Berberine protects against diabetic kidney disease via promoting PGC‐1α‐regulated mitochondrial energy homeostasis

Background and Purpose Disordered lipid metabolism and disturbed mitochondrial bioenergetics play pivotal roles in the initiation and development of diabetic kidney disease (DKD). Berberine is a plant alkaloid, used in Chinese herbal medicine. It has multiple therapeutic actions on diabetes mellitus and its complications, including regulation of glucose and lipid metabolism, improvement of insulin sensitivity, and alleviation of oxidative damage. Here, we investigated the reno‐protective effects of berberine. Experimental Approach We used samples from DKD patients and experiments with models of DKD (db/db mice) and cultured podocytes, to characterize energy metabolism profiles using metabolomics. Molecular targets and mechanisms involved in the regulation of mitochondrial function and bioenergetics by berberine were investigated, along with its effects on metabolic alterations in DKD mice. Key Results Metabolomic analysis suggested altered mitochondrial fuel usage and generalized mitochondrial dysfunction in patients with DKD. In db/db mice, berberine treatment reversed the disordered metabolism, podocyte damage and glomerulosclerosis. Lipid accumulation, excessive generation of mitochondrial ROS, mitochondrial dysfunction, and deficient fatty acid oxidation in DKD mouse models and in cultured podocytes were suppressed by berberine. These protective effects of berberine were accompanied by activation of the peroxisome proliferator‐activated receptor γ coactivator‐1α (PGC‐1α) signalling pathway, which promoted mitochondrial energy homeostasis and fatty acid oxidation in podocytes. Conclusion and Implications PGC‐1α‐mediated mitochondrial bioenergetics could play a key role in lipid disorder‐induced podocyte damage and development of DKD in mice. Restoration of PGC‐1α activity and the energy homeostasis by berberine might be a potential therapeutic strategy against DKD.

Kidney cells have high demands for energy to maintain their normal functions. The energy requirements of these cells are primarily satisfied by ATP generated via oxidative phosphorylation (OXPHOS) and fatty acid oxidation (FAO) contributes about 70% of the total supply (Vega, Horton, & Kelly, 2015). As the major power sources in kidney cells, mitochondria work through a set of carefully controlled gene regulation circuits (Bhargava & Schnellmann, 2017;Hock & Kralli, 2009). The peroxisome proliferator-activated receptor (PPAR) γ coactivator-1α (PGC-1α) is considered to be a crucial, upstream transcriptional regulator of mitochondrial biogenesis and function (Handschin & Spiegelman, 2006;Scarpulla, 2011). This role has been demonstrated in several gain-and loss-of-function experimental studies. For example, mice lacking PGC-1α displayed a significant reduction in oxidative metabolism and mitochondrial content (Leone et al., 2005;Tran et al., 2016). In contrast, transgenic overexpression of PGC-1α or drug-stimulated increase of its activity could promote mitochondrial biogenesis and FAO, increase the expression of mitochondrial genes, and inhibit kidney fibrosis and podocyte injury (Han et al., 2017;Lehman et al., 2000;Zhao et al., 2016).
Decreased PGC-1α expression and consequent defects in mitochondrial function directly threaten cell viability, leading to cell apoptosis and dedifferentiation, thereby contributing to various metabolic diseases including diabetes, renal failure, and cardiovascular diseases (Finck & Kelly, 2006;Youle & van der Bliek, 2012). Dysfunctional mitochondria and defective FAO have been described in DKD patients and animal models (Kang et al., 2015;Li & Susztak, 2018;Mootha et al., 2003;Sharma et al., 2013). Podocytes are glomerular cells that constitute the last filtration barrier to restrict the leakage of protein into urine. Mitochondrial OXPHOS is the energy source for the central cell body of podocytes, and they mainly rely on free fatty acids (FFA) as their primary fuel source (Abe et al., 2010). However, podocytes are extremely susceptible to high levels of FFA. Enhanced FFA uptake together with a reduction in FAO and in turn intracellular lipid accumulation are detrimental to podocytes, resulting in the overproduction of mitochondrial reactive oxygen species (mitoROS), imbalance of mitochondrial dynamics and bioenergetics (Imasawa & Rossignol, 2013;Mayrhofer et al., 2009). Therefore, the search for new compounds that would enhance FAO and protect mitochondrial function, in order to reduce lipid accumulation and metabolic disorders has become increasingly important.
Many strategies and drugs with hypolipidaemic and antidiabetic effects have been shown to increase FAO by targeting the transcription of PGC-1α (Ginsberg et al., 2010;Guo et al., 2015;Hong et al., 2014;Yuan et al., 2012). Among these, the plant-derived alkaloid, berberine, has attracted much attention. Particularly, berberine can regulate energy metabolism by targeting PGC-1α and this alkaloid exerted therapeutic effects in an AMP-activated protein kinase (AMPK)-dependent manner . Here, we have explored the defective FAO and mitochondrial dysfunction mediated by dysregulated PGC-1α, in patients and animal models with DKD. We have specifically investigated the molecular mechanisms by which berberine potently restored disturbed energy metabolism in cultured podocytes.

