Caffeic Acid Attenuates Diabetic Kidney Disease via Modulation of Autophagy in a High-Fat Diet/Streptozotocin- Induced Diabetic Rat

The aim of this study is to evaluate the anti-diabetic nephropathy effect of Caffeic acid and to prove our hypothesis for its mechanism of action that it may occur by reactivation of autophagy pathway via suppression of autophagy regulatory miRNAs. In vivo, high-fat diet and streptozotocin-induced (HFD-STZ) diabetic rats were treated with Caffeic acid once per day for 12 weeks before and after development of diabetic nephropathy. Blood and urine biochemical parameters, autophagy transcripts and their epigenetic regulators together with renal tissue morphology were investigated. In diabetic rats, Caffeic acid intake, caused improvement in albumin excretion,blood glucose, reduced renal mesangial matrix extension with increased vacuolation and reappearance of autophagosomes. Meanwhile, it resulted in autophagy genes up-regulation [RB 1-inducible coiled coil protein (RB1CC1), Microtubule-associated proteins 1A/1B light chain 3(MAP1LC3B), Autophagy related gene (ATG-12),] with simultaneous reduction in their epigenetic regulators; miRNA-133b, −342 and 30a, respectively. These above mentioned effects were more significant in the diabetic nephropathy Caffeic treated rats than in the prophylactic group. Based on our results we postulated that caffeic acid modulates autophagy pathway through inhibition of autophagy regulatory miRNAs, that could explain its curative properties against diabetic kidney disease.

the resumption or activation of autophagy. Proautophagic drugs are a promising class of compounds for diabetic nephropathy regression. Recent studies revealed many phytochemical constituents that can induce autophagy 8 .
Wang and his colleagues reported that caffeic acid extract of Artemisia dracunculus L. enhances insulin receptor signaling and modulates gene expression in skeletal muscle in KK-Ay mice 9 . Scientists recently reported that a semisynthetic compound derived from caffeic acid derivative induces DNA damage and apoptosis in tumor cells via induction of autophagy in cancer cells 10 . Also, Caffeic acid derivative have been reported to have a protective role in renal damage 11 . We sought to determine the efficacy of caffeic acid on modulating autophagy pathway via inhibiting expression of miRNA related DKD with subsequent upregulation of its target autophagy related genes utilizing HFD-STZ -diabetic rats.
The rational of this study was based on; (1) Bioinformatics analysis to retrieve a set of 3 DKD-characterestic autophagy genes(RB 1-inducible coiled coil protein (RB1CC1), Microtubule-associated proteins 1A/1B light chain 3(MAP1LC3B), Autophagy related gene (ATG- 12)]. Then to choose miRNAs (miR-133b, miR-342 and miR-30a) relevant to DKD and acting as epigenetic regulators of the former autophagy genes based on previous microarray studies; (2) Experimental validation to characterize the efficacy of caffeic acid on modulating the expression of chosen miRNA-autophagy target gene pairs in HFD-STZ induced rat model.

Results
Effect of caffeic acid on fasting blood glucose and other metabolic parameters in HFD-STZ induced diabetic rats within the first five weeks. STZ resulted in a significant increase in blood glucose level, compared to either control or naïve group. [F = 463.8, p < 0.01], with a significant increase of blood glucose level over time [F = 319.7, p < 0.01]. Bonferroni posttest was used to compare the both groups at different time points. Blood glucose level in HFD-STZ treated group was significantly higher from the naïve group starting at week 2, and progressively increased till week 4 (P < 0.01; Fig. 1).
After 12 weeks, the levels of FBG, serum total cholesterol, LDL, total triglycerides in the DM group were significantly higher than those of the NC group. Compared with DM group, Caffeic acid treatment and prophylactic groups showed marked lowering of FBG and lipid profile indicating a clear improvement in the glucose and lipid metabolism ( Table 1).
Effect of caffeic acid on the renal function of HFD-STZ induced diabetic rats. Urinary albumin, urinary creatinine, serum creatinine, blood urea nitrogen (BUN), urea and creatinine clearance in DN group were increased compared to NC group. Dose-response effects of CA on creatinine clearence in HFD-STZ-rats was performed. Increasing the dose of CA resulted in an improvement of creatinine clearence. Thus, we found that CA had a pronounced dose-dependent effect on improvement of creatinine clearance (Fig. 2) and 40-50 mg/kg of CA were the most effective doses for improving creatinine clearence in HFD-STZ rats.
The ameliorating effect of caffeic acid on renal function parameters was evident both in the treatment and prophylactic groups, ( Table 1).

