GRP78 Contributes to the Beneficial Effects of SGLT2 Inhibitor on Proximal Tubular Cells in DKD Free

Pathophysiological interactions among metabolic disorders, cardiovascular diseases, and chronic kidney disease have recently become evident in clinical and basic research (1,2). Therefore, there is a need for treatment strategies on diabetes that have additional benefits in preventing complications based on systemic organ cross talk besides controlling blood glucose levels. Although sodium–glucose cotransporter 2 (SGLT2) inhibitors were initially developed and marketed as antidiabetes agents, they have been established as a multibenefit drug for cardiorenal metabolic diseases beyond glucose-lowering effects. Large placebo-controlled trials demonstrated that several SGLT2 inhibitors reduced the risk of cardiovascular disease and hospitalization of heart failure in patients with type 2 diabetes (3). They also showed a protective effect on nondiabetic chronic kidney disease, especially IgA nephropathy, in addition to diabetic kidney disease (DKD) (3). This accumulating evidence led us to expect the existence of an unknown mechanism of SGLT2 inhibitors.

We identified an adipokine, named visceral adipose tissue-derived serine protease inhibitor (vaspin), that is upregulated in animal models with obesity and type 2 diabetes. As a compensation mechanism, it has protective roles on high-fat high-sucrose–induced obesity, insulin resistance, fatty liver, and atherosclerosis (4,5). In addition, vaspin is internalized to renal proximal tubular cells and inhibits endoplasmic reticulum (ER) stress response failure, autophagy impairment, and lysosomal membrane permeabilization with release of cathepsin B into cytoplasm, and subsequent NLRP3 (NOD-, LRR- and pyrin domain-containing protein 3) inflammasome activation (6). We revealed that these multifaced beneficial roles of vaspin are mediated by forming a complex with 78-kDa glucose-regulated protein (GRP78) (4–6).

GRP78 was originally identified as a protein highly induced when glucose was removed from the growth medium of normal cells. In turn, GRP78 protein is markedly suppressed when glucose is maintained at high levels in the growth medium (7). In healthy subjects, filtrated glucose from the glomerulus is almost completely reabsorbed by proximal tubular cells via SGLT2 and SGLT1. In diabetes, the expression of SGLT2 is increased and glucose reabsorption is enhanced in proximal tubular cells. These findings indicate that GRP78 and SGLT2 both reveal the common features that their expression is regulated by a high- or low-glucose milieu. In the current study, we found that the SGLT2 inhibitor, canagliflozin (CANA), possesses a GRP78-mediated protecting mechanism toward proximal tubular cells beyond glucose-lowering effect.

C57BL/6J male mice were obtained from CREA Japan and housed under a 12-h light-dark cycle with free access to water. Eight-week-old mice received five consecutive intravenous injections of 50 mg/kg streptozotocin (STZ) (Sigma-Aldrich, St. Louis, MO) in citrate buffer at pH 4.6 or citrate buffer only as a control (CTRL). The onset of diabetes was confirmed when blood glucose levels were >300 mg/dL at two different points after the third day of administration of STZ. CANA, provided by Mitsubishi Tanabe Pharma Corporation, was diluted in 0.5% hydroxypropyl methylcellulose (H7509, Sigma-Aldrich) and administrated to the CTRL and STZ groups at 30 mg/kg by oral gavage. This dose was used from previous studies that showed it is pharmacologically effective in mice (8,9). Vehicle-administrated mice were used as CTRL to CANA-treated mice. Four groups were generated: CTRL group (n = 8), STZ group (n = 13), CTRL+CANA group (n = 8), and STZ+CANA group (n = 12). Vehicle or CANA was administered daily for 2 weeks, body weight and food intake were measured, urine was collected, and blood and kidney samples were harvested under isoflurane anesthesia, and then mice were euthanized. All animal experiments were approved by the Animal Care and Use Committee of the Department of Animal Resources, Advanced Science Research Center, Okayama University (OKU-2019569, OKU-2019570).

HK2 cells were cultured in DMEM with low glucose (5.5 mmol/L glucose) or DMEM with high glucose (25 mmol/L glucose) that was supplemented with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin at 37°C in 5% CO₂. DMEM (042-32255, 041-29775, FUJIFILM Wako Chemicals) supplemented with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin was used for glucose-free culture experiments. HEK293T cells were cultured in DMEM with high glucose (25 mmol/L glucose) with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin.

For the knockdown of GRP78 in HK2 cells, MISSION shRNA lentivirus transduction particles for GRP78 (NM_005347) (shRNA-GRP78) or nontarget shRNA control lentivirus transduction particles (shRNA-CON) were used for transfection, as described previously (6). Lipofectamine LTX Reagent (Thermo Fisher Scientific) was used for transient transfection in HK2, and Polyethylenimine Max (Polysciences) in HEK293T cells.

Fluo-4 (Molecular Probes, Eugene, OR) was used to measure intracellular Ca²⁺ concentration by FlexStation (Danaher Corp., Washington, DC). CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega) was used for the proliferation assay of HK2 cells. Scratch assay was performed as previously described (10). In brief, a monolayer of confluent HK2 cells grown in a 6-well plate was scratched with a sterile 1,000-μm pipette tip. Cells were washed twice with PBS, cultured with medium containing 0.5% FBS and CANA for 24 h, and were fixed with methanol and stained with trypan blue solution (Sigma-Aldrich). We performed three independent replicates of in vitro experiments.