| RNA interference
PGC1α siRNA and control siRNA were provided by Ruibobio (Guangzhou, China) and transfected into podocytes using Liposomal Transfection Reagent (Hanheng, China) according to the manufacturer's protocol.

| Animals
All animal care and experimental procedures conformed to the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Committee for Animal Research of Huazhong University of Science and Technology (Wuhan, China). Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010) and with the recommendations made by the British Journal of Pharmacology.
Male C57BLKS/J db/db diabetic mice and their non-diabetic littermates (7 weeks old) were purchased from the Model Animal Research Center of Nanjing University. Mice were socially housed (2-3 mice per What is already known • Diabetic kidney disease is one of most serious and common complications of diabetes mellitus.
• There is no effective therapy for this disease at present.

What this study adds
• Metabolic changes associated with diabetic kidney disease in patients, animal and cellular models were identified • Berberine exerted therapeutic effects on the metabolic alterations in the progression of diabetic kidney disease.

What is the clinical significance
• Berberine might provide a new approach to the treatment of diabetic kidney disease. cage) at a constant temperature of 22 C ± 2 C, 40-60% humidity and a 12:12 hr light/dark cycle. All animals were maintained on a normal chow diet with free access to water. The db/db mice were used as DKD models (Sharma, McCue, & Dunn, 2003) and randomly separated into three groups: db/db + vehicle, db/db + 200 mgÁkg −1 Áday −1 of berberine (BBRL), and db/db + 300 mgÁkg −1 Áday −1 of berberine (BBRH), with 10 mice per group. The dose of berberine was based on our previous animal studies and other research (Dong et al., 2016;Qin et al., 2019;Zhou & Zhou, 2010). Intragastric administration of berberine or vehicle was started at 7 weeks of age and maintained for 8 weeks. Body weight and blood glucose were monitored every week. At the end of the intervention, 24-hr urine was collected and tested for microalbumin excretion. Glucose tolerance test (GTT) and insulin tolerance test (ITT) were performed after a 6-hr fast, as reported earlier . The blood glucose concentrations were measured in venous blood from the tail at 0, 15, 30, 60, 90, and 120 min after i.p. injection of glucose at 1 gÁkg −1 or i.p. injection of insulin at 1 unitÁkg −1 , respectively. Mice were killed (overdose of pentobarbital, i.p.) and kidneys removed. Kidney samples were either fixed in 4% paraformaldehyde overnight or preserved at −80 C. Mouse glomeruli and podocyte isolation were prepared according to the protocol reported previously . At least five mice per group were examined for assessment.

| Metabolomics
The clinical experiments were approved and supervised by the ethical review board of Tongji Medical College (#2016S192) and conformed to the international standards (US Federal Policy for the Protection of Human Subjects). The inclusion and exclusion criteria for DKD and healthy participants were in accordance with the guidelines of KDOQI (Li et al., 2017;Nelson et al., 2012). Fasting blood samples from participants were collected in EDTA-anticoagulated tubes. The separated plasma was stored at −80 C until assay. Extraction of metabolites from plasma was performed according to the previously described method followed by a two-step derivatization. The samples were then analysed on a TriPlus-RSH autosampler-Trace1300-ISQ GC-MS instrument (West Palm Beach, FL, USA). Parameters for sample injection, temperature programming, and MS were set according to the standard procedure (Lopez-Bascon, Priego-Capote, Peralbo-Molina, Calderon-Santiago, & Luque de Castro, 2016). Compound identification was performed using the Metabodetector software based on retention index and NIST library (version 11, 2011) searching (Hiller et al., 2009). SIMCA-P+ software (Version 13.0, Umetrics, Umea, Sweden) was used for multivariate projection modelling and plotting.

