WWOX somatic ablation in skeletal muscles alters glucose metabolism.

OBJECTIVE: WWOX, a well-established tumor suppressor, is frequently lost in cancer and plays important roles in DNA damage response and cellular metabolism.
METHODS: We re-analyzed several genome-wide association studies (GWAS) using the Type 2 Diabetes Knowledge Portal website to uncover WWOX's association with metabolic syndrome (MetS). Using several engineered mouse models, we studied the effect of somatic WWOX loss on glucose homeostasis.
RESULTS: Several WWOX variants were found to be strongly associated with MetS disorders. In mouse models, somatic ablation of Wwox in skeletal muscle (WwoxΔSKM) results in weight gain, glucose intolerance, and insulin resistance. Furthermore, WwoxΔSKM mice display reduced amounts of slow-twitch fibers, decreased mitochondrial quantity and activity, and lower glucose oxidation levels. Mechanistically, we found that WWOX physically interacts with the cellular energy sensor AMP-activated protein kinase (AMPK) and that its loss is associated with impaired activation of AMPK, and with significant accumulation of the hypoxia inducible factor 1 alpha (HIF1α) in SKM.
CONCLUSIONS: Our studies uncover an unforeseen role of the tumor suppressor WWOX in whole-body glucose homeostasis and highlight the intimate relationship between cancer progression and metabolic disorders, particularly obesity and type-2 diabetes. SUBJECT AREAS: Genetics, Metabolic Syndrome, Diabetes.

Introduction

Metabolic syndrome (MetS) is a major health problem, with an average worldwide prevalence of 25% among adults [1]. MetS encompasses a cluster of disorders including obesity, high blood pressure, and high fasting glucose levels that are caused by insulin resistance in peripheral tissues, mainly in skeletal muscles (SKM) and liver [2]. MetS is associated with fatal cardiovascular diseases and the development of type-2 diabetes mellitus (T2D) [2].

Comprising 40%–50% of the body mass, SKM represent the largest body organ. SKM are responsible for 85% of all insulin-mediated glucose clearance [3], highlighting their importance in maintaining whole-body glucose homeostasis [4]. SKM are composed of two types of fibers: oxidative slow-twitch fibers and glycolytic fast-twitch fibers, although muscle fibers are flexible and can transform from one state into another in response to energetic demands [5]. Notably, the SKM of T2D and obese patients have been shown to display a decreased oxidative capacity [6], by transitioning from oxidative slow-twitch muscle fibers into glycolytic fast-twitch fibers [7]. The factors that regulate this transition and its dynamics might modulate cellular metabolism, whereas their perturbations could lead to the development of metabolic disorders.

WWOX is a tumor suppressor that is commonly lost in several human malignancies [8]; its overexpression inhibits tumor growth and enhances apoptosis [9], [10]. WWOX encodes a 46 kDa protein that contains two N-terminal WW domains, known to mediate protein–protein interactions, and a central short-chain dehydrogenase/reductase (SDR) domain [10], whose function is still unknown. WWOX pleotropic tumor suppressor functions include promoting apoptosis and DNA repair and antagonizing aerobic glycolysis [11].

Several reports have implicated WWOX function in cellular metabolism [12], [13], [14]. In a previous work, we showed that WWOX regulates glucose metabolism in tissue culture cells via suppressing hypoxia-inducible factor 1-alpha (HIF1α) [15]. WWOX physically interacts with HIF1α and suppresses its activity, leading to activation of oxidative phosphorylation (OXPHOS) and inhibition of glycolysis to maintain a balanced cellular glucose metabolism [16]. Wwox null mice die by the age of 3 weeks due to severe hypoglycemia [17], hence precluding the study of WWOX physiological functions in adult mice. To overcome this issue, a conditional knockout mouse model in which Wwox can be deleted in a time- and tissue-specific manner was recently generated [18]. Using this novel model, we recently showed that WWOX somatic ablation in mammary epithelium is associated with mammary tumor formation and p53 impaired function [19]. Moreover, specific WWOX deletion in hepatocytes accelerates the development of hepatocellular carcinoma (HCC), partly due to the promotion of HIF1α activity [20].

In this report, we screened for the metabolic function of WWOX using engineered mouse models in which the murine Wwox gene was specifically deleted in the main metabolic peripheral organs including liver, adipose tissue, and SKM. Interestingly, we found that only mice with SKM-specific ablation of Wwox develop a phenotype resembling MetS, as manifested by hyperglycemia, obesity, and dyslipidemia. Remarkably, Wwox ablation in SKM is associated with decreased carbohydrate oxidation, fewer slow-twitch muscle fibers, and reduced mitochondrial mass. Mechanistically, WWOX loss is associated with impaired HIF1α and AMPK activity.

