Inhibition of Mitochondrial Fission by Drp-1 Blockade by Short-Term Leptin and Mdvi-1 Treatment Improves White Adipose Tissue Abnormalities in Obesity and Diabetes

Background Obesity and type 2 diabetes are chronic diseases characterized by insulin resistance, mitochondrial dysfunction and morphology abnormalities. Objective Herein, we investigated if dysregulation of mitochondrial dynamics and biogenesis is involved in an animal model of obesity and diabetes. Methods The effect of short-term leptin and mdivi-1 –a selective inhibitor of Drp-1 fission-protein– treatment on mitochondrial dynamics and biogenesis was evaluated in epididymal white adipose tissue (WAT) from male ob/ob mice. Results An increase in Drp-1 protein levels and a decrease in Mfn2 and OPA-1 protein expression were observed with enhanced and sustained mitochondrial fragmentation in ob/ob mice compared to wt C57BL/6 animals (p<0.05). The content of mitochondrial DNA and mRNA expression of PGC-1α –both parameters of mitochondrial biogenesis– were reduced in ob/ob mice (p<0.05). Leptin and mdivi-1 treatment significantly increased mitochondrial biogenesis, improved fusion-to-fission balance and attenuated mitochondrial dysfunction, thus inducing white-to-beige adipocyte transdifferentiation. Measurements of glucose and lipid oxidation in adipocytes revealed that both leptin and mdivi-1 increase substrates oxidation while in vivo determination of blood glucose concentration showed decreased levels by 50% in ob/ob mice, almost to the wt level. Conclusions Pharmacological targeting of Drp-1 fission protein may be a potential novel therapeutic tool for obesity and type 2 diabetes.


INTRODUCTION
Obesity and type 2 diabetes are chronic diseases that typically coexist and have grown in prevalence to become a global health problem. Both of them share a pathophysiological axis of insulin resistance (IR), oxidative stress, mitochondrial dysfunction and chronic inflammation. Obesity is characterized by an abnormal increase in white adipose tissue (WAT) mass depicting an imbalance between energy intake and consumption that leads to energy overload (1,2).
In mammals, there are three types of adipose tissue: white, brown and beige, with distinct functions and morphology, different protein expression patterns, and dissimilar developmental origins (3,4). While the function of WAT is to store energy in the form of lipid droplets that can be released to fuel other tissues, brown adipose tissue (BAT) has thermogenic properties for maintaining body temperature (5)(6)(7). Browning is the process by which some adipocytes within WAT depots acquire properties of brown adipocytes (an increase in mitochondrial content and oxidative metabolism) showing an intermediate phenotype between white and brown adipocytes ("beige" or "brite") (8).
Leptin, the satiety and anti-obesity hormone, is an adipokine released from white fat tissue that acts on the hypothalamus to induce satiety and reduce food intake, and on the peripheral tissues to control metabolism and to increase the metabolic rate (9).
Ob/ob mice suffer a mutation that determines a premature stop codon in the Lep gene (leptin) and constitute a well-known model of obesity and type 2 diabetes characterized by hyperglycemia, hyperinsulinemia, high food intake, mitochondrial dysfunction, and oxidative/nitrosative stress affecting mitochondrial complex I activity in WAT, liver and skeletal muscle (10)(11).
Mitochondria are dynamic organelles essential for cellular energy survival and a major site of ROS production. These organelles, whose morphology has a remarkable plasticity according to the energetic cell needs, undergo constant mitochondrial fusion and fission, and are replaced every 2-4 weeks in different tissues through mitophagy. All of these processes are necessary for mitochondrial quality control and clearance (12,13). Energetic homeostasis is sustained through the balance among mitochondrial biogenesis, dynamics, ultrastructure, function and degradation in a healthy mitochondrial population (14). While mitochondrial fusion is mediated by the mitofusin proteins (Mfn) 1 and 2 in the outer membrane and by OPA-1 (optic atrophy-1) in the inner membrane, mitochondrial fission requires the translocation of Drp-1, a member of the large GTPases dynamic family, from the cytosol to the mitochondrial outer membrane. When the cytosolic protein Drp-1 is activated, it translocates to the outer membrane of the mitochondrion where it multimerizes generating a ring-like structure that constricts and divides this organelle.
Post-translational modifications of the protein contribute to the regulation of mitochondrial fission mainly by phosphorylation of serine residues that increase or decrease its GTPase activity (15). Drp-1 activation is rapidly regulated by the phosphorylation of serine 616 and dephosphorylation of serine 637, which are targeted by different phosphatases and kinases. While phosphorylation of Drp-1 on Ser-637 prevents its mitochondrial translocation, phosphorylation of Drp-1 on Ser-616 promotes mitochondrial fragmentation during mitosis (16).
Mdivi-1 is a quinazolinone that selectively inhibits Drp-1 over other dynamin family members and prevents mitochondrial fission by pharmacological inhibition. This drug inhibits Drp-1 self-assembly into rings and its association with mitochondria (17-18).
Mitochondrial biogenesis is the process through which new mitochondria are generated, driven by the transcriptional activators NRF-1 and 2 and by PGC-1 alpha. It is activated by various signaling pathways, including nitric oxide (NO)/cyclic GMP, Akt and AMPK, among others (19). The AMP-activated protein kinase (AMPK) is a cellular energy sensor that when activated by phosphorylation leads to upregulation of PGC-1α, UCP-1, and mitochondrial biogenesis (20).
Alterations of mitochondrial dynamics and function have been implicated in neurodegeneration, aging, sepsis, cardiovascular disease, obesity and type 2 diabetes (21)(22)(23)(24). In this context, the twofold aim of the present study is to investigate if dysregulation of mitochondrial dynamics and biogenesis are involved in a relevant animal model of obesity and diabetes, and if Drp-1 can be pharmacologically targeted as a potential novel therapeutic tool for obesity and diabetes.

