The glycolytic system is selected for ATP synthesis not only in tumor cells but also in differentiated cells. Differentiated osteoblasts also switch the dominant metabolic pathway to aerobic glycolysis. We found that primary osteoblasts increased expressions of glycolysis-related enzymes such as Glut1, hexokinase 1 and 2, lactate dehydrogenase A and pyruvate kinase M2 during their differentiation. Osteoblast differentiation decreased expression of tumor suppressor p53, which negatively regulates Glut1 expression, and enhanced phosphorylation of AKT, which is regulated by phosphoinositol-3 kinase (PI3K). An inhibitor of PI3K enhanced p53 expression and repressed Glut1 expression. Luciferase reporter assay showed that p53 negatively regulated transcriptional activity of solute carrier family 2 member 1 gene promoter region. Inhibition of glycolysis in osteoblasts reduced ATP contents more significantly than inhibition of oxidative phosphorylation by carbonyl cyanide m-chlorophenyl hydrazine. These results have indicated that osteoblasts increase Glut1 expression through the down-regulation of p53 to switch their metabolic pathway to glycolysis during differentiation.

Introduction

Oxidative phosphorylation has higher efficiency of ATP production than glycolysis and is the main source of ATP in mammalian cells under aerobic conditions. In cancer cells, however, most glucose is metabolized to lactate even in the existence of oxygen, a phenomenon well known as the Warburg effect [1]. Aerobic glycolysis is also observed in proliferating cells such as stem cells [1] and some differentiated cell types [2]. For example, astrocytes, the most abundant glia cells in the central nervous system, were reported to utilize aerobic glycolysis in response to the incorporation of glutamate as a neurotransmitter [3]. Through glycolysis, astrocytes generate and secret lactate, which is incorporated into dendrites of neurons for ATP generation by TCA cycle and oxidative phosphorylation [4]. On the other hand, mature T cells up-regulate glucose uptake and glycolytic ATP production when activated by T cell receptor stimulation [5]. These observations suggest that differentiated cells also utilize aerobic glycolysis for proliferation and metabolite production.

Active aerobic glycolysis in osteoblasts has been demonstrated by numerous studies. Early studies in 1960s showed high production of lactate and decreased activity of TCA cycle in bone slices, and estimated that more than 80% of glucose was converted to lactate in bone tissues [6,7]. More recent studies have demonstrated that the extent of aerobic glycolysis becomes more evident during in vitro osteogenic differentiation of calvaria-derived osteoblasts [8]. The first rate-limiting process for glycolytic metabolism is glucose transport across the cell membrane. Tumor cells increase expressions of glucose transporters such as Glut1 and Glut4 to uptake glucose [2]. Especially, Glut1 and Glut3 are thought to be prime transporters of glucose in osteoblasts [9].

Transcription factors including, hypoxia-inducible factor 1-α (HIF-1α) and tumor suppressor p53, contribute to the regulation of Glut1 expression in tumor cells to select aerobic glycolysis for ATP generation [2,10]. The transcription of solute carrier family 2 member 1 (Slc2a1) mRNA encoding Glut1 protein is activated by HIF-1α, whose expression is induced under hypoxia condition of a solid tumor [10]. HIF-1α also increases expressions of glycolytic enzymes including hexokinase 2 (HK2), phosphofructokinase-1 (PFK1), and lactate dehydrogenase A (LDHA) to activate glycolytic metabolism [11]. On the other hand, the expression of Slc2a1 mRNA is negatively regulated by tumor suppressor p53, which is inactivated through many mechanisms such as point mutation, chromosome deletions and amplification of p53 specific ubiquitin ligase, mouse double minute 2 (Mdm2), during oncogenesis [12]. A tumor suppressor protein, p53, possesses apoptotic and anti-proliferative effects on not only cancer cells but also normal cells [13]. Interestingly, p53 in osteoblasts has been reported to play a negative regulatory role in bone formation [14]. The deficiency of p53 gene results in high bone mass, and the conditioned knockout mice of Mdm2 using collagen type 1, α1 (Col1a1) promoter for Cre transgene display reduction in bone formation and osteogenic differentiation in primary culture of osteoblasts [15].

The proliferation and differentiation of osteoblasts are positively regulated by insulin-like growth factor I (IGF-1), which is produced by the liver in response to growth hormone to promote longitudinal bone growth [16,17]. Recently, IGF-1 has been reported to induce Glut1 expression in rat bone-derived osteoblasts to uptake glucose [18]. This report prompted us to explore a possibility that differentiated osteoblasts select aerobic glycolysis similar to astrocytes and activated T cells. Moreover, Wnt signal is reported to induce aerobic glycolysis by increasing glycolytic enzyme expression levels in osteoblasts, and the Slc2a1 transcriptional activity is also positively controlled by an osteogenic transcription factor, Runt-related transcription factor 2 (Runx2), to increase glucose transport into mature osteoblasts [8,19]. However, the association of p53 with the glycolytic metabolic change during osteoblast differentiation remains unclear. In the present study, we show that p53 plays an important role in the regulation of Slc2a1gene expression in primary osteoblasts. Furthermore, we also found that mature osteoblasts select aerobic glycolysis for the production of ATP rather than oxidative phosphorylation.

Materials and methods

Materials

2-Deoxy-d-glucose (2-DG) and phloretin were obtained from Sigma–Aldrich. Carbonyl cyanide 3-chlorophenylhydrazone (CCCP) was from Wako pure Chemical Industries, Ltd (Osaka, Japan). The anti-Glut1 antibody was from Assay Biotechnology (San Francisco, CA). Anti-p53, anti-p21waf1, anti-p27kip1, anti-β-actin, anti-BAX, and anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) antibodies were from Santa Cruz Biotechnology. Anti-phosphoAKT(S473), anti-AKT, and anti-β-catenin were from Cell Signaling Technology (Danvers, MA, U.S.A.). Nutlin 3 was from Funakoshi Co., Ltd. (Tokyo, Japan). Brefeldin A (BFA) was from Nacalai Tesque, Inc. (Kyoto, Japan). 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA) and 4-methylumbel-liferyl-β-d-galactoside (4-MUG) were from Molecular Probe Inc. (Eugene, OR, U.S.A.). Oligomycin A and pifithrin-α (PFTα) was from Adipogen AG (Liestal, Switzerland). 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-d-glucose (2-NBDG) was from Cayman Chemical (Ann Arbor, MI, U.S.A.). Recombinant mouse Wnt3a was from Peprotech Inc. (Rocky Hill, NJ, U.S.A.).

Cell culture

C57BL/6 mice were obtained from Clea Japan (Tokyo, Japan). The animals were maintained and subjected to experiments in accordance with a protocol approved by the Kagoshima University Animal Care and Use Committee (D16025). All animal experiments were conducted in the animal laboratory of the Graduate School of Medical and Dental Sciences, Kagoshima University, and the room was approved for use as an animal experiment room by the Kagoshima University Animal Care Committee (08Shi020).

After euthanasia by cervical dislocation of newborn C57BL/6 mice, the calvaria were sequentially digested with trypsin and collagenase in order to isolate periosteal fibroblast-rich cell fraction (FB fr) and pre-osteoblast-rich cell fraction (pre-OB fr) [20]. Cells of either FB fr or pre-OB fr were cultured in Eagle's α-minimal essential medium (αMEM) (Nacalai, Tokyo, Japan) plus 10% fetal calf serum (FCS) containing 50 units/ml penicillin and 50 mg/ml streptomycin until confluence at 37°C with an atmosphere of 5% CO2. Cells were collected after several days of treatment with 280 µM l-ascorbic acid 2-phosphate trisodium (AA2P) and 5 mM β-glycerophosphate under normal glucose condition (100 mg/dl glucose) or high glucose condition (500 mg/dl glucose) for further analyses. For hypoxic stimulation, cells were incubated in an environment of 5% CO2, 2% O2, and 93% N2 at 37°C.

