|Home | About | Journals | Submit | Contact Us | Français|
PGC1α is a key transcriptional coregulator of oxidative metabolism and thermogenesis. Through a high throughput chemical screen, we found that molecules antagonizing the TRPVs (Transient Receptor Potential Vanilloid), a family of ion channels, induced PGC1α expression in adipocytes. In particular, TRPV4 negatively regulated the expression of PGC1α, UCP1 and cellular respiration. Additionally, it potently controlled the expression of multiple proinflammatory genes involved in the development of insulin resistance. Mice with a null mutation for TRPV4 or wild-type mice treated with a TRPV4 antagonist showed elevated thermogenesis in adipose tissues and were protected from diet-induced obesity, adipose inflammation and insulin resistance. This role of TRPV4 as a cell-autonomous mediator for both the thermogenic and proinflammatory programs in adipocytes could offer a new target for treating obesity and related metabolic diseases.
Brown adipose tissue (BAT) is specialized for the efficient dissipation of chemical energy in the form of heat. It does this by having an exceptionally high mitochondrial content and respiration that is uncoupled from ATP synthesis. This uncoupling is mainly due to the presence of UCP1, a protein that catalyzes proton leak across the inner mitochondrial membrane. Brown fat is very prominent in rodents and human infants but the presence of substantial brown fat deposits in adult humans has only recently been appreciated (Cypess et al., 2009; van Marken Lichtenbelt et al., 2009; Virtanen et al., 2009).
It is now known that there are two distinct kinds of brown adipocytes. The classical type is exemplified by the interscapular depot of rodents; these UCP1 expressing cells are derived from a muscle-like lineage that expressed Myf5/Pax7 during earlier development (Lepper and Fan, 2010; Seale et al., 2008). UCP1-positive cells can also emerge in many white fat depots under chronic exposure to cold or β-adrenergic stimulation (Cousin et al., 1992; Ghorbani and Himms-Hagen, 1997; Guerra et al., 1998; Himms-Hagen et al., 2000; Xue et al., 2005). These cells do not come from a myf5-positive lineage (Seale et al., 2008) and have been called beige or brite fat cells (Ishibashi and Seale, 2010; Petrovic et al., 2010). The regulation of UCP1 and the broader thermogenic gene program in both types of brown adipocytes has been studied in detail. Key transcriptional regulators include FOXC2 (Cederberg et al., 2001), C/EBPβ (Karamanlidis et al., 2007), LXR (Korach-Andre et al., 2011), PGC1α (Puigserver et al., 1998; Uldry et al., 2006) and PRDM16 (Kajimura et al., 2009; Seale et al., 2008; Seale et al., 2007).
PGC1α was originally identified as a coactivator of PPARγ in the control of the UCP1 promoter in brown adipocytes (Puigserver et al., 1998). Subsequent work has illustrated that it binds to and coactivates many transcription factors (Handschin and Spiegelman, 2006). PGC1α plays a key role in mitochondrial biogenesis and oxidative metabolism in many tissues, linking mitochondrial biogenesis to the extracellular and extraorganismal environment. PGC1α gene expression is induced in brown adipose tissue by cold exposure and by agents that activate the β-adrenergic system.
The responsiveness of PGC1α gene expression to external stimuli suggested that it might be possible to find chemical compounds that increase PGC1α expression and function. This in turn might be useful for the treatment of a variety of diseases that would benefit from increased PGC1α or from increased mitochondrial activity (Handschin and Spiegelman, 2008). Because elevated PGC1α in muscle plays an anti-dystrophic and anti-atrophic function, we previously screened for drugs and drug-like molecules that elevate PGC1α in murine myotubes (Arany et al., 2008). Several inhibitors of microtubules and protein synthesis were identified as PGC1α inducers. This illustrated that screening for activators of PGC1α expression could identify compounds capable of increasing mitochondrial action. Conversely, when a screen for chemicals that could alter mitochondrial function was carried out, an overlapping set of regulators of PGC1α was uncovered (Wagner et al., 2008). Unfortunately, none of these compounds had an activity/toxicity ratio that was favorable for animal or human studies.
