Plasma biomarkers for systemic inflammation in COVID-19 survivors

Abstract

Background: While the majority of COVID-19 patients fully recover from the infection and become asymptomatic, a significant proportion of COVID-19 survivors experience a broad spectrum of symptoms lasting weeks to months post-infection, a phenomenon termed "post-acute sequelae of COVID-19 (PASC)." The aim of this study is to determine whether inflammatory proteins are dysregulated and can serve as potential biomarkers for systemic inflammation in COVID-19 survivors.

Methods: We determined the levels of inflammatory proteins in plasma from 22 coronavirus disease 2019 (COVID-19) long haulers (COV-LH), 22 COVID-19 asymptomatic survivors (COV-AS), and 22 healthy subjects (HS) using an Olink proteomics assay and assessed the results by a beads-based multiplex immunoassay.

Results: Compared to HS, we found that COVID-19 survivors still exhibited systemic inflammation, as evidenced by significant changes in the levels of multiple inflammatory proteins in plasma from both COV-LH and COV-AS. CXCL10 was the only protein that significantly upregulated in COV-LH compared with COV-AS and HS.

Conclusions: Our results indicate that several inflammatory proteins remain aberrantly dysregulated in COVID-19 survivors and CXCL10 might serve as a potential biomarker to typify COV-LH. Further characterization of these signature inflammatory molecules might improve the understanding of the long-term impacts of COVID-19 and provide new targets for the diagnosis and treatment of COVID-19 survivors with PASC.

Keywords: COVID-19; SARS-CoV-2; biomarkers; inflammation; long haulers.

FIGURE 1

SARS‐CoV‐2 specific IgG concentrations and levels of inflammatory proteins in convalescent plasma from COVID‐19 survivors and HS. (A) The correlation between SARS‐CoV‐2 neutralization rates and log IgG concentrations was determined by Spearman correlation analysis in 22 COV‐AS and 22 COV‐LH. (B) SARS‐CoV‐2 specific IgG concentrations in 22 COV‐AS versus 22 COV‐LH, measured by ELISA. (C) Changes in SARS‐CoV‐2 IgG concentrations after COVID‐19 PCR diagnosis in 22 COV‐AS and 22 COV‐LH. (D) Heatmap showing the levels of 92 inflammatory proteins across 66 samples from HSs, COV‐AS, and COV‐LH, n = 22 per group. Each row represents one subject (y‐axis, right) and each column represents one particular protein (x‐axis, bottom). The alteration of color from red to blue indicates a change from a higher protein level to a lower protein level within one protein assay. COV‐AS, COVID‐19 asymptomatic survivors; COV‐LH, COVID‐19 long haulers; HS, healthy subject; IgG, immunoglobulin G; SARS‐CoV‐2, severe acute respiratory syndrome coronavirus 2

FIGURE 2

Dysregulated inflammatory proteins in convalescent plasma from COV‐AS. (A) PCA plot of plasma proteins in 22 COV‐AS versus 22 HS. The 44 samples were scatter plotted within two PCs, with PC1 on the x‐axis and PC2 on the y‐axis. A PC score of 28.08% means that 28.08% of variances are explained by PC1, which accounts for the majority of the variance in the original data; whereas 15.38% of variances are explained by PC2, which accounts for the second highest amount of variance in the NPX values. Each dot represents one subject, with orange color representing COV‐AS and blue color representing HS. All 92 assays from the Olink Target 96 Inflammation panel were used in the PCA analysis. Circles indicate the two clusters of subjects from COV‐AS and HS. (B) Volcano plot depicting 25 out of 92 proteins that displayed statistically significant differences in NPX values between COV‐AS and HS. Each dot represents one particular protein, with green color representing the statistically significant difference in adjusted p values and gray color representing non‐significant difference. The y‐axis shows log10 (p‐value), and the x‐axis shows the difference in NPX value between HS versus COV‐AS (HS_NPX – COV‐AS_NPX). The dotted line indicates a cut off p‐value (p = 0.05). Protein with greater than two‐fold changes (either upregulated or downregulated) in abundance levels are listed in the volcano plot. (C, D) Summary results of NPX for the upregulated (C) and downregulated proteins (D) in plasma from 22 COV‐AS versus 22 HS. Statistically significant p‐values are shown. COV‐AS, COVID‐19 asymptomatic survivors; HS, healthy subject; NPX, normalized protein expression; PC, principal component; PCA, principal component analysis

FIGURE 3

Dysregulated inflammatory proteins in convalescent plasma from COV‐LH. (A) PCA plot of plasma proteins in COV‐LH versus 22 HS. A total of 44 samples were scatter plotted within the two PCs, with PC1 on the x‐axis and PC2 on the y‐axis. Approximately 29.24% of variances were accounted for by PC1, and 13.11% of variances were explained by PC2. Each dot represents one subject, with red color representing COV‐LH and blue color representing HS. Circles indicate the data distribution in PC1 versus PC2 from two clusters of subjects (COV‐LH versus HS). (B) A volcano plot depicting the 43 out of 92 proteins that displayed statistically significant differences in adjusted p‐values between COV‐LH and HS. Each dot represents one particular protein, with green color representing statistically significant differences and gray color representing the non‐significant differences. The y‐axis shows the log10 (p‐value), and the x‐axis shows the difference in NPX value between HS versus COV‐LH (HS_NPX – COV‐LH_NPX). The dotted line indicates a cut off p‐value (p = 0.05). Proteins with more than two‐fold changes in abundance levels are listed. (C, D) Summary results of NPX for the upregulated (C) and downregulated proteins (D) in plasma from 22 COV‐LH versus 22 HS. Statistically significant p‐values are shown. COV‐LH, COVID‐19 long haulers; HS, healthy subject; NPX, normalized protein expression; PC, principal component; PCA, principal component analysis

FIGURE 4

Identification of three inflammatory proteins differentially expressed in HS, COV‐AS, and COV‐LH. (A–C) Summary results of 2^NPX for CXCL10, CCL20, and 4E‐BP1 protein levels in all three subject groups. (D) Concentrations of CXCL10 (pg/mL) amongst the three subject groups. Statistical differences among three groups were determined by one‐way ANOVA (data with normal distribution) or the Kruskal–Wallis test (data with non‐normal distribution), presented as an overall p‐value at the top. Differences between two groups were determined by Tukey's multiple comparisons test (A and C) or Dunn's multiple comparisons test (B and D), and p‐values are shown on top. (E) The correlation between CXCL10 concentrations and 2^NPX of CXCL10 was determined by Spearman correlation analysis in 21 COV‐AS, 21 COV‐LH, and 22 HSs. Each dot represents one subject. All outliers were identified by the ROUT method (Q = 1.000 %) and excluded from the analysis. ANOVA, analysis of variance; COV‐AS, COVID‐19 asymptomatic survivors; COV‐LH, COVID‐19 long haulers; HS, healthy subject; NPX, normalized protein expression