Ptpn6^+/mev mice maintained on a C57BL/6 background were purchased from Jackson Laboratories. These mice exhibit a T-to-A mutation at a splice consensus site. We designed primers (Supplementary Table 1) containing the mutation site in the middle of the PCR amplicon to genotype wild-type (^+/+), heterozygous (^+/mev), and homozygous (^mev/mev) mice with sequence maps. All animal experiments were undertaken following the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals with the approval of the institutional research ethics committee of Navy Medical University, Shanghai, China.

A mouse renal IRI model was established with three 7-week-old Ptpn6^+/mev mice and three wild-type littermates. The steps for I/R model establishment are shown in Figure 1A, as described previously (14). Before surgery, the mice were anesthetized by intraperitoneal injection with 100 mg/kg pentobarbital sodium. Each mouse was placed on a temperature-controlled heating blanket and fitted with a rectal thermometer probe connected to a thermal feedback controller (RWD Life Science, China), and the rectal temperature was maintained at 37 ± 0.5°C. An abdominal midline incision was made, and then the right kidney of the mice was resected. The pedicle of the left kidney was clamped for 34 min with a non-traumatic vascular clamp before intraperitoneal injection of heparin (50 U/kg). During the clamping time, the abdominal midline incision was covered with a surgical dressing to keep the kidney warm and hydrated. The sham controls underwent similar procedures except for left kidney pedicle clamping. At 24 h post-IRI, the mice were sacrificed, and the left kidneys were removed for histologic and RNA analyses. For fenofibrate intervention, three mice were i.p. injected with 100 mg/kg/d fenofibrate (Sigma-Aldrich, Germany) and three mice were i.p. injected with corn oil (MedChemExpress, United States) from day 0 to day 20 before IRI surgery.

Left renal kidney specimens were processed and sectioned as reported previously (15). Then, the sections were stained with hematoxylin-eosin (H and E) for histomorphological examination. A semiquantitative method was used to score the percentage of tubular damage according to the following metric: score 1, less than 10%; score 2, 10% to less than 25%; score 3, 25% to 75%; score 4, higher than 75% of cortex with tubular damage. At least three fields per section were analyzed by two different pathologists.

For western blot analysis, proteins were extracted from mouse kidney tissues and processed according to a routine procedure. The antibodies used in this experiment were SHP-1 (E1U6R) rabbit monoclonal antibody and β-actin (13E5) rabbit monoclonal antibody (Cell Signaling Technology, United States), which were diluted as 1:1000. Images were obtained with a DM IL LED microscope (Leica Microsystems, Germany).

Apoptosis was confirmed with a terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) staining assay. TUNEL staining was conducted according to the kit manufacturer’s instructions (Beyotime, China).

Total RNA samples from the left kidney tissues were isolated with TRIzol reagent (Thermo Fisher Scientific, United States). An Agilent 2100 Bioanalyzer was used to characterize the quality of the in vitro RNA transcripts. The RNA integrity numbers (RINs) of all samples were greater than 8.0.

The poly-A-containing mRNA molecules were purified with poly-T oligo-attached magnetic beads using two rounds of purification. A SuperScript Double-Stranded cDNA Synthesis Kit (Thermo Fisher Scientific, United States) was used to synthesize the double-stranded cDNA. Further library preparation was performed using a TruSeq™ RNA Sample Preparation Kit (cat# FC-122-1001, Illumina, United States). The libraries were sequenced as 2 × 150 bp single reads using an Illumina HiSeq 2000 according to the manufacturer’s instructions. We removed adaptor, low-complexity, and low-quality sequences from the raw reads. The remaining clean reads were used for further analyses.

TopHat v2.1.0 was used with the default parameters to generate acceptable alignments for Cufflinks, which was used to align the RNA sequencing (RNA-seq) paired-end reads against the reference genome, Ensembl release 90 GRCm38.p5 (16). The expression of the annotated genes in the RNA-seq data was evaluated in fragments per kilobase million (FPKM) using Cufflinks. The following formula was used to calculate the FPKM value: FPKM = (number of mapped fragments) × 10³ × 10⁶/[(length of transcript) × (total number of fragments)]. Log transformation and zero-mean normalization were used to normalize the expression data for comparisons. The false discovery rate (FDR) of <0.05, after applying Benjamini-Hochberg correction, was chosen for determining significant DEGs.

