for more information:ali.ma@wecistanche.com

Identification and verification of vascular cell adhesion protein 1 as an immune-related hub gene associated with the tubulointerstitial injury in diabetic kidney disease

Yan Jia, et al

ABSTRACT

Diabetic kidney disease (DKD) is the leading cause of chronic kidney disease (CKD) and end-stage renal disease (ESRD), but the pathogenesis is not completely understood. Tubulointerstitial injury plays a critical role in the development and progression of DKD (Diabetic kidney disease). The present study aimed to investigate the profile of tubulointerstitial immune cell infiltration and reveal the underlying mechanisms between tubular cell injury and interstitial inflammation in DKD (Diabetic kidney disease) using bioinformatics strategies. First, xCell analysis identified immune cells displaying significant changes in the DKD (Diabetic kidney disease) tubulointerstitium, including upregulated CD4+ T cells, Th2 cells, CD8+ T cells, M1 macrophages, activated dendritic cells (DCs), and conventional DCs, as well as downregulated Tregs. Second, pyroptosis was identified as the main form of cell death compared with other forms of programmed cell death. Vascular cell adhesion protein 1 (VCAM1) was identified as the top-ranked hub gene. The correlation analysis showed that VCAM1 was significantly positively correlated with pyroptosis and infiltrated immune cells in the tubulointerstitium. Upregulation of VCAM1 in the DKD tubulointerstitium was further verified in the European Renal cDNA Bank cohort and was observed to negatively correlate with the glomerular filtration rate (GFR). Our in vitro study validated increased VCAM1 expression in HK-2 cells under diabetic conditions, and pyroptosis inhibition by disulfiram decreased VCAM1 expression, inflammatory cytokine release, and fibrosis. In conclusion, our study identified upregulated VCAM1 expression in renal tubular cells, which might interact with infiltrated immune cells, thus promoting fibrosis. The FDA-approved drug disulfiram might improve fibrosis in DKD (Diabetic kidney disease) by targeting tubular pyroptosis and VCAM1 expression.

KEYWORDS: DKD; tubulointerstitium; immune cells; pyroptosis; VCAM1; disulfiram

Cistanche can treat Diabetic kidney disease

Click to Cistanche for kidney disease

Introduction

Diabetic kidney disease (DKD), one of the common complications of diabetes, is the main cause of chronic kidney disease (CKD) and end-stage renal disease (ESRD) in many developed and developing countries [1,2]. According to a report from the China Kidney Disease Network (CKNET), DKD (Diabetic kidney disease) accounts for 26.70% of all cases of CKD and imposes a large medical and economic burden [2]. The pathogenesis of DKD is complex and involves a multitude of different pathways. Clarifying the pathological characteristics and pathogenesis of DKD will help improve clinical management and identify new therapeutic targets.

Although glomerular damage is the main pathological feature of DKD (Diabetic kidney disease), increasing evidence has recently revealed that tubulointerstitial pathology [3,4], such as tubular atrophy, interstitial fibrosis, and interstitial infiltrating immune cells, plays a critical role in the development and progression of DKD (Diabetic kidney disease) [5–7]. Bioinformatics analysis represents an effective method to process large amounts of data within an extremely short time and provides valuable information about the disease. However, bioinformatics studies investigating tubulointerstitial gene expression and immune cell infiltration in DKD are relatively rare.

The purpose of the study was to describe the characteristics of DKD (Diabetic kidney disease) tubulointerstitial immune cell infiltration and to identify some key immune and inflammatory genes to provide novel insights into the pathogenesis and therapy of DKD. First, a tubulointerstitial microarray dataset of DKD (Diabetic kidney disease) was downloaded from the Gene Expression Omnibus (GEO) database. Using xCell [8], a web-based tool that performs a cell type enrichment analysis of gene expression data for 64 immune and stromal cell types, we first investigated the differences in immune cell infiltration between kidney tissues from individuals with DKD (Diabetic kidney disease) and normal controls. Second, a list of immune-related differentially expressed genes (DEGs) was screened, among which the top-ranked gene, vascular cell adhesion protein 1 (VCAM1), was identified as an immune hub gene in DKD (Diabetic kidney disease). Then, gene sets associated with various forms of cell death were obtained from GeneCards, and pyroptosis was evaluated as the main form of cell death using gene set variation analyses (GSVA) and gene set enrichment analysis (GSEA). Furthermore, a correlation analysis showed that VCAM1 was positively correlated with tubulointerstitial immune cell infiltration and pyroptosis. The upregulated expression of VCAM1 in the DKD tubulointerstitium was further verified in the European Renal cDNA Bank (ERCB) cohort and was observed to negatively correlate with renal function in patients with DKD. An in vitro study validated increased VCAM1 expression in HK-2 cells cultured under diabetic conditions, and pyroptosis inhibition by the FDA-approved drug disulfiram decreased VCAM1 expression, inflammatory cytokine release, and fibrosis. Therefore, tubular pyroptosis and upregulated VCAM1 expression might be targets for immunotherapy of DKD.

