Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Apr 1.
Published in final edited form as: Kidney Int. 2021 Dec 22;101(4):779–792. doi: 10.1016/j.kint.2021.11.031

Glomerular endothelial cell-podocyte stresses and crosstalk in structurally normal kidney transplants

Rajasree Menon 1, Edgar A Otto 2, Celine C Berthier 2, Viji Nair 2, Evan A Farkash 3, Jeffrey B Hodgin 3, Yingbao Yang 3, Jinghui Luo 3, Kenneth J Woodside 4, Haniyeh Zamani 5, Silas P Norman 2, Roger C Wiggins 2, Matthias Kretzler 2, Abhijit S Naik 2
PMCID: PMC9067613  NIHMSID: NIHMS1766451  PMID: 34952098

Abstract

Increased podocyte detachment begins immediately after kidney transplantation and is associated with long-term allograft failure. We hypothesized that cell-specific transcriptional changes in podocytes and glomerular endothelial cells after transplantation would offer mechanistic insights into the podocyte detachment process. To test this, we evaluated cell-specific transcriptional profiles of glomerular endothelial cells and podocytes from 14 patients of their first-year surveillance biopsies with normal histology from low immune risk recipients with no post-transplant complications and compared these to biopsies of 20 healthy living donor controls. Glomerular endothelial cells from these surveillance biopsies were enriched for genes related to fluid shear stress, angiogenesis, and interferon signaling. In podocytes, pathways were enriched for genes in response to growth factor signaling and actin cytoskeletal reorganization but also showed evidence of podocyte stress as indicated by reduced nephrin (adhesion protein) gene expression. In parallel, transcripts coding for proteins required to maintain podocyte adherence to the underlying glomerular basement membrane were downregulated, including the major glomerular podocyte integrin α3 and the actin cytoskeleton-related gene synaptopodin. The reduction in integrin α3 protein expression in surveillance biopsies was confirmed by immunoperoxidase staining. The combined growth and stress response of patient allografts post-transplantation paralleled similar changes in a rodent model of nephrectomy-induced glomerular hypertrophic stress that progress to develop proteinuria and glomerulosclerosis with shortened kidney life span. Thus, even among patients with apparently healthy allografts with no detectable histologic abnormality including alloimmune injury, transcriptomic changes reflecting cell stresses are already set in motion that could drive hypertrophy-associated glomerular disease progression.

Keywords: endothelial cells, glomerulosclerosis, integrin, kidney transplantation, podocytes

A reduction in nephron mass from any reason is known to cause glomerular hyperfiltration, hypertrophy, progressive proteinuria, glomerulosclerosis, and eventually leading to end-stage kidney disease. Allografts constitute a particular case wherein the supplied nephron mass is typically lower than that in the original 2-kidney state. This lower nephron mass relative to the size of the recipient’s body surface area is strongly associated with progressive proteinuria, increased glomerulosclerosis, and shorter allograft survival.1–3 Although chronic microcirculatory injury undoubtedly contributes to late graft failure, we have speculated that the relative lack of improvement in long-term allograft outcomes despite better immunosuppression may in part be due to nonimmune mechanisms driving late graft failure.4–6

Indication biopsy studies have noted chronic microcirculatory injury to be a key cause of late graft failure.4,6 Despite their importance, these studies were performed in failing allografts at 3 to 7 years post-transplantation. Because death censored allograft failure now exceeds 15 years, these studies do not capture key determinants of graft failure at >10 years post-transplantation in allografts that survive the initial immune-mediated injuries.7 In parallel, 2 studies that used surveillance biopsy data of stable allografts show progressive glomerulosclerosis to be an important cause of graft failure.8,9 Because glomerular disease progression requires podocyte depletion, we have investigated the potential role of podocyte loss in driving late graft failure.

Thus far, we have shown that accelerated podocyte loss, as measured in the urine pellet, begins immediately post-transplantation even in histologically “normal” allografts.5,10 Furthermore, modeling podocyte density reduction after transplantation predicted allograft life span with surprising fidelity.7 In addition, the average first-year rate of podocyte detachment predicted the allograft life at a median of 4.5 years post-transplantation, even in those patients without proteinuria or glomerular disease within the first year.5 In both our short- and long-term data, we observed that recipient-donor body size mismatch (a surrogate for hyperfiltration and hypertrophic stress) at the time of transplantation was associated with the rate of podocyte loss, thereby highlighting the possible role of podocyte hypertrophic stress at the time of transplantation in glomerular disease progression.5,11 However, the molecular mechanisms of what drives podocyte injury and loss early after transplantation remain unclear. Because allografts are inherently accompanied by hyperfiltration,12 we hypothesized that we should observe hyperfiltration-related transcript changes in cells that constitute the allograft glomerulus and evidence of cell-specific changes associated with the podocyte detachment process, even in allografts with no detectable histological abnormality. We, therefore, performed single cell transcriptomic analysis using protocol biopsy samples from allograft recipients who were at low immune risk, had no histological evidence of alloimmune injury, and with no significant post-transplant infectious or immune complications to minimize confounding variables.

METHODS

Tissue procurement and processing

We leveraged the surveillance biopsy program at the University of Michigan, where biopsies are performed at time 0, 3, 6, and 12 months. For this study, the time 0 biopsies from healthy living donors (controls) were obtained at the back table before perfusion. Surveillance biopsies were performed at 3 and 12 months post-transplantation (Figure 1). As part of our established Human Kidney Transplant Transcriptomic Atlas workflow along with standard of care biopsy cores, patients consent for a single research core obtained with a 16-G needle. The research core was collected following a standardized protocol.13,14 The research core is then cut into 3 pieces (~1 mg each), with piece 1 stored in RNAlater solution (Thermo-Fischer Scientific) for bulk transcriptome analysis and piece 2 and piece 3 put into in CryoStor CS10 freeze media (BioLife Solutions) for single cell RNA-seq analysis and for future use including single nuclear studies, respectively. All clinical and research activities being reported are consistent with the principles of the Declaration of Istanbul.

Figure 1 |.

Figure 1 |

Open in a new tab

Graphical representation of the study design.

Single cell processing of the research core

The generation of single cell preparations was accomplished by enzymatic (Liberase TL; Sigma Aldrich) and mechanical dissociation for 12 minutes at 37 °C following our protocol published previously.13 Cells are filtered through a 30 µm strainer and counted, and up to 10,000 viable cells are submitted to the University of Michigan Advanced Genomics Core facility to execute droplet-based high-throughput scRNA-seq on the 10x Genomics Chromium platform. After droplet encapsulation, the 10x Genomics approach allows cell lysis, individual cell RNA molecular barcoding, and reverse transcription. Subsequently, cDNA libraries are generated and sequenced on the NovaSeq 6000 platform (Illumina) as asymmetric paired-end (26 × 151) runs and generating >200 million raw sequence reads per sample. The sequencing data are preprocessed using the 10x Genomics software Cell Ranger. Downstream analysis is performed with the Seurat R package version 3 (Satija Lab).15 A combined analysis of the single cell data sets generated from the different sample sources (living donor and surveillance biopsies) using Seurat version 3 includes the following steps: filtering of cells with <500 genes or >50% mitochondrial content default normalization, scaling based on sample mRNA count and mitochondrial RNA content, dimensionality reduction principal component analysis and uniform manifold approximation and projection, sample integration using the Harmony algorithm, standard unsupervised clustering, and the discovery of differentially expressed cell type–specific markers.

