Piceatannol attenuates streptozotocin-induced type 1 diabetes inmice

As a natural analog of resveratrol, piceatannol (Pic) exhibits good antioxidant and anti-inflammatory activities in different disease models. However, the role of Pic in type 1 diabetes mouse model has not been reported yet. In this study, we investigated the in vivo effect of Pic in streptozotocin (STZ)-induced type 1 diabetic mice. Mice were injected with STZ to establish the type 1 diabetes mellitus (T1DM)model. After stable hyperglycemia was achieved, mice were then orally treated with Pic (40 mg/kg b.w., i.g.) for 30 days. The results indicated that Pic supplementation efficiently alleviated the typical symptoms associated with T1DM, including body weight loss, polydipsia, hyperglycemia, and hypoinsulinemia. Pic treatment also improved the glucose tolerance of STZ-induced diabetic mice. In addition, Pic supplementation markedly decreased the expression of pro-inflammatory molecules TNF-α and IL-6, the expression of endoplasmic reticulum (ER) stress markers GRP78 and CHOP, and the level of oxidative stress in T1DM mice. Moreover, Pic administration also partly reversed the metabolic profiles of STZ-treated mice as detected by H Nuclear Magnetic Resonance (NMR)-based metabolomics. Our study suggested that the therapeutic potential of Pic in type 1 diabetes and the anti-diabetic effects of Pic may be associated with its activities to suppress oxidative stress, inflammation, and ER stress.


Introduction
As one of the most common metabolic diseases, diabetes mellitus (DM) has become a global threat to public health. The World Health Organization (WHO) has estimated that more than 592 million people will suffer from diabetes in 2035 (Guariguata et al., 2014). Moreover, the global incidence of type 1 diabetes mellitus (T1DM) is also increasing and will be more than 90 million in 2035 (Guariguata et al., 2014). Extensive efforts have been devoted to investigating the underlying pathological mechanisms and promising therapeutics for diabetes (Dal Monte et al., 2019;Gencoglu et al., 2015;Ben Nasr et al., 2017). Accumulative evidence suggests that T1DM exhibits as a chronic autoimmune metabolic disease and many factors including both genetic and environmental factors contribute to its development and the associated complications (Alberti and Zimmet, 1999). Among these different factors, oxidative stress and inflammation are two critical mediators responsible for the pathogenesis of T1DM and its secondary complications (Negi and Jena, 2019), suggesting that compounds or natural products that possess antioxidant and anti-inflammatory properties might exhibit potential effects in treating diabetes.
Piceatannol (Pic, 3,5,3',4'-trans-tetrahydroxystilbene) is a natural polyphenol found in many fruits, including grapes, blueberries, and passion fruits, etc. (Piotrowska et al., 2012;Minakawa et al., 2012). As a naturally occurring analog and a metabolite of resveratrol, Pic exerts improved bioavailability and metabolism properties than resveratrol . Hence it is more attractive to investigate the benefits of applying Pic in different disease models. Our previous study reported that Pic showed hepatoprotective effect against D-GalN/LPS-induced acute liver injury via inhibiting the generation of oxidative stress and pro-inflammatory cytokines , suggesting that Pic might exhibit anti-oxidative and anti-inflammation properties in vivo. In addition, several studies also indicated that Pic was able to reduce the production of ROS and inflammation markers in vitro or in vivo (Minakawa et al., 2012;Jin et al., 2006). Therefore, the antioxidant and anti-inflammatory activities of Pic suggest its potential effect in diabetes models. However, to our knowledge, the effect of Pic on diabetes, especially in an animal model of T1DM, has not been well determined.
In this study, we investigated the effect of Pic on type 1 diabetes using an STZ-induced T1DM model. Our results indicated that Pic effectively lowered the blood glucose level, reversed the body weight loss and induction of polydipsia and hypoinsulinemia, improved the glucose tolerance in STZinduced diabetic mice, suggesting that Pic supplementation could alleviate the symptoms associated with T1DM.