Tunicamycin (T7765) and thapsigargin (T9033) were purchased from Sigma-Aldrich, and the stock solution was prepared in DMSO for the induction of ER stress. Recombinant human (rh)GRP78 protein (ab78432) was purchased from abcam. CANA, provided by Mitsubishi Tanabe Pharma Corporation, was dissolved in ethanol to make 100 mmol/L stock solutions and diluted with culture medium.

The measurement of glucose concentration of culture medium or cell lysate was performed using Glucose Assay Kit-WST (Dojindo, Kumamoto, Japan).

HK2 cells were transiently transfected with Path Detect pDR5-Luc (DR5/RARE) cis-Reporter Plasmid (no. 240120, Stratagene) and pGL4.74[hRluc/TK] Vector plasmid (E6921, Promega) using Lipofectamine LTX Reagent (Invitrogen), following the manufacturer’s protocol. Luciferase activities were measured with a dual-luciferase assay system and GloMax 20/20n Luminometer (Promega).

Cell samples were homogenized with lysis buffer (20 mmol/L Tris-HCl, pH 7.4; 100 mmol/L NaCl, 10 mmol/L benzamidine-HC, 10 mmol/L ε-amino-n-caproic acid, 2 mmol/L phenylmethylsulfonyl fluoride, and 1% Triton X-100), and centrifuged at 14,000 rpm for 30 min at 4°C. Supernatants were analyzed further. Kidney cortex tissues were excised and homogenized with lysis buffer. After centrifugation, the supernatants were collected for further analyses. For fractionation of kidney tissues, discontinuous iodixanol gradients method using OptiPrep (Invitrogen) was used with protocol S36 downloaded on 5 May 2008.

An equal amount of protein was subjected to SDS-PAGE under reducing conditions and electroblotted onto Hybond P polyvinylidene fluoride membranes (GE Healthcare Life Sciences). The membranes were immersed in blocking solution containing 5% nonfat dry milk and Tris-buffered saline (TBS) with Tween-20 (0.05% Tween-20, 20 mmol/L Tris-HCl, and 150 mmol/L NaCl, pH 7.6) or in 2% BSA and TBS with Tween-20. Then, the membranes were incubated with primary antibodies. Details of the primary antibodies are shown in Supplementary Table 1. They were then incubated with anti-rabbit or anti-mouse IgG conjugated with horseradish peroxidase (HRP), NA934 (AB_772206) and NA931 (RRID:AB_772210), 1:10,000 (GE Healthcare Life Sciences) or anti-rat IgG-conjugated HRP, 1:3,000 (no. 7077; RRID:AB_10694715; Cell Signaling Technology), or rabbit TrueBlot: anti-Rabbit IgG HRP, 1:1,000 (AB_2610848; Rockland). The blots were washed three times with TBS with Tween-20, immersed in ECL Plus Western Blotting Detection Reagents (GE Healthcare Life Sciences), and the chemiluminescence was analyzed using the LAS-4000 mini instrument (FUJIFILM).

Protein lysates were precleared for 1 h at 4°C using Sepharose 4B (Sigma-Aldrich) to remove nonspecific binding proteins. Immunoprecipitation of culture cells and kidney proteins was then performed using Catch and Release v2.0 (Millipore). Immunoprecipitation using FLAG tag fused protein was performed using ANTI-FLAG M2 Affinity Gel (Sigma-Aldrich).

Coding sequence without signal peptide of GRP78^19-654 was amplified by using PCR primers (5′-GGGGGGAAGCTTTGAGGAGGAGGACAAGAAG-3′ and 5′-GGGGGGGCGGCCGCCTACAACTCATCTTTTTC-3′) and subcloned into TOPO TA Cloning vector (Invitrogen). Then, the plasmid was digested with HindIII/NotI and ligated to p3XFLAG-CMV-10 (E 4401, Sigma-Aldrich) (p3xFLAG-GRP78).

HK2 cells were cultured on coverslips and fixed in 4% paraformaldehyde for 15 min. Cells were not permeabilized for GRP78 staining, but for staining of p65 nuclear factor (NF)-κB, were permeabilized with 0.1% Triton X-100 in PBS for 15 min, and followed by blocking with serum-free protein block (Dako) for 30 min. Mouse kidney samples were also blocked with serum-free protein block (Dako) for 30 min. Then these were incubated overnight with primary antibodies; GRP78/BIP monoclonal antibody, 1:200 (66574-1-Ig, RRID: AB_2881934, Proteintech); calnexin, 1:200 (C4731, RRID: AB_476845, Sigma-Aldrich); and NF-κBp65 (D14E12) XP rabbit monoclonal antibody, 1:500 (no. 8242, RRID: AB_10859369, Cell Signaling Technologies). The information on knockout and knockdown validation antibodies is provided in Supplementary Table 2. Cells or tissues were washed and incubated with secondary antibodies Alexa Fluor 488, 1:500 (A-11070, RRID: AB_142134, Invitrogen), or anti-rabbit IgG (H+L), 1:500, F(ab')2 Fragment (Alexa Fluor 594 Conjugate, no. 8889, RRID: AB_2716249, Cell Signaling Technologies). For double staining of SGLT2 and GRP78 on mouse kidney tissues, SGLT2 rabbit polyclonal antibody, 1:200 (24654-1-AP, RRID: AB_2750601, Proteintech) was used for the primary antibody with overnight incubation, and then Alexa Fluor 488, 1:500 (A-11070, RRID: AB_142134, Invitrogen) was used for 60 min at room temperature. They were then incubated overnight with BiP (C50B12) rabbit monoclonal antibody, 1:75 (Alexa Fluor 647 Conjugated no. 57163, Cell Signaling Technologies). Finally, samples were mounted by ProLong Diamond Antifade Mountant with DAPI (Thermo Fisher Scientific) and were imaged using a LSM780 microscope (Zeiss).

