Effects of Progranulin Deficiency on Inflammation and Fibrosis in the Kidneys and Liver of Diabetic Mice Fed a High-Fat Diet
Article information
Abstract
Background
Progranulin (PGRN) is an important regulator of inflammation, insulin resistance, and autophagy. However, the effects of PGRN deficiency on these processes in the kidneys and liver in diabetes remain unclear. In addition, the differential effects of PGRN deficiency and sodium-glucose co-transporter-2 (SGLT2) inhibitors on these organs are unknown.
Methods
Three diabetic mouse models were used: high-fat diet and nicotinamide/streptozotocin-induced diabetic wild-type (WT) and PGRN-knockout (KO) mice (WT-diabetes mellitus [DM] and KO-DM, respectively) and WT-DM mice treated with an SGLT2 inhibitor (tofogliflozin; WT-DM/Tofo).
Results
Despite similar glycemic control in WT-DM/Tofo and KO-DM mice, expression of inflammation- and fibrosis-related genes in the kidneys was highest in WT-DM mice, lower in KO-DM mice, and lowest in WT-DM/Tofo mice. WT-DM/Tofo mice also showed increased anti-microtubule-associated protein 1A/1B-light chain 3B and decreased p62 protein levels compared with KO-DM mice. In contrast, hepatic mRNA levels related to inflammation and fibrosis were improved in both WT-DM/Tofo and KO-DM mice. Moreover, hepatic protein levels of peroxisome proliferator-activated receptor γ (PPARγ) were elevated in both groups compared with WT-DM mice, while those of PPARα were increased in WT-DM/Tofo mice compared with both WT-DM and KO-DM mice.
Conclusion
Kidney inflammation and fibrosis were ameliorated in WT-DM/Tofo mice, but these improvements were limited by autophagy insufficiency in KO-DM mice. Additionally, both WT-DM/Tofo and KO-DM mice demonstrated improved liver inflammation and fibrosis; in the former, this was associated with enhanced fatty acid oxidation via PPARα activation, while in the latter, it appeared to result from improved insulin sensitivity and anti-inflammatory effects through PPARγ activation.
INTRODUCTION
Progranulin (PGRN) is a growth factor that plays a key role in regulating inflammation and glucose metabolism in various diseases through its effects on insulin resistance [1-5]. We previously reported that PGRN-knockout (KO) mice fed a high-fat diet (HFD) exhibited improved adipose tissue inflammation but exacerbated kidney inflammation compared with wild-type (WT) mice, suggesting that PGRN may exert organ- or tissue-specific effects [6]. Some reports indicate that administration of PGRN to mice increases blood glucose levels and aggravates insulin resistance in the liver, although the effects of PGRN on the liver have not been fully established [7,8].
Following traumatic brain injury, PGRN-KO mice exhibit reduced activation of mechanistic target of rapamycin complex 1 (mTORC1), resulting in elevated nuclear translocation of transcription factor EB, enhanced lysosomal biogenesis in activated microglia, and increased neuronal damage in the cerebral cortex [9]. Generally, mTORC1 promotes cell growth by inhibiting cellular catabolism through suppression of autophagy in states of overnutrition [9-14]. Several lines of evidence suggest that the pathogenesis of diabetic kidney disease is associated with impaired autophagic activity via activation of the mechanistic target of rapamycin (mTOR) pathway [15,16]. Inhibition of sodium-glucose co-transporter 2 (SGLT2) induces starvation- and hypoxia-like transcriptional changes, thereby activating autophagy in overnutrition-related diseases such as obesity and diabetes [17-21]. Indeed, several studies have demonstrated that SGLT2 inhibitors exert renoprotective effects in diabetic mice by suppressing the mTOR pathway and improving autophagic flux. However, it remains unclear whether PGRN deficiency has pathogenic or protective effects on the kidney and liver via the mTOR pathway in diabetic mouse models [17,22].
This study aimed to investigate the effects of PGRN deficiency on inflammation, fibrosis, and autophagy in the kidneys and liver of diabetic mice. In addition, it sought to determine how PGRN deficiency and SGLT2 inhibitors differentially affect these organs in diabetic mice.
