Proximal Tubule Autophagy Differs in Type 1 and 2 Diabetes

Significance Statement

Studies suggest that autophagy may be protective in kidney diseases, but understanding how the autophagic process is specifically altered in each disorder is important for applying it therapeutically. On the basis of the observation that autophagy in proximal tubule epithelial cells is mainly regulated by insulin, the authors used diabetic mouse models to investigate whether types 1 and 2 diabetic nephropathy differ in autophagic status. They found distinct patterns of autophagic dysregulation involved in the pathophysiology of types 1 and 2 diabetic nephropathy, with autophagy induction suppressed in the type 2 diabetic kidney (even under starvation) and basal autophagic activity enhanced in the type 1 diabetic kidney (even under fed conditions). They also provide evidence that activated autophagy protects the type 1 diabetic kidney, whereas autophagic suppression jeopardizes the kidney in type 2 diabetes.

Visual Abstract

Abstract

Background

Evidence of a protective role of autophagy in kidney diseases has sparked interest in autophagy as a potential therapeutic strategy. However, understanding how the autophagic process is altered in each disorder is critically important in working toward therapeutic applications.

Methods

Using cultured kidney proximal tubule epithelial cells (PTECs) and diabetic mouse models, we investigated how autophagic activity differs in type 1 versus type 2 diabetic nephropathy. We explored nutrient signals regulating starvation-induced autophagy in PTECs and used autophagy-monitoring mice and PTEC-specific autophagy-deficient knockout mice to examine differences in autophagy status and autophagy’s role in PTECs in streptozotocin (STZ)-treated type 1 and db/db type 2 diabetic nephropathy. We also examined the effects of rapamycin (an inhibitor of mammalian target of rapamycin [mTOR]) on vulnerability to ischemia-reperfusion injury.

Results

Administering insulin or amino acids, but not glucose, suppressed autophagy by activating mTOR signaling. In db/db mice, autophagy induction was suppressed even under starvation; in STZ-treated mice, autophagy was enhanced even under fed conditions but stagnated under starvation due to lysosomal stress. Using knockout mice with diabetes, we found that, in STZ-treated mice, activated autophagy counteracts mitochondrial damage and fibrosis in the kidneys, whereas in db/db mice, autophagic suppression jeopardizes kidney even in the autophagy-competent state. Rapamycin-induced pharmacologic autophagy produced opposite effects on ischemia-reperfusion injury in STZ-treated and db/db mice.

Conclusions

Autophagic activity in PTECs is mainly regulated by insulin. Consequently, autophagic activity differs in types 1 and 2 diabetic nephropathy, which should be considered when developing strategies to treat diabetic nephropathy by modulating autophagy.

Macroautophagy, hereafter autophagy, is a highly conserved degradation system that functions to regulate intracellular homeostasis.^1,2 Cytoplasmic components, such as organelles and proteins, are sequestered into double-membrane vesicles called autophagosomes.³ Then, the autophagosome fuses with the lysosome, resulting in the degradation of the sequestered materials. Essential roles of autophagy in maintaining kidney function have been elucidated using proximal tubular epithelial cell (PTEC)–, podocyte-, or endothelial cell–specific autophagy-deficient mice in both physiologic and pathologic settings.^4–11

Recently, we elucidated that renal autophagic activity is substantially dysregulated in aging and kidney diseases^12,13; for example, lipid overload basically stimulates autophagy for renovation of plasma and organelle membranes, which plays an essential role in maintaining the integrity of proximal tubules.¹³ However, this autophagic activation is inevitably accompanied by lysosomal stress and consequent stagnation of autophagy, contributing to renal lipotoxicity. Previous findings of the protective role of autophagy in various kidney diseases have led to the idea that upregulation of autophagy might represent a plausible therapeutic intervention in these disorders.^14–16 However, under lipid overload, simple autophagy “activation” may increase lysosomal stress, resulting in more downstream suppression of autophagy. Instead, pharmacologic correction that restores autophagic flux represents a novel therapeutic option. Thus, understanding the steps in the autophagic process that are altered in each disorder is critically important, and therapies that are tailored to autophagic status must be developed.¹⁷

Previous studies have demonstrated an overall suppression of autophagy in PTECs irrespective of type 1 and type 2 diabetic nephropathy, although most of these findings are on the basis of the reduction in expression of autophagy-related molecules and the number of autophagosomes and/or an accumulation of the autophagy substrate SQSTM1/p62.^18–20 This autophagic suppression during diabetic nephropathy has been attributed to enhanced activity of the mammalian target of rapamycin (mTOR) pathway, a potent suppressor of autophagy induction.²¹ However, the two types of nephropathy differ in plasma levels of insulin, which may influence autophagic activity in a tissue-dependent manner.²² Therefore, autophagy status in each disorder must be assessed more precisely.

A growing body of evidence suggests a primary role for PTECs in provoking nephron injury and kidney dysfunction during diabetes, which is represented by the term “diabetic tubulopathy.”^23,24 Because of their position and major reabsorptive role within the nephron, PTECs are exposed to various toxic factors in the glomerular filtrate, such as albumin, high glucose, and advanced glycation end products (AGEs), which are associated with tubular hypertrophy and progressive interstitial inflammation and fibrosis. Thus, in this study, we highlighted the autophagy status of proximal tubules during type 1 and type 2 diabetes. We hypothesized that the different patterns of autophagy status in each disorder, especially its dysregulation, may influence the progression of diabetic nephropathy, because autophagy protects PTECs via anti-inflammatory, antioxidant, and antifibrotic effects.²⁵

Accordingly, we explored herein the determinants of autophagy in PTECs as well as differences in the status and potential roles of autophagy in PTECs during types 1 and 2 nephropathy. We also investigated the effect of rapamycin treatment against types 1 and 2 nephropathy with or without ischemia-reperfusion (I/R) injury, in which enhanced autophagy is required to combat cellular stress.