Results

The WWOX gene is frequently altered in MetS

Genome-wide association studies (GWAS) have linked several single nucleotide polymorphisms (SNPs) in WWOX with obesity [21], [22] and T2D [23], [24]. To systematically study the role of WWOX in MetS, we reanalyzed GWAS datasets (n = 27) from the Type 2 Diabetes Knowledge Portal website (www.type2diabetesgenetics.org). Several WWOX variants were found to be strongly associated with MetS disorders including T2D (Figure S1A), high fasting glucose (Figure S1B), abnormal waist circumference (Figure S1C), high body mass index (BMI; Figure S1D), and dysregulated triglyceride levels (Figure S1E). Notably, several of these variants lie in the WWOX coding sequence and some are predicted to change its amino acid sequence (Supplementary Table 1). These data provide genetic evidence for the involvement of WWOX in MetS disorders.

Conditional ablation of murine Wwox in peripheral metabolic tissues

To investigate the role of WWOX in metabolic homeostasis, we conditionally deleted its gene in the main peripheral metabolic tissues: the liver, adipose tissue, and SKM (Figure 1A). To this end, the Wwox-floxed mouse (Wwox) [18] was bred with either Albumin-Cre [25], Adiponectine-Cre [26], or ACTA1-rtTAtetO-Cre [27] to delete the Wwox gene specifically in hepatocytes (Figure 1B), adipocytes (Figure 1C), or SKMs (Figure 1D), respectively. Thereafter, we tested the effects of WWOX loss on whole-body metabolism by monitoring fasting blood glucose levels and body weight. Although WWOX loss in hepatocytes and adipocytes had no significant effects on fasting glucose levels compared with their control littermate mice (Cre+, Wwox+/+), a significant increase was observed in mice with muscle-specific ablation of Wwox (Wwox) (Figure 1E). Likewise, only mice that lost WWOX in the muscle were defective in their ability to clear glucose, as determined by the glucose tolerance test (GTT) (Figure 1F–H). Finally, Wwox mice gained significantly more weight than their wild type controls (Figure 1K), with no similar trend observed in mice lacking the Wwox gene in the liver or adipocytes (Figure 1I and J). These data indicate that WWOX expression in SKM is essential for organismal glucose homeostasis.

Figure 1

Glucose homeostasis in Wwox conditional mouse models. (A) Work plan and animal model used in the study. Wwoxf/f genetically engineered mice were crossed with Albumin-Cre mice in order to delete Wwox specifically in hepatocytes (Hep-cKO) or Adiponectin-Cre mice in order to delete Wwox specifically in adipocytes (Adipo-cKO) or with ACTA1-etTA tetO-Cre transgenic mice in order to delete Wwox specifically in skeletal muscles (SKM-cKO). Validation of WWOX deletion in livers of Hep-cKO (B), in adipose tissue of Adipo-cKO (C), and in skeletal muscles SKM-cKO (D) using qRT-PCR and immunoblotting; n = 3 for each mouse group. (E) Blood glucose level (mg/dl) measured after 18 h of overnight fasting in control mice versus Hep-cKO, Adipo-cKO, or SKM-cKO mice; n = 8 of each group. Glucose tolerance test (GTT) was used for control mice versus Hep-cKO (F), Adipo-cKO (G), or SKM-cKO (H) mice at 6 months of age for all groups. Mice were injected with 2g/Kg glucose at zero time after 18 h overnight fasting and serum glucose was measured for up to 2.5 h; n = 8 for each group. Total body weight was measured in grams once a week for control mice versus Hep-cKO (I), Adipo-cKO (J), or SKM-cKO (K). The Y-axis represents the number of follow-up weeks. Mice weighing started at the age of 6–8 weeks; n = 8 for each group. * indicates P-value <0.05, ** P-value <0.01, *** P-value <0.001. Error bars indicate ± SEM. cKO: conditional knockout. Control mice in all analysis have Cre expression in Wwox+/+ background (Alb-Cre+;Wwox+/+, Adipo-Cre+;Wwox+/+, ACTA1-etTA tetO-Cre+;Wwox+/+, respectively).

