Autophagy-lysosome pathway in renal tubular epithelial cells is disrupted by advanced glycation end products in diabetic nephropathy

Autophagy was inactivated since membrane permeabilization and ABSTRACT It has been suggested that autophagy protects renal tubular epithelial cells (TECs) from injury in diabetic nephropathy (DN). However, the manner in which the autophagy-lysosome pathway is changed in this state remains unclear. In this study of DN, we investigated the autophagic activity and lysosomal alterations in vivo and


INTRODUCTION
In recent years, diabetic nephropathy (DN), a common consequence of type 1 and 2 diabetes mellitus, has garnered much attention because of its high prevalence, poor prognosis and heavy economic burden. In renal intrinsic cells, diabetes mellitus (DM)-mediated alterations of extra-and intra-cellular metabolism and hemodynamics influence DN (1). And advanced glycation end products (AGEs) play a key role during this process (2). Although some potential targets for renoprotective therapies, including AGEs inhibitors ， renin-angiotensin-aldosterone system blockers and peroxisome proliferator activated receptor-γ agonists, have been studied and tested in experimental and clinical models, the development of end-stage kidney disease remains a major concern (3). Therefore, more specific pathophysiologic mechanisms underlying DN should be identified to develop new treatment modalities and strategies for this devastating disease.
Autophagy activation is an important mechanism that degrades damaged organelles and abnormal and misfolded proteins to sustain cellular metabolism (4). Autophagy is initiated by the formation of a double-membraned autophagosome, which then fuses with the lysosome to form an autophagolysosome, in which the enclosed cargo is degraded by lysosomal acid hydrolases (5). Therefore, autophagy cannot function when any part of this process is interrupted. Lysosomal dysfunction is a major cause of autophagy inactivation that leads to the accumulation and aggregation of proteins, including defective organelles and autophagy substrates, resulting in cellular toxicity (6,7).
Renal tubular epithelial cells (TECs) are key players in orchestrating events in DN (8). TECs exhibit a certain level of autophagy under physiological conditions. Autophagy activation is an important mechanism that protects TECs against various injuries induced by aristolochic acid, cisplatin, cyclosporine A or endotoxemia (9-12). We also found that autophagy activation may reduce TEC injury induced by urinary proteins (13). Notably, several studies have suggested a role for the autophagy of TECs in DN (14). However, the manner in which the autophagic pathway changes during DN remains unclear. In our study, the changes of autophagy-lysosome pathway are investigated that occur in TECs after exposure to AGEs.

