Skip to main content
NCBI home page
Kidney360 logoLink to Kidney360
. 2023 Feb 10;4(4):e515–e524. doi: 10.34067/KID.0000000000000074

Bioinformatics Analysis Reveals a Shared Pathway for Common Forms of Adult Nephrotic Syndrome

DengFeng Li 1,✉, Liang Liu 2, Mariana Murea 3, Barry I Freedman 3, Lijun Ma 3,✉
PMCID: PMC10278839  PMID: 36763793

graphic file with name kidney360-4-e515-g001.jpg

Open in a new tab

Keywords: chronic nephropathy, nephrotic syndrome

Abstract

Key Points

  • Dysregulation of the focal adhesion pathway is present in the three most common forms of glomerular disease, that is, Focal segmental glomerulosclerosis, membranous nephropathy, and minimal change disease.

  • Zyxin is seen to be upregulated in the glomerular compartment of patients with the three most common forms of glomerular disease.

Background

Focal segmental glomerulosclerosis, membranous nephropathy, and minimal change disease are common causes of nephrotic syndrome. Although triggers for these diseases differ, disease progression may share common molecular mechanisms. The aim of this study was to investigate the presence of molecular pathways that are dysregulated across these glomerular diseases.

Methods

The gene expression dataset GSE200828 from the Nephrotic Syndrome Study Network study was obtained from the Gene Expression Omnibus database. R and Python packages, Cytoscape software, and online tools (DAVID and STRING) were used to identify core genes and topologically relevant nodes and molecular pathways. Single-cell RNA sequencing analysis was applied to identify the expression patterns of core genes across kidney cell types in glomerular compartments.

Results

A total of 1087 differentially expressed genes were identified, including 691 upregulated genes and 396 downregulated genes, which are common in all three forms of nephrotic syndrome compared with kidney donor controls (FDR P<0.01). A multiapproach bioinformatics analysis narrowed down to 28 similarly dysregulated genes across the three proteinuric glomerulopathies. The most topologically relevant nodes belonged to the adherens junction, focal adhesion, and cytoskeleton pathways, where zyxin covers all of those gene ontology terms.

Conclusions

We report that dysregulation of cell adhesion complexes was present in the three most common forms of glomerular disease. Zyxin could be a biomarker in all three common forms of nephrotic syndrome. If further functional studies confirm its role in their development, zyxin could be a potential therapeutic target.

Introduction

Focal segmental glomerulosclerosis (FSGS), membranous nephropathy (MN), and minimal change disease (MCD) are the most common forms of glomerular disease in adults and typically manifest with nephrotic syndrome.1 With the exception of monogenic forms of FSGS and steroid-resistant nephrotic syndrome, the etiology of these diseases is often unknown and believed to be multifactorial. For example, FSGS can be triggered by infection or drug toxicity and often has an underlying genetic predisposition. In recent years, risk variants in the apolipoprotein L1 gene (APOL1) have been identified as contributors to the development of FSGS in populations with recent African ancestry.2 Innate immunity pathways have been implicated in APOL1 expression,3 and collapsing FSGS can develop in patients who harbor APOL1 high-risk genotypes and contract HIV or coronavirus disease 2019.4,5 Similarly, MN and MCD have been related to alterations in the immune response. MN often develops as a result of autosensitization to podocyte-expressed proteins,6 and the risk of developing MN is significantly increased in those with HLA-DQA1 alleles.7 The pathophysiology of MCD is least well understood but is believed to be caused by cell-mediated immune dysregulation with a contribution of genetic susceptibility.8-10

Using machine learning methodology, the Nephrotic Syndrome Study Network (NEPTUNE), and the European Renal cDNA Bank database containing gene expression in microdissected kidney tissue, Sealfon et al. identified distinctive glomerular gene expression profiles in patients with MN, relative to those with other forms of nephrotic syndrome (namely, FSGS, MCD, IgA nephropathy, and others).11

We applied strategies for identifying shared molecular pathways across patients with FSGS, MN, and MCD. Mainstays in treating nephrotic syndrome are to slow down or halt processes that maintain and/or accelerate proteinuria and the loss of glomerular filtration. Identification of common molecular pathways, which could be triggered by similar (i.e., immune dysregulation) or different factors, may lead to the development of novel therapeutic targets.

Methods

Microarray Data and Patient Sample Selection

Gene Expression Omnibus (GEO) is a National Institutes of Health–supported functional genomics data repository (https://www.ncbi.nlm.nih.gov/geo/) housing gene expression arrays and next-generation sequence–based data for replication and novel in-depth analyses. We searched the GEO DataSets using the keywords “NEPTUNE” and “glomerular compartment” to identify suitable kidney gene expression datasets for glomerular compartments generated by the NEPTUNE consortium. The gene expression dataset GSE200828 consists of microarray profiles from glomerular compartment of kidney biopsies from adults screened for the NEPTUNE cohort (University of Michigan, Ann Arbor, MI) performed on the GPL19983 platform (Affymetrix Human Gene 2.1 ST Array).

