Introduction
Acute kidney injury (AKI) is a common clinical syndrome characterized by a sudden decline of kidney function, which is associated with high mortality, prolonged hospitalization trends, as well as high costs of medical care.[1,2] The incidence rate of AKI is about 10–15% in hospital inpatients, while its prevalence is sometimes more than 50% in intensive care unit (ICU) patients.[3,4] Despite considerable advances in the management of AKI, methods for early diagnosis and specific treatment of AKI are still lacking.
Although the causes of AKI are multifactorial, ischemia emerges as one of the primary causes.[5] Ischemic AKI can occur for a series of reasons, for example, hypotension linked to diarrhea, blood loss after trauma or surgery, as well as the use of vasoconstrictive drugs.[6,7] Until now, the underlying pathophysiology of ischemic AKI remains largely enigmatic. Hence, identifying the molecular mechanisms underlying ischemic AKI, and exploring novel early diagnostic and therapeutic targets are paramount to improve the outcome of ischemic AKI.
Single-cell RNA sequencing (scRNA-seq) is a powerful method for defining cell-type transcriptomes in given tissues and measuring gene expression in individual cells.[8,9] scRNA-seq has also comprehensively uncovered molecular targets and signaling pathways in human diseases at single-cell level.[10–12] This new technique has been used in several kidney diseases.[13–16] More recently, comprehensive cell-specific profiles and molecular mechanisms were revealed by scRNA-seq in ischemia/reperfusion (I/R) mice model.[17–19] However, these findings have yet to be validated in AKI patients. In this study, we employed scRNA-seq to renal biopsies of patients with ischemic AKI in order to provide insights into biological mechanisms and potential intervention targets for ischemic AKI.
Methods
Ethical approval
This study was performed in accordance with the Declaration of Helsinki and received approval from the Medical Ethics Committee of Xiangya Hospital of Central South University for Human Studies (No. 201711836). The animal study was reviewed and approved by The Institutional Animal Care and Use Committee of Central South University (No. 2021SYDW0206). All participants in our study provided written informed consent.
Clinical samples collection
Renal specimens were obtained from two patients diagnosed with ischemic AKI at the Department of Nephrology in Xiangya Hospital, Central South University. Kidney biopsy was performed with an 18-gauge core needle in consented subjects with ischemic AKI. The initial diagnosis of AKI and staging was defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Ischemic AKI was eventually diagnosed by renal histology. Urine neutrophil gelatinase-associated lipocalin (NGAL) concentration was measured by Enzyme-linked immunosorbent assay (ELISA), and the kit was provided by Chongqing Zhongyuan Biotechnology Company (Chongqing, China).
Public dataset collection and processing
The scRNA-Seq data of normal kidney samples from three human donors were acquired from the National Center of Biotechnology Information (NCBI) Gene Expression Omnibus database (GSM4145204, GSM4145205, and GSM4145206). Two out of the three specimens were acquired from patients who received radical nephrectomy, and the remaining specimen was from a patient who received radical nephroureterectomy.[20] Normal kidney tissues were collected at least 2 cm away from the tumor tissues. The downstream analysis was reproduced through the code supplied by the original manuscript. Harmony, which is a fast, sensitive, and accurate algorithm for performing integration of single-cell genomics datasets for batch correction and removing the influence of dataset-of-origin from the embedding,[21] was used to integrate samples and conduct downstream analysis using Seurat (version 3.1; Satija Lab,https://satijalab.org/seurat/). We normalized and scaled all the gene expressions through NormalizeData and ScaleData.[22] Clustering analysis using FindClusters [22] was conducted by first reducing the gene expression matrix to the first 20 principal compositions. Afterward, we used a resolution of 0.3 for graph-based clustering.
Tissue processing and single-cell isolation
The fresh renal tissue was washed with sterile phosphate-buffered saline (PBS, HyClone, Marlborough, MA, USA) after acquisition and transported into GEXSCOPE Tissue Preservation Solution (Singleron Biotechnologies, Nanjing, China) at 2–8°C rapidly. The extracted fresh kidney tissue samples were processed and dissociated into single-cell suspensions using a previously described method.[23] Cell viability was examined by trypan blue staining (Gibco, Grand Island, NY,USA). The sample processing could be conducted when cell viability exceeded 80%.
Library preparation and scRNA-seq procedure
The single-cell suspension at the concentration of 1 × 105 cells/mL was adjusted by PBS and was subsequently loaded onto the microfluidic chip. The scRNA-seq library was constructed according to the manufacturer's instructions (Singleron GEXSCOPE Single Cell RNA-seq Library Kit, Singleron Biotechnologies, China), including cell lysis, mRNA trapping, cell labeling (barcode), mRNA (Unique Molecular Identifiers [UMI]), mRNA reverse transcription into complementary deoxyribonucleic acid (cDNA) and amplification, and cDNA fragment. The captured libraries were sequenced by Illumina HiSeq ×10 sequencers with 150 bp paired-end reads. The internal pipeline was used to process raw reads to produce gene expression profiles. Poly(A) tails and adapters were trimmed with fastp (v 0.19.5) from R2 reads, and the trimmed R2 reads were aligned against GRCh38 using integrated version 92 gene annotation (featureCounts 1.6.2 and fastp 2.5.3a). The same cell barcode, gene, and UMI were grouped to quantify the number of UMIs per gene per cell. The scRNA-seq data were deposited in the GEO at NCBI (GSE174220).
Marker gene examination and cell type identification
In this study, we removed cells with fewer than 200 or higher than 5000 expressed genes. Cells with more than 30,000 UMIs and over 50% mitochondria content were also filtered out. The scRNA-seq data including cell population determination and clustering analysis were analyzed by the Seurat program (http://satijalab.org/seurat/, R package, v3.0.1). The shared nearest neighbor (SNN)-based model implemented in the Seurat package was applied for cluster analysis. Dimension reduction operation (principal component analysis [PCA], t-distributed stochastic neighbor embedding [tSNE], uniform manifold approximation and projection [UMAP]) was used to determine the location status of cells. After identification of each cell group, Wilcox (Wilcoxon rank-sum) test with Seurat FindAllMarkers function was applied for identifying the marker genes of each cell cluster in the kidney. The marker genes were chosen based on the expression exceeding 10% of cells per group, and the average log(Fold Change) exceeding 0.25. The heatmap was created by identification of the top 20 marker genes of each cell population.