Methods

Microarray data

The human gene expression dataset GSE30529 was downloaded from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO, database. GSE30529 (GPL571 [HGU133A_2] Affymetrix Human Genome U133A 2.0 Array) consisted of 10 DKD (Diabetic kidney disease) tubule samples and 12 control samples

Data processing

The raw data in the GSE30529 dataset were down-loaded and processed using the Limma package and Oligo package, respectively [9,10]. Data processing included background correction, normalization, and expression calculation. When multiple probes were mapped to one gene symbol, the average expression level of the probes was calculated and regarded as the gene expression level of that gene. The probes that did not map to genes were removed.

Cistanche can treat Diabetic kidney disease

Immune and stromal cell analyses

We applied a novel gene signature-based method, xCell [8], to estimate the cell type enrichment score and determine the profile of infiltrating immune cells in the tubulointerstitium of patients with DKD (Diabetic kidney disease). xCell is a method for cell-type enrichment analysis using single-sample gene set enrichment analysis (ssGSEA) that calculates the enrichment scores for 64 cell types, including 34 types of immune cells, 30 types of stromal cells, and other cells. It outperforms other extensively in silico analyses (including CIBERSORT) by encompassing more types of immune cells and employing a spillover compensation technique to reduce dependencies between closely related cell types. The 34 types of immune cells were categorized into nine groups, including CD4+T cell subpopulations, CD8+ T cell subpopulations, gamma delta T cells (Tγδ cells), NK cells, NKT cells, B cell subpopulations, monocyte/macrophage subpopulations, dendritic cell (DC) subpopulations, and granulocyte subpopulations.

Gene set enrichment analyses (GSEA) and gene set variation analyses (GSVA)

Using the hallmark gene sets as the reference gene set, we performed gene set variation analyses (GSVA) between DKD (Diabetic kidney disease) and control tissues using the GSVA Bioconductor package [11]. The thresholds were set to enrichment score change > 1.0, p-value < 0.05 and t value >2. Both upregulated and downregulated pathways were identified. GSEA was performed with the GSEA desktop application [12], and a false discovery rate (FDR) <0.25 and p-value <0.05 were set as the thresholds.

Construction of gene sets representing various forms of cell death and enrichment analysis

Gene sets representing various forms of cell death were constructed by searching the keywords ‘pyroptosis’, ‘necrosis’, ‘necroptosis’, ‘apoptosis’, ‘ferroptosis’, and ‘autophagy’ in the GeneCards database (https://www.genecards.org/). Nonprotein-coding genes were removed. The gene sets are listed in Table S2. Then, gene set GSEA was performed using the GSEA desktop application to evaluate the cell death form in the tubulointerstitium. We performed Pearson’s correlation analyses to explore the relationship between pivotal cell death pathways and infiltrating immune cells in the tubulointerstitium