Methods

Animals

Green fluorescent protein–microtubule-associated protein 1 light chain 3 (GFP-MAP1LC3) transgenic mice, Atg5^F/F;N-myc downstream-regulated gene 1 (NDRG) mice, and Atg5^F/F;kidney androgen–regulated protein (KAP) mice, all with a C57BL/6 background, have been described previously.^4,26 B6.BKS(D)-Lepr^db/J (db/+ mice) with a C57BL/6 background were purchased from the Jackson Laboratory (Bar Harbor, ME). To generate GFP-MAP1LC3 transgenic db/db mice, GFP-MAP1LC3 transgenic mice were crossed with db/+ mice, and the resultant GFP-MAP1LC3 db/+ mice were crossed with db/+ mice. To induce type 1 diabetes, mice were treated with streptozotocin (STZ; 50 mg/kg; Sigma-Aldrich, St. Louis, MO) intraperitoneally for 4 consecutive days at 8 weeks of age. Akita mice (Ins2^Akita) were purchased from Japan SLC (Shizuoka, Japan). The induction of autophagy deficiency in Atg5^F/F;NDRG mice and the assessment of the autophagic flux in vivo were previously described.¹² Only male mice were used in this study. Mice were housed in box cages, maintained on a 12-h light/12-h dark cycle, and fed a chow diet (Oriental Yeast, Osaka, Japan) ad libitum. All animal experiments were approved by the institutional committees of the Animal Research Committee of Osaka University and the Japanese Animal Protection and Management Law (no. 25).

Insulin, Amino Acid, or Glucose Administration to Starved Mice

To compare inhibitory effects on autophagic activity in PTECs, liver, and skeletal muscle, nondiabetic GFP-MAP1LC3 transgenic mice were refed or treated with insulin, amino acids, or glucose after 24 hours of starvation. Insulin (Humalin R; Lilly, Indianapolis, IN) was given at 400 mU/kg intraperitoneally every 10 minutes for 1 hour. The plasma glucose concentration was monitored every 5 minutes and maintained at 70–100 mg/dl by administration of 20% d-glucose (Wako, Osaka, Japan). The amino acid mixture (Amiparen; Otsuka, Tokyo, Japan) was given at 1092 mg/kg intraperitoneally every 30 minutes for 1 hour. Somatostatin (Sigma-Aldrich) was injected simultaneously (428.4 μg/kg). d-glucose was injected once at a dose of 2.5 g/kg intraperitoneally. To inhibit the mammalian target of rapamycin complex 1 (mTORC1) pathway, rapamycin (LC Laboratories, Woburn, MA) was injected twice at a dose of 0.4 mg/kg at 2 and 16 hours before sampling. Mice were euthanized, and samples were collected 1 hour after treatment.

Kidney I/R Injury

Twelve-week-old control, db/db, or STZ-treated mice were subjected to I/R injury. STZ was administered at the age of 8 weeks old. Left renal pedicles were exposed and clamped for 35 minutes or sham operated. Body temperature was monitored with a rectal probe and controlled in the range of 36.5°C–37°C. Rapamycin (0.4 mg/kg) or vehicle (0.1 ml DMSO (Sigma-Aldrich)/ethanol [1:1]) was administered intraperitoneally daily for 3 days before I/R injury. Mice were euthanized at 48 hours after I/R injury.

Antibodies

We used the following antibodies: antibodies for LRP2/MEGALIN (a gift from T. Michigami, Department of Bone and Mineral Research, Osaka Medical Center and Research Institute for Maternal and Child Health, Osaka, Japan), microtubule-associated protein 1 light chain 3B (MAP1LC3B; for Western blotting; Cell Signaling Technology, Beverly, MA; for immunostaining; Medical and Biologic Laboratory, Nagoya, Japan), phosphorylated Akt (Thr308; Cell Signaling Technology), LAMP1 (BD Biosciences, Oxford, United Kingdom), collagen I (Abcam, Cambridge, MA), SQSTM1/p62 (Medical and Biologic Laboratory), ubiquitin (Cell Signaling Technology), mTOR (Cell Signaling Technology), phospho-mTOR (Cell Signaling Technology), phospho-p70 S6 Kinase (Thr389; Cell Signaling Technology), phosphoribosomal protein S6 (p-RPS6; Ser235/236; Cell Signaling Technology), MUC1 (Bioss, Boston, MA), cathepsin D (Santa Cruz Biotechnology, Dallas, TX), ACTB (Sigma-Aldrich), biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA), horseradish peroxidase–conjugated secondary antibodies (DAKO, Glostrup, Denmark), and Alexa-conjugated secondary antibody (Invitrogen, Carlsbad, CA).

Histologic Analyses

Counting of the number of green fluorescent protein (GFP)–positive puncta on postfixed frozen kidney; immunohistochemical staining for SQSTM1/p62, ubiquitin, collagen I, p-RPS6, or LAMP1 on paraffin-embedded sections; double staining for SQSTM1/p62 (ubiquitin or p-RPS6) and LRP2/MEGALIN; quantification of the percentage of the immune-positive area for collagen I; and electron microscopy analysis were performed as previously described.⁷ Apoptotic cells were detected by the terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxy uridine nick-end labeling assay using an in situ apoptosis detection kit (Takara, Shiga, Japan) according to the manufacturer’s instructions. In all quantitative or semiquantitative analyses of histologic staining, at least ten high-power fields were reviewed for each tissue by two nephrologists (T.Y. and A.T.) in a blinded manner.

