Genetic Defect in Phospholipase Cδ1 Protects Mice From Obesity by Regulating Thermogenesis and Adipogenesis

PLCδ1^−/− mice were generated previously and genotyped with tail by PCR using a mixture of the following three primers: forward (5′-CAAGGAGGTGAAGGACTTCCTG-3′), reverse (5′- CTGGGTCAGCATCCTGTAGAAG-3′), and neomycin (5′- CCTGTGCTCTAGTAGCTTTACG-3′) (16). Mice had ad libitum access to water and either regular diet (RD) (CLEA Rodent diet CE-2; 12.6% of calories from fat; CLEA Japan, Tokyo, Japan) or high-fat diet (HFD) (CLEA Rodent diet Quick Fat; 30.6% of calories from fat; CLEA Japan). For diet-induced obesity, the mice were fed with HFD from the age of 6 weeks to 27 weeks. We performed experiments with male mice.

Blood glucose was measured directly with a blood glucose meter (Sanwa Kagaku Kenkyusho, Nagoya, Japan). Plasma insulin or leptin concentration was measured by an insulin ELISA kit or leptin ELISA kit (Shibayagi, Shibukawa, Japan). For glucose tolerance tests, mice were fasted for 16 h and injected intraperitoneally with glucose (2 g/kg body wt). For insulin tolerance tests, mice with ad libitum access to diets were intraperitoneally injected with human regular insulin (0.25 or 0.75 units/kg for RD or HFD, respectively; Eli Lilly, Indianapolis, IN).

The 24-week-old PLCδ1^+/− and PLCδ1^−/− mice fed with RD or HFD were subjected to metabolic analysis. Indirect calorimetry was performed with a computer-controlled open circuit calorimetry system (Oxymax; Columbus Instruments) composed of respiratory chambers. For measurement of oxygen consumption (VO₂) and carbon dioxide production (VCO₂), mice were individually housed in respiratory chambers to acclimate them for 1 day before measurement. Data were recorded for 2–3 days. Respiratory quotient was calculated as the VCO₂ to VO₂ ratio. Heat generation can also be calculated with the following expression: heat = calorie value × VO₂ (calorie value = 3.815 + 1.232 × VCO₂/VO₂).

Western blot analysis was carried out as described previously (18). Anti-PLCδ1 antibody was developed previously (18). Anti-UCP1 (Santa Cruz Biotechnology, Santa Cruz, CA), Tim23 (BD Biosciences, Tokyo, Japan), heat shock protein 60 (Hsp60) (Stressgen), cytochrome c (Cell Signaling), PKCβI (Santa Cruz Biotechnology), PKCε (Cell Signaling,), caveolin-1 (BD Biosciences), nuclear factor of activated T (NFAT)c4 (Santa Cruz Biotechnology), lamin B1 (Santa Cruz Biotechnology), and β-actin (Sigma, St. Louis, MO) antibodies were purchased, respectively.

Total RNAs from tissues and cells were isolated using an RNeasy Lipid tissue mini-kit or an RNeasy mini-kit (Qiagen, Venlo, the Netherlands). Template cDNA was synthesized from total RNA using the SuperScript III First-Strand Synthesis System (Invitrogen, San Diego, CA). Quantitative real-time PCR (qRT-PCR) was performed using a Thunderbird SYBR qPCR Mix (Toyobo, Osaka, Japan) with specific primer sets (Supplementary Table 1) in a CFX96 thermocycler (Bio-Rad, München, Germany). The relative amount of mRNA was normalized to 36B4 mRNA.

pMX-Ires Puro (IP) (19) was used to overexpress several genes, such as PLCδ1, or SV40 large T antigen into target cells. pSUPER retro puro (OligoEngine) was used for the expression of siRNA in target cells. The sequences used were as follows: scrambled (5′-GTAAGATGAGCTTCAAGGA-3′), 399i (5′-GGACCAGCGCAATACCCTA-3′), and 468i (5′-GGATAACAAGATGAACTTC-3′). For retrovirus preparation, indicated constructs were transiently transfected into the packaging cell line, PLAT-E cells (19) using Lipofectamine 2000 (Invitrogen). Target cells were maintained in a medium containing 1.5–2 μg/mL puromycin to select bulk cell populations stably transformed with the viruses.