An online analysis tool, the Protein ANalysis THrough Evolutionary Relationships (PANTHER) classification system, was used to annotate the DEGs into three major Gene Ontology (GO) domains: the molecular function, biological process, and cellular component domains (17). We also used the STRING system to analyze the most significantly (FDR < 0.005) enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in DEGs (18, 19). STRING is an online analysis tool that can provide customized protein-protein networks and functional characterization of user-uploaded gene/measurement sets. The results of enrichment analysis were presented after Bonferroni correction for multiple comparisons.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR) was carried out with a LightCycler^® 480 II real-time RT-PCR system (Roche, Switzerland). RNA samples from the kidneys of 3 knockdown mice and 3 wild-type littermates after renal IRI were used in reverse transcription reactions. cDNA was synthesized using an oligo-dT reverse primer and a PrimeScript™ RT reagent kit (Takara, Japan). The primers used for RT-PCR validation of genes are listed in Supplementary Table 1. The expression of the GAPDH housekeeping gene was used to normalize the data. Gene expression was quantified with the 2^–ΔΔCT method, and the results are expressed as fold change (FC) relative to the levels in the corresponding control samples (20).

Detection of ROS was performed on frozen sections using dihydroethidium (D7008, Sigma-Aldrich, United States) which was diluted as 1:500 for 30 min at 37°C in a dark incubator. DAPI (4′, 6-diamidino-2-phenylindole, Beyotime, China) was added after the slides were dry and incubated for 10 min in the dark at room temperature. Slides were washed three times for 5 min each in phosphate buffered saline (pH 7.4) on a shaker. After the plate was blocked with an antifade mounting medium (Servicebio, China), the images were observed and acquired under a fluorescence microscope.

GraphPad Prism 8.0 (GraphPad Software Inc., United States) was used to perform statistical analyses. All data are presented as the mean ± standard deviation. We performed Student’s unpaired t-tests for comparisons between two groups. All experiments were repeated more than three times. P < 0.05 was considered statistically significant.

First, we validated the mutation and protein expression of SHP-1 in Ptpn6^+/mev mice by Sanger sequencing and western blotting, respectively (Supplementary Figures 1A–C). According to the process shown in Figure 1A, we then established IRI models in the wild-type ^+/+ and Ptpn6^+/mev groups and performed RNA sequencing and data analysis. Mice in the Ptpn6^+/mev group exhibited more severe injury than those in the wild-type^+/+ group, with higher serum creatinine (CR) and urea nitrogen (BUN) levels (Figures 1B,C). H and E staining revealed that tubular injury was significantly worse in the Ptpn6^+/mev group than in the wild-type group (Figure 1D). The TUNEL-positive areas of the kidneys were more intensely stained in Ptpn6^+/mev mice than in wild-type mice, and the positive areas were larger (Figure 1E). Quantitative results of H&E staining and TUNEL staining were also shown in Figures 1G–H. ROS staining results showed that Ptpn6 ^+/mev mice produced more ROS than wild-type mice after IRI (Figure 1F). This finding suggests that the increased apoptosis of TECs in mice with SHP-1 deficiency contributes to renal IRI. Together, these results indicate that SHP-1 might be involved in apoptosis regulation and may exert a protective effect against renal IRI.

We extracted RNA from the kidneys of three C57BL/6J-Ptpn6^+/mev mice and three C57BL/6J control mice after renal IRI. Samples from individual mice were sequenced for bioinformatics analysis. The ischemia/reperfusion time was fixed at 34 min/24 h. After removing reads that could be mapped to ribosomes, a total of 152.67 million mRNA sequencing (mRNA-seq) reads were obtained. An average of 23.9 million reads of mRNA-seq data were generated per mRNA sample with a read length of 150 bp (paired-end reads) and an expected insertion size of 200 bp. The alignment rate against the reference genome (see section “Materials and Methods”) ranged from 94.07% (sample WT-3) to 92.00% (sample WT-2), with average mapping rates of 93.30 and 93.72% for the control and Ptpn6^+/mev samples, respectively (Supplementary Table 2).