ST2 cells, a mouse mesenchymal stem cell line, were obtained from RIKEN Cell Bank and maintained in RPMI1640 medium (Wako) plus 10% FCS containing 50 units/ml penicillin, and 50 mg/ml streptomycin. The cells were treated with 280 µM AA2P) and 5 mM β-glycerophosphate under normal glucose condition (5.5 mM glucose) to induce osteoblast differentiation [21].

To inhibit the transcriptional activity of p53, 0- or 3-day differentiated cells in pre-OB fr were treated with 1.2 μM PFTα for 6 h, or ST2 cells transfected with pcDNA3.1-p53 and Src2a1 promoter reporter plasmid were treated with 1.2 μM PFTα for 12 h.

Measurement of lactate concentration and uptake of glucose

Cells in pre-OB fr were cultured in αMEM +10% FCS with 280 µM AA2P and 5 mM β-glycerophosphate for 0, 3, 9, and 14 days, and then 2-NBDG was added to the media at 150 µg/ml. The cells were incubated for 5 or 65 min, and the fluorescence of cell lysates in RIPA buffer(20 mM Tris–HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, 1% sodium deoxycholate) was measured using a TriStar2 LB 942M microplate reader (Berthold Technologies, Baden Württemberg, Germany) at excitation/emission wavelengths of 485/535 nm. The value obtained by subtracting the fluorescence intensity of 5 min from that of 65 min was defined as glucose intake. The value was normalized by the amount of DNA in the cells.

Lactate in conditioned media was determined by using a commercial lactate assay kit (Dojindo Molecular Technologies, Kumamoto, Japan) according to the instruction manual. The lactate concentration of conditioned medium was determined by the absorbance at 405 nm, and normalized by the amount of DNA in the cells.

Cell lysates in RIPA buffer were diluted 5-fold with TNE buffer (10 mM Tris–HCl (pH 7.4), 1 mM EDTA, 2 M NaCl), and the DNA concentration in the diluted cell lysates was determined by adding 2 µg/ml Hoechst 33258 and measuring the fluorescence at excitation/emission wavelengths of 355/460 nm.

Western blot analysis

Preparation of whole-cell lysates, immunoblotting after electrophoresis and transfer onto PVDF membrane were performed as previously described [22].

Immunohistochemistry

Femur from male newborn mice was decalcified with 0.5 M EDTA in PBS(−) for 3 days after fixation with 3% formaldehyde in PBS(−) for 1 day. The paraffin section of the femur was stained with anti-Glut1 antibody, as described previously [20].

Reverse transcription (RT) of total RNA and real-time-polymerase chain reaction (PCR)

Total RNA was isolated from culture cells with Isogen II (Nippon gene Co. Ltd, Tokyo). Total RNA from calvaria was also extracted with Isogen II after periosteum was peeled off from calvariae of C57BL/6 mice at 0, 6, or 14 days old. The total RNAs were reverse transcribed with ReverTra Ace and oligo dT (Toyobo, Tokyo, Japan) according to the manufacture's protocol. Real-time PCR was performed using SYBR Green PCR master mix (Thermo Fisher Science). Primer sequences are either previously described [23], or listed in Table 1. The value for the specific genes was normalized based on the ubiquitin (Ubc) housekeeping control. Ubc was selected as the optimal housekeeping gene for normalization by NormFinder software (Supplementary Table S1).

Table 1.
Primer sequence for real-time PCR
Gene symbolPrimers 5′-3′GeneBankAmplified length
Slc2a1 TTCTCTGTCGGCCTCTTTGT
GAGAAGCCCATAAGCACAGC 
NM_011400 98 
Slc2a3 ACTCATTGGCATTTTCTGTGG
CTACCAGAATCCCAACAACGA 
XM_006505823 133 
Ndufaf1 GAGTAACGGTGCGGAGATGT
CGTTGGTGAATTTGTGGATG 
NM_019443 110 
Pkm1 CACAGAAGCTGCCGCCGTGG
ATTCAGCACGGCATCCTTAC 
NM_011099 217 
Pkm2 TGCTGTCTGGAGAAACAGCC
CCATGAGGTCTGTGGAGTGA 
NM_001253883 150 
hnRNPA1 GGAACTTCAGTGGTCGTGGT
ACCATCATTGCCAAATCCAT 
NM_010447 95 
hnRNPA2 GTAGCAGGAACATGGGAGGA
CTTCCACCATAGCCACCACT 
NM_182650 82 
Ldha AGGCTCCCCAGAACAAGATT
TCTCGCCCTTGAGTTTGTCT 
NM_001136069 135 
Hk1 ACGACACCCCAGAGAACATC
TCAGCGACGTGATCAAAAAG 
NM_010438 61 
Hk2 CAGGGTCTGAGCAAGGAGAC
AAGATCCAAGGCCAGGAAGT 
NM_013820 111 
Pfk1 CCATGTTGTGGGTGTCTGAG
ACAGGCTGAGTCTGGAGCAT 
NM_008826 80 
Gls CAAAGGGTGAAGTCGGTGAT
GGCTGACAGAGCAAACCTTC 
|NM_001081081 84 
c-Myc TGAAGGCTGGATTTCCTTTG
TTCTCTTCCTCGTCGCAGAT 
NM_001177354 140 
Grp78 CTGCTGAGGCGTATTTGGGA
TCG CTG GGCATCATTGAAGT 
NM_001163434 77 
Atf4 TGGCGAGTGTAAGGAGCTAGAAA
TCTTCCCCCTTGCCTTACG 
NM_009716 122 
Gapdh TCAAGAAGGTGGTGAAGCAG
GGTGGAAGAGTGGGAGTTGC 
NM_001289726 110 
Ubc CGTCGAGCCCAGTGTTACCACCAAGAAGG
CCCCCATCACACCCAAGAACAAGCACAAG 
NM_019639 112 
Cdkn1A(p21waf1TGTCTTGCACTCTGGTGTCT
ATCTGCGCTTGGAGTGATAGA 
NM_007669 117 
Gene symbolPrimers 5′-3′GeneBankAmplified length
Slc2a1 TTCTCTGTCGGCCTCTTTGT
GAGAAGCCCATAAGCACAGC 
NM_011400 98 
Slc2a3 ACTCATTGGCATTTTCTGTGG
CTACCAGAATCCCAACAACGA 
XM_006505823 133 
Ndufaf1 GAGTAACGGTGCGGAGATGT
CGTTGGTGAATTTGTGGATG 
NM_019443 110 
Pkm1 CACAGAAGCTGCCGCCGTGG
ATTCAGCACGGCATCCTTAC 
NM_011099 217 
Pkm2 TGCTGTCTGGAGAAACAGCC
CCATGAGGTCTGTGGAGTGA 
NM_001253883 150 
hnRNPA1 GGAACTTCAGTGGTCGTGGT
ACCATCATTGCCAAATCCAT 
NM_010447 95 
hnRNPA2 GTAGCAGGAACATGGGAGGA
CTTCCACCATAGCCACCACT 
NM_182650 82 
Ldha AGGCTCCCCAGAACAAGATT
TCTCGCCCTTGAGTTTGTCT 
NM_001136069 135 
Hk1 ACGACACCCCAGAGAACATC
TCAGCGACGTGATCAAAAAG 
NM_010438 61 
Hk2 CAGGGTCTGAGCAAGGAGAC
AAGATCCAAGGCCAGGAAGT 
NM_013820 111 
Pfk1 CCATGTTGTGGGTGTCTGAG
ACAGGCTGAGTCTGGAGCAT 
NM_008826 80 
Gls CAAAGGGTGAAGTCGGTGAT
GGCTGACAGAGCAAACCTTC 
|NM_001081081 84 
c-Myc TGAAGGCTGGATTTCCTTTG
TTCTCTTCCTCGTCGCAGAT 
NM_001177354 140 
Grp78 CTGCTGAGGCGTATTTGGGA
TCG CTG GGCATCATTGAAGT 
NM_001163434 77 
Atf4 TGGCGAGTGTAAGGAGCTAGAAA
TCTTCCCCCTTGCCTTACG 
NM_009716 122 
Gapdh TCAAGAAGGTGGTGAAGCAG
GGTGGAAGAGTGGGAGTTGC 
NM_001289726 110 
Ubc CGTCGAGCCCAGTGTTACCACCAAGAAGG
CCCCCATCACACCCAAGAACAAGCACAAG 
NM_019639 112 
Cdkn1A(p21waf1TGTCTTGCACTCTGGTGTCT
ATCTGCGCTTGGAGTGATAGA 
NM_007669 117 

Tetracycline-inducible (TetOn) expression system for p53 in ST2 cells

The TetOn inducible expression system for mouse p53 was established as previously described [22]. The constructed pTRE2Hyg-tp53 was stably transfected into ST2 pEF-1a-pTet-On cell line. The hygromycin B-selected clones were treated with or without 2 µg/ml doxycycline for 2 days, and protein and total RNA were extracted from cells as described above.