Wwox-specific ablation in skeletal muscles is associated with insulin resistance

To better understand the role of WWOX in SKM and its contribution to metabolic homeostasis, we further characterized the Wwox mice. Although no major differences were observed in SKM histology (Figure 2A), an analysis of their body composition revealed a significant increase in the fat content with a concomitant decrease in the lean body component (Figure 2B). Moreover, Wwox mice exhibited a decreased HDL/LDL ratio as well as high levels of triglycerides and cholesterol, compared with control littermates (Figure 2C and D), suggesting that Wwox mice are predisposed to MetS.

Figure 2

Wwox-specific ablation in skeletal muscles leads to insulin resistance. (A) H&E staining for quadriceps cross-sections of control and Wwox SKM-cKO mice (the upper image bar is 50 μm, the lower image bar is 20 μm). (B) Fat and lean percentage from the total body weight of control mice and Wwox SKM-cKO mice measured by MRI. (C) and (D) Serum HDL/LDL ratio. Triglycerides (TG) and total cholesterol (TC) of control mice and Wwox SKM-cKO mice were measured by Cobas analyzer. (I) Insulin tolerance test (ITT) for 10-month-old mice. Control and Hep-cKO (E), or Adipo-cKO (F), or SKM-cKO (G) mice were injected with 0.75 U/kg insulin at time zero followed by 4 h fasting and serum glucose was measured for up to 3 h. (H) Serum insulin levels of control and Wwox SKM-cKO mice (control n = 4; SKM-cKO n = 4) measured by ELISA. Glucose uptake assessment into gastrocnemius (I) and hamstring (J) skeletal muscles were assessed by PET-CT scan. * indicates P-value <0.05, ** P-value <0.01, *** P-value <0.001. Error bars indicate ± SEM. All assays were done on control mice n = 8; SKM-cKO mice n = 9.

Insulin resistance is a hallmark of metabolic deregulation and is tightly associated with MetS. Indeed, we found that, unlike control mice or mice lacking WWOX expression in the liver or adipocytes, Wwox mice suffer from insulin resistance (Figure 2E–G), and consequently, their basal insulin levels after fasting are much higher (Figure 2H). To further confirm these observations, we performed an in-vivo glucose uptake assay using radioactive (18F)-FDG and a PET-CT scan. We found that Wwox mice uptake less glucose into gastrocnemius (Figure 2I) and hamstring muscles (Figure 2J); however, their glucose uptake in the brain or the liver is not affected (Figure 2SA and B). These data suggest an autonomous defect in glucose uptake in SKM in the absence of WWOX.

Wwox-specific ablation in skeletal muscles is associated with metabolic reprograming

In order to assess the metabolic and physical activity of Wwox mice, we next monitored these mice using metabolic cages, as previously described [28]. Wwox mice exhibited decreased respiratory quotient (RQ) compared with control mice, in particular, during the active dark period of the day cycle (Figure 3A and B). Interestingly, Wwox mice displayed increased fat oxidation (Figure 3C and D), whereas carbohydrate oxidation decreased (Figure 3E and F). Although these mice gained body weight as a function of age (Figure 1K), the amount of food intake during the dark active period was decreased, compared with control mice (Figure 3G and H). Importantly, Wwox mice exhibited no difference in activity neither in the active dark period nor in the daylight period (Figure 3S A–H). Hence, the reduced locomotor activity can be excluded as a cause for fat storage, suggesting a metabolic rewiring at the organismal level.

Figure 3

Wwox-specific ablation in skeletal muscles is associated with decreased glucose oxidation. A–H: Metabolic cage analysis of control and Wwox SKM-cKO mice was measured every 0.5hr for 24hr (48 measurements); (n = 8 for each group). Respiratory quotient (RQ) chart (A), total RQ (B), fat oxidation (g/d/eff mass) chart (C), total fat oxidation (D), carbohydrate oxidation (CHO) chart (E), total CHO (F), food intake events chart (G), and total food intake (H). (I) mRNA levels of slow-twitch myosin Myh7, Tnni1, Tnnc1, and Tnnt1 vs fast-twitch troponin genes Tnni2, Tnnc2, and Tnnt3 in muscles of control and Wwox SKM-cKO mice quantified by qRT-PCR (n = 4 for each group). (J) Quantification of mitochondrial DNA in skeletal muscles of control and Wwox SKM-cKO mice as assessed by qRT-PCR (n = 4 for each group). (K) mRNA levels of Krebs cycle genes; Sdha, Ogdh, Mdh2, Idh2, Fh1, Ddlst, and Dlat in quadriceps of control and Wwox SKM-cKO mice quantified by qRT-PCR (n = 4 for each group). (L) Serum lactate fold change of control and Wwox SKM-cKO mice as assessed by MSPEC (n = 6 for each group). (M) mRNA levels of lactate transporter Mct4 and glycolytic HIF1α target genes; Ldha1, Gpi1, and Aldoa2 in quadriceps of control and Wwox SKM-cKO mice quantified by qRT-PCR (n = 4 for each group).