Functional literature-based analyses

Kidney-specific functional analysis was performed using HumanBase (https://hb.flatironinstitute.org/). It applies machine learning algorithms to learn biological associations from massive genomic data collections. Briefly, a tissue-specific functional network for the input genes is generated through regularized Bayesian integration of publicly available expression, physical interaction, and other omics experiments.16 Next, community clustering in the network is performed to identify tightly connected sets of genes using Human-Base.io module detection function.17 Biological literature-based network was built using Genomatix Pathway System software (www.genomatix.de) with the function word level as a minimum evidence level parameter. Major biological themes and upstream regulator analyses were identified using Ingenuity Pathway Analysis software (www.ingenuity.com).

Rat bulk isolated glomerular transcript analysis

The data used for analysis are those previously reported by Affymetrix analysis using a transgenic rat model whose podocytes expressed a modified 4E-binding protein 1 (4E-BP1) transgene under control of the human podocin/NPHS2 promoter. The expressed protein had 2 threonine residues changed to alanine that reduced the capacity of the rat podocyte mechanistic target of rapamycin complex 1 (mTORC1) to drive protein synthesis in response to hypertrophic stress induced by uninephrectomy. The phenotype of these rats is that they more rapidly develop proteinuria and focal segmental glomerular sclerosis (FSGS) in response to nephrectomy-induced hypertrophic stress than do wild-type podocytes. The data used compared the Affymetrix transcriptomic profiles of nephrectomized transgenic rats with those of sham-nephrectomized transgenic rats. Details of the pathophysiological correlates are provided in another study.18

Ligand-receptor interaction analysis

Ligand-receptor (LR) interaction analysis was used to infer inter-cellular communication using the R statistical package Single-CellSignalR.19 SingleCellSignalR uses a curated data set of 3250 LR pairs and generates an LR pair interaction score. To minimize the false-positive rate to <10%, we used an LR pair interaction score of >0.5 to infer intercellular crosstalk.

Immunoperoxidase studies

Integrin A3 (ITGA3) immunoperoxidase staining was performed on formalin-fixed paraffin-embedded kidney biopsies sectioned at 3-µm thickness. These sections were obtained from patients in this study who had tissue blocks available. Sections were deparaffinized, hydrated, and subjected to heated Antigen Retrieval Solution (pH 6.0, Abcam) followed by blocking with 0.3% hydrogen peroxide and then 10% normal goat serum in 5% bovine serum albumin. The tissue was then exposed to primary antibody anti-ITGA3 (HPA008572, Sigma) at 1:100 concentration overnight at 4 °C. The sections were then exposed to goat anti-rabbit Ig-HRP (4010–05, Southern Biotech) at 1:100 concentration for 1 hour at room temperature, and the signal was detected with 3,3′-diaminobenzidine (DAB) followed by hematoxylin counterstain. All immunohistochemistry studies were done in the same batch.

Study approval

The institutional review board approved all the studies. All the biopsies were obtained using the Human Kidney Transplant Transcriptomic Atlas under IRB HUM00150968 (Principal Investigator: ASN).

RESULTS

Population characteristics

Donor characteristics of the healthy living donor cohort.

Twenty samples from 18 living donors were available for analysis, with one of the donors providing a total of 3 samples. The mean age of the cohort was 45 ± 10 years, with 49% being male and 88% being white with a mean glomerular filtration rate of 100.7 ml/min per 1.73 m2. This cohort has been published in a recent report.20

Donor characteristics of protocol surveillance biopsies.

The mean age of donors of surveillance biopsies was 39.4 ± 13 years (range 17–57 years). Fifty percent of kidneys were from living donors (n = 7). Gender information was available in 13 of 14, of whom 8 were males (61.5%). Twelve of 14 donors had race information, with 9 (75%) being white, 1 African American, and 2 of “other” races. None of the donors had diabetes.

Recipient characteristics.

The underlying cause of end-stage kidney disease in the 14 recipients was diverse (Table 1). All 14 recipients received standard triple immunosuppression including tacrolimus, mycophenolate mofetil, and prednisone with no history of rejection, BK polyomavirus viremia, or nephropathy at any point in the study. The clinical biopsy cores were used to obtain detailed Banff scores using the Banff 2017 classification (Table 2).21 The biopsies revealed minimal to no glomerulitis, peritubular capillaritis, or interstitial inflammation. There was no evidence of recurrent glomerular disease, rejection, or infection. The mean estimated glomerular filtration rate at the time of biopsy was 64 ± 15.3 ml/min per 1.73 m2 (range 38–93 ml/min per 1.73 m2). Spot urine protein-to-creatinine ratio at the time of biopsy was 0.18 ± 0.9 g/g (range 0.08–0.34 g/g). The body mass index at the time of biopsy was 31 ± 5.2 kg/m2 (range 23.5–41 kg/m2). Thirteen of 14 recipients were male (93%). Ten of 14 recipients were white, 1 African American, 1 of Hispanic descent, and 2 of “other” races. Thirteen of 14 had class I and class II panel reactive antibodies <20 at the time of transplantation. These data were missing for 1 patient. None of the patients were on angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker therapies at the time of biopsy. Six of 14 patients were defined as having no diabetes after transplantation. Of these 6, 1 had type 1 diabetes but received a simultaneous kidney and pancreas transplant, thereby curing their diabetes. Eight patients had a diabetic milieu, including 7 who had diabetes before transplantation and 1 who developed post-transplant diabetes mellitus. Diabetic recipients were significantly older (age 57.9 years vs. 37.9 years), had higher proteinuria (urine protein-to-creatinine ratio 0.22 g/g vs. 0.11 g/g), and lower estimated glomerular filtration rate at the time of biopsy (54.1 ml/min per 1.73 m2 vs. 77.3 ml/min per 1.73 m2). There were no differences in body mass index, biopsy timing (3 months vs. 12 months), or living versus deceased donors by diabetes status.

Table 1 |.

Recipient cause of ESKD, type of donor kidney, and diabetic and nondiabetic milieus after kidney transplantation

Patient no. Cause of ESKD in the recipient Living or deceased donor kidney DM or nondiabetic milieu after transplantation
1 Reflux nephropathy DD No DM
2 DM type 2 DD DM
3 DM type 2 LD DM
4 HTN, DM type 1 LD DM
5 DM type 1 DD (SPK) No DM
6 Goodpasture, DM type 2 LD DM
7 FSGS DD DM
8 HTN DD No DM
9 Hepatitis C, GN LD DM
10 Unknown DD No DM
11 PKD LD No DM
12 DM type 2 DD DM
13 IgA nephropathy LD No DM
14 PKD, DM type 2 LD DM
Open in a new tab

DD, deceased donor; DM, diabetes mellitus; ESKD, end-stage kidney disease; FSGS, focal segmental glomerular sclerosis; GN, glomerulonephritis; HTN, hypertension; LD, living donor; PKD, polycystic kidney disease; SPK, simultaneous pancreas and kidney.