| Oil Red O (ORO) staining
Kidney cryosections and podocytes were washed in 60% isopropanol and then incubated with 0.5% ORO solution (Sigma-Aldrich) for 60 min. Samples were washed with isopropanol for 5 s and imaged.

| Mitochondrial function assays
Podocyte mitochondria were isolated using a Mitochondria Isolation Kit (Thermo Fisher Scientific, USA). Respiratory chain complex activities, the amounts of ATP, ADP, NAD and NADH were measured using commercial kits in accordance with manufacturer's protocols.

| mitoROS measurement
The mitoROS levels were measured using the MitoSOX Red Superoxide Indicator (Invitrogen, M36008) for 20 min at 37 C after drug intervention. Details were in accordance with the protocols (Robinson, Janes, & Beckman, 2008).

| Western blotting
The antibody-based procedures used in this study comply with the recommendations made by the British Journal of Pharmacology . Western blotting was conducted as described previously (Qin et al., 2019). Antibodies used were as follows: AMPK The bands were quantified using Image J software (NIH). Original scan images are available in Supporting Information.

| qRT-PCR analyses
The RNA extraction, reverse transcription, and qPCR analysis were conducted as previously described (Qin et al., 2019). For mitochondrial DNA (mtDNA) copy number, cytochrome c oxidase subunit 4I1 (Cox4i1) was used as mtDNA marker. The relative quantity was analysed using the 2 −ΔΔCT method. Sequences of primers used are presented in Table S1. and anti-goat antibodies were used for immunofluorescence. More than 10 fields from each sample were captured for assessment.

| Histological analysis
For Masson's trichrome staining and periodic acid-silver metheramine (PASM), samples were fixed in 4% paraformaldehyde overnight, and the sections were stained by Biossci Biotechnology, China.

| Electron microscopy
Mitochondrial morphology in cultured podocytes by transmission electron microscopy (TEM) was examined as previously described (Qin et al., 2019). Conventional scanning electron microscopic (SEM) examination for mouse glomeruli was conducted according to standard protocols. In brief, small cubes of kidney cortex were fixed in 2.5% glutaraldehyde solution and immersed in 2% osmium tetroxide for 2 hr. After dehydration with gradient ethanol, samples were freeze-dried, mounted on aluminium stubs, coated with gold and then viewed with a scanning electron microscope (Hitach su8010, Japan).

| Data and statistical analysis
The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). Data are presented as means ± SEM and tested for normality via Shapiro-Wilk test, which were then tested for homogeneity of variance via one-way ANOVA.
Comparisons between two groups were performed with the Student's t test. And one-way ANOVA with Tukey's test was conducted for comparisons among multiple groups. Post hoc tests were run only if F achieved P < .05, and there was no significant variance inhomogeneity. P values < .05 were considered statistically significant, and all tests were two-tailed. The declared group size is the number of independent values, and the statistical analysis was done using these independent values. When outliers were included or excluded in analysis, this is stated within the figure legend. Statistics were analysed using the GraphPad Prism 6 software.

| Materials
The compounds used in these experiments were supplied as follows:   (j) TG content in cultured podocytes treated with PA or BBR. Data were represented as mean ± SEM. *P < .05, significant effect of berberine or as indicated. AER, albumin excretion rate; BBRH, db/db mice treated with higher dose of BBR; BBRL, db/db mice treated with lower dose of BBR; FFA, free fatty acids; PA, palmitic acid; TG, triglyceride 3.2 | Berberine reduced overproduction of ROS and podocyte injury, induced by lipid overload Accumulation of FFA induces excessive ROS generation, which could then lead to lipid peroxidation, podocyte damage, and glomerulopathy. We first examined markers of podocytes, the slit diaphragm proteins (SDs) in glomeruli of DKD patients. As shown in  (Qin et al., 2019). Because ROS are mainly generated in the mitochondria during the process of electron transport, we assayed mtROS in PA-induced podocytes and found mtROS production was markedly increased by PA and reduced after berberine treatment ( Figure 2i). Together, these data showed that berberine could protect glomerular podocytes from FFA-induced oxidative damage.