METHODS
Animals and experimental protocols
Heterozygous PGRN-KO mice, developed by Kayasuga et al. [23], were obtained from the RIKEN Bioresource Center (Ibaraki, Japan). Genotyping of homozygous PGRN KO mice was performed as previously described [23]. Five-week-old male C57BL/6J mice were purchased from Sankyo Labo Service Corporation, Inc. (Tokyo, Japan) and used as WT mice in this study. WT mice were randomly assigned to one of three groups: WT-control (WT-CTRL) (n=5), WT-diabetes mellitus (WT-DM) (n=5), and WT-DM/tofogliflozin (WT-DM/Tofo) (n=5). PGRN-KO mice were divided into two groups: KO-CTRL (n=5) and KO-DM (n=5). CTRL mice were fed a normal diet (CE-2: 3.39 kcal/g; CLEA Japan, Tokyo, Japan), while DM mice were fed an HFD (HDF32: 5.1 kcal/g, 60% of calories from fat [32%/g crude fat content]; CLEA Japan) starting at 6 weeks of age. Additionally, nicotinamide (120 mg/kg) and streptozotocin (100 mg/kg) were injected intraperitoneally twice, following a 16-hour fast, every other day at 10 weeks of age. After 1 week, WT-DM/Tofo mice were fed an HFD containing 0.015% Tofo (an SGLT2 inhibitor) (Fig. 1). Tofo was generously provided by Kowa Company, Ltd. (Tokyo, Japan). Mice with blood glucose levels above 230 mg/dL 1 week after streptozotocin/nicotinamide administration were included in the study [24]. All mice were housed in plastic cages with free access to food and water, under specific-pathogen-free conditions of controlled temperature, humidity, and light (12-hour light: 12-hour dark cycle).
Experimental protocol of the study. The control groups were observed from 6 to 20 weeks of age on a standard diet (SD), while the streptozotocin (STZ)/nictotinamide (NA) groups were observed over the same period on a high-fat diet (HFD). STZ and NA were administered intraperitoneally twice at 10 weeks of age. After 1 week, the wild-type (WT)-diabetes mellitus (DM)/tofogliflozin (Tofo) group received 0.015% Tofo from 11 to 20 weeks of age. CTRL, control; KO, knockout.
At 20 weeks of age, all mice were sacrificed by intraperitoneal pentobarbital anesthesia (60 mg/kg) after 16 hours of fasting. All animal experiments were approved by the Ethics Review Committee for Animal Experimentation of the Juntendo University Faculty of Medicine (document No. 1312), and all animals were treated according to the institutional guidelines for experimentation of Juntendo University in Tokyo, Japan. Blood glucose levels were measured using a glucometer (Glutest Mint; Sanwa Kagaku Kenkyusho Co. Ltd., Aichi, Japan) at 11, 15, and 20 weeks of age. At 20 weeks, glycated hemoglobin (HbA1c) was measured using a DCA 2000 HbA1c immunoassay cassette and a DCA Vantage Analyzer (Siemens Healthcare, Erlangen, Germany). Serum triglycerides were measured using the TG E-Test Wako (Wako Pure Chemical Industries Ltd., Osaka, Japan). Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were quantified using the Fuji Drychem 7000V (Fujifilm Co., Tokyo, Japan). Hepatic triglyceride content was assessed using a Triglyceride Colorimetric Assay Kit (Cayman Chemical, Ann Arbor, MI, USA).
Quantitative polymerase chain reaction
Total RNA was extracted from kidney and liver tissues using the RNeasy Mini kit (Qiagen, Hilden, Germany) and reverse-transcribed into complementary DNA according to the manufacturer’s protocols. Quantitative polymerase chain reaction (qPCR) was performed using TaqMan real-time PCR (Applied Biosystems, Foster City, CA, USA) on a 7,500 Fast Real-Time PCR system to analyze the expression of related genes. The mRNA levels of actin alpha 2 (Acta2), serpin family E member 1 (Serpine1), fibronectin 1 (Fn1), tumor necrosis factor (Tnf), and C-C motif chemokine ligand 2 (Ccl2) in kidney and liver tissues, as well as stearoyl-CoA desaturase 1 (Scd1), sterol regulatory element-binding protein 1 (Srebf1), and sterol regulatory element-binding protein 2 (Srebf2) in liver tissue, were measured using commercially available assays from Applied Biosystems (Supplemental Table S1). Expression levels of target genes were normalized to 18S rRNA, and the ΔΔcycle threshold method was used for quantitative analyses.
