GDE5 inhibition accumulates intracellular glycerophosphocholine and suppresses adipogenesis at a mitotic clonal expansion stage

Abstract

Mammalian glycerophosphodiesterases (GDEs) were recently shown to be involved in multiple cellular signaling pathways. This study showed that decreased GDE5 expression results in accumulation of intracellular glycerophosphocholine (GPC), showing that GDE5 is actively involved in GPC/choline metabolism in 3T3-L1 adipocytes. Using 3T3-L1 adipocytes, we further studied the biological significance of GPC/choline metabolism during adipocyte differentiation. Inhibition of GDE5 suppressed the formation of lipid droplets, which is accompanied by the decreased expression of adipocyte differentiation markers. We further showed that the decreased GDE5 expression suppressed mitotic clonal expansion (MCE) of preadipocytes. Decreased expression of CTP: phosphocholine cytidylyltransferase (CCTβ), a rate-limiting enzyme for phosphatidylcholine (PC) synthesis, is similarly able to inhibit MCE and PC synthesis; however, the decreased GDE5 expression resulted in accumulation of intracellular GPC but did not affect PC synthesis. Furthermore, we showed that mRNAs of proteoglycans and transporters for organic osmolytes are significantly upregulated and that intracellular amino acids and urea levels are altered in response to GDE5 inhibition. Finally, we showed that reduction of GDE5 expression increased lactate dehydrogenase release from preadipocytes. These observations indicate that decreased GDE5 expression can suppress adipocyte differentiation not through the PC pathway but possibly by intracellular GPC accumulation. These results provide insight into the roles of mammalian GDEs and their dependence upon osmotic regulation by altering intracellular GPC levels.

INTRODUCTION

Adipocytes are primary depots for triacylglycerol (TAG) storage within the mammalian body. Neutral lipid droplets (LDs) in adipocytes represent intracellular storage for TAG, which is hydrolyzed to release fatty acids via the lipolysis pathway during nutritional deficiency (2, 3). The neutral lipid core of an LD is surrounded by a phospholipid monolayer and associated proteins that are functionally important for LD stability and lipid metabolism. During LD expansion, phosphatidylcholine (PC) must be produced and added coordinately with the increased core TAG amount to maintain LD stability and prevent LD coalescence (15, 24) because, in the case of mammalian adipocytes, the monolayer of phospholipids contains 50–60% PC (23). In addition to adipocyte hypertrophy with LD expansion, an increase in the number of preadipocytes is involved in the increased white adipose tissue mass during the development of obesity (17, 18). Because PC, the major phospholipid component of membranes, has been confirmed to be closely related to cell proliferation (21), increased PC synthesis could be essential for not only adipocyte hypertrophy but also preadipocytes proliferation.

Depending on the cell type, PC homeostasis is regulated by two independent pathways as follows: 1) the CDP-choline (Kennedy) pathway, which is a main route for PC synthesis dependent on intracellular choline and 2) the phosphatidylethanolamine (PE) methylation pathway in which PC is newly generated from PE by sequential methylation reactions by phosphatidylethanolamine N-methyltransferase (PEMT) (19). In the CDP-choline pathway, a rate-limiting enzyme, CTP:phosphocholine cytidylyltransferase-α (CCTα), is reversibly associated with membranes of the nuclear envelope and/or cytoplasmic LDs and regulates PC synthesis (24). Alternatively, in the PE methylation pathway, a reaction catalyzed by PEMT consumes S-adenosylmethionine as the major biological methyl donor. Intracellular choline is irreversibly oxidized to betaine, which functions as a methyl donor for the conversion of homocysteine to methionine, which is used as an intermediate of S-adenosylmethionine (14, 35). Taken together, these observations indicate that intracellular choline is an essential determinant of PC synthesis and methylation reactions; however, the choline supply route in mammalian cells has been only slightly elucidated.

It has been reported that intracellular choline concentration is regulated by two pathways in the yeast Saccharomyces cerevisiae (28, 29). The first pathway is a direct uptake through choline transporter proteins, and the second pathway is through the action of glycerophosphocholine (GPC)-glycerophosphodiesterase (GDE), Gde1, in which GPC catabolized from PC is further hydrolyzed to choline (9, 28, 29). Reversibly, intracellular choline generated by Gde1 is reused back into PC through the CDP-choline pathway (9). To date, seven mammalian GDEs have been identified by us and other groups and investigated for their physiological functions, particularly as an intracellular determinant for glycerophosphoinositol or GPC levels inside cells (6, 33, 43). We previously isolated GDE5 and showed that GDE5 is a unique GDE because it does not have a transmembrane region, as seen for other mammalian GDEs; instead, it has an N-terminal carbohydrate binding domain that is often found in glycosylhydrolases (34). We further demonstrated that the GDE5 protein is localized in the cytoplasm and selectively hydrolyzes GPC in an in vitro reaction, suggesting that GDE5 possibly plays a variety of intracellular roles via GPC regulation (34). In this study, we showed that GDE5 is actually expressed in mouse adipocytes and first examined the involvement of GDE5 in GPC/choline metabolism. As described above, because choline is an important nutrient for PC synthesis within adipocytes, we hypothesized a critical role of GDE5 as a supply route of intracellular choline and further explored the biological significance of GDE5 as an osmotic regulator through the control of intracellular GPC levels.

MATERIALS AND METHODS

Materials.

3T3-L1 and NIH-3T3 cells were obtained from Health Science Research Resources Bank (Sennan, Japan). Dulbecco’s modified Eagle’s medium and fetal calf serum were purchased from Invitrogen (Carlsbad, CA). Glycerophosphocholine, choline, and phosphocholine were products of Sigma (St. Louis, MO).

Cell culture, siRNA transfection, and cell volume.