Analysis of DEGs and screening of immune hub genes

DEGs between patients with DKD (Diabetic kidney disease) and healthy controls were identified using the ‘Limma’ package [13] in R software. An adjusted P value<0.05 and log fold change (FC) ≥ 1 were set as the thresholds. A list of 1793 immune-related genes was down-loaded from ImmPort Shared Data (https://www. immport.org/shared/genelists), and shared genes from the two datasets (DEGs and immunerelevant gene list) were determined by constructing a Venn diagram to identify immune-related hub genes. An analysis of the functional interactions between proteins might provide insights into the mechanisms of DKD (Diabetic kidney disease). In the present study, the protein-protein interaction (PPI) networks of ninety-four shared genes were generated using the STRING database (https://string-db.org/) and visualized with Cytoscape software (version 3.7.1, http://www.cytoscape.org/). An interaction score greater than 0.4 (medium confidence) was considered statistically significant. CytoHubba, a Cytoscape plugin, was utilized to explore PPI network hub genes [14]. Among them, VCAM1 was the top-ranked gene. Enrichment analyses of the thirty immune hub genes identified above were performed using the online tool Metascape [15]. Furthermore, dominant modules in the PPI network were identified using the MCODE plugin (version 1.4.2, http://apps.cytoscape.org/apps/ MCODE) in Cytoscape software.

VCAM1 verification and correlation analysis with renal function

The differential expression of VCAM1 in the tubulointerstitium of patients with DKD (Diabetic kidney disease) was validated in the ERCB cohort (31 healthy controls and 17 patients with DKD). Then, we used the Nephroseq v5 online database (http://v5.nephroseq.org), an integrated data-mining platform for gene expression datasets of kidney diseases, to validate the correlation between VCAM1 expression and clinical traits of patients with DKD using Pearson’s correlation analysis in ERCB. A p-value of <0.05 was considered statistically significant

Immunofluorescence staining for VCAM1

Paraffin kidney sections from a patient clinicopathologically diagnosed with DKD (Diabetic kidney disease) were stained with VCAM1. This was approved by the Committee on Research Ethics of Peking University First Hospital (NO. 20171280). Nonspecific binding was blocked with 3% BSA following fixation and heat treatment for antigen retrieval. Kidney sections were then incubated with a primary antibody against VCAM1 (Abcam, ab134047, 1:200) overnight at 4°C. The Cy3-labeled secondary antibody was obtained from Jackson ImmunoResearch. Tumor-adjacent kidney tissue from nephrectomy samples was used as normal control. Representative images were captured using a Leica DFC 7000 T camera via Leica Application Suite V4.7.1 software.

Cell culture

Human kidney 2 (HK-2) cells, a human kidney proximal tubular epithelial cell line, were cultured in DMEM (11885084, Gibco, USA) supplemented with 10% fetal bovine serum (10099141 C, Gibco, USA) and 1% penicillin/streptomycin (10378016, Gibco, USA) at 37°C with 5% CO2. High glucose (30 mM D-glucose, G8644, Sigma) [16] and TNFα (40 ng/ml, 300–01A, PeproTech) [17] was used to induce HK-2 cell injury in vitro. Disulfiram (DSL, 0.3 μM [18,19], NSC190940, Selleck), a pore-formation inhibitor, was added 3 hours prior to the induction of HK-2 cell injury.

Western blot

Total protein was extracted from HK-2 cells using RIPA buffer (Sigma, R0278), and the protein concentration was determined using a Pierce BCA Protein Assay kit (Thermo Fisher Scientific, 23227). Next, denatured proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS– PAGE) and then electrically transferred to polyvinylidene difluoride membranes (Millipore, IPVH00010). The membranes were blocked for 60 minutes with 5% fat-free milk dissolved in Tris-buffered saline containing 0.1% Tween 20 (TBST). The blots were incubated with the following relevant primary antibodies overnight at 4°C: GSDMD (Abcam, ab209845, 1:1000), VCAM1 (Abcam, ab134047, 1:2000), and tubulin (ZSBio, TA-10, 1:5000). Incubation with a 1:1000 dilution of the HRP-conjugated secondary antibody was carried out for 1 hour at room temperature. After five washes with TBST, the membranes were incubated with the chemiluminescence substrate (Millipore, WBKLS0100) for 5 minutes, and images were captured using an Image Quant LAS 4000 Mini system (GE Healthcare). The semiquantitative analysis was conducted using ImageJ software (Media Cybernetics, Silver Spring, MD).

RNA isolation and RT–PCR analysis

Total RNA was extracted from HK-2 cells using an RNAsimple Total RNA Kit (Tiangen, DP419) according to the manufacturer’s instructions and reverse transcribed into cDNAs using a FastKing RT Kit (Tiangen, KR116). Quantitative PCR was performed using a StepOne Real-Time PCR System (Applied Biosystems, USA) with two-step methods. The sequences of the primers used are shown in Table S4. Comparative gene expression was calculated using the 2−ΔΔCt method as described previously.