Effect of caffeic acid on renal histology and autophagy induction in HFD-STZ induced diabetic rats.
Parallel to the development of DM in HFD-STZ treated rats, histopathological examination of H&E stained kidney tissues by light microscopy revealed a focal inflammatory cells infiltration as well as detection of aggregation in-between the degenerated tubules and glomeruli at the cortex in diabetic group. Moreover, experimentally induced DKD group showed severe congestion in the glomeruli associated with swelling in the tubular lining epithelium at the cortex. Furthermore, prophylactic administration of caffeic acid in HFD-STZ induced diabetic rats showed mild congestion in the glomeruli associated with swelling in the lining tubular epithelium. Lastly treatment with caffeic acid in HFD-STZ induced diabetic rats resulted in glomerular vacuolization in the lining endothelium of the tufts associated with degeneration in the tubular lining epithelium at the cortex. All in all, prophylactic and caffeic acid treated group, significantly ameliorated the renal changes with extensive vacuolation( Fig. 2A-E). To visualize autophagy induction by caffeic acid in DKD, TEM was performed. Electron micrographs of autophagosomes was demonstrated in figure (Fig. 3A-K). Rare autophagic vacuoles were detected

Continued
in the tubular cells of the control and diabetic rats. However, an increased level of autophagy was observed in the groups treated with caffeic and prophylactic group. Autophagic vacuoles including autophagosomes and autophagolysosomes were markedly increased in the proximal tubular cells. Double membrane vacuoles containing electron-dense material, degenerating cytoplasmic organelles and cytosol, mitochondria with loss of visible cristae were frequently observed in the tubules.

Effect of caffeic acid on the expression of autophagy transcripts in DKD.
At the end of the 12 weeks, RB1CC1, ATG12, MAP1LC3B mRNA expression were estimated in the kidney of the 6 groups and were downregulated in the DKD group compared to the naïve group (3.8, 5.2 and 13 fold, respectively). Caffeic acid administration in the treated and prophylactic groups resulted in a significant induction of renal RB1CC1, ATG12, MAP1LC3B mRNA expression, compared to the control group [F = 6.3,4.9 and 4.7, respectively, at p < 0.01]. However, we could not find any significant difference between the treated and prophylactic groups ( Table 2  Effect of caffeic acid on the renal expression of miR-133b, −342,30a in DKD. At the end of the 12 weeks, miR-133b, -342, 30a expression were estimated in the kidney of the 6 groups and were up-regulated in the DKD group compared to the naïve group (1900, 175 and 125 fold, respectively).
Caffeic acid administration in treated and prophylactic groups resulted in a significant down regulation of renal miR-133b, -342, 30a expression, compared to the control group [F = 6.7,7.896 and 7.814, respectively at p < 0.01]. Notably, there was no significant difference between the treated and prophylactic groups (Table 3 and Fig. 5a-c). Of note, there was a significant negative correlation between the chosen DN-characteristic miRNAs and their corresponding autophagy target genes(p < 0.01) among all the groups of the study (Table 4).