EXPERIMENTAL PROCEDURES
Patients The Institutional Review Board of the Affiliated Hospital of Guangdong Medical College approved this study. All of the clinical data from 17 patients (35-80 years) at the Affiliated Hospital of Guangdong Medical College were de-identified. Kidney tissue specimens were obtained from biopsy-proven DN patients (n = 11) with mild, diffuse or nodular glomerulosclerosis. The kidney specimens (n = 6) obtained from patients with mild urinary protein excretion or only hematuria and characterized with a minimal change in histology were used as controls.
Enzymatic assay To measure the activity of CB, CD or CL, fluorescence-based assay kits (K140-100, K143-100 and K142-100, BioVision) were used. After cleavage of the synthetic substrate by the cell lysate, the released fluorescence was quantified using a fluorescence plate reader, according to the manufacturer's instructions.
Immunofluorescence study An immunostaining analysis of the tissues and cells was performed as described previously (15). Rabbit anti-LC3B, rabbit anti-LAMP1, mouse anti-CB, mouse anti-LAMP1 (ab51520, ab24170, ab58802, ab25630; Abcam), mouse anti-SQSTM1, mouse anti-ubiquitin (sc28359, sc8017; Santa Cruz, CA, USA), rabbit anti-SQSTM1 (PM045; MBL, Nagoya, Japan), Alexa Fluor ® 488 donkey anti-rabbit and Alexa Fluor ® 594 donkey anti-mouse IgG antibodies (A21206 and A21203; Invitrogen, Carlsbad, CA, USA) were used for staining. The images were taken with a TCS SP5 II confocal microscope (Leica Microsystems, Wetzlar, Germany). The LC3-II-or SQSTM1-positive puncta/cell were counted in at least 30 HK-2 cells for each group. The puncta were assessed blindly by two independent investigators and the slides were mixed with a set of normal slides to ensure that the examiner was blinded to its information. The expression level of LC3-II or SQSTM1 in each proximal renal tubule was first graded on a scale of 0 to 4, and the average of these scores was subsequently computed as described previously (13).
Reactive oxygen species (ROS) assay After incubated by 10 μ Ovalbumin dequenching assay and LTR uptake test After exposure to AGE-BSA for 12 h, the HK-2 cells were incubated with 10 µg/ml DQ-ovalbumin (D12053; Invitrogen) for an additional 2 h or with 50 nM LTR (L-7528; Invitrogen) for an additional 15 min in complete DMEM at 37°C. For the ovalbumin dequenching assay, the cells were washed with PBS and fixed with 4% paraformaldehyde. Green-fluorescent DQ-ovalbumin puncta in individual HK-2 cells were counted and the average numbers of dots in at least 30 cells are presented in the figures. For the LTR uptake test, the cultures were washed with PBS to remove any excess lysosomal marker, and the images were taken with a TCS SP5 II confocal microscope. The mean fluorescent intensity value of the LTR was measured, and the data are expressed as fold changes relative to the control. In addition, HK-2 cells were trypsinized and re-suspended in PBS for FACS analysis to measure the mean fluorescent intensities of fluorescent DQ-ovalbumin and LTR. To determine whether the dequenched DQ-ovalbumin vesicles were co-localized with lysosomes, after DQ-ovalbumin incubation the cells were fixed with 4% paraformaldehyde at room temperature and permeabilized with ice-cold methanol at −20 °C for 10 min, respectively. Following washed in PBS, the cells were labeled with mouse anti-LAMP1 and visualized by second antibody.
Statistical analysis All of the statistical tests were performed using SPSS 16.0. All of the data are expressed as the means ± standard error of the mean (S.E.M.), counts or percentage. Two-group comparisons were performed using an independent-sample t-test unless otherwise indicated. Multiple-group comparisons were performed using ANOVA followed by Bonferroni or Dunnett post-hoc tests. Differences with a P value less than 0.05 were considered statistically significant.

Autophagic vacuoles are accumulated in
TECs during the progression of DN The clinical characteristics of the 17 enrolled patients were presented in Table 1. There was no statistical significance between the DN group and the control group with regard to sex or average age. Compared with the controls, the hemoglobin level decreased, whereas the serum creatinine, blood urea nitrogen and incidence of diabetic retinopathy and neuropathy increased in DN patients. The HbA1c, blood fat, serum uric acid, and urinary protein levels and the incidence of hypertension were higher, but the serum albumin was lowered in the DN patients than controls.
In our study, we utilized immune-fluorescent technology to examine the expression of the LC3-II, a key marker of autophagy, in TECs both in vivo and in vitro. As shown in Fig. 1A, C, compared with the control, there were more LC3-II-positive puncta in the renal TECs from patients with DN. To mimic the pathological process of kidney disease induced by DM, the human renal tubular epithelial cell line HK-2 was exposed to AGE-BSA in vitro. The AGE-BSA concentration at 100 µg/ml was utilized, since this dose was reported to represent AGEs level in the serum of diabetic patients (17). Before incubation with AGE-BSA, only a few LC3-II puncta were detected in the HK-2 cells through immunofluorescence staining (Fig. 1B, D). After exposure to 100 µg/ml AGE-BSA, the number of LC3-II-positive puncta tended to increase at 3 h (data not shown), and significantly increased at 6, 12 and 24 h (Fig. 1B, D). A similar pattern of LC3-II expression was observed by western blotting (Fig. 1E).