| Berberine rescued mitochondrial function and improved FAO
Increased ROS production is strongly associated with mitochondrial dysfunction, impaired mitochondrial OXPHOS, insufficient FAO, and disturbed energy homeostasis in DKD (Bonnard et al., 2008;Imasawa & Rossignol, 2013). In our case-control study, plasma metabolite profiles from DKD patients and healthy controls were measured using gas chromatography-MS (GC/MS). Clinical and biochemical characteristics of the participants are shown in Table S2. DKD patients have higher levels of fasting blood glucose (FBG), haemoglobin A1c (HbA1c), AER, and lower high-density lipoproteins-cholesterol (HDL-C) than healthy controls. The orthogonal partial least squares discriminant analysis (OPLS-DA) scatter plot ( Figure S1) and heatmap analysis ( Figure 3a) were conducted to visualize the metabolic profiles and showed a distinct segregation among the two groups. The relative abundance of plasma metabolites is shown in Table 1. A total of 106 metabolites were detected, and 27 metabolites were identified to be significantly different between cases and controls, according to the VIP analysis and t test analysis. Particularly, the DKD group showed higher levels of long chain fatty acids, uric acid, citric acid, succinate, and lower levels of serine, glycine, and certain tricarboxylic acid cycle (TCA) intermediates (malate, cis-aconitate, fumarate, and isocitric acid). In addition, the DKD patients had significantly decreased levels of 3-hydroxybutyrate, acetoacetate, and acetone, which indicated impaired FAO. Alterations in metabolic pathways correlated with the progression of DKD are summarized in Figure 3b,c. Most of these metabolites can be broadly classified into three pathways: lipid metabolism and fatty acid β-oxidation, amino acid metabolism, and carbohydrate metabolism. Particularly, abnormal levels of mitochondrial metabolites involved in the TCA cycle was observed. All these data suggested altered mitochondrial fuel usage and generalized mitochondrial dysfunction in DKD patients.
Given these abnormalities in mitochondrial metabolic markers, we next assessed mitochondrial energy metabolism in mouse glomeruli and cultured podocytes. Key genes responsible for mitochondrial FAO and the activity of enzymes related with OXPHOS and TCA cycle were screened. As shown in Figure 3d-i, the expression of genes controlling β-oxidation, the activity of complex I, IV, V involved in OXPHOS and citrate synthase involved in TCA cycle were decreased both in mitochondrial homogenates of kidneys from db/db mice and PA-treated podocytes. Treatment of the mice with berberine normalized these variables and restored the mitochondrial energy homeostasis.

| Berberine promoted mitochondrial biogenesis and increased energy output
The direct consequence of decreased mitochondrial FAO and OXPHOS activity is a reduction in energy output and mitochondrial biogenesis. To elucidate this, kidney tissue sections from DKD patients and controls were analysed with double immunofluorescence of mitochondrial VDAC and the podocyte marker, nephrin. We observed a significant reduction in podocyte numbers and mitochondria contents in glomeruli derived from DKD patients (Figure 4a). Similar results were obtained in samples from db/db mice and these changes were improved by berberine treatment (Figure 4b). Then we determined energy generation in podocytes of db/db mice as reflected by NAD + /NADH ratios. Our results indicated that berberine could improve the NAD + /NADH ratio in diabetic mouse kidneys

| Berberine regulated mitochondrial energy metabolism through AMPK/PGC1α
PGC-1α controls and is controlled by several well-known regulators involved in mitochondrial dynamics and bioenergetics (Li & Susztak, 2018). Given the profound effects of AMPK on energy metabolism, we tested whether berberine could activate AMPK, which then mediates mitochondrial energy homeostasis through activating PGC-1α. As shown in Figure 5 We next ascertained the regulatory relationship between AMPK and PGC-1α by using the inhibitor of AMPK, Compound C. The results showed that protein levels of PGC-1α and phosphorylated ACC were decreased after inhibiting AMPK, and this were reversed by berberine treatment (Figure 6a,b). To further elucidate the contribution of PGC-1α in berberine-regulated mitochondrial function and FAO, PGC-1α siRNA (with scrambled oligonucleotide acting as controls) was used to down-regulate its expression. As shown in