Cytochrome c Oxidase and Succinate Dehydrogenase Staining

Cytochrome c oxidase (COX)/succinate dehydrogenase (SDH) staining was performed as previously described.¹²

Biochemical Measurements

Plasma concentrations of glucose, insulin, BUN, and creatinine were measured using the Glucose CII-test (Wako), ELISA kit (Shibayagi, Gunma, Japan), BUN-Test-Wako (Wako), and CRE-EN Kainos test (Kainos, Tokyo, Japan), respectively. Urinary albumin concentrations were determined using the MICROFLUORAL Microalbumin test (Progen, Heidelberg, Germany). All kits were used in accordance with the manufacturers’ protocols.

Cell Culture

PTEC lines were previously described.⁴ HK2 cells (human renal PTECs) were obtained from American Type Culture Collection (Manassas, VA). To compare the inhibitory effects of glucose, insulin, or amino acid treatment in cultured PTECs on autophagy, cells were starved in HBSS for 24 hours followed by incubation for 1 hour with (1) low-glucose DMEM (100 mg/dl glucose; Gibco, Grand Island, NY) containing 0.1% FBS, (2) low-glucose DMEM containing 10% FBS, (3) low-glucose DMEM containing 0.1% FBS and insulin (0.25, 2.5, or 25 μg/ml; Sigma-Aldrich), (4) low-glucose DMEM containing 0.1% FBS and amino acids (1.5 or 3 times higher than original concentration), or (5) high-glucose DMEM (450 mg/dl glucose) containing 0.1% FBS. For amino acid administration, l-glutamine solution (Sigma-Aldrich) and MEM amino acids solution (50×; Sigma-Aldrich) were added to DMEM to be 1.5- or 3-fold higher than standard concentrations. To assess autophagic flux, PTECs incubated with low- or high-glucose DMEM for up to 24 hours were treated with 200 nM bafilomycin A1 (BafA1; Wako) for 1 hour before harvest. To inhibit the mTORC1 pathway, PTECs incubated with low-glucose DMEM supplemented with 0, 2.5, or 5 μg/ml of insulin for 30 minutes were treated with 0.4 μg/ml rapamycin 30 minutes before harvest.

Statistical Analyses

All results are presented as means±SEM. Statistical analyses were conducted using JMP software (SAS Institute, Cary, NC). The difference between two experimental values was assessed by a t test. Multiple group comparisons were performed using one-way ANOVA followed by the Dunnett test to detect intergroup differences. Statistical significance was defined as P<0.05.

Results

Insulin-mTORC1 Pathway Regulates Starvation-Induced Autophagy in PTECs

Although insulin and amino acids are known to strongly suppress autophagy,¹ a recent paper reported that autophagy is differentially regulated by insulin and amino acids in a tissue-dependent manner.²² This led us to investigate which nutrient signals regulate starvation-induced autophagy in PTECs. We compared the inhibitory effects of refeeding, insulin, amino acids, or glucose on autophagic activity in PTECs of 24-hour starved nondiabetic GFP-MAP1LC3 transgenic mice, in which GFP-positive puncta represent autophagosomes (Figure 1A). Refeeding induced a marked elevation in plasma glucose and insulin levels (Supplemental Table 1). Differential contributions of insulin and amino acids to the mTORC1-autophagy pathway in the liver and muscle were demonstrated by a decrease in the number of GFP-positive puncta (Supplemental Figure 1, A and B). In the kidney, both insulin and amino acids, but not glucose administration, significantly decreased the number of GFP-positive puncta in PTECs (Figure 1B).

Figure 1.

Starvation-induced autophagy is suppressed by insulin in proximal tubular epithelial cells (PTECs). (A) Experimental protocol (see Methods). Green fluorescent protein (GFP)–microtubule-associated protein 1 light chain 3 (MAP1LC3) transgenic mice were starved for 24 hours followed by starvation for 1 hour (“Starved”); refeeding (“Refed”); or treatment with insulin (“Insulin”), amino acids (“Amino acid”), or glucose (“Glucose”; n=3–5 in each group). ip, intraperitoneal. (B) Confocal microscopy images of kidney sections after immunostaining for LRP2/MEGALIN (red). The number of GFP–positive puncta was counted in the proximal tubules (right panel). (C) Relationship between autophagosome formation and mammalian target of rapamycin (mTOR) signaling was assessed by immunostaining for p-RPS6 (Ser235/236), a marker for mTOR complex 1 activation (red), in the kidney sections. Relative intensity of the p-RPS6–positive area was quantified by densitometry (n=16; R²=0.43; P<0.001 as determined by linear regression analysis). (D) The effects of rapamycin treatment (0.4 mg/kg) 2 and 16 hours before sampling in addition to each treatment on GFP-positive puncta formation were assessed in kidney sections after immunostaining for LRP2/MEGALIN. The number of GFP-positive puncta was counted in the proximal tubules (right panel). DAPI, 4′,6-diamidino-2-phenylindole. (B–D) Representative images are presented. Statistically significant differences are indicated. *P<0.05. Scale bars, 20 μm in B–D. (E) Autophagosome formation (assessed by MAP1LC3-II) and activation of the mTOR and Akt pathways (assessed by p-RPS6 or p-Akt, respectively) were investigated by Western blot analysis using cultured PTECs. PTECs were starved with HBSS for 1 hour followed by treatment with either 0.1% or 10% FBS or 0.1% FBS supplemented with insulin, amino acids, or glucose at the indicated concentrations for 1 hour. Actin, beta was used as a loading control. Densitometric analysis was performed (right panel). The mean value of starved controls (0.1% FBS) at the baseline was adjusted to one as a reference. (B, D, and E) Data are presented as means±SEM. Statistically significant differences are indicated. *P<0.05 versus starved controls (0.1% FBS).