Discussion
Until the current study, whether caffeic acid exerted any curative or prophylactic effect in DN was unknown. To address this question, we investigated the effect of caffeic acid on urinary albumin excretion in HFD-STZ induced diabetic rats. Obesity was induced in rats with high-fat diet and streptozotocin injection, which affects islet β-cells and results in insulin resistance 12,13 . Urinary albumin has been used as a predictive biomarker for prognosis of DN. Furthermore, the reduction in urinary albumin in DN reportedly were correlated with renal protection 14 .
Herein, we show that DN caused significant increases in the urinary albuminuria, whereas caffeic acid suppressed this effect.
Caffeic acid treatment and prophylaxis markedly increased RB1CC1, ATG12, MAP1LC3B mRNA expression in diabetic kidneys and decreased congestion with marked vacuolation in diabetic glomeruli. We used transmission electron microscopy (TEM) to monitor the appearance of autophagosomes. As shown in Fig. 5A, no obvious autophagic vacuoles were found in control with numerous autophagic vacuoles appeared in caffeic acid treated and prophylactic group. In addition, Caffeic acid inhibited miR-133b, -342, 30a expression (epigenetic regulators of the chosen autophagy genes) in diabetic nephropathy rats. Our findings suggested that caffeic acid resulted in marked induction of autophagy in diabetic kidneys with potential curative and preventative effect. In agreement with previous studies that addressed the role of caffeic acid in improving glucose utilization in insulin-resistant        Table 3. miRNA Expression among different groups of animal model. stress 16,17 . Moreover, previous studies explored the role of caffeic acid and its derivative in induction of autophagy 10,18 .
Accumulating evidences indicate that change in the nutrient-sensing paths in diabetic states possibly will alter the autophagic response stimulated by cellular stress, which could subsequently result in diabetic nephropathy 19 . Impaired autophagy lead to accumulation of p62/SQSTM1 protein in proximal tubular cells 20 , activation of the mTOR pathway 21 and inactivation of AMPK 22 . Modulation of the autophagy pathway has a great impact on development of a new nephro protective and therapeutic option 23 . Caffeic acid triggered induction of AMPK, class III PI3-kinase and the autophagic response in tumor 11,28 .
Autophagy involves a series of dynamic membrane rearrangements controlled by a set of ATG proteins 24 . RB1-inducible coiled-coil protein 1(RB1CC1 or ATG 17), which are needed for phagophore formation and initiation of autophagy 25,26 . Microtubule-associated proteins 1A/1B light chain 3B is a central gene in the autophagy pathway where it functions in autophagosome membrane expansion and fusion events and have structural homology with ubiquitin 27 . ATG12, forms the ATG12-ATG5-ATG1 complex involved in autophagosome maturation 28 . Zahng et al., reported that, erlotinib increased renal autophagy, as indicated by altered expression and activity of ATG12, and LC3A II, in diabetic mice 29 . Fang and his colleagues demonstrated that the expression of autophagy related proteins such as Beclin-1, ATG12-ATG5 and LC3-II was markedly inhibited in DKD 30 . Recent evidence showed that induction of autophagy may be linked to maintaining renal homeostasis in diabetic kidney 31 . In the light of our results, it seems that caffeic acid might improve diabetic nephropathy through the restoration of autophagy activity in diabetic kidneys.
Because miRNAs are recently linked to regulation of autophagy pathway 32 , we applied combined bioinformatics analysis to retrieve DN related autophagy genes and their miRNA regulators. Accordingly, miR-133b, -342 and -30a target the above mentioned autophagy genes and were previously reported by our research group as urine markers for DN 33 . miR-30a inhibits autophagy by selectively down regulating ATG5 and Beclin 1 expression 34 . Targeting miR-30a, induces autophagy in response to imatinib treatment in chronic myeloid leukemia 35 . miR-30a and miR-34a play their key roles in the regulation of autophagy pathway such as PI3KCI/Akt/mTORC1 signaling pathway, and Ras-Raf-MAPK cascade 36 . Interestingly, miR-133b induced autophagic cell death in colon cancer cells 37 . Vitamin E-based therapy for hyperglycemia & T2DM triggers the expression of AMPK via regulation of miRNA-133b 38 . Moreover, miR-342-5p is coupled to the antiviral IFN response 39 which in turn linked to autophagy.
Our Experimental model revealed that treatment with caffeic acid suppressed the expression of these miR-NAs with subsequent induction of autophay which ameliorated glomerular changes, albuminuria with reducing blood glucose levels in HFD-STZ -induced diabetic rats. We hypothesize that caffeic acid seems to trigger AMPK signaling, PIK3 pathway via regulation of miR-30a,-342,-133b which in turn induces autophagy that ameliorates diabetic nephropathy.