| Berberine inhibited FFA uptake via downregulating the expression of CD36
We next examined proteins associated with FFA uptake and synthesis. The main mediator of the cellular uptake of FFA is CD36 (scavenger receptor B2) and up-regulation of CD36 is known to lead to lipid accumulation, increased ROS production, cell apoptosis, and kidney fibrosis (Hua et al., 2015;Yang et al., 2017). In accordance with these findings, we found that the gene expression of CD36, but not the fatty acid transport protein 1 (FATP1), was up-regulated in PA-  (Li & Susztak, 2018;Long et al., 2016;Zhu et al., 2014). Other studies have suggested that the mRNA level of PGC-1α was significantly decreased in patients with chronic kidney disease (CKD), compared with levels in control subjects (Kang et al., 2015  immunofluorescence. Data shown are means ± SEM. *P < .05, significantly different as indicated. ACC, acetyl-CoA carboxylase; BBRH, db/db mice treated with higher dose of BBR; BBRL, db/db mice treated with lower dose of BBR; FATP1, fatty acid transport protein 1; FFA, free fatty acid; NS, not significant; PA, palmitic acid; pACC, phosphorylated acetyl-CoA carboxylase up-regulate the expression of PGC-1α, thereby tipping the balance of energy metabolism further toward higher FAO to more ATP generation (Lee et al., 2018). In podocytes, PGC-1α activation attenuated mitochondrial dysfunction, restored expression of the SD protein and reduced cell apoptosis (Long et al., 2016;Zhou et al., 2011). Other strategies have been exploited to promote mitochondrial FAO to treat metabolic diseases either via targeting CPT1 or ACC, for example, using PPAR agonists, metformin or certain natural drug treatment, or overexpressing a CPT1 mutant to boost the activities (Dai et al., 2018;Hong et al., 2014;Lee et al., 2018). In agreement with these observations, we demonstrated that the inhibition of AMPK or PGC-1α alone is sufficient to drive a marked reduction in FAO and to increase lipid accumulation. Our data provide further support for promoting PGC-1α activity and FAO-related gene expression, in order to regulate mitochondrial bioenergetics and provide benefits in DKD.
In this regard, the results with berberine showed this alkaloid to be a naturally occurring, direct activator of energy regulators, thus providing a unique pharmacological tool to promote mitochondrial energy output and FAO, specifically in kidney podocytes. Studies have previously attempted to explore the mechanisms of hypolipidemic and hypoglycaemic activity of berberine, and multiple targets and signalling pathways were observed to have transcriptional or posttranslational alterations (Ni, Ding, & Tang, 2015;Zhang et al., 2014).
For instance, berberine directly stimulated the AMPK signalling pathway in fat, liver and kidney cells (Ni et al., 2015. Activated AMPK could promote the phosphorylation of ACC to suppress the expression of lipid synthesis genes in cells, which significantly decreased TG and cholesterol concentration (Lee et al., 2018).
Besides, berberine appears to regulate the expression of multiple genes associated with energy metabolism in mitochondria. Berberine can be as a direct activator of PGC-1α, whose up-regulation could induce the expression of mitochondrial and FAO and thermogenic genes . Combined with previous research, our findings have strengthened the evidence for berberine as a regulator of mitochondrial function, dynamics and energy metabolism, by upregulating the level of PGC-1α. These changes brought about by berberine treatment could protect renal cells from lipotoxicity, thereby decreasing extracellular matrix accumulation, alleviating glomerular sclerosis, and improving the clinical symptoms of DKD.
However, our study has some limitations. One is that failure to compare the metabolic profiles before and after berberine intervention in vivo precluded us from characterizing the role of berberine in mitochondrial function and energy metabolism, as well as other aspects of berberine's bioactivity. Another limitation is that our research could not answer whether berberine treatment could bring benefits to DKD patients, including protecting against podocyte damage, decreasing AER and restoring the GFB. These questions are of great relevance to its possible use in treating DKD in patients. We are optimistic that berberine has high potential for further exploitation because similar effects of berberine have been observed in patients with metabolic disorders such as DM. Further clinical trials and experimental research are needed to resolve these problems.
In summary, we believe that there may exist other transcription factors regulating FAO and cellular bioenergetics. Here, we have demonstrated that PGC-1α is essential for the comprehensive regulation of energy homeostasis and might be a key target in the actions of berberine. Our present findings have revealed the important role of berberine in regulating kidney energy homeostasis and protecting glomerular podocytes from metabolic stress, and have identified berberine as a promising drug for the treatment of DKD, in the future.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

RIGOUR
This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis, Immunoblotting and Immunochemistry, and Animal Experimentation, and as recommended by funding agencies, publishers and other organisations engaged with supporting research.