Temporal multi-omics analysis of COVID-19 in end-stage kidney disease

Abstract

Patients with end-stage kidney disease (ESKD) are at high risk of severe COVID-19. We performed longitudinal single-cell immune profiling of ESKD patients with COVID-19. Transcriptome, surface proteome, and immunoreceptor sequencing data were generated on 580,040 high-quality cells, derived from 187 samples from 61 patients. For a subset of individuals, we obtained samples before and during infection, allowing intra-individual comparison. Longitudinal profiling demonstrated distinct temporal gene expression trajectories in severe/critical versus mild/moderate COVID-19. We identified a population of transcriptionally distinct monocytes that emerged in peripheral blood following glucocorticoid treatment. Evaluation of clonal T cell dynamics showed that the fastest expanding clones were enriched in known SARS-CoV-2-specific sequences and shared across multiple patients. Comparison with external datasets revealed upregulation of immune cell TGF-β pathway expression in ESKD, irrespective of COVID-19 status. Our data delineate the temporal dynamics of the immune response in COVID-19 in a high-risk population.

Keywords: CITE-seq; COVID-19; ESKD; SARS-CoV-2; T cell receptor; TCR; end-stage kidney disease; glucocorticoids; interferon; longitudinal; monocytes; sequencing; single-cell transcriptomics.

Graphical abstract

Figure 1

Study overview (A) Schematic of the study design. Neg. control = negative control (i.e., COVID-19⁻ patient with ESKD). Created using Biorender.com. (B) Timing of blood sampling in relation to COVID-19 onset. Colors indicate COVID-19 severity in a given individual over time. “X” with an adjacent arrow indicates death during the hospital admission occurring at >30 days (C) Uniform manifold approximation and projection (UMAP) showing the major cell-type annotations of B, myeloid and progenitor, and T cells, respectively. ASC, antibody-secreting cell; ASDC, Axl Siglec dendritic cell; CM, central memory; CTL, cytotoxic T lymphocyte; EM, effector memory; EMRA, terminally differentiated effector memory T cell; ILC, innate lymphoid cell; int, intermediate; MAIT, mucosal-associated invariant T cell; mono, monocyte; sw mem, switched memory; T g/d, gamma delta T cell.

Figure 2

Cell-type abundance and differential gene expression and pathway analysis comparing COVID-19 positive versus COVID-19 negative ESKD patients (A–J) Bar charts displaying cell proportions that significantly changed in week 1 or 2 of COVID-19 infection compared to a control group of COVID-19⁻ ESKD patients. FDR-adjusted p values, with two significant digits shown. (K) Heatmap of gene expression pathways significantly (FDR <0.05) associated with COVID-19 positivity. Pathways were defined using the Reactome database. ELEC., electron; NES, normalized enrichment score; PROD., production; RESP., respiratory; TRANSP., transport. Log-p = −log10 BH-adjusted p value. (L) Dot plot displaying the expression of the leading-edge subset of genes that contributed to the term “orexin receptor pathway” for COVID-19⁺ and COVID-19⁻ ESKD patients. Mono = monocytes.

Figure 3

Cellular and transcriptomic changes associated with COVID-19 severity in ESKD (A) Proportion of antibody-secreting B cells in mild/moderate versus severe/critical COVID-19. BH-adjusted p-value 0.00031 (linear mixed model). (B) Heatmap of pathways significantly (FDR <0.05) associated with COVID-19 severity (Reactome database). (C) Expression of selected genes associated with COVID-19 severity in total monocytes. (D) Left: monocyte TNF mRNA expression (single-cell RNA-seq dataset). TPM, transcripts per million. Center: plasma TNF-α protein abundance (Olink immunoassays). n = 57 samples from 21 individuals with both RNA and plasma protein levels measured. Right: correlation between monocyte TNF gene expression and TNF-α plasma protein levels. (E) Heatmap displaying the number of viral reads from different viral genomes, including SARS-CoV-2 (NC_045512.2), detected across cell types.