Statistical analysis

GraphPad Prism 6.0 software was used for statistical analyses, and data are presented as the means ± SEM. A two-tailed unpaired t-test was applied for comparisons between two groups. Differences at the P < 0.05 level were considered statistically significant.

Cistanche can treat Diabetic kidney disease

Results

Accumulating evidence has recently revealed that tubulointerstitial pathology plays a critical role in the development and progression of DKD (Diabetic kidney disease). The purpose of the study was to describe the characteristics of DKD (Diabetic kidney disease) tubulointerstitial immune cell infiltration and to identify some key immune and inflammatory genes to provide novel insights into the pathogenesis and therapy of DKD. In this study, we investigated the profile of tubulointerstitial immune cell infiltration and identified pyroptosis as the main form of programmed cell death in DKD (Diabetic kidney disease)using bioinformatics analysis. VCAM1 was identified as the top-ranked immune-related hub gene and was positively correlated with pyroptosis and infiltrated immune cells. Furthermore, VCAM1 expression was validated to be elevated in renal tubular cells cultured under diabetic conditions. The FDA-approved drug disulfiram inhibited renal tubular cell pyroptosis and decreased VCAM1 expression, inflammatory cytokine levels, and fibrosis in vitro.

1. Bioinformatics analysis workflows

The main steps of the workflow are shown in Figure 1. We first evaluated immune cell infiltration in tubulointerstitial tissues from patients with DKD (Diabetic kidney disease). Then, pyroptosis, the main form of cell death in the tubulointerstitium, was discovered using GSEA and GSVA. Next, we screened a list of genes closely related to immune cell infiltration and identified hub genes. A correlation analysis was performed between infiltrated immune cells, pyroptosis, and the top hub gene VCAM1. In addition, clinical validation and immunostaining validation of VCAM1 were performed. Finally, we further conducted an in vitro study to validate the results of the bioinformatics analysis.

2. Immune and stromal cell deconvolution analyses

We used xCell, which generates cell type enrichment scores using bulk gene expression data, to determine the cell types potentially involved in tubulointerstitial injury in diabetic kidney disease. The enrichment scores of 64 cell types, including 34 types of immune cells and 30 types of stromal and other cells, were obtained for each sample (Figure 2(a)). The heatmap of 34 immune cell enrichment scores was illustrated to identify the immune landscape of the DKD (Diabetic kidney disease) tubulointerstitium (Figure 2(b)). Compared with normal controls, the majority of T cell enrichment scores were relatively higher in patients with DKD (Diabetic kidney disease), except for CD4+ Tcm, Tregs, Th1 cells, CD8+ naïve T cells, and NKT cells. The enrichment scores of aDCs and cDCs were higher in patients with DKD (Diabetic kidney disease) than in controls, while Pro-B cell enrichment scores were lower in patients with DKD. The enrichment scores of most monocyte/macrophage and granulocyte subsets were not significantly different between DKD (Diabetic kidney disease) and normal tissues, except for M1 macrophages and neutrophils, and eosinophils (Table 1).

Figure 2. Immune cell infiltration analysis by xCell in the tubulointerstitium of DKD (Diabetic kidney disease). (a) Comparison of cell scores of 64 cell types in the tubulointerstitium between DKD and control renal tissue in the GSE30529 dataset. (b) The cellular landscape of immune microenvironment in DKD (Diabetic kidney disease). The heatmap represents the cell type enrichment score of each immune cell type for all samples. DC, dendritic cell; Tcm, central memory T cell; Tem, effective memory T cell; Tγδ cell, gamma delta T cell.

3. GSVA and GSEA of hallmark gene sets

Weconducted GSVA and GSEA of hallmark gene set in the GSE30529 dataset to explore the biological processes in the DKD (Diabetic kidney disease) tubulointerstitium. The ‘interferon-alpha response’, ‘interferon-gamma response’ and ‘complement’ signaling pathways were significantly activated in the pathogenesis of tubulointerstitial injury in patients with DKD (Figure 3(a)), and a summary of the GSVA results is displayed in Table S1. Then, we performed GSEA of the three gene sets and found that they were positively enriched in the dataset. The normalized enrichment scores (NESs) were 1.89, 1.89 and 1.70, with normalized p values (NOM p values) of 0.000, 0.002 and 0.019, respectively (Figure 3(b-d)). All of the significantly enriched pathways identified using GSEA are shown in Table 2. These data suggest that the tubulointerstitial immune response plays important role in the pathogenesis of DKD (Diabetic kidney disease).