Study Design

Figure 1 presents approaches for narrowing down the potential overlapping pathways in FSGS, MN, and MCD for this study. To detect shared pathways, differentially expressed genes (DEGs) were analyzed for each form of nephrotic syndrome and compared with the control group (CTL) of healthy living kidney donors using GEO2R. GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) is an interactive web software tool that allows operators to compare two or more different groups of samples in a GEO series to identify DEGs across experimental conditions via the GEO query and “limma” R package (Bioconductor project). DEGs with a false discovery rate (FDR) P-value<0.01 in each form of nephrotic syndrome (versus CTL) were selected and merged to identify common upregulated or downregulated transcripts, where DEGs of |log2FC|<0.75 were excluded. R packages (“gage” and “pathfindR”) were used to identify pathways enriched for shared transcripts of the DEGs and map them into the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database.

Figure 1.

Figure 1

Open in a new tab

Study design.

Transcripts enriched in the top KEGG pathway were further analyzed using machine learning technology to identify the feature importance of individual transcripts. Because of the imbalanced dataset, balanced random forest classification (oversampling) and weighted logistic regression algorithms were performed using the Python package scikit-learn. We performed feature selection and train-test split with k-fold stratified cross-validation to ensure the performance of the models. The feature importance score for the random forest model was computed on the basis of the decrease in impurity (Gini) and coefficient for the logistic regression model.

Overlapping upregulated transcripts and downregulated transcripts for all three forms of nephrotic syndrome versus CTL were implemented to Cytoscape12 and DAVID Bioinformatics tools13 to uncover the enriched transcripts pointing to the top gene ontology (GO) terms from the top functional annotation clustering.

To reduce the large numbers of DEGs, the top 200 upregulated transcripts for each of the FSGS, MN, and MCD DEGs were selected and merged by using gene Entrez ID to form an overlapping transcript list, which was subsequently inner-merged with the enriched transcripts for the top GO term identified by DAVID. This narrowed-down gene list was imported to a protein-protein interaction (PPI) web prediction tool STRING (https://string-db.org/) to visualize physical interactions of top gene encoded proteins. The enriched nodes (gene products) indicated by STRING and supported by top KEGG pathways and top GO terms (from DAVID) revealed central hub (highlighted nodes) of the most significant common pathways contributing to FSGS, MN, and MCD.

Genes highlighted in the central hub were queried for their distribution patterns in kidney cell types using the Python “Scanpy” package according to our single-cell RNA sequencing (scRNA-seq) analysis results from a mixture of glomerular-enriched cells obtained from the normal kidney tissue of a patient who underwent nephrectomy. The scRNA-seq data are openly available at the National Center for Biotechnology Information (NCBI) sequence read archive (https://www.ncbi.nlm.nih.gov/sra/PRJNA781289, accession: SRX13160927). The protein distribution patterns of encoded hub genes on kidney sections were referenced in the human protein atlas (https://www.proteinatlas.org/).

Other kidney biopsy gene expression datasets GSE104948, GSE129973, GSE115857, and GSE200818 from the GEO database were acquired for the validation of differential gene expression of key transcripts using GEO2R.

Statement of Ethics

This bioinformatics research was conducted based on publicly available data on NCBI, with no direct patient involvement.

Results

Target Samples and Microarray Information

The gene expression dataset of GSE200828, covering the gene expression profile of glomerular compartment kidney biopsies from six healthy living kidney donors, 31 patients with FSGS, 51 patients with MN, and 19 patients with MCD, was selected for analysis. The dataset was submitted on April 14, 2022, and last updated on April 17, 2022.

Identification of DEGs and Relevant GO Terms and KEGG Pathways

Differential gene expression of FSGS versus CTL, MN versus CTL, and MCD versus CTL was assessed. First, the distribution of samples was viewed by GEO2R, and the median-centered log2 transformed gene expression values (relative log expression plot) are displayed in Supplemental Figure 1. The data appeared normally distributed and comparable across diseases. Hence, the sample quality was felt suitable for differential gene expression analysis. We then identified a total of 1087 significant DEGs (all FDR<0.01), all of which appeared to show the same direction for log2(fold change) among three differential gene expression groups. This resulted in 691 upregulated genes (Supplemental Table 1-1) and 396 downregulated genes (Supplemental Table 1-2) common in FSGS, MN, and MCD. Using R package “pathfindR,” a tool for enrichment analysis via active subnetworks, the package offers functionalities to cluster enriched terms and identify representative terms in each cluster. We identified focal adhesion (KEGG: hsa04510) and Rap1 signaling pathway (KEGG: hsa04015) with P-values<10−10 as the highest support values (Supplemental Table 2). A minimum |log2fold change| of the three was selected to represent the overall fold change of differential gene expression for FSGS, MN, and MCD versus CTL. This was an attempt to avoid exaggeration of differential gene expression effects. Similar results (adherens junction [GO cellular component] and focal adhesion and Rap1 signaling pathway [GO biological process] were the most relevant, and correlated GO terms across hierarchies [Figure 2]) were obtained from a different R package “gage.” The KEGG pathway analysis results were visualized using “pathview” (Figure 3). Zyxin (ZYX), actinin (ACTN), CT10 regulator of kinase (CRK), and integrin (ITGA) were among the most upregulated transcripts.

