Abstract
Background AKI is a common sequela of infection with SARS-CoV-2 and contributes to the severity and mortality of COVID-19. Here, we tested the hypothesis that kidney alterations induced by COVID–19–associated AKI could be detected in cells collected from urine. Methods We performed single-cell RNA sequencing (scRNAseq) on cells recovered from the urine of eight hospitalized patients with COVID-19 with (n55) or without AKI (n53) as well as four patients with non–COVID–19 AKI (n54) to assess differences in cellular composition and gene expression during AKI. Results Analysis of 30,076 cells revealed a diverse array of cell types, most of which were kidney, urothelial, and immune cells. Pathway analysis of tubular cells from patients with AKI showed enrichment of transcripts associated with damage-related pathways compared with those without AKI. ACE2 and TMPRSS2 expression was highest in urothelial cells among cell types recovered. Notably, in one patient, we detected SARS-CoV-2 viral RNA in urothelial cells. These same cells were enriched for transcripts associated with antiviral and anti-inflammatory pathways. Conclusions We successfully performed scRNAseq on urinary sediment from hospitalized patients with COVID-19 to noninvasively study cellular alterations associated with AKI and established a dataset that includes both injured and uninjured kidney cells. Additionally, we provide preliminary evidence of direct infection of urinary bladder cells by SARS-CoV-2. The urinary sediment contains a wealth of information and is a useful resource for studying the pathophysiology and cellular alterations that occur in kidney diseases.
ORGANIC HEALTH FOODS FOR KIDNEY HEALTH
Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection triggers pathology across multiple systems, including the kidney, and AKI is associated with significant morbidity and mortality in coronavirus disease 2019 (COVID-19) (1–5). Multiple studies have demonstrated high rates of AKI among hospitalized patients with COVID-19 (6), with some reporting up to 50% of infected individuals developing AKI (7–9). The primary SARS-CoV-2 receptor ACE2 is expressed on epithelial cells throughout the urinary system, including proximal tubule cells and urothelial cells (10–13), although it is unclear if AKI in patients with COVID-19 is due to direct viral infection of the proximal tubules or is a result of the systemic response to SARS-CoV-2 (14–22). Similarly, it is unclear if SARS-CoV-2 can cause viral cystitis via direct infection of urothelial cells, although this possibility has been proposed (23–26). Despite the high prevalence of COVID–19–associated AKI, the underlying cellular alterations that occur in the setting of AKI remain poorly understood.
COVID–19–associated AKI has remained largely understudied as access to kidney tissue requires biopsy or postmortem analysis (27). Recent studies have detected a diverse array of kidney, bladder, and immune cells in the urine (28–30). Thus, the urine may offer valuable insight into the noninvasive study of kidney changes during COVID–19–associated AKI. Here, we performed single-cell RNA sequencing (RNA-Seq) to characterize the cellular diversity in the urine of hospitalized patients with COVID-19 with and without AKI. We tested the hypothesis that kidney alterations in COVID–19–associated AKI could be detected in cells collected from urine. We also collected samples from patients without COVID-19 and with AKI (non–COVID–19 AKI). We found several inflammatory immune cell populations and differentially activated pathways in COVID–19–associated AKI as well as preliminary evidence for direct infection of urothelial cells by SARS-CoV-2.
Materials and Methods
Participants and Variables Adults aged 18 years old and older were screened during admission or transfer to the University of Alabama at Birmingham (UAB) hospital between March and May 2021. Cases of AKI were identified using the Kidney Disease Improving Global Outcomes definition as a rise in serum creatinine (sCr) .0.3 mg/dl within 48 hours or .1.53 baseline creatinine. Controls with no change in creatinine were selected based on age and sex matching where possible and processed with each respective AKI sample. Baseline sCr was determined using the most recent sCr value 7–365 days before hospitalization. Additional clinical data regarding demographics, medical history, clinical characteristics, and laboratory values were extracted from patient charts through the UAB Center for Clinical and Translational Sciences i2b2 team. Eight hospitalized patients with COVID-19, five with AKI and three without AKI, were included. This study and the specimen collection were approved by the UAB Institutional Review Board. The UAB Acute Nephrology Consult Team also collected samples from patients with non–COVID–19 AKI (n54) to compare cellular changes with COVID–19–associated AKI. These were collected under a different institutional review board protocol that allows for the collection of remnant urine samples and thus, is anonymous.