4. GSEA and GSVA of gene sets associated with various forms of cell death

Cell death includes pyroptosis, necrosis, necroptosis, apoptosis, ferroptosis, and autophagy, which have different pathophysiological mechanisms and signaling pathways. Various forms of programmed cell death have been reported to be associated with inflammatory and immune responses. We constructed gene sets of several types of cell death from GeneCards (Table S2) and then performed both GSEA and GSVA to explore the involvement of cell death in tubulointerstitial inflammation in patients with DKD (Diabetic kidney disease). GSEA showed that only the pyroptosis gene sets were significantly enriched in patients with DKD (Diabetic kidney disease) compared to normal controls (NES = 1.65, FDR q value = 0.023, FWER = 0.013) (Figure 3(e)). Meanwhile, the GSVA score for pyroptosis was the highest among all of the various types of cell death analyzed (Table S3). Based on this result, pyroptosis might play a role in the pathogenesis of tubulointerstitial injuries in patients with DKD (Diabetic kidney disease).

Figure 3. Gene set enrichment analyses (GSEA) and gene set variation analyses (GSVA). (a) Barplot of GSVA results of 50 hallmark gene sets. (b)-(d) GSEA results of ‘Hallmark_ interferon-alpha response’, ‘Hallmark_ interferon-gamma response’ and ‘Hallmark complement’. (e) GSEA result of pyroptosis.

5. Correlation between pyroptosis enrichment scores and immune cell infiltration

GSVA enrichment scores for pyroptosis were correlated with the proportion of immune cell infiltration by performing Pearson’s correlation analysis to evaluate the correlation between pyroptosis and infiltration of immune cells. We observed a significant correlation between 15 types of immune cells and pyroptosis (Figure 4), suggesting a close interaction between immune cells and pyroptosis.

6. Discovery of core genes

Differential expression analysis was performed to further explore the relationship between renal tubular cells and interstitial immune cells. Four hundred eight DEGs were identified between the DKD (Diabetic kidney disease) group and the control group, among which 65 were downregulated and 343 upregulated (Figure 5). An immune-related gene list was downloaded to identify immune-related DEGs, and ninety-four hub immune genes were acquired from two datasets (DEGs and the immune-related gene list) via a Venn diagram (Figure 6(a)). The PPI analysis of these ninety-four hub immune genes revealed 97 nodes and 694 interactions (Figure S1). The top 30 nodes calculated by the MCC algorithm were identified as hub genes (Figure 6(b) and Table 3). Then, an enrichment analysis of hub genes was performed. 'Allograft rejection', 'interferon-gamma response', 'inflammatory response', 'interferon-alpha response', 'epithelial-mesenchymal transition' and 'complement' were the most enriched items in the hallmark analysis. ‘Rheumatoid arthritis', 'cytokine receptor interaction’, 'Toll-like receptor signaling pathway', 'pertussis' and 'NF-kappa B signaling pathway' was the most enriched KEGG terms. 'Cellular response interferon gamma', 'chemokine activity', 'leukocyte cell-cell adhesion', and 'regulation of cytokine production' were the most enriched GO terms (Figure 6(c)). Furthermore, the MCODE scoring system identified two clusters with a score ≥5. PTRPC, TRIM22, PSMB8, KNG1, and CXCL6 were hub nodes with higher node degrees in module 1 (Figure 6(d)), and CD1C, EGF, FCER1G, CD3D, and CD48 were hub nodes in module 2 (Figure 6(e)).