In this study, we have screened a chemical library for compounds that could increase PGC1α gene expression in white adipocytes. We show here that TRPV4, a member of a family of chemically tractable ion channels, is a negative regulator of PGC1α and the thermogenic gene program. Furthermore, TRPV4 positively regulates a host of proinflammatory genes in white adipocytes. Genetic ablation and pharmacological inhibition of TRPV4 in mice modulate both thermogenic and proinflammatory pathways in fat, resulting in a robust resistance to obesity and insulin resistance.
We performed a quantitative PCR-based chemical screen to identify small molecules that can induce Pgc1α mRNA expression in white adipocytes. Fully differentiated 3T3-F442A adipocytes were treated with a chemical library of 3,000 drugs and drug-like compounds for 20 hours; mRNA from treated cells was then analyzed to quantify the expression of Pgc1α (Figure S1). AM-251, a cannabinoid receptor 1 (CB1) antagonist was identified as one of the primary hits (Figure 1A). AM-251 is a structural analogue of a well-known CB1 antagonist rimonabant (Lan et al., 1999), an anti-obesity drug that was in clinical use in Europe. Although AM-251 is annotated as a CB1 antagonist, two other CB1 antagonists, SLV319 and CAY10508, failed to induce Pgc1α at any dose tested (Figure 1A). Importantly, other molecular targets of AM251 or rimonabant have been reported, including TRPV1 (De Petrocellis et al., 2001; Zygmunt et al., 1999). As shown in Figure 1B, two TRPV1 antagonists, AMG9810 and BCTC, increased Pgc1α mRNA expression in adipocytes in a dose-dependent manner (Figure 1B).
AMG9810 antagonizes TRPV1 but can also antagonize closely related TRPVs, such as TRPV2, TRPV3 and TRPV4, at the micromolar doses used here (Gavva et al., 2005). We therefore compared the mRNA expression of Trpv1–4 in 3T3-F442A adipocytes. mRNAs encoding Trpv1, Trpv2 and Trpv4 were expressed in 3T3-F442A adipocytes, with Trpv4 being expressed at the highest level (Figure 1C). To determine which of these channels were regulating Pgc1α expression, we used shRNA-mediated knockdown of each of the expressed TRPVs with lentiviral vectors (Figure S1B). None of the shRNAs appeared to affect adipose differentiation per se, as indicated by the similar expression of the adipose-selective gene aP2 (Figure 1D). Pgc1α mRNA was strongly induced by the shRNA against TRPV4; shRNA against TRPV1 also had a small effect (Figure 1E). This functional data, along with the fact that the expression of Trpv4 mRNA was 10 times higher than that of Trpv1 in these cells, strongly suggest that TRPV4 was the dominant TPRV family member regulating the induction of Pgc1α mRNA by the chemical inhibitors.
TRPV4 is a calcium permeable ion channel that was first identified as an osmolality sensor (Liedtke et al., 2000; Strotmann et al., 2000). Since then, many physical and chemical stimuli have been shown to activate TRPV4, including heat, mechanical stress, anandamide, arachidonic acid and its derivatives (Everaerts et al., 2010; Nilius et al., 2007). Adipose tissue was shown to have one of the highest levels of Trpv4 mRNA expression (Liedtke et al., 2000). We found that Trpv4 expression was higher in white adipose tissues than brown fat tissue (Figure S2A).
We used retroviral vectors expressing an shRNA against TRPV4 or GFP to make stable cells with altered TRPV4 expression for biochemical and bioenergetic analyses. Again, the ectopic retroviral shRNA did not appear to effect adipocyte differentiation per se (Figure S2B). We first examined if there were functional TRPV4 channels present in 3T3-F442A adipocytes. TRPV4 protein was detected by western blot (Figure 2A). In addition, we used Ca2+ imaging to test for TRPV4 activity. Both hypotonicity and the TRPV4 agonist GSK1016790A (Thorneloe et al., 2008; Willette et al., 2008) induced a TRPV4-dependent increase in intracellular calcium in adipocytes (Figure 2B). Consistent with Ca2+ imaging results, TRPV4 agonist evoked a TRPV4-like current in control adipocytes but not in cells expressing an shRNA against TRPV4 (Figure 2C, D).