In accordance with these findings, Wwox mice exhibited decreased mitochondrial-OXPHOS-dependent slow-twitch fibers, with no change in the glycolytic fast-twitch-dependent fibers (Figure 3I), as determined by the measurement of myosin and troponins mRNA expression [29]. Interestingly, we found that WWOX levels are higher in the slow-twitch soleus fibers, compared with the fast-twitch extensor digitorum longus (EDL) fibers in Wwox wild-type muscle (Figure 4S). These findings were also consistent with the decreased mitochondrial mass and Krebs cycle gene levels in the SKMs of Wwox (Figure 3J and K). Taken together, our data suggest a defect in mitochondrial oxidation. Indeed, serum lactate levels were significantly elevated in Wwox mice (Figure 3L). Likewise, we also observed an upregulation of lactate production (LDHA1) and exportation genes (MCT4), coupled with elevation of other HIF1α-regulated glycolytic genes (Figure 3M). Thus, our findings indicate that Wwox ablation in SKM disrupts the intra-muscular energetic balance by suppressing mitochondrial glucose oxidation and activating lactate production, leading to disruption of whole-body glucose homeostasis and development of MetS.

WWOX regulates central metabolic proteins in skeletal muscles

In order to better understand how WWOX regulates the metabolic state in skeletal muscles, we tested the effects of its loss on a number of key metabolic pathways. We have previously shown that WWOX modulates HIF1α by inhibiting its accumulation and function under normoxic conditions [15]. We therefore set out to examine whether HIF1α levels and function are enhanced in Wwox mice. Indeed, we found elevated HIF1α protein levels in quadriceps muscles of Wwox mice (Figure 4A and B). Moreover, the levels of Pkm2 and Pdk1, key glycolytic genes, regulated by HIF1α, were also elevated in SKM of these mice (Figure 4A and B), indicating that HIF1α is functionally active in Wwox-deficient skeletal muscles, consistent with previous observations in Wwox-null mice (Abu-Remaileh and Aqeilan, 2014).

Figure 4

WWOX regulates central metabolic proteins in skeletal muscles. (A) Immunoblotting of HIF1α, PKM2, PDK1, pAMPK, total-AMPK, pACC, total-ACC, and WWOX protein levels in quadriceps of control and Wwox-SKM-cKO mice. Representative blots of two mice for each group is shown. (B) Quantification of the pAMPK/tAMPK ratio, the HIF1α, PKM2, PDK1, and pACC/tACC ratio of 3 different control and SKM-cKO mice presented as fold change. (C) pAMPK levels in C2C12 cells after WWOX knockdown using siRNA. (D) Quantification of the pAMPK/tAMPK ratio, the WWOX and pACC/tACC ratio of control and si-WWOX-transfected cell lines presented as the fold change. (E) Co-immunoprecipitation experiments showing physical interaction between WWOX and AMPK. Flag-AMPK and HA-WWOX were co-transfected in HEK293T cells. Lysates were immunoprecipitated with anti-HA or anti-FLAG, followed by immunoblotting with the indicated antibodies. Anti-IgG was used as a negative control. (F) Immunoprecipitation experiment showing physical interaction between endogenous WWOX and pAMPK in C2C12 muscle cells. Lysates were immunoprecipitated with mouse monoclonal anti-WWOX or anti-IgG, followed by immunoblotting with polyclonal anti-WWOX or anti-pAMPK. Low and high exposure blots for WWOX antibody are shown. (G) A model showing WWOX inhibition of HIF1α and proper activation of AMPK, resulting in enhanced mitochondrial function and antagonizing of MetS. Band quantification of western blots was done by the ImageJ program. * indicates P-value <0.05, ** P-value <0.01, *** P-value <0.001. Error bars indicate ± SEM. Protein levels were quantified from control mice n = 3; SKM-cKO mice n = 3 or control cell lines n = 4; treated cell lines n = 4. Representative blots are shown.

