Peripheral Transcriptomics in Acute and Long-Term Kidney Dysfunction in SARS-CoV-2 Infection

Abstract

Key Points:

1. AKI in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is linked to mitochondrial dysfunction and cellular stress, highlighting novel biomarkers for severe AKI.

2. Molecular similarities between AKI in SARS-CoV-2 and sepsis suggest that current therapeutic strategies may be applicable across both conditions.

3. Molecular changes in acute SARS-CoV-2 infection are tied to long-term inflammation, immune dysregulation, and lasting cardiac and renal dysfunction.

Background: AKI is common in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and coronavirus disease 2019, often leading to long-term kidney dysfunction. However, the transcriptomic features of AKI severity and its long-term effects are underexplored.

Methods: We performed bulk RNA sequencing on peripheral blood mononuclear cells from hospitalized patients with SARS-CoV-2 infection and complemented these findings with proteomic data from the same cohort. We compared the functional enrichment findings with historical sepsis–AKI data and subsequently examined the association between molecular signatures and long-term kidney function changes.

Results: In 283 patients, 57 had mild AKI (stage 1) and 49 had severe AKI (stage 2 or 3). After adjustments for age, sex, severity of infection, and preexisting CKD, we identified 6432 differentially expressed genes in the severe AKI versus control comparison, 840 in the mild AKI versus control comparison, and 1213 in the severe versus mild AKI comparison (false discovery rate <0.05). Common pathways included unfolded protein response, cellular response to stress using eIF2, and IFN-γ–mediated inflammatory response. Severe AKI was linked to pathways involved in mitochondrial dysfunction and endoplasmic reticulum stress. Proteomic analysis confirmed 40 established AKI and inflammation biomarkers, whereas gene-set enrichment of transcription regulators revealed additional biomarkers for severe AKI. Comparison with peripheral blood mononuclear cell transcriptomics from sepsis-related AKI showed significant functional overlap (30%). Analysis of postdischarge eGFR data in 115 patients identified 177 differentially expressed genes for severe AKI versus control, 106 for mild AKI versus control, and 46 for severe versus mild AKI. Key associations included kidney function decline related to carbohydrate and mitochondrial metabolism, inflammatory-response, and cardiovascular regulation.

Conclusions: We demonstrate that severe AKI in SARS-CoV-2 infection is linked to mitochondrial dysfunction and endoplasmic reticulum stress. The functional overlap with sepsis–AKI suggests potential broader therapeutic applicability. Long-term kidney dysfunction is influenced by disruptions in cellular energy metabolism and immune response.

Keywords: AKI; CKD; COVID-19; congestive heart failure; fibrosis; kidney failure; mRNA; mitochondria; oxidative stress; transcriptional profiling.

Graphical abstract

Figure 1

Flow chart detailing selection of study cohort. (A) Patients and samples included in this study and (B) visualization of the sample collection time line. (A) Flow chart of patients included in this study. (B) Time line of sample collection for this study. COVID-19, coronavirus disease 2019; QC, quality control.

Figure 2

Volcano plots for each of the comparison groups show that most of the genes have fold changes >1.5 across the comparison groups. (A) Volcano plot of the DEGs for the severe AKI (AKI stage 2/3) versus control (no AKI) comparison group. (B) Volcano plot of the DEGs for the severe AKI (AKI stage 2/3) versus mild AKI (AKI stage 1) comparison group. (C) Volcano plot of the DEGs for the mild AKI (AKI stage 1) versus control (no AKI) comparison group. DEG, differentially expressed gene; log FC, log-fold change.

Figure 3

Identifying and comparing transcription regulators and pathways across all 3 contrasts. (A) Venn diagram of the common transcription regulators across the three comparison groups uncovers four common regulators that are also known biomarkers for AKI. (B) Venn diagram of the common canonical pathways across the three comparison groups uncovers 31 common pathways, which also include six pathways unique to severe or worsening AKI and five pathways that are common to all forms of AKI severity. (C1–3) Bar plots of significantly enriched (−log10 adjusted P values) canonical pathways across the three comparison groups shows eukaryotic translation and elongation and EIF2AK4 GCN2 amino acid deficiency being common across all three comparison groups. KEGG, kyoto encyclopedia of genes and genomes; MIL–CTRL, mild AKI versus no AKI; SEV–CTRL, severe AKI versus no AKI; SEV–MIL, severe AKI versus mild AKI.

Figure 4

Molecular signatures of COVID-19–AKI (across all three comparison groups) from the COVID-19–AKI cohort against β-estimates of the linear mixed model analyzing the effect of these expressions on overall kidney function decline as measured by eGFR (over 1 year after discharge). Gene expressions are plotted on the x axis (as a function of their log₂-fold change), and the y axis is the β-estimate of the overall change in long-term eGFR. Genes with a significant fold change that positively correlated β-estimate of association (both positive or both negative) with mean eGFR conferred a protective effect on overall long-term eGFR. On the contrary, genes with a significant fold change that inversely correlated β-estimate of association (positive fold change, negative β-estimate, or negative fold change, positive β-estimate) worsened kidney function in the long term. (A–C) Association of top DEG signatures in AKI in the COVID-19–AKI cohort and overall decline in eGFR in the long-term cohort 1 year after discharge across all three comparison groups. DSP, desmoplankin.