Pgc1α mRNA expression was 3 times higher in adipocytes expressing shRNA against TRPV4 compared to controls (Figure 2E). Increased PGC1α protein was confirmed by western blot (Figure 2F). β-adrenergic signaling is important for the induction of PGC1α and its thermogenic targets; when cells were exposed to norepinephrine, mRNA expression of Pgc1α and Ucp1 was robustly increased in the TRPV4 knock-down cells compared to controls (Figure 2E). PGC1α is known to drive the expression of many genes involved in mitochondrial oxidative phosphorylation, including cytochrome c (CytC), and the cytochrome C oxidative (COX) subunits (CoxIII, Cox4il, Cox5b, Cox7a and Cox8b). We observed higher mRNA (Figure 2G) and protein (Figure S2C) expression of these genes in TRPV4 knock-down adipocytes. These changes were dependent on the induction of PGC1α, as the increased expression of these genes was attenuated by expression of an shRNA against PGC1α (Figure S2D).
The increased expression of Pgc1α, Ucp1 and other mitochondrial genes suggested that TRPV4 inhibition caused the white adipocytes to develop brown fat-like characteristics, which we termed “browning” here. To determine the physiological impact of this browning gene program, oxygen consumption was measured in adipocytes. As shown in Figure 2H, TRPV4 knock-down adipocytes showed increase in basal, uncoupled and FCCP-stimulated maximal respiration, compared to controls.
We next examined whether chemical activation of TRPV4 would have the opposite impact on these pathways. When added to 3T3-F442A adipocytes, the TRPV4 agonist GSK1016790A repressed the expression of mRNAs encoding Pgc1α, Ucp1 and Cox8b in a TRPV4-dependent manner (Figure 2I). Taken together, these data strongly suggest that TRPV4 functions as a negative regulator of PGC1α and oxidative metabolism in white adipocytes.
To fully understand the function of TRPV4 in adipocytes, microarray analysis was performed with mRNA from 3T3-F442A adipocytes expressing shRNAs against TRPV4 or GFP. As expected, many genes whose expression was strongly increased were those involved in brown fat function (Table S1). Pgc1α was one of the most highly regulated genes on these arrays.
Strikingly, many genes whose expression was decreased by the TRPV4 knock-down were chemokines/cytokines or genes related to proinflammatory pathways (Table S1). We further analyzed the expression of 22 genes that are either highly regulated by TRPV4 (from the array) or are known to be important in adipose inflammation based on literature. Knock-down of TRPV4 had a profound inhibitory effect on a whole array of chemokines, such as Ccl2 (Mcp1), Ccl3 (Mip1α), Ccl5 (Rantes), Ccl7 (Mcp3), Cxcl1 (KC), Ccl8, Cxcl5 and Cxcl10 and cytokines such as Il6, Saa3 and Thrombospondin (Figure 3A). A similar effect was observed on the expression of other genes important for inflammatory processes, such as Tlr2, Timp1, Socs3, Socs5, Mmp2, Fas and Vcam (Figure 3B). Conversely, mRNA expression of Mip1α, Cxcl1, Il6, Timp1 and Tlr2, can be induced by the TRPV4 agonist (Figure 3C). This effect is specific and dependent on TRPV4, as shRNA against TRPV4 fully abolished the induction caused by the agonist (Figure 3C). mRNA expression changes for other adipokines, such as Adiponectin, Leptin, RBP4 and Resistin were also observed (Figure S3).
To determine if these effects on gene expression resulted in alterations in chemokine secretion, we measured levels of secreted MCP1, MIP1α, CXCL1, IL6 and RANTES in culture medium. Similar to what we observed at the mRNA level, the concentrations of MCP1, MIP1α, CXCL1, IL6 and RANTES were each reduced by more than 85% in the culture medium from the TRPV4 knockdown adipocytes, compared to controls (Figure 3D). The TRPV4 agonist potently induced MIP1α and IL6 secretion; the secretion of MCP1 was mildly reduced. These data indicate a powerful role for TRPV4 in the regulation of a proinflammatory pathway in adipocytes.