Another key player in glucose homeostasis and cellular metabolism is AMPK, which has been reported to be inactivated in MetS disorders [30], and its reactivation has been recently suggested as a new therapeutic approach to MetS [31]. Hence, we examined AMPK activation and found that Thr172 phosphorylation of AMPK (p-AMPK) is completely disrupted in skeletal muscles of Wwox mice (Figure 4A and B). Notably, the phosphorylated form of acetyl-CoA carboxylase (p-ACC), a canonical substrate of AMPK, was also decreased, indicating an AMPK decreased activity upon Wwox-specific deletion in SKM (Figure 4A and B). Interestingly, fat tissues of Wwox mice display no major changes in the previously mentioned metabolic proteins (Figure 5S). These results suggest that WWOX could act as a positive regulator of AMPK activity in SKM.

To address whether WWOX regulates AMPK activation, we transfected the mouse myoblast C2C12 cell line with siRNA against murine Wwox and tested AMPK levels and activity. As seen in Figure 4C and D, p-AMPK is decreased upon acute WWOX depletion, consistent with a possible role of WWOX in directly regulating AMPK activation.

To further delineate this crosstalk, we investigated whether WWOX, a known adaptor protein, is physically associated with AMPK protein. To this end, HEK293T cells, which can be transfected easily and are frequently used in our lab to check WWOX interactome [32], were co-transfected with Flag-AMPK1α and HA-WWOX, followed by immuno-precipitation with anti-Flag or anti-HA, or anti-IgG as a negative control. We indeed found that both exogenous proteins interact with each other (Figure 4E). Moreover, we detected specific endogenous WWOX and p-AMPK protein interaction in C2C12 muscle cells (Figure 4F), implying functional crosstalk between the two metabolic proteins. Taken together, these results indicate that WWOX deletion in SKM leads to a failure in proper AMPK activation, and an enhanced activation of HIF1α, leading to disruption of metabolic balance in SKM and the development of MetS disorders (Figure 4G).

Discussion

SKM is a fundamental peripheral tissue that is involved in glucose homeostasis. Impaired glucose uptake and insulin resistance in SKM is a major cause of T2D and other metabolic disorders [33]. In fact, a disruption in OXPHOS, mitochondrial activity, fatty acid oxidation, and glycolysis in SKM is suggested as the main cause of failure of glucose homeostasis [34]. Therefore, identifying the main regulators of SKM glucose metabolism is of great interest.

Over the last decade, many GWAS studies have reported several genetic variants in WWOX as among the most significant SNPs and variants associated with obesity and T2D [22], [24]. However, a direct role of WWOX involvement in metabolic disorders is unclear. Here, we demonstrated a novel function for the tumor suppressor WWOX in SKM metabolism. We found that specific ablation of murine Wwox in SKM could lead to obesity, hyperglycemia, and insulin resistance. These metabolic abnormalities were associated with decreased mitochondrial mass and reduced OXPHOS, hence impairing the bioenergetic balance. At the molecular level, our findings reveal an important functional crosstalk between WWOX and HIF1α, as well as between WWOX and AMPK in SKM. These observations may suggest that germline mutations in the WWOX gene result in impaired glucose homeostasis as a result of SKM metabolic dysfunction.

MetS and impaired glucose homeostasis are commonly caused by defects in peripheral organs involved in glucose uptake and utilization including liver, adipose tissue, and SKM. Characterizing the metabolic phenotypes of different mouse models that harbor Wwox deletion in each of these organs revealed that only WWOX's loss of function in SKM is associated with phenotypes resembling MetS. Although WWOX inactivation in liver and adipocytes did not reveal a defect in glucose homeostasis and body weight regulation, this does not exclude a function of WWOX in other metabolic pathways or nutrient states, which were not addressed in this study. In fact, specific deletion of Wwox in hepatocytes was shown to be associated with impaired lipid metabolism [13], consistent with the susceptibility of certain WWOX variants to cardiovascular diseases [13]. More recently, we demonstrated that WWOX ablation in hepatocytes is associated with acceleration of HCC development due to enhanced proliferation mediated by a glycolytic switch [20].