Notably, the number of GFP-positive puncta was inversely correlated with both mTOR activity (assessed by the intensity of p-RPS6 [Ser235/236] staining) and plasma insulin levels (Figure 1C, Supplemental Figure 1C, left panel). Conversely, GFP-positive puncta were positively correlated with plasma glucose levels (Supplemental Figure 1C, right panel). To clarify the importance of mTORC1 activity in autophagic regulation in PTECs in vivo, we administered the mTORC1 inhibitor rapamycin to refed or insulin-treated mice. Pretreatment with rapamycin abolished both the increase in p-RPS6 staining (data not shown) and the reduction in the number of GFP-positive puncta induced by refeeding or insulin (Figure 1D).

To confirm that insulin directly suppresses autophagic activity via mTOR signaling, we investigated the effect of insulin, amino acids, and glucose on the mTORC1-autophagy pathway in cultured PTECs. To induce autophagy, PTECs were subjected to starvation by incubating with HBSS for 1 hour and then, treated with either 0.1% or 10% FBS or 0.1% FBS supplemented with insulin, amino acids, or glucose for 1 hour. Western blot analysis demonstrated that the amount of MAP1LC3-II was increased under the starvation condition, whereas it was significantly decreased by insulin treatment, especially at moderate concentrations, and the amount of p-RPS6 was also significantly increased (Figure 1E). Rapamycin treatment suppressed the insulin-induced reduction of MAP1LC3-II levels, further confirming the involvement of mTOR signaling in insulin-induced suppression of autophagy (Supplemental Figure 1D). Amino acid treatment was less effective in reducing the amount of MAP1LC3-II compared with insulin treatment (Figure 1E). However, glucose treatment did not decrease the amount of MAP1LC3-II (Figure 1E). These data were largely recapitulated in HK2 cells (Supplemental Figure 2). Collectively, these results indicate that insulin, but not glucose itself, regulates autophagic activity via the mTORC1 pathway in PTECs.

High Glucose Enhances Autophagic Activity In Vitro

The observation that glucose did not suppress starvation-induced autophagy led us to investigate precisely how high glucose affects autophagic activity in PTECs. We cultured PTECs in medium containing high (450 mg/dl) or low (100 mg/dl) glucose and assessed autophagic flux by measuring autophagic flux index (defined as the proportion of MAP1LC3-II levels in the presence and absence of BafA1). PTECs treated with high glucose showed higher autophagic flux than those with low glucose at early stages (6 and 12 hours), with seemingly, suppression of autophagy at a later stage (24 hours) (Figure 2A). Higher autophagic flux induced by high glucose (at 6 hours) was confirmed in cultured PTECs stably expressing GFP-MAP1LC3 by comparing the number of GFP-positive puncta in the presence and absence of BafA1 (Supplemental Figure 3A). Next, we examined whether high glucose–induced autophagy could be suppressed by insulin. The combination of high glucose and insulin decreased the conversion from MAP1LC3-I to MAP1LC3-II (assessed by Western blot analysis) and the number of autophagosomes (assessed by electron microscopy) compared with high glucose alone (Figure 2B, Supplemental Figure 3, B and C). The effects of high glucose with or without insulin treatment on autophagic activity were recapitulated in HK2 cells (Supplemental Figure 4). These results indicate that high glucose itself does stimulate autophagic activity, whereas insulin suppresses high glucose–mediated autophagy induction.

Figure 2.

High glucose (HG) stimulates autophagic activity with downstream suppression of autophagic flux in vitro. Autophagic flux in proximal tubular epithelial cells (PTECs) during normal and HG treatment was estimated by the conversion from microtubule-associated protein 1 light chain 3-I (MAP1LC3-I) to MAP1LC3-II as a readout of autophagosome formation (n=4–5, respectively). (A) Autophagic flux index (defined as the proportion of the levels of MAP1LC3-II in the presence of bafilomycin A1 (BafA1) to the levels of MAP1LC3-II in the absence of BafA1) is calculated at the indicated times. (B) Effect of insulin on autophagic flux in PTECs during HG exposure was assessed. Representative immunoblots are shown. Data are presented as means±SEM. Statistically significant differences are indicated. ACTB, actin, beta; LG, low glucose. *P<0.05

Autophagy Induction Is Suppressed in PTECs of Type 2 Diabetic Mice

From the above observation, we deduced that the autophagic status in PTECs is quite different between type 1 and type 2 diabetic nephropathy. To measure autophagic flux in the kidneys of type 2 diabetic mice, GFP-MAP1LC3 transgenic db/+ or db/db mice under fed or starvation conditions were administered chloroquine 6 hours before euthanasia.¹² A greater gain in body weight and higher plasma glucose and insulin levels were observed in db/db mice compared with nondiabetic db/+ mice (Supplemental Table 2). GFP-positive puncta were rarely observed in LRP2/MEGALIN-positive PTECs of both db/+ and db/db mice under the fed condition, regardless of chloroquine administration (Figure 3A). In contrast, chloroquine administration significantly increased the number of GFP-positive puncta in 24-hour starved db/+ mice but not in 24-hour starved db/db mice, suggesting that autophagic activity is suppressed even under starvation in PTECs of db/db mice (Figure 3A).

Figure 3.

Autophagy induction is suppressed in proximal tubular epithelial cells of type 2 diabetic db/db mice. (A) Autophagic flux was assessed by comparing the number of green fluorescent protein (GFP)–positive puncta with or without chloroquine administration in the proximal tubules of db/+ or db/db GFP–microtubule-associated protein 1 light chain 3 (MAP1LC3) transgenic mice at 12 weeks of age that were fed or subjected to 24 hours of starvation (n=3–6 in each group). Kidney sections were immunostained for LRP2/MEGALIN (red) and counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). The number of GFP-positive puncta per proximal tubule under each condition was counted in at least ten high-power fields (×600). (B) The relationship between autophagosome formation and mammalian target of rapamycin signaling was assessed. Kidney sections were immunostained for p-RPS6 (Ser235/236; red) and counterstained with DAPI (blue). Images are representative of multiple experiments. Data are presented as means±SEM. Statistically significant differences are indicated. Scale bars, 20 μm. *P<0.05.