Histochemical studies
For histological assessment, kidney and liver tissues were fixed in 20% neutral-buffered formalin, embedded in paraffin, and sectioned at 4 μm thickness. Kidneys were dewaxed using standard sequential techniques and stained with periodic acid–Schiff. Paraffin-embedded liver sections were stained with hematoxylin and eosin to quantify steatosis, fat droplets, and hepatic inflammation. Frozen kidney and liver tissues were sectioned at 10 μm thickness, fixed in 10% buffered formalin, and subjected to Oil Red O and Nile blue staining for lipid droplet visualization. Triglycerides are stained red by both Oil Red O and Nile blue, whereas phospholipids are stained blue by Nile blue [25].
Immunohistochemistry
For immunohistochemistry, kidney sections were embedded in paraffin to prepare for light microscopy. Immunohistochemical staining was performed using a commercially available rabbit polyclonal anti-microtubule-associated protein 1A/1B-light chain 3B (LC3B) antibody (ab51520; Abcam, Cambridge, UK). Nuclei were counterstained with hematoxylin. Photomicrographs were acquired under an optical microscope at ×200 magnification. LC3B staining was quantified in at least five randomly selected fields (×200 magnification) from each mouse using the KS-400 image analysis system version 4.0 (Carl Zeiss Vision, Munich, Germany). For immunofluorescence, a guinea pig polyclonal anti-p62 antibody (GP62-C; Progen Biotechnik GmbH, Heidelberg, Germany) was used. Kidney tissue was fixed with 4% paraformaldehyde for 10 minutes, blocked in 2% bovine serum albumin/phosphate-buffered saline (PBS) for 30 minutes, and incubated overnight with primary anti-p62 antibodies at 4°C. After washing with PBS, the tissue was incubated with secondary antibodies conjugated to Alexa Fluor 488 (A11073; Invitrogen, Waltham, MA, USA) for 30 minutes, then washed again with PBS and examined using a fluorescence microscope (Olympus FluoView version 4.2b, Olympus, Tokyo, Japan). The positively stained area in at least five glomeruli from each mouse was quantified using the KS-400 system.
Transmission electron microscopy and morphometry
For transmission electron microscopy, small fragments of kidney tissue were fixed overnight in 2% glutaraldehyde and postfixed in 1% osmium tetroxide. The tissues were then dehydrated with graded ethanol and embedded in epoxy resin. Semithin sections were stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate and examined using an electron microscope (H-7700, Hitachi, Tokyo, Japan) at 100 kV. Five glomeruli and proximal tubules were randomly selected from each mouse, and 10 electron micrographs were obtained from each.
Western blot analysis
Kidney and liver tissues were lysed in radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors, separated by 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred to polyvinylidene fluoride membranes. Membranes were blocked with a blocking reagent (0.1% Tween 20 and 5% bovine serum albumin in Tris-buffered saline) for 1 hour and subsequently incubated overnight at 4°C with the following primary antibodies (1:1,000): phospho-S6 (#2211; Cell Signaling Technology, Danvers, MA, USA), S6 (#2217; Cell Signaling Technology), phosphorylated AMP-activated protein kinase (p-AMPK) (#2531; Cell Signaling Technology), AMPK (#2532; Cell Signaling Technology), phospho-Akt (#4060; Cell Signaling Technology), Akt (#4691; Cell Signaling Technology), peroxisome proliferator-activated receptor α (PPARα) (#SC-398394; Santa Cruz Biotechnology, Dallas, TX, USA), and PPARγ (#2435; Cell Signaling Technology). After incubation with secondary antibodies (1:10,000) for 1 hour at room temperature, signals were developed using Super Signal West Dura Extended Duration Substrate (#34075; Thermo Fisher Scientific, Waltham, MA, USA) and detected with an enhanced chemiluminescence detection system (Fusion FX7; Fisher Biotec, Wembley, Australia). Protein bands were analyzed densitometrically using Fusion Capt Advance (Vilber Bio Imaging, Collégien, France). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (G8795, Sigma-Aldrich, Darmstadt, Germany) was used as an internal control.