Mouse 3T3-L1 preadipocytes and mouse NIH-3T3 fibroblasts were cultured in a maintenance medium (10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin in Dulbecco’s modified Eagle’s medium) at 37°C in 5% CO₂-95% humidified air. Confluent 3T3-L1 preadipocytes were treated with differentiation medium [maintenance medium plus 0.5 mM 3-isobutyl-1-methylxanthine (Sigma), 5 μg/ml insulin, and 1 μM dexamethasone (Sigma) (MDI)] and incubated for 2 days. Then, differentiation medium was replaced with growth medium (maintenance medium supplemented 5 μg/ml insulin), which was refreshed every 2 days. Stealth siRNA duplex oligoribonucleotides against mouse GDE5 and mouse CCTβ were synthesized by Invitrogen. The sequences were as follows: mouse GDE5, sense 5′-CAGUGUGUUGUGGAAAGCAGUGAUU-3, antisense 5′-AAUCACUGCUUUCCACAACACACUG-3′ and mouse CCTβ, sense 5′-AAACUCUGGAGUAAGAGUCCAUGGC-3′, antisense 5′-GCCAUGGACUCUUACUCCAGAGUUU-3′. 3T3-L1 and NIH-3T3 cells were transfected with siRNAs at a final concentration of 20 nM using LipofectAMINE RNAimax (Invitrogen). Average cell volume was measured by using Sceptor (Miliipore). After 2 days of siRNA transfection, at least > 10,000 cells were trypsinized and subjected to the measurement of average cell volume.

LC-MS/MS analyses.

3T3-L1 adipocytes were homogenized using a mixture of water and methanol with a ratio of 1:4 (vol/vol) and transferred to a screw capped glass tube. The mixture of water and methanol (1:4, vol/vol) was added to the tube to bring the volume up to 400 μl. Chloroform was added to yield a water/methanol/chloroform ratio of 1:4:2 (vol/vol/vol). Samples were shaken for 1 h. Chloroform (160 μl) and water (160 μl) were added to achieve the phase separation. After vortexing for 1 min, the samples were centrifuged at 1,800 g for 15 min. The upper aqueous layer was dried, dissolved in MeOH, and subjected to UPLC-MS analysis. The UPLC-MS analysis was carried out using an Acquity UPLC system (Waters, Milford, MA) coupled to Acquity TQD tandem quadrupole mass spectrometer (Waters) with electrospray ionization (ESI) in the positive ion-mode electrospray ionization (29). Samples were injected onto a BEH HILIC column (Waters; 2.1 × 50 mm, 1.7 μm) at a flow rate of 0.4 ml/min using gradient elution with 5 mM ammonium formate (pH 3.0) in 80% acetonitrile (A) and 5 mM ammonium formate (pH 3.0) in 5% acetonitrile (B) as follows: 0–0.1 min 100% A, 0.1–2.75 min 0% B, 2.75–5.0 min 100% B. Analyte detection was performed using single ion recording: mass-to-charge ratio (m/z) 104, m/z 258, and m/z 184.5 for choline, GPC, and phosphocholine, respectively. Cone voltage was 25 V. Dwell time was 50 ms.

The lower layer was dried, dissolved in MeOH, and subjected to UPLC-MS analysis with ESI in the positive modes. Sample solution was injected onto a reversed-phase BEH C18 column (Waters, 2.1 × 50 mm, 1.7 μm) at a flow rate of 0.4 ml/min using gradient elution with 10 mM ammonium acetate (pH 5.0) in 40% acetonitrile (A) and 10 mM ammonium acetate (pH5.0) in acetonitrile and isopropanol with a ratio of 1:9 (vol/vol) (B) as follows: 0–3.0 min 20% B → 100% B, 3.0–4.0 min 100% B. Column temperature was maintained at 40°C. Analyte detection was performed using multiple reaction monitoring with the following transition: m/z 758.7 → 184.2 for 16:0/18:2 PC. Cone voltage was 60 V and collision energy 5 V. Dwell time was 200 ms. Peak area ratios of the analyte to the internal standard (IS) were calculated as a function of the concentration ratios of the analyte (QuanLynx, Waters).

Western blotting.

A polyclonal antibody for GDE5 was previously obtained by injecting rabbits with mouse GDE5 protein (amino acids 1–163) and validated by Western blotting using the homogenates from C2C12 myoblasts transfected with luciferase siRNA or GDE5 siRNA (34). 3T3-L1 adipocytes were washed with ice-cold PBS and scraped in ice-cold homogenization buffer (40 mM Tris-HCl, pH 7.5, 15 mM benzamidine, 5 μg/ml pepstatin A, and 5 μg/ml leupeptin). The cell suspension was disrupted with a handheld sonicator for 15 s (three times with 1-min intervals), and the homogenates were centrifuged at 10,000 × g for 30 min. Supernatants (10 μg) were subjected to SDS-PAGE. Proteins separated by SDS-PAGE were transferred to an Immobilon P filter (Millipore). The blots were blocked for 18 h at 4°C by soaking in 5% nonfat dried milk in PBS and incubated for 18 h at 4°C with anti-GDE5 antibody (diluted 1:1,000), anti-β-actin (2F3) monoclonal antibody (Wako, Osaka, Japan), or anti-α-tubulin (MAS 077) monoclonal antibody (Harlan, Indianapolis). Signals were detected using horseradish peroxidase-conjugated anti-rabbit IgG and enhanced chemiluminescence systems (GE Healthcare).

Oil Red O staining.

3T3-L1 adipocytes were fixed with 10% formalin solution for 5 min and washed with water and then 60% isopropyl alcohol. After being washed, cells were incubated in Oil Red O (Sigma) staining solution: 1.8 mg/ml in 60% isopropyl alcohol at 37°C for 30 min. The excess staining solution was removed by washing with 60% isopropyl alcohol and water. Spectrophotometric measurement for Oil-red O staining was performed by dissolving the stained LDs in the cells with isopropyl alcohol, and then the absorbance was measured at 520 nm.

RNA analyses.