Autophagy is inhibited by AGEs in TECs during DN
To explore the accumulation of autophagic vacuoles was either a sign of autophagic activation or a hint of impaired autophagic degradation, the autophagy substrate SQSTM1/p62 was next examined. As shown in Fig. 1A, C, more SQSTM1-positive dots were observed in the tubular cells from DN patients than the controls. The in vitro immunofluorescence and western blot assays also confirmed that the SQSTM1 protein was increased in HK-2 cells after exposure to AGE-BSA for 6, 12 and 24 h (Fig. 1B, D, E), which is most likely an indicator of autophagy inhibition.
To separately evaluate the extent of autophagosome and autolysosome accumulation, tfLC3 was transfected into HK-2 cells. In HK-2 cells overexpressing tfLC3, green and red puncta were clearly visible. The numbers of both GFP (green) and mRFP (red) dots were significantly increased after exposure to rapamycin (an autophagy inducer) or AGE-BSA for 12 h ( Fig. 2A, B). There were more yellow and red dots in cells after rapamycin exposure than the controls ( Fig. 2A, C), indicating an increase in the formation of both autophagosomes and autolysosomes. However, only yellow, but not red dots were found increased in AGE-BSA treated cells, suggesting a large number of autophagosomes were not degraded by lysosomes ( Fig. 2A, C). We then treated cells with the lysosomal inhibitor chloroquine, and found that the numbers of yellow and free red puncta were not significantly changed in AGE-BSA-plus-chloroquine-treated cells compared with the cells treated with AGE-BSA only ( Fig. 2A, C). It further indicates that AGE-BSA results in the inhibition of autophagosome degradation.
It is well known that bafilomycin A1 ， leupeptin and chloroquine block the fusion of autophagosomes and lysosomes, inhibit protease activity and increase intralysosomal pH, respectively (18). So, LC3-II turnover was also examined in presence and absence of these lysosomal inhibitors (19). We found that the protein levels of LC3-II were significantly increased after exposure to AGE-BSA for 12 h. However, LC3-II expression was not further elevated by the addition of bafilomycin A1， leupeptin or chloroquine by western blot assay (Fig. 3A). Consistent with the western blot results, an additive increase in LC3-II accumulation was not observed by immunostaining in the bafilomycin A1-， leupeptin-or chloroquine-plus-AGE-BSA-treated group compared to the group treated with AGE-BSA only (Fig. 3B, C). These data indicate that LC3-II is not degraded by lysosomes and that autophagic activity is inhibited upon the exposure of HK-2 cells to AGE-BSA.
To determine whether AGE-BSA blocked the fusion of autophagosomes and lysosomes, LC3 and LAMP1 were double-stained. We found that LAMP1 granules had a uniform distribution in Co-BSA-treated cells. After treated with AGE-BSA, cytoplasmic LAMP1 granules became irregular and larger and most of them were co-localized with the LC3-II-positive puncta (Fig. 3D). It seems that the increased LC3-II dots can fuse with the irregular LAMP1 granules.
Autophagy inhibition is attributed to impaired lysosomal activity after exposure to AGEs We next examined the lysosomal activity since the lysosome-mediated degradation system plays a key role in autophagy activation. For enzyme activity in lysosomes, AGE-BSA but not the non-glycated Co-BSA induced a significant decrease in CB activity at 6, 12 and 24 h (Fig. 4A). Similar results were obtained for CL activity, although CD activity was not changed significantly at any time points (just showed a decreasing tendency) (Fig. 4A). Different from the activities of CB and CL, CD is reported to remain active at neutral pH (20), which might explain the reason behind the unchanged activity of CD. We subsequently used LTR to label and track the acidic intracellular compartments (lysosomes) in live cells. Co-BSA treatment did not change the fluorescence at any time points (data not shown). However, AGE-BSA abolished the LTR staining from 6 to 24 h when examined by fluorescence microscopy (Fig. 4B, C). LTR signal evidenced by mean fluorescent intensity was also declined after exposure to AGE-BSA for 6, 12 or 24 h by FACS analysis (Fig. 4D). These data indicate a reduction in the intracellular acidic components and lysosomal impairment after AGE treatment. To further evaluate the efficiency of lysosome-mediated proteolytic degradation, a self-quenched substrate for proteases, DQ-ovalbumin, was used. A large number of dequenched DQ-ovalbumin dots were co-localized with LAMP1-positive puncta in Co-BSA-treated HK-2 cells, indicating the lysosomal degradation of DQ-ovalbumin. However, accompanying an occurrence of irregular and larger LAMP1 granules, few DQ-ovalbumin vesicles were detected in AGE-BSA-treated cells (Fig. 5A). After exposure to AGE-BSA, the mean DQ-ovalbumin puncta per cell were significantly decreased at 6, 12 or 24 h when compared with 0 h (Fig. 5B, C). Also, the mean fluorescent intensity of DQ-ovalbumin was lower in AGE-BSA-treated cells than non-treated control, as assessed by flow cytometry (Fig. 5D). All of these data suggest a decrease in the degradative potential of the lysosome after treatment with AGE-BSA.
RAGE-mediated oxidative stress contributes to the lysosomal impairment after exposure to AGEs It is well known that AGEs can mediate multiple effects via a specific receptor RAGE (21), so the possible action of RAGE on lysosomes was subsequently tested. AGE-BSA could significantly abolish LTR staining and DQ-ovalbumin fluorescence. However, these actions were inhibited partially by anti-RAGE antibody pre-treatment under the microscope (Fig. 6A and 7A). Similar results were obtained when assessed the mean fluorescent intensities of LTR and DQ-ovalbumin by flow cytometer (Fig. 6B and  7B), indicating that AGEs-RAGE interaction is an important mechanism underlying lysosomal dysfunction. Furthermore, we found that the production of reactive oxygen species (ROS) was significantly enhanced after exposure of HK-2 cells to AGE-BSA for 12 h. However, similar to the effect of antioxidant NAC or catalase, anti-RAGE antibody notably suppressed the ROS production mediated by AGE-BSA (Fig.  7C).
Interestingly, pre-treatment with antioxidant NAC or catalase also improved the lysosomal dysfunction in AGE-BSA-treated cells, characterized by an increase in LTR and DQ-ovalbumin fluorescence signals (Fig. 6A, B and 7A, B). These data suggest that AGEs-RAGE axis evokes oxidative stress generation, which plays an important role in lysosomal impairment. Finally, the endocytosis inhibitors genistein and DC (22,23) were also used in our study to test the possible effect of endocytosed AGE-BSA on lysosomes. Lysosomal dysfunction was not significantly affected by pre-treatment with the endocytosis inhibitors (data not shown), indicating that the effect of AGEs is endocytosis-independent.
Lysosomal membrane permeabilization (LMP) is triggered by AGEs in TECs during DN LMP causes the release of cathepsins from the lysosomal lumen to the cytosol. Therefore, we ascertained the cathepsin immunoreactivity to determine whether LMP occurs in TECs. After exposure to Co-BSA, HK-2 cells displayed fine, granular, perinuclear CB immunoreactivity at all time points. However, in the AGE-BSA-treated cells, the CB granules were irregular and dispersed, suggesting the release of CB from the lysosomes into the cytosol (Fig. 8A, B). To further clarify the relationship between CB leakage and LMP, CB and LAMP1 were double-stained. In Co-BSA-treated HK-2 cells, LAMP1-immunoreactive granules were regularly distributed in the cytosol and most of those were co-localized with CB. Accompanying the dispersed CB expression, irregular and larger cytoplasmic LAMP1 granules were detected in many cells after exposure to AGE-BSA (Fig. 8A, C). In kidney preparations from the controls, the majority of TECs presented regular CB granules, which were co-localized with LAMP1 puncta. However, diffuse CB immunostaining, accompanied by dispersed LAMP1 immunofluorescence, was observed in many TECs and in even whole tubules (Fig. 8D). This finding indicates that LMP occurs in the TECs of DN patients, which may contribute to lysosomal dysfunction.
Ubiquitinated proteins accumulate after exposure to AGEs To determine the functional importance of autophagy inhibition and lysosomal impairment in TECs, SQSTM1 proteins and ubiquitinated proteins were double-labeled. Both proteins were rarely detected in the Co-BSA-treated group (data not shown). Compared to the non-treated group, the ubiquitinated proteins significantly accumulated in the cytoplasm after exposure to AGE-BSA, particularly at 12 and 24 h. Interestingly, these proteins predominantly co-localized with the SQSTM1-positive puncta (Fig. 9A). indicating the disruption of the autophagy-lysosome pathway because autophagic-lysosomal degradation is required for removing ubiquitinated proteins (24). Furthermore, the western blot assay revealed the accumulation of SQSTM1 and ubiquitinated proteins after treatment with AGE-BSA (Fig. 9B).