DeadEnd Colorimetric TUNEL System (Promega, Madison, WI) was used for TUNEL assay in mouse kidney tissues. TUNEL-positive cells were regarded as the occurrence of DNA damage (11).

Kidney tissue samples were fixed in 10% formaldehyde, embedded in paraffin, and sectioned at thickness of 4 μm. They were deparaffinized, rehydrated, and pretreated by microwave for 10 min in citrate buffer, pH 6.0. Endogenous peroxidase activity was blocked by 3% hydrogen peroxide. Nonspecific binding was blocked by incubation in 10% goat serum for 30 min. The tissue sections were incubated at 4°C overnight with anti-BiP, 1:200 (C50B12; no. 3177, Cell Signaling Technology); anti-fibronectin antibody, 1:500 (ab2413, abcam); TIM-1/kidney injury molecule 1 (KIM-1)/HAVCR antibody, 1:100 (NBP1-76701, Novus Biologicals); and integrin-β1/CD29 antibody, 1:1,000 (GTX128839; GeneTex). Tissue sections were washed in PBS and incubated with a goat anti-rabbit IgG ImmPRESS Secondary Antibody (HRP Polymer; Vector Laboratories). The information on knockout and knockdown validation antibodies is provided in Supplementary Table 2. Immunochemical staining was performed with the ImmPACT DAB Substrate kit (Vector Laboratories). Fibronectin-positive area and KIM-1–positive area were measured using ImageJ.

An ELISA kit for heat shock 70-kDa protein 5 (Cloud-Clone Corp.) was used following the manufacturer’s manual.

All values are presented as the mean ± SD. Statistical analyses were conducted using GraphPad Prism 8.0 software. Unpaired t test and one-way ANOVA with Tukey-Kramer was used to determine the differences. P < 0.05 was considered statistically significant.

All data supporting the findings of this study are available upon reasonable requests to the authors.

STZ-induced diabetic mice revealed decreased body weight and increased blood glucose, kidney weight-to-body weight ratio, urine volume, food intake, and serum creatinine level. CANA administration for 2 weeks suppressed STZ-induced hyperglycemia, but there were no significant differences in the kidney weight-to-body weight ratio, urine volume, food intake, or serum creatinine level in the STZ-induced diabetic states (Supplementary Fig. 1A–F). There were no differences of creatinine clearance among the four groups (Supplementary Fig. 1G). Urinary albumin excretion tended to increase in STZ mice and tended to decrease by CANA administration; however, there was no statistical significance (Supplementary Fig. 1H and I). Immunofluorescent staining showed GRP78 was localized in both SGLT2-positvie proximal tubular cells and SGLT2-negative distal tubular cells. In STZ-induced diabetic mice, SGLT2, shown as a green signal, was colocalized with red-colored GRP78 (Fig. 1). GRP78 is known to have multiple binding partner proteins on the cell surface (12), and we hypothesized that SGLT2 may also form a complex with GRP78. By immunoprecipitation using total lysate of STZ-induced diabetic mouse kidneys, we detected the complex formation of GRP78 and SGLT2 proteins (Supplementary Fig. 2A and B).

To investigate the localization of GRP78, we performed fractionation of mouse kidney cortex by iodixanol gradient ultracentrifugation. Western blot analysis confirmed the expression of GRP78 in the plasma membrane fraction, although its expression was very low compared with ER fractions (Fig. 2A). Analysis of concentrated mixed-membrane fraction samples (fractions 1 to 5) showed that the expression of GRP78 increased in STZ-induced diabetic mice in parallel with SGLT2 and that CANA suppressed the elevated expression in the STZ-induced diabetic model (Fig. 2B and C). In Western blot analysis using whole-kidney cortex samples, the expression of GRP78 was not altered in STZ mice, and CANA administration rather increased the expression (Fig. 2D and E). In immunofluorescence staining, the merged area of GRP78 and the ER marker calnexin was increased by CANA administration in both CTRL and STZ groups (Supplementary Fig. 3A). Therefore, there is a discrepancy in the distribution of GRP78: it is increased on the plasma membrane in STZ mice compared with CTRL, whereas CANA suppressed GRP78 localized on the plasma membrane and increased GRP78 on the ER.

CANA is known to be nephroprotective in clinical meta-analyses and in basic medical investigations in vivo and in vitro (13–15). Given the altered localization of GRP78, the following two hypotheses have been proposed as mechanism for the renoprotective effect of CANA. First, cellular stress in DKD alters the subcellular localization of GRP78 (i.e., ER stress induces the shift of GRP78 to cell surface and CANA restores GRP78 localization). Second, CANA may modify the regulation of ER stress itself via decreasing intracellular glucose levels.