Figure 2.

Figure 2

Open in a new tab

Dot plot of top GO terms implicated by enriched transcripts. The focal adhesion pathway was supported by GO terms in cellular component and biological process.

Figure 3.

Figure 3

Open in a new tab

Differentially expressed genes are enriched in focal adhesion and actin cytoskeleton pathways. Key components in (A) the focal adhesion pathway are regulated in the same direction for FSGS, MS, and MCD and (B) the actin cytoskeleton pathway (adherens junction/focal adhesion) are regulated in the same direction for FSGS, MS, and MCD.

Feature Importance Identification Using Machine Learning Algorithms

The expression level of 25 transcripts (Supplemental Table 2) enriched in the focal adhesion pathway was analyzed using machine learning algorithms in 107 individuals: 101 (31 FSGS, 51 MN, and 19 MCD) cases versus six controls. A correlation matrix was developed using the expression level of 25 transcripts highlighted in the focal adhesion pathway (Supplemental Table 2). After performing the feature selection with correlation threshold set to 0.7, the algorithm automatically dropped six transcripts highly correlated with the others. The built-in random forest feature importance analysis (based on the decrease in impurity [Gini]) and coefficient of logistic regression were used to identify the most important contributing factor in the random forest and logistic regression models, respectively. Of the 19 transcripts that remained in machine learning models, ZYX was identified as the most important contributor (Figure 4) to FSGS, MN, and MCD in both the random forest (balanced) (Figure 4A) and weighted logistic regression (Figure 4B) analyses. We performed train-test split with 0.7/0.3 and k-fold stratified cross-validation. The precision recall score for testing data and cross-validation score were high, indicating that the performance of both models was reliable.

Figure 4.

Figure 4

Open in a new tab

Feature importance ranking for transcripts enriched in the focal adhesion pathway. (A) Random forest analysis (Scikit-Learn in Python) identified ZYX as the highest-ranking transcript contributing to development of the common forms of nephrotic syndrome (FSGS, MN, and MCD). (B) Weighted logistic regression (Scikit-Learn in Python) confirmed that ZYX was the most important contributor for the three common forms of nephrotic syndrome.

GO Analysis Using Cytoscape and DAVID Bioinformatics Tools

Because transcripts identified in the focal adhesion-adherens junction pathway using R package (above) were primarily upregulated, we separated the upregulated transcripts (Supplemental Table 1-1) from downregulated transcripts (Supplemental Table 1-2) common to FSGS, MN, and MCD. Data were imported to Cytoscape and DAVID, respectively. Supplemental Figure 2 revealed that upregulated transcripts highlighted the GO Cellular Component hierarchical terms in Cytoscape (BiNGO v3.0.5): membrane → plasma membrane → plasma membrane part → cell junction → anchoring junction → adherens junction; and the GO Molecular Function hierarchical terms: protein binding → cytoskeletal protein binding → actin binding || protein binding → enzyme binding → GTPase binding → small GTPase binding → Ras GTPase binding → Rho GTPase binding. However, the downregulated transcripts did not result in any significant hits in Cytoscape (BiNGO v3.0.5) GO term mapping.

We performed the same procedure using the DAVID bioinformatics tool. GO Cellular Component term “plasma membrane” replicated and appeared to be the top hit (FDR=2.0×10−12) for the functional annotation clustering (Supplemental Table 3), covering 254 transcripts enriched in GO term “plasma membrane” (Supplemental Table 4).

GO Term Plasma Membrane-Focused Core Gene Search and Protein-Protein Interactions

To concentrate on core nodes of shared pathway networks covering FSGS, MN, and MCD, we chose a more stringent prioritization approach by selecting top 200 upregulated transcripts for each of FSGS, MN, and MCD (compared with CTL; ranked by FDR P-values in each group from Supplemental Table 1-1). The results were merged by gene Entrez ID yielding 72 gene IDs (Supplemental Table 5) with maximum FDR P<6.0×10−5 and minimum log2FC≥1.0. These represent the highest-ranked upregulated genes common to all three forms of adult nephrotic syndrome. We then explored the physical interactions of the proteins encoded by a smaller number of gene IDs to cover the central nodes that contribute to common mechanisms in these diseases. When Entrez IDs in Supplemental Tables 4 and 5 were merged, a concentrated gene list of 28 Entrez IDs for GO term “plasma membrane” was prioritized (Supplemental Table 6) for subsequent PPI analysis using the STRING bioinformatics tool (https://string-db.org/). Figure 5 shows PPI of the group of proteins encoded by the top 28 genes on the basis of existing knowledge. Since GO terms “membrane” and “plasma membrane” are located in the hierarchy of ancestor levels compatible to those gene IDs in Supplemental Table 4, we chose GO terms relevant to those implicated by Figure 3 and highlighted them in Figure 5 (see the legend). ZYX links to both the NOTCH1-TGFB1 functional cluster and the focal adhesion-cytoskeleton network, where GO terms regarding focal adhesion, adherens junction, anchoring junction, cell-cell junction, and actin cytoskeleton were highlighted.