We investigated the signaling pathways by which TRPV4 carries out these functions in adipocytes. It has been reported that the protein kinases ERK1/2 can be activated by TRPV4 signaling (Li et al., 2009; Thodeti et al., 2009). We therefore examined TRPV4 agonism and the activation of three MAP kinases that have been implicated in adipose biology: ERK1/2, JNK1/2 and p38 MAPK. Addition of the TRPV4 agonist to 3T3-F442A adipocytes caused a rapid phosphorylation of both ERK1/2 and JNK1/2 at sites known to reflect activation of these kinases. In contrast, no activating phosphorylation on p38 MAPK was detected with TRPV4 agonism; while the β3-agonist CL316243 led to the expected p38 MAPK activation (Cao et al., 2001) (Figure 4A). The activation of ERK1/2 appeared to be dependent on TRPV4 and the presence of extracellular calcium. The phosphorylations caused by the TRPV4 agonist were largely attenuated by the shRNA against TRPV4, or by the depletion of extracellular calcium; in contrast, the phosphorylations caused by TNFα were not affected (Figure 4B).
Inhibitors of MEK1/2 (U0126) and JNK (SP600125) were used to determine if the activation of these two MAP kinases was required for the key TRPV4-mediated gene regulation events. Pretreatment of cells with U0126 and SP600125 blocked the TRPV4 agonist-induced phosphorylation of ERK1/2 and JNK1/2, respectively (Figure 4C). Interestingly, U0126 effectively reversed the repression on Pgc1α caused by the agonist (Figure 4D). In contrast, SP600125 had only a small effect. Concordantly, the induction of Mip1α and Cxcl1 by the TRPV4 agonist was abolished by U0126; SP600125 had no effect. These data strongly suggest that the ERK1/2 protein kinases mediate much of the effect of TRPV4 activation on both the repression of Pgc1α expression and the induction of many chemokines/cytokines in adipocytes.
To investigate the function of TRPV4 in adipose tissues in vivo, we studied mice with a genetic deletion of Trpv4. These mice are grossly similar to wild-type animals in morphology, behavior and breeding (Liedtke and Friedman, 2003). On a chow diet, their body weight is indistinguishable from WT littermates (Figure 5A). In light of the effect of TRPV4 on ”browning” and proinflammatory programs in white adipocytes, we examined gene expression in white adipose tissues from Trpv4−/− and WT mice.
Subcutaneous (SubQ) adipose tissue has been shown to have a greater thermogenic capacity than other white adipose tissues (Barbatelli et al., 2010) and can significantly contribute to whole body energy homeostasis (Seale et al., 2011). Strikingly, SubQ fat from Trpv4−/− mice expressed 30 fold higher Ucp1 mRNA and more UCP1 protein compared to controls (Figure 5B, D). A trend towards increased Pgc1a mRNA (Figure 5B) and protein (Figure 5C) was also detected in Trpv4−/− adipose tissues, however, with substantial variability between individual animals. These mice also have elevated mRNA levels for many genes known to be enriched in BAT, such as Cidea, Cox4il, and Cox8b (Figure 5B). In general, epididymal (EPI) fat has a lower thermogenic capacity; nonetheless, mRNA levels for Adrb3, Pgc1β, CytC, Cox4il and Cox5a, were significantly higher in EPI fat from the Trpv4−/− mice compared to controls (Figure S4D).
We also measured the expression of proinflammatory chemokines, identified from the analysis of TRPV4 knock-down 3T3-F442A adipocytes. These included Mcp1, Mip1α, Mcp3, Rantes and Vcam. These genes were expressed at very low levels in the adipose tissues of lean animals and no significant differences were observed between the mutants and controls on a chow diet (Figure 5F).
To further understand the role of TRPV4 under metabolic stress, we challenged these mice with a high fat diet (HFD) which would induce obesity and provoke adipose inflammation. There was no significant body weight difference between the Trpv4−/− and control mice until the animals were on the HFD for 9 weeks (Figure 5A). Male and female mice showed a similar pattern of weight gain (Figure S4A).