The reverse transcriptase reaction was carried out with 1 μg total RNA as a template to synthesize cDNA using ReverTra Ace (TOYOBO, Osaka, Japan) and random hexamers (TaKaRa Bio, Kyoto, Japan), according to the manufacturer’s instructions. For quantitative PCR analysis, cDNA and primers were added to the THUNDERBIRD SYBR qPCR Mix (TOYOBO), to give a total reaction volume of 20 µl. PCR reactions were then performed using StepOnePlus (Applied Biosystems, Foster City, CA). Conditions were set to the following parameters: 2 min at 95°C, followed by 40 cycles each of 15 s at 95°C and 1 min at 60°C. The primers used for PCR analyses were as follows: GDE5, forward, 5′-TTTGATGTCCACCTTTCAAAGGAC-3′ and reverse, 5′-CTCCATCCCTGTGTTGGCAAATCC-3′; L19, forward, 5′-GGGCATAGGGAAGAGGAAGG-3′ and reverse, 5′-GGATGTGGTCCCATGAGGATGC-3′; CCTα, forward, 5′-GCCAGCTCCTTTTTTCTGATG-3′ and reverse, 5′-TCATCACATGAAGCCCTTG-3′; CCTβ, forward, 5′-TGCCATGGGAGTTACTAGGG-3′ and reverse, 5′-GTGCAAGGCTCTTTGAGGAC-3′; choline kinase-α (CKα), forward, 5′-GCCATTCTTGCAGAGAGGTC-3′ and reverse, 5′-GGATCTTGTGCAGTTGGTGA-3′; CKβ, forward, 5′-TCTGCCACAATGACATCCAG-3′ and reverse, 5′-TCCCGTTTCTTCTGTTCCTC-3′; Cept, forward, 5′-TTCAGTTACCGACACCACCA-3′ and reverse, 5′-AATGAAAGGCCACAAGCAC-3′; UT-B, forward, 5′-GAAGGAATGGAAGCAGCAAC-3′ and reverse, 5′-TGAGTGTGGCAGTGGAAAAC-3′; osteoglycin, forward, 5′-GACCTGGAATCTGTGTGCCTCC-3′ and reverse, 5′-GTTTGGATGCTTTCCCAGAG-3′; decorin, forward, 5′-TGAGCTTCAACAGCATCACC-3′ and reverse, 5′-AAGTCATTTTGCCCAACTGC-3′; epiphycan, forward, 5′-AAGAGCTGGTGGTTCTGGGTGAC-3′ and reverse, 5′-TAGAGGGATGTGGTCCAAGC-3′; lumican, forward, 5′-TGCAGTGGCTCATTCTTGAC-3′ and reverse, 5′-GCAGCTTGCTCATCATGATTG-3′; C/EBPα, forward, 5′-TGGACAAGAACAGCAACGAG-3′ and reverse, 5′-CCTTGACCAAGGAGCTCTCA-3′; aP2, forward, 5′-AGCGTAAATGGGGATTTGGT-3′ and reverse, 5′-TCGACTTTCCATCCCTCTTC-3′; PPARγ, forward, 5′-ATCAGCTCCTGTGGAGCCTCTC-3′ and reverse, 5′-GATGCTTTAATCCCCACAGAC-3′; slc6a9, forward, 5′-CTGTCTGGCAACCTGTCTCA-3′ and reverse, 5′-GAAGACAACCACCCAGGAGA-3′; FAS, forward, 5′-TGGGTTCTAGCCAGCAGAGT-3′ and reverse, 5′-ACCACCAGAGACCGTTATGC-3′.

Bromodeoxyuridine incorporation assay.

3T3-L1 preadipocytes were induced to differentiate on a cover glass in a 6-well plate. Eighteen hours later, BrdU (Sigma) was added to the culture to a final concentration of 50 μM, and the cells were incubated for an additional 2 h. Then, cells were fixed with ice-cold methanol for 10 min, denatured with 1.5 N HCl, and neutralized with 0.1 m sodium borate (pH 8.5). After blocking with ImmunoBlock (Dainipponseiyaku) at 4°C overnight, cells were incubated with anti-BrdU antibody (DAKO) in phosphate-buffered saline at room temperature for 2 h. Cells were next incubated with anti-mouse IgG-Cy3 (GE Healthcare) in PBS at room temperature for 1 h. The cover glass was mounted with PBS/glycerol (1:1) containing DAPI on a slide glass and fluorescence was observed. Cy3 and DAPI-positive cells were counted in six different microscopic views from three different cover glasses for each sample.

DNA microarray.

Total RNAs were isolated using RNeasy lipid tissue kit (Qiagen Sciences, Germantown, MD), and pooled RNAs were subjected to cRNA synthesis for a DNA microarray analysis according to the manufacturer’s instructions (44K whole-mouse genome 60-mer oligo microarray, Agilent Technologies, Palo Alto, CA). All procedures of fluorescence labeling, hybridization, and imaging were carried out according to the manufacturer’s instructions (38). In this experiment, each comparison was hybridized to two arrays employing a DyeSwap method to eliminate the bias between dyes because the difference between Cyanine 3-CTP and Cyanine 5-CTP altered the efficiency of hybridization. Gene expression data were obtained using Agilent Feature Extraction software, using defaults for all parameters except ratio terms, which were changed according to the Agilent protocol to fit the direct labeling procedure. Files and images, including error values and P values, were exported from the Agilent Feature Extraction Program (version 9.5). The microarray data are also deposited in the National Center for Biotechnology Information Gene Expression Onmibus database (available on the World Wide Web at https://www.ncbi.nlm.nih.gov/geo) under accession no. GSE83644.

Measurement of amino acid and urea concentration.

3T3-L1 adipocytes were washed twice with ice-cold PBS, suspended in 500 μl of distilled water, and boiled for 15 min. The suspension was centrifuged for 3 min at 5,000 rpm, and the supernatant was collected. This extract was analyzed with an amino acid analyzer JLC-500/V2 (JEOL, Tokyo, Japan).

Statistical analyses.

Values are presented as means ± SE. Statistical significance was determined by one-way ANOVA and Duncan’s multiple-range test. Differences were considered significant if P < 0.05.

RESULTS

GDE5 controls GPC/choline metabolism in 3T3-L1 adipocytes.

