S4). Moreover, Kaplan-Meier survival curve also showed that COVID-19 patients with early detected high level of ORF7b specific IgM antibody (log2FI ≥ 7.5) on admission had lower risk of morality than the patients with low levels (log2FI < 7.5) during the following-up observation period of 66 days (p < 0.001, Figure 1b). The linear association between the levels of IgM response to ORF7b and the mortality risk of COVID-19 was further demonstrated by spline regression analysis (p = 0.0013, Figure 1c). Taken together, our results suggest that high levels of ORF7b IgM antibody upon admission are negatively correlated with the mortality risk of COVID-19.
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of SARS-CoV-2 specific IgM response
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of SARS-CoV-2 specific IgM response after exclusion of patients who died within the first 7 days
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of anti-SARS-CoV-2 specific IgM/IgG responses
High levels of IgG antibody against NSP7 or NSP9 predict disease death
To establish the associations between anti-SARS-CoV-2 IgG responses with risk of death, the relationship between the levels of IgG antibody against 20 proteins of SARS-CoV-2 with the mortality risk of COVID-19 was shown in Table 2 and Table S5, respectively. After adjusting for age and gender in model 1, we observed that the levels of IgG responses to N, NSP1, NSP4, NSP7, NSP8, NSP9, NSP10, RdRp (NSP12), NSP14, ORF3a, ORF3b, ORF7b, and ORF9b were significantly associated with the risk of COVID-19 mortality, respectively (all with p trend < 0.05). After further adjustment for potential confounders in model 2, the association of the multivariable adjusted HRs (95% CI) of COVID-19 mortality with these IgG responses remained statistically significant, except IgG responses to N, ORF3a and ORF7b (Table 2). After excluding patients who died within the first 7 days after admission, the associations between IgG responses to NSP4, NSP7, NSP9, NSP10, RdRp (NSP12), NSP14, and ORF3b and the mortality risk of COVID-19 remained statistically significant (Table S6). As shown in Figure 2, Kaplan-Meier survival curve also supported that COVID-19 patients with higher levels of specific IgG responses against NSP1 (log2FI ≥ 8.2), NSP4 (log2FI ≥ 7.9), NSP7 (log2FI ≥ 9.4,), NSP8 (log2FI ≥ 7.8), NSP9 (log2FI ≥ 8.7), NSP10 (log2FI ≥ 6.3), RdRp (log2FI ≥ 8.1), NSP14 (log2FI ≥ 7.4), ORF3b (log2FI ≥ 8.7), and ORF9b (log2FI ≥ 8.0) had higher risk of morality after admission, respectively (all p < 0.05). Most importantly, the linear association between high levels of IgG antibodies against only NSP9 or NSP10 with COVID-19 death was demonstrated by spline regression analysis with p values 0.0073 and 0.0003, respectively (Figure 3). Taken together, our results suggest that high levels of IgG responses to 6 non-structural proteins such as NSP4, NSP7, NSP9, NSP10, RdRp (NSP12), NSP14 and 1 accessory protein ORF3b are positively correlated with the mortality risk of COVID-19.
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of SARS-CoV-2 specific IgG response
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of anti-SARS-CoV-2 specific IgG response after exclusion of patients who died within the initial 7 days after hospitalization
The lines represent adjusted hazard ratios based on restricted cubic splines for the levels of NSP9 and NSP10 IgG antibodies in Cox regression model. Knots were placed at the 5th, 50th, and 95th percentiles of the distribution of NSP9 and NSP10 IgG antibodies, and the reference value was set at the 10th percentile. Age, sex, hypertension, diabetes, lymphopenia, increased alanine aminotransferase and lactate dehydrogenase were used as adjustment factors.