Conclusion
We adopt alternative strategies for better management of DN with the interest in the search of new drugs from natural sources and determining their mechanism of action which will be of great value in developing countries with limited resources and high incidence rates of diabetes mellitus. Also, there is a profitable concern in the  Table 4. Correlation between the selected miRNAs and autophagy transcript among the investigated groups of rats. **Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed). ***Indicates P < 0.001, Unpaired, two tailed t test, N = 8/group. r:Spearman correlation coefficient.
Scientific RepoRts | 7: 2263 | DOI:10.1038/s41598-017-02320-z advancement of tools to suppress or stimulate miRNA expression associated with autophagy markers in vivo as an advance in the management of diabetes complications.

Material and Methods
Animal Experiments. All procedures for the care and use of laboratory animals were approved by the Institutional Animal Ethics Committee for Ain Shams University, Faculty of Medicine. All the methods were performed in accordance with the relevant Ethical guidelines and regulations (Ethical committee approval no FWA 000017585). Male Wistar rats (weighing 250-300 g) were purchased from National Research Institute (Cairo, Egypt) and were accommodated in an animal house with temperature (22 ± 1C) and lighting (12 h light-dark cycle) control. Before the start of the experimental work, an adaptation one week during which rates were administered vehicle (tap water).
Experimental protocol. Caffeic acid and STZ were purchased from Sigma Aldrich 40 . Forty eight male Wistar rats weighing 250-300 g were divided into six groups (Fig. 6), 8 animals each as shown in Fig. 1. After one week adaptation, rats were splitted into a high-fat group (32 rats) which obtained a high-fat diet for four weeks and a normal age-matched control group (16 rats) which received a standard diet and subdivided into control and naïve groups. All high fat group after 4 weeks had received STZ 30 mg/kg. i.v. once with high fat diet for another 4 weeks). HFD-STZ developed type II DM with fasting blood glucose above 16.7 mmol/L. Afterwards, they were haphazardly divided into two groups: diabetic control group (8 rats), DKD model group, 24 rats (They were further subdivided into three groups: diabetic nephropathy control group (8 rats); herbal extract caffeic acid (CA) treated group, 8 rats (CA 40 mg/kg body weight/day orally for 4 weeks) and pretreated group: 8 rats (received the herbal extract CA after induction of diabetes for 4 weeks). CA was dissolved in cold water and administrated  Scientific RepoRts | 7: 2263 | DOI:10.1038/s41598-017-02320-z via intra-gastric gavage (i.g.) one time daily for twelve weeks according to Jayanthi et al.40 and Dhungyal et al. 41 (Fig. 1). To study the effectiveness of CA at different doses in modulating renal function, creatinine clearance was estimated after CA treatment period in the HFD-STZ rats (Fig. 7).
Induction of DKD. Type 2 diabetes mellitus was provoked according to Zhang et al. 42,43 . In the normal group, rats were fed a standard chow diet of a total kcal value of 20 kJ/kg (52% carbohydrate, 20% protein, 5% fat), while diabetic group rates were fed a high-fat diet of a total kcal value of 40 kJ/kg (45% carbohydrate, 22% protein, 20% fat). The two groups were kept on their diets for 8 weeks duration. In the 4 th week, a single low dose of STZ (30 mg/kg, dissolved in 0.1 M sodium citrate buffer at pH 4.4) was injected into each rat of the diabetic group intraperitoneally. After the STZ injection, rats obtained drinking water containing sucrose (15 g/L) for 48 hours, to