Statistical analyses
Results are expressed as mean±standard deviation or median (interquartile range). Group means were compared using one-way analysis of variance (ANOVA). Data with two independent variables were analyzed by two-way ANOVA. Tukey multiple comparisons tests were performed as appropriate. P values less than 0.05 were considered to indicate statistical significance. Statistical analyses were performed using GraphPad Prism 7 (GraphPad Software Inc., San Diego, CA, USA).
RESULTS
Phenotypic characterization and inflammatory and fibrotic gene expression according to PGRN and DM status
To investigate how PGRN and DM interact to influence metabolic and inflammatory states, we first assessed the general phenotypes of the experimental mice (Table 1). Relative to controls, DM mice had significantly higher body weight; levels of HbA1c, AST, ALT, and serum triglycerides; and hepatic triglyceride content. WT mice had significantly higher HbA1c, ALT, and lipid levels than their PGRN-deficient KO counterparts, suggesting that PGRN may exacerbate metabolic disturbances.
Phenotypic and Metabolic Characterization of Each Group
When both DM and PGRN deficiency were present (KO-DM mice), the phenotype of metabolic disturbance was ameliorated: HbA1c, AST, ALT, and serum triglycerides were all significantly lower than in WT-DM mice. This observation suggests that the absence of PGRN may confer partial protection against DM-induced metabolic derangement.
To further clarify the anti-inflammatory and anti-fibrotic potential of PGRN deficiency, we examined the expression of relevant genes. In the kidney, DM mice exhibited significantly higher expression of pro-inflammatory (Ccl2, Tnf) and fibrotic (Serpine1, Fn1) markers compared with control mice. WT mice displayed significantly higher expression of Acta2 and Ccl2 than KO mice, and KO-DM mice had significantly lower expression of Acta2, Fn1, and Ccl2 than WT-DM mice. A similar pattern was observed in the liver, where KO-DM mice exhibited significantly lower expression of Tnf, Acta2, Fn1, and Ccl2 than WT-DM mice (Table 2).
mRNA Levels of Inflammation- and Fibrosis-Related Genes in the Kidney and Liver
These findings raised an important question: were the improvements in KO-DM mice driven primarily by the absence of PGRN, or were they a byproduct of superior glycemic control? To address this, we introduced a third group—WT-DM mice treated with the SGLT2 inhibitor Tofo (WT-DM/Tofo)—which exhibited HbA1c levels comparable to those of KO-DM mice (Fig. 2).
Phenotypic and metabolic characterization of each group. The figure compares (A) body weight; (B) glycated hemoglobin (HbA1c); serum levels of (C) aspartate aminotransferase (AST), (D) alanine aminotransferase (ALT), and (E) triglycerides; and (F) hepatic triglyceride content at 20 weeks. Comparisons among groups were performed using one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. DM, diabetes mellitus; Tofo, tofogliflozin; WT, wild-type; KO, knockout. aP<0.05, bP<0.01 vs. WT-DM; cP<0.05 vs. WT-Tofo.
Phenotypic comparisons among WT-DM, KO-DM, and WT-DM/Tofo mice
As shown in Fig. 2, both KO-DM and WT-DM/Tofo mice had significantly lower HbA1c and ALT levels compared with WT-DM mice, while body weight remained similar across groups. AST levels were also significantly lower in WT-DM/Tofo mice than in WT-DM mice. Although serum triglyceride levels did not differ significantly among the three groups, hepatic triglyceride content was significantly lower in WT-DM/Tofo mice compared with the WT-DM group. A similar, non-significant trend was observed in KO-DM mice compared with WT-DM animals.
These results suggest that while both interventions improved glycemic control, Tofo exerted a more pronounced effect on hepatic lipid accumulation, prompting further investigation of underlying histological and molecular changes.