Since GRP78 was originally identified as an upregulated molecule induced by glucose depletion (7), the inhibition of glucose influx into proximal tubular cells by CANA may mimic the state of glucose depletion. Intracellular glucose levels in HK2 cells cultured with 5.5 mmol/L glucose were decreased by treatment with CANA to the same level as when the cells were cultured in glucose-free medium (Fig. 3A). In contrary, glucose levels in culture medium were increased by treatment with CANA in a dose-dependent manner (Fig. 3B). In HK2 cells cultured with 5.5 mmol/L glucose medium, CANA increased the levels of GRP78 in Western blot analysis using total lysate (Fig. 3C and D). In fluorescent staining, GRP78, colocalized in the ER marker calnexin, was increased by CANA (Fig. 3E and F).

Although we could not evaluate statistical differences in the appearance of the ER by electron microscopy (Supplementary Fig. 3C), immunofluorescence double staining showed that GRP78 was increased in calnexin-positive ER by CANA in vivo (Supplementary Fig. 3A) and in vitro (Fig. 3E and F). We hypothesized that the significance of the increased expression of GRP78 was preconditioning of the unfolded protein response (i.e., the enhancement of ER robustness), and next performed GRP78 knockdown experiments.

We investigated whether CANA-related signals are attenuated by GRP78 knockdown in Western blotting (Fig. 4A). In HK2 cells, in which we confirmed the expression of both SGLT1 and SGLT2 (Supplementary Fig. 4A), under 5.5 mmol/L glucose, CANA promoted phosphorylation of AMPK in a dose-dependent manner, while its phosphorylation was reduced in shRNA-GRP78 knockdown cells (Fig. 4A, C, and D). The expression of E-cadherin was increased by CANA, and its expression was clearly suppressed in shRNA-GRP78 cells (Fig. 4B and D). In addition, shRNA-GRP78 cells showed increased expression of N-cadherin (Fig. 4B and D). This is consistent with a previous report that an activation of AMPK signaling protects proximal tubular cells from high glucose-induced epithelial-mesenchymal transition (EMT) (16). These results suggest that GRP78 is required for CANA-induced maintenance of the epithelial phenotype of proximal tubular cells.

GRP78 is known as an ER stress-related molecule that modifies unfolded and misfolded proteins. We examined the effect of CANA on ER stress using HK2 cells cultured in medium containing 5.5 mmol/L glucose.

ER stressors, such as thapsigargin and tunicamycin, increased the expressions of phosphorylated eukaryotic translation initiation factor 2α (eIF2α), activating transcription factor 4 (ATF4), and C/EBP homologous protein (CHOP) in HK2 cells, and thapsigargin-induced upregulation of these molecules was suppressed by administration of CANA, while tunicamycin-induced ER stress enhancement was not altered by CANA with statistical significance (Supplementary Fig. 4B and C). These findings suggest that CANA decreased thapsigargin-induced ER stress response failure independently of glucose concentration.

ER stress is known to be attenuated by AMPK activation, such as 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) and metformin (17). Tunicamycin inhibited AMPK phosphorylation, which was slightly restored by CANA, whereas thapsigargin enhanced AMPK phosphorylation (Supplementary Fig. 4B and C). Although thapsigargin and tunicamycin are both ER stressors, previous reports showed that thapsigargin increased spontaneous Ca²⁺ release from the ER, whereas tunicamycin did not significantly affect Ca²⁺ release from the ER into the cytosol (18). These results suggest that CANA reduces the failure of the ER stress response and affects intracellular Ca²⁺ concentration. Then, thapsigargin increased the expression of apoptosis inducer Bax, which was inhibited by CANA (Supplementary Fig. 4B and C).

Next, to demonstrate the role of Ca²⁺ dynamics in the ER stress response, we examined the effect of CANA on intracellular Ca²⁺ levels. Thapsigargin, which inhibits sarco/ER Ca²⁺-ATPase (SERCA), leads to an increase in intracellular Ca²⁺ levels. In experiments using Fluo-4, CANA suppressed thapsigargin-induced intracellular Ca²⁺ elevation (Fig. 5A and B). In the luciferase assay, we confirmed that CANA inhibited the increase in intracellular Ca²⁺ concentration induced by thapsigargin. It is reported that retinoic acid (RA) induces RA response element (RARE) activation at least in part by Ca²⁺ signaling (19). Therefore, we examined whether CANA inhibits an activation of RARE induced by thapsigargin or not in HK2 cells transfected with the DR5/RARE cis-reporter plasmid (Fig. 5C). The thapsigargin-induced firefly-to-Renilla luciferase activity ratio was suppressed by CANA (Fig. 5D). In contrast, tunicamycin reduced RARE activity, and CANA did not have additive effects on RARE activity (Fig. 5E). Tunicamycin is an N-linked glycosylation inhibitor with no inhibitory effects on SERCA activity. However, tunicamycin and a variety of ER stress inducers not known to alter the ER Ca²⁺ homeostasis also secondarily cause a marked reduction in ER Ca²⁺ levels (18). In contrast, thapsigargin induces ER stress via direct depletion of Ca²⁺ in the ER by suppressing SERCA activity (18).