The phenotypes observed upon WWOX ablation in SKM were very prominent. Increased body weights, fat-to-lean ratios, and circulating fats were evident in Wwox mice. Moreover, these mice displayed impaired GTT and ITT, and manifested hyperglycemia and hyperinsulinemia, indicating a state of insulin resistance. Furthermore, an in-vivo 18F-FDG glucose uptake assay demonstrated a specific decrease in SKM glucose uptake. Although these mice developed hyperglycemia, they mainly relied on fatty oxidation as a main source of cellular energy. Consistent with these findings, we observed a decrease in the quantity of oxidative-dependent slow-twitch fibers in SKM of Wwox mice. These observations are consistent with differential WWOX function in soleus slow-twitch muscles, expressing relatively higher levels of WWOX, when compared to EDL fast-twitch muscles. Moreover, decreased mitochondrial mass suggests inhibition of OXPHOS and carbohydrate oxidation. In accordance, the levels of lactate serum, the lactate exporter MCT4, and the lactate dehydrogenase enzyme LDHA1 were all found to be elevated in Wwox mice, indicating enhanced glycolysis. In general, these metabolic perturbations mimic the metabolic status of hyperglycemic T2D human patients [35]. In fact, the SKM of T2D human patients present a downregulated mitochondrial OXPHOS and decreased mitochondrial DNA content [36], [37]. Furthermore, T2D patients exhibit an increased rate of glycolysis and lactate production [38], further suggesting high similarities between human T2D patients and Wwox deletion in SKM. These data imply that WWOX in SKM plays a central role in regulating glucose catabolic pathways and antagonizes the development of the MetS.

Although Wwox mice weighed more, a snapshot analysis of their 24-hour food consumption revealed that these mice eat less during their dark active period. These results might be explained by the high systemic insulinemia, which is known to affect central pathways that modulate feeding [39], as well as play a lipogenic role in adipose tissue [40], suggesting that MetS and insulin resistance have a SKM cell-autonomous effect.

At the molecular level, WWOX is known to play pleotropic roles in different contexts including DNA damage response and cellular metabolism [11]. In our current study, we demonstrate that WWOX loss could lead to significant changes in two main pathways: HIF1α and AMPK. Consistent with previous observations [15], we show that upon WWOX loss, HIF1α levels are elevated in SKM of Wwox mice. This elevation in HIF1α levels was accompanied by increased transactivation activity of the main glycolytic genes, resulting in their upregulation. Although the increased glycolysis does not seem massive, still, upregulated PDK1 levels blocked glucose influx into the Krebs cycle, leading to decreased glucose oxidation. Moreover, HIF1α was reported to inhibit Krebs cycle genes and mitochondrial mass [41], [42], which is in accordance with the phenotype observed in Wwox mice.

AMPK is a serine/threonine kinase sensor that is activated upon a drop of the energy status in the cell, namely, by elevated levels of AMP [43]. Once phosphorylated (activated), AMPK phosphorylates a number of target proteins and alters their functions, resulting, for example, in inhibition of lipid metabolism and activation of fatty acid oxidation through ACC1 and ACC2 [44]. Upon WWOX loss, AMPK activation is hampered in SKM of Wwox mice, as revealed by its decreased phosphorylation on residue T-172. AMPK activation increases fatty acid oxidation in SKM [45]. The increased systemic fat oxidation in Wwox mice, as assessed by metabolic cages, may be explained as a compensatory effect for impaired catabolic pathways of carbohydrate oxidation. Moreover, the inhibition of ACC2 in SKM by phosphorylation, known to enhance fatty acid oxidation in SKM [46], is indeed impaired and may explain the decreased glucose oxidation and increased fatty acid oxidation in Wwox mice, which is also demonstrated by knocking down Wwox in C2C12 muscle cells. Since AMPK is an activator of mitochondrial respiration [47] and inhibits lipogenesis through inhibition of ACC [48], WWOX loss could result in improper activation of AMPK and hence decreased OXPHOS. This could affect many metabolic pathways in other active metabolic organs, such as adipose tissue and liver, to increase their dependence on fatty acid oxidation [49].

We propose that WWOX regulates the proper activation of AMPK in order to respond to nutrient stress conditions and maintain the homeostasis of the skeletal muscles. Under normal physiological conditions, AMPK is mostly inactive; however, under stress (for example, starvation), WWOX protein levels are elevated, leading to enhanced activation of AMPK. Therefore, it is reasonable to assume that elevated levels of WWOX mimic the poor energetic status of the cells, resulting in AMPK activation and hence, increased glucose oxidation. Under normal conditions, we failed to observe an AMPK activation in WWOX-depleted cells, suggesting that WWOX is required for this important cellular response. In fact, we found that WWOX knockdown resulted in impaired activation of AMPK, suggesting that WWOX is most likely an upstream regulator of AMPK. Furthermore, we provided evidence that WWOX physically binds AMPK, indicating a functional interaction between these two proteins. Therefore, it is possible that WWOX deficiency signals a state of excess nutrients and thus AMPK is not activated, leading to the accumulation of extracellular glucose and hyperglycemia. Further studies will be necessary to examine the functional association between WWOX and AMPK.