Table 2 |.

Banff 2017 scoring of 3- and 12-mo surveillance biopsies

ID g cg mm i ti t ci ct v cv Ah ptc c4d
1 0 0 0 0 0 0 0 0 0 0 0 0 0
2 0 0 0 0 0 0 0 0 0 2 2 0 0
3 1 0 0 0 0 0 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0 1 0 1 1 0 0
5 0 0 0 0 0 0 0 0 0 0 0 0 0
6 0 0 0 0 0 0 1 1 0 0 0 0 0
7 0 0 0 0 0 0 1 0 0 2 1 0 0
8 0 0 0 0 0 0 0 0 0 0 0 0 0
9 0 0 0 0 0 0 0 0 0 0 0 0 0
10 0 0 0 0 0 0 1 0 0 2 0 0 0
11 0 0 0 0 0 0 0 0 0 0 0 0 0
12 0 0 1 0 0 1 0 1 0 1 0 1 0
13 0 0 0 0 0 0 0 1 0 0 1 0 0
14 0 0 0 0 0 0 0 0 0 0 0 0 0
Open in a new tab

ah, arteriolar hyalinosis; c4d, c4d score in peritubular capillaries; cg, transplant glomerulopathy; ci, interstitial fibrosis; ct, tubular atrophy; cv, arterial fibrointimal thickening; g, glomerulitis; i, interstitial inflammation; mm, mesangial matrix increase; ptc, peritubular capillary inflammation; t, tubulitis; ti, total inflammation; v, vascular inflammation.

Single cell analysis

Figure 1 shows the study overview. A total of 40,254 cells were analyzed, including 22,964 cells from healthy living donors at the time of transplantation and 17,290 cells from surveillance biopsies obtained within the first post-transplant year. At a resolution of 1.4 (parameter for cluster granularity) in the unsupervised clustering analysis, we defined a total of 29 clusters (Figure 2a). All clusters were well represented in both healthy living donor biopsies (Figure 2b) and surveillance biopsies (Figure 2c).

Figure 2 |. Uniform manifold approximation and projection (UMAP) plots resulted from the unsupervised clustering using Seurat R package version 3.

Figure 2 |

Open in a new tab

(a) Unsupervised clustering of a total of 40,254 cells resulted in 29 clusters at a resolution of 1.4 (parameter for cluster granularity). (b) UMAPs of 22,964 cells from healthy living donors and 17,290 cells from surveillance biopsies. ATL, ascending thin limb; CNT, connecting tubule; DCT, descending thin limb; EC1, endothelial cell cluster 1; EC2, endothelial cell cluster 2; EC3, endothelial cell cluster 3; FIB, fibroblast; IC1, intercalated cells cluster 1; IC2, intercalated cells cluster 2; MC, mesangial cells; MYL, myeloid cells; NKT, natual killer T cells; PC1, principal cells cluster 1; PC2, principal cells cluster 2; PEC, parietal epithelial cells; POD, podocytes; PT1, proximal tubule cluster 1; PT2, proximal tubule cluster 2; PT3, proximal tubule cluster 3; TAL1, thick ascending limb cluster 1; TAL2, thick ascending limb cluster 2; vSMC, vascular smooth muscle cells.

Glomerular endothelial cells

A total of 993 cells were identified as glomerular endothelial cells (GECs). EH Domain-Containing 3 (EHD3) expression was used to confirm the identity of GECs and distinguish them from peritubular capillary endothelial cells and arteriolar endothelial cells as previously reported (Figure 3a).14 Using the FindMarkers functionality embedded in the Seurat R package, we performed differential expression analysis between histologically normal surveillance biopsies and healthy living donor biopsies. We also assessed the transcriptional response of GECs in the context of recipient diabetes status. We then performed kidney-specific functional module analyses using the HumanBase web-based tool to evaluate the differentially expressed genes. This included 273 genes in all surveillance biopsies versus controls, 323 genes in nondiabetic allografts versus controls, 368 genes in diabetic allografts versus controls, and 245 genes in diabetic allografts versus nondiabetic allografts. The top 5 enriched pathways (Figure 4a) included (i) fluid shear stress response, (ii) angiogenesis, (iii) interferon-related Janus kinase–signal transducer and activator of transcription (STAT) signaling, (iv) regulation of cell proliferation, and (v) release of sequestered calcium in the cytosol. Importantly, “response to fluid shear stress” pathway was a common feature of allografts in both a diabetic and a nondiabetic milieu (Figure 4b and c), suggesting that all allografts are undergoing glomerular endothelial shear stress as would be expected from hyperfiltration that uniformly accompanies all allografts.12 We further investigated the similarities and differences in allografts in a diabetic versus nondiabetic milieu (Figure 4d). Because transplantation is a common factor in both groups, this analysis would be expected to identify key diabetes/hyperglycemia-related genes. We identified lower interferon-related Janus kinase–STAT signaling but an increase in genes related to energy utilization in diabetic versus nondiabetic allograft GECs.

Figure 3 |. Dot plots showing the expression pattern of cluster-specific genes.

Figure 3 |

Open in a new tab

(a) Dot plot showing the expression of PLVAP, EH Domain-Containing 3 (EHD3), and SERPINE2 in the 3 endothelial clusters identified: EC1, EC2, and EC3. The dot plot indicates EC2 with a large number of cells expressing EHD3. Prior reports document the glomerular endothelial cell–specific expression of EHD3.14 (b) Dot plot showing the podocyte-specific expression of NPHS1, NPHS2, and SYNPO. ATL, ascending thin limb; CNT, connecting tubule; DCT, descending thin limb; EC1, endothelial cell cluster 1; EC2, endothelial cell cluster 2; EC3, endothelial cell cluster 3; FIB, fibroblast; IC1, intercalated cells cluster 1; IC2, intercalated cells cluster 2; MC, mesangial cells; MYL, myeloid cells; NKT, natual killer T cells; PC1, principal cells cluster 1; PC2, principal cells cluster 2; PEC, parietal epithelial cells; POD, podocytes; PT1, proximal tubule cluster 1; PT2, proximal tubule cluster 2; PT3, proximal tubule cluster 3; TAL1, thick ascending limb cluster 1; TAL2, thick ascending limb cluster 2; vSMC, vascular smooth muscle cells.

Figure 4 |. Functional network modules generated by HumanBase analysis for the genes differentially expressed in glomerular endothelial cells.

Figure 4 |

Open in a new tab

(a) All surveillance biopsies versus healthy controls. (b) Nondiabetic allografts versus healthy controls. (c) Diabetic allografts versus healthy controls. (d) Diabetic allografts versus nondiabetic allografts. Each blob is a gene—the node size (blob size) corresponds to the degree of the node in the network (number of edges connecting it to other genes). So, larger the blob, the more connections it has with other genes in that module. ATP, adenosine triphosphate.