A recent study reported that the reduction of intracellular Ca²⁺ levels via SERCA activity by CANA leads to vasodilation of the rabbit thoracic aorta (20). Our findings and other reports indicate that CANA inhibits intracellular Ca²⁺ increase via maintaining SERCA activity. Taken together, CANA maintains intracellular Ca²⁺ homeostasis and protects cells from cell damage caused by excess intracellular Ca²⁺.

In the previous section, we illustrated the possibility that phenotypic change of HK2 cells from epithelial to mesenchymal was suppressed by CANA (Fig. 4). The high-glucose condition (25 mmol/L glucose) decreased E-cadherin expression and increased α-smooth muscle actin (α-SMA) expression in HK2 cells, indicating a trend of phenotypic change from epithelial to mesenchymal (Fig. 6A and B). Administration of CANA maintained the E-cadherin protein level and suppressed the expression of α-SMA (Fig. 6A and B). In scratch assay using HK2 cells, 24-h incubation under 25 mmol/L glucose promoted the migration of cells into the scratch-free area, and CANA suppressed the invasion (Fig. 6C and D). In a proliferation assay, HK2 cells cultured under 25 mmol/L glucose conditions for 24 h actively proliferated compared with the 5.5 mmol/L glucose condition, and CANA inhibited the proliferation induced by high glucose (Fig. 6E). These results indicate the possibility of CANA inhibiting phenotypic change of HK2 cells induced by high glucose.

Incubation with high glucose for 48 h decreased GRP78 expression, but CANA reversed the change (Fig. 6A and B). Since shRNA-GRP78 HK2 cells promoted the phenotypic change from epithelial to mesenchymal type (Fig. 4A, B, and D), reduced GRP78 expression due to high glucose condition may also be involved in the mechanism of EMT.

We also investigated whether CANA inhibits tubulointerstitial damage in vivo. Enhanced ER stress and unfolded protein response failure induce cell apoptosis (21), but since CANA reduced ER stress enhancement in the previous section, we evaluated the number of TUNEL-positive cells in renal cortex of mice. TUNEL-positive DNA-damaged cells were increased in STZ group and were decreased by CANA treatment (Supplementary Fig. 5A). KIM-1 is known as a marker associated with inflammation in renal tubular injury (22). In immunohistochemical staining, the expression of KIM-1 in renal tissue of STZ-induced diabetic mice was increased and tended to be inhibited in CANA administration (Supplementary Fig. 5B). Western blot analysis using kidney proteins showed increased expression of Bax in STZ-induced diabetic wild-type mice, and it was ameliorated in CANA-treated diabetic mice (Supplementary Fig. 5C). Furthermore, 2 weeks after induction of diabetes by STZ, the fibronectin-positive area was increased compared with CRTL mice, whereas CANA administration decreased the area (Supplementary Fig. 6). In summary, CANA inhibited high glucose-induced EMT-like phenotypic changes of proximal tubular cells and ameliorated DNA damage of tubular cells and interstitial fibrosis in the STZ-induced diabetic kidney.

Western blot analysis of urine samples detected secreted (s)GRP78 in STZ-induced diabetes, and CANA administration suppressed urinary secretion of GRP78 in diabetic mice (Supplementary Fig. 7). We hypothesized that GRP78 secreted into the tubular lumen would act on the surrounding or downstream tubular cells and induce a proinflammatory response. HK2 cells cultured with 25 mmol/L glucose increased the secretion of GRP78 into the culture media, and CANA inhibited the secretion of GRP78 (Fig. 7A). In contrast, administration of BSA also increased the secretion of GRP78, and CANA failed to inhibit the release of GRP78 (Fig. 7B).

We further investigated sGRP78-induced intracellular signaling in HK2 cells. The administration of rhGRP78 protein into HK2 cells increased phosphorylation of p65NF-κB associated with decreased AMPK phosphorylation. In HK2 cells treated with rhGRP78, CANA inhibited the expression of phosphorylated p65NF-κB and phosphorylated Akt and tended to reverse the inhibition of AMPK phosphorylation in Western blotting analysis (Fig. 7C–F). Phosphorylation of p65NF-κB and Akt are reported to promote EMT (23,24). Immunofluorescence staining showed that accumulation of p65NF-κB in nuclei was observed in HK2 cells cultured with 25 mmol/L glucose-containing medium, and nuclear staining of p65NF-κB was suppressed by CANA treatment (Fig. 7G and H). It is plausible that sGRP78 induces phosphorylation of p65NF-κB and Akt in surrounding or distal tubular cells, which subsequently promotes tubulointerstitial injury with fibrosis.

Tissue fibrosis is caused by the interaction between cells and the surrounding extracellular matrix (ECM). Several cellular adhesion molecules mediate the interaction, and integrins are of particular relevance to renal fibrosis (25). We therefore investigated the inhibitory effects of CANA via GRP78 and integrin-β1 on EMT of proximal tubular cells and renal fibrosis.