Figure 5.

Figure 5

Open in a new tab

PPI nodes highlighted by impacted GO terms. A concentrated gene list of 28 gene IDs for GO term “plasma membrane” was prioritized (Supplemental Table 6) for PPI analysis using the STRING bioinformatics tool (https://string-db.org/). ZYX links to both the NOTCH1-TGFB1 functional cluster and the focal adhesion-cytoskeleton network, where GO terms regarding focal adhesion, adherens junction, anchoring junction, cell-cell junction, and actin cytoskeleton were highlighted.

Transcript and Protein Expression Patterns of Core Genes Mapped to the Implicated Pathway

To reveal the kidney cell types that express those core genes in the KEGG pathways (Figure 3) and gene products in the PPI network (Figure 5), we used a scRNA-seq dataset obtained from glomerular-enriched cells in normal kidney tissue from a patient who underwent nephrectomy to highlight those transcripts in known groups of kidney cells. The dataset has been openly available at the NCBI sequence read archive (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA781289) starting November 17, 2021, when submitted by Wake Forest University School of Medicine (Winston Salem, NC). The cells were classified into 14 categories using “leiden” algorithm with Python “scanpy” package (Supplemental Figure 3A). Thirteen of the 14 categories of cells were determined based on known cell markers (Supplemental Figure 3B). Figure 6 displays core transcript enrichment patterns in different kidney cell types using the “scanpy” package. ZYX is widely expressed in most kidney cells. ENG is mainly expressed in glomerular endothelial cells and mesangial cells. CRK is expressed in most kidney cells with lower levels in glomerular endothelial cells. Most of these transcripts appeared in the top hits shown in the focal adhesion pathway (Figure 3A), among which ACTN4 and ITGA3 are expressed in most known kidney cell types, particularly podocytes and mesangial cells.

Figure 6.

Figure 6

Open in a new tab

Gene expression profile of key nodes in predicted PPI and KEGG pathway top hits in scRNA-seq featured kidney cell types.

We queried the Human Protein Atlas (https://www.proteinatlas.org) image database for evidence of glomerular expression of core classifier genes (Figure 7) and found that many of them appear to exhibit podocyte, endothelial cell, parietal cell, or mesangial cell–specific expression in the glomerulus supported by scRNA-seq data (Figure 6).

Figure 7.

Figure 7

Open in a new tab

Protein distribution profile of key nodes in predicted PPI and KEGG pathway top hits in the human protein atlas.

Validation of Differential Expression of ZYX in Human Glomerulopathies Versus Normal Controls

Supplemental Table 7 shows that ZYX was upregulated in Europeans with FSGS and MCD (versus European controls) and Chinese with FSGS (versus Chinese controls) in the glomerular compartment gene expression datasets and in Italians with MN (versus Italian controls) in a renal biopsy gene expression dataset. ZYX was also upregulated in the tubular interstitial compartment gene expression dataset in the original NEPTUNE cohort for patients with FSGS, MN, and MCD (versus living kidney donor controls).

Discussion

Although FSGS, MN, and MCD have differing etiologies, they all display heavy proteinuria. The bioinformatics analyses herein identified dysregulation of adherens junction, focal adhesion, and actin cytoskeleton pathways across all three diseases. Compared with controls, actinin (ACTN4), filamin (FLNA), talin (TLN1), myosin II (MYH9), Arp2/3 (ARPC2/ARPC3), CRK, ZYX, integrin (ITGA), and extracellular matrix (ECM1) were upregulated in the glomerular compartment of patients with FSGS, MN, and MCD. All encoded proteins are major components of the focal adhesion and actin cytoskeleton pathways. This is consistent with the notion that actin binding and regulatory proteins, as well as integrins, are essential components of signaling and actin dynamics at focal adhesions in podocytes and suggest that they form a common path for development of proteinuric kidney disease.14 Focal adhesions are sites where cells connect to the ECM. They contain clusters of transmembrane integrin receptors tethered at one end to the ECM and the other to actin stress fibers, which are responsible for cell traction and ECM reorganization. Animal models of glomerular disease and podocyte cell culture firmly established the role of focal adhesion molecules and the actin cytoskeleton in maintaining proper podocyte functional structure with foot processes attached to the glomerular basement membrane (GBM).14