Positive staining for p-RPS6 was observed in PTECs of fed mice; it was more pronounced in db/db mice and is in agreement with recent papers showing hyperactivation of mTORC1signaling in PTECs of high-fat diet–induced obese mice (Figure 3B).^13,27 Twenty-four hours of starvation, significantly reduced mTORC1 signaling in PTECs of db/+ mice, but not of db/db mice, was observed. Notably, PTECs positive for p-RPS6 staining exhibited suppressed autophagic activity as assessed by the reduced number of GFP puncta (Figure 3B). Collectively, these results demonstrate that starvation-induced autophagic upregulation is impaired in db/db mice due to mTORC1 hyperactivation.

Autophagic Flux Is Enhanced in PTECs of Type 1 Diabetic Mice but Stagnates under Starvation

Next, to assess alterations in autophagic activity in type 1 diabetic mice, STZ-treated GFP-MAP1LC3 transgenic mice were analyzed. STZ-treated mice exhibited significantly lower body weight and higher plasma glucose without an increase in plasma insulin compared with control mice (Supplemental Table 2). In contrast to the above findings in db/db mice, a number of GFP-positive puncta were observed in PTECs of STZ-treated mice even under the fed condition, which was significantly increased after chloroquine administration, indicating high basal autophagic activity in STZ-treated mice (Figure 4A). Staining for p-RPS6 was largely negative in PTECs of STZ-treated mice, even under the fed condition (Figure 4B). Enhanced autophagic activity was also observed in Akita mice, another type 1 diabetic model (Supplemental Figure 5A). Furthermore, we found that insulin treatment suppressed this autophagic activation in Akita mice and that STZ treatment restored autophagic activity in db/db mice (Supplemental Figure 5), confirming that insulin regulates autophagic activity in PTECs. Starvation further increased the number of GFP-positive puncta in PTECs of STZ-treated mice (Figure 4A); however, chloroquine administration did not increase the number of GFP-positive puncta, suggesting that autophagic activity cannot be boosted under starvation. Collectively, basal autophagic activity is high, but under the starvation condition, it stagnates in PTECs of type 1 diabetic mice.

Figure 4.

Starvation-induced autophagic flux is stagnated in proximal tubular epithelial cells of streptozotocin (STZ)-treated mice. Vehicle- or STZ-treated green fluorescent protein (GFP)–microtubule-associated protein 1 light chain 3 (MAP1LC3) transgenic mice at 12 weeks of age were fed or subjected to 24 hours of starvation (n=3–6 in each group). (A) Autophagic flux was assessed by comparing the number of GFP-positive puncta with or without chloroquine administration in the proximal tubules. Kidney sections were immunostained for LRP2/MEGALIN (red) and counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). The number of GFP-positive puncta per proximal tubule under each condition was counted in at least ten high-power fields (×600). (B) The relationship between autophagosome formation and mammalian target of rapamycin signaling was assessed by immunostaining kidney sections for p-RPS6 (Ser235/236; red) and counterstaining with DAPI (blue). Images are representative of multiple experiments. Data are provided as means±SEM. Statistically significant differences are indicated. Scale bars, 20 μm. *P<0.05.

Prolonged High-Glucose Exposure without Insulin Leads to Lysosomal Stress in PTECs

Next, we tried to clarify how autophagy is suppressed after prolonged high-glucose exposure in PTECs (Figure 2A) and in STZ-treated mice (Figure 4A). We focused on the function and morphology of lysosomes, because we previously observed that high-fat diet–induced lysosomal dysfunction leads to suppression of autophagy.¹³ We first assessed lysosomal activity in PTECs treated with low glucose or high glucose by measuring the proportion of the levels of cathepsin D heavy chain (component of mature form) to procathepsin D.²⁸ The lysosomal activity was higher in high glucose than in low glucose at an early stage (6 hours), but it decreased at a later stage (24 hours), which was not accompanied by activation of mTOR pathway (Supplemental Figure 6A). Immunostaining for LAMP1 revealed only a mild increase in the number of lysosomes in PTECs of db/db mice, whereas intense staining was observed in STZ-treated mice (Supplemental Figure 6, B and C). Notably, we found an increase in the number of enlarged lysosomes in starved STZ-treated mice (Supplemental Figure 6B). Electron microscopy revealed an increased number of lysosomes containing onion-like structures (Supplemental Figure 6D), which are similar to the lysosomes of high-fat diet–fed mice.¹³ Consistently, lysosomal activity was decreased in the kidney of STZ-treated mice (Supplemental Figure 6E). Collectively, prolonged high-glucose exposure without insulin leads to lysosomal stress in PTECs, which may suppress autophagy.

Differences in Basal Autophagic Substrate Degradation between Type 1 and Type 2 Diabetic Kidneys