Principal component analyses support the predictive role of IgG antibody as disease death
To further establish the association among IgG responses to different proteins with the outcome of COVID-19, we further conducted principal component analyses (PCs) and screened hypothetical new variables that account for as much as possible of the variance, in order to reduce the dimension of data and the complexity of data with the least loss of original information. The HRs (95%CIs) for the COVID-19 mortality according to PCs tertiles are presented in Table 3. Four PCs with eigen values > 1 were extracted, accounting for 71.95% of the total variance. Of four PCs, we found that only PC1 had the statistical association with the COVID-19 mortality (p trend = 0.004, Table 3), whatever adjusting age and sex, or further for hypertension, diabetes, lymphopenia, increased alanine aminotransferase and lactate dehydrogenase. Interestingly, IgG responses to NSP1, NSP2, NSP4, NSP7, NSP8, NSP9, NSP10, RdRp (NSP12), NSP14, NSP15, NSP16, ORF3b, and ORF9b remained main contributors of PC1 (Table S7), in line with our above findings.
Factor loadings of 20 proteins of anti-SARS-CoV-2 specific IgG responses among the study participants
Hazard ratio (95%CI) for COVID-19 mortality according to tertiles of principal components of anti-SARS-CoV-2 specific IgG responses
IgG responses positively correlate with laboratory biomarker measurements related with severity factors
Previous studies established the associations between COVID-19 death with several laboratory biomarker measurements related with severity factors, such as lymphocyte count, procalcitonin, C-reactive protein, lactate dehydrogenase, D-dimer, IL-2R, and IL-6 [15-17]. Therefore, linear correlation among SARS-CoV-2 specific IgM/IgG responses and these biomarkers was further analyzed (Table S8). Interestingly, the biomarker of lymphocyte count was positively correlated with the ORF7b specific IgM antibody (rs=0.21, p < 0.01) but negatively correlated with IgG responses to NSP1, NSP4, NSP7, NSP8, NSP9, NSP10, RdRp (NSP12), NSP14, ORF3b, and ORF9b, respectively (all p < 0.01). Moreover, IgM response to ORF7b was negatively correlated with pro-inflammatory factors such as procalcitonin, C-reactive protein, lactate dehydrogenase, D-dimer, IL-2R, and IL-6, respectively (all p < 0.01). However, these pro-inflammatory factors except IL-2R and IL-6 were positively correlated with all of these IgG responses. The levels of IL-2R were also positively correlated with these IgG responses except NSP9 and ORF9b specific IgG antibodies, while NSP8, NSP10, RdRp (NSP12) and NSP14 specific IgG antibodies were positively correlated with IL-6, respectively.
Correlations between the levels of anti-SARS-CoV-2 specific IgM/IgG responses and other laboratory biomarkers related with severity factors
High prediction efficacy for clinical outcome using validation models
It is a common practice to validate “potential biomarker” by independent sample cohort. However, it is very difficult to collect more COVID-19 samples at this moment, because of very strict regulations of sample handling and very few COVID-19 patients in China. To assure the reliability of our finding, alternatively, we performed computational cross-validation based on the large sample cohort that we have already analyzed, by following a protocol that we have established previously [18]. ORF7b IgM, NSP9 IgG and NSP10 IgG were confirmed as three potential biomarkers for predicting clinical outcome (Figure 4a). Through computational cross validation (validation1), the AUC of ORF7b IgM for predicting COVID-19 survival was determined as 0.74 (cut-off = 7.59, Figure 4b). The AUCs of NSP9 and NSP10 for predicting COVID-19 death were 0.66 and 0.68, respectively (cut-off = 9.13 for NSP9 and 6.29 for NSP10, Figure 4b). Furthermore, we evaluated the prognosis efficacy of these three potential biomarkers using the samples collected at the second time point (n = 633, Validation2A) and the third time point (n = 382, Validation2B) after hospitalization from the 1,034 patients. In validation 2A and validation 2B, the accuracies of ORF7b IgM (Figure 4c-d), NSP9 IgG, and NSP10 IgG (Figure 4c-d) were 0.69 and 0.66, 0.77 and 0.76, 0.76 and 0.77, respectively. Besides, we depicted the dynamic changes of ORF7b IgM (Figure 5a), NSP9 IgG (Figure 5b), and NSP10 IgG (Figure 5c) for 1,034 COVID-19 patients with 2,973 consecutive samples. They all maintained relatively stabilized levels after symptoms onset.