There are three layers of focal adhesion complex: bottom integrin signaling layer comprising proteins paxillin and focal adhesion kinase (FAK); middle force transduction layer proteins comprising talin (TLN1) and vinculin; and top layer proteins comprising ZYX and vasodilator-stimulated phosphoprotein (VASP).15 Although several molecules of adherens junction, focal adhesion, and cytoskeleton malfunction have been studied in common forms of nephrotic syndrome, including integrin,16 the role of ZYX has not been explored. In this study, ZYX was identified as the most important molecule within the focal adhesion pathway contributing to these common forms of nephrotic syndrome. ZYX is a LIM-domain containing protein found prominently at sites of cell adhesion. The LIM region of ZYX displays robust targeting to focal adhesion.17 When overexpressed in cells, the LIM region of ZYX causes displacement of endogenous ZYX from focal adhesions. Upon mislocalization of full-length ZYX, at least one member of the enabled (Ena)/VASP family is also displaced, and the organization of the actin cytoskeleton is perturbed.18 ZYX plays a key role in force-induced actin polymerization at focal adhesions through recruitment of Ena/VASP.19,20

Beyond serving in a structural role, ZYX has been shown to regulate actin dynamics. According to our KEGG pathway analysis for focal adhesion, ZYX interacts with α-actinin (ACTN4), which is a family of proteins that cross-link actin filaments, are highly enriched in podocyte foot processes, and serve as scaffolds for the assembly of large protein complexes. The interaction site has been identified as an extreme N-terminus of ZYX.21 With ZYX deletion of the ACTN4 binding site, the association of ZYX with focal adhesion plaques is impaired.22 Loss of ACTN4 disrupts interactions between the actin cytoskeleton and integrins, decreasing the overall strength of podocyte attachment to the GBM. Mutations in ACTN4 are known to cause FSGS.23,24 In addition, dysregulation of ACTN4 was noted to occur in early stages of MCD and MN.25-27 This suggests that ZYX may function through ACTN4 to affect all three forms of nephrotic syndrome.

In the single-cell analysis, ZYX was widely expressed in podocytes, glomerular mesangial cells, and parietal cells. Thus, ZYX-associated focal adhesion malfunction might implicate multiple kidney cell types in development of nephrotic syndrome. Jefferson et al. and Shankland et al. implicated glomerular parietal cells in development of FSGS, collapsing glomerulopathy, and crescentic glomerulopathies.28,29 Cultured mesangial cells possess abundant F-actin–containing fibers demonstrating the characteristic disposition of stress fibers at the edge of the cytoplasmic extensions, representing focal adhesion plaques where ZYX is primarily localized.30

Our results also show that ZYX interacts with CRK, upregulated in FSGS, MN, and MCD glomeruli (Figure 3). CRK is a family of adaptors required for formation of focal adhesions31 and expressed in podocytes and glomerular parietal cells. Several signaling transduction proteins are localized in focal adhesions, including FAK, CRK-associated substance (Cas), integrin-linked kinase, and small GTPases of both Ras (Figure 3B) and Rho families, including Rap1 (Figure 2), Rac1, Cdc42, and Rho GTPases,32 some of which were upregulated in our pathway analysis. Deletion of CRK in mice prevents proteinuria and foot process effacement.33 In humans, FAK-Cas–mediated CRK-dependent signaling is present in MN and MCD.33 In addition, ZYX also interacts with endoglin (ENG) and junction plakoglobin (JUP). ENG is expressed in glomerular mesangial cells and endothelial cells and regulates cytoskeletal organization through binding to zyxin-related protein 1.34 JUP is a junctional and membrane-associated plaque protein that influences the arrangement and function of both cytoskeleton and cells within the tissue.35,36

It is worth noting that podocyte-specific Tln1-knockout mice developed proteinuria and kidney failure, resulting in death within 10 weeks,37 although TLN1 was upregulated in the focal adhesion pathway in this study. TLN1, located at the middle force transduction layer of the focal adhesion complex, is key for integrin activation and directly relaying signals from focal adhesion to the ECM. However, mice with podocytes lacking Tln1 expression demonstrated mild decreases in adhesion and β1 integrin activation compared with control littermates. The dramatic and defective organization of the F-actin cytoskeleton in podocyte-specific Tln1-knockout mice was thought to be the culprit.37 The upregulation of TLN1 in our analysis is not contradictory to Tln1-knockout mice developing kidney failure. It likely relates to the fact that the three groups of patients in this study (FSGS, MCD, MN) were enrolled at the screening stage with median eGFR≥70 ml/min per 1.73 m2 (https://neptune-transmart.med.umich.edu/), when kidney function was fairly well compensated. It could also be the result of negative feedback from the mobilization of zyxin from focal adhesion to actin filament (ZYX upregulation) and the disturbed interaction between ZYX and the actin-cytoskeleton during early-stage nephrotic syndrome.

ZYX upregulation was consistently observed in several independent study cohorts (Supplemental Table 7), where TLN1 displayed a weaker trend of upregulation that missed statistical significance (Supplemental Table 8). This highlights the important role of ZYX in the human focal adhesion pathway for early stages of the three common forms of nephrotic syndrome and warrants further functional investigation of this molecule.