Reagents
Pic was obtained from Adamas Reagent Co., Ltd. (Switzerland). Pic was dissolved in 0.25% ethanol within 1 h before administration. Streptozotocin (STZ) was obtained from Sigma-Aldrich, St. Louis, MO (USA). All other chemicals were of analytical grade.
Mouse studies C57BL/6 mice (males, 6-7 week-old) were obtained from the Model Animal Research Center of Nanjing University (Nanjing, China) and housed under standard laboratory conditions (22 ± 2°C, 12 h light/dark cycle) with free access to water and food. All mice experiments were performed in accordance with the protocols approved by the Animal Care and Use Committee at Nanjing University of Science and Technology.
All mice were randomly divided into four groups: (1) Control group (Con, n = 6), (2) Pic treatment group (Pic, n = 6), (3) diabetes mellitus group (DM, n = 8), and (4) diabetes plus Pic treatment group (DM + Pic, n = 8). The mice of DM and DM + Pic groups were injected with STZ (40 mg/kg b.w., i.p., dissolved in 10 mM citrate buffer, pH 4.4) for 5 consecutive days (Liu et al., 2017;Bai et al., 2016). Then mice with stable hyperglycemia (fasting blood glucose ≥11.1 mM) for at least 10 days were regarded as diabetic animals, and half of them (DM + Pic group) were treated with Pic (40 mg/kg b.w., i.g.) for 30 days. Mice from the DM group were treated by vehicle control accordingly. The daily food and water intake of mice were detected in the last week of Pic treatment. The dose of Pic was selected based on our previous in vivo study  as well as our preliminary experiments. The rationale for the experimental design was illustrated in Fig. 1(A).

Oral glucose tolerance test (OGTT)
The mice were fasted for 12 h and then administrated with glucose (2 g/kg b.w.) by gavage on the 30th day after Pic treatment. The tail-vein blood was harvested at 0, 30, 60, 120 min after the administration of glucose, respectively. Then the glucose levels were measured by a blood glucose meter (ACCU-CHEK Ò Performa, Roche Diagnostics) according to the manufacturer's instruction.

Biochemical analysis
Collecting blood samples, centrifuged at 3000 g for 10 min, then the serum was harvested for insulin detection. Malondialdehyde (MDA) Assay kit (A003-1), the Glutathione Peroxidase (GSH-PX) Assay kit (A005), and the Total Superoxide Dismutase (SOD) Assay Kit (A001-1) were purchased from Jiancheng Bioengineering Institute (Nanjing, China). The levels of MDA, activities of GSH-PX and SOD were measured using the above kits following the manufacturer's instructions.
Quantitative real-time PCR Total RNA was extracted from the liver tissue using TRIzol (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. The reverse transcriptase kit (Invitrogen) was used for the reverse transcription reaction according to the manufacturer's protocol. Quantitative realtime PCR amplification was conducted using an ABI 7300 Detection System with SYBR Green dye (Kapa Biosystems Pty, Ltd.). Primers sequences were listed in Tab. 1.

H NMR sample preparation and analysis
All 1 H NMR samples were prepared with reference to the previous works (Sun et al., 2017). Frozen liver tissue samples were homogenized in ice-cold acetonitrile solution (50% v/v, 5 mL/g tissue). After centrifugation (13000 g, 10 min at 4°C), the supernatant was collected, and the acetonitrile was removed by nitrogen blowing. Then the samples were frozen, lyophilized to dryness, and kept at -80°C. When subjected to NMR analysis, the dried extract samples were dissolved in 600 μL 99.8% D 2 O phosphate buffer (0.2 M, pH = 7.0) containing 0.05% (w/v) sodium3-(trimethylsilyl) propionate-2,2,3,3-d4(TSP). D 2 O was used for field frequency locking and TSP acted as a chemical shift reference ( 1 H, 0.00 ppm). After centrifugation (12000 g, 10 min), the transparent supernatant was pipetted into 5-mm NMR tubes for further detection. 1 H NMR spectra of all samples were recorded on Bruker AVANCE III 500 MHz NMR spectrometer at 25°C.