(a) The workflow of validation. Firstly, three potential biomarkers were confirmed using 1,034 samples. Secondly, the three candidates were then computationally validated by 1,000 times random sampling. Finally, the accuracy of prognosis efficacy of the three biomarkers was evaluated in two other sample sets, i. e., 633 samples and 382 samples. (b) The prognosis efficacy determined by computational cross-validation. The receiver operating characteristic curve was conducted for the prediction of COVID-19 survival and death, and 1,000 times computational cross-validations were conducted. For each cross-validation procedure, 477 survivors and 39 non-survivors were randomly selected as the training set. The rest of the samples were treated as the testing set (478 survivors and 40 non-survivors). The average cutoff values were shown. (c) The prognosis efficacy in Validation 2A. The second set of samples (n = 633) were analyzed to evaluate the prognosis efficacy of ORF7b IgM, NSP9 IgG, and NSP10 IgG. The accuracy was calculated. (d) The prognosis efficacy in Validation 2B. The third sample set (n = 388) were analyzed to evaluate the prognosis efficacy of ORF7b IgM, NSP9 IgG, and NSP10 IgG. The accuracy was calculated.
Dynamic changes of ORF7b IgM (a), NSP9 IgG (b), and NSP10 IgG (c) for 1034 COVID-19 patients (2,973 samples) with the days after symptoms onset. The lines show the mean value expected from a Loess regression model, the ribbons indicate the 95% confidence interval.
Discussion
In this study, we aimed to understand the role of anti-SARS-CoV-2 IgM and IgG responses on COVID-19 disease prognosis and outcome. Our findings have important indications for better control of COVID-19 pandemic.
Firstly, we established a rapid and high-throughput assay platform based on proteome microarrays to measure IgM and IgG responses against 20 SARS-CoV-2 proteins in COVID-19 patients. After analyzing 1,034 hospitalized patients, we established that COVID-19 is associated with high levels of IgM and IgG responses to 11 non-structural proteins and 3 accessory proteins of SARS-CoV-2 at presentation. Importantly, our observations indicate that antibody patterns are predictive of COVID-19 survival and mortality, independently of demographics and comorbidities, but also of standard clinical biomarkers of disease severity. We found that OFR7b IgM response is independently the prognostic marker of survival, and IgG antibodies against 6 non-structural proteins NSP4, NSP7, NSP9, NSP10, RdRp (NSP12), NSP14, and 1 accessory protein ORF3b, especially NSP9 and NSP10 are predictors of death after adjusting for the demographic features and comorbidities. Early antibody measurements based on our established serum proteome microarray analysis as predictors of survival and death, therefore, raise the importance of using antibody levels for rapidly improving clinical management, treatment decisions and rational allocation of medical resources in short supply during the process of dealing with COVID-19 pandemic.
ORF7b is an accessory protein of SARS-CoV-2 with a length of 43 amino acids, which is also highly similar to the SARS-CoV ORF7b but absent from MERS-CoV [6]. Previous studies reported that the SARS-CoV ORF7b is not only an accessory protein but also a structural component of the virion, which is a viral attenuation factor during early phase of infection [19-21]. Our results implicate that anti-SARS-CoV-2 ORF7b specific antibody might play a protective role against COVID-19 infection and disease, further supported by the evidence that there was no significant association between levels of ORF7b IgG response and the risk of COVID-19 mortality. Therefore, ORF7b might be a promising target antigen for vaccine development. Unfortunately, high mutation rate was observed for current prevalent SARS-CoV-2 strains worldwide (Table S9). The effects of ORF7b and its mutations on immunization and host-virus interaction deserve further exploration. Although the function of each non-structural proteins of the SARS-CoV-2 is not yet clear, their protein sequences are highly similar to that of SARS-CoV. Most non-structural proteins always locate in the core of virion and play important roles in the pathogenesis. For example, RdRp, also called NSP12 of SARS-CoV, can catalyze the synthesis of viral RNA and plays an important role in the replication and transcription cycle of the virus [22, 23]. RdRp itself performs the polymerase reaction with limited efficiency, whereas NSP7 and NSP8 as co-factors can significantly stimulate its polymerase activity [22]. Previous studies based on cryogenic electron microscopy (cryo-EM) indicated that the viral polymerase RdRp-NSP7-NSP8 complex might be an excellent target for new therapeutics of SARS and COVID-19 [23, 24]. NSP1 of the SARS-CoV may promote viral gene expression and immune escape by affecting interferon-mediated signal transduction [25]. NSP4 is a multichannel membrane protein, which is an essential protein for viral replication [26]. NSP9 plays a role of dimeric ssRNA binding protein during viral replication [27, 28]. NSP10 interacts with NSP14 and regulates ribose-2’-O-MTase activities involved in mRNA capping [28-30]. Therefore, the relationship between these IgG responses and COVID-19 mortality indicates that IgG antibodies against these non-structural proteins might involve in the pathogenesis of SARS-CoV-2.