In our previous study, to characterize the substrate specificity for GDE5, mouse recombinant GDE5 protein was purified from baculovirus-infected Sf9 cells and showed a marked substrate preference for GPC in vitro (34). We first characterized the expression patterns of GDE5 mRNA and protein in 3T3-L1 development. The GDE5 transcripts and proteins were increased at day 3 and expressed during 3T3-L1 adipocyte maturation (Fig. 1, A and B). Because we have previously shown that GDE5 selectively hydrolyzes GPC in vitro, we examined intracellular GPC level in 3T3-L1 development. Intracellular GPC level was decreased at day 3 (Fig. 1C). Next, to examine the involvement of GDE5 in the hydrolysis of GPC in differentiated 3T3-L1 cells, we used siRNA-mediated GDE5 knockdown in 3T3-L1 adipocytes. The amount of reduced protein and mRNA expression was assessed by Western blotting and quantitative RT-PCR (Fig. 2). As shown in Fig. 2, GDE5 RNAi gene silencing resulted in significant intracellular GPC accumulation, whereas phosphocholine and choline concentrations were lower than in control 3T3-L1 adipocytes. These observations strongly suggest a regulatory role of GDE5 for GPC metabolism in differentiated 3T3-L1 cells. We measured the cell volume of 3T3-L1 adipocytes and calculated the concentrations of these choline metabolites, showing that the intracellular concentration of GPC was much higher than those of choline and phosphocholine and that the reduced GDE5 expression resulted in the accumulation of intracellular GPC (Fig. 3).

Fig. 1.

GDE5 mRNA and protein expression and intracellular GPC level during differentiation of 3T3-L1 cells. A: on days 0, 3, and 9 after induction of differentiation, total RNAs were extracted and subjected to quantitative PCR to examine mRNA expression level of GDE5. All values were normalized to L19 levels. The data (mean ± SE) are representative of two independent experiments. *P < 0.05 compared with day 0. B: on days 0, 3, and 9 after induction of differentiation, whole cell lysates were obtained, and total protein extracts (10 μg/lane) were subjected to SDS-PAGE followed by Western blotting using an anti-GDE5 or anti-β-actin antibody. C: intracellular GPC was measured as described in materials and methods (n = 4). *P < 0.05, compared with day 0. GDE, glycerophosphodiesterase; GPC, glycerophosphocholine.

Fig. 2.

Effects of GDE5 siRNA on intracellular levels of choline metabolites in differentiated 3T3-L1 adipocytes. 3T3-L1 preadipocytes treated with MDI for 2 days were differentiated into mature adipocytes for 7 days, and then transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5) for 2 days. At day 9 of differentiation, total protein extracts (10 µg/lane) from transfected 3T3-L1 adipocytes were subjected to SDS-PAGE followed by Western blotting using anti-GDE5 or anti-α-tubulin antibody. Total RNAs were extracted and subjected to quantitative PCR analyses to examine expression levels of GDE5 mRNA. All values were normalized to L19 levels. Intracellular GPC, phosphocholine (P-choline), and choline levels were measured as described in materials and methods. All values were normalized to protein levels and are expressed as mean ± SE (n = 5). The data are representative of two independent experiments. *P < 0.05, **P < 0.01 compared with those of cells transfected with control siRNA (siLuc). GDE, glycerophosphodiesterase; GPC, glycerophosphocholine.

Fig. 3.

Effects of GDE5 interference on intracellular GPC metabolism in differentiated 3T3-L1 adipocytes. 3T3-L1 adipocytes were treated with MDI and differentiated into mature adipocytes for 7 days. Differentiated 3T3-L1 adipocytes were transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5). After 2 days of transfection (day 9), 3T3-L1 adipocytes were homogenized to measure intracellular GPC, phosphocholine (P-Cho), and choline (Cho) levels (n = 5). In this experiment, 3T3-L1 adipocytes were separately subjected to the measurement of the cell volume as described in materials and methods (n = 5). The average cell volume of 3T3-L1 adipocytes transfected with siLuc or siGDE5 showed 5.50 μl and 5.55 μl, respectively. In this figure, intracellular concentration of choline metabolites was expressed. GDE, glycerophosphodiesterase; GPC, glycerophosphocholine; P-Cho, phosphocholine.

Inhibition of GDE5 expression suppresses 3T3-L1 adipocyte differentiation.

To explore the biological significance of GDE5 expression in 3T3-L1 adipocytes, we examined the effect of siRNA-mediated GDE5 knockdown on 3T3-L1 adipose differentiation (Fig. 4). Transfected 3T3-L1 preadipocytes were grown to confluence and then induced to differentiate with a standard adipogenic cocktail, MDI. Nine days after induction of adipose differentiation, lipid droplets in 3T3-L1 adipocytes were visualized by Nile Red (Fig. 4A) or Oil Red O (Fig. 4B) staining, showing that lipid accumulation was severely impaired in adipocytes transfected with siRNA directed for GDE5 with no observable change in the number of DAPI-positive cells (Fig. 4A). On the other hand, when we conducted siRNA transfection at 6 days after induction of adipose differentiation, 3T3-L1 adipocytes transfected with GDE5 siRNA showed no change in LD formation compared with control 3T3-L1 adipocytes (Fig. 4C). Consistent with the defects in adipogenesis, as shown in Fig. 4D, expression of C/EBPα and PPARγ, transcription factors required for adipogenesis, and of adipocyte marker aP2, a fatty acid-binding protein (FAS), was reduced in 3T3-L1 adipocytes transfected with siRNA directed toward GDE5. These results imply that the inhibitory effect of siRNA-mediated GDE5 knockdown on adipogenesis is primarily attributable to its inhibitory effect during the early phase of adipogenesis. The molecular mechanisms underlying hormonal and nutritional regulation of adipocyte differentiation have been extensively studied. In response to MDI, growth-arrested confluent 3T3-L1 preadipocytes are known to reenter the cell cycle for an additional two rounds of division, known as mitotic clonal expansion (MCE) (7). Within 24 h after induction by MDI, activated MCE concomitantly initiates a well-programmed series of transcriptional activation events with expression of early adipogenic transcription factors, including C/EBPα, C/EBPβ, and PPARγ. During the MCE, phosphatidylcholine must be produced and added coordinately with the increased phospholipid membrane, because PC is the major membrane phospholipid in mammalian cells, and the cellular phospholipid mass is regulated by its biosynthesis and turnover during cell division (23, 32, 40). Because after the MCE confluent preadipocytes are rearrested during the cell cycle for differentiation into mature adipocytes (41), PC biosynthesis is supposed to be downregulated to meet the low demands of phospholipid membrane. We compared the mRNA expression level of enzymes that are involved in the CDP-choline (Kennedy) pathway (Fig. 5A) between growth-arrested confluent preadipocytes and mature adipocytes. This study showed that CTP: phosphocholine cytidylyltransferase β (CCTβ) mRNA expression is dramatically decreased during 3T3-L1 adipocyte differentiation. In addition, previous studies showed that endogenous CCTα knockdown did not alter PC synthesis in preadipocytes (1), suggesting that CCTβ is a key regulatory step in PC biosynthesis for the proliferation of 3T3-L1 preadipocytes (Fig. 5).