Next, differences in the amount of substrates degraded by basal autophagy between type 1 and type 2 diabetic kidneys were evaluated in diabetic tamoxifen-inducible PTEC-specific autophagy-deficient mice (Atg5^F/F;NDRG1).^12,29 To produce mouse models of type 1 and type 2 diabetes, Atg5^F/F;NDRG1 and Atg5^F/F control mice were treated with STZ at 8 weeks of age or crossed with db/db mice, respectively. Three weeks before euthanasia, nondiabetic control, STZ-treated, or db/db mice at 12 weeks of age received tamoxifen to induce genetic ablation of Atg5. The accumulation of SQSTM1/p62- or ubiquitin-positive aggregates represents the amount of substrate requiring degradation over the 3-week period, because SQSTM1/p62 is a ubiquitin and MAP1LC3 binding protein, and it is removed by autophagy, thus serving as an index of autophagic degradation.^30,31 Autophagy deficiency induced by tamoxifen treatment was confirmed (Supplemental Figure 7). PTECs of STZ-treated Atg5^F/F;NDRG1 mice administered tamoxifen exhibited massive cytosolic swelling compared with those of nondiabetic Atg5^F/F;NDRG1 mice (Figure 5A). In contrast, autophagy deficiency had little effect on PTEC morphology in db/db mice (Figure 5A). Moreover, 3-week ablation of autophagy triggered a significant increase in SQSTM1/p62- and ubiquitin-positive protein aggregates in tamoxifen-treated nondiabetic Atg5^F/F;NDRG1 mice, which was more pronounced in STZ-treated Atg5^F/F;NDRG1 mice but not in db/db Atg5^F/F;NDRG1 mice (Figure 5, B and C). These results indicate that STZ-treated mice use autophagy for the degradation of increasing substrates, whereas autophagy is intrinsically suppressed in db/db mice.

Figure 5.

Streptozotocin (STZ)-treated mice use autophagy for the degradation of increasing substrates, whereas autophagy is intrinsically suppressed in db/db mice. (A–C) Representative images of (A) periodic acid–Schiff-stained or (B) immunostained for SQSTM1 and (C) ubiquitin in kidney cortical regions of STZ-treated mice (left panel), db/db mice (right panel), or respective control (Atg5^F/F and Atg5^F/F;N-myc downstream-regulated gene 1 [NDRG1]) mice at 12 weeks of age, which were treated with vehicle or tamoxifen 3 weeks before euthanasia (n=4 or 5 in each group). Scale bar, 50 μm. Magnified images are presented in bottom panels. (D) Cell size of proximal tubules and the number of SQSTM1/p62- or ubiquitin-positive puncta were assessed in at least ten high-power fields (×400). Data are provided as means±SEM. Statistically significant differences are indicated. F/F, Atg5^F/F mice; F/F;NDRG1, Atg5^F/F;NDRG1 mice. *P<0.05.

Differences in the Consequences of Genetic Ablation of Autophagy in Type 1 and Type 2 Diabetic Kidneys

We examined the differences in consequences of genetic ablation of autophagy in PTECs of type 1 and type 2 diabetic kidneys by monitoring PTEC-specific Atg5-deficient mice. We observed STZ-treated Atg5^F/F;KAP, db/db Atg5^F/F;KAP, and respective control mice for up to 10 months. Physiologic and biochemical parameters of these mice are shown (Supplemental Table 3). Plasma glucose levels were significantly increased in both STZ-treated and db/db mice. Plasma insulin levels were decreased in STZ-treated mice, whereas insulin levels were increased in db/db mice, independent of autophagy competency. Intense p-RPS6 staining was observed in db/db mice but not in STZ-treated mice, suggesting insulin-mediated mTOR activation (Supplemental Figure 8A). Intense staining for LAMP1 observed in STZ-treated mice was diminished under autophagy deficiency, indicative of reduced overload of the autophagy-lysosome pathway, whereas we observed no change in LAMP-positive lysosomes between db/db Atg5^F/F and db/db Atg5^F/F;KAP mice (Supplemental Figure 8B). Urinary albumin excretion was increased in STZ-treated Atg5^F/F;KAP mice compared with STZ- or vehicle-treated Atg5^F/F or vehicle-treated Atg5^F/F;KAP mice, whereas it was comparably increased in db/db Atg5^F/F and db/db Atg5^F/F;KAP mice (Figure 6A).

Figure 6.

Activation of autophagy guards against mitochondrial damage and fibrosis in streptozotocin (STZ)-treated mice, whereas these protective effects are jeopardized in db/db mice. Streptozotocin (STZ)-treated, db/db, or respective control nondiabetic (Atg5^F/F and Atg5^F/F;kidney androgen–regulated protein [KAP]) mice were monitored up to 10 months. (A) Age-dependent changes in urinary albumin levels in STZ-treated (left panel) and db/db (right panel) mice (Atg5^F/F;KAP and Atg5^F/F; n=6–8 in each group). (B) Representative images of periodic acid–Schiff-stained kidney sections (left panel) and quantitation of proximal tubular size (right panel). (B and C) Representative images of cytochrome c oxidase (COX; B, upper panel), succinate dehydrogenase (SDH; B, lower panel), and collagen I (COL1; C) staining and kidney cortical regions (n=5 in each group). Data are presented as means±SEM. (C) The mean value of nondiabetic Atg5^F/F control mice was adjusted to one as a reference. Statistically significant differences are indicated. F/F, Atg5^F/F mice; F/F;KAP, Atg5^F/F;KAP mice. Scale bars, 500 μm. *P<0.05.

Electron microscopy analysis revealed extensive aggregation of fragmented and swollen mitochondria with derangement of the cristae in STZ-treated Atg5^F/F;KAP mice compared with STZ-treated Atg5^F/F control mice, whereas mitochondrial morphology abnormalities were comparably observed in db/db Atg5^F/F and db/db Atg5^F/F;KAP mice (Supplemental Figure 8C). COX and SDH staining was performed to evaluate alterations in mitochondrial respiration activity caused by autophagy deficiency in STZ-treated or db/db mice. Intensity of tubular COX and SDH staining was somewhat decreased in STZ-treated Atg5^F/F mice compared with vehicle-treated Atg5^F/F mice; however, a more prominent decline in the intensity was observed in STZ-treated Atg5^F/F;KAP mice. In contrast, db/db Atg5^F/F;KAP and db/db Atg5^F/F mice comparably exhibited a significant decrease in the intensity of tubular COX and SDH staining compared with nondiabetic mice (Figure 6B). Immunostaining for collagen I and Masson trichrome staining demonstrated that fibrosis was significantly increased in STZ-treated Atg5^F/F;KAP mice compared with STZ-treated Atg5^F/F mice, whereas comparable fibrosis was observed in db/db Atg5^F/F and db/db Atg5^F/F;KAP mice (Figure 6C, Supplemental Figure 8D). Collectively, these data suggest that activation of autophagy guards against mitochondrial damage and fibrosis in the kidneys of STZ-treated diabetic mice, whereas these protective effects of autophagy are intrinsically jeopardized in db/db mice.