3T3L1 preadipocytes were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 10% calf serum. For adipocyte differentiation, 2-day postconfluent cells were maintained in DMEM containing 10% FBS, 5 μg/mL insulin, 0.3 mmol/L 3-isobutyl-1-methylxanthine (IBMX; Sigma), and 0.25 μmol/L dexamethasone for 3 days and incubated in DMEM with 10% FBS and 5 μg/mL insulin for an additional 4 days (20). Differentiated adipocytes were fixed with 4% paraformaldehyde and stained with 0.5% Oil Red O. For quantitative analysis, the lipid droplets were eluted with isopropylalcohol, and the absorbance was measured at 510 nm.

SVF was prepared as reported previously (21). SVF cells were plated at 8 × 10⁵ per well of a 24-well plate and grown in DMEM supplemented with 10% FBS and 10 ng/mL βFGF (R&D Systems, Minneapolis, MN). After 2 days of incubation, the cells were incubated in differentiation medium with 10% FBS, 1 μg/mL insulin, 0.5 mmol/L IBMX, and 0.25 μg/mL dexamethasone for 3 days and then with 10% FBS for an additional 4 days. Fluorescence-activated cell sorting (FACS) analysis was performed using FACSCanto (Becton Dickinson, Franklin Lakes, NJ) to define stem cells from WAT SVF cells by staining with Ter119-FITC and CD45-APC.

Immortalized brown preadipocytes were obtained from interscapular BAT of newborn PLC1^+/− mice and PLC1^−/− mice littermates by collagenase digestion, immortalized by infection with SV40 large T antigen retrovirus, and selected by 2 μg/mL puromycin for at least 3 weeks (22). Seven immortalized brown preadipocyte lines were established from independent littermates. For differentiation, 2-day-confluent brown preadipocytes were incubated for 3 days in culture medium supplemented with 20 nmol/L insulin, triiodothyronine (T3; Sigma, St. Louis, MO), 0.125 mmol/L IBMX, 0.5 μmol/L dexamethasone, and 0.5 μmol/L indomethacin. Subsequently, the cells were maintained in culture medium supplemented with 20 nmol/L insulin and 1 nmol/L T3 for 4 days. To stimulate thermogenesis, differentiated brown adipocytes at the same degree were treated with 0.5 mmol/L cAMP and 0.1 mmol/L forskolin for 4 h.

Data are expressed as the mean ± SEM. Statistical significance was assessed using the Student t test. A P value of < 0.05 was considered statistically significant.

Because we noticed that PLCδ1^−/− mice seem to be leaner than PLCδ1^+/− mice, we measured the body weights of mice fed with RD or HFD from the age of 6 weeks to 20 weeks. PLCδ1^−/− mice had a decreased body weight compared with PLCδ1^+/− littermate mice on both RD and HFD (Fig. 1A), suggesting that the absence of the PLCδ1 gene conferred protection from obesity. Decreased body fat mass was also observed in PLCδ1^−/− mice. The weights of epididymal WAT (eWAT) and BAT were extremely lower in PLCδ1^−/− mice on both diets. The weight of the liver was also lower in PLCδ1^−/− mice on HFD, whereas those of most other tissues were almost the same (Fig. 1B). Chronic exposure of mice to HFD causes enlarged body mass and accumulation of lipids in eWAT, BAT, and liver. HFD induced increased mass of eWAT in PLCδ1^+/− mice but not in PLCδ1^−/− mice (Fig. 1C). Moreover, BAT and liver in PLCδ1^−/− mice on HFD were darker than those in PLCδ1^+/− mice, indicating less accumulation of lipid in these tissues (Fig. 1D and E). Hematoxylin/eosin staining revealed that adipocyte sizes of eWAT and BAT were extremely smaller in PLCδ1^−/− mice than in PLCδ1^+/− mice on HFD (Fig. 1F and G). All these data suggest that the absence of the PLCδ1 gene prevents obesity.