In conclusion, the present bioinformatics analyses identified a common pathway for the three major forms of adult nephrotic syndrome (FSGS, MN, MCD) encompassing focal adhesion and the actin cytoskeleton. To our knowledge, this is the first linkage of this pathway to these three common glomerular diseases through a differential gene expression analysis. This analysis was cross-sectional and cannot establish whether focal adhesion dysfunction caused or was a consequence of glomerular pathology. However, this study provides novel insights on ZYX in common forms of nephrotic syndrome. Future functional studies of ZYX protein are essential to illuminate its role in the development of FSGS, MN, and MCD and could lead to a potential therapeutic target.

Acknowledgment

D.L. is currently a graduate student of the Informatics and Analytics program at UNCG. He would like to thank Drs. David Bickel and Prashanti Manda for Bioinformatics course instructions. He would also like to thank program specialist Ms. Genevieve Smith for her academic advising. The authors are grateful to the NEPTUNE investigators who shared kidney biopsy gene expression data on the NCBI GEO database with the public. The authors are thankful to the National Science Foundation ACCESS program (project ID BIO220154) for computational resource support.

Footnotes

See related editorial, “Understanding Nephrotic Syndrome Using Kidney Transcriptome Profiling and Computational Studies,” on pages 431–433.

Disclosures

B.I. Freedman reports the following: Consultancy: AstraZeneca Pharmaceuticals, RenalytixAI, and Xinthera; Research Funding: AstraZeneca Pharmaceuticals and RenalytixAI; Patents or Royalties: Wake Forest University Health Sciences and have rights to a US patent related to APOL1 genetic testing; Advisory or Leadership Role: Editorial Boards: American Journal of Nephrology, JASN, and Kidney International; and Other Interests or Relationships: Chief Medical Officer, Health Systems Management, Inc. The remaining authors have nothing to disclose.

Funding

None.

Author Contributions

Conceptualization: Lijun Ma, Mariana Murea.

Formal analysis: DengFeng Li.

Investigation: Lijun Ma.

Methodology: DengFeng Li.

Software: DengFeng Li.

Supervision: Barry I. Freedman, Liang Liu, Lijun Ma.

Validation: Liang Liu.

Visualization: DengFeng Li.

Writing – original draft: Barry I. Freedman, Lijun Ma, Mariana Murea.

Writing – review & editing: Barry I. Freedman, Lijun Ma, Mariana Murea.

Data Sharing Statement

Previously published data were used for this study: Gene Expression Omnibus (GEO): GSE200828, NCBI sequence read archive (https://www.ncbi.nlm.nih.gov/sra/PRJNA781289, Accession: SRX13160927), Verification datasets in GEO: GSE104948, GSE129973, GSE115857, and GSE200818. Code Availability: R and Python codes are available at https://github.com/DennyLi-Github/Nephrotic-Syndrome.

Supplemental Materials

This article contains the following supplemental material online at http://links.lww.com/KN9/A316.

Supplemental Table 1-1. Upregulated genes common in FSGS, MN, and MCD.

Supplemental Table 1-2. Downregulated genes common in FSGS, MN, and MCD.

Supplemental Table 2. Top 3 KEGG pathways for GO all enriched terms.

Supplemental Table 3. Top 5 annotation clusters from 691 upregulated transcripts common in FSGS, MN, and MCD in Supplementary Table 1.

Supplemental Table 4. Gene IDs falling into GO Cellular Component “plasma membrane” from 691 upregulated transcripts in Supplementary Table 1-1.

Supplemental Table 5. Common gene IDs of top upregulated 200 transcripts for each of FSGS, MN, and MCD ranking by FDR from Supplementary Table 1-1.

Supplemental Table 6. Common gene IDs between Supplementary Table 4 and Supplementary Table 5.

Supplemental Table 7. ZYX expression in kidney biopsies from patients with glomerulopathies vs. normal living donor controls in geographically different study samples.

Supplemental Table 8. TLN1 expression in kidney biopsies from patients with glomerulopathies vs. normal living donor controls in geographically different study samples.

Supplemental Figure 1. RLE plot of gene expression for FSGS vs. CTL, MN vs. CTL, and MCD vs. CTL.

Supplemental Figure 2. Highlighted GO terms implicated by upregulated transcripts common in FSGS, MN, and MCD using Cytoscape “BinGO” application.

Supplemental Figure 3A. Kidney cell clusters grouped by distinguished markers using Python “scanpy” package.

Supplemental Figure 3B. Kidney cell types featured by unique transcripts from scRNA-seq (Python “scanpy” package).