Rapamycin Exerts Opposite Effects on I/R Injury in STZ-Treated and db/db Mice

Lastly, we subjected diabetic mice to kidney I/R injury to determine whether autophagy status in types 1 and 2 diabetic kidneys leads to differences in vulnerability to I/R injury, because autophagy has a protective role against I/R injury. After I/R injury, severely injured tubules with massive tubular sediments and vacuolation were comparably observed in nondiabetic or STZ-treated mice, and they were significantly pronounced in db/db mice (Figure 7A). The number of apoptotic, terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxy uridine nick-end labeling–positive tubular cells paralleled the tubular injury (Figure 7B). We treated mice with rapamycin to investigate the effect of enhanced autophagy induction on kidney damage after I/R injury in types 1 and 2 diabetic nephropathy. In sham-operated mice, periodic acid–Schiff staining demonstrated that rapamycin treatment has no effect among nondiabetic and STZ-treated mice and db/db mice (Figure 7A). Daily rapamycin treatment for 3 days before I/R injury significantly ameliorated kidney injury in db/db mice, whereas it significantly exacerbated kidney injury in STZ-treated mice (Figure 7). These data suggest that rapamycin exerts opposite effects on I/R injury in STZ-treated and db/db mice.

Figure 7.

Rapamycin-induced pharmacological autophagy produces opposite effects on ischemia-reperfusion (I/R) injury in streptozotocin (STZ)-treated and db/db mice. Twelve-week-old mice (nondiabetic control, streptozotocin [STZ]-treated, or db/db mice) were subjected to unilateral I/R injury. Rapamycin or vehicle was given intraperitoneally daily for 3 days before I/R injury. Representative images of (A) periodic acid–Schiff and (B) terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxy uridine nick-end labeling (TUNEL) staining on the kidney cortexes 48 hours after I/R injury or sham operation. (A, lower panel) The tubular injury score is shown. (B, lower panel) The number of TUNEL-positive proximal tubular epithelial cells was calculated in at least ten high-power fields. Data are presented as means±SEM. Statistically significant differences are indicated. Original magnification, ×400. Scale bars, 500 μm. *P<0.05; ^#P<0.05 versus sham-operated mice in each group.

Discussion

In this study, we demonstrated that autophagy status in kidney proximal tubules is quite different between type 1 and type 2 diabetic nephropathy, which affects the fate of PTECs and disease progression. The difference seems to apply also to podocytes but does not seem to apply to distal tubules (Supplemental Figure 9). A simple schematic drawing is shown in Figure 8.

Figure 8.

Distinct patterns of autophagic dysregulation are involved in the pathophysiology of type 1 and type 2 diabetic nephropathies. Insulin mainly regulates autophagic activity in kidney proximal tubular epithelial cells by activating mammalian target of rapamycin (mTOR) signaling. Therefore, autophagic activity is suppressed in type 2 diabetic nephropathy by mTOR signaling due to high plasma insulin level, whereas high basal autophagic activity is observed in type 1 diabetic nephropathy and leads to lysosomal stress. Rapamycin treatment for ischemia-reperfusion (I/R) injury in type 1 diabetic mice leads to stagnation of autophagic activity and exaggerated I/R injury, although it resorts autophagic activity in type 2 diabetic mice and attenuates I/R injury. An mTOR-independent pathway may be involved in the regulation of autophagy in type 1 and type 2 diabetic nephropathies (see Discussion). DM, diabetes mellitus.

Nutrition signals are known to contribute differently to the mTOR-autophagy pathway in a tissue-dependent manner; mTORC1 and autophagy are mainly regulated by amino acids in the liver, whereas insulin regulates mTORC1 and autophagy in skeletal muscle.²² To identify the determinants of autophagic flux in the kidney, we investigated the responses of PTECs to insulin, amino acids, and glucose, and we determined that insulin and less potently, amino acids but not glucose significantly suppressed autophagic activity in PTECs. Moreover, we found that insulin regulates autophagy mainly via the mTOR pathway, at least in healthy PTECs.

Previous studies have reported that autophagy is suppressed overall, irrespective of type 1 and type 2 diabetic nephropathy.^18–20 However, the above in vitro findings indicate that insulin is a major determining regulator of autophagy in PTECs; thus, we deduced that autophagy status is quite different between types 1 and 2 diabetic nephropathy depending on the presence or absence of insulin. Although it is possible that many confounding factors other than insulin, such as metabolites, growth factors, lipid status, and hormones, affect autophagic activity, the in vivo flux assay using chloroquine suggests that autophagy is substantially activated in PTECs of STZ-treated mice even under the fed condition, whereas the kidney cannot boost autophagic activity against starvation. In addition to nutritional cues, such as starvation, autophagy can be evoked by various cellular stresses,³² such as reactive oxygen species, endoplasmic reticulum stress, and hypoxia, all of which are involved in the pathogenesis of diabetic nephropathy^33–35 and may induce autophagy in type 1 diabetic kidney. We and another group recently reported that autophagy induced by AGEs upregulates the biogenesis and function of lysosomes in PTECs, contributing to the degradation of AGEs,³⁶ and that AGE-induced lysosomal membrane permeabilization and lysosomal dysfunction induce stagnation in autophagic activation in PTECs.³⁷ Consistently, we found that prolonged high-glucose exposure to PTECs leads to decreased lysosomal activity (as assessed by cathepsin D maturation) and deformed lysosomes. Thus, it is plausible that high glucose–induced constitutive stress exhausts lysosomes in type 1 diabetic nephropathy, which may lead to a reduced capacity for additional autophagy stimulation.