H NMR data analysis
The 1 H NMR spectra were pre-processed by MestReNova (version 11.0, Mestrelab Research SL). After phase and baseline correction and peak alignment, each spectrum was segmented into bins between 0.5 and 9.5 ppm with the excision of the regions from 4.7 to 5.2 ppm to remove the signals of water and its neighboring regions. The total spectral area of each spectrum was normalized to unity to facilitate the comparison among samples. The processed 1 H NMR data were exported to CSV files for further analysis. After mean-centered and Pareto-scaled by SIMCA (version 14.1, Umetrics), the software Chenomx NMR suite 7.7 (Chenomx Inc., Edmonton, AB, Canada) was utilized to identify the metabolites in the 1 H NMR spectra of the liver extracts. The data were also analyzed by orthogonal signal correction partial least squares discriminant analysis (OPLSDA). OPLSDA is an improved method to minimize the influence of unrelated variables between groups. Color-coded loading plots and Score plots of the OPLSDA model were obtained through MetaboAnalyst (www.metaboanalyst.ca) to show the separation between groups. The fold-change values of metabolites and their p-values were calculated and visualized in the table.

Statistical analysis
The statistical analysis of the results was performed using GraphPad Prism ® 7.01 software (San Diego, CA, USA). All the data are expressed as the mean ± SE of the mean (SEM). p < 0.05 was considered to be a statistically significant difference.

Results
Pic alleviates the symptoms and improves glucose tolerance in type 1 diabetes mice To evaluate the effect of Pic on type 1 diabetic mice, the STZinduced T1DM mouse model was used, and the experimental rationale was illustrated in Fig. 1(A). During the whole experiment process, we monitored the body weight, food and water intake in all groups of mice. As Figs. 1(B)-1(C) showed, firstly the body weight gain of the model group (DM) was lower than that of the control group. In the 7th week, the body weight of the diabetic mice started to decline in contrast to the control group. Moreover, the diabetic mice (DM group) took much more food and water than the control group (Figs. 1(E)-1(F)). The body weight change and food/water intake of the model group were all consistent with the symptoms of type 1 diabetes. However, oral administration of diabetic mice with Pic not only attenuated the decline of body weight but also significantly reduced the daily food and water intake of diabetic mice (DM + Pic group, Figs. 1(B)-1(F)), suggesting that Pic could alleviate the typical symptoms of type 1 diabetes to a certain extent. Fasting blood glucose levels in all groups of mice were also measured before and after STZ administration as well as Pic treatment. As Fig. 1(D) indicated, before STZ injection, the fasting blood glucose levels of all mice were in the normal range, around 5.5 mmol/L. However, at week 4 (3 weeks after STZ injection but prior to Pic treatment), the fasting blood glucose increased to 14 mmol/L in both DM and DM + Pic groups, indicating the successful establishment of the diabetes animal model. Pic supplementation significantly decreased the blood glucose level to 10 mmol/L, in contrast to the 17.5 mmol/L in the DM group. At