References

  • 1.Hull RP, Goldsmith DJA. Nephrotic syndrome in adults. BMJ. 2008;336(7654):1185–1189. doi: 10.1136/bmj.39576.709711.80 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Genovese G Friedman DJ Ross MD, et al. Association of trypanolytic ApoL1 variants with kidney disease in African Americans. Science. 2010;329(5993):841–845. doi: 10.1126/science.1193032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Nichols B Jog P Lee JH, et al. Innate immunity pathways regulate the nephropathy gene Apolipoprotein L1. Kidney Int. 2015;87(2):332–342. doi: 10.1038/ki.2014.270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Abid Q Best Rocha A Larsen CP, et al. APOL1-associated collapsing focal segmental glomerulosclerosis in a patient with stimulator of interferon genes (STING)-associated vasculopathy with onset in infancy (SAVI). Am J Kidney Dis. 2020;75(2):287–290. doi: 10.1053/j.ajkd.2019.07.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kudose S Santoriello D Bomback AS, et al. Longitudinal outcomes of COVID-19-associated collapsing glomerulopathy and other podocytopathies. J Am Soc Nephrol. 2021;32(11):29582021–29582969. doi: 10.1681/ASN.2021070931 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Beck LH Bonegio RGB Lambeau G, et al. M-type phospholipase A2 receptor as target antigen in idiopathic membranous nephropathy. New Engl J Med. 2009;361(1):11–21. doi: 10.1056/nejmoa0810457 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Stanescu HC Arcos-Burgos M Medlar A, et al. Risk HLA-DQA1 and PLA(2)R1 alleles in idiopathic membranous nephropathy. New Engl J Med. 2011;364(7):616–626. doi: 10.1056/nejmoa1009742 [DOI] [PubMed] [Google Scholar]
  • 8.Bertelli R, Bonanni A, Caridi G, Canepa A, Ghiggeri GM. Molecular and cellular mechanisms for proteinuria in minimal change disease. Front Med (Lausanne). 2018;5:170. doi: 10.3389/fmed.2018.00170 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Jia X Yamamura T Gbadegesin R, et al. ; Research Consortium on Genetics of Childhood Idiopathic Nephrotic Syndrome in Japan, Korean Consortium of Hereditary Renal Diseases in Children, Midwest Pediatric Nephrology Consortium Genetics of Nephrotic Syndrome Study Group, NEPHROVIR. Common risk variants in NPHS1 and TNFSF15 are associated with childhood steroid-sensitive nephrotic syndrome. Kidney Int. 2020;98(5):1308–1322. doi: 10.1016/j.kint.2020.05.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Lane BM Chryst-Stangl M Wu G, et al. Steroid-sensitive nephrotic syndrome candidate gene CLVS1 regulates podocyte oxidative stress and endocytosis. JCI Insight. 2022;7(2):e152102. doi: 10.1172/jci.insight.152102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Sealfon R Mariani L Avila-Casado C, et al. Molecular characterization of membranous nephropathy. J Am Soc Nephrol. 2022;33(6):1208–1221. doi: 10.1681/ASN.2021060784 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Shannon P Markiel A Ozier O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–2504. doi: 10.1101/gr.1239303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2008;4(1):44–57. doi: 10.1038/nprot.2008.211 [DOI] [PubMed] [Google Scholar]
  • 14.Sever S, Schiffer M. Actin dynamics at focal adhesions: a common endpoint and putative therapeutic target for proteinuric kidney diseases. Kidney Int. 2018;93(6):1298-1307. doi: 10.1016/j.kint.2017.12.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Legerstee K, Houtsmuller AB. A layered view on focal adhesions. Biology (Basel). 2021;10(11): 1189. doi: 10.3390/biology10111189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lennon R, Randles MJ, Humphries MJ. The importance of podocyte adhesion for a healthy glomerulus. Front Endocrinol. 2014;5:160. doi: 10.3389/fendo.2014.00160 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Luo S, Schaefer AM, Dour S, Nonet ML. The conserved LIM domain-containing focal adhesion protein ZYX-1 regulates synapse maintenance in Caenorhabditis elegans. Development. 2014;141(20):3922–3933. doi: 10.1242/dev.108217 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Nix DA Fradelizi J Bockholt S, et al. Targeting of zyxin to sites of actin membrane interaction and to the nucleus. J Biol Chem. 2001;276(37):34759–34767. doi: 10.1074/jbc.m102820200 [DOI] [PubMed] [Google Scholar]
  • 19.Barzik M Kotova TI Higgs HN, et al. Ena/VASP proteins enhance actin polymerization in the presence of barbed end capping proteins. J Biol Chem. 2005;280(31):28653–28662. doi: 10.1074/jbc.m503957200 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hirata H, Sokabe M, Lim CT. Molecular mechanisms underlying the force-dependent regulation of actin-to-ECM linkage at the focal adhesions. Prog Mol Biol Transl Sci. 2014;126:135–154. doi: 10.1016/B978-0-12-394624-9.00006-3 [DOI] [PubMed] [Google Scholar]