Fig. 4.

Effects of GDE5 siRNA on 3T3-L1 adipocyte differentiation. A: 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5) and followed by treatment with MDI for 48 h. 3T3-L1 cells were differentiated into mature adipocytes for 9 days and intracellular lipid droplets were visualized by Nile Red (NR). Nuclei were stained with DAPI. Scale bar, 20 μm. B: 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5) and followed by treatment with MDI for 48 h. 3T3-L1 cells were differentiated into mature adipocytes for 9 days and intracellular lipid droplets were visualized by Oil-Red-O staining (n = 5). C: after 3T3-L1 cells were differentiated into mature adipocytes for 6 days, 3T3-L1 adipocytes were transfected with luciferase siRNA or GDE5 siRNA. At day 9 of differentiation, intracellular lipid droplets were visualized by Oil-Red-O staining (n = 5). Scale bar, 20 μm. D: 3T3-L1 preadipocytes were transfected with luciferase siRNA or GDE5 siRNA and followed by treatment with MDI for 48 h. 3T3-L1 cells were differentiated into mature adipocytes. At day 2 of differentiation, total RNAs were extracted and subjected to quantitative PCR analyses to examine expression levels of C/EBPα, PPARγ, aP2, and FAS mRNAs. All values were normalized to L19 levels (n = 5). The data (mean ± SE) are representative of two independent experiments. *P < 0.05, **P < 0.01 compared with those of cells transfected with control siRNA (siLuc). GDE, glycerophosphodiesterase; GPC, glycerophosphocholine.

Fig. 5.

Altered mRNA expression related to phosphatidylcholine synthesis during 3T3-L1 adipogenesis. A: CDP-choline (Kennedy) pathway. B–F: 3T3-L1 preadipocytes were treated with MDI for 48 h. 3T3-L1 cells were differentiated into adipocytes as described in materials and methods. At day 3 of differentiation, total RNAs were extracted from 3T3-L1 preadipocytes (pre) and differentiated 3T3-L1 cells (d) and subjected to quantitative PCR to examine mRNA expression level of choline kinase α (CKα) (B), choline kinase β (CKβ) (C), CTP:phosphocholine cytidylyltransferase α (CCTα) (D), CTP:phosphocholine cytidylyltransferase β (CCTβ) (E) and choline/ethanolamine phosphotransferase (Cept) (F). All values were normalized to L19 levels (n = 4). The data (mean ± SE) are representative of two independent experiments. *P < 0.05 compared with those of 3T3-L1 preadipocytes (pre).

Next, we examined the effect of siRNA-mediated GDE5 or CCTβ knockdown on the MCE in the early phase of adipogenesis. We assessed DNA synthesis during the MCE by measuring the incorporation of bromodeoxyuridine (BrdU). Knockdown of endogenous GDE5 or CCTβ by siRNA in 3T3-L1 adipocytes resulted in reduction of BrdU-positive cells compared with control cells (Fig. 6). In these experiments, we showed that GDE5 knockdown did not affect the number of DAPI-positive cells (Fig. 6). Then, we measured the concentrations of choline metabolites, showing that PC level was actually lower than that of control cells by siRNA-mediated CCTβ knockdown (Fig. 7). Consistent with previous reports (40), the reduction of PC level by CCTβ knockdown suppressed cell replication during the MCE. However, interestingly, GDE5 knockdown did not affect the PC level but resulted in a significant accumulation of intracellular GPC instead. Taken together, these observations indicated that siRNA-mediated GDE5 has an inhibitory effect on the early phase of adipogenesis by the suppression of the MCE independently of the regulation of PC biosynthesis.

Fig. 6.

GDE5 siRNA suppressed mitotic clonal expansion of preadipocytes. 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc), GDE5 siRNA (siGDE5) or CCTβ (siCCTβ) and followed by treatment with MDI for 18 h, and then subjected to bromodeoxyuridine (BrdU) incorporation assay as described in materials and methods. Anti-BrdU antibody was visualized by Cy3-labeled goat anti-mouse IgG and nuclei were stained with DAPI. Cy3 and DAPI-positive cells were counted in six different microscopic views for each sample as described in materials and methods. The data (mean ± S.E.) are representative of two independent experiments. *P < 0.05 compared with those of cells transfected with control siRNA (siLuc). Scale bar, 20 μm. GDE, glycerophosphodiesterase.

Fig. 7.

Effects of GDE5 interference on choline metabolism in 3T3-L1 preadipocytes. 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc), GDE5 siRNA (siGDE5) or CCTβ (siCCTβ) and treated with MDI for 18 h. Intracellular GPC, choline, and PC levels were measured as described in materials and methods (n = 4). The data (mean ± SE) are representative of two independent experiments. *P < 0.05, **P < 0.01 compared with those of cells transfected with control siRNA (siLuc). GDE, glycerophosphodiesterase; GPC, glycerophosphocholine.

Inhibition of GDE5 expression affects gene expression in 3T3-L1 adipocyte differentiation.