The adipose tissues were first examined at 8 weeks of HFD, before the body weight of Trpv4−/− mice diverged from controls. Although the HFD tended to blunt the difference in thermogenic gene expression, SubQ fat from the Trpv4−/− animals expressed higher levels of these genes (Figure S4B). Histological analysis showed that mutant mice have smaller and more UCP1-positive adipocytes compared to controls (Figure 4E).
8 weeks of the HFD significantly elevated the mRNA expression of proinflammatory chemokines in EPI fat in WT mice, such as Mcp1, Mip1α, Rantes and Mcp3, compared to animals on a chow diet. Interestingly, without a difference in adiposity at this time point, Trpv4−/− mice showed a substantial decrease in the mRNA expression of Mcp1, Mip1α and Mcp3 (Figure 5F). Similarly, the induction of those genes in the SubQ fat was blunted in the Trpv4−/− mice (Figure S4E).
As exposure to the HFD extended to 16 weeks, there was no longer a difference in Ucp1 mRNA between the Trpv4−/− and control animals. However, the Trpv4−/− mice still had elevated expression of mRNAs encoding Adrb3 and Pgc1α in both the SubQ and EPI depots (Figure S4C, D). Chemokine expression in EPI fat continued to rise in WT mice. The expression of mRNAs for Mip1α and Vcam remained low in the Trpv4−/− mice, but the differences in Mcp1 and Mcp3 were blunted (Figure 5F).
Since the Trpv4−/− mice have a whole-body TRPV4 deficiency, we asked if the phenotype observed in vivo was associated with cell-autonomous alterations in adipocytes. To examine this, stromal-vascular cells from the adipose tissue of young, lean Trpv4−/− and WT mice were isolated and stimulated to differentiate into adipocytes in vitro. After 8 days, greater than 90% of the cells were fully differentiated. Compared to those from WT controls, adipocytes from Trpv4−/− mice showed elevated mRNA expression for “browning” gene program (Figure 5G). Importantly, they had much greater induction in terms of Pgc1α and Ucp1 expression in response to norepinephrine. Conversely, the mRNA expression of proinflammatory genes, such as Mcp1, Mip1α, Mcp3, Tnfα and Vcam were substantially reduced in Trpv4−/− adipocytes (Figure 5H).
Moreover, we analyzed other metabolically active tissues and organs including interscapular brown fat, skeletal muscle and liver; minimal changes in expression of genes of oxidative metabolism were observed in these tissues between the WT and Trpv4−/− mice (Figure S5A, B and C). Together, these data indicate that TRPV4 controls key gene expression programs in adipocytes in an apparently cell autonomous manner.
The TRPV4 mutant mice began to gain significantly less weight after 9 weeks on the HFD, compared to WT controls. Body composition analysis showed that Trpv4−/− mice had gained less fat mass compared to controls (Figure 6A).
Energy expenditure (EE) in these mice was measured after 7 weeks of HFD, before the body weight of mutants diverged from controls. By normalizing oxygen consumption to body weight (Figure 6B) or by ANCOVA analysis (Tschop et al., 2012) using body weight and genotype as covariance (p=0.02 for genotype effect, Table S2), a highly significant increase in EE was observed in the Trpv4−/− mice compared to WT controls. Importantly, there was no significant difference in food intake, physical activity or body temperature between the two groups (Figure S6A–C). Together, these data strongly suggest that the reduced weight gain upon HFD in Trpv4−/− mice was due, at least in part, to an increased EE associated with an elevated thermogenic gene program in the white adipose tissues.
Obesity is associated with chronic “metainflammation” in adipose tissue (Hotamisligil, 2006). Although the causes of the recruitment of immune cells has not been well understood, elevated chemokine expression has been suggested to be critical for the development of inflammation and insulin resistance (Sell and Eckel, 2009).