Differentially expressed genes (DEGs) analysis
Genes differentially expressed in each cell cluster in kidney was defined by comparing transcription profiling of patients with ischemic AKI and controls. The DEGs of each cell group between two groups were identified by the Wilcox test with the Seurat FindAllMarkers function. The DEG was determined based on a gene with the average log(Fold Change) threshold higher than 0.25 and the P value lower than 0.05.
Enrichment and cell-cell crosstalk communication analysis
Gene Ontology (GO) enrichment analysis of altered genes was carried out to clarify molecular functions and biological processes by the clusterProfiler software (version 3.8.1, Bioconductor, https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html). Likewise, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of DEGs was performed to find key pathways using the clusterProfiler software. Ligand-receptor interaction between different cell clusters was analyzed to determine cell-cell communication by CellphoneDB (v2.1.7, https://github.com/Teichlab/cellphonedb).
Renal ischemia/reperfusion injury model
8–10-week-old wild-type male C57BL/6 (B6) mice were included in this study. AKI was induced by renal I/R injury as described by our previous study.[24] Mice were anesthetized by intraperitoneal injection of pentobarbital sodium 50 mg/kg (Tianjin Yifang Technology, Tianjin, China), then put on the homeothermic blanket to keep body temperature at 36.5°C. The kidney pedicles were exposed via a lateral incision and bilaterally clamped by microarterial clips (Fine Science Tools, Germany) for 30 min to induce renal ischemia. After the duration of reperfusion for 24 h, the mice were anesthetized again and sacrificed. Kidneys were collected after the left ventricle was perfused with PBS. The sham groups received the same operation without clamping the renal pedicles.
Immunohistochemistry (IHC) staining
In brief, IHC staining was routinely conducted on paraffin-embedded renal sections according to the manufacturer's instructions. For primary antibodies, the dilute proportions were performed in human kidney: immediate early response 3 (IER3; Abcam, ab65152, Cambridge, UK): 1:200; early growth response 1 (EGR1; Abcam, ab194357): 1:400; SAM and SH3 domain-containing protein 1 (SASH1; Abcam, ab111901): 1:600; paralemmin-3 (PALM3; Abcam, ab189996): 1:800; and protein disulfide isomerase family A Member 6 (PDIA6; Abcam, ab227545): 1:400. For primary antibodies, the dilute proportions were conducted in kidney of I/R injury mice: ubiquitin-specific peptidase 47 (USP47; Abclonal, A15461, Boston, USA): 1:300; ras-association domain family 4 (RASSF4; GeneTex, GTX117153, San Antonio, USA): 1:100; estrogen receptor-binding fragment-associated antigen 9 (EBAG9; Bioss, bs-19557R, Beijing, China): 1:400; IER3 (Abcam, ab65152): 1:400; EGR1 (Proteintech, 22008-1-AP, Chicago, USA): 1:400; SASH1 (Bioss, bs-6099R): 1:600; PALM3 (Abcam, ab189996): 1:500; and PDIA6 (Abcam, ab227545): 1:400. Images of sections were acquired by a H-7700 transmission electron microscope (Hitachi, Tokyo, Japan). Semiquantitative calculation of the staining degree was conducted by Image-Pro plus 6.0 immunohistochemical analysis software. We used the average optical density (A) quantitative analysis index to compare the protein expression.
Results
Clinical and renal pathological characteristics of ischemic AKI patients
Two patients with ischemic AKI underwent a kidney biopsy and were included in our study. The major morphological features of kidney in patients with ischemic AKI were acute renal tubular necrosis, swelling, and varying degrees of degeneration, accompanied by inflammatory cell infiltration in renal interstitium by light microscopy, as well as normal or slight glomerular lesions. Dilation of lumens, loss of brush border, and shedding of degenerating tubular epithelial cells into the lumens also could be seen in some renal tubules [Supplementary material file 1A–D,https://links.lww.com/CM9/B526]. There was no obvious immune complex deposition was found on immunofluorescence. Furthermore, acute renal tubular injury was determined by electron microscopy [Supplementary material file 1E, F, https://links.lww.com/CM9/B526].
One AKI patient was a 45-year-old male, and his serum creatinine (Scr) was 3.28 mg/dL at the time of renal biopsy. Another AKI patient was a 53-year-old male, and his Scr was 2.84 mg/dL at the time of renal biopsy. Urinary AKI biomarker–NGAL was elevated (482 ng/mL and 603 ng/mL, respectively) in two AKI patients, and the urinalysis showed normal or trace proteinuria. Both AKI patients were diagnosed with AKI stage 3. Patients with AKI had no disease history including primary renal diseases, hypertension, diabetes, or other secondary renal diseases. Combined with clinical, laboratory, and renal pathological data, the causes of two AKI patients in this study were mainly attributed to ischemia induced by severe diarrhea. Scr of the first AKI patient decreased to 1.23 mg/dL 10 days after renal biopsy. Scr of the second AKI patient went down to 1.12 mg/dL 7 days post renal biopsy. During half a year of follow-up, the level of Scr remained in the normal range in two AKI patients.
Classification of cell lineage in human renal cortical tissue
We sequenced 19,715 single kidney cells totally in kidney samples from ischemic AKI patients. The cell viability was 81.84% and 89.22% [Supplementary material file 1,https://links.lww.com/CM9/B526. We also integrated and compared the results of two ischemic AKI patients with three controls from public datasets in the Gene Expression Omnibus.[20] As these five kidney data sets were not well integrated to be represented as a distinct batch, we used the Harmony algorithm[21] to mitigate the batch effect [Supplementary material file 2A, B,https://links.lww.com/CM9/B526]. Single-cell transcriptomes were generated after data handling and quality control. Cell clusters were classified based on the anchor gene expression of each cell type and the literature. UMAP plot of 15 distinct kidney cell populations was displayed in Figure 1A, including mesangial cells, podocytes, endothelial cells (EC), loop of Henle (LOH) cells, proximal and distal tubule (DT) cells, principal and intercalated cells (IC) from collecting ducts, macrophages (MC), monocytes (MON), dendritic cells, B cells, natural killer T (NKT) cells, smooth muscle cells (SMC), and fibroblasts. The number of distinct cell clusters in kidney from each subject was illustrated in Dataset 1, http://links.lww.com/CM9/B532.UMAP plot of major cell types distribution from each subject was illustrated in Figure 1B. The proportions of distinct cell clusters from individual AKI patients and control subjects were determined [Figure 1C]. In order to identify mutually exclusive profiles and kidney cell types, the top 20 most differential expression marker genes analysis for each cell population was conducted [Figure 1D and Dataset 2, http://links.lww.com/CM9/B533]. The representative cell-type-specific marker gene was represented as a violin plot [Figure 1E].