Animals experimentation.
Male ob/ob (± 70 g) and wt C57BL/6 (± 25 g) 5-month-old mice were purchased from Jackson Labs (USA). Animal experiments were performed in accordance with the Principles of Laboratory Animal Care. The animal experiments were approved by the local Scientific and Technology Ethics Committee at the University of Buenos Aires (UBA). All efforts were made to minimize animal suffering and to reduce the number of animals used. Mice were maintained under controlled temperature (21º C +/-2º C), humidity (50-60%), and air-flow conditions, with a fixed 12-h light/dark cycle.
Until the experiment onset, all animals were fed with a standard mice laboratory chow and had free access to food and water to standardize their nutritional status. Animals were sacrificed by cervical dislocation at the end of the treatment with leptin or mdivi-1.

Blood glucose determination.
On the day of sacrifice, blood samples were collected using cardiac puncture. Blood glucose concentration was determined by Accu-check performance glucometer (Roche Lab). Animals were fasted for 10h before any procedure.
White adipose tissue extraction. Mice were sacrificed by cervical dislocation and epididymal white adipose tissue (6-10 g) was immediately extracted and homogenized in sucrose buffer 20% (TRIS 10 mM, EDTA 0.1mM, sucrose 20%, 2% protease inhibitor cocktail) (Sigma Aldrich, St. Louis, MO, USA). The homogenate was centrifuged at 800g for 10 min at 4°C, frozen in liquid nitrogen and stored at 80ºC until further analysis processed with this pipeline to identify and measure more than 300 mitochondria. All morphometric measurements of mitochondria were exported to a standard database and were further analysed with Cell Profiler Analyst (Broad Institute, USA). A standard random forest classification algorithm was used to implement a machine learning approach to classify mitochondria intro three groups: round, elongated and branched. The training procedure was repeated until global classification accuracy achieved 90%, with 100% classification accuracy of the round mitochondria subgroup. Then, the whole set was scored for the amount of mitochondria belonging to each of the three groups (26-28).
Adipocytes siRNA transfection. Adipocytes were transfected with siRNAsDrp-1 or empty-vector siRNA (Santa Cruz Biotechnology) 50nM using lipofectamine (Invitrogen Corp. California, USA) in Opti-MEM reduced serum medium and were incubated at 37ºC in 5% CO 2 for 10 hours, according to the protocol provided by the manufacturer (11).
The tubes were sealed and incubated for 2 h; then, 10 N HCl was added to release 14 (29).