Slc2a1 promoter assay in osteoblast like cells

Slc2a1 promoter constructs from −395, −262, −109, −56, or −5 to +125 were amplified from mouse C57BL/6 genomic DNA by PCR with primers (5′-CCAGCCTCTCTGGTCTTCAC-3′, 5′-TAGCGCTGGCTTACAGC-3′, 5′-AGAGCCAGACTGTGGTC-3′, 5′-CC TCT CAG AGT CCC GCC-3′, 5′-CC ACCTACACCCCAGAA C-3′, 5′-GACGCACTTAAGACCCCGTA-3′, respectively). The PCR products were digested with Kpn I and Xho I, and then ligated into pGL3-basic (Promega) predigested with the same enzymes.

Using KOD-Plus mutagenesis kit (Toyobo, Osaka, Japan), the tandem repeat-mutated constructs of the Slc2a1 promoter (pGL3-Slc2a1 −395/+125 del and pGL3-Slc2a1 −395/+125 AA mut) were generated based on the wild-type Slc2a1 promoter construct (pGL3-Slc2a1 −395/+125) according to the manufacturer's instructions. Briefly, inverse PCR was performed using tandem repeat deletion primers (5′-GACGGAGGGCTACTGACCACAGTCTG-3′ and 5′-AGAACCACGCCTCTCAGAGTCCCGCC-3′) or tandem repeat mutation primers (5′-GACGGAGGGCTACTGACCACAGTCTG-3′ and 5′-AATCCCCACCTCCCCACAGAACCACGCC-3′) with KOD-plus DNA polymerase. Following the PCR, products were digested by Dpn I, followed by self-ligation using T4 ligase.

ST2 cells were cultured in αMEM plus 10% FCS on 12 well plate, and then transfected with 1 µg of each of Slc2a1 promoter reporter plasmids described above using HilymaxTM transfection reagent (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). The cells were co-transfected with pcDNA3.1 (+) to be selected as neomycin resistant cells. After cloning using 0.2 mg/ml G418, the cells were cultured in DMEM + 10% FCS in the presence or absence of AA2P and β-glycerophosphate for 21 days with the medium change every 3 days. The whole cell extracts were prepared in Reporter Lysates bufferTM (Promega) and the luciferase activity was measured by One-Glo assay systemTM (Promega) following the instruction manual. The measured values were normalized by protein concentrations.

ST2 cells were cultured in αMEM plus 10% FCS on 12 well plate, and then transfected with each of Glut1 promoter reporter vector described above, pSV-β-galactosidase (Promega) and pcDNA3.1 (−) or pcDNA-p53 using HilymaxTM transfection reagent for 12 h. After change a medium to αMEM plus 10% FCS, the cells were incubated for 2 days. The luciferase activity in whole cell extracts were measured as described above. The values were normalized by β-galactosidase activities, which were measured with 4-MUG as the substrate, as described previously [24].

Measurement of cellular ATP

Cells in pre-OB fr or FB fr were cultured in α-MEM plus 10% FCS in the presence or absence of AA2P and β-glycerophosphate for several days. The cells were stimulated with 5 mg/ml 2-DG, 5 µM CCCP, or 0.5 mM phloretin for 2 h, and then treated with 500 µl of 1% trichloroacetic acid to extract ATP for 30 min, and neutralized by addition of 60 µl of 1 M Tris–HCl (pH 9.0). Contents of ATP in the extracts were measured with ATP luciferase assay kit (Promega). The ATP contents were normalized by protein concentration.

Measurement of pH in conditioned media

Confluent primary undifferentiated osteoblasts were cultured in αMEM plus 10% FCS in the presence or absence of AA2P and β-glycerophosphate under high glucose or normal glucose condition for 9 days. Conditioned media were collected, and the pH was measured by pH meter F-52 (Horiba Ltd, Kyoto, Japan).

Statistical analysis

The results obtained for each group were expressed as the mean ± standard deviation (n = 3). Student's t-test or A one-way analysis of variance (ANOVA), followed by Fisher's least significant difference test, was used to evaluate the significant differences among the treatments. Statistical significance was indicated as *P < 0.05.

Results

Increased expressions of glycolytic metabolism associated genes in primary osteoblasts during differentiation

To isolate fibroblast-rich and osteoblast-rich cell fractions, neonatal mouse calvaria were sequentially digested with collagenase/trypsin, and the obtained cells were cultured for the mRNA expression analysis of osteoblast differentiation markers such as Spp1 and Osteocalcin. Since the cells in the 1st digestion fraction had low expressions of the differentiation markers, they were used as FB fr, and the second and third digested fractions containing cells with high expressions of the differentiation markers were used as pre-OB fr (Figure 1A). When we examined if high glucose condition affected osteoblast differentiation, we found that phenol red in conditioned media turned yellow, which indicated acidic state, during primary cultured osteoblast differentiation. Therefore, we compared the pH of the conditioned medium of differentiated osteoblasts from pre-OB fr with that of undifferentiated osteoblasts. It was found that osteoblast differentiation induced pH drop of the conditioned medium not only under high glucose but also normal glucose condition, whereas neither high glucose nor differentiation condition increased the amount of DNA per one well (Figure 1B). Lactate is known as a major acidic product by glycolysis within solid tumors [25], which prompted us to examine the possibility that differentiation of primary osteoblasts increased lactic acid fermentation in aerobic glycolytic metabolism. Therefore, we measured lactate production and glucose uptake by cells in pre-OB fr during osteoblast differentiation. Lactate content in conditioned media and glucose uptake were found to increase with osteoblast differentiation, and lactate production and glucose uptake were suppressed by a glucose uptake inhibitor, 2-DG (Figure 1C).

Glut1 expression in primary osteoblasts during differentiation.

Figure 1.
Glut1 expression in primary osteoblasts during differentiation.

(A) Neonatal mouse calvariae were sequentially digested with collagenase/trypsin. The obtained cell fractions after 1st, 2nd, 3rd, and 4th digestion were cultured for 7 days. (B) The cells in pre-OB fr were cultured in αMEM plus 10% FCS with 100 mg/dl (N gluc) or 600 mg/dl (H gluc) glucose in the presence or absence of AA2P and β-glycerophosphate for 14 days. The pH of conditioned media was measured by a pH meter. Each bar represents an individual pH value of the culture medium (upper panel). (n = 3). *P < 0.05. DNA contents in cell lysates were determined using Hoechst 33258 (lower panel). (C) Cells in the pre-OB fr were cultured with 100 mg/dl glucose in the presence of AA2P and β-glycerophosphate for 0, 3, 9, or 14 days, and incubated for 6 h with or without 5 mM 2-DG. Lactate content in the conditioned medium of the cells was measured (upper panel). After incorporation of 2-NBDG into the cells, the fluorescence of the cell lysate was measured as glucose uptake (lower panel). Both lactate and glucose uptake values were normalized by the DNA content of the cells. (DF) Cells in the pre-OB fr (pre-OB) or the FB fr (FB) were cultured in media containing AA2P and β-glycerophosphate for 0, 3, 9, 14 days (D), or 0, 7, 14 days (E,F). Total RNA isolated from the cells or the calvariae to detect expressions of Spp1, Ocn, Slc2a1, Slc2a3, or mitochondrial Ndufa1 mRNA. The mRNA levels were normalized by Ubc mRNA levels. *P < 0.05, versus time 0 (A,D,E). Glut1, GAPDH, and β-actin expressions in the cell lysates were analyzed by western blot (F).

