The animal experiments were performed in the Experimental Animal Platform of Renji Hospital, and all animal research was approved by the Institutional Animal Ethics Committee of Renji Hospital. Briefly, 8-week-old male C57BL/6 mice and BKS.Cg-Dock7m +/+ Leprdb/J(db/db) mice were both purchased from SLAC Laboratory Animal Co., Ltd. (Shanghai, China). Mice were housed under controlled ambient conditions with standard chow, a 12-h light/dark cycle (lights on at 7 a.m.), and were maintained at 25°C. For experiments, mice were divided into three groups: control group (Ctrl, C57BL/6 mice, n=5), diabetic group (Case, db/db mice, n=6), and treatment group (Dapa, db/db mice treated with dapagliflozin [1.0 mg/kg/day, AstraZeneca, Cambridge, UK], n=6) for 6 weeks. Dapagliflozin was administrated in the drinking water by oral gavage once daily in the treatment group, and mice in the Ctrl and Case groups were administrated with the same volume of saline. When the experiment was completed, all mice in the three groups were sacrificed by cervical dislocation after CO₂-induced narcosis. Immediately afterward, the kidneys and blood were collected.

For ischemia/reperfusion-induced chronic kidney disease mouse models, 14-week-old male C57BL/6 mice were anesthetized and subjected to right kidney nephrectomy. Right kidneys were collected and used as controls, and the left renal artery was clamped for 30 minutes at 37°C followed by reperfusion. At 7 days post surgery, kidney tissues and blood and urine samples were collected for further analysis.

Total RNA was isolated using an RNeasy mini kit (Qiagen, Germany). Strand-specific libraries were prepared using the TruSeq Stranded Total RNA Sample Preparation kit (Illumina, USA), following the manufacturer’s instructions. Briefly, mRNA was enriched with oligo(dT) beads. Following purification, the mRNA is fragmented into small pieces using divalent cations under 94°C for 8 min. The cleaved RNA fragments are copied into first-strand cDNA using reverse transcriptase and random primers. This is followed by second-strand cDNA synthesis using DNA Polymerase I and RNase H. These cDNA fragments then go through an end repair process, adding a single’A’base, and then ligating the adapters. The products are then purified and enriched with PCR to create the final cDNA library. Purified libraries were quantified by Qubit 2.0 Fluorometer (Life Technologies, USA) and validated by Agilent 2100 bioanalyzer (Agilent Technologies, USA) to confirm the insert size and calculate the mole concentration. The clusters were generated by cBot with the library diluted to 10 pM and then were sequenced on the Illumina NovaSeq 6000 (Illumina, USA). The library construction and sequencing were performed at Shanghai Biotechnology Corporation.

For each sample, 33-95M RNA-sequencing (RNA-seq) clean reads were obtained that mapped to the mouse reference genome (mm10) using hierarchical indexing for spliced alignment of transcripts (HISAT2) v2.0.477. Sequencing read counts were calculated using Stringtie78,79 (v.1.3.0). Then, expression levels from different samples were normalized by the Trimmed Mean of M values (TMM) method. The normalized expression levels of different samples were converted to Fragments Per Kilobase of transcript per Million mapped fragments (FPKM). The edgeR package of R was used to analyze the difference between intergroup gene expression, the P-values were calculated, and multiple hypothesis tests were performed. The P-value threshold was determined by controlling the False Discovery Rate (FDR) with the Benjamini algorithm. The corrected P-value is called the q-value. Differentially expressed genes (DEGs) were defined as transcripts with a fold change in expression level (according to the FPKM value) greater than 2.0 and a q-value less than 0.05. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed with the clusterProfiler package of R, and the enrichment criteria included a q-value < 0.05. Heatmaps of specific genes were generated using the heatmap package of R. K-means analysis was performed with Euclidean distances in Multiple Experiment Viewer 4.9, and yielded 10 clusters based on input FPKM values (11).