Statistical analysis
Data are presented as mean ± SEM according to the normal or skewed distribution.
Significant differences between groups were accepted at p < 0.05. One-way ANOVA (multigroup comparisons) followed by Bonferroni's multiple comparison test or Dunnett's test were carried out with GraphPad Prism 5.01 (La Jolla, CA).

Short-term leptin and mdivi-1 treatment reverts mitochondrial morphological defects in WAT from ob/ob mice
Mitochondrial morphology defects were detected by TEM in WAT slices in ob/ob mice in comparison to wt C57BL/6 and ob/ob treated with leptin or mdivi-1 mice. While wt mitochondria were predominantly tubular-shaped (80%), ob/ob mitochondria were aberrantly small and spherical in shape (85%) (Fig. 1 A and B). Leptin or mdivi-1 treatment resulted in a significant increase in tubular-shaped mitochondria (55% and 49%, respectively) ( Fig. 1 C, D and E). As mitochondrial fragmentation and dysfunction lead to mitophagy, we analyzed WAT slices by TEM and observed mitochondria engulfed by double-membrane vacuoles with mitochondrial fragments indicating an active process of mitophagy in ob/ob mice (10 images per field). These alterations were reduced after treatment with leptin (5 images per fields) or mdivi-1 (3 images per field) ( Fig. 1F). We also assessed the LC3 II protein levels, which increased in ob/ob mice and decreased with leptin or mdivi-1 treatment (Fig.1 G). No changes were observed in wt C57BL/6.

Short-term leptin and mdvi-1 treatment reverts mitochondrial morphological defects in isolated adipocytes from WAT
We investigated mitochondria morphological changes on isolated adipocytes from WAT wt C57BL/6 or ob/ob mice treated with leptin or mdvi-1 intraperitoneally using a machine learning algorithm to allow automated classification of mitochondria shape into three groups: branched, round, or elongated detected by fluorescence microscopy images (Suppl. Fig. 1 A)."Branched" mitochondria denoted complex shapes including T-shapes, Y-shapes, P-shapes, O-shapes, among others, and have been associated with particular bioenergetic profiles (29). The results of mitochondrial classification into each of the three groups through machine learning is shown in Fig. 2 A

Short-term leptin treatment modulates the mitochondria fission-to-fusion ratio in WAT from ob/ob mice
In order to examine whether leptin can modulate the mitochondria fission in leptindeficient ob/ob mice with mitochondrial dynamics impaired, we analyzed the expression of the mitochondrial fission (Drp-1) and fusion (Mfn2 and OPA1) proteins in WAT with and without leptin treatment. Drp-1 protein level is increased by ~50% in the WAT of ob/ob mice compared to wild type animals (Fig. 3A). Drp1 activity is regulated by the opposing effects of phosphorylation at two key serine residues: phosphorylation of Ser 616 increases Drp-1 activity whereas it is decreased by phosphorylation of Ser 637 (30).
We measured the phosphorylation of both residues and found an increase in p-Ser 616-Drp-1 (active form) whereas p-Ser 637-Drp-1 (inactive form) decreased in ob/ob mice ( Fig. 3 B and C). Administration of leptin reverted the changes in Drp1 expression and phosphorylation forms in ob/ob mice stimulating the Drp-1 inactive form (Figs. 3 A-C).
In addition, we observed a reduction in Mfn2 and OPA-1 protein levels in ob/ob, which were restored after leptin treatment (Fig. 3 D and E). No changes were observed in wt C57BL/6 under any treatment.