GRP78 is reported to form a complex with integrin-β1 on the cell surface (26), and we confirmed the binding of GRP78 and integrin-β1 in p3xFlag-GRP78 overexpressed HEK293T cells and mouse kidney samples (Fig. 8A and B). In Western blotting using total kidney lysate, integrin-β1 was increased in STZ-induced diabetic mice, and CANA suppressed this integrin-β1 level (Fig. 8C and D). A previous study reported that the localization of GRP78 and its interaction with integrin-β1 on the cell surface of mesangial cells promotes ECM protein synthesis by Akt activation in response to high glucose (27). In our study, Western blot analysis of whole kidney samples from STZ-induced diabetic mice showed that phosphorylated Akt was enhanced and that phosphorylation was suppressed by CANA administration (Fig. 8C and D). In the mixed plasma membrane fraction sample of kidney tissue, integrin-β1 was also increased in STZ mice (Fig. 8F) in parallel with GRP78 (Fig. 2B), and integrin-β1 and GRP78 were both suppressed in CANA-treated mice. In immunohistochemical staining of kidney samples of control mice, integrin-β1 staining was found primarily at the cell periphery, whereas cytoplasmic integrin-β1 was increased in the tubules of STZ-induced diabetic mice. CANA restored integrin-β1 localization, cytoplasmic staining was reduced, and cell surface linear staining was maintained (Fig. 8E).

In summary, CANA contributed to the maintenance of the appropriate amount and localization of integrin-β1 in tubular cells.

In recent clinical trials of A Study to Evaluate the Effect of Dapagliflozin on Renal Outcomes and Cardiovascular Mortality in Patients With Chronic Kidney Disease (Dapa-CKD) (28), Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) (13), and The Study of Heart and Kidney Protection With Empagliflozin (EMPA-KIDNEY) (29), SGLT2 inhibitors demonstrated favorable effects on renal outcome of both DKD and non-DKDs. The beneficial actions of SGLT2 inhibitors have been shown, such as a correction of tubule-glomerular feedback, increased sodium excretion into urine, amelioration of fluid retention, lowering inflammation, oxidative stress and fibrosis, increased erythropoietin-producing cells, and amelioration of tubulointerstitial hypoxia, beyond glucose-lowing effects and improvement of obesity (30,31). Although these renoprotective impacts of SGLT2 inhibitors have been well demonstrated, the plausible molecular mechanisms are still unraveled. Other groups have suggested mechanisms underlying renoprotection of SGLT2 inhibitors in STZ-induced diabetes through activation of the YAP/TAZ (31) or HMGB1 pathways (32) or inhibition of the p53 pathway (33).

Low-grade microinflammation is a risk for atherosclerosis, cardiovascular diseases, and chronic kidney disease. A meta-analysis of experimental animal models showed that SGLT2 inhibition resulted in reduction of inflammatory markers, interleukin-6 and C reactive protein, tumor necrosis factor-α (TNF-α), and monocyte chemoattractant protein-1 (MCP-1), independent of whether they were diabetic animal models or not (34). In a human study, dapagliflozin therapy for 6 weeks decreased albuminuria, urinary excretion of KIM-1 and interleukin-6, which was caused by a reduction of glomerular hyperfiltration and improved integrity of proximal tubular cells (35). In a post hoc mediation analysis of the Canagliflozin Cardiovascular Assessment Study (CANVAS) trial, CANA also reduced urinary KIM-1, suggesting decreased tubular damage (14). In this report, the authors estimated the direct and indirect effects of CANA on urine albumin-to-creatinine ratio, monocyte chemoattractant protein-1/creatinine, and KIM-1/creatinine using g-computation, showing that half of the beneficial effects on tubular damage are a consequence of a decrease in albuminuria and/or monocyte chemoattractant protein-1, the other half are mediated by direct effects on proximal tubular cells. These reports indicate that SGLT2 inhibitors have anti-inflammatory effects as well as yet uncharacterized beneficial actions on proximal tubular cells.

We investigated the significance of GRP78 in the favorable effects of CANA on proximal tubular cells. GRP78 was originally identified as a molecule upregulated under the glucose-deprived condition (7). GRP78 has been known as an ER stress-related molecule that facilitates protein folding and assembly, controls protein quality, and regulates ER stress signaling. In the recent efforts to search the subcellular localization of GRP78, the cell-surface and secreted forms have been identified, and it demonstrates the unique roles in various pathological conditions besides the role of a chaperone in the ER (36,37). In response to cellular stress, GRP78 translocates from ER to plasma membrane and functions as a receptor for several ligands, including activated α2-macroglobulin, plasminogen kringle 5, microplasminogen, prostate apoptosis response-4 protein (Par-4), vaspin, and sGRP78 itself (12,37,38).

We previously reported that the ER stress inducer, thapsigargin, promotes translocation of GRP78 to the plasma membrane in hepatocytes and vascular endothelial cells and functions as a receptor for vaspin (4,5). Vaspin is an adipokine that we previously identified, and in vaspin transgenic mice, insulin resistance and fatty liver induced by a high-fat high-sucrose diet were ameliorated (4). In WKY rats in which vaspin was overexpressed in the carotid intima using adenovirus, vaspin gene delivery suppressed intimal thickening induced by balloon-injury. In the cuff injury model of the femoral artery, vaspin transgenic mice again demonstrated the reduction of intimal thickening (5). During the efforts to identify the interacting partners, we found that vaspin forms a complex with cell surface GRP78, is internalized into proximal tubular cell by clathrin-mediated endocytosis, and protects cells from organelle stresses associated with obesity and diabetes (6). In the current study, we focused on GRP78 and SGLT2 interaction, because the expression of both molecules is tightly regulated by glucose concentrations in the microenvironment of the cells.