Briefly, a total of 20,000 cells per sample was washed twice with 50 µL of cold Dulbecco’s Phosphate-Buffered Saline and re-suspended in 50 µL of lysis buffer [10 mM tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, and 0.1% (v/v) NP-40 substitute (Sigma-Aldrich, 11332473001)]. The suspension of nuclei was then centrifuged for 10 min at 500g at 4°C, followed by the addition of 50 µL of transposition reaction mix (10 µL of 5×TTBL buffer, 4 µL of TTE mix, and 36 µL of nuclease-free H2O) of TruePrep DNA Library Prep Kit V2 for Illumina (Vazyme Biotech, TD501). Samples were then incubated at 37°C for 30 min. DNA was isolated using the QIAquick PCR Purification Kit (QIAGEN, 28106). ATAC-seq libraries were first subjected to five cycles of pre-amplification using NEBNext High-Fidelity 2X PCR Master Mix (New England Biolabs, M0541S). To determine the suitable number of cycles required for the second round of PCR, the library was assessed by qPCR using NEBNext High-Fidelity 2X PCR Master Mix with SYBR Green I Nucleic Acid Gel Stain (Thermo Fisher Scientific, S7563) and then PCR amplified for the appropriate number of cycles. Libraries were purified with the QIAquick PCR Purification Kit. Library quality was checked using the High Sensitivity DNA Analysis Kit (Agilent, 5067-4626). Last, 2 X 150 paired-end sequencing was performed on an Illumina Nova seq (Illumina, San Diego, CA). The library construction and sequencing were performed at Shanghai Biotechnology Corporation.

Alignment and peak-calling of ATAC sequencing data were performed using the ENCODE-DCC ATAC-Seq pipeline (https://github.com/ENCODE-DCC/atacseq-pipeline) (ENCODE Project Consortium, 2012). Briefly, data were aligned to the mouse reference genome (mm10) using Bowtie2. Aligned data were filtered for PCR duplicates, blacklisted regions, and shifted +4bp for the + strand and -5bp for the - strand. Peaks were called using MACS2. Differential peak calling was performed using Diffbind (DiffBind,http://bioconductor.org/packages/release/bioc/vignettes/DiffBind/inst/doc/DiffBind.pdf.). Motif enrichment was performed using Hypergeometric Optimization of the Motif EnRichment program (HOMER) version 4.1.

Pathway enrichment analysis was performed on the DEGs using the Gene Ontology (www.geneontology.org) and the Kyoto Encyclopedia of Genes and Genomes (https://www.kegg.jp/) databases.

Mouse urine albumin concentration and urine creatinine concentration were measured using the Albumin Creatinine Ratio Assay Kit (ab241018) from Abcam. The excretion of albumin is expressed as the urinary albumin creatinine ratio (UACR). Urine albumin concentration was confirmed by loading 1 μl urine sample per lane on a sodium dodecyl sulfate-polyacrylamide gel for electrophoresis, followed by Coomassie blue staining. 1 and 5 μg bovine serum albumin (BSA) were loaded as controls for albuminuria. BUN and plasma creatinine levels were measured using a Liquid Urea Nitrogen Reagent Set and Creatinine Assay kit (Nanjing Jiancheng Bioengineering Institute, China) according to the manufacturer’s protocol.

Renal tissue harvested from animals was washed with 0.9% saline, fixed in 10% neutral buffered formalin, and then embedded in 10% paraffin. Sections (4 µm thick) were stained with periodic acid-Schiff staining (PAS) for further microscopic analysis. Three consecutive glomerular cross-sections from each sample were photographed by a nephropathologist, blinded to the source of the tissue, using a digital camera (Olympus DP74; Olympus America, Melville, New York), and imported into Image-Pro Plus (Media Cybernetics, Silver Spring, Maryland).

Immortalized mouse macrophage cell line (RAW264.7) was cultured in Dulbecco’s modified Eagle’s medium (DMEM, HyClone) supplemented with 10% heat-inactivated fetal bovine serum (Gibco), 100 U/mL penicillin, and 100 μg/mL streptomycin (Sigma) at 37°C under 5% CO₂. Immortalized human renal proximal tubular cell line (HK-2) was cultured in DMEM/F12 medium (HyClone) supplemented with 10% heat-inactivated fetal bovine serum (Gibco), 100 U/mL penicillin, and 100 μg/mL streptomycin (Sigma) at 37°C under 5% CO₂.

