Integrated cell-free DNA and cytokine analysis uncovers distinct tissue injury and immune response patterns in solid organ transplant recipients with COVID-19

Abstract

COVID-19 pathogenesis is associated with an exuberant inflammatory response. However, the tissue injury pattern and immune response in solid-organ transplant recipients (SOTRs) taking immunosuppressive therapy have not been well characterized. Here, we perform both cfDNA and cytokine profiling on plasma samples to map tissue damage, including allograft injury and delineate underlying immunopathology. We identified injuries from multiple-tissue types, including hematopoietic cells, vascular endothelium, hepatocyte, adipocyte, pancreas, kidney, heart, and lung in SOTRs with COVID-19 that correlates with disease severity. SOTRs with COVID-19 have higher plasma levels of cytokines such as IFN-λ1, IFN-γ, IL-15, IL-18 IL-1RA, IL-6, MCP-2, and TNF-α as compared to healthy controls, and the levels of GM-CSF, IL-15, IL-6, IL-8, and IL-10 were associated with disease severity in SOTRs. Strikingly, IFN-λ and IP-10 were markedly increased in SOTRs compared to immunocompetent patients with COVID-19. Correlation analyses showed a strong association between monocyte-derived cfDNA and inflammatory cytokines/chemokines in SOTRs with COVID-19. Moreover, compared to other respiratory viral infections, COVID-19 induced pronounced injury in hematopoietic, vascular endothelial and endocrine tissues. Allograft injury, measured as donor-derived cfDNA was elevated in SOTRs with COVID-19, including allografts distant from the primary site of infection. Allograft injury correlated with inflammatory cytokines and cfDNA from immune cells. Furthermore, longitudinal analysis identified a gradual decrease of cfDNA and inflammatory cytokine levels in patients with a favorable outcome. Our findings highlight distinct tissue injury and cytokine features in SOTRs with COVID-19 that correlate with disease severity.

Figure 1.

Study design and experimental workflow. a. Schematic diagram of study subjects (including 30 Healthy controls [HC], 21 Stable solid organ transplant [SOT] controls, 44 SOT COVID-19 [32 Mild/moderate and 12 severe] patients, 40 Non-SOT COVID-19 [24 Mild/moderate and 16 severe] patients, and 18 Lung transplant recipients with other respiratory viral infections [Other RVIs]). b. Schematic representation of plasma cfDNA and cytokine quantification. c. Heatmap representation of tissue-specific cfDNA profile across the groups.

Figure 2.

Elevation of circulating cfDNA levels in solid organ transplant recipients (SOTRs) with COVID-19. Comparison of absolute plasma cell-free mitochondrial (mtcfDNA) (a) and nuclear DNA (ncfDNA) (b) and tissue-specific cfDNA derived from monocytes (c), B cells (d), NK cells (e), neutrophils (f), erythroblasts (g), vascular endothelium (h), adipocytes (i), hepatocytes (j), lung (k), pancreas (l), cardiac myocytes (m), kidney (n), bladder (o), and squamous epithelium (p) from Healthy controls (HCs) (n=19), stable SOT controls (n=21), and SOTRs with COVID-19 patients (n=44). Plasma ncfDNA and mtcfDNA concentrations were measured by digital droplet PCR. cfDNA Whole-genome bisulfite sequencing was performed to measure tissue-specific cfDNA profiles, leveraging tissue-specific DNA methylomes and deconvolution algorithms. Median [interquartile range (IQR)] of cfDNA copies per mL (cp/mL) are reported. Statistical significance was determined by the Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). A p-value ≤0.05 and FDR≤0.25 was considered statistically significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 3.

Comparable cfDNA profiles between SOT and Non-SOT COVID-19 patients. Absolute plasma mtcfDNA (a) and ncfDNA (b) levels in Non-SOT (n=40) and SOT COVID-19 patients. c–p. Quantification of tissue-specific cfDNA levels: monocytes (c), B cells (d), NK cells (e), neutrophils (f), erythroblasts (g), vascular endothelium (h), adipocytes (i), hepatocytes (j), lung (k), pancreas (l), cardiac myocytes (m), kidney (n), bladder (o), and squamous epithelium (p) from Non-SOT and SOT COVID-19 patients. Statistical significance was determined by the Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 4.

SOTRs with COVID-19 displayed dysregulated cytokine responses. Comparison of plasma cytokine/chemokine levels on healthy controls (n=30), Non-SOT COVID-19 patients (n=38) and SOTRs with COVID-19 (n=44). Cytokine values are reported as picograms per milliliter (pg/mL). Statistical significance was determined by the Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex & BMI). A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 5.