Excessive lipid accumulation and hypertrophy in adipose tissues cause insulin resistance, which leads to a compensatory increase in insulin secretion and a decrease in glucose uptake (1–4). Chronic exposure to HFD remarkably increased blood glucose and plasma insulin levels in the fasting state in PLCδ1^+/− mice (Fig. 2A and B). However, these increases were not observed in PLCδ1^−/− mice, demonstrating that PLCδ1^−/− mice might be protected from HFD-induced insulin resistance. Intraperitoneal injection of glucose induced comparable levels of increase in blood glucose in both PLCδ1^+/− and PLCδ1^−/− mice on RD; however, PLCδ1^−/− mice were more glucose tolerant than PLCδ1^+/− mice on HFD (Fig. 2C). Similarly, PLCδ1^−/− mice fed with HFD were more sensitive to intraperitoneal injection of insulin compared with PLCδ1^+/− mice in the insulin tolerance test (Fig. 2D). Plasma leptin was significantly lower in PLCδ1^−/− mice (Fig. 2E).

Obesity is primarily caused by an excess food intake relative to energy expenditure. PLCδ1^−/− mice had similar food intake compared with PLCδ1^+/− mice on either RD or HFD (Fig. 3A). We then examined whether PLCδ1^−/− mice would have a higher energy expenditure. PLCδ1^−/− mice showed a significant increase in oxygen consumption and heat production compared with PLCδ1^+/− mice (Fig. 3B). This result indicates that PLCδ1^−/− mice have a higher energy expenditure and therefore a higher metabolic rate. Because these enhancements were observed throughout the light and dark phases, an increase in locomotion may not be involved. The respiratory quotient was also examined as a measure of fuel-partitioning patterns. No significant differences were observed between PLCδ1^+/− and PLCδ1^−/− mice either on RD or HFD (Fig. 3B).

Because WAT, BAT, liver, and skeletal muscle are involved in energy metabolism, we examined the expression pattern of PLCδ1 among these metabolic tissues. qRT-PCR analysis showed that the relative expression level of PLCδ1 was very high in WAT and BAT, low in muscles, and very low in the liver (Supplementary Fig. 1). This tissue-specific expression suggests that PLCδ1 possibly contributes to the pathogenesis of obesity-related metabolic disorders in adipose tissues.

Given that we observed inhibition of WAT hypertrophy, better glucose tolerance, and increased insulin sensitivity in PLCδ1^−/− mice, we examined the expression patterns of genes related to energy metabolism in WAT from mice aged 8 and 24 weeks on HFD by qRT-PCR analysis (Fig. 4A and Supplementary Fig. 2A). In 24-week-old PLCδ1^−/− mice, the expression of peroxisome proliferator–activated receptor γ (PPARγ) was increased, whereas p21, which is related to WAT hypertrophy (23), was extremely reduced compared with PLCδ1^+/− mice. The expression of genes related to glucose uptake, including glucose transporter 4 (GLUT4), Krüppel-like zinc finger transcription factor (KLF15), and adiponectin, which is correlated with insulin sensitivity (20,24), was enhanced. On the other hand, the expression of genes related to insulin resistance, such as tumor necrosis factor (TNFα) and heparin-binding epidermal growth factor (HB-EGF)-like growth factor (1,2,25), was decreased in PLCδ1^−/− mice.

In BAT of PLCδ1^−/− mice at both ages 8 and 24 weeks, the expression of thermogenic genes UCP1 and PPARγ coactivator 1α (PGC1α) (5,6) was increased (Fig. 4B). The expression of fatty acid synthase (Fasn) and adrenergic receptor β3 (Adrb3) was also enhanced in BAT of PLCδ1^−/− mice (Supplementary Fig. 2B). These data, along with the gross appearance of BAT (Fig. 1D), suggest that the functions of BAT are well sustained in PLCδ1^−/− mice, even after long-term HFD feeding.

In the liver of PLCδ1^−/− mice, decreases in expression of genes related to glucose and lipid metabolism, such as acyl-CoA oxidase 1 (Acox1), carnitine palmitoyl transferase 1a (Cpt1a), glucokinase, and acyl-CoA dehydrogenase medium chain (Acadm), were observed at the age of 24 weeks (Fig. 4C and Supplementary Fig. 2C). Although these results are consistent with the observation that PLCδ1^−/− mice are resistant to liver adiposity (Fig. 1E), it may be an adipose tissue–dependent secondary effect, because PLCδ1 is not expressed in the liver (Supplementary Fig. 1). Little change in gene expression was observed in muscles of PLCδ1^−/− mice (Fig. 4D).