In contrast, we demonstrated that autophagy induction is significantly suppressed in PTECs of type 2 diabetic mice, which is in good agreement with a recent report.³⁸ Consistently, genetic ablation of Atg5 has little effect on the pathophysiology of type 2 diabetic kidney, because autophagy is already suppressed. Insulin-mediated mTOR activation may contribute to the suppression of autophagy; however, given that high glucose can induce autophagy via an mTOR-independent pathway, it is plausible to think that autophagy induction can also be suppressed independent of mTOR pathway in type 2 diabetic nephropathy. Indeed, it is reported that activation of Akt pathway (downstream of insulin signaling) suppresses autophagy not only in an mTOR-dependent manner but also, in an mTOR-independent manner by inhibiting forkhead box O 3a,³⁹ a transcription factor previously shown to positively regulate autophagy.⁴⁰ Moreover, cytokines and growth factors, such as IGF1, many of which are enriched in type 2 diabetic mice, inhibit type 3 phosphatidylinositol 3-kinase, leading to an inhibition of autophagy independent of the mTOR pathway.³⁹ The precise mechanism by which autophagy is suppressed in the type 2 diabetic kidney—whether mTOR dependent or independent—needs additional study. Previous reports have indicated that the mTOR pathway is hyperactivated in both type 1 and type 2 diabetic kidneys^41–43 and that treatment with mTOR inhibitors is effective in preventing the progression of nephropathy as judged by a reduction in glomerular hypertrophy, mesangial matrix expansion, and apoptosis and an increased expression of profibrotic genes.^44–46 However, most of these studies are on the basis of data from podocytes or glomeruli, and it is unclear whether the mTOR pathway is also hyperactivated in proximal tubules during diabetic nephropathy. Our data suggest that the mTOR pathway is hyperactivated in PTECs of type 2 diabetic nephropathy but not in PTECs of type 1 nephropathy. Different autophagic adaptations between type 1 and type 2 diabetes have also been reported in diabetic cardiomyopathy.⁴⁷

Patients with diabetes have an increased risk for AKI.⁴⁸ Given that autophagy counteracts AKI,⁴ it is easy to infer that autophagy upregulation will contribute to improved kidney prognosis among patients with diabetes. However, the above observation made us aware that the nature of the autophagic defect and the cellular response to that defect should be taken into account during the implementation of these therapeutic approaches before simply upregulating autophagy. We speculated that treatment with rapamycin, the classic autophagy inducer, has different effects on renal outcome in type 1 and type 2 diabetic nephropathy after AKI. As expected, rapamycin treatment significantly ameliorated kidney injury in db/db mice, whereas it significantly exacerbated kidney injury in STZ-treated mice; it is speculated that mTOR inhibition during STZ-treated diabetic nephropathy places a greater burden on the autophagy-lysosome system, because autophagy is activated, whereas in db/db mice, rapamycin partially alleviates kidney injury by restoring autophagy. Significant variation in the effect of mTOR inhibitors on renal outcomes during various pathologic settings has been reported,^49,50 which may be derived from differences in autophagy status. Rapamycin-induced deterioration of type 1 diabetic nephropathy is reminiscent of a previous report that rapamycin treatment leads to cell death in Gaucher disease induced pluripotent cell–derived neuronal cells, in which autophagic flux is significantly blocked due to defective lysosomal clearance of autophagosomes.⁵¹ Therapies that facilitate the overall autophagy process, especially reducing lysosome overload, are required in the treatment of type 1 diabetic nephropathy.

In conclusion, autophagic activity in kidney PTECs is mainly regulated by insulin via the mTOR pathway, and consequently, distinct patterns of autophagic activity are observed in type 1 and type 2 diabetic nephropathy. These patterns should be taken into consideration when treating diabetic nephropathy by modulating autophagy.

Disclosures

None.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2018100983/-/DCSupplemental.

Supplemental Figure 1. Differential contributions of insulin, amino acids, or glucose to the autophagic activity in the liver, skeletal muscle, and kidney of GFP-MAP1LC3 transgenic mice.

Supplemental Figure 2. Starvation-induced autophagy is suppressed by insulin in HK-2 cells.

Supplemental Figure 3. Insulin-induced suppression of autophagy is dependent on activation of the mTOR pathway.

Supplemental Figure 4. High glucose stimulates autophagic activity with downstream suppression of autophagic flux in HK-2 cells.

Supplemental Figure 5. Autophagy is activated in PTECs of Akita mice and STZ-treated db/db mice.

Supplemental Figure 6. Prolonged high-glucose exposure without insulin leads to lysosomal stress in PTECs in vitro and in vivo.

Supplemental Figure 7. Autophagy deficiency in kidneys of tamoxifen-treated Atg5^F/F NDRG1 mice.

Supplemental Figure 8. Differential consequences of autophagy deficiency in types 1 and 2 diabetic kidneys.

Supplemental Figure 9. Autophagic activity of podocytes and distal tubules in STZ-treated or db/db mice.

Supplemental Table 1. Plasma glucose and insulin levels in nondiabetic 24-hour starved GFP-MAP1LC3 mice at 8 weeks of age.

Supplemental Table 2. Plasma glucose and insulin levels in db/+ or db/db GFP-MAP1LC3 mice and vehicle- or STZ-treated GFP-MAP1LC3 mice at 12 weeks of age.

Supplemental Table 3. Physiologic parameters of vehicle- or STZ-treated mice, and db/+ or db/db mice at 10 months of age.