Figure 1.
Glut1 expression in primary osteoblasts during differentiation.

(A) Neonatal mouse calvariae were sequentially digested with collagenase/trypsin. The obtained cell fractions after 1st, 2nd, 3rd, and 4th digestion were cultured for 7 days. (B) The cells in pre-OB fr were cultured in αMEM plus 10% FCS with 100 mg/dl (N gluc) or 600 mg/dl (H gluc) glucose in the presence or absence of AA2P and β-glycerophosphate for 14 days. The pH of conditioned media was measured by a pH meter. Each bar represents an individual pH value of the culture medium (upper panel). (n = 3). *P < 0.05. DNA contents in cell lysates were determined using Hoechst 33258 (lower panel). (C) Cells in the pre-OB fr were cultured with 100 mg/dl glucose in the presence of AA2P and β-glycerophosphate for 0, 3, 9, or 14 days, and incubated for 6 h with or without 5 mM 2-DG. Lactate content in the conditioned medium of the cells was measured (upper panel). After incorporation of 2-NBDG into the cells, the fluorescence of the cell lysate was measured as glucose uptake (lower panel). Both lactate and glucose uptake values were normalized by the DNA content of the cells. (DF) Cells in the pre-OB fr (pre-OB) or the FB fr (FB) were cultured in media containing AA2P and β-glycerophosphate for 0, 3, 9, 14 days (D), or 0, 7, 14 days (E,F). Total RNA isolated from the cells or the calvariae to detect expressions of Spp1, Ocn, Slc2a1, Slc2a3, or mitochondrial Ndufa1 mRNA. The mRNA levels were normalized by Ubc mRNA levels. *P < 0.05, versus time 0 (A,D,E). Glut1, GAPDH, and β-actin expressions in the cell lysates were analyzed by western blot (F).

Next, we determined the expressions of glucose transporters, Glut1 and Glut3, which are dominant in osteoblasts [9]. Cells in pre-OB fr showed increased Slc2a1 mRNA coding Glut1 protein during osteoblast differentiation, whereas mRNAs for Slc2a3 mRNA coding Glut3 and a mitochondrial gene, NADH dehydrogenase [ubiquinone] 1 α subcomplex assembly factor 1 (Ndufaf1), did not changed (Figure 1D). Unlike cells in the FB fr, the treatment of cells in pre-OB fr with AA2P and β-glycerophosphate increased expressions of Slc2a1 mRNA and bone matrix protein, and Glut1 protein production (Figure 1E,F). We also found that osteoblast differentiation slightly increased a glycolytic enzyme, GAPDH (Figure 1F).

We analyzed Glut1 expression in osteoblasts in vivo. After removing periosteum from calvaria of mice at 0, 3, 6, or 14 days of age, we examined expressions of Spp1 encoding OPN and Ocn mRNAs (osteoblast differentiation markers), as well as Slc2a1 mRNA in calvaria extracts. Consequently, we found that the development of mice increased Slc2a1 mRNA in calvaria accompanied by the induction of osteoblast marker expressions (Figure 2A–C). Besides, immunohistochemical staining of mouse endochondral ossification showed that Glut1 was detected in bone lining cells and hypertrophic chondrocytes (Figure 2D). These observations suggest that Glut1 expression is increased during bone development.

Glut1 expression in osteoblasts from mouse calvariae.

Figure 2.
Glut1 expression in osteoblasts from mouse calvariae.

(AC) Periosteum-removed calvariae were obtained from C57BL/6 mice at 0, 3, 6, or 14 days. Total RNA isolated from the calvariae to detect expressions of Spp1, Ocn, or Slc2a1 mRNA (n = 4). The mRNA levels were normalized by Ubc mRNA levels. *P < 0.05 versus 0 day. (D) Femur from mice at 3 days old were fixed with 3% formaldehyde and decalcified with 0.5 M EDTA for 2 days. Paraffin sections were stained with anti-Glut1 antibody. (a) Low magnification (×100), (bd) high magnification (×400). (b) resting zone, (c) hypertrophic zone, (d) ossification zone. Arrows and arrowheads indicate osteoblasts and hypertrophic chondrocytes, respectively.

Figure 2.
Glut1 expression in osteoblasts from mouse calvariae.

(AC) Periosteum-removed calvariae were obtained from C57BL/6 mice at 0, 3, 6, or 14 days. Total RNA isolated from the calvariae to detect expressions of Spp1, Ocn, or Slc2a1 mRNA (n = 4). The mRNA levels were normalized by Ubc mRNA levels. *P < 0.05 versus 0 day. (D) Femur from mice at 3 days old were fixed with 3% formaldehyde and decalcified with 0.5 M EDTA for 2 days. Paraffin sections were stained with anti-Glut1 antibody. (a) Low magnification (×100), (bd) high magnification (×400). (b) resting zone, (c) hypertrophic zone, (d) ossification zone. Arrows and arrowheads indicate osteoblasts and hypertrophic chondrocytes, respectively.

Increased gene expression of glycolytic enzymes and Slc2a1 is observed in tumor cells as Warburg effect [26]. We examined if gene expressions of glycolytic enzymes increase in primary osteoblasts during differentiation. Real-time RT-PCR showed that osteoblast differentiation increased mRNA expressions of hexokinases 1/2 (Hk1/2) and lactate dehydrogenase A (Ldha), which catalyze the initial and last steps of glycolysis, respectively (Supplementary Figure S1A–C). In addition, mRNA expression of phosphofructokinase (Pfk), which is a key enzyme for the control of glycolysis, also increased (Supplementary Figure S1D). The GAPDH mRNA level in 11-day differentiated osteoblasts was slightly higher than that of undifferentiated pre-osteoblast cells, that was consistent with the protein expression level (Supplementary Figure S1E). Concerning pyruvate kinase, expression of isotype M2 (Pkm2) increased during osteoblast differentiation, whereas that of isotype M1 (Pkm1) did not (Supplementary Figure S1F,G). Alternative splicing of Pk gene transcripts, which produces Pkm1 and Pkm2 mRNAs, is associated with tumor metabolic phenotype [27,28]. The alternative splicing is regulated by heterogeneous nuclear ribonucleoproteins (hnRNPs) such as hnRNPA1, A2, and polypyrimidine tract-binding protein (PTB) in cancer cells [28]. We found that osteoblast differentiation significantly increased expression of hnRNPA1 mRNA, but not hnRNPA2 mRNA, in cells of pre-OB fr (Supplementary Figure S1H,I). On the other hand, the expression of Ptb mRNA was undetectable by quantitative RT-PCR (data not shown). These observations of glycolysis associated enzyme and hnRNPA1 expressions are very similar to the metabolic features of tumor cells [29].

Glut1, Hk2, Ldha, and Pkm2 expressions are known to be positively regulated by c-Myc oncogenic transcription factor in tumor cells [11]. However, c-Myc expression did not increase during osteoblast differentiation, (Supplementary Figure S1J). Also, osteoblast differentiation did not increase the expression of Gls1, a transporter of glutamate into mitochondria, which is also known to be increased by c-Myc (Supplementary Figure S1K) [11]. These results suggested that c-Myc did not affect glycolysis associated enzyme expression in osteoblasts. Therefore, glycolytic phenotype of differentiated osteoblasts is not entirely comparable with Warburg effect in tumor cells.