Tofo treatment and PGRN deficiency alleviate renal phospholipid accumulation and hepatic steatosis
Histological analysis of the kidney revealed severe vacuolation of proximal tubules in WT-DM mice (Fig. 3A). These vacuoles, identified as phospholipid-rich by Nile blue staining, were largely absent in WT-DM/Tofo mice and only mildly present in KO-DM mice. In the liver (Fig. 3B), both intervention groups (WT-DM/Tofo and KO-DM) exhibited reduced lipid droplet formation compared with WT-DM mice. Notably, Oil Red O staining indicated that in KO-DM mice, it was droplet size, rather than number, that was primarily reduced. This result was consistent with the absence of a significant decrease in hepatic triglyceride content in these mice.
Histological findings in the kidney and liver of each group. (A) Periodic acid–Schiff (PAS), Oil Red O, and Nile blue staining of kidney tissue (scale bar: 100 µm). Vacuolation was observed in the proximal tubules of the wild-type (WT)-diabetes mellitus (DM) group (closed arrowhead) but not in the other groups, as shown by PAS staining of kidney sections from 20-week-old mice. Vacuoles were stained blue with Nile blue. (B) Hematoxylin and eosin (HE), Oil Red O, and Nile blue staining of liver tissue (scale bar: 200 µm). Fatty liver was observed in the WT-DM group but not in the other groups. Vacuoles were stained red with Oil Red O and Nile blue (open arrowhead). KO, knockout; Tofo, tofogliflozin.
These findings indicate that both PGRN deficiency and SGLT2 inhibition ameliorate renal phospholipid accumulation, but only Tofo robustly reduces hepatic steatosis, suggesting an organ-specific mechanism.
Inflammatory and fibrotic gene expression in the kidney and liver
To determine whether histological improvements were accompanied by molecular changes, we next assessed the expression of inflammatory and fibrotic genes in the kidney and liver (Fig. 4). In the kidney, WT-DM/Tofo mice exhibited significantly lower expression of inflammatory and fibrotic markers than WT-DM mice, with the exception of Acta2. KO-DM mice had significantly lower Ccl2 and Serpine1 expression compared with WT-DM mice. While no significant differences were observed between WT-DM/Tofo and KO-DM mice, a trend toward lower expression in the Tofo group was noted.
mRNA levels of fibrosis- and inflammation-related genes in the (A) kidney and (B) liver of each group. mRNA expression levels of fibrosis- and inflammation-related genes. Comparisons among groups were performed with one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. DM, diabetes mellitus; Tofo, tofogliflozin; WT, wild-type; KO, knockout; Tnf, tumor necrosis factor; Ccl2, C-C motif chemokine ligand 2; Acta2, actin alpha 2; Serpine1, serpin family E member 1; Fn1, fibronectin 1. aP<0.05, bP<0.01 vs. WT-DM.
To complement these findings, transmission electron microscopy was performed (Fig. 5). WT-DM kidneys displayed mitochondrial swelling and podocyte foot process effacement, which were markedly ameliorated in WT-DM/Tofo mice and partially improved in KO-DM mice. However, the KO-DM mice continued to exhibit more ultrastructural abnormalities, suggesting that PGRN deficiency alone may not fully restore renal integrity.
Electron microscopy analysis of mitochondria and podocytes. (A) Electron microscopy analysis of abnormal mitochondrial morphology (loss of electron-dense mitochondrial matrix and cristae) (scale bar: 500 nm). (B) Electron microscopy analysis of foot process effacement (scale bar: 2.0 µm). Comparisons among groups were performed with one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. WT, wild-type; DM, diabetes mellitus; Tofo, tofogliflozin; KO, knockout. aP<0.01 vs. WT-DM; bP<0.05, cP<0.01 vs. WT-Tofo.
In the liver (Fig. 4B), Ccl2 expression was significantly lower in both intervention groups (WT-DM/Tofo and KO-DM) compared with WT-DM mice. Relative to the WT-DM group, Serpine1 expression was significantly lower in WT-DM/Tofo mice, while Fn1 expression was significantly lower in KO-DM mice. These results highlight distinct molecular signatures induced by SGLT2 inhibition and PGRN deficiency, even though both interventions improved hepatic lipid accumulation to varying degrees.
Divergent effects on AMPK–mTORC1 signaling and autophagy
Given previous reports that SGLT2 inhibitors activate autophagy via the AMPK–mTORC1 pathway and that PGRN deficiency impairs the autophagy–lysosome axis, we investigated whether changes in these pathways might underlie the observed tissue-level responses.