the end of the experiment, the concentration of blood insulin was detected. The insulin levels in STZ-induced diabetic mice were much lower than that in the control group ( Fig. 1(G)), demonstrating the hypoinsulinemia phenotype of type 1 diabetes. Pic treatment efficiently reversed the level of insulin of DM mice ( Fig. 1(G)). Oral glucose tolerance test (OGTT) was performed to evaluate the glucose tolerance. As Fig. 1(H) demonstrated, the blood glucose of all mice reached the peak 30 min after glucose administration, and diabetic mice (DM group) showed obviously impaired glucose tolerance compared with the control group. The blood glucose of all time points, as well as the AUC value in the DM + Pic group, were significantly lower than that in the DM group ( Fig. 1(I)), suggesting that Pic supplementation notably improved the glucose tolerance in STZ-induced diabetic mice. Taken together, these results indicated that Pic supplementation efficiently alleviates the diabetes phenotypes in STZ-induced type 1 diabetic mice, including body weight loss, polydipsia, hyperglycemia, hypoinsulinemia, and pancreatic injury.
Pic attenuates the oxidative stress in type 1 diabetic mice To explore the potential mechanisms, we first measured the different parameters of oxidative stress since previous studies have suggested that oxidative stress contributes to the pathogenesis of type 1 diabetes (Sueishi et al., 2017). In STZinduced diabetic mice, malondialdehyde (MDA), which represents the level of lipid peroxidation and is often used as a marker of oxidative stress in vivo, was markedly increased in the liver tissue compared with the control group (Fig. 2(A)). Consistently, the activity of antioxidant enzymes superoxide dismutase (SOD) and glutathione peroxidase (GSH-PX) were both significantly reduced in the DM group but not in the control mice (Figs. 2(B)-2(C)). Pic administration not only significantly down-regulated the MDA level but also increased the activity of SOD and GSH-PX in diabetic mice (Figs. 2(A)-2(C)). Thioredoxin interacting protein (TXNIP), which regulates the antioxidant functions of thioredoxin (Trx), plays a critical role in redox homeostasis and has been identified as an important component linking redox regulation and the pathogenesis of different diseases, including diabetes (Zhou and Chng, 2013;Yoshihara et al., 2014). Our results indicated that the expression of TXNIP was markedly increased in STZ-induced diabetic mice, and Pic treatment completely inhibited the increase (Fig. 2(D)). These results suggest that Pic might exert the anti-diabetic effect via its anti-oxidative properties.
Pic reduces the inflammation and ER stress in type 1 diabetic mice As chronic inflammation and ER stress have also been suggested to be closely involved in the pathogenesis of diabetes, we analyzed whether Pic treatment affected the status of inflammation and ER stress in STZ-induced diabetic mice. In agreement with previous studies, the
transcriptional expression of pro-inflammatory cytokines including TNF-α and IL-6 and ER stress markers including GRP78 and CHOP were all significantly up-regulated in DM mice, while the increase in the expression of these genes were all inhibited by Pic supplementation (Figs. 3