We have reported that PLCδ1^−/− mice have a hair defect (16). Because there are limited data on the effect of hairlessness on metabolic rate, we studied this relationship using nude mice and C57BL/6 mice with hair removed. As shown in Supplementary Fig. 3, nude mice and mice with hair removed showed increases in oxygen consumption and heat production compared with control mice, indicating that hairlessness at least partially affects metabolic rate through the change in thermogenesis.

A more important question is whether PLCδ1 is directly involved in thermogenesis. We first compared the UCP1 expression in BAT mitochondria from 27-week-old mice fed HFD. The protein levels of UCP1 as well as Tim23, an inner mitochondrial membrane protein, and heat shock protein (HSP)-60 were extremely enhanced in BAT of PLCδ1^−/− mice (Fig. 5A). Amount of mitochondrial DNA was also increased in PLCδ1^−/− mice (Supplementary Fig. 3). We next examined the change in expression of PLCδ1 mRNA in BAT from PLCδ1^+/− mice placed in cold surroundings (4°C) for 3 h. As predicted, the expression levels of thermogenic genes UCP1, deiodinase iodothyronine type II (DIO2), and PGC1α in BAT were increased by cold exposure (5,6,26,27) (Fig. 5B). In contrast, interestingly, the expression level of PLCδ1 in BAT of PLCδ1^+/− mice was decreased by cold exposure (Fig. 5B), strongly suggesting that PLCδ1 has a role in cold exposure–induced thermogenesis.

Similar results were obtained by using immortalized brown preadipocytes. Immortalized brown preadipocytes from PLCδ1^+/− and PLCδ1^−/− mice littermates were developed, differentiated, and then treated with cAMP and forskolin to induce thermogenesis (27). It is worth noting that the increase in UCP1 expression is more remarkable in PLCδ1^−/− immortalized brown adipocytes than in PLCδ1^+/− adipocytes (Fig. 5C) and that this phenomenon is independent from hairlessness. With an inverse correlation, the expression of PLCδ1 was reduced by induction of thermogenesis. Furthermore, the enhancement of UCP1 expression in PLCδ1^−/− adipocytes induced by thermogenesis was canceled by ectopic expression of PLCδ1 in PLCδ1^−/− immortalized brown adipocytes (Fig. 5D). These results clearly indicate that PLCδ1 is especially involved in adaptive thermogenesis in BAT and immortalized brown adipocytes and thereby in energy expenditure.

We next tried to examine the involvement of PLCδ1 in brown adipocyte differentiation in vitro. The expression of UCP1 and PGC1α after differentiation was remarkably enhanced in PLCδ1^−/− adipocytes (Fig. 5E). The expression of PPARγ, PR domain containing 16 (PRDM16), and Cpt1b was also increased in PLCδ1^−/− adipocytes. These data suggest that PLCδ1 is negatively involved in the regulation of UCP1 or PGC1α expression in differentiation.

We next examined the effect of PLCδ1 on differentiation of WAT in vitro. Two sequence segments of PLCδ1 for RNA interference effectively reduced PLCδ1 expression in 3T3L1 preadipocytes by infection of the retrovirus (Fig. 6A). In PLCδ1-KD 3T3L1 adipocytes, the accumulation of lipid droplets was largely inhibited compared with cells infected by control retrovirus (Fig. 6B). In contrast, when PLCδ1 was ectopically expressed into 3T3L1 preadipocytes by the retrovirus (Fig. 6C), lipid accumulation after the induction of adipocyte differentiation was promoted (Fig. 6D), indicating that PLCδ1 positively regulates the differentiation of 3T3L1 preadipocytes.