In the kidney, WT-DM/Tofo mice exhibited significantly higher phosphorylation of AMPK and significantly lower phosphorylation of S6 compared with WT-DM mice (Fig. 6), consistent with AMPK activation and mTORC1 suppression. These changes were not observed in KO-DM mice. LC3B and p62 levels were significantly higher and lower, respectively, in WT-DM/Tofo mice than in WT-DM mice, indicating upregulation of autophagic flux; no such effect was observed in the KO-DM mice (Fig. 7). In the liver, AMPK activation was observed only in the WT-DM/Tofo group, which showed significantly higher p-AMPK expression than the WT-DM group; p-S6 expression indicated that mTORC1 signaling was unaltered across all groups (Fig. 8). Notably, the p-Akt/Akt ratio, representing an alternative regulator of mTORC1, was unchanged among all groups in both organs, suggesting that the Akt–mTORC1 axis was not a major contributor under these conditions (Figs. 6B, 8B).
Sodium-glucose co-transporter-2 inhibitor- and progranulin knockout-induced autophagy via the mechanistic target of rapamycin complex 1 pathway in the kidneys of each group. Western blot analysis of the protein levels of (A) phosphorylated AMP-activated protein kinase (p-AMPK) and AMPK, (B) p-Akt and Akt, and (C) p-S6 and S6 in the kidneys. Comparisons among groups were performed with one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; WT, wild-type; DM, diabetes mellitus; Tofo, tofogliflozin; KO, knockout. aP<0.05, bP<0.01 vs. WT-DM; cP<0.05 vs. WT-Tofo.
Immunohistochemistry analysis of microtubule-associated protein 1A/1B-light chain 3B (LC3B) and p62 in the kidneys. (A) LC3B staining was primarily localized to the proximal tubules (scale bar: 100 µm). (B) p62 accumulation was observed in glomeruli (scale bar: 50 µm). Comparisons among groups were performed with one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. WT, wild-type; DM, diabetes mellitus; Tofo, tofogliflozin; KO, knockout. aP<0.05, bP<0.01 vs. WT-DM; cP<0.05, dP<0.01 vs. WT-Tofo.
Sodium-glucose co-transporter-2 inhibitor- and progranulin knockout-induced autophagy via the mechanistic target of rapamycin complex 1 pathway in the livers of each group. Western blot analysis of the protein levels of (A) phosphorylated AMP-activated protein kinase (p-AMPK) and AMPK, (B) p-Akt and Akt, and (C) p-S6 and S6 in the liver. Comparisons among groups were performed with one-way analysis of variance followed by Tukey tests. Data are expressed as mean±standard deviation. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; WT, wild-type; DM, diabetes mellitus; Tofo, tofogliflozin; KO, knockout. aP<0.05.
Collectively, these data suggest that SGLT2 inhibition, but not PGRN deficiency, enhances autophagy via AMPK–mTORC1 signaling. This effect is particularly notable in the kidney, which may contribute to the observed differences in tissue protection.
PPAR expression and lipid metabolism markers in the liver
To elucidate the mechanisms underlying the differential effects on hepatic steatosis, we assessed mRNA and protein markers of lipid metabolism. Although the expression of Scd1, Srebf1, and Srebf2 tended to be lower in both WT-DM/Tofo and KO-DM mice compared with WT-DM mice (Fig. 9A), these differences were not statistically significant.
Protein and mRNA levels of hepatic fatty acid and lipid synthesis/oxidation-related markers. (A) mRNA expression levels of hepatic fatty acid and lipid synthesis/oxidation-related markers. (B) Levels of hepatic peroxisome proliferator-activated receptor α (PPARα) and PPARγ evaluated by Western blot analysis. Comparisons among groups were performed with one-way analysis of variance. DM, diabetes mellitus; Tofo, tofogliflozin; WT, wild-type; KO, knockout; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; PGRN, progranulin; Scd1, stearoyl-CoA desaturase 1; Srebf1, sterol regulatory element-binding protein 1; Srebf2, sterol regulatory element-binding protein 2. aP<0.05, bP<0.01 vs. WT-DM; cP<0.05 vs. WT-Tofo.
