Different immunological characteristics of asymptomatic and symptomatic COVID-19 patients without vaccination in the acute and convalescence stages

Study subjects

A total of 10 COVID-19 patients and four healthy controls (HCs) were recruited from the Fifth Hospital of Shijiazhuang from January 2021 to April 2021. All individuals were unvaccinated against SARS-CoV-2. The definitions and classifications of COVID-19 patients comply with the guidelines of the “World Health Organization, Chinese Diagnosis and Treatment Protocol for COVID-19 Patients (Tentative 8th Edition)” and Chinese Prevention and Control of COVID-19. The COVID-19 patients confirmed by positive SARS-CoV-2 nucleic acid test were divided into two groups: (1) symptomatic cases (four moderate cases and two severe cases), (2) asymptomatic cases (n = 4). Patients with moderate disease have symptoms such as fever, respiratory symptoms, etc., and imaging showed pneumonia. Patients who meet any of the following are classified as severe patients (1) Shortness of breath, RR ≥ 30 per min. At rest, oxygen saturation ≤93% when sucking air; Arterial partial pressure of oxygen (PaO₂)/oxygen absorption concentration (FiO₂) ≤300 mmHg (1 mmHg = 0.133 kPa). Asymptomatic cases are defined as those with positive for SARS-CoV-2 test and without any clinically recognizable or self-perception symptoms over the latest 14 days. HCs were defined as individuals without any clinical symptoms and had two consecutive negative SARS-CoV-2 nucleic acid tests (interval >24 h). Individuals who had other infections, tumors, severe hepatic and renal system diseases and previous major diseases, such as autoimmune diseases and metabolic diseases, were excluded.

A total of 24 peripheral blood mononuclear cell (PBMCs) samples were collected, including samples from 10 patients at both acute and convalescent time points and four samples from HCs. Blood samples were collected within 7 days of onset (acute) and 21 days after admission when they reached the standard of discharge (convalescence).

Patients in the acute stage was defined as individuals who tested positive for SARS-CoV-2 nucleic acid through quantitative real-time polymerase chain reaction, and were within 7 days of symptom onset. Among these, cases exhibiting no symptoms and those presenting symptoms were designated as asymptomatic_1 and symptomatic_1, respectively. On the other hand, patients in the convalescent stage were characterized by having two consecutive negative SARS-CoV-2 tests with an interval of more than 24 h once they met the discharge criteria. Additionally, all the collected samples were obtained 21 days after the onset of symptoms. The subgroups within the convalescent stage were labeled as asymptomatic_2 and symptomatic_2.

Demographic features, clinical and laboratory testing data were extracted from electronic medical records. Detailed information on the participants was shown in Table S1. This study was discussed and approved by the Ethics Committee of the Fifth Hospital of Shijiazhuang (IRB number: 2020008) and obtained the paper version of informed consent.

Preparation of PBMCs

Peripheral blood samples from all subjects were collected into blood collection tubes containing ethylenediaminetetraacetic acid (EDTA). PBMCs were isolated by density gradient centrifugation using human PBMC isolation buffer (produced by Serumwerk Bernburg AG for Alere Technologies AS, Oslo, Norway). The mononuclear cells were collected and washed twice with phosphate buffered saline (PBS). The cells were stored at liquid nitrogen until analysis.

Antibodies and antibody labeling for mass cytometry

A total of 42 antibodies were designed to characterize the immune cell phenotype. All antibodies utilized for analysis were documented in Table S2, inclusive of their specificity information. Purified primary antibodies were purchased from Fluidigm were already labeled with metal isotopes. Antibodies from other companies were labeled with metal isotopes using a Maxpar antibody labeling kit (Fluidigm, South San Francisco, CA, USA). Antibody stabilization buffer (Candor Bioscience, Allgäu, Germany) was added, and the antibodies were stored at 4 °C. Prior to the experiment, the conjugated antibodies were titrated to determine the optimal concentration.