Podocytes

A total number of 179 podocytes were identified by WT1, NPHS1, and NPHS2 gene expression (Figure 3b). Differential expression gene studies between all surveillance biopsies and healthy controls shown in Supplementary Figure S1 revealed the most enriched pathways to be (i) response to growth factor, (ii) actin filament reorganization, (iii) nucleocyto-plasmic transport, and (iv) positive regulation of protein transport. Similar to previously published data, we observed transcriptional evidence of podocyte hypertrophic stress as noted by an increased NPHS2/NPHS1 ratio driven primarily by a relative reduction in NPHS1 expression (vs. NPHS2) (Figure 5c and d) that is known to accompany podocyte stress and injury.22 Furthermore, we observed a significant reduction in the average expression of SYNPO (Figure 5b) after transplantation. Because we have previously observed an increased rate of podocyte loss after transplantation, we evaluated the effect of transplantation on molecules that are responsible for maintaining the podocyte attachment to the underlying glomerular basement membrane. In surveillance biopsies, we observed a significant reduction in ITGA3 expression (but not ITGB1) that codes for the laminin-binding integrin heterodimer protein ITGα3β1, which is the major integrin responsible for keeping podocytes attached to the glomerular basement membrane (Figure 5e and f). This was also confirmed at a protein level by ITAG3 immunoperoxidase staining of tissue sections at 3 months (vs. time 0) in one of the study patients who had both samples (Figure 6).

Figure 5 |. Podocyte-specific gene expression patterns in surveillance biopsies and healthy living donors.

Figure 5 |

Open in a new tab

(a–f) Plots showing the expression of Wilms Tumor 1 (WT1), Synaptopodin (SYNPO), Nephrin (NPHS1), Podocin (NPHS2), Integrin A3 (ITGA3), and Integrin B1 (ITGB1).

Figure 6 |. Glomerular ITGA3 protein expression.

Figure 6 |

Open in a new tab

(a,a′) Immunoperoxidase staining for Integrin A3 (ITGA3) reveals strong cytoplasmic staining of podocytes in the glomeruli of kidney biopsies at the time of transplantation. (b,b′) At 3 months post-transplantation (post-TP), the staining intensity indicates a significant reduction of ITGA3 expression in podocytes. Pictures were taken at original magnification ×40. Bars = 50 μm in length.

Functional literature-based analysis, major themes from regulated genes, and upstream regulator analysis of GECs

GECs.

Literature-based network (Genomatix Pathway System software) analysis from the genes regulated in the GECs of all surveillance biopsies compared with all healthy controls show the 100 best connected genes (Figure 7a). Similar to the differentially expressed gene analysis of GECs, the most interconnected genes are those associated with angiogenesis and interferon-γ and STAT1. We then used Ingenuity Pathway Analysis to identify major themes and their relationships from the genes regulated in the GECs of all surveillance biopsies (vs. living donors) (Figure 7b). Upstream regulator analysis was performed to identify which upstream pathways can explain the differentially regulated gene sets observed in GECs between healthy living donor biopsies and all surveillance biopsies (Figure 7c). We identified the cytokine IFNG (interferon-γ) as the top upstream regulator together with the transcription factor IRF3 (interferon regulatory factor 3). Upstream regulator analysis of allografts in a nondiabetic milieu identified similar regulators operating as above. In contrast, among allografts in a diabetic milieu, we identified interleukin-6 as the top cytokine together with growth factors Vascular endothelial growth factor A (VEGFA) and Fibroblast growth factor 2 (FGF2).

Figure 7 |. Functional literature-based analyses.

Figure 7 |

Open in a new tab

(a) Literature-based network (Genomatix Pathway System software) analysis from the genes regulated in the glomerular endothelial cells (GECs) of all surveillance biopsies compared with all healthy controls. The image displays the top 100 best connected genes co-cited in PubMed abstracts in the same sentence linked by a function word (most relevant genes/interactions). Orange represents the genes that are upregulated and green represent the genes that are downregulated in all biopsies compared with all controls. (b) Ingenuity Pathway Analysis (IPA) used to identify major biological themes and their relationships from the genes regulated in the GECs of all surveillance biopsies compared with all healthy controls. (c) IPA was used to identify a mechanistic network from interferon-γ (IFNG) as an upstream regulator from the genes regulated in the GECs of all surveillance biopsies (combined diabetic and nondiabetic) compared with all healthy controls.

Podocytes.

Comparing the differentially expressed gene sets between podocytes in surveillance biopsies versus controls, we identified NRCP1, Oncostatin M (OSM), and DEAD box helicase 5 (DDX5) as top upstream regulators.

Inferences from LR interaction studies

LR interaction analysis was performed to infer intercellular communication mechanisms. We specifically chose to focus on the podocyte–GEC interaction given their spatial proximity and known importance.23 Surveillance biopsies were separated by diabetes status because we expected possible differences caused by exposure to a diabetic milieu. We examined (i) GEC ligand to podocyte receptors and (ii) podocyte ligand to GEC receptors as shown below.

GECs (ligand expression) and potential crosstalk with podocytes (receptor expression).

In controls, we identified 14 LR pairs (GEC ligands and corresponding podocyte receptors) with interaction scores of >0.5. In contrast, we identified 47 and 44 LR pairs, respectively, in surveillance biopsies obtained from allografts in a nondiabetic and diabetic milieu (Figure 8). This suggests an increase in the bidirectional GEC and podocyte crosstalk occurring after kidney transplantation. In GECs exposed to a nondiabetic milieu, we observed an increase in ligand expression that have putative receptors in podocytes. The receptors identified in podocytes are involved in mediating cell-cell and cell–extracellular matrix interactions (ITGA8), tyrosine kinase cell surface receptors (DDR1), cell growth regulation (GPC1), uptake of various macromolecules such as lipoproteins, sterols, and hormones (LRP6), and regulation of apoptosis (TNFRSRF11B) by the activation of the tumor necrosis factor-α/nuclear factor κB pathway. In podocytes exposed to a diabetic milieu, there was an increased receptor expression for G protein–coupled adenosine receptor 1 activity (ADORA1), tyrosine kinase receptor responsible for the downstream activation of the Akt/mTORC1 pathway (AXL), and increased fibroblast activation (FGFR2) signaling.

Figure 8 |. Putative ligand-receptor interactions in healthy living donors and allografts in a nondiabetic and diabetic milieu.

Figure 8 |

Open in a new tab

An interaction score of 0.5 was set as the minimum threshold. The light blue to dark pink color scale indicates the range of the interaction score where the light blue is 0.5 and dark pink is >0.5. In each case, the top arches in all figures are the putative ligands and the bottom arches are their putative receptors. Interactions between putative glomerular endothelial cell ligands and their podocyte receptors in healthy controls, allografts in a nondiabetic milieu, allografts in a diabetic milieu are shown in (a), (b), and (c), respectively. Interactions between putative podocyte ligands and their glomerular endothelial cell receptors in healthy controls, allografts in a nondiabetic milieu, allografts in a diabetic milieu are shown in (d), (e), and (f), respectively.

Podocytes (ligand expression) and potential crosstalk with GECs (receptor expression).

Increased expression of LR pairs was observed in which the podocyte ligand would be expected to modulate GEC receptors involved in angiogenesis, regulation of endothelial cytoskeletal structure, integrity, and vascular maturation (KDR, NOTCH4, and APLNR), cell adhesion (PECAM1), endothelial cell differentiation (SIPR1), smooth muscle relaxation (VIPR1), increased transforming growth factor β (SMAD3), and fibroblast growth factor signaling (FGFR3). In GECs exposed to a diabetic milieu, we observed putative LR interactions associated with increased intercellular signaling (CD93), vasoconstriction and vasodilation (EDNRB), cell surface adhesion, and signaling (ITGA5).