By analyzing human data sets in the T2D portal, we revealed a significant association between WWOX variants and a number of metabolic disorders including T2D, obesity, BMI, and many others. Interestingly, several of these variants are located in the WWOX coding region, and they are expected to generate mutations in WWOX protein. Future studies should investigate the effect of these variants on WWOX protein function and stability, as well as how they modulate AMPK and HIF1α functions.

Although it has been reported that AMPK is a critical modulator of HIF1α transcriptional activity in hypoxic DU145 prostate cancer cells [50], very little is known about the cross-talk between AMPK and HIF1α in SKM [51]. Our results, presented here, provide, for the first time, clear evidence that WWOX plays a key role in glucose homeostasis in SKM by regulating AMPK and HIF1α. WWOX maintains a proper AMPK activation and restricts HIF1α activity, resulting in balanced glucose metabolism circuits in SKMs. Upon WWOX deficiency, AMPK activation becomes dysfunctional and HIF1α accumulates, leading to reduced mitochondrial oxidation, augmented glycolysis, and fatty acid oxidation, which can lead to the development of MetS (Figure 4G).

Methods

Animal models

Wwox-floxed (Wwoxfl/fl) C57BL6/J; 129sv mixed genetic background mice [18] were crossed with the following transgenic mice: Albumin-Cre transgenic mice [Jackson stock No. 003574] to generate Wwox conditional knock-out in hepatocytes (Hep-cKO) [25], Adiponectine-Cre transgenic mice [Jackson stock No. 028020] to generate Wwox conditional knock-out in adipose tissue (Adipo-cKO) [26], and ACTA1-rtTAtetO-Cre transgenic mice [Jackson stock No. 012433] to generate Wwox conditional knock-out in skeletal muscles (SKM-cKO) [27]. Control mice for each genotype exhibited Cre expression in Wwox+/+ background (Alb-Cre+;Wwox+/+, Adipo-Cre+;Wwox+/+, ACTA1-etTA tetO-Cre+;Wwox+/+). Male mice were weighed once every week from the age of 6 weeks up to 37 weeks. To induce Wwox deletion in SKM, 2 mg/mL doxycycline and 1% sucrose were added to the drinking water of both ACTA1-rtTAtetO-Cre;Wwoxf/f mice and ACTA1-rtTAtetO-Cre;Wwox+/+ control littermates in breeding cages, which were changed every 3 days and this protocol lasted until mice were euthanized for analysis. Before euthanasia, the metabolic profiles of the mice were assessed by using the Promethion High-Definition Behavioral Phenotyping System (Sable Instruments, Inc., Las Vegas, NV, USA), as described previously [28]; total body fat and lean masses were determined by EchoMRI-100H™ (Echo Medical Systems LLC, Houston, TX, USA). All experiments involving mice were approved by the Hebrew University Institutional Animal Care and Use Committee (IACUC).

MicroPET-CT scan

PET-CT scans were carried out in the Cyclotron Unit at Hadassah Medical Center according to IACUC rules. Following an overnight (12–16 h) food deprivation, mice were anesthetized with isoflurane (1%–2.5% in O2) and maintained normothermic using a heating pad. Following a CT attenuation-correction scan, PET acquisitions were carried out in list-mode using an Inveon™ MM PET-CT small animal-dedicated scanner (Siemens Medical Solutions, USA). PET scans were initiated at the time of [18F]-FDG injection via the lateral tail vein (7.8 ± 1 MBq), and lasted 1 h. Emission sinograms were normalized and corrected for attenuation, scatter, randoms, dead time, and decay. Images were reconstructed using Fourier rebinning and two-dimensional ordered-subsets expectation maximization (2D-OSEM), with a voxel size of 0.776 × 0.776 × 0.796 mm3. Image analysis and quantification were performed using Inveon Research Workplace 4.2 (Siemens). Delineation of volumes of interest (VOIs) in the right and left gastrocnemius muscles of each animal was performed by manual segmentation, based on the corresponding CT image, and [18F]-FDG time activity curves (TACs) were generated. Distribution of activity was calculated as the percentage of injected dose per mL of tissue (%ID/mL). Standardized uptake values (SUVs) were calculated as the product of %ID/mL and the total body weight of the animal.