Staining of PBMCs

PBMCs were stained with 50 nM cisplatin solution (Life Sciences, Amsterdam, Netherlands) for 1 min at room temperature to reject suitability. Cisplatin was quenched with five volumes of MaxPar Cell Staining Buffer (CSB; Fluidigm, South San Francisco, CA, USA), and the cells were stained with monoclonal antibodies conjugated to metal isotopes for 30 min at 4 °C. Cells were fixed overnight at 4 °C with MaxPar Fix and Perm Solution (Fluidigm, South San Francisco, CA, USA) containing 125 nM Cell-ID Intercalator-Ir (Fluidigm, South San Francisco, CA, USA). After washing with CSB and Milli-Q water, the cells were resuspended in cell acquisition solution and stored at 4 °C until collection.

Mass cytometry data acquisition and analysis

On average, $1.07\pm 0.489\times {10}^5$ cells for each sample were identified into mass cytometry. Mass cytometry data were randomized using the Fluidigm acquisition algorithm (V6.0.626). Individual samples were artificially gated using FlowJo to exclude normalizing beads, cell debris, dead cells, and duplexes for further analysis to identify cells. The removal of EQ beads and isolation of cells were achieved by setting the x-axis to Event_length, adjusting the range to 0–60, and the y-axis to the EQ beads channel, selecting regions where the x-values are below 40 and y-values are below 102. Isolate individual cells by setting both axes to DNA, with the brightest area in the graph indicating the target cells, and continue with further processing of these cells. Isolate live immune cells by setting the x-axis to CD45 and the y-axis to Pt platinum dye, typically selecting around 102 or below on the y-axis and generally above 101 on the x-axis for immune cells. Finally, $1.2\times {10}^4$ for each sample were used for data analysis. Data were subsequently transformed using asinh (cofactor 5) and normalized to the 99.9th percentile for each channel in the R environment (V3.6.1). For dimensionality reduction analysis, the Barnes-Hut implementation of the T-distributed stochastic neighbor embedding (t-SNE) algorithm was applied using the Rtsne package. Then, they were clustered with the phenograph algorithm by the cytofkit package using k = 30.

Detection of SARS-CoV-2 nucleic acid

Nasopharyngeal swab or throat swab samples from all subjects were collected according to the “Technical Guidelines for Laboratory Testing of COVID-19” formulated by the National Health Commission of China (Chinese Center for Disease Control and Prevention, 2020). Nucleic acid testing was performed within 24 h of sample collection. Viral RNA was extracted using an RNA extraction kit (Cat No. DA0930-DA0932; DaAnGene, Guangzhou, China) according to the manufacturer’s instructions. Real-time reverse-transcriptase quantitative PCR (RT‒qPCR) analysis was performed using a novel coronavirus nucleic acid detection kit (DaAnGene, Guangzhou, China). The cycle threshold (CT) values of nucleic acids of the SARS-CoV-2 ORF1ab/N gene were less than or equal to 40, and there was a clear amplification curve, which could be judged as positive according to the manufacturer’s instructions.

Serum IgM and IgG antibody detection

Prior to testing, all serum samples were stored at −20 °C and subsequently inactivated using a 56 °C water bath or oven for 30 min at ambient temperature following thawing. IgM and IgG antibody levels were assessed utilizing the Novel Coronavirus (2019-nCoV) IgM Antibody Detection Kit (approval number: 20203400769) and the novel coronavirus (2019-nCoV) IgG antibody detection kit (approval number: 20203770) supplied by Shenzhen Yahuilong Biotechnology Co., Ltd (Shenzhen, China). Subsequently, a calibration curve was constructed to quantify the levels of IgM and IgG antibodies. Results with a detection value ≥10.0 AU/mL were classified as positive.

Statistical analysis

Statistical analyses were carried out using the SPSS 25.0 statistical package (IBM, Armonk, NY, USA). Differences between groups were determined by Wilcoxon rank-sum test and visualized using the R package ggplot2. The results were presented as the mean ± standard deviation or median and interquartile range. Statistical significance was assessed using the Mann‒Whitney U test for two groups or Kruskal‒Wallis test with Dunn’s posttest for multiple groups. P value ≤ 0.05 was considered to indicate significance.