Parallel podocyte and endothelial transcriptomic changes occur in the single kidney state in humans (human allografts vs. pretransplant) and a rodent model of hypertrophy-induced FSGS, also caused by transition from the 2 kidneys to the single kidney state

The transcriptional changes observed in allograft recipients noted above were compared with those observed in a transgenic rodent model that develops podocyte stress and FSGS. Table 318 shows notable parallels between the differential gene expression of genes induced in human allografts 3 to 12 months post-transplantation and in the remaining rodent kidney 3 weeks after uninephrectomy. Even when a wild-type rodent model was used, similar parallels are noted, although not to the levels of significance for the transgenic rat model (data not shown). This result is compatible with the concept that allograft podocytes and endothelial cells undergo parallel hypertrophy-associated stresses similar to those occurring in the rat hypertrophy-induced FSGS model.

Table 3 |.

Comparison of podocyte and endothelial transcriptomic changes occurring in 2 kidneys vs. single kidney states for humans undergoing kidney transplantation and a rat transgenic model of hypertrophy-induced FSGS

Variable Human kidney single cell transcriptome: single kidney vs. 2 kidneys
 Rat isolated glomerular transcriptome: single kidney vs. 2 kidneys
Log fold change P Log fold change P
Podocytes
 NPHS1 −0.36 0.01 −0.88 <0.000001
 NPHS2 0.18 0.35 −0.56 0.00004
NPHS2/NPHS1 ratio 0.62 0.001 0.27 0.02
 SYNPO −0.31 0.01 −0.45 0.0007
 ITGA3 −0.42 0.0002 −0.60 0.00001
Glomerular endothelium
 FLT1 −0.03 0.89 0.29 0.03
 KDR 0.3 2.54 × 10−10 0.219 0.09
 APLNR 0.38 2.95 × 10−6 0.58 0.00002
Open in a new tab

4E-BP1, 4E-binding protein 1; APLNR, Apelin Receptor; FLT1, Fms Related Receptor Tyrosine Kinase 1; FSGS, focal segmental glomerular sclerosis; ITGA3, Integrin A3; KDR, Kinase Insert Domain Receptor; NPHS1, Nephrin; NPHS2, Podocin; SYNPO, Synaptopodin.

For human data (at left), differentially expressed podocyte-specific and endothelial cell single cell transcriptomes are from allograft surveillance biopsies (1 kidney) vs. healthy kidney donors (2 kidneys) are shown. For rat data (at right), the differentially expressed transcriptomes are from transgenic rat glomeruli isolated from nephrectomized rats (1 kidney) vs. sham-nephrectomized rats (2 kidneys) 3 weeks after either uninephrectomy or sham uninephrectomy. In the transgenic rat model, podocytes specifically overexpressed a podocin promoter–driven modified 4E-BP1 transcript coding for a 4E-BP1 protein with threonine residues changed to alanine to downregulate mTORC1 activation of protein synthesis in response to hypertrophy selectively in podocytes.18 Expression of this transgene rendered these rats susceptible to hypertrophy-induced stress and development of FSGS caused by uninephrectomy-induced compensatory kidney hypertrophy (single kidney state). Shown in this table are podocyte and endothelial-selective transcripts to allow a comparison of bulk glomerular with single cell transcripts. There is a similarity between the differential gene expression profiles of podocytes and glomerular endothelial cells across the human and rat models as they adapt from the 2 kidneys to the single kidney state and where, in the case of the rat model, glomerular hypertrophy–associated FSGS develops.

DISCUSSION

In this observational study, we report an unbiased single cell RNA-seq analysis of stable and histologically normal allografts in the first post-transplant year. Several key observations are made as follows:

  1. Among GECs, we identified pathways that were enriched for endothelial shear stress, angiogenesis, and interferon-γ signaling with downstream activation of the STAT pathway.

  2. Among podocytes, we identified response to growth factor and actin filament reorganization as key enriched pathways. In addition, we observed an increase in podocyte hypertrophic stress and a parallel reduction in genes that code for proteins responsible for keeping podocytes attached to the glomerular basement membrane.

  3. From LR interaction studies, we inferred a significant increase in bidirectional crosstalk between podocytes and GECs after kidney transplantation.

  4. We identified parallels in genes related to podocyte and endothelial stress between allografts and rat models of uninephrectomy that progressed to FSGS compatible with the idea that cell stresses have already been set in motion that can drive hypertrophy-associated glomerular disease progression.

  5. Lastly, we offer a comprehensive collection of single cell transcriptional profiling of healthy kidney allografts and living donors that can be used as a healthy transplant and healthy human single cell RNA-seq controls for future studies for all investigators.

Among GECs, we identified fluid shear stress as a highly enriched pathway in all allografts. This finding is not surprising given the well-documented increase in blood flow (and thus hyperfiltration) in allografts after transplantations that imposes laminar and circumferential shear stresses on the endothelial cells.12 Hemodynamic shear stresses in remnant kidney models have been associated with the progression of glomerular disease.24 We identified reduced GEC expression of KLF2 and KLF4, both of which have been associated with worsening glomerular disease phenotype.25 Endothelial shear stress is typically accompanied by increasing angiogenesis because of increased KDR (VEGFR2) expression, a finding that was also observed in our data, thereby linking shear stress with angiogenesis.26 Increased KDR expression in GECs has been associated with the progression of kidney disease.27 Together, these data highlight fluid shear stress and angiogenesis as crucial pathways that are activated even in histologically normal allografts after kidney transplantation. Both pathways have been associated with progressive disease in model systems. In addition to shear stress and angiogenesis, we observed increased interferon-related signaling (including interferon-γ being the primary upstream regulator) and downstream activation of the STAT pathway. The Janus kinase–STAT pathway has been established to be critical in the development of diabetic nephropathy and other glomerular diseases, including FSGS and IgA nephropathy.28–30 Whether this is the typical compensatory response of the allograft to transplantation or suggestive of activation of deleterious progression pathways remains unclear and cannot be resolved till more long-term data of this cohort are obtained.