As presented in the principal component analysis (PCA) plot, after IRI, the kidneys of three Ptpn6 ^+/mev mice and control mice showed differences in overall gene expression (Figure 2A). A total of 336 genes were differentially expressed in the kidney samples (FDR < 0.05). Of the 336 DEGs, 169 were upregulated, and 167 were downregulated in Ptpn6^+/mev mice after IRI. A total of 161 genes showed | FC| > 2 with FDR < 0.05, of which 42 genes with | FC| > 4 (Figure 2B). Of the 161 genes significantly changed after IRI, 99 were upregulated, and 62 were downregulated in Ptpn6^+/mev mice (Figure 2C). These 161 genes were DEGs in the following analysis (Supplementary Table 3).

The FDR and the largest | FC| were used to filter the gene list. Fourteen genes were differentially expressed (FDR < 0.0001) with | FC| > 4 in the kidneys of Ptpn6^+/mev mice (Figure 2D). Different types of cells are present in the kidney, and distinct translational signatures were identified in the nephron, interstitial cell populations, vascular endothelium, and macrophages in mice (21). By analyzing the Gene Expression Omnibus (GEO) dataset (GSE52004), we found that in the 11 highly overexpressed genes in Ptpn6^+/mev mice after IRI, Cyp27b1, Mmp10, Hmgcs2, Azgp1, Cyp2d12, and Slc38a3 were expressed mostly in nephrons, and Ckap2l was expressed mostly in macrophages. The genes Ccna2, Kif18b, and Lqgap3 were expressed in both the endothelium and macrophages. Gpnmb was expressed mostly in interstitial cells (Figure 2E).

The list of DEGs was uploaded to the online PANTHER version 16 classification system, which annotated the genes into three major GO domains: the molecular function, biological process, and cellular component domains (17). Among 161 DEGs, 104 genes were annotated to molecular functions. The PANTHER classification system annotated 49 genes (47.1%) to the catalytic activity category (GO:0003824) and 33 genes (31.7%) to the binding category (GO:0005488) in the molecular function domain. The most enriched terms in these two categories were the hydrolase activity (GO:0016787) and protein binding (GO:0005515) terms, respectively (Supplementary Figure 2A). Moreover, DEGs were also enriched in molecular function categories such as the molecular function regulator (GO:0098772) and transporter activity (GO:0005215) categories. In the biological process domain, a total of 66 genes were annotated to the cellular process category (GO:0009987), and 44 genes were annotated to the metabolic process category (GO:0008152) (Supplementary Figure 2B). The two most represented terms in the cellular process category were the cellular component organization (GO:0016043) and cell cycle (GO:0007049) terms. The organic substance metabolic process (GO:0071704) and primary metabolic process (GO:0044238) terms were the two most enriched terms in the metabolic process category. In the cellular component domain, 45 genes were annotated to the organelle category (GO:0043226), 33.3% of which were related to the cytoskeleton (GO:0005856), and 64 genes were annotated to the cell category (GO:0005623), 58.3% of which were related to the intracellular space (GO:0005622) (Supplementary Figure 2C). The PANTHER classification system (protein classes) sorted 33 genes to the metabolite interconversion enzyme (PC00262), 18 genes to oxidoreductase (PC00176) and 12 genes to cytoskeletal protein (PC00085) (Supplementary Figure 2D). Additional information on all gene annotations and protein classifications is available upon further request.