Before experiments, cells were seeded onto 6-well plates to reach a confluence of 70% and then starved in serum-free DMEM overnight. Cells were then divided into three groups: for the osmotic control group (Ctrl), cells were cultured in DMEM medium containing 5mM glucose and 30mM mannitol; for the high-glucose group (Case), cells were cultured in DMEM medium containing 35mM glucose; for the dapagliflozin treatment group (Dapa), cells were cultured in DMEM medium containing 35mM glucose and 20µM dapagliflozin. Macrophages were incubated for 48 hours, and renal proximal tubular cells were incubated for 72 hours.

Proteins from mice kidneys or cultured cells were extracted using RIPA lysis buffer supplemented with a cocktail of protease and phosphatase inhibitors. The total protein concentration was measured using a BCA assay, and equal amounts of protein were loaded on PAGE 4-20% gradient gels (GenScript) and transferred to PVDF membranes (Millipore). The primary antibodies were mouse anti-CSF2RB (Santa Cruz Biotechnology), rabbit anti-BTLA (Invitrogen), mouse anti-ISG15 (Santa Cruz Biotechnology), rabbit anti-SGLT2 (Proteintech), and mouse anti-β-actin (Cell Signaling Technology). Blots were developed using Omni ECL reagent (EpiZyme).

Total RNA from cultured cells was extracted using TRIzol Reagent (Invitrogen). RNA quality was checked by controlling the OD at 260 and 280 nm. cDNA was synthesized from 1 µg RNA using the Hifair II 1st Strand cDNA Synthesis Super Mix (Yeasen, China). Real-time PCR was performed with the Roche Light Cycler 480 detection system using SYBR green PCR master Mix (Roche Diagnostics). Specific primers for target mRNAs (listed in the Supplementary Materials) were used for amplification, and a standard curve was generated for each targeted transcript. All relative expression levels were normalized to β-actin expression, and the results were analyzed using the ΔΔCt method.

Mouse whole kidney lysate and cell culture medium supernatant of RAW264.7 were scanned using the QAM-INF-1 array (Raybiotech), and cell culture medium supernatant of HK-2 was scanned using the Quantibody Human TH17 array (Raybiotech) to obtain the raw data. Then, the Raybiotech software was used to remove the chip background and normalize the original data. Differentially expressed proteins (DEPs) between different groups were identified using the limma package (Version 3.42.2) in R software (RStudio, PBC, Boston, MA, USA). Individual p-values were calculated and converted to adjusted p-values (adj. p. val) for comparisons by false discovery rate correction of the Benjamini and Hochberg test. A cutoff point of adj. p. val < 0.05 and |fold change (FC)| > 1.2 were used to select DEPs. Then, the heatmap, and scatter plot of the DEPs were drawn using the ggplot2 package in R software. The expression levels of a specific protein of interest were determined by ELISA. The raw data were detailed in the Supplementary Material.

The DEPs were uploaded to an online bioinformatics database, the Database for Annotation, Visualization, and Integrated Discovery (DAVID) version 6.8 Beta (https://david-d.ncifcrf.gov/), for enrichment analysis, including gene ontology (GO) pathways. The obtained results were visualized in the R ggplot2 package and a p-value of <0.05 was considered statistically significant.

Data analysis was performed using GraphPad Prism software. Differences between two groups were analyzed using unpaired Student’s two-tailed t-test, and differences between more than two groups were analyzed using a one-way ANOVA test. Each experiment was replicated for more than three times. P values are presented as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, not significant.

Transgenic db/db mice (Case) of 14-week-old showed spontaneous type 2 diabetes mellitus phenotype: compared with the healthy mice (Ctrl), diabetic mice had augmented body weight (22.9 ± 2.9 g in Ctrl versus 47.8 ± 9.2 g in Case, P<0.0001) and hyperglycemia (3.9 ± 0.7 mM in Ctrl versus 22.7 ± 4.9 mM in Case, P<0.0001) (Figures 1A, B). In contrast, db/db mice given oral administration of 1 mg/kg dapagliflozin for 6 weeks (Dapa) presented significantly lower blood glucose levels (11.8 ± 2.0 mM in Dapa, P=0.0006 compared with Case), though their bodyweight displayed no significant changes (55.3 ± 1.7 g in Dapa, P=0.1854 compared with Case) (Figures 1A, B).