To understand the impact of the changes in chemokine gene expression caused by TRPV4 deficiency, we analyzed the expression of macrophage markers (F4/80, CD68 and CD11b) to quantify the macrophage infiltration in Trpv4−/− and WT adipose tissues from all three diet groups. As expected, HFD increased the expression of macrophage markers in WT adipose tissue (Figure 6C), indicating that macrophages were actively recruited into adipose tissues. Consistent with the reduction in chemokine expression, Trpv4−/− adipose tissue showed a 40% or 60% reduction in the expression of the macrophage marker after 8 or 16 weeks of HFD, respectively (Figure 6C). Indeed, histologic analysis also showed there were far fewer “crown-like-structures”, previously shown to represent macrophages in fat tissues (Cinti et al., 2005), in the Trpv4−/− adipose tissues (Figure 6D). Importantly, macrophages from Trpv4−/− mice did not have altered program of gene expression either at the basal level or in response to LPS or free fatty acid (Figure S6D, E); this strongly suggests that the difference in macrophage infiltration in adipose tissues was likely due to the altered adipocyte function rather than alterations in the macrophages per se.
To further assess the inflammation in adipose tissues, the mRNA expression of Tnfα, a key cytokine for obesity-induced insulin resistance was measured. HFD significantly increased Tnfα mRNA in WT adipose tissue, while the induction was largely reduced in Trpv4−/− mice (Figure 6F). Furthermore, phosphorylation of serine 273 on PPARγ, a modification that is associated with obesity and insulin resistance (Choi et al., 2010), was strongly attenuated in the Trpv4−/− adipose tissue compared to WT controls (Figure 6E).
Adipose inflammation is associated with insulin resistance. Consistent with the changes in inflammatory markers, Trpv4−/− mice have improved insulin sensitivity, as indicated by the reduced fasting and glucose-stimulated insulin levels (Figure 6G) as well as an improved glucose tolerance (Figure 6H) compared to WT controls. We also observed an improved insulin tolerance (Figure 6I) in the Trpv4−/− mice. Importantly, these changes preceded the alterations in body weight (Figure S6F, G).
The role of TRPV4 in adipose energy metabolism and inflammation makes it a potential target for obesity and insulin resistance. We used theTRPV4 antagonist GSK205 (Phan et al., 2009) to test if pharmacological modulation of TRPV4 could provide metabolic benefits in a mouse model of obesity and insulin resistance.
GSK205 potently antagonized TRPV4 in 3T3-F442A adipocytes, as it effectively blocked the calcium influx caused by TRPV4 agonist (Figure 7A). Treating these adipocytes with GSK205 for 4 days resulted in increased expression of thermogenic genes (Figure 7B) and was also accompanied by a decrease in the proinflammatory gene program. This shift resembled the gene expression changes seen in TRPV4-deficient adipocytes.
GSK205 has a relatively short half-life of 2 hours in the plasma and adipose tissues (Figure S7A). As proof-of-principle, we treated diet-induced obese (DIO) mice with 10mg/kg GSK205 or vehicle twice daily for a short period (7 days). The compound was relatively well tolerated as there were no apparent sign of sickness or weight loss in either group over this period (Figure S7B). Compared to controls, GSK205-treated mice showed significantly increased expression of thermogenic genes such as Ucp1, Pgc1a, Cidea and Cox8b. Drug treatment also caused a reduced expression of the proinflammatory chemokines, macrophage marker and Tnfa (Figure 7C) in the EPI fat. These changes largely recapitulated the molecular phenotypes seen in the Trpv4−/− mice. A less significant trend in the expression of these genes was observed in the SubQ fat (Figure S7C). In contrast, no significant change was observed in the interscapular BAT (Figure S7D). Consistent with these gene expression changes, strikingly, this short term GSK205 treatment significantly improved glucose tolerance in DIO mice, compared to controls (Figure 7D).