Plasma cfDNA features stratify SOT COVID-19 patients by severity. COVID-19 subjects were grouped based on disease severity as mild/moderate and severe. Comparisons of absolute total mtcfDNA (a) ncfDNA and mtcfDNA (b) and tissue-specific cfDNA levels derived monocytes (c), B cells (d), NK cells (e), neutrophils (f), erythroblasts (g), vascular endothelium (h), adipocytes (i), hepatocytes (j), lung (k), pancreas (l), cardiac myocytes (m), kidney (n), bladder (o), and squamous epithelium (p) from SOTRs with mild-moderate (n=12) and severe disease (n=32). (q). ROC curve analyses using admission cfDNA profile was performed to identify SOT patients with severe COVID-19. Statistical significance was determined by the Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). Statistical significance was determined by the Mann–Whitney test. A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 6.

Plasma cytokine levels associated with COVID-19 disease severity in SOTRs. (a). Comparison of plasma cytokine/chemokine levels in mild/moderate (n=32) verses severe SOTRs with COVID-19 (n=12). Cytokine values are reported as picograms per milliliter (pg/mL). (b). ROC curve analyses using admission cytokine profile was performed to identify SOT patients with severe COVID-19. Statistical significance was determined by Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 7.

A strong association between circulating cfDNA and inflammatory cytokine/chemokine signatures in SOTRs with COVID-19. Admission plasma cfDNA and cytokine were measured from SOTRs with COVID-19 (n=44) and Non-SOT COVID-19 patients (n=38). Pearson correlation matrix analysis of cfDNA profiles and cytokines/chemokines levels in SOTRs (a) with COVID-19 and Non-SOT COVID-19 (b). Scatter Plot of the relationship between monocyte derived cfDNA and cytokines in SOT patients with COVID-19 (c). Correlation scatter plot plots between neutrophil derived cfDNA and cytokines in SOT patients with COVID-19 (d). Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). Statistical significance was determined by the Spearman correlations test. A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 8.

Linear regression analysis reveals an important association between cfDNA and cytokine signatures. Hierarchical Clustering Heatmap of circulating cfDNA and cytokine signatures in SOT (a) and Non-SOT patients COVID-19 (b). t-statistics scatterplot plots between monocyte derived cfDNA and cytokines in SOT patients with COVID-19 (c). t-statistics scatter plot plots between neutrophil derived cfDNA and cytokines in SOT patients with COVID-19 (d). Linear regression analysis between cfDNA features and cytokine profile was conducted. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). A p-value ≤0.05 and FDR≤0.25 was considered significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 9.

High plasma cfDNA levels in lung transplant recipients with COVID-19 compared to other respiratory viruses. Comparisons of absolute total mtcfDNA (a) ncfDNA and mtcfDNA (b) and tissue-specific cfDNA levels derived monocytes (c), B cells (d), NK cells (e), neutrophils (f), erythroblasts (g), vascular endothelium (h), adipocytes (i), hepatocytes (j), lung (k), pancreas (l), cardiac myocytes (m), kidney (n), bladder (o), and squamous epithelium (p) among lung transplant recipients with COVID-19 (n=18) and other respiratory viral infections (other RVIs) (n=21). Statistical significance was determined by the Mann–Whitney test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex & BMI). A p-value ≤0.05 and FDR≤0.25 was considered statistically significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.

Figure 10.

Elevated ddcfDNA level in SOTRs with COVID-19. (a) comparison of ddcfDNA level between stable transplant recipients (n=18) and SOTRs with COVID-19 (n=39). (a) comparison of ddcfDNA level between stable controls (n=21) and SOTRs with COVID-19 (n=44). (b) comparison of ddcfDNA in SOT patients with mild/moderate (n=28) and severe COVID-19 disease (n=12). (c) comparison of ddcfDNA level in lung transplant recipients with COVID-19 (LTR) (n=17) and other SOT types (n=22). (c) comparison of ddcfDNA level in lung transplant recipients with COVID-19 (LTR) (n=17) and other RVIs (n=19). (d) Correlation scatter plot plots between ddcfDNA and cytokines in SOT patients with COVID-19. (e) Correlation scatter plot plots between ddcfDNA level and total and tissue-specific cfDNA profiles in SOT patients with COVID-19. Statistical significance was determined by Mann–Whitney test and Spearman correlations test. Adjusted p values are reported (for multiple comparison and demographic factors (age, sex, and BMI). A p-value ≤0.05 and FDR≤0.25 was considered statistically significant;*: FDR ≤ 0.25 and p-value ≤ 0.05, **: FDR ≤ 0.1 and p-value ≤ 0.05, ***: FDR ≤ 0.05 and p-value ≤ 0.05, NS: FDR > 0.25 or p-value > 0.05.