We next examined the gene expression patterns during differentiation of PLCδ1KD 3T3L1 preadipocytes (Fig. 6E). Although expression levels of early differentiation genes, such as C/EBPδ and C/EBPβ, seemed to be similar between control and PLCδ1KD 3T3L1 preadipocytes, the expression levels of PPARγ and C/EBPα were markedly decreased in PLCδ1KD 3T3L1 preadipocytes at days 3 and 6 after differential induction. Similarly, remarkable decreases in expression of KLF15, GLUT4, aP2, and resistin were detected. These data suggest that PLCδ1 has important roles in the differentiation of 3T3L1 preadipocytes around from the early stage to the middle stage.

We further examined the effect of PLCδ1 on adipogenesis using WAT SVF from mice. SVF is considered to be an enriched fraction of stem cells in WAT (21). SVF was first isolated and then differentiated into adipocytes. Interestingly, lipid accumulation in PLCδ1^−/− WAT SVF was reduced to less than half of that in PLCδ1^+/− SVF (Fig. 6F). FACS analysis indicated that the cell number of WAT SVF and population of lineage-negative (Ter119^–, CD45^–) cells in WAT SVF were almost the same between PLCδ1^−/− and PLCδ1^+/− WAT SVF, indicating that the reduced lipid accumulation of PLCδ1^−/− WAT SVF was caused by the differentiation potential, but not the number of lineage-negative cells. Taken together, these results indicate that PLCδ1 positively regulates both adipogenesis and hypertrophic lipid accumulation in WAT model culture cells.

We further tried to examine the role of PLCδ1 in WAT development in mice. To exclude the effect of hairlessness of PLCδ1^−/− mice, we analyzed mice at 6 days of age, before hair growth. Even as early as 6 days of age, inguinal WAT mass was significantly decreased in PLCδ1^−/− mice (Fig. 7A). We also detected the decreases in the expression of C/EBPβ, PPARγ, C/EBPα, KLF15, aP2, GLUT4, and lipoprotein lipase (LPL) in inguinal WAT of PLCδ1^−/− mice compared with those of PLCδ1^+/− mice. These expression profiles are generally consistent with those of 3T3L1 cells, indicating that PLCδ1 is involved in adipose development and has functional roles in adipose tissues in mice.

To provide mechanisms downstream of PLCδ1 in differentiation of adipocytes, we focused on PKC and NFAT, which are targets of the second messengers diacylglycerol and/or inositol 1,4,5-triphosphate/calcium. Among PKC isozymes, PKCβ—a conventional type of PKC—was reported to be involved in adipocyte differentiation (28), and PKCβKO mice were leaner and more resistant to HFD-induced obesity (29). When PKCβ is activated, PKCβ is translocated from the cytosol to the plasma membrane (28). An increase in PKCβI expression at the plasma membrane was significantly observed at 48 h after the induction in control 3T3L1 adipocytes, whereas this increase was less detected in PLCδ1KD adipocytes (Fig. 8A). In addition, PKCε—a novel type of PKC—began to be expressed in the nuclei and is required for adipocyte differentiation (30,31), and functional ablation of PKCε in mice show improved glucose homeostasis in models of type 2 diabetes (32). PKCε expression was increased in the nuclei of control 3T3L1 cells, but not in PLCδ1KD cells at 48 h after the induction (Fig. 8B).

Mice with the NFATc2/NFATc4 gene disruption exhibit defects in fat accumulation and are protected from diet-induced obesity (33). Immunostaining indicated that NFATc4 expressions were induced in control 3T3L1 adipocytes, whereas they were less induced in PLCδ1KD adipocytes (Fig. 8C). Taken together, we identified for the first time PKCβI, PKCε, and NFATc4 as downstream molecules of PLCδ1 in adipogenesis of 3T3L1 cells.

Phosphoinositol (PI) 3-kinase–mediated phosphorylation of insulin receptor substrate or Akt is essential for GLUT4 translocation and glucose uptake (34–36). Therefore, we predicted that PLCδ1 is directly involved in insulin signaling. However, we have not found any relation between PLCδ1 and PI 3-kinase, such as increased phosphorylation of Akt in WAT of PLCδ1^−/− mice. Therefore, the enhanced insulin sensitivity observed in PLCδ1^−/− mice may be explained by the condition of less obesity.