Selection of aerobic glycolysis to produce ATP by differentiated osteoblasts

To confirm metabolic switch for ATP production in primary osteoblast differentiation, we examined intracellular ATP contents after treatment of differentiated or undifferentiated cells in pre-OB fr with 2-DG (a glucose analog), phloretin (a glucose uptake inhibitor), CCCP (an uncoupler of oxidative phosphorylation), or oligomycin A (an ATP synthase inhibitor). In the result, 2-DG and phloretin significantly reduced ATP concentration in cells of pre-OB under osteoblast differentiation condition, but not in cells under undifferentiation condition as well as FB fr cells (Figure 3A,B,D). The reduction in ATP synthesis with phloretin in differentiated cells of pre-OB fr was abolished by a high glucose condition (Figure 3C). On the other hand, CCCP and oligomycin A did not reduce ATP concentration in differentiated cells of pre-OB fr. In contrast, CCCP and oligomycin A inhibited ATP production in cells of undifferentiated pre-OB fr and FB fr (Figure 3B,D). These observations suggested that differentiated osteoblasts selected aerobic glycolytic pathway to produce ATP rather than oxidative phosphorylation in mitochondria.

Effects of metabolic inhibitors on ATP contents and an osteogenic marker in primary osteoblasts.

Figure 3.
Effects of metabolic inhibitors on ATP contents and an osteogenic marker in primary osteoblasts.

(AC) Cells in the pre-OB fr were cultured in the presence or absence of AA2P and β-glycerophosphate for 7 days. After treatment with 5 mM 2-DG, 5 µM oligomycin A, 5 µM CCCP, or 0.5 mM phloretin (Phl) under normal glucose or high glucose (Phl + Gluc) condition for 2 h, the ATP contents in the cells were measured by luciferase assay. *P < 0.05 (B,D) Cells in FB fr were cultured in the presence of AA2P and β-glycerophosphate for 7 days. After treatment with 5 mM 2-DG, 5 µM CCCP, 5 µM oligomycin A, or 0.5 mM phloretin (phl) for 2 h, ATP contents in the cells were measured by luciferase assay. *P < 0.05 versus control. (E) Total RNA was isolated from the pre-OB cells to detect the expression of Spp1 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05.

Figure 3.
Effects of metabolic inhibitors on ATP contents and an osteogenic marker in primary osteoblasts.

(AC) Cells in the pre-OB fr were cultured in the presence or absence of AA2P and β-glycerophosphate for 7 days. After treatment with 5 mM 2-DG, 5 µM oligomycin A, 5 µM CCCP, or 0.5 mM phloretin (Phl) under normal glucose or high glucose (Phl + Gluc) condition for 2 h, the ATP contents in the cells were measured by luciferase assay. *P < 0.05 (B,D) Cells in FB fr were cultured in the presence of AA2P and β-glycerophosphate for 7 days. After treatment with 5 mM 2-DG, 5 µM CCCP, 5 µM oligomycin A, or 0.5 mM phloretin (phl) for 2 h, ATP contents in the cells were measured by luciferase assay. *P < 0.05 versus control. (E) Total RNA was isolated from the pre-OB cells to detect the expression of Spp1 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05.

Additionally, we examined whether inhibition of glycolysis affects osteoblast differentiation, and found that 2-DG partially inhibited mRNA expression of an osteogenic marker, Spp1. In contrast, an uncoupler of mitochondrial oxidative phosphorylation, CCCP, did not affect its expression. (Figure 3E).

Negative regulation of Glut1 expression by p53 in osteoblasts

Previously, p53 expression has been shown to be decreased when Runx2 expression is slightly enhanced during osteoblast differentiation [30]. We also observed that p53 protein appeared in cells in pre-OB fr under differentiation condition within 3 days after confluence, whereas Glut1 protein expressed after 10 days of differentiation (Figures 1E and 4A). However, the expression of p53 was low before confluence, and the expression of p53 was slightly increased at 25 days of osteoblast differentiation (Figure 4B). Osteoblast differentiation for 25 days was observed to partially reduce Glut1 and Ldha expressions (Supplementary Figure S2). Furthermore, p21waf1, transcription of which is activated by p53, was also expressed within 5 days, but not p27kip1 (Figure 4A). Hence, we stimulated cells in pre-OB fr with nutlin 3, an activator of p53. Nutlin 3 increased p53 expression in 8-day differentiated osteoblasts, and inhibited Slc2a1 expression in more than 6-day differentiation osteoblasts, whereas it did not affect Spp1 expressions (Figure 4C,D). Furthermore, we found that nutrin 3 inhibited 2-NBDG incorporation, which represents glucose uptake, into differentiated osteoblasts (Figure 4E). We then prepared a tetOn inducible p53 expression system in ST2 cells, which are able to differentiate into osteoblasts by treatment with AA2P and β-glycerophosphate [21]. We found that the forced expression of p53 reduced Glut1 expression under a differentiation condition (Figure 4F). Additionally, since gene ablation of p53 was reported to increase osterix expression but not Runx2 [31], we investigated whether the inhibition of p53 transcriptional activity could affect the expression of osteoblast differentiation markers. PFTα, an inhibitor of p53 transcriptional activity, reduced mRNA expression of cyclin kinase inhibitor, p21waf1, and enhanced Ocn mRNA expression but not Spp1 mRNA expression in pre-OB fr of 0- or 3-day differentiated cells with p53 expression (Figure 4G), suggesting that reduced p53 levels partially promotes osteoblast differentiation.

Regulation of Slc2a1 gene expression in mouse osteoblasts by p53.

Figure 4.
Regulation of Slc2a1 gene expression in mouse osteoblasts by p53.

(A) After confluent, the cells in the osteoblast fr were cultured with AA2P and β-glycerophosphate, and the cell lysates were collected at 0, 5, 10, or 14 days. (B) Cells in the pre-OB fr were cultured without AA2P and β-glycerophosphate until confluency. Cell lysates were prepared at 1 or 3 days prior to confluency (PC). After confluency, the cells were cultured with AA2P and β-glycerophosphate for 0, 4, 8, 14, or 25 days. Prepared cell lysates were subjected to western blot analysis to detect expressions of p53, Glut1, p21, p27, and β-actin (A,B,F). (C,D) Cells in pre-OB fr were cultured with or without AA2P and β-glycerophosphate for 0, 3, 6, or 8 days, and then were treated with or without 10 µM nutlin-3 (N3) for 8 h. Total RNA was isolated from the cell to detect expressions of Spp1 or Slc2a1 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05 versus control. (E) Cells in the pre-OB fr were cultured with or without AA2P and β-glycerophosphate for 7 days, and then treated with or without 10 µM nutlin-3 (N3) for 8 h. After incorporation of 2-NBDG into the cells, the fluorescence of the cell lysate was measured as glucose uptake. Fluorescence intensity was normalized by intracellular DNA content. (F) ST2 cells were stably transfected with pTRE2Hyg-tp53, and treated with doxycycline. Forty-eight hours later, p53 and Glut1 expressions were analyzed by western blot. (G) After becoming confluent, cells in pre-OB fr were culture for 0 or 3 days with AA2P and β-glycerophosphate, and then treated with 1.2 µM PFTα for 6 h. Total RNA was isolated from the cell to detect expressions of Spp1, Slc2a1, Ocn, or p21 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05 versus control (DG). (H) Cells in the pre-OB fr were cultured in the presence of AA2P and β-glycerophosphate for 0, 7, or 14 days. Following the induction of differentiation for 0 or 7 days, the cells were incubated under a hypoxic condition for 6 or 24 h. Glut1, HIF-1α, p53, and β-actin expressions in the cell lysates were analyzed by western blot.

Figure 4.
Regulation of Slc2a1 gene expression in mouse osteoblasts by p53.