;)
Figure 1: Cell lineage analysis in kidney determined by scRNA-seq technology in AKI and control subjects. (A) Fifteen distinct cell populations in kidney were identified by UMAP plot, and each cell was colored representing different subclusters. (B) UMAP plot of cell clusters from distinct subjects with AKI or control, and cells were colored for the individual origin. (C) Bar plot displayed the percentage of cell clusters in kidney from different subjects. Blocks represented distinct subjects, and the cell quantity was in proportion to the block height. (D) Heatmap represented the top 20 most differentially expressed genes of each cell type to identify the cell lineage. Each column reflected an individual cell population, and each row showed marker genes of a single cluster. (E) Violin plot indicated selected marker genes for each cell cluster, and the colors represented marker genes for each cell type. AKI: Acute kidney injury; Con: Control subjects; DC: Dendritic cells; DT: Distal tubule cells; EC: Endothelial cells; FIB: Fibroblasts; scRNA-seq: single-cell RNA sequencing; UMAP: Uniform manifold approximation and projection; SMC: Smooth muscle cells; POD: Podocytes; MES: Mesangial cells; PC: Principal cells; NKT: Natural killer T; IC: Intercalated cells; MON: Monocytes; MC: Macrophages; LOH: Loop of Henle cells; PT: Proximal tubule cell.
Cell-specific marker genes were combined to determine each kidney cell type [Table 1]. For example, high expression of FHL2, CTGF, MYL9, and AKAP12 defined the cluster of mesangial cells (MES). The marker genes NPHS2, PTPRO, and PODXL identified podocytes (POD). EC highly expressed PECAM1, VWF, CLDN5, and ACKR1, whereas LUM, DCN, and COL1A1 characterized fibroblasts. For renal tubule cells, proximal tubule (PT) cells were labeled by CUBN, ALDOB, LRP2, and SLC13A1. SLC8A1, CALB1, and SLC12A3 were used to define DT cells. UMOD, SLC12A1, and CLDN16 were applied to identify LOH cells. High expression of AQP2 and AQP3 indicated principal cells (PC), while SLC26A7, SLC4A1, and ATP6V1G3 were used to designate IC. SMC distinctlyexpressed TAGLN, MYH11, MYLK, and ACTA2. Additionally, leukocytes uniquely expressed B cells, MON, and MC marker genes (MS4A1, CD14 and CD68, respectively).
Table 1 -
Cell-cluster-specific marker genes of different cell types in kidney.
|
Cell types
|
Abbreviations
|
Marker genes
|
|
Natural killer Tcells |
NKT |
CD3D, CD3E, KLRD1, NKG7
|
| B cells |
B cells |
MS4A1, CD79A, CD79B
|
| Monocytes |
MON |
LYZ, VCAN, CD14, FCN1
|
| Macrophages |
MC |
LYZ, CD68, MRC1, C1QA, C1QB
|
| Dendritic cells |
DC |
CD1C, CD1E, FCER1A, CLEC10A
|
| Fibroblasts |
FIB |
LUM, DCN, COL1A1
|
| Smooth muscle cells |
SMC |
ACTA2, TAGLN, MYLK, MYH11
|
| Mesangial cells |
MES |
CTGF, MYL9, FHL2, AKAP12
|
| Podocytes |
POD |
NPHS2, PODXL, PTPRO
|
| Endothelial cells |
EC |
PECAM1, VWF, CLDN5, ACKR1
|
| Proximal tubule cell |
PT |
CUBN, SLC13A1, LRP2, ALDOB
|
| Loop of Henle cells |
LOH |
UMOD, SLC12A1, CLDN16
|
| Distal tubule cells |
DT |
SLC12A3, CALB1, SLC8A1
|
| Intercalated cells |
IC |
SLC26A7, SLC4A1, ATP6V1G3
|
| Principal cells |
PC |
AQP2, AQP3
|
Cell-cluster-specific gene expression and enrichment analysis in PT cells from ischemic AKI subjects
To elucidate the molecular differences in ischemic AKI patients, we compared transcriptional changes in major kidney parenchymal cells between ischemic AKI and control subjects. Comprehensive DEGs in cell clusters from the tubules and glomerulus were shown in Datasets 3 [https://links.lww.com/CM9/B534] and 4 [https://links.lww.com/CM9/B535], respectively. Representative DEGs in tubular cells and glomerular intrinsic cells were displayed in Figure 2A,B. The central renal intrinsic cell types involved in the pathogenesis of ischemic AKI were the tubular epithelial cells and EC.
Figure 2: DEGs in distinct cell clusters from kidney between ischemic AKI and control subjects. (A) Representative DEGs in glomerulus from AKI patients in comparison with control subjects. (B) Representative DEGs in tubules from AKI patients compared to control subjects. (C, D) Representative DEGs in innate immune cells and adaptive immune cells, respectively, between AKI patients and controls. pct.exp:proportion of cell expressing gene; AKI: Acute kidney injury; DEGs: Differentially expressed genes; IC: Intercalated cells; LOH: Loop of Henle cells; MC: Macrophages; MES: Mesangial cells; MON: Monocytes; NKT: Natural killer T cells; PC: Principal cells; POD: Podocytes; PT: Proximal tubule cells; DC: Dendritic cells; DT: Distal tubule cells; EC: Endothelial cells.