We further explored the biological significance of decreased GDE5 expression that is accompanied by an accumulation of intracellular GPC. We hypothesized that characterization of the gene expression profile with siRNA-mediated GDE5 knockdown would be highly informative for understanding the molecular basis of the suppression of the MCE by decreased GDE5 expression. In this study, we performed microarray analysis using RNA samples from 3T3-L1 adipocytes and showed that numerous gene clusters are altered in 3T3-L1 adipocytes with knockdown of endogenous GDE5 (Fig. 8). Because GDE5 knockdown has a strong inhibitory effect in the early phase of adipogenesis, we expected that the genes whose expression is related to the adipocyte differentiation would be included. Mouse fibroblastic NIH-3T3 cells, which are not differentiated into adipocytes, have been well studied to show a similar phenotype to 3T3-L1 cells when PPARγ is expressed (30, 44). To isolate the genes whose expression is regulated directly by the knockdown of endogenous GDE5, we used NIH-3T3 cells as the control cells that are not committed to adipocytes and performed a similar siRNA experiment. After transfection of siRNA for GDE5 into NIH-3T3 cells, we confirmed that GDE5 RNAi gene silencing decreased GDE5 mRNA levels and resulted in accumulation of intracellular GPC in NIH-3T3 cells (Fig. 8A). Moreover, we performed gene expression analysis on NIH-3T3 cells transfected with GDE5 siRNA and control siRNA. Then, we carried out a comparative analysis based on these 3T3-L1 and NIH-3T3 transcriptomic data to isolate candidate genes that were directly affected by GDE5 RNAi gene silencing. In this study, as shown in Table 1 and Fig. 8B, we identified 54 genes whose expression was significantly upregulated more than twofold in 3T3-L1 adipocyte and NIH-3T3 cell expression profiles. On the contrary, the expression of 33 genes was downregulated more than twofold in both 3T3-L1 adipocytes and NIH-3T3 cells. Interestingly, the expression of several proteoglycan genes was strongly upregulated with the siRNA-mediated GDE5 knockdown. Quantitative RT-PCR also showed that mRNA transcripts for osteoglycin, decorin, epiphycan, and lumican were more abundant in both 3T3-L1 adipocytes and NIH-3T3 cells with GDE5 siRNA than in control cells (Fig. 8C). Because previous reports showed that GPC accumulates in renal inner medullary cells in response to high NaCl and urea and functions as a counteracting organic osmolyte to protect cells against a hypertonic environment (4, 5, 13), we further examined the expression of transporters that can transport intracellular osmolytes such as glycine and urea. mRNA expression of a glycine transporter GLYT1 (slc6a9) and an urea transporter UT-B were regulated by GDE5 RNAi gene silencing in 3T3-L1 adipocytes (Fig. 9A). As shown in Fig. 9A, the intracellular glycine and urea levels are significantly altered in response to the decreased GDE5 expression, strongly suggesting that the intracellular GPC accumulation stimulates an osmotic stress response. To examine the possibility that osmotic stress suppresses the MCE during 3T3-L1 adipogenesis, we calculated the average intracellular concentration of GPC in 3T3-L1 adipocytes transfected with siGDE5 as shown in Fig. 3 and then added GPC at that same concentration in the medium to maintain the osmotic balance against accumulated GPC inside of the 3T3-L1 adipocytes. Because GPC exogenously added in the medium partially cancelled the MCE suppression induced by decreased GDE5 expression (Fig. 9B), it can be suggested that osmotic changes by altering intracellular GPC levels affected the MCE of 3T3-L1 adipogenesis. Finally, we measured lactate dehydrogenase (LDH) release into the medium from 3T3-L1 adipocytes to assess whether downregulation of GDE5 induces cell necrosis. A significant increase in plasma membrane permeabilization, as assessed by LDH release, was observed in the presence of GDE5 siRNA, suggesting that the decrease in GDE5 expression consequently induced cell necrosis (Fig. 9C).

Fig. 8.

Analysis of two transcriptomes to isolate genes whose expression is upregulated by GDE5 interference. A: NIH-3T3 fibroblasts were transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5) (n = 3). After 48 h, intracellular GPC was measured as described in materials and methods. Total RNAs were extracted and subjected to quantitative PCR analyses to examine expression level of GDE5 mRNA. All values were normalized to L19 levels (n = 4). The data (mean ± SE) are representative of two independent experiments. *P < 0.05 compared with those of NIH-3T3 fibroblasts transfected with control siRNA (siLuc). B, The Venn diagram shows genes that are upregulated in 3T3-L1 adipocytes and NIH-3T3 fibroblasts (n = 2). Of a total of 318 genes upregulated in 3T3-L1 adipocytes, expression of 54 genes was also increased in NIH-3T3 fibroblasts by GDE5 siRNA (P < 0.05). C: different sets of total RNA samples were prepared from 3T3-L1 adipocytes 48 h after transfection of luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5) and subjected to quantitative PCR to show mRNA levels of osteoglycan, decorin, epiphycan, and lumican in 3T3-L1 adipocytes (n = 4). All values were normalized to L19 levels. The data (mean ± SE) are representative of two independent experiments. **P < 0.01 compared with those of cells transfected with control siRNA (siLuc). GDE, glycerophosphodiesterase; GPC, glycerophosphocholine.

Table 1.

Analysis of two transcriptomes to isolate genes whose expression is upregulated by GDE5 interference