We found increased expression of SGLT2 in the plasma membrane fraction of STZ-induced diabetic mice, and it was suppressed by CANA administration. Simultaneously, an expression of GRP78 also increased on the plasma membrane fraction of STZ-induced diabetic mice and was suppressed by CANA, although protein levels of GRP78 in total lysates were similar. We presumed that GRP78 was translocated to the plasma membrane fraction of STZ-induced diabetic kidney due to enhanced cellular stress. A previous report demonstrated that the increase in cell-surface GRP78 due to ER stress is not simply paralleled by an increase in the total amount of GRP78 in the cell, but rather, that ER stress promotes a mechanism that facilitates the localization of GRP78 to the cell surface (i.e., escape from the KDEL receptor-mediated recovery to ER may promote cell surface expression and formation of the partner protein complex) (39). Immunofluorescence staining revealed that GRP78 was partially colocalized with SGLT2 in STZ-induced diabetic kidney, and we also confirmed a complex formation of SGLT2 and GRP78. For the searching the renoprotective mechanism of SGLT2 inhibitors, we hypothesized unexplained roles of SGLT2 may be mediated by GRP78. We found that phospho-AMPK and E-cadherin were increased by CANA, while the increases in phospho-AMPK and E-cadherin were partially abolished in GRP78-knockdown HK2 cells. In line with this evidence, we further investigated the effects of CANA on the interplay between AMPK and GRP78 and subsequent cellular events such as EMT.

CANA has been reported to activate AMPK, induce autophagy, and protect proximal tubular cells from cisplatin-induced acute kidney injury (40). Increased expression of GRP78 also activates AMPK and contributes to initiation of autophagy (41). GRP78, an ER stress-related molecule, has been reported to be protective against cellular stress as follows. Ischemic preconditioning-induced GRP78 upregulation is involved in autophagy activation and protection against ischemic injury in neural cells (42). Valproate was reported to increase the expression of GRP78 associated with Ca²⁺ binding and chaperone potentials and to enhance neuronal resistance to cytotoxicity (43). GRP78 is involved in various fundamental cellular functions, including regulation of calcium homeostasis (41). Ca²⁺ is a second messenger that regulates various intracellular signaling events. Under unstimulated conditions, the intracellular Ca²⁺ concentration remains low. Upon stimulation, Ca²⁺ is released from ER and it increases the intracellular Ca²⁺ levels. Subsequently, Ca²⁺ is returned from cytosol to the ER by an ion pump SERCA to maintain the basal Ca²⁺ levels. Thapsigargin is a SERCA inhibitor that blocks Ca²⁺ reuptake into the ER and is the established inducer of ER stress. Intracellular Ca²⁺ efflux, induced by thapsigargin or calcium ionophore, enhances ER stress with increased expression of GRP78 as a compensatory mechanism for cell survival, while persistent or excessive stress leads to compensatory failure and cell apoptosis (44). In calcium assays, CANA suppressed the acute-phase intracellular Ca²⁺ elevation prompted by thapsigargin, suggesting that CANA maintains the SERCA activity. At a later stage, the ER stressors, tunicamycin and thapsigargin, increased the known ER stress markers such as phosphorylation of eIF2α, AFT4, and CHOP, whereas CANA suppressed these changes. Since GRP78 is upregulated by glucose deprivation, we initially hypothesized that CANA reduces the influx of glucose into the cells, increases the expression of GRP78, and compensates ER stress response failure induced by tunicamycin and thapsigargin in HK2 cells. In the current investigation, the high glucose-induced shift of GRP78 to plasma membrane was reversed by CANA, and one can speculate that CANA corrects the ER stress response failure and recovers ER robustness. ER stress is increased in the proximal tubular cells under pathophysiological conditions, such as oxidative stress, dyslipidemia, hyperglycemia, and hypoxia, and CANA protects the cells from ER stress by maintaining SERCA activity.

Recently, many clinical studies have clearly shown that SGLT2 inhibitors have cardioprotective effects and prevent the progression of heart failure. In basic experiments, SGLT2 inhibitors are reported to have direct effects on cardiomyocytes by altering activity of ion channels and/or transporters, including the sodium-calcium exchanger and sodium-hydrogen exchanger 1 (45). However, the expression of SGLT2 in cardiomyocytes is controversial, and some studies reported a lack of SGLT2 expression, in contrast to SGLT1 being highly expressed in cardiomyocytes (46,47). CANA is a selective SGLT2 inhibitor with modest SGLT1 inhibitory activity at clinical doses (48). Therefore, maintaining SERCA activity may be one of the beneficial mechanisms of CANA for cardiomyocytes, although further investigation is needed to confirm the effects of SGLT2 inhibitors on cardiomyocytes. Since expression of both SGLT1 and SGLT2 was observed in HK2 cells in this experiment, we cannot completely negate the possibilities that the cytoprotective action of CANA may involve the inhibitory effects on SGLT1 besides SGLT2.