7. Correlation between VCAM1 expression and infiltrating immune cells

Among the hub genes identified, the top ranked gene, VCAM1, was identified as an immune hub gene in DKD (Diabetic kidney disease). VCAM1 expression was significantly increased in patients with DKD (Diabetic kidney disease) (Figure 7 (a)). The correlation analysis revealed that VCAM1 expression was positively correlated with CD4+ T cells (r = 0.6877, P = 0.0004), CD4+ T memory cells (r = 0.7102, P = 0.0002), Th2 cells (r = 0.7358, P < 0.0001), CD8+ Tcm (r = 0.6026, P = 0.0030), Tγδ cells (r = 0.5864, P = 0.0041), NK cells (r = 0.5692, P = 0.0057), macrophages (r = 0.5068, P = 0.0161), M1 macrophages (r = 0.4441, P = 0.0384), aDCs (r = 0.4814, P = 0.0233) and cDCs (r = 0.6562, P = 0.0009) but negatively correlated with naïve CD8+ T cells (r = −0.4909, P = 0.0204) and NKT cells (r = −0.5548, P = 0.0074) (Figure 7(b)).

8. Correlation between VCAM1 expression and pyroptosis

Pearson’s correlation analysis was performed to explore whether the upregulated VCAM1 expression in the tubulointerstitium of patients with DKD (Diabetic kidney disease) is related to pyroptosis. VCAM1 expression was closely related to pyroptosis (r = 0.6118, P = 0.0025) (Figure 7(c)), indicating that pyroptosis might induce VCAM1 expression.

9. Verification, clinical relevance, and immunofluorescence validation of VCAM1 expression

We performed a relevant analysis of patients with DKD (Diabetic kidney disease) from ERCB to validate the change in VCAM1 expression in the DKD (Diabetic kidney disease) tubulointerstitium. Consistently, VCAM1 mRNA levels were significantly higher in patients with DKD (Diabetic kidney disease) (Figure 8(a)). VCAM1 expression was negatively correlated with the glomerular filtration rate (GFR) (r = −0.5920, P = 0.0123) (Figure 8(b)) and positively correlated with serum creatinine levels (r = 0.5189, P = 0.0328) (Figure 8(c)). Furthermore, we performed immunofluorescence staining for VCAM1 in the kidney tissue of a patient with DKD (Diabetic kidney disease) and found that renal tubular VCAM1 expression was significantly upregulated compared with that in normal kidneys (Figure 8(d)).

10. DSL inhibited HG-induced HK-2 cell pyroptosis, VCAM1 expression, and inflammation

We exposed cultured HK-2 cells to high glucose (HG) and TNF-α for 48 h to further determine the relationship between tubular cell pyroptosis, VCAM1 expression, and the tubulointerstitial immune response in individuals with DKD (Diabetic kidney disease). Since the core of cellular pyroptosis is membrane pore formation induced by cleaved GSDMD, we detected the levels of the GSDMD N-terminal domain (GSDMD-NT) and VCAM1 using western blotting. Both HG and TNF-α induced a significant increase in the GSDMD-NT level, indicating the occurrence of GSDMD-mediated pyroptosis (Figure 9(a)). VCAM1 expression was increased statistically significantly in the HG and TNF-α groups (Figure 9(b)). Disulfiram (DSL) is an effective inhibitor of GSDMD pore formation [18]. Pretreatment with DSL alleviated the increase in GSDMD-NT levels, indicating that cell pyroptosis was suppressed. Meanwhile, VCAM1 expression was decreased in the DSL-pretreated HG and TNF-α groups (Figure 9(a-b)). Next, the inflammatory cytokine profile was determined using qPCR. The results showed that the upregulated levels of the MCP-1, IL-1β, and IL-18 mRNAs induced by HG were significantly decreased after DSL pretreatment. In the TNF-α group, the level of IL-1β decreased significantly in the DSL-pretreated group, and the levels of MCP-1 and IL-18 did not change significantly. No significant changes in the levels of the IL-6 and IL-8 mRNAs were found in any group. Importantly, TGF-β1 mRNA levels were also decreased in the DSL pretreated groups. Thus, pyroptotic tubular cells increased VCAM1 expression under diabetic conditions, and a treatment abrogating pyroptosis inhibited VCAM1 expression, inflammation, and fibrosis in HK-2 cells.

Cistanche can treat Diabetic kidney disease

Click here to Part II

From: 'Identification and verification of vascular cell adhesion protein 1 as an immune-related hub gene associated with the tubulointerstitial injury in diabetic kidney disease' by Yan Jia, et al

---BIOENGINEERED 2021, VOL. 12, NO. 1, 6655–6673