(A) After confluent, the cells in the osteoblast fr were cultured with AA2P and β-glycerophosphate, and the cell lysates were collected at 0, 5, 10, or 14 days. (B) Cells in the pre-OB fr were cultured without AA2P and β-glycerophosphate until confluency. Cell lysates were prepared at 1 or 3 days prior to confluency (PC). After confluency, the cells were cultured with AA2P and β-glycerophosphate for 0, 4, 8, 14, or 25 days. Prepared cell lysates were subjected to western blot analysis to detect expressions of p53, Glut1, p21, p27, and β-actin (A,B,F). (C,D) Cells in pre-OB fr were cultured with or without AA2P and β-glycerophosphate for 0, 3, 6, or 8 days, and then were treated with or without 10 µM nutlin-3 (N3) for 8 h. Total RNA was isolated from the cell to detect expressions of Spp1 or Slc2a1 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05 versus control. (E) Cells in the pre-OB fr were cultured with or without AA2P and β-glycerophosphate for 7 days, and then treated with or without 10 µM nutlin-3 (N3) for 8 h. After incorporation of 2-NBDG into the cells, the fluorescence of the cell lysate was measured as glucose uptake. Fluorescence intensity was normalized by intracellular DNA content. (F) ST2 cells were stably transfected with pTRE2Hyg-tp53, and treated with doxycycline. Forty-eight hours later, p53 and Glut1 expressions were analyzed by western blot. (G) After becoming confluent, cells in pre-OB fr were culture for 0 or 3 days with AA2P and β-glycerophosphate, and then treated with 1.2 µM PFTα for 6 h. Total RNA was isolated from the cell to detect expressions of Spp1, Slc2a1, Ocn, or p21 mRNA. The mRNA levels were normalized by Ubc mRNA level. *P < 0.05 versus control (DG). (H) Cells in the pre-OB fr were cultured in the presence of AA2P and β-glycerophosphate for 0, 7, or 14 days. Following the induction of differentiation for 0 or 7 days, the cells were incubated under a hypoxic condition for 6 or 24 h. Glut1, HIF-1α, p53, and β-actin expressions in the cell lysates were analyzed by western blot.

Hypoxia induces Glut1 expression in osteoblasts to convert metabolism for ATP synthesis to glycolysis through HIF-1α [32]. We found that hypoxia increased expression of HIF-1α by undifferentiated cells in pre-OB fr (Figure 1H). However, no enhancement of HIF-1α expression was observed in cells of pre-OB fr induced to differentiate with an ascorbic acid derivative, regardless of the oxygen condition (Figure 1H), which may be consistent with a previous report that ascorbic acid increases prolyl hydroxylation of HIF-1α to be polyubiquitinated by pVHL and degraded by the proteasome [33]. It is also noteworthy that hypoxia time dependently increased the expression of p53, which attenuated Glut1 expression in this study (Figure 1H).

Glut1 expression in osteoblasts has been reported to be up-regulated by Wnt ligands such as Wnt3a [34]. To examine if p53 was involved in Wnt-induced Glut1 expression, Glut1 expression was induced by Wint3A in the presence or absence of Nutlin 3 (Supplementary Figure S3). Induction of Glut1 expression by Wnt3a was observed in p53 expressing undifferentiated cells in pre-OB fr (Supplementary Figure S3). Nutlin 3 treatment to increase more expression of p53 did not alter Wnt3a-induced Gut1 expression in the cells (Supplementary Figure S3). These results suggested that p53 is not involved in Wnt3a-induced Glut1 expression.

Next, to examine the transcriptional activity of Slc2a1 gene during osteoblast differentiation, we generated luciferase reporter constructs containing five serial deletion constructs of Slc2a1 gene promoter with the fixed 3′ end at the +125 position (relative to transcription start site) (Figure 5A), and isolated stably transfected ST2 cells. Searching for putative transcription factor binding sites in Slc2a1 promoter sequence by MatchTM predicts binding sites of glycolytic genes transcriptional activator 1 (GCR1) and nuclear factor-1 (NF-1) [35] (Figure 5A). NF1 recognition element has been reported to function as a transcriptional repressor [36,37]. Osteoblast differentiation increased luciferase activities of −395 and −262 constructs (Figure 5C). Increased luciferase activity was also observed for −109 construct, which contains neither GCR1 nor NF-1 binding site (Figure 5A). These observations suggested that GCR1 and NF-1 binding sites in Slc2a1 promoter region are not associated with the increased promoter activity during osteoblast differentiation. In contrast, osteoblast differentiation failed to increase the promoter activity of −54 or −5 construct (Figure 5C). In the promoter region between −108 and −54, we found a tandem repeat, CCTCCCCACCTCCCC at −83 bp (Figure 5A). Although neither substitution nor deletion mutation of the tandem repeat negated the inhibition of luciferase activity by p53 overexpression, both of these mutations reduced the luciferase activity (Figure 5B,D,E). These observations suggested that p53 represses Slc2a1 gene transcription through its promoter site (−395/ + 125), but the precise location of the promoter element is not clear.

Regulation of Slc2a1 gene promoter activity by p53.

Figure 5.
Regulation of Slc2a1 gene promoter activity by p53.

(A,B) The schematic structure of luciferase reporter vectors containing 5′-terminal serial deletion regions of Slc2a1 gene (A) and tandem repeat deletion or substitution from CC to AA in −395/+125 site (B). The transcription initiation site of Slc2a1 promoter is denoted as +1. (CF) ST2 cells were stably transfected with each luciferase reporter construct, containing the −395 upstream of Slc2a1 gene and its deletion derivatives (−262, −109, −84, and −5) (C,D,F) or tandem repeat deletion or substitution in −395/+125 (E), and then were cloned. Cells were differentiated with AA2P and β-glycerophosphate for 3 weeks. Luciferase activities in the cell lysates were measured by One-Glo luciferase assay systemTM (DF). Luciferase reporter construct-transfected ST2 cells were transiently transfected with pcDNA3.1 or pcDNA3.1-p53 along with pSV-βgal. And then the cells were treated with or without PFTα (F). The values were normalized by the protein concentration in the cell lysates. N = 3, P < 0.05 versus control (C). Luciferase and β-galactosidase activities in the cell lysates were detected with beetle luciferin in One-Glo luciferase assay systemTM and 4-MUG, respectively. The value of luciferase activity was normalized by β-galactosidase activity. N = 3, P < 0.05 versus pcDNA3.1 (DF).

Figure 5.
Regulation of Slc2a1 gene promoter activity by p53.

(A,B) The schematic structure of luciferase reporter vectors containing 5′-terminal serial deletion regions of Slc2a1 gene (A) and tandem repeat deletion or substitution from CC to AA in −395/+125 site (B). The transcription initiation site of Slc2a1 promoter is denoted as +1. (CF) ST2 cells were stably transfected with each luciferase reporter construct, containing the −395 upstream of Slc2a1 gene and its deletion derivatives (−262, −109, −84, and −5) (C,D,F) or tandem repeat deletion or substitution in −395/+125 (E), and then were cloned. Cells were differentiated with AA2P and β-glycerophosphate for 3 weeks. Luciferase activities in the cell lysates were measured by One-Glo luciferase assay systemTM (DF). Luciferase reporter construct-transfected ST2 cells were transiently transfected with pcDNA3.1 or pcDNA3.1-p53 along with pSV-βgal. And then the cells were treated with or without PFTα (F). The values were normalized by the protein concentration in the cell lysates. N = 3, P < 0.05 versus control (C). Luciferase and β-galactosidase activities in the cell lysates were detected with beetle luciferin in One-Glo luciferase assay systemTM and 4-MUG, respectively. The value of luciferase activity was normalized by β-galactosidase activity. N = 3, P < 0.05 versus pcDNA3.1 (DF).

To confirm that the transcriptional activity of p53 was involved in the suppression of Slc2a1 promoter activity, p53 expressing cells were transfected with the Slc2a1 promoter reporter plasmid and treated with an inhibitor of p53 DNA binding, PFTα. PFTα partially attenuated the inhibition of Slc2a1 promoter activity due to p53 forced expression in cells with a reporter plasmid containing the Slc2a1 promoter region (−351/+126) (Figure 5F). Since there is no p53 binding sequence in the Slc2a1 promoter region, p53 was thought to be indirectly involved in inhibiting Slc2a1 transcriptional activity.