reduce early death due to insulin discharge from partially injured pancreatic islets. Seventy two hours later, rats were checked for hyperglycemia and those with FBS more than 250 mg/dL were included in studies of diabetic nephropathy. Diabetic rats received long-acting insulin (2-4 U/rat) via S.C injection to maintain blood glucose levels in a desirable range (300 mg/dL) and also to prevent subsequent development of ketonuria 44 . Nephropathy was noted in rats (4-8 weeks) after the administration of STZ and was assessed in terms of significant increase in proteinuria, serum creatinine, blood urea nitrogen (BUN), extracellular matrix deposition and thickening of glomerular basement membrane. At the end of the eight-week pretreated and treated, rats were sacrificed using an intraperitoneal injection of sodium pentobarbital (50 mg kg −1 ). Blood samples and kidney were collected for biochemical, histopathological analyses and TEM. A 24-h urine was collected on the day before scarification.
Metabolic parameters, urinary albumin excretion, and renal function analysis. Fasting blood sugar(FBG) was measured using the glucose oxidase method. Serum total cholesterol, LDL, HDL, Triglycerides, creatinine, and urine creatinine were detected by automated clinical chemistry analyzer (Olympus-2000, Tokyo, Japan). Samples from rat kidneys were snap frozen at −80 °C for further RNA extraction, and another samples were processed for histopathological examination and TEM.
Histopathological Examination. Kidneys were fixed (10% neutral buffered formalin (NBF)), paraffin embedded, cut into 4 μm sections and then stained with Hematoxylin and Eosin to be examined at 400 magnification by 2 independent histopathologists.
Transmission electron microscopy (TEM). Kidney tissue samples were fixed in glutaraldehyde and osmium tetroxide, washed with PBS& dehydrated. After exchange through acetone, the samples were later embedded in Epon 812. The kidney tissue were made into ultra-thin (70-80 nm) after observation and positioning, and were double stained following standard methods. All of the kidney samples were examined using TEM.
Afterwards, 500 ng total RNA from kidney was changed to cDNA by miScript II RT Kit (Qiagen, Valencia, CA) using miScript HiSpec buffer which was used in real-time PCR analysis using a miScript primer assay and the miScript SYBR Green Kit. The miScript Universal Primer (reverse primer) and QuantiTect SYBR Green PCR Master Mix, respectively were used as per manufacturer's protocol to detect 3 miRs (miR-133b, miR-342, miR-30a) and 3 autophagy genes(RB1CC1, ATG12 and MAP1LC3B), respectively. RNA quality control and housekeeping gene (RNU6-2, GADPH) were also included in the assay for miRNA and autophagy gene expression, respectively. The qPCR tubes were run on a Step One Plus ™ System (ABi). The primer of selected miRNAs, autophagy transcript and endogenous control were purchased from Qiagen. Data Analysis were done by using the ΔΔCT method of relative quantification 46 on a Light-cycler, software v2.2.2 (StepOne ™ Software). Data were presented as fold change in expression and were calculated as 2 (−ΔΔCT) . Where ΔCT = CT target gene -AVG CT reference gene and ΔΔCT = ΔCT (sample 2) − ΔCT (sample 1) where sample 1 is the control sample and sample 2 is the experimental sample.
Statistical tests. Data are presented as mean ± S.D. Statistical differences between the groups were estimated by one-way analysis of variance (ANOVA) followed by Bonferroni's post-tests. A p-Value < 0.05 was considered significant. Statistical analyses carried out using Graph Pad Prism (GraphPad software) and with SPSS version 21.0 (SPSS, Chicago, IL, USA).