(A)-3(D)).
Pic alters the metabolic profiles of liver tissue in type 1 diabetic mice The results suggested that Pic might affect the metabolism of diabetic mice by targeting oxidative stress, inflammation, and ER stress. We therefore investigated how Pic affected the metabolic homeostasis in mouse liver using NMRbased metabolomics. Typical 1 H NMR spectra of liver extracts from four groups of mice were showed in Fig. 4, and 30 metabolites were assigned (Fig. 4). The detailed information was presented in Tab. 2. The 1 H NMR data from four groups of mice were evaluated using OPLSDA analysis. The results indicated that the DM and DM + Pic groups were clearly separated in the OPLSDA score plot (Fig. 4A). According to the OPLSDA S-plot and colorcoded loadings plots (Fig. 4), alanine, glutamate, taurine, inosine, histidine, oxypurinol were markedly increased in STZ-induced diabetic mice, while glucose was significantly decreased (Fig. 4, Tab. 2). Pic supplementation reversed the changes of alanine, taurine, inosine, histidine, and fumarate, but had no effect on the changes of succinate and creatine (Tab. 2).

Discussion
Pic is a natural analog and a metabolite of resveratrol. In recent years, the function and effect of resveratrol have caused intensive controversies. Although tons of studies have provided compelling evidence that resveratrol exhibits beneficial effects in various animal disease models, including the prevention of cardiovascular diseases, metabolic syndrome and diabetes, etc. (Cao et al., 2018;Lee et al., 2015;Simas et al., 2017;Yang and Kang, 2018), the argument arises from the fact that it is difficult to translate these findings to clinical use for treating human diseases (Smoliga et al., 2011). The failure in the clinical translation of resveratrol has been largely attributed to its poor bioavailability and metabolism. Pic, with improved bioavailability and metabolism properties than resveratrol (Piotrowska et al., 2012;Minakawa et al., 2012), might overcome this issue. Hence, it would be interesting to investigate the potential health-promoting and diseaselimiting capabilities of Pic in different disease models. To the best of our knowledge, the effect of Pic in STZ-induced type 1 diabetic mice has not been reported yet. This study presents evidence indicating the beneficial effects of Pic in a mouse model of T1DM. Supplementation with Pic in T1DM mice efficiently preserved the body weight alleviated the symptoms of polydipsia and hyperglycemia and improved glucose tolerance. The decreased levels of serum insulin in diabetic mice caused by the destruction of pancreatic β-cells were also restored by Pic treatments. It is well known that insulin resistance is the main characteristic of type 2 diabetes mellitus (T2DM), while the destruction of the pancreatic β-cells is the main cause of T1DM, consequently leading to the impairment of insulin production. These results suggest that Pic might exhibit its anti-diabetic effect through protecting the pancreatic islets and thus reserving the function of the pancreas. More importantly, these beneficial effects were achieved when Pic was administered at the advanced diabetic stage with stable hyperglycemia, suggesting the therapeutic potential of Pic in treating T1DM patients. Diabetes is a complicated chronic condition and mediated by multiple factors. Oxidative stress and inflammation are two critical factors mediating the pathogenesis of diabetes and its associated complications (Forbes et al., 2008;Li et al., 2013). During the progression of diabetes mellitus, persistent FIGURE 3. Pic treatment reduces the expression of pro-inflammatory cytokines and ER stress markers in type 1 diabetic mice. The transcriptional expression levels of pro-inflammatory cytokines TNF-α (A) and IL-6 (B), ER stress markers GRP78 (C) and CHOP (D) in liver tissues of all groups of mice. Con: control group; Pic: control mice treated by Pic only; DM: diabetes mellitus group; DM + Pic: diabetic mice treated by Pic. Data are expressed as mean plus SEM. ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001 (unpaired Student's t-test).
hyperglycemia induces the production of free oxygen radicals via several mechanisms, including autoxidation of glucose and non-enzymatic glycation between sugars and proteins to generate intracellular advanced glycation end products (AGEs). These free radicals induce peroxidation and damage to protein, DNA, and lipid, resulting in oxidative stress in different tissues. Regarding the underlying molecular mechanism, our recent study has investigated the interaction of Pic with bovine serum albumin (BSA) using fluorescence spectroscopy, ultraviolet-visible absorption spectroscopy, circular dichroism spectroscopy and molecular simulation (Xu et al., 2019). Results indicated that Pic could inhibit the nonenzymatic glycosylation of BSA to a certain extent; Pic can effectively inhibit the formation of AGEs and protect BSA from undergoing structural changes induced by glycation, suggesting that Pic might exhibit its antidiabetic effect through inhibiting the glycosylation of protein and the formation of AGEs. Accumulating evidence from both experimental and clinical studies demonstrated that oxidative stress plays a critical role in the pathogenesis of T1DM through inducing cellular injury and chronic inflammation (Kobayashi and Schmid-Schönbein, 2006). Several studies have reported that hyperglycemia and hyperlipidemia contribute to the accumulation of ROS and antioxidants deficiency in diabetic animals (Ma et al., 2018;Domingueti et al., 2016).