When mice are exposed to cold temperatures, upregulation of UCP1 and DIO2 in BAT is essential for adaptive thermogenesis to maintain body temperature; cold-exposed DIO2^−/− mice became hypothermic because of impaired BAT thermogenesis (26). UCP1^−/− mice are also cold sensitive and show temperature-dependent obesity (37,38). It is noteworthy that PLCδ1 expression had dramatically decreased in BAT from mice exposed to cold and immortalized brown preadipocytes treated with thermogenic inducers, and it showed an inverse correlation with the upregulation of thermogenic genes such as UCP1 and PGC1α (Fig. 5B–D). These observations indicate that PLCδ1 participated in adaptive thermogenesis mediated by UCP1 and PGC1α in a physiological manner.

In vitro analysis further confirmed that PLCδ1 regulates adipogenesis. We indicated that PLCδ1 negatively regulates UCP1 or PGC1α expression during differentiation of brown adipocytes (Fig. 5E), whereas PLCδ1 positively regulates differentiation of preadipocytes in WAT culture cells (Fig. 6B, E, and F). The opposing regulation of PLCδ1 in the WAT and BAT models seems interesting. Detailed future works could provide insights into the mechanism of differentiation decision or conversion between WAT and BAT.

A similar gene expression pattern with 3T3L1 cells was observed in inguinal WAT of 6-day-old PLCδ1^−/− mice, at which age the effect of hair is negligible. This result shows that PLCδ1 is involved in adipose development in adipose tissues in mouse pups. On the other hand, the elevation of PPARγ, GLUT4, or KLF15, and reduced expression of genes related to WAT hypertrophy or insulin resistance in the case of 24-week-old PLCδ1^−/− mice fed with HFD (Fig. 4A), indicate that PLCδ1^−/− WAT sustains normal adipose functions, even under diet-induced hypertrophic conditions. A similar observation was reported in IκB kinase ε (IKKε)^−/− mice. IKK positively regulates the nuclear factor (NF)-κB pathway by phosphorylation and release of inhibitory IκB from NFκB. IKKε^−/− mice are protected from diet-induced obesity and show increased expression of PPARγ, GLUT4, or adiponectin in WAT, as well as enhanced energy expenditure at the age of 22–26 weeks (39). Because NFκB is a downstream effector of PLC, a relationship between PLCδ1 and NFκB would be predicted. It is noteworthy to identify NFATc4, as well as PKCβI and PKCε, as downstream molecules of PLCδ1. PLCδ1 was recently reported to act as an anti-oncogene and regulate the expression of p21 in esophageal squamous cell carcinoma (40). Taken together with these reports, our results indicate the possibility that the loss of PLCδ1 in WAT induces a decrease in p21 expression and inhibits the development of hypertrophy. Reduced hypertrophy development in turn results in a decrease in expression of insulin resistance–inducing genes, TNFα, or HB-EGF and increases in expression of the insulin-sensitive gene adiponectin.

Here, we show for the first time that PLCδ1 contributes to thermogenesis and adipogenesis, and thereby in developing obesity. It is important to further elucidate how PLCδ1 participates in these various pathways, since obesity is an important issue in present society.

This work was supported by a Grant-in-Aid for General Scientific Research and a High-Tech Research Center Project for private universities matching the fund subsidy (to K.F.) from the Japan Ministry of Education, Culture, Sports, Science and Technology. This work was also supported in part by the Mochida Memorial Foundation for Medical and Pharmaceutical Research and Takeda Science Foundation (to K.F.). This work was carried out by the joint research program of the Institute for Molecular and Cellular Regulation, Gunma University (#22002). No other potential conflicts of interest relevant to this article were reported.

M.H. and M.S. researched data, contributed to discussion, and reviewed and edited the manuscript. R.I., R.S., T.U., and To.K. researched data and contributed to discussion. T.S. and Ta.K. researched data, contributed to discussion, and reviewed and edited the manuscript. H.Y. contributed to discussion. Y.N. contributed to discussion and reviewed and edited the manuscript. K.F. contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript.

The authors thank Dr. Toshio Kitamura (Tokyo University) for providing the Plat-A cells and pMX vectors; Shoichi Matsuoka, Fumiko Yoshino, and Kaori Kanemaru (Tokyo University of Pharmacy and Life Sciences) for technical assistance; and Dr. Masataka Asagiri (University of California, San Diego) for valuable advice.