Adipocytes play a number of key roles in systemic energy balance and metabolic regulation. First, white adipocytes are the primary depot for energy storage in mammals. This important function is highlighted in the tissue steatosis and illnesses that occurs in individuals with lipodystrophy. Second, in the context of obesity, where energy intake chronically outstrips energy expenditure, adipocytes become enlarged and adipose tissue becomes inflamed. This was first recognized as a greatly increased expression of TNFα and other cytokines in rodent models of obesity (Hotamisligil et al., 1993). While it was originally believed that fat cells themselves made these cytokines, it is now appreciated that most of the secretion of these molecules comes from immune cells, especially macrophages, that infiltrate adipose tissue in elevated numbers in obesity (Weisberg et al., 2003; Xu et al., 2003). Hence, a critical question now is what are the physiological and pathological signals secreted by fat cells that regulate the infiltration and function of these immune cells. Finally, brown adipocytes are an important component in whole body energy homeostasis through the dissipation of stored chemical energy in the form of heat. The role of brown fat as a defense against both hypothermia and obesity, at least in rodents, is well established (Feldmann et al., 2009; Lowell et al., 1993). Adult humans have significant depots of brown fat but the contribution made by these deposits to total energy metabolism is not known.
Despite both being important for obesity and related diseases, thermogenesis and inflammation are ordinarily considered as two separate aspects of adipose biology. As a common mediator for both programs, TRPV4 is one of the first examples indicating that these two programs are connected at the molecular and physiological levels. Moreover, the striking changes of chemokines gene expression in an adipocyte cell line (Figure 3A) suggest that signaling from TRPV4 could serve as an early trigger of immune cell chemoattraction. Hence, the convergence of thermogenesis and inflammation identified here not only suggests a novel angle for targeting obesity, but also provides a new perspective on understanding the origin of adipose inflammation and insulin resistance.
It is interesting that TRPV4 has been shown to be activated by cellular swelling (Liedtke et al., 2000; Strotmann et al., 2000) and by cellular stretch (Gao et al., 2003; Mochizuki et al., 2009; Thodeti et al., 2009). Since adipocytes become very large in obesity, it is possible that this cellular distention activates TRPV4 and leads to the changes in gene programs. The precise mechanisms by which TRPV4 signals are obscure but it is clear that ERK1/2 activation is very important for the effects.
TRPV4 deficiency protected mice from diet-induced obesity and insulin resistance. Although the animals studied here have a global Trpv4 deletion, there are several reasons to believe that white adipose tissues contribute significantly to the phenotypes. First, the gene expression changes in Trpv4−/− adipose tissue largely recapitulated what we observed in cultured adipocytes, and many of those changes preceded metabolic intervention and the difference in physiological parameters. In contrast, minimal differences in these key metabolic pathways were observed in other metabolically active organs (liver, muscle) in the same time frame (Figure S5A, B and C). These data strongly suggest that the phenotypes seen in white adipose tissue are unlikely to be secondary to those organs. However, with current data, we certainly cannot exclude potential contributions from the CNS and the sympathetic nervous system in particular. Also, it is interesting that a very recent study of Trpv4−/− mice has also shown a resistance to diet induced obesity (Kusudo et al., 2011). Although this paper did not examine adipose tissues in detail, they showed alterations in muscle biology and fiber-type switching in the soleus muscle. It is not clear how this could affect energy balance and obesity, but the role of TRPV4 in multiple tissues will be important for future studies.
Our proof-of-principle study suggests that pharmacologic inhibition of TRPV4 leads to an elevation of the thermogenic gene program and a reduction in adipose tissue inflammation; both of which could provide therapeutic benefits for obesity and metabolic diseases. Although TRPV4 is highly expressed in fat, it is also expressed in many other tissues and has been implicated in osmotic regulation (Liedtke and Friedman, 2003), bone formation (Masuyama et al., 2008) and bladder dysfunction (Gevaert et al., 2007). Hence, the size of the therapeutic window of TRPV4 antagonists in metabolic diseases may depend on those functions. In particular, several TRPV4 mutations associated with neurodegenerative disease have been recently identified in humans (Jia et al., 2010; Landoure et al., 2010; Phelps et al., 2010). Further characterization on the nature of these mutations would be valuable in determining the therapeutic value of TRPV4 antagonists.