Clinical characteristics and single immune cell map design of COVID-19 patients

To chase the immunological responses in different stages after SARS-CoV-2 infection, 10 patients and four health control were involved in this study. The demographic and clinical characteristics of all patients and HCs were presented in Table S1. The 10 patients including two males and eight females and ranged from age 30 to 89 years old, with a median of 60 years old. A total of five (50%) people suffered from chronic diseases, such as hypertension and diabetes. The age of the HCs group was 50.50 (47.50–53.75) years old and without chronic diseases (Tables S1, S4). Cycle threshold (CT) values were shown in Tables S1, S5. In acute stage, the CT values of ORF1ab gene in symptomatic and asymptomatic patients were 23.18 (19.95–34.95) and 29.6 (26.25–30.82), respectively. The CT values of the N gene in symptomatic and asymptomatic cases were 22.09 (18.54–33.75) and 30.12 (26.79–30.62), respectively. A detailed laboratory examination results were shown in Fig. 1A, Tables S1 and S5. Lymphocyte counts were significantly lower in the acute phase than in the convalescence phase only among those with symptomatic cases (Fig. 1A). As for IgM and IgG antibodies against SARS-CoV-2, IgG antibody level in the symptomatic group was significantly higher in the convalescent period than that in the acute period (Fig. 1A). Meanwhile, IgM antibody levels in all COVID-19 patients were significantly higher in the convalescent period than that in the acute period, but IgM antibody level in the symptomatic group was significantly higher than in the HCs group (Fig. 1A).

To reveal the immunological responses at the single-cell level, CyTOF was used in this study. A schematic overview of the CyTOF workflow is depicted in Fig. 1B. The CD45⁺ cells of COVID-19 patients and HCs were divided into 29 immune cell subsets (Fig. S1A, Table S3), including four NK-cell subsets, two B-cell subsets, three DC subsets, two monocyte subsets, eight CD4⁺ T-cell subsets, four CD8⁺ T-cell subsets, one neutrophil subset, one gdT cell subset, one OX40⁺ T cell subset, one Ki67⁺ T cell subset, one mast subset and one plasma cell subset (Fig. S1B). t-SNE analysis was used to visualize the high-dimensional data and the expression level of each protein maker were illustrated by color scale from blue to red in Fig. 2A. So the immune maps of 12 types of major immune cell subgroups of all patients and HCs were obtained (Fig. 2B). Expression pattern of each protein markers was summarized for major cell subsets of immune cell clusters (Fig. S1C, Table S3). Variances in the distribution and composition of the primary immune cell subsets were noted between COVID-19 patients exhibiting varying symptoms and healthy controls (Figs. 2C, S1E). Simultaneously, the composition of different cell subsets in each sample was described (Fig. S1D).

By comparing the percentage profiles of major immune cell subsets in COVID-19 patients with different clinical manifestations in the acute and convalescence phases and HCs, we found that the immune cell subsets in patients with COVID-19 were distinct from HCs, (Figs. 2D, S1E). Specifically, we focused on the variations observed in T cells, B-cell subsets, NK cells, DCs, monocyte subsets, neutrophil cells, and mast cells.

Altered immune signatures of T-cell subsets

Human T cells are the main immune cells against viral infection and play an important role in viral clearance. In this study, fifteen T lymphocyte subsets were differentiated into five types of CD4⁺ T cells, CD8⁺T cells, gdT, OX40⁺ T cells and Ki67⁺T cell subsets according to the expression of typical lineage markers (Fig. S1C), and were mapped into t-SNE of all participants (Fig. 3A). Moreover, we analyzed the changes in major T-cell subsets in patients with different stages and clinical presentations and healthy controls (Fig. 3B).