As for kidney functions, db/db mice presented higher serum creatinine levels (51.1 ± 2.8 µM in Ctrl versus 63.7 ± 6.0 µM in Case, P<0.0001) and elevated urinary albumin creatinine ratio (UACR, 0.0011 ± 0.0003 in Ctrl versus 0.0037 ± 0.0006 in Case, P<0.0001) (Figures 1C–E). Treatment with dapagliflozin remarkably ameliorated azotemia (serum creatinine=53.4 ± 5.0 µM in Dapa, P=0.0215 compared with Case) and albuminuria (UACR=0.0024 ± 0.0004 in Dapa, P=0.0059 compared with Case) in diabetic mice (Figures 1C–E). Periodic acid-Schiff staining of renal tissue from the Case group exhibited mild-to-moderate diabetic renal structural alterations, including mesangial area broadening, mesangial matrix increase, and mesangial hyperplasia, while these changes were alleviated in mice receiving dapagliflozin treatment (Figure 1F). In short, dapagliflozin successfully rescued both hyperglycemia and diabetic nephropathy in diabetic mice.

To determine potential candidate genes that might participate in the pathogenesis of DKD, we performed a combined analysis of RNA sequencing and Assay for Transposase Accessible Chromatin with high-throughput sequencing (ATAC-seq) on mice whole kidney lysates. Mice displayed distinguished RNA expression patterns between the Ctrl group (Con.1-5) and the Case group (Db.1-6)(Figure 2A). Functional enrichment analysis utilizing the gene ontology (GO) database and Kyoto Encyclopedia of Gene and Genome (KEGG) database suggested that, in addition to metabolic disorders, diabetic mice showed immune system alterations that might involve response to interferon (Figures 2B–D). To be precise, analysis based on the KEGG database revealed pathways involving multiple material metabolism, inflammation and autoimmune-related diseases, and antigen processing and presentation (Figure 2C); while GO terms indicated material metabolism, antigen processing and presentation via MHC complex, and response to interferon (Figure 2D). Moreover, despite individual discrepancies, mice treated with dapagliflozin (Db.7-12) also presented different RNA expression patterns to the Case group (Figure 2E). Similarly, the top DEGs of these two groups were enriched in the immune-associated pathways, especially the response to interferon (Figures 2F–H). In the analysis of Dapa versus Case, the KEGG database unraveled pathways involving inflammation and autoimmune-related diseases, helper T cell differentiation, and antigen processing and presentation (Figure 2G); GO terms also suggested antigen processing and presentation, and response to interferon (Figure 2H).

Furthermore, we performed a K-means analysis with the aim of revealing the co-ordinance between Ctrl and Dapa groups. According to their different expression patterns among the three groups, the DEGs were divided into 10 clusters. We noticed cluster 2 and cluster 3, in which DEGs showed upregulation in the Case group but downregulated by dapagliflozin treatment (Figure 3). GO terms and the KEGG database indicated that pathways enriched in these two clusters were immune response, response to interferon-gamma (IFN-γ), antigen processing and presentation, and inflammation-related diseases (Figure 3). In addition, DEGs in cluster 5 and cluster 9 also presented co-ordinance between Ctrl and Dapa groups: pathways enriched in cluster 5 mainly involved material metabolism, endocrine regulation, and renal intrinsic cells; while pathways enriched in cluster 9 were associated with immunoglobulin, PPAR, and FoxO pathways (Supplemental Figure 1).

Combining the results of RNA sequencing and ATAC analysis, we identified three inflammation-associated genes that showed opposite regulation trends in Case versus Ctrl and Dapa versus Case, i.e., Isg15, Btla, and Csf2rb (Figure 4). Isg15 (Figures 4A–C) and Csf2rb (Figures 4A, F–G) presented upregulated chromatin accessibility and mRNA levels in diabetic mice and attenuated by dapagliflozin administration. In contrast, Btla (Figures 4A, D, E) displayed downregulated chromatin accessibility and mRNA levels in the Case group and restored in the Dapa group. These genes were further analyzed due to their significant correlation with the clinical features of DKD (Figure 5).

We first noticed Isg15, a pro-inflammatory gene, as functional enrichment analysis of RNA sequencing highlighted interferon-related pathways (Figure 2). In the current study, the univariate analysis including all mice from three groups demonstrated that Isg15 expression levels were positively correlated to hyperglycemia and UACR (Figures 5A, B), suggesting its deleterious role in DKD. We also found positive correlations between another pro-inflammatory gene, Csf2rb mRNA levels in the kidney and the blood glucose concentration, serum urea nitrogen levels, and UACR of diabetic mice (Figures 5C–E), implying that CSF2RB should be detrimental in DKD. On the contrary, the anti-inflammatory gene Btla presented a negative correlation to serum creatinine concentration and UACR (Figures 5F–H). Therefore, we concluded that DKD is associated with renal inflammation, as pro-inflammatory cytokines would exacerbate while anti-inflammatory might attenuate kidney injury.