In podocytes, after transplantation, we identified evidence of podocyte injury and stress as noted by an increased NPHS2/NPHS1 ratio.22 This ratio was driven by the relative downregulation of NPHS1 (nephrin) expression. NPHS1 transcript downregulation is a common feature observed in model systems of nephrectomy as well as many conditions accompanied by hyperfiltration, including diabetes and obesity, and is considered a marker of early glomerular injury.31–36 The downregulation of the NPHS1 transcript observed in our data should be viewed in the context of there being no concurrent histopathological evidence of recurrent, de novo glomerular disease, or clinically significant proteinuria. In model systems, an elevated NPHS2/NPHS1 ratio has been strongly correlated with higher podocyte depletion, glomerulosclerosis, and interstitial fibrosis score where it performed better or was at least equal to the traditionally used urine protein-to-creatinine ratio.18,22

Two separate studies have shown that podocyte loss begins early after transplantation, even in allografts with no glomerular disease or proteinuria.5,10 Kriz and Lemley have speculated that biomechanical forces on the glomeruli are in part responsible for the detachment of viable podocytes from the glomerular basement membrane.37 Because podocytes are attached to the glomerular basement membrane chiefly by integrins, we investigated the expression patterns of transcripts that code for these key proteins involved in podocyte adhesion to the glomerular basement membrane. Compared to healthy living donors, we noted a significant reduction in the expression of the ITGA3 transcript coding integrin α3, which forms part of the heterodimer integrin α3β1, an essential protein complex involved in maintaining podocyte adhesion to the glomerular basement membrane.38 Importantly, this reduced expression pattern was noted in both allografts exposed to an early diabetic and nondiabetic milieu, suggesting that this a common process independent of diabetes status. The change in ITGA3 transcript expression was also accompanied by reduction in ITGA3 protein expression confirmed on ITGA3 immunoperoxidase staining from a paired sample (Figure 6). In addition, another key transcript SYNPO (synaptopodin) was significantly reduced. SYNPO is an actin-associated protein that is thought to play a key role in modulating the shape and motility of podocyte foot processes.39 Several other transcripts that code for crucial adhesion proteins were reduced specifically in diabetic allografts. For example, there was also a significant reduction in the expression of genes associated with ITGB5 (integrin beta 5) and TALN (talin). Talin is a cytoplasmic protein that binds to the integrin tail, resulting in β-integrin activation.40

Several key insights were inferred from LR interaction studies. The fundamental assumption used for these studies is that this putative crosstalk is occurring between cells that are in close spatial proximity to one another and thus likely to have significant paracrine interactions. First, we observed a remarkable increase in the putative crosstalk between GECs and podocytes after transplantation in allografts of both diabetic and nondiabetic recipients compared with controls. Second, there was a substantial overlap of expected LR signaling (cell-cell adhesion, angiogenesis, and intracellular signaling) among GECs and podocytes of allografts in a diabetic and nondiabetic milieu. However, there were some novel observations. For example, in diabetic patients, there was an increased expression of glomerular endothelial CD93, a transmembrane receptor that activates β1-integrin and is involved in cell-cell adhesion, angiogenesis, and fibronectin fibrillogenesis important for laying down the extracellular matrix.41 CD93 signaling has recently been identified as a therapeutic target to reduce neovascular retinopathy and as a biomarker of diabetic nephropathy.42,43 Caution should, however, be taken in the interpretation of LR interaction studies. These studies are based on the increased expression of receptors and their corresponding ligands that are based on expertly curated data sets obtained from the literature. These studies do not account for as yet unidentified LR interactions. Furthermore, because of our desire to exclude false-positive results, ligands and receptor interactions with low interaction scores (<0.5) were excluded and so specific interactions that are occurring at low levels might be missed. False-positive interactions could also be identified where a commonly expressed ligand or receptor could be paired with a corresponding ligand or receptor that has become activated in response to endocrine (as opposed to paracrine) signals from other cells/cell types that are not in immediate proximity. Importantly, these interactions capture expression levels only at 1 point in time (at the time of biopsy) and thus do not exclude the possibility of other unobserved dynamic interactions.

We then tested whether key transcriptional changes observed in our single cell transcriptomic data might have longer-term implications. To do this, we compared the transcriptional profile of podocytes and GECs from the allografts with the bulk transcriptional profile of podocyte- and GEC-specific transcripts from isolated glomeruli of a rat model that is triggered by uninephrectomy to develop overt proteinuria and FSGS. We observed a striking similarity between key genes associated with podocyte stress, hypertrophy, binding of the podocyte to the glomerular basement membrane, and angiogenesis between allografts and the progressive rat model that, like allografts, is induced by the transition from the 2 kidneys to the single kidney state. These data are compatible with the concept that even in apparently healthy human allografts with no detectable histological abnormality, transcriptomic changes reflecting cell stresses have already been set in motion that could potentially drive hypertrophy-associated glomerular disease progression in the allograft as it does in the rat model. The long-term surveillance biopsy data in the reports by Nankivell et al. and Stegall et al. in human kidney allografts without chronic rejection further support this finding.8,9

Our study design allowed us to test transcriptional profile differences between diabetic and nondiabetic kidney allografts. Conceptually, because these groups both have the common element of transplantation, this approach should uncover the differences in transcriptional profile between diabetic and nondiabetic milieus, provided other donors and recipient factors were well balanced between the groups. Because all donated kidneys were from nondiabetic individuals, this approach should provide an early transcriptional profile of kidneys exposed to a diabetic milieu. Such studies are especially important because currently most available transcriptional profiling in humans has been performed in kidneys with already established diabetic nephropathy.44 Thus, whether the findings from such studies genuinely represent the underlying drivers of the diabetic process as opposed to the cumulative insults incurred by the kidney up to the time of the biopsy remains unclear. However, much larger studies that balance additional key factors such as underlying primary diseases, detailed medication exposures, and other post-transplant complications are needed before one could confidently attribute a transcriptional milieu to the early diabetic environment. Therefore, the data presented here should be viewed only as “preliminary and proof of concept.”

Although our study offers important insights into the transcriptional landscape of a “normal” allograft, there are several limitations, including those that are related to single cell transcriptomic analysis that have been reported elsewhere.45–47 First, the transcriptional profile of a surveillance biopsy represents, to some extent, the baseline transcriptional profile of the donor’s kidney, which comprised a combination of living and deceased donors. Therefore, the transcriptional profile could represent, to some extent, the underlying disease states in deceased donors and the ischemia-reperfusion injury accompanying transplants. However, since biopsies were at or beyond the 3-month post-transplantation, peritransplant events, such as ischemia-reperfusion injury, are less likely to be driving the transcriptional profile. Furthermore, there were no differences in drug exposure at the time of biopsy because all allografts, whether at 3 or 12 months, have similar immunosuppression protocols at our institution. None of the patients had any known infectious or noninfectious complications after transplantation, including rejection. In addition, despite advances in cell dissociation technology, the yield for epithelial cells such as podocytes remains suboptimal in single cell analysis, and as such, this remains a limitation of this study.48,49 However, reassuringly the patterns of gene expression in the single cell data of glomerular sub-components in allograft kidneys and the bulk RNA-seq data from isolated glomeruli of rat nephrectomy models were comparable, thereby supporting the validity of the findings.

Taking the known allograft hyperfiltration and previous animal models of single kidney states, 1 key assumption we make is that the transcriptional profile of these cells is related to the degree of hyperfiltration. However, there was no direct experiment comparing the transcriptional profile of cells with a measure of allograft hyperfiltration, so this question merits more testing. It is indeed possible that the higher rate of podocyte detachment might be due to an inherent susceptibility of allograft podocytes to an immune milieu after transplantation and this independently, or together with post-transplant hyperfiltration, might drive the higher-than-expected podocyte detachment as was the case with the transgenic animals in Nishizono et al.,18 whose podocytes had a heightened susceptibility to hypertrophic and hyperfiltration-related stresses. Animal models testing this very concept will need to be performed to investigate these questions.