At the protein level, expression of PPARα—a key regulator of fatty acid oxidation—was significantly higher only in WT-DM/Tofo mice compared with WT-DM mice. Expression of PPARγ, which is commonly associated with adipogenesis and lipid storage, was significantly higher in both WT-DM/Tofo and KO-DM mice than in WT-DM animals (Fig. 9B). Notably, recent evidence suggests that PPARγ may also exert anti-inflammatory effects and enhance insulin sensitivity [26]. In this context, its upregulation could represent a compensatory response that promotes resolution of hepatic steatosis rather than exacerbation.
These findings suggest that Tofo improves hepatic lipid handling primarily through enhanced fatty acid oxidation, while PGRN deficiency may ameliorate steatosis indirectly by reducing inflammation and improving insulin sensitivity.
DISCUSSION
In this study, inflammation- and fibrosis-related markers in the kidneys were improved in both KO-DM and WT-DM/Tofo mice compared with WT-DM mice. However, KO-DM kidneys exhibited more severe structural damage. As glycemic control was comparable between KO-DM and WT-DM/Tofo mice, the autophagy insufficiency observed in KO-DM kidneys may have contributed to the greater renal injury. In the liver, Tofo treatment activated both PPARα and PPARγ, whereas PGRN deficiency selectively activated PPARγ. Thus, improvements in hepatic inflammation and fibrosis in KO-DM and WT-DM/Tofo mice may involve enhanced fatty acid oxidation via PPARα and/or improved insulin sensitivity and anti-inflammatory effects via PPARγ.
The differential responses to PGRN deficiency in the liver and kidney may reflect organ-specific lysosomal specializations [27,28]. In the diabetic kidney, lipid accumulation and impaired autophagy are associated with lysosomal dysfunction and tubular injury [29]. In our study, KO-DM mice displayed reduced tubular vacuolation compared with WT-DM mice, with an even greater reduction observed in WT-DM/Tofo animals [6]. This may be attributable to our previous finding that PGRN deficiency reduces megalin expression and fatty acid uptake in tubular cells, as megalin mediates the reabsorption of albumin-bound fatty acids [30]. Given the key role of renal tubular lysosomes in protein reabsorption and degradation, impaired lysosomal function from PGRN deficiency may disrupt these processes and thus contribute to kidney damage. By contrast, hepatic lysosomes are primarily involved in metabolizing exogenous lipids. While lysosomal dysfunction can cause lipid accumulation, the liver’s diverse lipid-processing capacity and distinct lysosomal enzyme profile may confer resilience or, alternatively, unique susceptibility to such dysfunction [31]. Indeed, PGRN deficiency affects lysosomal lipid metabolism differently across tissues [32], supporting the concept of a liver-specific response. These organ-specific lysosomal roles likely underlie the varied effects of PGRN deficiency. Beyond the liver and kidney, PGRN may influence other tissues with specialized metabolic or degradative functions. Further research is needed to clarify these systemic effects.
Park et al. [33] highlighted the complex and context-dependent roles of PGRN, particularly in metabolic and inflammatory conditions. For example, mice with microglia-specific deletion of the Grn gene exhibit detrimental effects when fed a standard diet, including fasting hyperglycemia and abnormal activation of hypothalamic microglia. In contrast, feeding these mice an HFD results in beneficial outcomes, such as reduced obesity, improved glucose tolerance, and attenuated hypothalamic inflammation. In a previous study, we used a unilateral ureteral obstruction model that induced pronounced renal inflammation and fibrosis, in contrast to the chronic low-grade inflammation observed in our current diabetic model [34]. In the previous investigation, we demonstrated that both PGRN and its proteolytically cleaved products, granulins, are upregulated in the kidney [34]. This upregulation was associated with exacerbated tubulointerstitial inflammation and fibrosis. Collectively, these findings suggest that PGRN exerts dual and tissue-specific effects, functioning as either a pro-inflammatory or an anti-inflammatory mediator depending on the target organ and physiological context. Therefore, a comprehensive understanding of this context-dependent duality is crucial for elucidating the physiological and pathological roles of PGRN, particularly in metabolic and inflammatory diseases.