Figure 4

Longitudinal gene expression and TCR trajectories (A) Temporal cell-type abundance changes in COVID-19 over time, stratified by whether the cells exhibit an interferon (IFN)-stimulated state. Log₂ fold change relative to COVID-19⁻ samples. Significance measured by the local true sign rate with FDR control. Only cell types that have an IFN- and non-IFN-stimulated counterpart are shown. Sample n: 37 COVID-19⁻, 138 COVID-19⁺, 10 recovery. (B) Temporal changes in IFN pathway gene expression, stratified by peak illness severity. Estimated marginal mean (line) and 95% confidence intervals (shade). Cell types with significant time × severity interaction from LMM shown. n = 139 COVID-19⁺ samples. (C) Heatmap displaying 10 genes from multiple KEGG immunological disease-associated pathways that had a significantly different temporal profile in mild versus severe COVID-19 (LMM, FDR < 0.05) in CD14 monocytes. Color indicates LMM estimated marginal means over time, stratified by patient group (n = 130 samples from 37 individuals). Genes selected to represent different temporal dynamics and clustered based on the temporal profile of the discordance between mild/moderate and severe/critical disease. For (A)–(C), time since onset of disease represents either time since display of first symptom or positive test (whichever is earliest). (D) Absolute numbers of clones considered for longitudinal analysis and expanded clone counts (from n = 139 COVID-19⁺ samples, and the same samples analyzed for (E)–(J). (E) Proportion of SARS-CoV-2-specific clones among all clones, stratified by whether the clone expanded after day 10 following positive PCR test. Specificity defined as a perfect match with a TCR alpha chain from the SARS-CoV-2 database VDJDB. Two-sided Mann-Whitney test. (F) As for (E) but stratifying by whether a clone was expanded after day 2 and further after day 10. (G) SARS-CoV-2-specific clone proportion among fastest-increasing clones. Clones sorted by decreasing expansion magnitude after day 10 following positive PCR test. Dashed line: baseline of matches with database from pre-pandemic samples. (H) Sequence logos of three most shared paired-chain TCR motifs, showing the number of individuals and number of unique clones sharing the motif. Letter height indicates frequency of amino acid (aa) at that position across T cells pertaining to the motif. Each aa is colored by side chain chemistry: acidic (red), basic (blue), hydrophobic (black), neutral (purple), and polar (green). (I) Distribution of predicted activated T cells across days since positive swab result. T cell-type frequency averaged per sample and aggregated across time points. Cell states predicted using Celltypist. (J) Activated T cell state proportion among fastest-increasing clones. Clones sorted by decreasing expansion magnitude pre-/post-day 10 following positive PCR test. Dashed line: baseline proportion of activated T cells from pre-pandemic samples.

Figure 5

Dexamethasone treatment promotes a distinct monocyte subset (A) UMAPs displaying monocytes; colored by subset (top), patient cohort (bottom left), COVID-19 status (bottom center), and severity (bottom right). (B) Proportions of total monocytes stratified by COVID-19 severity. (C) Gene expression (left) and protein expression (right) across monocyte subsets. (D) Differential abundance of monocyte subsets for samples from patients given glucocorticoids pre- and post-treatment. (E) Intra-individual changes in monocyte subsets pre- and post-glucocorticoids. Line colors represent different patients. (F) Expression of monocyte marker genes across monocyte subsets. (G) Gene module scores for CD14 monocytes, IFN-stimulated CD14 monocytes and the dexamethasone-associated monocytes (dex. mono).

Figure 6

Comparison of COVID-19-associated transcriptomic changes in ESKD versus non-ESKD cohorts (A) UMAPs displaying major PBMC cell types for each of the datasets individually and after integration. (B) Number of individuals per study, separated by COVID-19 status. (C) As for (B) but showing the number of samples. (D) TGF-β signature score for COVID-19⁺ samples aggregated across cell compartments and stratified by ESKD status. (E) Combined score for TGFB1, TGFB2, and TGFB3 genes, for COVID-19⁺ samples aggregated across cell compartments and stratified by ESKD status. (F) As for (D), but for COVID-19⁻ samples. (G) As for (E), but for COVID-19⁻ samples.