Another potential confounder is the timing of biopsies at 3 versus 12 months. Although there are likely to be some time-dependent effects between 3 and 12 months post-transplantation, analysis of enriched pathways of GECs in 3- and 12-month biopsies identified the same top 3 pathways in both groups (data not shown). For analysis, we therefore combined 3- and 12-month biopsies into a biopsy cohort within the first year for further analysis. Larger numbers will be required to evaluate this point in more detail. It is important to note that a change in the transcriptional profile does not necessarily imply that protein expression is altered in the same direction; however, at least in the case of ITGA3 we observed concordance between transcript and protein expression. However, the focus of the present study was to examine an unbiased transcriptional profile of glomerular subcomponents, specifically the podocytes and GECs. Thus, detailed validation studies focusing on potentially druggable pathways, while ongoing, are beyond the planned scope of this report. Although measuring protein-protein interactions might have been theoretically preferable, RNA-seq data sets are numerous, easier to access, and easier to analyze at the present time. Furthermore, proteomics, specifically at single cell resolution, is still a technology in its infancy. Long-term allograft outcomes will be required to determine whether cell-specific biomarkers of disease progression can be identified at very early time points before the onset of histologically apparent pathology as a step toward precision medicine.

Supplementary Material

1

Supplementary File (TIFF)

Figure S1. Functional network modules generated by HumanBase for the genes differentially expressed in podocytes of all surveillance biopsies versus all healthy controls. Each blob is a gene—the node size (blob size) corresponds to the degree of the node in the network (number of edges connecting it to other genes). So larger the blob, the more connections it has with other genes in that module.

ACKNOWLEDGMENTS

We acknowledge the contribution of John Magee, the former director of the Comprehensive Transplant Program, without whose backing this study would not be possible. Kelly Shaffer, Derek Green, and Grant Chappel were instrumental in consenting patients and collecting samples. Lastly, we acknowledge all the patients who willingly allowed the investigators to obtain additional samples for the purpose of the study.

ASN acknowledges support of the George M. O’Brien Michigan Kidney Translational Core Center (P30 DK081943) and the Michigan Institute for Clinical and Health Research First Pathway Award (UL1TR002240) as well as startup funds from the Department of Internal Medicine and the National Institutes of Health (NIH; grant K23 DK 125529). RCW acknowledges the support of the NIH (grants R01 DK 46073 and R01 DK 102643) and University of Michigan O’Brien Kidney Translational Core Center (P30 DK081943). MK acknowledges the support of the Kidney Precision Medicine Project (KPMP) UG-DK-114907. KPMP is a multiyear project funded by the National Institute of Diabetes and Digestive, and Kidney Diseases (NIDDK) with the purpose of understanding and finding new ways to treat chronic kidney disease and acute kidney injury. KPMP is funded by the following grants from the NIDDK: U2C DK114886, UH3DK114861, UH3DK114866, UH3DK114870, UH3DK114908, UH3DK114915, UH3DK114926, UH3DK114907, UH3DK114920, UH3DK114923, UH3DK114933, and UH3DK114937. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Please note that 3 samples used in this analysis were processed under KPMP.

Footnotes

DISCLOSURE

MK reports grants from the National Institutes of Health (NIH); nonfinancial support from the University of Michigan; grants from JDRF, AstraZeneca, Novo Nordisk, Eli Lilly, Gilead, Goldfinch Bio, Merck, Janssen, Boehringer Ingelheim, Moderna, European Union Innovative Medicine Initiative, Chan Zuckerberg Initiative, Certa, Chinook, amfAR, Angion Pharmaceuticals, Renalytix, Travere Therapeutics, Regeneron, and Ionis Pharmaceuticals; and other grants from Astellas, Poxel, and Shire Pharmaceuticals (outside the submitted work). In addition, MK has a patent PCT/EP2014/073413 “Biomarkers and methods for progression prediction for chronic kidney disease” licensed. EAF is a consultant for Novartis. JBH reports grants from the NIH, AstraZeneca, Novo Nordisk, Eli Lilly, Gilead, and Janssen. All the other authors declared no competing interests.

DATA STATEMENT

All the biopsy samples used in this study were uploaded to Gene Expression Omnibus under accession number GSE 169285 and are thus publicly available.