SGLT2 inhibitors enhance autophagy via mTOR inhibition [17,22], whereas PGRN deficiency impairs the autophagy–lysosome system [35-39]. Consistent with this, autophagy markers indicated that autophagosome formation was reduced in KO-DM kidneys compared with WT-DM/Tofo mice, which likely contributed to worsened histology despite similar glycemic control. However, LC3B expression alone cannot determine whether autophagy is upregulated or blocked, and flux assays using lysosomal inhibitors were not conducted in this study. Nonetheless, even under fasting conditions, which promote autophagy, LC3B expression was not elevated in KO-DM kidneys, suggesting impaired autophagy induction in the absence of PGRN.
The podocyte effacement observed in KO-DM mice may have resulted from impaired recovery mechanisms rather than direct injury due to PGRN deficiency. Previous studies have shown that PGRN protects podocytes from high-glucose injury by enhancing mitophagy and mitochondrial biogenesis, supporting a reparative role [40]. However, future studies using conditional or inducible KO models are needed to clarify this mechanism.
The improvement in hepatocyte vacuolation observed in WT-DM/Tofo mice likely involves mechanisms distinct from those underlying renal autophagy. Hepatic triglyceride levels, but not serum triglycerides, were reduced in WT-DM/Tofo mice, suggesting enhanced hepatic β-oxidation; elevated hepatic p-AMPK/AMPK and PPARα expression support this mechanism. By contrast, KO-DM mice showed no increase in PPARα but exhibited the highest PPARγ expression among all groups. PPARγ activation improves insulin sensitivity and reduces hepatic inflammation, and PGRN deficiency has been associated with increased PPARγ in adipose tissue and macrophages [41,42]. These effects could underlie the improved hepatic steatosis in KO-DM mice, despite unchanged hepatic triglyceride levels. Interestingly, hepatocyte vacuole size was reduced in KO-DM mice, similar to the pemafibrate-induced shifts in lipid droplet morphology reported previously; this may represent a PPARγ-dependent, PPARα-independent mechanism of inflammation amelioration [43].
This study has several limitations. First, the tissue-specific effects of PGRN deficiency remain unclear. Second, although autophagy was enhanced in WT-DM/Tofo kidneys via the AMPK–mTORC1 pathway, assessment of autophagy in the liver was technically limited by inconsistent LC3B and p62 staining, which precluded reliable evaluation of hepatic autophagy status. Third, it is unknown whether PGRN supplementation could reverse the observed phenotypes. Fourth, lysosomal function was not evaluated, despite its potential contribution to renal inflammation in KO-DM mice and its known regulation by PGRN. Finally, urinary albumin, a key marker of renal injury, could not be measured because urine was diluted as a result of SGLT2 inhibition.
In summary, PGRN deficiency exacerbated renal inflammation and fibrosis in diabetic mice, likely due to impaired autophagy; however, it also ameliorated hepatic steatosis through increased PPARγ expression and associated anti-inflammatory effects.
Supplementary Material
Supplemental Table S1.
Taqman Probes Used for Quantitative Polymerase Chain Reaction Analysis
Notes
CONFLICTS OF INTEREST
This research was partially funded by Novo Nordisk Pharma Ltd. (Grant Number 2021-46; awarded to Tomohito Gohda).
ACKNOWLEDGMENTS
This research was partially funded by JSPS KAKENHI (Grant Number JP23K07681) and the 1st Japan Diabetes Foundation.
We thank Kowa Company Ltd., the Laboratory of Molecular and Biochemical Research, the Laboratory of Biomedical Research Resources, the Laboratory of Morphology and Image Analysis, the Research Support Center, and the Juntendo University Graduate School of Medicine for the technical assistance provided.
We also thank Philippa Gunn, D.Phil., from Edanz (https://jp.edanz.com/ac), for editing a draft of this manuscript.
AUTHOR CONTRIBUTIONS
Conception or design: M.M., T.G. Acquisition, analysis, or interpretation of data: H.S., M.M., T.S., T.G. Drafting the work or revising: H.S., M.M., S.H., Y.S., T.G. Final approval of the manuscript: H.S., M.M., S.H., T.S., Y.S., T.G.