Proteomic Analysis of 442 Clinical Plasma Samples From Individuals With Symptom Records Revealed Subtypes of Convalescent Patients Who Had COVID-19

Abstract

After the coronavirus disease 2019 (COVID-19) pandemic, the postacute effects of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection have gradually attracted attention. To precisely evaluate the health status of convalescent patients with COVID-19, we analyzed symptom and proteome data of 442 plasma samples from healthy controls, hospitalized patients, and convalescent patients 6 or 12 months after SARS-CoV-2 infection. Symptoms analysis revealed distinct relationships in convalescent patients. Results of plasma protein expression levels showed that C1QA, C1QB, C2, CFH, CFHR1, and F10, which regulate the complement system and coagulation, remained highly expressed even at the 12-month follow-up compared with their levels in healthy individuals. By combining symptom and proteome data, 442 plasma samples were categorized into three subtypes: S1 (metabolism-healthy), S2 (COVID-19 retention), and S3 (long COVID). We speculated that convalescent patients reporting hair loss could have a better health status than those experiencing headaches and dyspnea. Compared to other convalescent patients, those reporting sleep disorders, appetite decrease, and muscle weakness may need more attention because they were classified into the S2 subtype, which had the most samples from hospitalized patients with COVID-19. Subtyping convalescent patients with COVID-19 may enable personalized treatments tailored to individual needs. This study provides valuable plasma proteomic datasets for further studies associated with long COVID.

Keywords: COVID‐19 convalescents; plasma; proteomics; subtype; symptoms.

Figure 1

Symptom transformation during coronavirus disease 2019 (COVID‐19) recovery. Transformation of symptoms from the 6‐month follow‐up (F1, first follow‐up) to the 12‐month follow‐up (F2, second follow‐up) during COVID‐19 recovery. Further details are provided in Supporting Information S5: Table S2.

Figure 2

Workflow chart. Workflow showing the grouping of participants, questionnaire‐based symptom records, and plasma proteome detection using liquid chromatography‐mass spectrometry (LC‐MS).

Figure 3

Significantly correlated symptoms and proteins at the first follow‐up (F1 group) or the second follow‐up (F2 group). The left panel shows the significantly correlated symptoms. Spearman's correlation coefficients were calculated for symptom grade, disease level, and age, and the threshold was set at p < 0.01. Red lines represent positive correlations, and blue lines represent negative correlations. The thicker the lines, the higher the absolute values of the correlation coefficients. At least one correlation coefficient between symptoms was > 0.3 in the red dotted circles (symptom clusters 1, 2, and 3). The right panel shows proteins that were significantly correlated with symptom clusters 1, 2, or 3. The symptom grades of one patient in each symptom cluster were summed. Spearman's correlation coefficients were calculated between summed symptom grades and protein expression values, and proteins with p < 0.05 were used for functional enrichment analysis. Further details are provided in Supporting Information S8: Table S5 and Supporting Information S9: Table S6.

Figure 4

Plasma protein expression trends considering recovery time after coronavirus disease 2019 (COVID‐19). In total, 442 samples were used in this study. Differentially expressed proteins (Student's t‐test, p < 0.05 with FC_HO/HC > 1.5 or FC_HO/HC < 0.67) between the HC and HO groups were kept. These differentially expressed proteins were then separated into 12 clusters using the fuzzy c‐means algorithm, based on protein expression trends among the HC, HO, F1, and F2 groups. Reversibly upregulated clusters are indicated with red frames. The reversibly downregulated cluster is indicated with a blue frame. Irreversibly upregulated clusters are indicated with red dotted frames. Irreversibly downregulated clusters are indicated with blue dotted frames. Further details are provided in Supporting Information S10: Table S7. F1, first follow‐up group; F2, second follow‐up group; HC, healthy control group; HO, hospitalized patient group.

Figure 5

Subtypes of convalescent patients with coronavirus disease 2019 (COVID‐19) based on plasma proteome data. The plasma samples were classified into three subtypes (S1, S2, and S3) using the NMF algorithm. Sample groups, symptoms (upper panel), and expression values of subtype signature proteins (lower panel; see Supporting Information S12: Table S9) are presented as heat maps. NMF, non‐negative matrix factorization.

Figure 6

Different symptom characteristics of convalescent patients with coronavirus disease 2019 (COVID‐19) in each subtype. Six symptoms were significantly different among subtypes S1, S2, and S3. Fisher's exact test was used to compare the incidence of mMRC and appetite in subtypes S1, S2, and S3, and the Chi‐square test was used to compare the other symptoms. The thresholds was set at p < 0.05. Further details are provided in Supporting Information S13: Table S10. mMRC, modified British Medical Research Council scale.