Association of PI3K/AKT pathway with Glut1 expression in primary osteoblasts

Glut1 expression and its glucose transport are regulated by various factors, including IGF-1 through PI3K/AKT pathway [38,39]. On the other hand, AKT is known to regulate p53 protein expression by controlling MDM2, which ubiquitinates p53 for degradation in tumor cells [40]. Hence, we examined the possibility that PI3K/AKT pathway was associated with Glut1 expression in osteoblasts. Consequently, a PI3K inhibitor, LY29006, decreased phosphorylation of AKT through osteoblast differentiation, and inhibited Glut1 protein and mRNA levels, as well as LDH mRNA expression at day 8 (Figure 6A,B). In contrast, the PI3K inhibitor increased p53 protein expression at day 8 of osteoblast differentiation (Figure 6A). However, the PI3K inhibitor did not affect OPN protein and expressions of Ocn and Spp1 mRNA (Figure 6A,B). These results and previous reports suggested PI3K/AKT pathway is associated with Glut1 expression during osteoblast differentiation.

Effects of a PI3K inhibitor on Glut1 expression in primary osteoblasts.

Figure 6.
Effects of a PI3K inhibitor on Glut1 expression in primary osteoblasts.

Cells in the pre-OB fr were cultured in the presence or absence of AA2P and β-glycerophosphate for 0, 4, or 8 days, and then were treated with or without 5 µM LY294002 (LY). (A) PhosphoAKT (pAKT), Glut1, p53, osteopontin (OPN), and β-actin expressions in the cell lysates were analyzed by western blot. (B) Total RNA isolated from the cells or the calvariae to detect expressions of Spp1, Ocn, Slc2a1, or Ldha mRNA. The mRNA levels were normalized by Ubc mRNA levels.

Figure 6.
Effects of a PI3K inhibitor on Glut1 expression in primary osteoblasts.

Cells in the pre-OB fr were cultured in the presence or absence of AA2P and β-glycerophosphate for 0, 4, or 8 days, and then were treated with or without 5 µM LY294002 (LY). (A) PhosphoAKT (pAKT), Glut1, p53, osteopontin (OPN), and β-actin expressions in the cell lysates were analyzed by western blot. (B) Total RNA isolated from the cells or the calvariae to detect expressions of Spp1, Ocn, Slc2a1, or Ldha mRNA. The mRNA levels were normalized by Ubc mRNA levels.

Discussion

Recently, metabolic switch to glycolysis for ATP production has been known to occur in differentiated cells such as CD8+ T cells, which secrete cytotoxic proteins such as perforin, TNF-α, and granzyme B as immune responses to infection. However, some effector CD8+ T cells later become resting memory T cells, which repress granzyme B expression and possess extramitochondrial respiratory capacity [5,41]. In this study, the change of metabolism to glycolysis was also observed in primary osteoblasts, which secrete large amounts of bone matrix proteins including collagen type I, osteopontin, and osteocalcin. The active osteoblasts become quiescent osteocytes encoated into bone matrix to play the role of mechanosensor. A recent report has addressed that osteocytes select oxidative phosphorylation for ATP production rather than glycolysis [42]. Consistently, the late differentiation (25 days) of primary osteoblasts slightly restored expression of p53 with the reduction in Slc2a1 mRNA expression (Figure 4B, Supplementary Figure S2). In summary, these findings suggest that osteoblast/osteocyte differentiation and CD8+ T cell activation process possess similar metabolic changes from glycolysis to oxidative phosphorylation during induction to quiescence.

PI3K/ATK pathway is required for BMP-derived osteoblast differentiation. LY294002, a PI3K inhibitor, has been reported to suppress expressions of BMP-2-induced osteoblastic markers such as ALP and osteocalcin in osteoblastic cell lines [43,44]. However, we found that LY294002 had little effect on ascorbic acid and β-glycerophosphate induced expressions of Spp1 and Ocn mRNA in primary osteoblasts. These observations suggested that ascorbic acid and β-glycerophosphate derived osteoblast differentiation mediated different signaling pathways from BMP-derived osteoblastic differentiation. Furthermore, although BMP-9 induces mRNA expressions of HIF-1α, which induces PDK expression to inhibit pyruvate dehydrogenase activity and TCA cycle, ascorbic acid-induced osteoblast differentiation did not increase HIF-1α expression (Figure 4H) [19]. These phenomena also suggested that different signals existed between BMP and ascorbic acid-derived osteoblast differentiation.

Glucose is transported into osteoblasts via Glut protein family. Glut1 is a class I glucose transporter that possesses a high affinity for glucose and the highest expressed transporter among the Glut family in osteoblasts [19]. Also, we detected Glut1 expression in both osteoblasts and hypertrophic chondrocytes of growth plate from newborn mice. Glut1 has been reported to be expressed in hypertrophic chondrocytes of Meckel's cartilage and epiphysis from fetal mice at embryonic day 16, but not in proliferating chondrocytes or osteoblasts [45]. In the present study, we observed, however, Glut1 expression in both osteoblasts and hypertrophic chondrocytes from newborn mice. The difference of mouse age might cause this discrepancy.

Our results have shown that Glut1 gene expression increases during primary osteoblast differentiation, similar to osteopontin expression. It has previously been reported that Glut1 expression is regulated by several transcription factors, including p53, HIF-1α, and c-Myc, which are associated with carcinogenesis [2]. In this study, we focused on p53 as a negative regulator of Glut1 expression because the expression of p53 decreased during osteoblast differentiation (Figure 4), while c-myc mRNA and HIF-1α protein expressions did not change (Figure 4H, Supplementary Figure S1J). It is well known that p53 regulates the expression of glycolytic enzymes in cancer cells as a Warburg effect. p53 as a transcription factor negatively regulates Glut3 expression through IKK/NF-κB pathway, and increase TP53-inducible glycolysis and apoptosis regulator (TIGAR) expression to lower fructose-2,6-bisphosphate levels as fructose-2,6-bisphosphatase in cells, resulting in an inhibition of glycolysis [26]. Furthermore, p53 suppresses the expression of Glut1, Glut4, and HKII, and the expression of Slc2A1 mRNA has been reported to be down-regulated by p53 transcription factor via a decrease of Glut1 promoter activity [46]. Although sequence analysis revealed that the 5′-flanking region of Slc2a1 gene does not contain p53 transcription factor binding site, p53 expression was negatively correlated with Slc2a1 expression [35]. To investigate if p53 signal indirectly affects transcriptional regulation of Scl2a1, we co-transfected p53 expression plasmid with each of the serial deletion constructs of Slc2a1 promoter reporter vectors. Co-transfection of p53 expression plasmid reduced luciferase activities in deletion constructs of −395/+125, −262/+125, and −109/+125, although did not reduce those of deletion constructs −54/+125 and −5/+125, which did not contain the tandem repeat sequence (Figure 5). These results were consistent with the promoter assay data for osteoblast differentiation reducing Slc2a1 expression, in which −54/+125 construct did not show decreased transcriptional activity in osteogenic differentiation (Figure 5), suggesting that p53 reduced the transcription of Slc2a1 gene, which was not associated with GCR1 or NF-1 binding site. Furthermore, the promoter assay of Slc2a1 suggested that a tandem repeat (CCTCCCC-A-CCTCCCC) containing domain is essential for decrease the promoter activity. Although a tandem repeat-like sequence (CCTCACC-C-CCTCCCC) exists in human SLC2A1 transcription site at 466 bp downstream of the transcription start site, both p53 induced and repressed binding sites are different from the tandem repeat. Mutations of the tandem repeat reduced transcriptional activity of SLC2A1, while p53 did not affect reduction in the transcriptional activity (Figure 5). However, the results that inhibition of p53 DNA binding induces suppression of Slc2a1 promoter activity (Figure 5F) suggested that p53 is indirectly involved in Slc2a1 promoter activity.