In particular, our research mainly focused on PT, since PT were most susceptible to ischemic injury and suffered the most damage during the process of AKI. RASSF4 and EBAG9, which have been reported to participate in cell cycle control and apoptosis,[25,26] were overexpressed in PT and not evaluated in kidney pathology before. Up-regulation of pro-apoptotic factors including IER3 and SASH1 in PT of AKI, which have not been implicated in AKI pathogenesis yet, was further validated at the protein level in paraffined sections from ischemic AKI patients by IHC staining [Supplementary material file 3,https://links.lww.com/CM9/B526]. USP47 and SEPTIN7, which have also been implicated in promoting apoptosis,[27,28] were upregulated in PT with AKI. EGR1, which has been reported to participate in cell apoptosis and fibrosis,[29] was highly expressed in PT of AKI at both gene and protein levels [Supplementary material file 3,https://links.lww.com/CM9/B526]. PT with AKI also overexpressed NUB1, an interferon (IFN)-inducible factor, which mediates cell apoptosis.[30]PALM3, belonging to the paralemmin protein family, is associated with toll-like receptor 4 signaling and inflammatory response.[31]PALM3 expression was increased in PT with AKI at transcriptome and protein levels. DEGs upregulated in PT such as PDIA6, ATF6, HSPA5, and DNAJC3, were involved in response to endoplasmic reticulum stress by GO enrichment analysis, whereas ANKRD17, BIRC3, PUM1, and LSM14A were enriched in retinoic acid-inducible gene I (RIG-I) signaling. The transcriptomic finding of PDIA6 was also verified at the protein level in PT with ischemic AKI [Supplementary material file 3,https://links.lww.com/CM9/B526]. Correspondingly, PT displayed increased expression of autophagy-related genes such as S100A11, CLDN1, TMBIM6, RB1CC1, and VMP1, which may be protective compensatory reactions in ischemic AKI. Furthermore, antioxidants (NQO1, GPX1, SOD2, and TXNRD1) and stress response gene GDF15, which may exert renoprotective effects,[32] were elevated in PT. Moreover, DEGs in PT between ischemic AKI and control subjects were mainly enriched in response to endoplasmic reticulum stress, regulation of apoptotic signaling pathway, as well as interleukin (IL)-12-mediated signaling pathway [Figure 3A–D].
Figure 3: Enrichment analysis in distinct cell clusters from kidney between ischemic AKI and control subjects. (A, C) GO and KEGG enrichment analysis indicated upregulated DEGs involved in biological processes or signaling pathways in different kidney cells. The left side of the circle represented distinct cell clusters, whereas the right side represented distinct biological processes or signaling pathways. The inner circle represented gene numbers enriched in cell clusters or biological processes and signaling pathways. (B, D) The bubble chart exhibited the GO and KEGG enrichment analysis of PT comparing AKI patients to control subjects. count and number of genes annotated to GO terms or KEGG pathways; AKI: Acute kidney injury; DEGs: Differentially expressed genes; GO: Gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; IC: Intercalated cells; LOH: Loop of Henle cells; MC: Macrophages; MON: Monocytes; NKT: Natural killer T cells; PC: Principal cells; POD: Podocytes; PT: Proximal tubule cells; DC: Dendritic cells; DT: Distal tubule cells; EC: Endothelial cells.
We also constructed renal I/R injury mice model and verified several hub genes in mice kidney at the protein level by IHC staining. AKI mice displayed elevated expression of USP47, RASSF4, and EBAG9 compared with control mice. Consistent with the results of human kidney of AKI, we found IER1, EGR1, SASH1, and PALM3 protein expressions were increased in kidney of I/R injury mice compared with control mice. However, there was no significant difference in the expressions of RASSF4 and PDIA6 between the two groups [Supplementary material file 4, https://links.lww.com/CM9/B526].
Cell-cluster-specific gene expression and enrichment analysis in other renal intrinsic cells from ischemic AKI subjects
As listed in Figure 3A,C, DT of AKI displayed overexpressed genes including PDIA6, HSPA5, CALM1, and FOS, involved in protein processing in endoplasmic reticulum, estrogen signaling, and IL-17 signaling pathway, whereas DEGs such as CTSB, USP47, TMBIM6, and CD63 upregulated in LOH were enriched in protein kinase A signaling, integrin-mediated signaling pathway, and response to endoplasmic reticulum stress. Comparison of DEGs in PC between ischemic AKI and control subjects were enriched in nucleotide-binding and oligomerization domain (NOD)-like receptor signaling, IL-17 signaling, and response to endoplasmic reticulum stress, of which TXNIP, IL17RB, and ATF3 were involved. In addition, IC with ischemic AKI had elevated expression of HIF1A, JAK1, and DDIT3, which contribute to Th17 cell differentiation, IL-12-mediated signaling, and response to endoplasmic reticulum stress, respectively. Detailed data of enrichment analysis in kidney between ischemic AKI and healthy subjects were illustrated in Datasets 5 [https://links.lww.com/CM9/B536] and 6 [https://links.lww.com/CM9/B537].
Because few cell numbers of glomerulus especially MES and POD obtained from public control group, we used kidney datasets from two healthy living donors as control. The two healthy control kidney specimens were obtained from renal transplant donors by needle biopsy at the Department of Nephrology in Xiangya Hospital, Central South University, and the specimens were harvested between the time of removal from donors and implantation into recipients. As shown in Figure 2B, EC of ischemic AKI had elevated expression of IL-32, which has pro-inflammatory properties and can trigger endoplasmic reticulum stress.[33]TSPAN1, promoting endoplasmic reticulum stress and apoptosis,[34] was upregulated in EC with ischemic AKI. EC had upregulated genes including cell adhesion molecules SPP1, interferon-induced factor (IFI6), apoptosis and lipid metabolism regulator (PLIN2), and chemoattractant (CXCL14). TM4SF1, a key regulator of EC function important for growth and motility,[35] was reduced in EC of ischemic AKI. The complement-inhibitory protein CD59 was decreased in the EC of AKI. The down-regulation of CD59 was involved in boosted susceptibility to hypoxic conditions and complement activation.[36]
Cell-cluster-specific gene expression and enrichment analysis in immune cells of kidney from ischemic AKI subjects
Immune cells also participate in the development and misguided repair of ischemic AKI. Several types of leukocytes in the kidney of ischemic AKI were determined by DEGs, including MC, MON, dendritic cells, NK-T cells, and B cells. The innate immune cells could be further subdivided into classical MON, non-classical MON, MC, proliferating MC, MC-IL1Bhigh S100A8high, MC-IL1Bhigh, and DC-IL1Bhigh [Supplementary material files 5A–C, https://links.lww.com/CM9/B526]. The DEGs dataset of leukocytes was shown in Figure 2C,D and Dataset 7 [https://links.lww.com/CM9/B538].