It should be considered that other closely related TRPVs, such as TRPV1, may also regulate one or both pathways controlled by TRPV4 in fat. Several reports have suggested that TRPV1 and TRPM8 could affect adipose function (Ma et al., 2012; Motter and Ahern, 2008; Zhang et al., 2007). Nonetheless, the fact that the genetic ablation and chemical inhibition of TRPV4 has a cell-autonomous effect on both thermogenic and proinflammatory programs in white adipocytes in vivo, makes TRPV4 in particular a very promising target for treating obesity and type 2 diabetes.
Antibody sources are: UCP1, tubulin and OXPHOS (Abcam), TRPV4 (Alomone), p-ERK1/2, ERK1/2, p-JNK, JNK, p-p38, p38 (Cell Signaling), PGC1α (Calbiochem). BCTC was from Tocris. SLV319, CAY10508 and rosiglitazone were from Cayman. U0126 and SP600125 were from Cell Signaling. GSK205 was synthesized at the Scripps Research Institute. Other chemicals are from Sigma. shRNA constructs were in pLKO vectors or pMKO vectors. Calcium-free DMEM was made by adding 2.5mM EGTA into DMEM (Cellgro). Sequences for all shRNA and primers are listed in Supplemental Table 3.
Briefly, after two days of differentiation, 3T3-F442A adipocytes were trypsinized and split into 384 well plates (3000 cell/well). At day 6, cells were treated with the bioactive library (Broad Institute) for 20 hours. mRNA was harvested using the TurboCapture kit (Qiagen), reverse transcribed to cDNA, and quantified by qPCR. All values were normalized to DMSO treated cells.
All animal experiments were performed according to procedures approved by the IACUC of Dana-Farber Cancer Institute. Mice were on a standard chow or a 60% high-fat diet (Research Diets) with 12-hour light cycles. Trpv4−/− mice were provided by Dr. Liedtke and back-crossed to C57BL/6J background. Unless specified, male mice were used for experiments. Each group contains 9–16 animals. For drug treatment, C57BL/6J mice were on HFD for 14–15 weeks before treatment. GSK205 (in DMSO) was dissolved in a vehicle contains 5% Tween80 and 90% saline before intraperitoneal injection. Each group contains 10–12 mice and experiments were performed twice in independent cohorts.
For glucose tolerance tests, animals were fasted overnight. Glucose levels in tail blood were measured with a standard glucometer prior to and at indicated intervals following an intraperitoneal injection of D-glucose. For insulin tolerance tests, animals were fasted for 4 hours before experiments. Fat and lean mass was measured by MRI. Energy expenditure was evaluated using a Comprehensive Lab Animal Monitoring System (Columbia Instruments). Mice were acclimated in the metabolic chambers for 2 days before the experiments. CO2 and O2 levels were collected every 32 minutes for each mouse over a period of 2 days and normalized to total body weight. Movement and food intake are measured more frequently at regular intervals.
For virus production, 293T (for lentivirus) or phoenix cells (retrovirus) were transfected with Fugene 6 (Roche) with viral vectors. Viral supernatant was harvested 48 hours later. 3T3-F442A pre-adipocytes were infected for 4 hour (lenti) or overnight (retro), followed by puromycin selection (2ug/ml). 3T3-F442A adipocyte differentiation was induced s by treating confluent cells with 850 nM insulin for 8–10 days. To stimulate thermogenesis, cells were incubated with norepinephrine for 4 hours. For primary adipocytes, SVF from inguinal fat of 5-week-old male mice were prepared and differentiated for 8 days as previously described (Kajimura et al., 2009).
Student’s t-test was used for single comparisons. Two-way ANOVA (repeated measurement) was used for GTT and ITT. Unless specified, * P<0.05, ** P<0.01, *** P<0.001, n.s. P>0.05.
We thank Drs. Zoltan Arany (Beth Israel Deaconess Medical Center) and Bridget Wagner (The Broad Institute) for help with screen setup and Dr. Kai Cui (Duke University) for assistance with statistical analysis. We are grateful to Drs. Patrick Seale and Chih-Hao Lee for useful discussion. We also thank Yingying Zhang, Diti Bhowmick and Lingling Dai for technical assistance. L.Y. was supported by the Interdisciplinary Training grant 5R90DK071507. This work was supported by NIH grants DK31403, DK080261 (B.M.S.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.