  • 21.Bartram MP Habbig S Pahmeyer C, et al. Three-layered proteomic characterization of a novel ACTN4 mutation unravels its pathogenic potential in FSGS. Hum Mol Genet. 2016;25(6):1152–1164. doi: 10.1093/hmg/ddv638 [DOI] [PubMed] [Google Scholar]
  • 22.Li B, Trueb B. Analysis of the alpha-actinin/zyxin interaction. J Biol Chem. 2001;276(36):33328–33335. doi: 10.1074/jbc.m100789200 [DOI] [PubMed] [Google Scholar]
  • 23.Kaplan JM H Kim S North KN, et al. Mutations in ACTN4, encoding alpha-actinin-4, cause familial focal segmental glomerulosclerosis. Nat Genet. 2000;24(3):251–256. doi: 10.1038/73456 [DOI] [PubMed] [Google Scholar]
  • 24.Khurana S Chakraborty S Lam M, et al. Familial focal segmental glomerulosclerosis (FSGS)-linked α-actinin 4 (ACTN4) protein mutants lose ability to activate transcription by nuclear hormone receptors. J Biol Chem. 2012;287(15):12027–12035. doi: 10.1074/jbc.m112.345421 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Blaine J, Dylewski J. Regulation of the actin cytoskeleton in podocytes. Cells. 2020;9(7):1700. doi: 10.3390/cells9071700 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Liu Z Blattner SM Tu Y, et al. α-Actinin-4 and CLP36 protein deficiencies contribute to podocyte defects in multiple human glomerulopathies. J Biol Chem. 2011;286(35):30795–30805. doi: 10.1074/jbc.m111.255984 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Goode NP, Shires M, Khan TN, Mooney AF. Expression of alpha-actinin-4 in acquired human nephrotic syndrome: a quantitative immunoelectron microscopy study. Nephrol Dial Transplant. 2004;19(4):844–851. doi: 10.1093/ndt/gfg620 [DOI] [PubMed] [Google Scholar]
  • 28.Jefferson JA, Shankland SJ. The pathogenesis of focal segmental glomerulosclerosis. Adv Chronic Kidney Dis. 2014;21(5):408–416. doi: 10.1053/j.ackd.2014.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Shankland SJ, Anders H-J, Romagnani P. Glomerular parietal epithelial cells in kidney physiology, pathology, and repair. Curr Opin Nephrol Hypertens. 2013;22(3):302–309. doi: 10.1097/mnh.0b013e32835fefd4 [DOI] [PubMed] [Google Scholar]
  • 30.Cortes P Méndez M Riser BL, et al. F-actin fiber distribution in glomerular cells: structural and functional implications. Kidney Int. 2000;58(6):2452–2461. doi: 10.1046/j.1523-1755.2000.00428.x [DOI] [PubMed] [Google Scholar]
  • 31.Antoku S, Mayer BJ. Distinct roles for Crk adaptor isoforms in actin reorganization induced by extracellular signals. J Cell Sci. 2009;122(22):4228–4238. doi: 10.1242/jcs.054627 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Epifano C, Perez-Moreno M. Crossroads of integrins and cadherins in epithelia and stroma remodeling. Cell Adh Migr. 2012;6(3):261–273. doi: 10.4161/cam.20253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.George B Verma R Soofi AA, et al. Crk1/2-dependent signaling is necessary for podocyte foot process spreading in mouse models of glomerular disease. J Clin Invest. 2012;122(2):674–692. doi: 10.1172/jci60070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Sanz-Rodriguez F, Guerrero-Esteo M, Botella L-M, Banville D, Vary CPH, Bernabéu C. Endoglin regulates cytoskeletal organization through binding to ZRP-1, a member of the Lim family of proteins. J Biol Chem. 2004;279(31):32858–32868. doi: 10.1074/jbc.m400843200 [DOI] [PubMed] [Google Scholar]
  • 35.Swope D, Li J, Radice GL. Beyond cell adhesion: the role of armadillo proteins in the heart. Cell Signal. 2013;25(1):93–100. doi: 10.1016/j.cellsig.2012.09.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Fang Y Li F Qi C, et al. Plakoglobin is involved in cytoskeletal rearrangement of podocytes under the regulation of UCH-L1. Biochem Biophys Res Commun. 2020;529(1):112–118. doi: 10.1016/j.bbrc.2020.05.093 [DOI] [PubMed] [Google Scholar]
  • 37.Tian X Kim JJ Monkley SM, et al. Podocyte-associated talin1 is critical for glomerular filtration barrier maintenance. J Clin Invest. 2014;124(3):1098–1113. doi: 10.1172/jci69778 [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.

Data Availability Statement

Previously published data were used for this study: Gene Expression Omnibus (GEO): GSE200828, NCBI sequence read archive (https://www.ncbi.nlm.nih.gov/sra/PRJNA781289, Accession: SRX13160927), Verification datasets in GEO: GSE104948, GSE129973, GSE115857, and GSE200818. Code Availability: R and Python codes are available at https://github.com/DennyLi-Github/Nephrotic-Syndrome.

Articles from Kidney360 are provided here courtesy of American Society of Nephrology

RESOURCES