Even if not detected in preclinical and clinical trials, memory B cell responses to our reported non-structural proteins might be induced after a healthy individual vaccinated with inactivated or attenuated SARS-CoV-2 vaccine candidates. After exposure to or reinfection with SARS-CoV-2, IgG antibodies to these proteins might evolve in a higher, faster and stronger fashion in recipients of immunization. However, beneficial evidences from current animal models vaccinated with inactivated SARS-CoV-2 vaccine candidates against infection or reinfection [31, 32], attention should be paid to the risk of immunization with these kinds of vaccines to increase SARS-CoV-2 vulnerability, as demonstrated by a previous study that a double-inactivated SARS-CoV vaccine could elicit eosinophilic and immunoenhancing pathology, as well as poor protection, especially in aged animals upon challenge with virulent strains [33]. In addition, a newly reported COVID-19 case in USA with secondary infection with SARS-CoV-2 had more serious illness [34], also indicating the potential risk of the preexisting immunity. Although antibody-dependent enhancement (ADE) of SARS-CoV-2 infection has not been issued as yet [35-37], decay of antibodies was found to be related with the loss of Bcl-6-expressing T follicular helper cells and germinal centers influencing memory B cells in critically ill patients with COVID-19 [38], which might significantly decrease the risk of administration with convalescent plasma containing IgG antibodies to the above non-structural proteins to treat critically ill patients with COVID-19, especially when used our microarray to screen suitable donors of convalescent plasma before treatment.
In conclusion, we provided a novel application of SARS-CoV-2 proteome microarray to detect serum IgM/IgG responses for early predicting COVID-19 survival and death. Our results demonstrate that high level of IgM antibody against ORF7b at the time of hospitalization is an independent predictor of patient survival, while IgG responses to NSP9 and NSP10 possess significant predictive power for patient death. Our research might improve clinical management and guide the development of effective medical interventions and vaccines by enhancing the further understanding of pathogenesis of COVID-19.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Author Contributors
X-L.F., S-C.T., and F.W. performed experiments and designed the study. Q.L., C-Z.Y., and Y.L. performed the experiments. H-Y.H., Z-Y.S., and B.Z. collect specimens. D-Y.L., J-B.X., Z-Q.O., J-L.B., and Z-W.X. prepared the reagents. Y-D.Z., Z-J.Y., H.C., Y-X.Z., and X-S.L. analyzed the data. M-A.H, X-N.W., H-W.J., H-N.Z., H.Q., and S-J.G. took responsibility for the accuracy of the data analysis. X-L.F., Q.L., and C-Z.Y. wrote the manuscript with suggestions from other authors.
Conflict of Interest Disclosures
The authors declare no conflicts of interest.
Funding/Support
This work was supported by grants from the National Mega-Projects of Science Research for the 13th Five-year Plan of China (No. 2018ZX10302302002-001), the Natural Science Foundation of China (No. 81971909), and the Fundamental Research Funds for the Central Universities (HUST COVID-19 Rapid Response Call No. 2020kfyXGYJ040). This work was partially supported by National Key Research and Development Program of China Grant (No. 2016YFA0500600), National Natural Science Foundation of China (No. 31970130, 31670831, 31370813 and 31501054)
Acknowledgments
We thank Prof. H. Eric Xu (Shanghai Institute of Materia Medica) for providing RdRp protein. We also thank Healthcode Co., Ltd., Hangzhou Bioeast biotech Co., Ltd. and Vacure Biotechnology Co.,Ltd. for providing the proteins.