Next, we tested these inflammation-related molecules at the protein level in the mice kidneys. Diabetic mice presented a significantly higher level of ISG15 than the normal controls, which was alleviated by the dapagliflozin treatment (Figures 6A, B). In contrast, reduced expression of the immunosuppressor BTLA was detected in the Case group that was restored by dapagliflozin administration (Figures 6A, B). Although CSF2RB was nearly undetectable at the protein level in the whole kidney lysates, these inflammatory proteins were consistent with their mRNA levels and the univariate analysis.

To further explore the effect of dapagliflozin on renal inflammation in vivo, we performed an inflammation-associated antibody array assay analysis of mice whole kidney lysates, and found significant differences in the protein components of mice kidneys from different groups. In comparison with the Ctrl group, diabetic mice presented both upregulated and downregulated proteins (Figure 6C). Gene ontology analysis revealed that the Case group displayed an apparent profile of inflammatory cell proliferation and activation, as well as JAK-STAT pathway activation (Figure 6D). In contrast, mice treated with dapagliflozin showed overall downregulated inflammatory proteins in the kidney lysates (Figure 6E). Dapagliflozin administration triggered significant downregulation of two well-characterized inflammatory markers, tumor necrosis factor-alpha (TNF-α) and interleukin-1beta (IL-1β) (Figure 6F). Furthermore, dapagliflozin reduced the expression of multiple pro-inflammatory molecules in the diabetic kidney: CD30L, also known as CD153, can bind CD30 to promote senescence-associated tertiary lymphoid tissue expansion, facilitating local T and B cell interactions in the chronically inflamed kidney (12); IL-3 can act as the ligand of the aforementioned receptor CSF2RB (13); IL-15 is a pluripotent cytokine that supports both myeloid immune cells and lymphocytes, regulating innate and adaptive immune system (14); B lymphocyte chemoattractant, also called CXCL13, plays a role in diabetic pain and production of pro-inflammatory cytokines IL-1β, TNF-α, and IL-6 (15). In summary, dapagliflozin administration attenuated the overall inflammatory profile of the diabetic kidney in mice.

With the aim of dissection of the renoprotective role of dapagliflozin at the cell level, we performed a cell culture of the human proximal tubular cell line (HK-2) with high expression of the target receptor SGLT2 (16). We found augmented mRNA levels of ISG15 in tubular cells cultured in the high-glucose medium (Case group) versus cells cultured in normal-glucose medium (Ctrl group), which was alleviated by dapagliflozin treatment under the same high glucose concentration conditions (Dapa group) (Figure 7A). However, there was no expression of CSF2RB in proximal tubular cells, and the level of BTLA was kept almost unchanged regardless of the glucose or dapagliflozin treatment, indicating that dapagliflozin did not directly impact BTLA expression in renal tubular cells (Figure 7A). Correspondingly, immunoblot presented elevated ISG15 expression in the Case group that was attenuated in the Dapa group (Figures 7B, C). Subsequently, we performed an inflammation-associated antibody array assay analysis of the cell culture medium to further explore dapagliflozin’s effect on tubular cells under hyperglycemia. Renal tubular cells tended to produce more inflammatory molecules in the high-glucose medium (Figure 7D), activating the STAT pathway and promoting inflammation (Figure 7E). Incubation with dapagliflozin attenuated HK-2 secretion of the inflammatory proteins (Figure 7F), especially the well-known marker TNF-α and the fibrosis marker transforming growth factor-beta (TGF-β) into the cell culture medium (Figure 7G), implying an anti-inflammatory, anti-fibrotic role of dapagliflozin in the diabetic kidney. Moreover, we detected amelioration of two other well-characterized inflammation-associated molecules by dapagliflozin treatment, i.e., TNF-β and IFN-γ (Figure 7G). Of note, the aforementioned effects of dapagliflozin on HK-2 cells were independent of glucose concentrations, as the Case group and the Dapa group were cultured under the same glucose concentration conditions. We concluded that dapagliflozin served as an anti-inflammatory, anti-fibrotic agent to protect renal proximal tubular cells against DKD independent of glucose concentrations.