Glucose and insulin tolerance tests (GTT and ITT)

For GTT, mice (6 months old) were fasted for 18 h, followed by an intraperitoneal (i.p.) injection of 2 g/kg body weight of glucose [52], [53]. Blood glucose was measured at 0, 30, 60, 120, and 150 min. ITT was performed on 6-month-old mice after a fasting of 4 h, followed by an i.p. injection of 0.75 U/kg body weight of insulin (Biological Industries, Beit Haemek, Israel). Blood glucose was measured at 0, 30, 60, 120, and 150 min. All blood glucose measurements were carried out using an Accu-Check glucometer (Roche Diagnostics, Mannheim, Germany).

Histology of skeletal muscles

Skeletal muscles were harvested from the scarified mice, covered with OCT (Scigen Scientific Gardena, CA, USA), dipped in chilled 2-methylbutane (Isopentane) for 25 s, then transferred to frozen tissues onto dry ice for 20 min. The tissues were wrapped in aluminum foil and stored at −80 °C until sectioning. Tissues were sectioned (4 μm) in a cryostat microtome at −20 °C, and stained with hematoxylin and eosin.

Cell culture and transient transfection

C2C12 and HEK293 cells were grown in DMEM supplemented with 10% FBS (Gibco, Grand Island, USA), glutamine, and penicillin/streptomycin (Beit-Haemek, Israel). Cells were routinely authenticated, and cell aliquots from early passages were used. Transient transfections of HEK293 cells were achieved using Mirus TransLTi (Mirus Bio LLC, Wisconsin, USA) according the manufacturer's instructions. Transient transfections of C2C12 cells by WWOX siRNA (Sigma Aldrich, Germany) were achieved using lipofectamine 2000 (Thermo Fisher Scientific, CA, USA) according to the manufacturer's instructions.

Immunoprecipitation assays

Cells were lysed using NP-40 lysis buffer containing 50 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 10% glycerol, 0.5% Nonidet P-40, and protease inhibitors. Lysates were pre-cleared with mouse IgG, immunoprecipitations (IP) were carried out in the same buffer, and lysates were washed 4 times with the same buffer containing 0.1% Nonidet P-40. Antibodies used for IP were monoclonal anti-WWOX antibody [54], anti-HA (Cat # ab9110, Abcam, USA), and anti-Flag (Cat # F3165, Sigma–Aldrich, Germany).

Immunoblot analysis

For immunoblot analysis, cells were lysed using lysis buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 10% glycerol, 0.5% Nonidet P-40, and protease inhibitors (1:100). Lysates were resolved on SDS/PAGE. Antibodies used were polyclonal anti-WWOX antibody (Cat# S2603, Epitomics, CA, USA), anti-GAPDH mouse mAB (Cat # CB 100 1, CALBIOCHEM, MA, USA), anti-HIF1α Rabbit (Cat # NB100-479, Novus Biological, CO, USA), anti-HSP90 Rabbit (Cat # CA1016, CALBIOCHEM, MA, USA), anti-pACC (Ser79) Rabbit (Cat # 3661, Cell Signaling Technology, MA, USA), anti-AMPKa Rabbit (Cat # 2532, Cell Signaling Technology, MA, USA), anti-pAMPKa (Thr 172) Rabbit (Cat # 2535, Cell Signaling Technology, MA, USA), anti-HA (Cat # ab9110, Abcam, USA), and anti-Flag (Cat # F3165, Sigma–Aldrich, Germany).

RNA extraction and real-time PCR

Total RNA was prepared using TRI reagent (Sigma Aldrich), following the manufacturer's instructions. One microgram of RNA was used for cDNA synthesis using the First-Strand cDNA Synthesis kit (Bio-Rad, Hercules, CA). Quantitative real-time PCR was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). All measurements were performed in triplicate and standardized to the levels of the Ubc gene. A list of primers used is provided as Supplemental Table 1.

Blood biochemistry

Serum levels of cholesterol, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) were determined using the Cobas C-111 chemistry analyzer (Roche, Switzerland).

Statistics

Results of the experiments were expressed as mean ± SD or SEM. Student's t-test was used to compare the values of the test and control samples. P < 0.05 indicates a significant difference. * P-value <0.05, ** P-value < 0.01. Human data analysis was carried out by Type 2 Diabetes Knowledge Portal site (www.type2diabetesgenetics.org), which included 27 datasets in Dec. 2017.