			3T3-L1		NIH-3T3
Gene Name	Systematic Name	Description	Fold	P value	Fold	P value
Lum	NM_008524	Lumican	9.3	0.00	6.3	0.00
Slc14a1	NM_028122	Solute carrier family 14 (urea transporter), member 1	7.9	0.00	2.6	0.00
Chgb	NM_007694	Chromogranin B	6.4	0.00	2.1	0.00
Lama2	U12147	Laminin-2 α2 chain	6.4	0.00	2.8	0.00
Epyc	NM_007884	Epiphycan	6.1	0.00	10.3	0.00
Sgk3	NM_133220	Serum/glucocorticoid regulated kinase 3	5.4	0.00	2.2	0.03
Kera	NM_008438	Keratocan	5.4	0.00	2.2	0.03
Hlf	NM_172563	Hepatic leukemia factor	5.2	0.00	3.2	0.00
Dcn	AK052759	Decorin	5.1	0.00	6.3	0.00
Me2	AK033595	Adult male cecum cDNA, clone:9130022D06	4.5	0.00	3.2	0.00
Egfr	AK033431	Adult male colon cDNA, clone:9030024J15	4.4	0.00	3.8	0.00
Cyp2f2	NM_007817	Cytochrome P450, family 2, subfamily f, polypeptide 2	4.3	0.00	6.8	0.00
Il33	NM_133775	Interleukin 33	4.3	0.00	2.2	0.00
Ypel1	NM_023249	Yippee-like 1 (Drosophila)	4.3	0.00	2.6	0.00
C130026I21Rik	NM_175219	C130026I21 gene	4.1	0.01	2.1	0.00
Cdh26	NM_198656	Cadherin-like 26	4.1	0.01	2.9	0.00
Aldh1a7	NM_011921	Aldehyde dehydrogenase family 1, subfamily A7	3.6	0.00	12.5	0.00
A530050D06Rik	NM_001081169	A530050D06 gene	3.5	0.00	2.2	0.00
Fgfr2	NM_010207	Fibroblast growth factor receptor 2	3.4	0.00	3.2	0.00
Ablim1	AK029371	0 Day neonate head cDNA, clone:4833406P10	3.2	0.00	2.9	0.00
6330442E10Rik	NM_178745	6330442E10 gene	3.1	0.00	2.8	0.00
Sort1	NM_019972	Sortilin 1	3.1	0.00	2.5	0.00
Lbp	NM_008489	Lipopolysaccharide binding protein	3.0	0.00	2.3	0.00
Prelp	NM_054077	Proline arginine-rich end leucine-rich repeat	2.9	0.00	3.0	0.00
Glrb	NM_010298	Glycine receptor, β- subunit	2.9	0.02	2.1	0.00
Colec10	NM_173422	Collectin subfamily member 10	2.9	0.01	3.7	0.00
Agt	NM_007428	Angiotensinogen (serpin peptidase inhibitor, clade A, member 8)	2.9	0.00	4.3	0.00
C3	NM_009778	Complement component 3	2.8	0.00	4.1	0.00
Selenbp2	NM_019414	Selenium binding protein 2	2.6	0.00	3.1	0.00
Nid2	NM_008695	Nidogen 2	2.5	0.00	2.5	0.00
Svep1	NM_022814	Sushi, von Willebrand factor type A, EGF, and pentraxin domain containing 1	2.4	0.00	4.5	0.00
Bmf	NM_138313	BclII modifying factor	2.4	0.00	2.0	0.00
2900064A13Rik	AK033552	Adult male colon cDNA, clone:9030618C06	2.3	0.00	2.2	0.00
Zdhhc8	NM_172151	Zinc finger, DHHC domain containing 8	2.3	0.00	2.3	0.00
Calml4	NM_138304	Calmodulin-like 4	2.3	0.00	3.7	0.00
AI256396	AK134520	11 Days embryo head cDNA, clone:6230431N17	2.2	0.00	2.1	0.00
Ctso	NM_177662	Cathepsin O	2.2	0.00	3.0	0.00
Aytl1	NM_173014	Acyltransferase like 1 (Aytl1), mRNA NM_173014	2.2	0.00	2.2	0.00
Col5a3	NM_016919	Procollagen, type V, α3	2.1	0.00	2.8	0.00
Igtp	NM_018738	Interferon γ−induced GTPase	2.1	0.00	3.9	0.00
Setd7	NM_080793	SET domain containing (lysine methyltransferase) 7	2.1	0.00	2.1	0.00
Adam19	NM_009616	A disintegrin and metallopeptidase domain 19 (meltrin β)	2.1	0.00	2.2	0.00
Sgcb	NM_011890	Sarcoglycan, β (dystrophin-associated glycoprotein)	2.1	0.00	2.1	0.00
Edg3	NM_010101	Endothelial differentiation, sphingolipid G protein-coupled receptor, 3	2.1	0.00	2.2	0.00
Grpr	NM_008177	Gastrin releasing peptide receptor	2.1	0.00	2.2	0.00
Irs1	AK141842	12 Days embryo spinal ganglion cDNA, clone:D130017A09	2.1	0.00	2.3	0.00
Iigp2	NM_019440	Interferon inducible GTPase 2	2.0	0.00	4.2	0.00
Rasl11b	NM_026878	RAS-like, family 11, member B	2.0	0.00	2.0	0.00
Selenbp1	NM_009150	Selenium binding protein 1	2.0	0.01	7.4	0.00
Sned1	NM_172463	Sushi, nidogen and EGF-like domains 1	2.0	0.00	2.1	0.00
Nbl1	NM_008675	neuroblastoma, suppression of tumorigenicity 1	2.0	0.00	3.2	0.00
AI836003	NM_177716	Expressed sequence AI836003	2.0	0.00	4.7	0.00
Rab36	NM_029781	RAB36, member RAS oncogene family	2.0	0.00	2.1	0.00
Aspn	NM_025711	Asporin	2.0	0.00	4.7	0.00

DNA microarray analysis was repeated with the Cy3 and Cy5 dyes reversed (a dye swap). Fold change represents the average of mRNA expression level in 3T3-L1 adipocytes and NIH-3T3 cells transfected with GDE5 siRNA relative to cells transfected with control siRNA. GDE, glycerophosphodiesterase.

Fig. 9.

Osmotic balance is possibly changed by GDE5 interference and affects the MCE of 3T3-L1 adipogenesis. A: UT-B and slc6a9 expression in 3T3-L1 adipocytes is affected by GDE5 interference. Total RNAs from 3T3-L1 adipocytes were prepared as shown in Fig. 8C and subjected to quantitative PCR to examine mRNA expression level of UT-B and slc6a9 (n = 4). *P < 0.05 compared with those of cells transfected with control siRNA (siLuc). Forty-eight hours after transfection, intracellular fraction of 3T3-L1 adipocytes was analyzed with an amino acid analyzer. All values were normalized to protein concentration (n = 5). B: 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc) or GDE5 siRNA (siGDE5), followed by treatment with MDI for 18 h in the presence or absence of GPC and then subjected to BrdU incorporation assay. The data (n = 6, mean ± SE) are representative of two independent experiments. C: 3T3-L1 preadipocytes were transfected with luciferase siRNA (siLuc), GDE5 siRNA (siGDE5), or CCTβ (siCCTβ) and treated with MDI for 18 h. Condition mediums from 3T3-L1 preadipocytes were subjected to measurement of lactate dehydrogenase release (n = 5). The data (n = 5, mean ± SE) are representative of two independent experiments. *P < 0.05 compared with those of cells transfected with control siRNA (siLuc). GDE, glycerophosphodiesterase; GPC, glycerophosphocholine; MCE, mitotic clonal expansion.