Leptin and mdvi-1 increase the expression of PGC-1α and AMPK-P as well the mitochondrial mass in WAT from ob/ob mice
After observing an overexpression of Drp-1 protein in ob/ob mice and to determine the effect of leptin or mdvi-1 treatment in mitochondrial biogenesis, we assessed PGC-1α gene expression. The mRNA level of PGC-1α was significantly lower in ob/ob mice compared to wt C57BL/6, leptin or mdvi-1 treatment (Fig 4 A). mtDNA content showed a reduction in ob/ob mice compared to wt C57BL/6. Inhibition of Drp-1 by mdvi-1 led to increased mtDNA content indistinguishable from leptin (Fig. 4 B). WAT obtained from ob/ob mice revealed that AMPK expression was devoid in its phosphorylated form (active) as compared to wt, while leptin or mdvi-1 treatment increased AMPK-P protein expression ( Fig. 4 C and D).
The activity of the mitochondrial electron transport chain (ETC) complex I was evaluated in mitochondria. An impairment of mitochondrial function was observed in ob/ob (-50%), while leptin, mdvi-1 and siRNADrp-1 caused a 5-fold increase (Fig. 5 C). No changes were found in wt C57BL/6 under any treatment.

Effect of leptin and mdvi-1 in blood glucose level, body weight and food intake
Both leptin and mdvi-1 decreased the blood glucose concentration by 50% in ob/ob mice almost to the wt level. We did not observe any changes in body weight and food intake over the course of the treatments (Table 1).

Leptin and mdvi-1 stimulate the "browning" process in WAT from ob/ob mice
In order to study whether the leptin or mdvi-1 treatments in ob/ob mice stimulate the "browning" process in WAT, we measured the adipocyte area and vascular density and observed that leptin and mdvi-1 treatment reduced the adipocyte area and increased the vascular density ( Fig. 6 A and B). In addition, we determined that mRNA UCP-1 expression was increased by mdvi-1 more than leptin in ob/ob WAT (Fig. 6 C).