As macrophages are considered the major leukocyte in the chronically inflamed kidney in diabetes (5), and our data highlighted the role of antigen-presenting cells and inflammation in DKD, we further investigated the role of dapagliflozin in macrophages. The immortalized mouse macrophage cell line (RAW 264.7) was cultured in vitro, and we detected upregulation of ISG15, BTLA, and CSF2RB in the high-glucose medium at both mRNA and protein levels (Figures 8A, C). Surprisingly, with our immunoblot analysis showing SGLT2 expression on macrophages (Supplemental Figure 2D), dapagliflozin treatment significantly alleviated the upregulation of these inflammatory molecules (Figures 8A–C). As macrophages might secret various cytokines to interact with multiple cell types, we also detected a series of molecules in the cell culture medium supernatant utilizing an inflammation-associated antibody array assay. Results of DEPs were enriched in the endothelial cell apoptosis process and leukocyte chemotaxis, suggesting that macrophages in hyperglycemia might induce endothelial injury and recruit other leukocytes to exacerbate inflammatory responses (Figures 8D, E). However, dapagliflozin treatment seemed to have little impact on the cytokines in the macrophage cell culture medium supernatant according to the antibody array assay. To deepen our investigations, we tested the role of glucose levels and dapagliflozin in macrophage polarization. High-glucose incubation activated macrophages to express a higher level of Tnf-α, Il-1β, iNos, Il-10, and Tgf-β, implying complicated macrophage activation and polarization process with a combination of different phenotypes (Figure 8F) (17). Dapagliflozin treatment attenuated the activation of these macrophages, exerting a seemly “calm down” effect (Figure 8F). Altogether, these results revealed a novel, unrecognized role of dapagliflozin in regulating inflammatory response via its effect on macrophages and independent of glucose concentrations, and its specific signaling pathways deserve further exploration.

Considering that our cell culture experiments suggested a blood glucose-independent, anti-inflammatory effect of dapagliflozin, we subsequently explored the potential role of dapagliflozin in healthy mice and mice with CKD induced by ischemia/reperfusion (I/R). Dapagliflozin administration in healthy mice caused no apparent changes in serum creatinine (48.05 ± 4.7 µM in Ctrl versus 45.88 ± 4.9 µM in Ctrl+Dapa, P=0.9294), blood urea nitrogen (8.45 ± 0.4 mM in Ctrl versus 7.83 ± 0.8 mM in Ctrl+Dapa, P=0.6332), and blood glucose levels (6.43 ± 0.5 mM in Ctrl versus 7.40 ± 0.8 mM in Ctrl+Dapa, P=0.2799) (Figures 9A–C). At 7 days post I/R surgery, serum creatinine (74.35 ± 3.8 µM in I/R, P<0.0001) and blood urea nitrogen levels (15.68 ± 0.6 mM in I/R, P<0.0001) elevated while blood glucose (7.05 ± 0.7 mM in I/R, P=0.6293) remained unaltered in CKD mice compared with Ctrl mice. In contrast, treatment with dapagliflozin significantly restored kidney function (serum creatinine=55.15 ± 4.2 µM in I/R+Dapa, P=0.0009 compared with I/R; blood urea nitrogen=12.55 ± 0.7 mM in I/R+Dapa, P=0.0003 compared with I/R) and reduced blood glucose levels (4.63 ± 0.4 mM in I/R+Dapa, P=0.0024 compared with I/R) in I/R-induced CKD mice (Figures 9A–C). Then, we detected the protein levels of BTLA and ISG15 in the mice kidneys. Compared with the Ctrl group, healthy mice treated with dapagliflozin displayed augmented BTLA expression, and their ISG15 expression seemed to decrease but did not reach a significant threshold, probably due to the low ISG15 protein levels at the baseline (Figures 9D, E). Furthermore, I/R surgery downregulated BTLA while upregulated ISG15 in the kidney, and such a pro-inflammatory state was ameliorated by dapagliflozin (Figures 9D, E). As a result, dapagliflozin played a renoprotective, anti-inflammatory role in I/R-induced CKD mice.