MC in AKI highly expressed PDIA3 plays critical roles in promoting oxidative stress, inflammation, and apoptosis.[37]SPP1, which has been shown to correlate with macrophage infiltration and M2 polarization,[38] was elevated in MC with ischemic AKI. Chemokines including CCL3, CCL4, CCL4L2, and CXCL2, which induce chemotaxis and adhesion of immune cells, were upregulated in MC of AKI. MON displayed increased expression of leukocyte-recruiting chemokines (CCL4, CCL4L2, CXCL14, CXCL12, and CXCL3). DEGs overexpressed in MON with ischemic AKI such as PLIN2, ACSL1, and ME1 were involved in PPAR signaling pathway. CD84, which promotes T-B cell adhesion and macrophage activation,[39] was upregulated in MON. Enrichment of genes in dendritic cells of ischemic AKI mainly participated in NF-kappa B signaling, toll-like receptor signaling, IL-17 signaling as well as NOD-like receptor signaling pathway. NKT cells highly expressed inflammatory regulators such as PALM3 and CD46, responsible for the regulation of inflammatory cytokine production and complement activation. Additionally, B cells in ischemic AKI had increased expression of genes which facilitate neutrophil and monocyte recruitment and contribute to inflammation of the injured kidney.
Cell–cell communication in kidney from ischemic AKI subjects
Next, we performed ligand–receptor interaction analysis to determine the intercellular crosstalk of distinct cells in kidney from ischemic AKI subjects. The potential interactions of ligands and receptors between different cell clusters of ischemic AKI were displayed in Figure 4A–I. The intercellular interplays were mainly involved in MC or MON with other kidney resident cells.
Figure 4: Ligand–receptor interactions between different cell types in kidney of ischemic AKI patients. (A) Ligand–receptor interactions between different cell types in kidney of AKI subjects, and the frequency of cell–cell communication ranged from low (blue) to high (red). (B–I) Visualized diagrams indicated representative ligand–receptor interactions in MC-MON, MC-EC, MC-MES, MON-DC, MON-DT, MON-EC, EC-MES, and MES-PC, respectively. Lines represented interrelations between the ligand and corresponding receptor. Genes depicted in blue represented ligands, and those displayed in red were receptors. Only AKI patients (n = 2) were analyzed. AKI: Acute kidney injury; MC: Macrophages; MES: Mesangial cells; MON: Monocytes; IC: Intercalated cells; LOH: Loop of Henle cells; PC: Principal cells; POD: Podocytes; PT: Proximal tubule cells; DC: Dendritic cells; DT: Distal tubule cells; EC: Endothelial cells; SMC: Smooth muscle cells; FIB: Fibroblasts.
MC expressed chemokines CCL2, CCL5, CXCL1, and CXCCL8 interacted with the atypical receptor ACKR1 in EC, which facilitate monocyte and neutrophil recruitment to injured kidney and inflammatory response.[40] Also, chemokines including CCL3, CCL3L1, and CCL8 expressed by dendritic cells or MC interacted with receptors CCR2 and CCR1 expressed by MON. TNFRSF14 expressed by MC, which can induce cytokine production and inflammatory response, interacted with MIF expressed in MES and LOH. PTPRC, an essential modulator of T and B cell antigen receptor-mediated activation,[41] was expressed by dendritic cells, interacting with MRC1 expressed by MC. MON-derived ligand CD44, which participates in the adhesion of renal resident cells and leukocytes,[42] interplayed with adhesion molecule SELE expressed by EC. Besides, fibroblasts expressed TGFB1, TGFB2, and TGFB3, which may promote inflammation and fibrosis via transforming growth factor beta (TGF-β) signaling, interacted with TGFBR3 expressed in EC.
Studies have suggested that Notch signaling activation was elevated in both acute and chronic kidney injuries.[43] Accordingly, we observed JAG1/Notch signaling was expressed between distinct cell clusters in ischemic AKI. The ligand JAG1 expressed by PC or MES became interaction pair of NOTCH2 expressed by MON. Notch receptor NOTCH4 expressed in EC interacted with JAG1 expressed in MES. The complete data of ligand–receptor interactions between kidney cells from ischemic AKI subjects were shown in Dataset 8 [https://links.lww.com/CM9/B539].
Discussion
This study comprehensively presents the single-cell transcriptome atlas of kidney specimens from human ischemic AKI through unbiased scRNA-seq analysis. We analyzed cell-cluster-specific gene expression and signaling pathways in kidneys from ischemic AKI and control subjects, thereby providing important findings and clues for further research. Particularly, the injured PT exhibited a proapoptotic and proinflammatory phenotype. Overexpression of genes in PT with ischemic AKI was involved in apoptosis, endoplasmic reticulum stress, autophagy, and RIG-1 signaling. The DEGs in other renal resident cells were mainly enriched in inflammatory signatures including NF-kappa B signaling, toll-like receptor signaling, IL-17 signaling as well as IL-12-mediated signaling. Moreover, the cell–cell communication analysis in ischemic AKI underlined the extensive crosstalk of immune cells especially MC and MON.
PT exhibited different transcriptome phenotypes in control and ischemic AKI, which reflects the crucial pathological mechanisms of ischemic AKI. As an important parenchymal cell type, PT emerge as not only passive victim, but also active perpetrator during the pathophysiology of AKI. Emerging evidence indicates that PT are especially susceptible to ischemia, and acute renal tubular cells' injury promotes the initiation and progress of AKI.[44] PT of AKI had up-regulation of novel pro-apoptotic genes including USP47, RASSF4, EBAG9, IER3, SASH1, SEPTIN7, and NUB1, which have not been recognized in ischemic AKI pathogenesis yet. Further, PT highly expressed endoplasmic reticulum stress-related genes, as prolonged endoplasmic reticulum stress promotes apoptosis in AKI.[45] Autophagy is cytoprotective in response to multiple stresses, and it can remove damaged macromolecules and organelles in AKI.[46] Notably, autophagy-related genes were elevated in PT from ischemic AKI patients, which maybe a protective reaction. Interestingly, RIG-1 signaling participating in inflammatory response[47] was upregulated in PT, which has not yet been explored in pathogenesis of ischemic AKI. PALM3 was overexpressed in PT of ischemic AKI, indicating the potential pathogenic role of PALM3 in ischemic AKI. We verified several upregulated genes including IER3, EGR1, SASH1, PALM3, PDIA6 at the protein level in ischemic AKI patients. Moreover, USP47 and EBAG9 protein expressions were elevated in renal I/R injury mice. Thus, the PT played a pivotal role in the initiation and progression of ischemic AKI.