By comparing the changes in the main T lymphocyte subsets in different disease stages, the frequency of OX40⁺ T cells (CD3⁺OX40⁺) in the convalescence stage of COVID-19 patients were significantly lower than healthy controls (Figs. 3B, S2B). According to the different functional characteristics of T cells, we divided CD4⁺ T cells into five categories: naïve CD4⁺ T cells (CD4 TN) (CD3⁺CD4⁺CD45RA⁺CD45RO⁻CD127⁺CCR7⁺), effector memory CD4⁺ T cells (CD4 TEM) (CD3⁺CD4⁺CD45RA⁻CD45RO⁺CD127⁺CCR7⁻), central memory CD4⁺T cells (CD4 TCM) (CD3⁺CD4⁺CD45RA⁻CD45RO⁺CD127⁺), type 1 T helper (Th1) (CD3⁺CD4⁺Foxp3⁺ CXCR3⁺CD45RO⁺ki67⁺CD127⁻) and regulatory T cells (Tregs) (CD3⁺CD4⁺Foxp3⁺ CD25⁺CD45RO⁺ki67⁻CD127⁻) (Fig. S1C, Table S3). According to Figs. 3C and S2C, cluster 10 (CXCR3⁺) of CD4 TN cell subsets in both symptomatic and asymptomatic patients decreased during convalescence stage than that at acute stage, while the frequency of cluster 15 (CD25⁺CD27^dimCCR7^dim) in CD4 TCM and cluster 13 (CD127⁺Granzyme_B⁺) in CD4 TEM showed the opposite trend. In addition, the frequency of Th1 cells was increased obviously in the acute stage of COVID-19 compared to healthy controls.

In addition, in the CD8⁺ T cell subsets, there were effector CD8⁺ T (CD8 TE) cells (CD3⁺CD8⁺CD45RA⁻CCR7⁻Tbet⁺), effector memory CD8⁺ T (CD8 TEM) cells(CD3⁺CD8⁺CD45RO⁺CD45RA⁻CCR7^-CD27⁺PD1⁺), naïve CD8⁺ T (CD8 TN) cells (CD3⁺CD8⁺CD45RA⁺CD127⁺CD27⁺CCR7⁺CXCR3⁺), mucosa-associated invariant T (MAIT) (CD3⁺CD8⁺CD45RA⁻CCR7⁻CD127⁺TCR_Va7.2⁺). Also it was found that cluster 21 as gdT cluster (CD3⁺TCRgd⁺), cluster 22 as OX40⁺T cells (CD3⁺OX40⁺), and cluster 26 (CD3⁺CD45⁺Ki67⁺) as Ki67⁺T cells. Cluster 16 (CD45RA⁺CCR7⁺Tbet⁺) in CD8 TE cell subsets was significantly increased in convalescence stage than that in acute stage. The frequency of MAIT cell subset and OX40⁺ T cell were decreased significantly in patients at convalescent stage than that in healthy controls (Figs. 3B, 3D). In contrast, no differences were observed in the frequencies of all the CD8⁺ T-cell subsets and other T-cell subsets in acute or convalescence COVID-19 patients with different clinical presentations (Figs. S2D, S2E).

We also observed differences in surface markers of major T-cell subsets between symptomatic and asymptomatic groups of patients with COVID-19 during acute and convalescence stage vs healthy controls (Fig. S3A), the overall expression of cell-defining markers was consistent with changes in cell subsets.

Alterations in the immune profile of B-cell subsets, NK cells, DCs, monocyte subsets, neutrophil cells and Mast cells