Our results indicate that Pic could not only substantially alleviate the oxidative stress, but also markedly inhibit the expression of inflammatory molecules in STZ-induced diabetic mice, that is consistent with previous studies using different animal disease models by our and others' groups . The health-promoting and diseasepreventing ability of resveratrol and its analogs have previously been attributed to their antioxidant activity. However, Prysyazhna et al. (2019) recently reported that resveratrol exerts its blood pressure-lowering effect in hypertensive mice through directly inducing the oxidation of protein kinase 1α (PKG1α) instead of its antioxidant role (Prysyazhna et al., 2019). Therefore, although the final outcome in our experimental model demonstrated that the liver oxidative stress in diabetic mice was suppressed by Pic supplementation, the underlying molecular action still needs to be further investigated.
In addition to oxidative stress and inflammation, ER stress has been identified as another important factor involved in the pathogenesis of diabetes. The disruption of ER homeostasis may contribute to β-cell dysfunction and diabetes. Many studies have detected the elevated ER stress markers in different animal models of diabetes as well as in diabetes patients (Özcan et al., 2004;Bhatta et al., 2015;Zhong et al., 2012). Hyperglycemia has also been shown to prompt the induction of ER stress (Özcan et al., 2004;Zhong et al., 2012). Treatment with ER stress-inhibiting chemicals could efficiently lead to the normalization of hyperglycemia, improved glucose and insulin tolerance, and enhancement of insulin action in diabetic mice (Liu et al., 2015;Hosoi and Ozawa, 2016). Moreover, increasing evidence demonstrated that ER stress is able to induce oxidative stress and inflammatory responses (Grootjans et al., 2016;Cao et al., 2016;Ochoa et al., 2018), suggesting that ER stress, oxidative stress, and inflammation are closely interconnected, the interventions that regulate the ER stress response offer other opportunities for preventing and treating diabetes. Our previous study found that Pic reduced the expression of ER stress markers in D-GalN/LPS-induced acute liver injury model and could also inhibit the inflammation induced by ER stress-inducing drugs in vitro . Combined with the results in this current study that Pic markedly down-regulated the
expression of ER stress markers in diabetic mice, it suggests that Pic could alleviate ER stress in different disease models and it should take this mechanism into account when investigating the effect of Pic.
As one of the metabolic diseases, T1DM affects the metabolism homeostasis significantly (Abu Bakar Sajak et al., 2017), and this was reflected by our results obtained from NMR-based metabolomics analysis. In STZ-induced fold change (FC) = A group/B group; Color-coded according to the log 2 (FC), red means "increased" and blue means "decreased" concentrations in the former group when two groups were compared. Color bar -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 b p-values were calculated by unpaired Student's t-test. *p < 0.05, **p < 0.01, ***p < 0.001. diabetic mice, the metabolic profiles that were involved in energy metabolic pathway (glucose, succinate, acetate, creatine), amino acid metabolism pathway (isoleucine, leucine, valine, alanine, tyrosine, phenylalanine) and oxidative stress pathway (glutamate, glutamine, betaine, taurine), were markedly disturbed. The increased levels of glucose in STZ-treated mice suggested a lowered glycolysis, while the decreased level of succinate and fumarate, which are the intermediates of the tricarboxylic acid (TCA) cycle, suggested an inhibited TCA cycle, indicating that STZ disturbed the energy homeostasis in vivo (Sun et al., 2017). Betaine and taurine have been found to have anti-oxidative potential and betaine plays an important role in fatty acid metabolism. Hence, the decrease of betaine and taurine in STZ-treated mice suggests that STZ may affect lipid metabolism by regulating the choline metabolic pathway (Nam et al., 2017). However, Pic supplementation attenuated these changes, suggesting that Pic might target on the energy metabolism pathway and redox regulation pathway to restore the metabolism homeostasis disturbed by STZ injection. However, it also has several drawbacks as it did not improve the metabolite levels of leucine and valine, which are involved in protein synthesis pathways. In summary, oral treatment of STZ-induced type 1 diabetic mice with Pic effectively attenuated the symptoms associated with T1DM. The anti-diabetic effects of Pic might be associated with its activities in inhibiting ER stress, inflammation, and oxidative stress. Our study provides evidence that treatment with Pic is a potentially useful strategy for the treatment of diabetes. Considering the improved bioavailability properties of Pic than resveratrol, it would be interesting to investigate the potential use of Pic for treating TIDM, although further studies regarding the underlying mechanisms are still required.