Glomerular cells, such as EC, upregulated DEGs participating in cell adhesion, apoptosis, and chemotaxis. Also, key regulators for function of normal EC were reduced in EC with ischemic AKI. These findings indicate that EC dysfunction may be involved in pathogenesis of ischemic AKI. Other tubular cells including DT, IC, and PC also increased the expression of genes participating in endoplasmic reticulum stress, apoptosis, neutrophil activation as well as Th17 cell differentiation. Moreover, DEGs overexpressed in tubular cells were primarily enriched in NOD-like receptor signaling, estrogen signaling, IL-12-mediated signaling, and IL-17 signaling. NOD-like receptors serve as intracellular pattern recognition receptors which exclusively recognize pathogen-associated molecular patterns and engage in innate immune responses.[48] Depletion of NLRP3 inflammasome, a NOD-like receptors family member, dramatically protected mice from ischemic AKI.[49] Particularly, Th17 cell and IL-17 signaling promote inflammatory cascades and emerge as key players in the pathogenesis of AKI.[50] Recent studies have shown that IL-17 over-release induces leukocyte migration and kidney oxidative stress and disrupted the integrity of kidney in AKI.[51] Instead, the roles of other pathways including estrogen signaling and IL-12-mediated signaling in the development of ischemic AKI need to be explored by further research.
Current studies have shown that both innate and adaptive immunities participate in mediating renal tubular epithelial cells' injury and recovery in AKI.[52] Soon after tubular epithelial or EC damage, activation and recruitment of inflammatory leukocytes occur in AKI. Previously, studies indicated that immune cells including MON/MC, dendritic cells, T lymphocytes, and B lymphocytes were pivotal in the development of AKI.[53] Consistently, dataset integration identified MC, NK-T cells, MON, dendritic cells, and B cells in AKI kidneys. Infiltrating MC highly expressed genes involved in apoptosis and oxidative stress (PDIA3), leukocytes activation (PDIA3, SPP1), as well as chemotaxis (CCL3, CCL4, CCL4L2, and CXCL2), whereas MON expressed responsible for PPAR signaling pathway, cell adhesion, and leukocyte-recruiting. NK-T cells overexpressed genes such as PALM3 and CD46 responsible for inflammatory cytokines production and complement activation. Our observation indicates that infiltrating immune cells contribute to the inflammatory signature production and fuel the inflammatory phase in ischemic AKI.
The importance of cell–cell communication in the pathogenesis of AKI has been reported.[54] We focused on the regulation of inflammation between distinct cells in kidneys in ischemic AKI through ligand–receptor interactions analysis. Our findings indicate that the interplay between distinct kidney cell clusters in ischemic AKI was most significant in the MC or MON with other cell populations in kidney. MC or dendritic cells expressed several types of chemokines that interacted with the receptors in EC or MON. These interactions facilitate leukocyte recruitment to the injured kidney and propagate inflammatory responses. The ligand CD44 expressed by MON interacts with the adhesion molecule SELE expressed by EC, which participates in the adhesion of resident cells and leukocytes. Notch signaling serves as a highly conserved cell–cell crosstalk mechanism which modulates tissue homeostasis, inflammation, and repair.[55] Our data also indicated that JAG1/Notch signaling was expressed between distinct cell clusters in kidney through ligand–receptor interactions.
This research had several limitations. First, the number of patients involved in this study was relatively small, and increasing the sample size is needed to reduce individual differences. Second, a small part from the needle biopsy specimens might not be enough to represent the pathological damage of the whole kidney, due to the heterogeneity of the kidney damage in ischemic AKI. Third, despite verifying several molecules at a protein level in ischemic AKI patients, further animal and in vitro studies need to be performed to validate these scRNA-seq findings.
Overall, this study revealed a comprehensive transcriptional profile of distinct cell types in kidney of patients with ischemic AKI. Moreover, we identified novel hub genes, signaling pathways, and cell–cell crosstalk, which indicated relevant molecular mechanisms related to the development of ischemic AKI. These findings offered insights into pathophysiological mechanisms and provide potential treatment targets for ischemic AKI. These data in our work will be further validated through in vitro and in vivo studies.
Funding
This work was supported by the National Key Research and Development Program of China (No. 2020YFC2005000), the Key Research and Development Program of Hunan province (No. 2020WK2008), the science and technology innovation Program of Hunan Province (No. 2020RC5002), the Natural Science Foundation of Hunan Province (Nos. 2022JJ30070, 2021JJ31130 and 2021JJ31057), the Project of Health Commission of Hunan Province (Nos. A202303050036 and 202104101009), "Yiluqihang Shenmingyuanyang" medical development and Scientific Research Fund project on Kidney Diseases (No. SMYY20220301001).
Conflicts of interest
None.
References
1. Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract 2012;120: c179–c184. doi: 10.1159/000339789.
2. Mehta RL, Cerda J, Burdmann EA, Tonelli M, Garcia-Garcia G, Jha V, et al. International Society of Nephrology's 0by25 initiative for acute kidney injury (zero preventable deaths by 2025): A human rights case for nephrology. Lancet 2015;385: 2616–2643. doi: 10.1016/S0140-6736(15)60126-X.
3. Hoste EAJ, Kellum JA, Selby NM, Zarbock A, Palevsky PM, Bagshaw SM, et al. Global epidemiology and outcomes of acute kidney injury. Nat Rev Nephrol 2018;14: 607–625. doi: 10.1038/s41581-018-0052-0.
4. Hoste EA, Bagshaw SM, Bellomo R, Cely CM, Colman R, Cruz DN, et al. Epidemiology of acute kidney injury in critically ill patients: The multinational AKI-EPI study. Intensive Care Med 2015;41: 1411–1423. doi: 10.1007/s00134-015-3934-7.
5. Ronco C, Bellomo R, Kellum JA. Acute kidney injury. Lancet 2019;394: 1949–1964. doi: 10.1016/S0140-6736(19)32563-2.
6. Sharfuddin AA, Molitoris BA. Pathophysiology of
ischemic acute kidney injury. Nat Rev Nephrol 2011;7: 189–200. doi: 10.1038/nrneph.2011.16.