We also performed enrichment analysis of KEGG pathways in DEGs using the online STRING classification system (18, 19). The most significantly enriched (FDR < 0.005) KEGG pathways for the 62 genes that were significantly down-regulated more than twofold were the peroxisome-proliferator-activated receptor alpha (PPARα) signaling pathway and retinol metabolism (Figures 3A–C). In these two pathways, we identified the nine PPARα signaling pathway-related genes CD36, Fabp1, Acaa1b, Ehhadh, Cyp4a10, Cyp4a14, Cyp4a31, Pck1, and Hmgcs2. Among these genes, CD36 and Fabp1 are related to fat absorption, while Acaa1b, Ehhadh, Cyp4a10, and Cyp4a14 are related to fatty acid degradation. The expression of genes encoding Ptpn6-interacting proteins in kidney samples from Ptpn6^+/mev mice after IRI is shown in Figure 3D. Then we found that the expression of Ppara was also significantly decreased in Ptpn6^+/mev mice than in wild-type mice after IRI (FC −1.63, FDR = 0.003). We validated the expression of the nine genes involved in the PPARα signaling pathway in the Ptpn6^+/mev mice and wild-type mice after IRI by RT-PCR. We also compared the expression of these genes in the control kidneys without IRI in Ptpn6 ^+/mev and wild-type mice and found that there were no significant changes (| FC| < 1.5 with FDR > 0.05) (Supplementary Table 4). Moreover, for the 99 significantly upregulated genes in Ptpn6^+/mev mice after IRI, the most significantly enriched KEGG pathways were the mitotic cell cycle process and cell division pathways. Previous studies have revealed that cell cycle function-associated genes are generally upregulated at the peak of injury (24 h after IRI) (22). In this study, key kinases involved in the proper spindle and kinetochore assemblies, such as Aurkb, Plk1, Cdk1, Tpx2, and Cdca3, were upregulated in the Ptpn6^+/mev group compared with the wild-type^+/+ group at 24 h after IRI.

Src homology region 2 domain-containing phosphatase-1 has two SH2 binding domains and one PTP catalytic domain. SHP-1 inhibits signal transduction from receptors in many cell types¹ through interaction with the immunoreceptor tyrosine-based inhibitory motif (ITIM) (23–25). After IRI, we found five upregulated and nine downregulated genes in Ptpn6^+/mev mice that encoded proteins with ITIMs (FDR < 0.05), of which three genes were upregulated and five genes were downregulated by more than twofold (Supplementary Table 5). Of these eight genes, six genes were related to lipid metabolism. In the kidneys, proximal TECs use fatty acids as their main energy source due to the high energy demand. A previous study indicated that overexpression of ATF6a transcriptionally downregulated PPARα, leading to reduced fatty acid β-oxidation, enhanced apoptosis and reduced cell viability in a human proximal TEC line (HK-2) (26). Since apoptosis in TECs was also enhanced in our study, we suggest that the PPARα signaling pathway plays an important role in the development of renal IRI in SHP-1-knockdown mice.

To confirm that PPARα signaling plays a central role in the aggravation of renal IRI induced by SHP-1 insufficiency, we activated PPARα signaling by administrating the agonist fenofibrate or vehicle control (corn oil) to six Ptpn6^+/mev mice for 3 weeks to activate the PPARα signaling pathway, followed by IRI (Figure 4A). When the mice were killed 24 h after IRI, the serum CR and BUN of three mice taking fenofibrate were significantly lower than those of three mice taking corn oil (Figures 4B,C). Furthermore, as assessed using H and E staining and ROS staining analysis, the kidneys of mice in the fenofibrate-treated group showed attenuation of renal injury and reduction of ROS levels (Figures 4D–G). We extracted RNA from the kidney after IRI for RT-PCR to confirm that the expression of Acaa1b, Ehhadh, Cyp4a10, and Cyp4a14, which are related to fatty acid degradation, was significantly increased in mice treated with fenofibrate but not corn oil (Figure 4H). Taken together, these in vivo results suggested that SHP-1 knockdown may lead to the downregulation of fatty acid oxidation, along with an increase in ROS production and renal tubular damage. This phenotypic combination ultimately contributes to extracellular matrix production and promotes mesangial matrix or basement membrane thickening after IRI in Ptpn6^+/mev mice. A schematic of SHP-1 insufficiency leading to the development of renal IRI via suppression of PPARα signaling is shown in Figure 5.