DISCUSION
Diabetes-induced decrease in the lysosomal enzyme activity (25,26) and reduction in the renal lysosome-mediated degradation of albumin (27) have been implicated in several studies. We found that AGEs-RAGE interaction, but not endocytosis of AGEs, induced a significant decline of lysosomal activity and degradative potential for DQ-ovalbumin. In addition to the cathepsin leakage and irregular LAMP1 expression, the decreased LTR fluorescence also suggested the occurrence of LMP (20), which is likely a primary cause of lysosomal dysfunction. LMP is susceptible to oxidative stress suggesting that overproduced ROS plays a key role in governing LMP (28). Furthermore, the cytoplasmic active cathepsins released from lysosomes might play an important role in triggering TEC apoptosis as reported previously (29). Nearly all intrinsic renal cells have a basal level of autophagy (14, 30,31). However, the manner in which the autophagy process changes in these cells in DN remains unclear. In TECs from diabetic animals, inhibited autophagy was implicated by Zhan's investigation (32), while autophagy activation was suggested by Zhao's study (33). In podocytes, Fang reported that hyperglycemia induced impaired autophagy (34), whereas Ma demonstrated that high glucose promoted autophagy (35). This discrepancy is most likely because these studies mainly assessed the upstream, but not the downstream of autophagic pathway. In our study, we found that autophagic vacuoles accumulate in TECs during the progression of DN. However, an increase in autophagic vacuoles may be due to an increased formation and/or decreased clearance of autophagic vacuoles (36). A separate assessment of autophagosome/autolysosome accumulation and an accurate examination of LC3-II turnover are the principal methods for monitoring autophagic flux (16,19). The inhibition of the lysosomal degradation of autophagosomes and the blockage of the lysosomal turnover of LC3-II were shown in our study of TECs exposed to AGE-BSA, suggesting a decrease in autophagic activity. Therefore, the lysosomal degradation pathway may be more important for assessing the autophagy activity in DN patients.
Diabetes-induced changes in renal lysosomal processing were reported to link several initial events of DN, including reduced albumin reabsorption in lysosomes and an increase in lipid peroxidation in the tubules (25,37). Further study showed that lysosomal impairment contributed to a decrease in kidney protein degradation and, hence, cellular hypertrophy in diabetes (26). However, it is not clear whether autophagy inactivation induced by lysosomal impairment is involved in the accumulation of abnormal proteins. In our study, the ubiquitinated proteins was assessed in TECs, since the ubiquitination of mutant huntingtin or ubiquitin-related degradation is one of the most important ways to detoxify an abnormal cellular protein. As expected, we found that ubiquitinated proteins were accumulated after exposure of HK-2 cells to AGE-BSA. It is well known that ubiquitinated proteins target SQSTM1/p62 for autophagic degradation, and disruption of the autophagy-lysosome pathway results in ubiquitinated protein aggregates (6,38). In our study, an obvious colocalization of ubiquitinated proteins and SQSTM1 was observed, indicating that the ubiquitinated protein bodies were delivered to the autophagy machinery but not degraded by the impaired lysosomes (39). Therefore, the impaired autophagic-lysosomal degradation of ubiquitinated and aggregated proteins may cause abnormal protein accumulation, which probably recapitulates many features of DN including diabetic renal hypertrophy.
Taken together, the results show that the TEC autophagy-lysosome pathway is disrupted by AGEs during the development of DN, which results in the accumulation of abnormal proteins (Fig. 10). The autophagy-lysosome pathway should be emphasized to elucidate the mechanisms underlying DN.