Specimen Collection and Processing
All steps were performed on ice. Urine was collected either as a voided specimen or from a urinary catheter. Samples were immediately transferred to a biosafety level 21 laboratory for processing. Urine samples were transferred to a 50-ml conical tube and centrifuged at 10003g for 10 minutes at 4 C. Cell pellets were washed with ice-cold PBS, filtered through a 40-mm filter, and centrifuged again. Live cells were purified using the MACS Debris Removal Kit (Miltenyi Biotec) followed by the EasySep Annexin V Dead Cell Removal Kit (StemCell). Briefly, cells were resuspended in ice-cold PBS and mixed with a debris removal solution. Cold PBS was overlaid on the mixture and centrifuged at 3003g for 10 minutes at 4 C. The top two phases were aspirated, and then, the remaining cells were washed with PBS. Cells were resuspended and mixed with Dead Cell Removal Cocktail, Biotin Selection Cocktail, and then, RapidSpheres before separation in an EasySep magnet. Cells were washed and resuspended in 52 ml of PBS (no calcium or magnesium) 10.04% BSA (Fisher Scientific) for scRNAseq processing
scRNAseq
Purified cells were transferred to the UAB Flow Cytometry and Single Cell Core and immediately processed using the Chromium 39 Single-Cell RNA Sequencing Kit (103 Genomics) according to the manufacturer's instructions. The cell suspension was counted, combined with the 103 Chromium reagent mixture, and loaded into a microfluidic single-cell partitioning device in which lysis and reverse transcription occur in microdroplets. The resulting cDNA was amplified by PCR and subsequently processed to yield
bar-coded sequencing libraries. Paired-end sequencing was carried out on the Illumina NovaSeq6000 or NextSeq500 sequencing platform (Illumina). Reads were processed using the 103 Genomics Cell Ranger Single-Cell Software Suite version 6.0 on the UAB Cheaha High-Performance Computing Cluster. BCL files were converted to FASTQ files using the CellRanger fast function. CellRanger fast was used to align FASTQ files to a custom genome consisting of the hg38 human genome (GRCh38.p13) with
the SARS-CoV-2 genome (NC_045512.2) inserted as an exon (31). The genes table, barcode table, and transcriptional expression matrices were created for the analysis indicated below.
Data Analyses
Analyses were carried out using packages created for the R statistical analysis environment (version 4.06). Data were primarily analyzed in Seurat version 3.2.3 (32,33) and its associated dependencies. Data from each patient were imported using the Read103 function and then structured into a Seurat object using CreateSeuratObject. For quality control, cells with unique feature counts over 2500 or under 200 and cells with mitochondrial proportions of .15% were filtered out. Data were normalized using Log Normalize and scaled to prepare for linear dimensional reduction. Objects from individual patients were labeled with unique group identifications and then merged into a single object using the Seurat merge function. Patient samples were integrated with the RunHarmony function using the Harmony R package (34). Principal component analysis was performed, and then, cells were clustered based on differential gene expression as determined by the Seurat FindAllMarkers function set to a resolution of one. Dimensional reduction was performed using uniform manifold approximation and projection. Cell types were identified by comparing the differentially expressed transcripts for each cluster with known transcripts associated with specific cell types (29,35,36). The Escape R package was used to run Gene Set Enrichment Analysis (GSEA) (37). WebGestaltR was used for gene ontology analysis to identify pathways using the Biologic Process and Kyoto Encyclopedia of Genes and Genomes databases (38).