B cells play a crucial role in the immune response to COVID-19 by producing antibodies that target the virus and aid in its clearance from the body. The frequency of B cells was significantly higher in acute phase than that in convalescence phase (Fig. 4A). Symptomatic COVID-19 patients at the acute phase had higher B cell levels than HCs. To further investigate the variation in B-cell subsets, we detected three major peripheral blood B-cell subsets by clustering, namely IgD⁻ B cell subset (CD3⁻IgD⁻CD19⁺CD20⁺CD25⁺CD24⁺CCR6⁺), IgD⁺ B cell subset (CD3⁻IgD⁺CD19⁺CD20⁺CD25⁻CD24⁻CCR6⁺CCR7⁺) and plasma cell subset (Fig. 4A). By analyzing the t-SNE plots of all B-cell frequencies in patients with different disease stages (Fig. S4A), we found that the frequency of B-cell subsets in COVID-19 patients decreased significantly in the convalescent phase compared to the acute phase, especially the IgD⁺ B-cell subsets and plasma cell subsets (Fig. 4A). While the proportion of IgD⁺ B-cell subsets in the symptomatic group increased significantly in the acute phase compared to the convalescent phase and HCs (Fig. 4A). The proportion of IgD⁻ B-cell subsets in the asymptomatic group increased significantly in convalescence phase compared to the acute phase (Fig. S4B). In addition, the expression levels of CD19, CD24, CD27, IgD and CXCR5 in B cells were significantly different between the asymptomatic group and the healthy control group in the acute and convalescent phases of COVID-19 patients (Fig. 4B).

NK cells are the main components of cytotoxic lymphocytes and innate immunity. From the t-SNE plots of all NK cell subpopulations in all participants (Fig. S4C), the percentage of NK cells decreased significantly in the acute phase of COVID-19 compared to healthy controls and increased significantly in the convalescence phase compared to the acute phase, and then returned to normal levels (Fig. 5A). While the frequency of CD3⁻CD56^dimCD16^-NK cell subsets was obviously lower in acute phase than in convalescence phase, especially in the symptomatic group (Figs. 5A, S4D). In addition, the expression levels of CD56, CD11c, CD127 and CD25 in NK cells were significantly different between asymptomatic and symptomatic groups of COVID-19 patients in the acute and convalescence phases with healthy controls (Fig. 5B).

Monocytes are the precursors of antigen presenting DCs during inflammation. Generally, monocytes are mainly characterized by the expression of CD14 and CD16 and are further divided into classic monocyte subsets, non-classical monocyte subsets, and intermediate monocytes. We further identified two classical monocyte subsets by PhenoGraph and t-SNE as cluster 3 (CD3⁻CD4⁺CD11c⁺CD33⁺CD11b⁺CD14⁺HLA_DR⁺Granzyme_B⁻ CD16⁻), Cluster 4 (CD3⁻CD4⁺CD11c⁺CD33⁺CD11b⁺CD14⁺HLA_DR⁺CD16⁻). This study showed a significant reduction of CD14⁺CD16⁻GB⁻CXCR3⁺ classical monocyte in convalescence phase of COVID-19 patients, especially in the asymptomatic group. DCs are antigen-presenting cells that initiate and lead CD4⁺ and CD8⁺ T-cell immune responses. In this study cluster 5 (CD3⁻CD4⁺CD11c⁺CD33⁺CD272⁺CD16⁺), cluster 6 (CD3⁻CD4⁺CD33⁺CD11c⁺CD11b^dimCCR3⁺Ki67⁺HLA_DR⁺) and cluster 7 (CD3⁻CD4⁺CD11c⁺CD33^dimCD11B⁻CD123⁺HLA_DR⁺) were defined as subsets of DCs. The frequencies of DC subsets in COVID-19 patients in the acute and convalescence periods were not significantly different compared to healthy controls (Figs. 6A, S4E, S4F).

We also identified cluster 20 (CD3^dimCD11b⁺CD16⁺CD66b⁺CD24⁺) as a neutrophil subpopulation and cluster 28 (CD3⁺CD4⁻CD11c⁻CD74⁺CD117⁺) as a mast cell subpopulation (Fig. S4G). Compared to healthy controls, the frequency of neutrophils was significantly increased in COVID-19 patients in both the acute and convalescence phases. However, there was no significant difference between the acute and convalescence phases. In contrast, the frequency of mast cells was significantly lower in patients recovering from COVID-19, especially in symptomatic patients, compared to the acute phase (Figs. 6B, S4H).