REFERENCES

  • 1.Kasiske BL, Snyder JJ, Gilbertson D. Inadequate donor size in cadaver kidney transplantation. J Am Soc Nephrol 2002;13:2152–2159. [DOI] [PubMed] [Google Scholar]
  • 2.Naik AS, Sakhuja A, Cibrik DM, et al. The impact of obesity on allograft failure after kidney transplantation: a competing risks analysis. Transplantation 2016;100:1963–1969. [DOI] [PubMed] [Google Scholar]
  • 3.Giral M, Foucher Y, Karam G, et al. Kidney and recipient weight incompatibility reduces long-term graft survival. J Am Soc Nephrol 2010;21:1022–1029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Einecke G, Sis B, Reeve J, et al. Antibody-mediated microcirculation injury is the major cause of late kidney transplant failure. Am J Transplant 2009;9:2520–2531. [DOI] [PubMed] [Google Scholar]
  • 5.Naik AS, Afshinnia F, Aqeel J, et al. Accelerated podocyte detachment early after kidney transplantation is related to long-term allograft loss of function. Nephrol Dial Transplant 2019;34:1232–1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Gaston RS, Cecka JM, Kasiske BL, et al. Evidence for antibody-mediated injury as a major determinant of late kidney allograft failure. Transplantation 2010;90:68–74. [DOI] [PubMed] [Google Scholar]
  • 7.Naik AS, Afshinnia F, Cibrik D, et al. Quantitative podocyte parameters predict human native kidney and allograft half-lives. JCI Insight 2016;1: e86943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Nankivell BJ, Borrows RJ, Fung CL, et al. The natural history of chronic allograft nephropathy. N Engl J Med 2003;349:2326–2333. [DOI] [PubMed] [Google Scholar]
  • 9.Stegall MD, Cornell LD, Park WD, et al. Renal allograft histology at 10 years after transplantation in the tacrolimus era: evidence of pervasive chronic injury. Am J Transplant 2018;18:180–188. [DOI] [PubMed] [Google Scholar]
  • 10.Yang Y, Hodgin JB, Afshinnia F, et al. The two kidney to one kidney transition and transplant glomerulopathy: a podocyte perspective. J Am Soc Nephrol 2015;26:1450–1465. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Naik AS, Aqeel J, Wang SQ, et al. Urine marker analysis identifies evidence for persistent glomerular podocyte injury across allograft lifespan. Clin Transplant 2021;35:e14457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Bosma RJ, Kwakernaak AJ, van der Heide JJ, et al. Body mass index and glomerular hyperfiltration in renal transplant recipients: cross-sectional analysis and long-term impact. Am J Transplant 2007;7: 645–652. [DOI] [PubMed] [Google Scholar]
  • 13.Arazi A, Rao DA, Berthier CC, et al. The immune cell landscape in kidneys of patients with lupus nephritis. Nat Immunol 2019;20:902–914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Menon R, Otto EA, Hoover P, et al. Single cell transcriptomics identifies focal segmental glomerulosclerosis remission endothelial biomarker. JCI Insight 2020;5:e133267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Butler A, Hoffman P, Smibert P, et al. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 2018;36:411–420. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Greene CS, Krishnan A, Wong AK, et al. Understanding multicellular function and disease with human tissue-specific networks. Nat Genet 2015;47:569–576. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Krishnan A, Zhang R, Yao V, et al. Genome-wide prediction and functional characterization of the genetic basis of autism spectrum disorder. Nat Neurosci 2016;19:1454–1462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Nishizono R, Kikuchi M, Wang SQ, et al. FSGS as an adaptive response to growth-induced podocyte stress. J Am Soc Nephrol 2017;28:2931–2945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Cabello-Aguilar S, Alame M, Kon-Sun-Tack F, et al. SingleCellSignalR: inference of intercellular networks from single-cell transcriptomics. Nucleic Acids Res 2020;48:e55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Menon R, Otto EA, Sealfon R, et al. SARS-CoV-2 receptor networks in diabetic and COVID-19–associated kidney disease. Kidney Int 2020;98: 1502–1518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Haas M, Loupy A, Lefaucheur C, et al. The Banff 2017 Kidney Meeting Report: revised diagnostic criteria for chronic active T cell-mediated rejection, antibody-mediated rejection, and prospects for integrative endpoints for next-generation clinical trials. Am J Transplant 2018;18: 293–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Fukuda A, Wickman LT, Venkatareddy MP, et al. Urine podocin:nephrin mRNA ratio (PNR) as a podocyte stress biomarker. Nephrol Dial Transpl 2012;27:4079–4087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Siddiqi FS, Advani A. Endothelial-podocyte crosstalk: the missing link between endothelial dysfunction and albuminuria in diabetes. Diabetes 2013;62:3647–3655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Lee LK, Meyer TW, Pollock AS, Lovett DH. Endothelial cell injury initiates glomerular sclerosis in the rat remnant kidney. J Clin Invest 1995;96:953–964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zhong F, Mallipattu SK, Estrada C, et al. Reduced kruppel-like factor 2 aggravates glomerular endothelial cell injury and kidney disease in mice with unilateral nephrectomy. Am J Pathol 2016;186:2021–2031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kang DH, Joly AH, Oh SW, et al. Impaired angiogenesis in the remnant kidney model: I. Potential role of vascular endothelial growth factor and thrombospondin-1. J Am Soc Nephrol 2001;12:1434–1447. [DOI] [PubMed] [Google Scholar]
  • 27.Sato W, Tanabe K, Kosugi T, et al. Selective stimulation of VEGFR2 accelerates progressive renal disease. Am J Pathol 2011;179:155–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Tao J, Mariani L, Eddy S, et al. JAK-STAT signaling is activated in the kidney and peripheral blood cells of patients with focal segmental glomerulosclerosis. Kidney Int 2018;94:795–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Marrero MB, Banes-Berceli AK, Stern DM, Eaton DC. Role of the JAK/STAT signaling pathway in diabetic nephropathy. Am J Physiol Renal Physiol 2006;290:F762–F768. [DOI] [PubMed] [Google Scholar]
  • 30.Tao J, Mariani L, Eddy S, et al. JAK-STAT activity in peripheral blood cells and kidney tissue in IgA nephropathy. Clin J Am Soc Nephrol 2020;15: 973–982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Benigni A, Gagliardini E, Tomasoni S, et al. Selective impairment of gene expression and assembly of nephrin in human diabetic nephropathy. Kidney Int 2004;65:2193–2200. [DOI] [PubMed] [Google Scholar]
  • 32.Kandasamy Y, Smith R, Lumbers ER, Rudd D. Nephrin—a biomarker of early glomerular injury. Biomark Res 2014;2:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Fornoni A, Jeon J, Varona Santos J, et al. Nephrin is expressed on the surface of insulin vesicles and facilitates glucose-stimulated insulin release. Diabetes 2010;59:190–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ruggenenti P, Porrini EL, Gaspari F, et al. Glomerular hyperfiltration and renal disease progression in type 2 diabetes. Diabetes Care 2012;35: 2061–2068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Sato Y, Wharram BL, Lee SK, et al. Urine podocyte mRNAs mark progression of renal disease. J Am Soc Nephrol 2009;20:1041–1052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Fukuda A, Wickman LT, Venkatareddy MP, et al. Angiotensin II-dependent persistent podocyte loss from destabilized glomeruli causes progression of end stage kidney disease. Kidney Int 2012;81:40–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Kriz W, Lemley KV. A potential role for mechanical forces in the detachment of podocytes and the progression of CKD. J Am Soc Nephrol 2015;26:258–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Pozzi A, Zent R. Integrins in kidney disease. J Am Soc Nephrol 2013;24: 1034–1039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Mundel P, Heid HW, Mundel TM, et al. Synaptopodin: an actin-associated protein in telencephalic dendrites and renal podocytes. J Cell Biol 1997;139:193–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Wegener KL, Partridge AW, Han J, et al. Structural basis of integrin activation by talin. Cell 2007;128:171–182. [DOI] [PubMed] [Google Scholar]
  • 41.Lugano R, Vemuri K, Yu D, et al. CD93 promotes b1 integrin activation and fibronectin fibrillogenesis during tumor angiogenesis. J Clin Investig 2018;128:3280–3297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Tosi GM, Caldi E, Parolini B, et al. CD93 as a potential target in neovascular age-related macular degeneration. J Cell Physiol 2017;232:1767–1773. [DOI] [PubMed] [Google Scholar]
  • 43.Lee M, Park HS, Choi MY, et al. Significance of soluble CD93 in type 2 diabetes as a biomarker for diabetic nephropathy: integrated results from human and rodent studies. J Clin Med 2020;9:1394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Wilson PC, Wu H, Kirita Y, et al. The single-cell transcriptomic landscape of early human diabetic nephropathy. Proc Natl Acad Sci U S A 2019;116: 19619–19625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Angerer P, Simon L, Tritschler S, et al. Single cells make big data: new challenges and opportunities in transcriptomics. Curr Opin Syst Biol 2017;4:85–91. [Google Scholar]
  • 46.Hicks SC, Townes FW, Teng M, Irizarry RA. Missing data and technical variability in single-cell RNA-sequencing experiments. Biostatistics 2018;19:562–578. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Kiselev VY, Andrews TS, Hemberg M. Challenges in unsupervised clustering of single-cell RNA-seq data. Nat Rev Genet 2019;20:273–282. [DOI] [PubMed] [Google Scholar]
  • 48.Wu H, Kirita Y, Donnelly EL, Humphreys BD. Advantages of single-nucleus over single-cell RNA sequencing of adult kidney: rare cell types and novel cell states revealed in fibrosis. J Am Soc Nephrol 2019;30:23–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Wu H, Malone AF, Donnelly EL, et al. Single-cell transcriptomics of a human kidney allograft biopsy specimen defines a diverse inflammatory response. J Am Soc Nephrol 2018;29:2069–2080. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1

Supplementary File (TIFF)

Figure S1. Functional network modules generated by HumanBase for the genes differentially expressed in podocytes of all surveillance biopsies versus all healthy controls. Each blob is a gene—the node size (blob size) corresponds to the degree of the node in the network (number of edges connecting it to other genes). So larger the blob, the more connections it has with other genes in that module.

NIHMS1766451-supplement-1.tiff (1.1MB, tiff)

RESOURCES