7. Bhargava P, Schnellmann RG. Mitochondrial energetics in the kidney. Nat Rev Nephrol 2017;13: 629–646. doi: 10.1038/nrneph.2017.107.
8. Andrews TS, Kiselev VY, McCarthy D, Hemberg M. Tutorial: Guidelines for the computational analysis of
single-cell RNA sequencing data. Nat Protoc 2021;16: 1–9. doi: 10.1038/s41596-020-00409-w.
9. Svensson V, Natarajan KN, Ly LH, Miragaia RJ, Labalette C, Macaulay IC, et al. Power analysis of single-cell RNA-sequencing experiments. Nat Methods 2017;14: 381–387. doi: 10.1038/nmeth.4220.
10. Huang B, Chen Z, Geng L, Wang J, Liang H, Cao Y, et al. Mucosal profiling of pediatric-onset colitis and IBD reveals common pathogenics and therapeutic pathways. Cell 2019;179: 1160–1176.e24. doi: 10.1016/j.cell.2019.10.027.
11. Patel AP, Tirosh I, Trombetta JJ, Shalek AK, Gillespie SM, Wakimoto H, et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science 2014;344: 1396–1401. doi: 10.1126/science.1254257.
12. Han J, DePinho RA, Maitra A.
Single-cell RNA sequencing in pancreatic cancer. Nat Rev Gastroenterol Hepatol 2021;18: 451–452. doi: 10.1038/s41575-021-00471-z.
13. Rao DA, Arazi A, Wofsy D, Diamond B. Design and application of
single-cell RNA sequencing to study kidney immune cells in lupus nephritis. Nat Rev Nephrol 2020;16: 238–250. doi: 10.1038/s41581-019-0232-6.
14. Wilson PC, Wu H, Kirita Y, Uchimura K, Ledru N, Rennke HG, et al. The single-cell transcriptomic landscape of early human diabetic nephropathy. Proc Natl Acad Sci U S A 2019;116: 19619–19625. doi: 10.1073/pnas.1908706116.
15. Liu Y, Hu J, Liu D, Zhou S, Liao J, Liao G, et al. Single-cell analysis reveals immune landscape in kidneys of patients with chronic transplant rejection. Theranostics 2020;10: 8851–8862. doi: 10.7150/thno.48201.
16. Zheng Y, Lu P, Deng Y, Wen L, Wang Y, Ma X, et al. Single-cell transcriptomics reveal immune mechanisms of the onset and progression of IgA nephropathy. Cell Rep 2020;33: 108525. doi: 10.1016/j.celrep.2020.10852.
17. Kirita Y, Wu H, Uchimura K, Wilson PC, Humphreys BD. Cell profiling of mouse acute kidney injury reveals conserved cellular responses to injury. Proc Natl Acad Sci U S A 2020;117: 15874–15883. doi: 10.1073/pnas.2005477117.
18. Legouis D, Ricksten SE, Faivre A, Verissimo T, Gariani K, Verney C, et al. Altered proximal tubular cell glucose metabolism during acute kidney injury is associated with mortality. Nat Metab 2020;2: 732–743. doi: 10.1038/s42255-020-0238-1.
19. Dixon EE, Wu H, Muto Y, Wilson PC, Humphreys BD. Spatially resolved transcriptomic analysis of acute kidney injury in a female murine model. J Am Soc Nephrol 2022;33: 279–289. doi: 10.1681/ASN.2021081150.
20. Xu J, Shen C, Lin W, Meng T, Ooi JD, Eggenhuizen PJ, et al. Single-cell profiling reveals transcriptional signatures and
cell–cell crosstalk in anti-PLA2R positive idiopathic membranous nephropathy patients. Front Immunol 2021;12: 683330. doi: 10.3389/fimmu.2021.683330.
21. Korsunsky I, Millard N, Fan J, Slowikowski K, Zhang F, Wei K, et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat Methods 2019;16: 1289–1296. doi: 10.1038/s41592-019-0619-0.
22. Satija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015;33: 495–502. doi: 10.1038/nbt.3192.
23. Tang R, Meng T, Lin W, Shen C, Ooi JD, Eggenhuizen PJ, et al. A partial picture of the single-cell transcriptomics of human IgA nephropathy. Front Immunol 2021;12: 645988. doi: 10.3389/fimmu.2021.645988.
24. Li H, Chen W, Chen Y, Zhou Q, Xiao P, Tang R, et al. Neferine attenuates acute kidney injury by inhibiting NF-κB signaling and upregulating Klotho expression. Front Pharmacol 2019;10: 1197. doi: 10.3389/fphar.2019.01197.
25. Eckfeld K, Hesson L, Vos MD, Bieche I, Latif F, Clark GJ. RASSF4/AD037 is a potential ras effector/tumor suppressor of the RASSF family. Cancer Res 2004;64: 8688–8693. doi: 10.1158/0008-5472.CAN-04-2065.
26. Sonoda K. Novel therapeutic strategies to target RCAS1, which induces
apoptosis via ectodomain shedding. Histol Histopathol 2011;26: 1475–1486. doi: 10.14670/HH-26.1475.
27. Hu Y, Ma Z, Chen Z, Chen B. USP47 promotes
apoptosis in rat myocardial cells after ischemia/reperfusion injury via NF-kappaB activation. Biotechnol Appl Biochem 2021;68: 841–848. doi: 10.1002/bab.2000.
28. Li R, Dong W, Zhang S, Yang H, Zhang L, Ye Z, et al. Septin 7 mediates high glucose-induced podocyte
apoptosis. Biochem Biophys Res Commun 2018;506: 522–528. doi: 10.1016/j.bbrc.2018.10.081.
29. Ho LC, Sung JM, Shen YT, Jheng HF, Chen SH, Tsai PJ, et al. Egr-1 deficiency protects from renal
inflammation and fibrosis. J Mol Med (Berl) 2016;94: 933–942. doi: 10.1007/s00109-016-1403-6.
30. Hosono T, Tanaka T, Tanji K, Nakatani T, Kamitani T. NUB1, an interferon-inducible protein, mediates anti-proliferative actions and
apoptosis in renal cell carcinoma cells through cell-cycle regulation. Br J Cancer 2010;102: 873–882. doi: 10.1038/sj.bjc.6605574.
31. Li S, Guo L, Zhao Y, Qian P, Lv X, Qian L, et al. Silencing of paralemmin-3 protects mice from lipopolysaccharide-induced acute lung injury. Peptides 2016;76: 65–72. doi: 10.1016/j.peptides.2016.01.001.