DISCUSSION
This study focuses on mitochondrial dynamics and biogenesis as both processes are involved in WAT in obesity and diabetes and their modulation by leptin and mdivi-1.
Mitochondrial fragmentation and mitophagy prevail in ob/ob mice WAT; however, both mdivi-1 and leptin treatment increase the percentage of tubular mitochondria (fusion) and decrease the amount of fragmented mitochondria (fission), disclosing that the defects observed in mitochondrial morphology are linked to fission activation. Mitophagy plays an important part in removing damaged mitochondrial structure, which is facilitated by prior fission (32). We found a high number of images compatible with mitophagy in WAT from ob/ob mice. During autophagosome formation, cytosolic LC3-I is conjugated to phosphatidylethanolamine with the resulting LC3-II localized to autophagosomes (33).
According to that, we observed an increase in LC3 II protein expression. This abnormality in fusion-fission-mitophagy ratio improves with leptin and mdvi-1. To confirm matching changes of mitochondrial morphology in cells, we studied the effect of leptin or mdvi-1 in isolated adipocytes from treated mice, and we observed an increment of longer mitochondria, as a result of increased fusion and decreased fission. The stronger profusion and anti-fission effect of leptin and mdvi-1 could explain the proportional increase of complex shaped mitochondria as observed in the "branched mitochondria" group.
Upon observing the presence of mitochondrial fragmentation and mitophagy in WAT from ob/ob mice and its modulation by leptin and mdvi-1, we studied their effect on mitochondrial dynamics. We observed an increase in the expression of mitochondrial fission-protein level in ob/ob respect to wt WAT (total Drp-1 and Drp1-pSer 616 active isoform), while concomitantly reducing the expression of proteins responsible for mitochondrial fusion (Mfn2 and OPA-1).
The mechanism of controlling mitochondrial biogenesis, promoting mitochondrial DNA replication and integrity, and increasing insulin sensitivity is associated with PGC-1α (34,35). Our data show a decreased mRNA PGC-1α expression and mtDNA/nDNA ratio in ob/ob, while leptin and mdivi-1 induce mitochondrial biogenesis increasing PGC-1α levels and mitochondrial mass. In agreement with our data, evidence supports alterations in mitochondrial dynamics and reduced expression of PGC-1α and Mfn2 resulting in the loss of mitochondrial mass and structure abnormalities in obesity and type 2 diabetes in humans, rats, leptin-deficient ob/ob and high-fat-diet-fed dietary mice with hyperleptinemic state (36)(37)(38).
The short-term leptin hormone replacement in hypoleptinemic state is not only sufficient to normalize the levels of biogenesis and dynamics proteins but also inactivates Drp-1 by Ser 637 phosphorylation.
The accumulation of three major nutrients -glucose, fat acids and aminoacids-in obesity and diabetes with glucose levels suppresses AMPK through mechanisms that do not affect the AMP/ATP ratio and contributes to insulin resistance, oxidative stress and mitochondria dysfunction to impede substrate oxidation and to promote fat accumulation and obesity (39). It has also been reported that AMPK activation indirectly phosphorylates Drp-1 at Ser 637, and this phosphorylation has been linked to the inhibition of Drp-1 and a decrease in mitochondrial fission. In addition, AMPK and aerobic exercise downregulate the phosphorylation level of Drp1 at Ser 616 which is upregulated by ROS. Likewise, AMPK plays a role in mitochondrial homeostasis phosphorylating PGC-1α on specific serine and threonine residues leading to an increased mitochondrial gene expression. AMPK is activated by leptin in the peripheral tissues increasing fatty acid oxidation and glucose uptake. Leptin produces Drp-1 inhibition modulating mitochondrial fission and PGC-1α stimulation favoring mitochondrial biogenesis; both are protective effects mediated by AMPK (40,41). In our study, leptin and mdvi-1 induce AMPK-P supporting the idea that AMPK can work in the context of different cellular signals that are distinguished from energetic demand, to promote a variety of mitochondria function.
Elongated and branched mitochondria have been previously associated with improved oxidative phosphorylation and several bioenergetic advantages, while round mitochondria have been associated with a predominance of oxidative stress and dysfunction (42). In the present study, we detected decreased complex I activity and oxidative metabolism in ob/ob adipocytes, counteracted by leptin, mdvi-1 and siRNAs Drp-1, which increased complex I activity followed by an increment of glucose and palmitic acid oxidation, depicting that mitochondrial function is closely related to Drp-1 inhibition. In accordance with the metabolic and mitochondrial effects, leptin and mdvi-1 treatments produce a significant decrease in blood glucose levels in ob/ob mice which typically exhibit high levels of blood glucose concentration. Normalization of blood glucose levels were not accompanied by a decrease in body weight and food intake, suggesting that blood glucose concentration was related to increased glucose oxidative metabolism secondary to changes in mitochondrial homeostasis.
"Browning" or "beiging" WAT process contributes to high metabolic rates and a high level of mitochondrial content in these cells. UCP-1-mediated thermogenesis is a hallmark of brown/beige adipocytes development (43). According to numerous cold exposure studies in mice, subcutaneous adipose tissue depots are more susceptible to browning than the metabolically unfavorable visceral fat. However, it has been reported that in obese subjects, omental (visceral) fat expressed higher transcript levels of browning markers and genes involved in mitochondrogenesis compared to abdominal subcutaneous fat (44). We studied whether epididymal white adipose tissue would be capable of undergoing the browning process. The adipocytes termed "beige" or "brite" Leptin production by adipocytes is critical for energy homeostasis activating AMPK in peripheral tissues and is the promoter of BAT thermogenesis in both UCP-1-mediated and independent thermogenesis mechanisms (47). Leptin acute and chronic effects have been studied in different models and pathological situations. While short-term leptin exposure produces lipolytic effects, a lipogenic effect has been observed with leptin chronic stimulus in mature adipocytes (48). Some studies reported that leptin directly stimulated lipolysis in adipocytes isolated from fasted wild type or ob/ob mice (49). There is also evidence that lipolysis inhibition in adipose tissue under longer periods of leptin exposure reduced endothelial cell expression of PPARγ, while chronic leptin treatment restored PPARγ expression in mice (50). We hypothesize that in mature adipocytes shortterm leptin and mdvi-1 treatment directly induces lipolysis inhibiting PPARγ expression.
In summary, mitochondrial dysfunction has been related with obesity and type 2 diabetes,

body weight and food intake
Blood glucose concentration (mg/dl) was measured in all studied groups. Body weight (g) and food intake (g/d) were determined in all animals during the study. Values are means ± SD of 8 mice per group *p<.0.5 wt C57BL/6 vs ob/ob, #p < 0.05 ob/ob vs ob + leptin or mdvi-1. One-way analysis of variance (ANOVA) and Bonferroni post hoc.