CONFLICT OF INTEREST
The authors declare that they have no conflicts of interest with the contents of this article.

AUTHOR CONTRIBUTIONS
WJL and HL conceived and coordinated the study and wrote the paper. TTS and RHC designed, performed and analyzed the experiments shown in Figures 3, 5, 6 and 7. HW and YJW designed, performed and analyzed the experiments shown in Figures 1, 2 and 4 (E) Western blot analysis of LC3 or SQSTM1 after exposure to AGE-BSA for 0, 6, 12 and 24 h. Densitometry was performed for the quantification, and the ratio of LC3-II or SQSTM1 to tubulin is expressed as the fold change compared with the level in the control. * P < 0.05, ** P < 0.01 and *** P < 0.001. or free red puncta per cell. * P < 0.05 and *** P < 0.001 (for green or red puncta); ^^^P < 0.001 (for free red puncta); # P < 0.05 and ### P < 0.001 (for yellow puncta); ns, no significance.

FIGURE 3. Autophagosome and lysosome fusion analysis and LC3 turnover assay in vitro. (A)
Western blot analysis of LC3 and the difference in the LC3-II levels between samples with and without bafilomycin A1 (Baf A1), leupeptin (LEU) and chloroquine (CQ) was compared after exposure to AGE-BSA or Co-BSA. (B and C) Immunofluorescence analysis of LC3 and the difference in LC3-II puncta between samples with and without Baf A1, LEU and CQ was compared after exposure to AGE-BSA or Co-BSA. LC3-II turnover (comparing lanes 5, 6, 7 and 8 with 4) was not observed after exposure to AGE-BSA. (D) Immunofluorescent signals doubled labeled by LC3 (green) and LAMP1 (red). Scale bar, 20 μm. ** P < 0.01 and *** P < 0.001.