32. Liu J, Kumar S, Heinzel A, Gao M, Guo J, Alvarado GF, et al. Renoprotective and immunomodulatory effects of GDF15 following AKI invoked by ischemia-reperfusion injury. J Am Soc Nephrol 2020;31: 701–715. doi: 10.1681/ASN.2019090876.
33. Aass KR, Kastnes MH, Standal T. Molecular interactions and functions of IL-32. J Leukoc Biol 2021;109: 143–159. doi: 10.1002/JLB.3MR0620-550R.
34. Wang J, Huang J, Huang L. TSPAN1 silencing protects against cerulein-induced pancreatic acinar cell injury via targeting AGR2. Drug Dev Res 2022;83: 158–166. doi: 10.1002/ddr.21855.
35. Shih SC, Zukauskas A, Li D, Liu G, Ang LH, Nagy JA, et al. The L6 protein TM4SF1 is critical for endothelial cell function and tumor angiogenesis. Cancer Res 2009;69: 3272–3277. doi: 10.1158/0008-5472.CAN-08-4886.
36. Weinstock C, Anliker M, von Zabern I. CD59: A long-known complement inhibitor has advanced to a blood group system. Immunohematology 2015;31: 145–151. doi: 10.21307/immunohematology-2019-083.
37. Shen J, Zhan Y, He Q, Deng Q, Li K, Wen S, et al. Remifentanil promotes PDIA3 expression by activating p38MAPK to inhibit intestinal ischemia/reperfusion-induced oxidative and endoplasmic reticulum stress. Front Cell Dev Biol 2022;10: 818513. doi: 10.3389/fcell.2022.818513.
38. Zheng Y, Hao S, Xiang C, Han Y, Shang Y, Zhen Q, et al. The correlation between SPP1 and immune escape of EGFR mutant lung adenocarcinoma was explored by bioinformatics analysis. Front Oncol 2021;11: 592854. doi: 10.3389/fonc.2021.592854.
39. Cannons JL, Qi H, Lu KT, Dutta M, Gomez-Rodriguez J, Cheng J, et al. Optimal germinal center responses require a multistage T cell: B cell adhesion process involving integrins, SLAM-associated protein, and CD84. Immunity 2010;32: 253–265. doi: 10.1016/j.immuni.2010.01.010.
40. Girbl T, Lenn T, Perez L, Rolas L, Barkaway A, Thiriot A, et al. Distinct compartmentalization of the chemokines CXCL1 and CXCL2 and the atypical receptor ACKR1 determine discrete stages of neutrophil diapedesis. Immunity 2018;49: 1062–1076.e6. doi: 10.1016/j.immuni.2018.09.018.
41. Al Barashdi MA, Ali A, McMullin MF, Mills K. Protein tyrosine phosphatase receptor type C (PTPRC or CD45). J Clin Pathol 2021;74: 548–552. doi: 10.1136/jclinpath-2020-206927.
42. Ponta H, Sherman L, Herrlich PA. CD44: From adhesion molecules to signalling regulators. Nat Rev Mol Cell Biol 2003;4: 33–45. doi: 10.1038/nrm1004.
43. Sirin Y, Susztak K. Notch in the kidney: Development and disease. J Pathol 2012;226: 394–403. doi: 10.1002/path.2967.
44. Chiba T, Peasley KD, Cargill KR, Maringer KV, Bharathi SS, Mukherjee E, et al. Sirtuin 5 regulates proximal tubule fatty acid oxidation to protect against AKI. J Am Soc Nephrol 2019;30: 2384–2398. doi: 10.1681/ASN.2019020163.
45. Yan M, Shu S, Guo C, Tang C, Dong Z. Endoplasmic reticulum stress in ischemic and nephrotoxic acute kidney injury. Ann Med 2018;50: 381–390. doi: 10.1080/07853890.2018.1489142.
46. Kaushal GP, Shah SV. Autophagy in acute kidney injury. Kidney Int 2016;89: 779–791. doi: 10.1016/j.kint.2015.11.021.
47. Kolmychkova KI, Zhelankin AV, Karagodin VP, Orekhov AN. Mitochondria and
inflammation. Patol Fiziol Eksp Ter 2016;60: 114–121.
48. Platnich JM, Muruve DA. NOD-like receptors and inflammasomes: A review of their canonical and non-canonical signaling pathways. Arch Biochem Biophys 2019;670: 4–14. doi: 10.1016/j.abb.2019.02.008.
49. Kim HJ, Lee DW, Ravichandran K, Keys DO, Akcay A, Nguyen Q, et al. NLRP3 inflammasome knockout mice are protected against ischemic but not cisplatin-induced acute kidney injury. J Pharmacol Exp Ther 2013;346: 465–472. doi: 10.1124/jpet.113.205732.
50. Collett JA, Ortiz-Soriano V, Li X, Flannery AH, Toto RD, Moe OW, et al. Serum IL-17 levels are higher in critically ill patients with AKI and associated with worse outcomes. Crit Care 2022;26: 107. doi: 10.1186/s13054-022-03976-4.
51. Wang F, Yin J, Lin Y, Zhang F, Liu X, Zhang G, et al. IL-17C has a pathogenic role in kidney ischemia/reperfusion injury. Kidney Int 2020;97: 1219–1229. doi: 10.1016/j.kint.2020.01.015.
52. Radi ZA. Immunopathogenesis of acute kidney injury. Toxicol Pathol 2018;46: 930–943. doi: 10.1177/0192623318799976.
53. Kinsey GR, Okusa MD. Role of leukocytes in the pathogenesis of acute kidney injury. Crit Care 2012;16: 214. doi: 10.1186/cc11228.
54. Rudman-Melnick V, Adam M, Potter A, Chokshi SM, Ma Q, Drake KA, et al. Single-cell profiling of AKI in a murine model reveals novel transcriptional signatures, profibrotic phenotype, and epithelial-to-stromal crosstalk. J Am Soc Nephrol 2020;31: 2793–2814. doi: 10.1681/ASN.2020010052.
55. Mukherjee M, Fogarty E, Janga M, Surendran K. Notch signaling in kidney development, maintenance, and disease. Biomolecules 2019;9: 692. doi: 10.3390/biom9110692.