DISCUSSION

In our previous study, we identified a novel cytosolic glycerophosphodiesterase, GDE5, and showed that the purified recombinant GDE5 protein selectively hydrolyzed GPC to produce choline in vitro (34), suggesting that intracellular GPC produced from PC is reused as a source of choline upon hydrolysis by GDE5 (9). Because intracellular levels of PC are critical for cell proliferation and LD formation in adipocytes (15, 24), which are involved in enlargement of adipose tissue mass during obesity development, we investigated the role of GDE5 as a regulator of intracellular PC levels during adipocyte differentiation by using 3T3-L1 preadipocytes. This study showed that GDE5 RNAi gene silencing in preadipocytes suppressed the expression of adipogenic factors, C/EBPα and PPARγ, and synthesis of lipid droplets and the inhibition of adipogenesis by GDE5 knockdown was due to the reduction in cell replication during MCE, which was accompanied by a significant increase in intracellular GPC levels. Although MCE is a unique step that occurs in in vitro cell culture, it has been shown that not only the volume of adipocytes but also the number of adipogenic progenitor cells contribute to the expansion of white adipose tissue mass in vivo (10, 22). In this study, we showed that inhibition of GDE5 suppressed MCE through osmotic stress, and thus it may affect the proliferation of adipogenic progenitors during obesity development in vivo.

GPC is widely recognized as an important organic osmolyte abundant in cells in the renal inner medulla (4, 5, 11, 13). GPC accumulates in renal medullary cells in response to either high extracellular NaCl or urea (5, 13). Previous studies by Gallazzini et al. (12) have shown that a mammalian membrane-bound glycerophosphodiesterase, GDE2, contributes to osmotic regulation as a GPC phosphodiesterase. High NaCl levels reduced GDE2 mRNA abundance and inhibited GPC degradation in response to hyperosmotic stress due to a high extracellular NaCl concentration in mIMCD3 cells (12, 42). This study showed that GDE5 can hydrolyze intracellular GPC efficiently because GDE5 is a cytosolic glycerophosphodiesterase without a transmembrane domain and possibly regulates intracellular osmotic balance by regulating intracellular GPC levels. Because GDE2 mRNA expression was undetectable in 3T3-L1 adipocytes by RT-PCR analysis (data not shown), it is possible that GDE5 might be the sole cytosolic enzyme responsible for intracellular GPC hydrolysis in 3T3-L1 adipocytes. To date, numerous studies have been conducted to characterize the cellular response to hyperosmotic stress because hyperosmotic conditions are easily established by the addition of high amounts of NaCl or urea to the extracellular space (4, 5) In contrast, this study showed an intriguing possibility that GDE5 inhibition mimics the cellular response to hypo-osmotic stress in mammalian cells through significant accumulation of intracellular GPC and provides novel information on how mammalian cells react to hypo-osmotic stress. In both 3T3-L1 adipocytes and NIH-3T3 cells, serum/glucocorticoid regulated kinase-3 (sgk-3) mRNA was identified to be upregulated by siRNA-mediated GDE5 knockdown by using microarray analysis. Sgk has been identified as a novel protein kinase family that is transcriptionally stimulated by serum and glucocorticoids (27). Interestingly, expression of the sgk-1 gene has been reported to be induced by various stress stimuli, such as hyperosmotic or hypo-osmotic stress (31). In particular, Rozansky et al. (37) have shown that hypotonic conditions stimulated sgk-1 mRNA expression in A6 cells, a cultured cell line derived from the distal nephron of Xenopus laevis. These findings suggest that sgk-3 mRNA is also expressed and plays a part in the response to the hypo-osmotic stress with intracellular GPC accumulation. Here, GDE5 inhibition actually altered intracellular glycine levels, accompanied by the regulation of the related transporter mRNA expression. GLYT1, a glycine transporter, regulates cell volume by regulation of the steady-state level of intracellular glycine, an organic osmolyte, in the early stage of the mouse preimplantation embryo. In the current study, a significant decrease in glycine concentration in the cells, possibly through the downregulation of GLYT1 expression, may have counteracted the intracellular accumulation of GPC by decreased GDE5 expression via a volume-regulatory mechanism. On the other hand, previous studies have shown that UT-B enhances osmotically driven water transport when expressed in Xenopus oocytes (45). UT-B-facilitated water transport was blocked by urea transport inhibitors, suggesting that UT-B upregulation may be involved in water permeability in response to a significant accumulation of intracellular GPC (46).

Another intriguing feature is that GDE5 inhibition significantly upregulated the mRNA expression of several proteoglycans such as lumican, decorin, and epiphycan, which possibly respond to GPC accumulation in cells and are small leucine-rich proteoglycans that participate in extracellular matrix (ECM) assembly (20, 39). In particular, these proteoglycans bind to various types of collagens and elastic fibril components, thereby regulating the activity and stability of proteins and signaling molecules within the matrix (20, 39). These proteoglycans regulate cell differentiation and proliferation of various types of cells. Previous studies have shown that overexpression of decorin enhances the differentiation of myoblasts (26) and increased lumican expression promotes monocyte differentiation into fibrocytes (36); both decorin and lumican inhibit cell proliferation (8, 25). Furthermore, proteoglycans in the articular cartilage provide osmotic pressure within the tissue to resist compressive loads (16); however, to our knowledge, this is the first study to show the possibility that proteoglycans play an important role in the maintenance of osmotic balance in response to hyperosmotic or hypo-osmotic stress in mammalian cells. The increase in the expression of these proteoglycans upon suppression of GDE5 expression may counteract the osmotic stress caused by intracellular GPC accumulation, mainly through ECM rearrangement at the cell surface (20, 39). Thus, this study provides a novel approach to investigate proteoglycan functions for the maintenance of osmotic balance through modulation of GDE5 activity in mammalian cells.

GRANTS

J. Sands and J. Klein were supported by NIH grant no. R01-DK41707. This work was supported in part by grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to N. Yanaka).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.O. and T.I. conceived and designed research; K.N., S.T., I.Y., F.Y., and N.O. performed experiments; K.N., S.T., I.Y., F.Y., and N.Y. analyzed data; Y.O., J.D.K., J.M.S., and N.Y. interpreted results of experiments; N.Y. prepared figures; Y.O., J.D.K., T.K., J.M.S., and N.Y. drafted manuscript; Y.O., T.K., and N.Y. edited and revised manuscript; Y.O., K.N., S.T., I.Y., and N.Y. approved final version of manuscript.