Previously, it was reported that in proximal tubular cells in vitro, high glucose induces premature senescence and EMT, which is accompanied by downregulation of AMPK/mTOR signaling (16). EMT plays an important role in renal tissue fibrosis exposed to chronic hyperglycemia (24). Hyperglycemia causes sustained renal injury and increased ECM deposition, which ultimately leads to renal fibrosis. Fibroblasts and myofibroblasts are the major contributors of tissue fibrosis; however, the origin of myofibroblasts in DKD remains controversial (49). Recently, single-cell RNA sequencing revealed that pericytes and fibroblasts are the major cellular sources of myofibroblasts, while tubular epithelial cells, endothelial cells, and monocytes produced small amount of ECM (50). However, injured tubular epithelial cells could also release inflammatory cytokines and chemokines on surrounding cells, such as fibroblasts, causing them to transform into myofibroblasts and produce ECM (50). We found that CANA increased AMPK phosphorylation and simultaneously inhibited the high glucose-induced phenotypic change of HK2 cells from epithelial to mesenchymal cells.

Furthermore, sGRP78 from HK2 cells was increased by high glucose, and CANA inhibited the release of GRP78 from the cells. Extracellular vesicles carry important information that communicates cells to cells or organs to organs, including kidney. In kidney disease, podocytes and tubular cells secrete extracellular vesicles and might have a role in tubulointerstitial inflammation (51).

We previously reported that coculture with BSA promoted the secretion of HSP family A (Hsp70) member 1 like (HSPA1L) from HK2 cells and that vaspin inhibited its secretion (6). HSPs are secreted from cells as free HSPs, exosomes, or microvesicles. The roles of extracellular HSPs have been known in cancer, immunity, and various pathological conditions (52). However, the role of extracellular HSPs in kidney disease is barely elucidated. GRP78 also belongs to the HSP family, and sGRP78 is known as proapoptotic ligand of cell-surface GRP78.

In addition, recombinant GRP78 induces apoptosis in cytokine-treated pancreatic β-cells (53), and anti-GRP78 antibody inhibits extracellular GRP78-induced apoptosis. These findings suggest that soluble extracellular GRP78 activates cell-surface GRP78 in cytokine-exposed cells (53). We speculated that sGRP78 from proximal tubular cells under high glucose or some stressed conditions may promote inflammation and proapoptotic signaling in surrounding or downstream tubular cells. In this study, CANA activated AMPK in HK2 cells and inhibited recombinant GRP78-induced phosphorylation of p65NF-κB, which is a significant mediator of inflammation. In summary, our results suggest that CANA inhibits GRP78 secretion in proximal tubular cells as well as suppressing the p65NF-κB pathway activated by soluble GRP78.

This study has some limitations. We found the interaction between SGLT2 and GRP78; however, we could not conclude whether it was a direct physical binding or not. SGLT1 is reported to interact with Hsp70 to increase SGLT1 activity without altering SGLT1 expression (54). It was shown that Hsp70 forms a complex with SGLT1 and increases the expression level of SGLT1 in the apical membrane, resulting in upregulation of glucose uptake (54). Therefore, it is reasonable to speculate that SGLT2 may form a complex with GRP78, which is also classified as the Hsp70 family, and translocate to the apical membrane. Further investigations are needed.

Next, the current investigation is an observational study at the time of induction of diabetes with STZ and 2 weeks after the administration of CANA. The short observation period after the onset of diabetes may be one reason for the lack of significant differences in urinary albumin excretion. A previous study showed that C57BL/6 mice develop albuminuria within 25 weeks after the injection of STZ (55). However, as the renoprotective effect of SGLT2 inhibitors has recently been observed in the absence of albuminuria, our model shows the beneficial effect of CANA beyond albuminuria-lowering effects. In addition, the STZ-induced diabetes model shows very high glucose levels, and STZ itself is toxic. The investigation of different animal models of DKD or chronic kidney disease are further required.

In conclusion, in addition to inhibiting glucose uptake, CANA protects proximal tubular cells from high glucose–induced EMT-like phenotypic changes, inhibits secretory GRP78, which spreads inflammatory signals to surrounding cells, and inhibits interstitial fibrosis. In nondiabetic conditions, reduction of intracellular glucose levels by CANA promotes ER robustness and ameliorates ER stress response failure via upregulation of GRP78 protein (Supplementary Fig. 8). These effects are beneficial in both DKD and non-DKD (Supplementary Fig. 9).

Acknowledgments. The authors acknowledge support from Central Research Laboratory, Okayama University Medical School, use of Beckman Coulter XL 80, LSM780 (Zeiss), FlexStation (Danaher Corp.), and the production of paraffin blocks and sections. The authors also acknowledge excellent technical support from Masumi Furutani at Central Research Laboratory, Okayama University Medical School, for performing morphological analyses by electron microscopy.

Funding. The study was conducted in collaboration with the Japan Kidney Foundation and Mitsubishi Tanabe Pharma Corporation. This work was supported by a Grant-in-Aid for Scientific Research (C) 23K07673 to A.N.

Duality of Interest. The study was conducted in collaboration with Mitsubishi Tanabe Pharma Corporation. A.N. receives grant support from Mitsubishi Tanabe Pharma Corporation. J.W. receives speaker honoraria from AstraZeneca, Bayer, Boehringer Ingelheim, Daiichi Sankyo, Kyowa Kirin, Novo Nordisk Pharma, and Mitsubishi Tanabe Pharma Corporation and receives grant support from Bayer, Chugai, Kyowa Kirin, Otsuka, Shionogi, Sumitomo Pharma, and Mitsubishi Tanabe Pharma Corporation. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. A.N. and S.Y. performed experiments and analyzed and interpreted data. A.N. and J.W. designed the project and experiments and wrote the manuscript. A.N and J.W are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.