The other transcription factors for induction of Glut1 expression are known to be HIF-1α, c-Myc, and Runx2. First, Runx2 is a master transcription factor for osteoblast differentiation. Runx2 binds to its consensus sequence ‘ACCACA’ in the promoter region of Spp1 gene to increase its transcriptional activity [47]. Slc2a1 promoter region also contains ‘ACCACA’ (−811/−796) and Runx2 promotes Slc2a1 transcriptional activity [19]. In our Slc2a1 gene promoter assay, we used Slc2a1 gene deletion constructs of −395/+125, −262/+125, and −109/+125, which did not contain the Runx2 recognition site, but increased promoter activity of these constructs was observed during osteoblast differentiation. Next, HIF-1α promotes angiogenesis and osteogenesis during bone development to induce osterix, Runx2, and type I collagen in osteoblasts [32]. Recently, pyruvate dehydrogenase kinase 1 (PDK1) expression induction by HIF-1α is reported to contribute to bone formation. [48] Noteworthy observation in the report is that OCN-positive cells in chondro-osteo junction area of primary ossification exist under a hypoxic condition [48]. The observation suggested that HIF-1α might also regulate the expression of Glut1 in osteoblasts of chondro-osteo junction area, being consistent with our result. Polyubiquitination by von Hippel–Lindau tumor suppressor protein (pVHL) promotes proteasome-dependent degradation of HIF-1α [33]. This process requires hydroxylation of proline residues in HIF-1α, and ascorbic acid acts as a coenzyme for the hydroxylation [33]. As ascorbic acid promotes HIF-1α degradation, it is expected to attenuate HIF-1α expression in ascorbic acid-induced osteoblast differentiation, in contrast with increased HIF-1α expression during BMP-derived differentiation. In this study, we found that ascorbic acid-derivatives induced Glut1 expression without an increase of HIF-1α expression in differentiated primary osteoblasts, whereas hypoxia induced Glut1 expression with an increase of HIF-1α expression in undifferentiated osteoblasts (Figure 4H). Lastly, c-Myc has the ability to indirectly increase Glut1 expression mediated through the promotion of HIF expression and inhibit the degradation of HIF-1α. [49]. In addition, the trafficking of Glut1 to the cell surface is Akt can promote glycolysis through increased glucose transporter [50]. In perichondrium, c-Myc is expressed in proliferating chondrocytes of growth plate and osteoblasts [51]. c-Myc promotes the proliferative expansion of osteoblasts to regulate endochondral growth and ossification rather than control of osteoblast differentiation [51]. We observed that primary osteoblasts did not change the expression of c-Myc and Gls, the transcriptional activity of which is regulated by c-Myc in cancer cells, during osteoblast differentiation [11] (Supplementary Figure S1J,K).

We observed that mRNA expressions of several glycolytic enzymes including Hk1, Hk2, Ldha, Pfk, and Pkm2 other than Slc2a1 increased during ascorbic acid-induced osteoblast differentiation. In this study, we found that p53 negatively regulated Glut1 transcriptional activity to convert osteoblast metabolism to glycolysis. Wnt signal also controls osteoblast differentiation and activity, and induces aerobic glycolysis increasing the translational levels of glycolytic enzymes, including Glut1, HK2, Ldha, and Pdk1 through mTOR complex 2 (mTORC2) activation [34] (Supplementary Figure S4). These previous reports and our data prompted us to examine if p53 is involved in Wnt-signaling that mediates the induction of Glut1 expression. However, we found that increased p53 level did not alter Wnt3a-induced Glut1 expression in undifferentiated cells of pre-OB fr, suggesting that p53 did not suppress Wnt signal-induced Glut1 expression in undifferentiated osteoblasts (Supplementary Figure S3). In addition, there is a report that p53 attenuates Wnt signal by inhibiting β-catenin through miR-34, which is a member of the tumor suppressor miRNA family [52]. However, this report does not suggest that p53 suppresses the Wnt signal-induced Glut1 expression because the Wnt signal-induced Glut1 expression is mediated through mTORC2, but not β-catenin (Supplementary Figure S4) [52,53].

Tumor cells also switch expression isoforms of pyruvate kinase, which is a glycolytic enzyme for the production of ATP, from Pkm1 to Pkm2. It is known that glycolytic energy production by PK can inhibit oxidative phosphorylation in mitochondria, which is described as Crabtree effect [54]. Alternative splicing of PKM pre-mRNA derives both PKM1 and PKM2 isoforms. M2 isoform has been found in cancer cells with an absence of M1 form, which is expressed in muscle and brain [55]. However, PKM2 is usually in the form of dimmer, which is inactive in the enzymatic activity [26]. Recently, PKM2 has been the focus of attention as a signaling molecule, as the phosphorylation of PKM2 causes interaction with fms-related tyrosine kinase, Janus kinase 2, protein kinase C (PKC) δ, transcription factor HIF or histone H3 [26,56,57]. In this study, we found that osteoblast differentiation enhanced mRNA expression of hnRNPA1, a mediator of Pkm2 mRNA splicing.

We found that the inhibition of glucose transport decreased ATP synthesis in mature osteoblasts, but not in pre-mature osteoblasts, indicating that glycolysis is the essential metabolic pathway for mature primary osteoblasts. Glut1 expression in osteoblasts has been reported to be regulated by Wnt-signaling molecules and Runx2, both of which are osteogenic factors. In this study, we found that p53 repressed Glut1 transcriptional activity during osteoblast differentiation.

Competing Interests

The authors declare that there are no competing interests associated with the manuscript.

Funding

This work was supported by KAKENHI from the Japan Society for the Promotion of Science (Grant 17K11646, 17K11676).

Author Contributions

T.O., K.B., and T.M. designed the experiments; T.O. and J.K. carried out the experiments. T.O. and T.M. wrote the manuscript, and all authors contributed to data analysis. All authors have approved the manuscript.

Acknowledgement

This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Abbreviations

     
  • 2-DG

    2-deoxy-d-glucose

  •  
  • 4-MUG

    4-methylumbel-liferyl-β-d-galactoside

  •  
  • AA2P

    l-ascorbic acid 2-phosphate trisodium

  •  
  • CCCP

    cyanide 3-chlorophenylhydrazone

  •  
  • Col1a1

    collagen type 1, α1

  •  
  • FCS

    fetal calf serum

  •  
  • GAPDH

    glyceraldehyde 3-phosphate dehydrogenase

  •  
  • GCR1

    glycolytic genes transcriptional activator 1

  •  
  • HIF-1α

    hypoxia-inducible factor 1-α

  •  
  • HK2

    hexokinase 2

  •  
  • IGF-1

    insulin-like growth factor I

  •  
  • LDHA

    lactate dehydrogenase A

  •  
  • Mdm2

    mouse double minute 2

  •  
  • PCR

    polymerase chain reaction

  •  
  • PDK1

    pyruvate dehydrogenase kinase 1

  •  
  • PI3K

    phosphoinositol-3 kinase

  •  
  • PKC

    protein kinase C

  •  
  • RT

    reverse transcription

  •  
  • Runx2

    runt-related transcription factor 2

  •  
  • αMEM

    α-minimal essential medium

References

References
1
Heiden
,
V.
,
Cantley
,
M.G.
,
& Thompson
,
L.C.
and
B
,
C.
(
2009
)
Understanding the Warburg effect: the metabolic requirements of cell proliferation
.
Science
324
,
1029
1033
2
Ward
,
P.S.
and
Thompson
,
C.B.
(
2012
)
Metabolic reprogramming: a cancer hallmark even Warburg did not anticipate
.
Cancer Cell
21
,
297
308
3
Kasischke
,
K.A.
,
Vishwasrao
,
H.D.
,
Fisher
,
P.J.
,
Zipfel
,
W.R.
and
Webb
,
W.W.
(
2004
)
Neural activity triggers neuronal oxidative metabolism followed by astrocytic glycolysis
.
Science
305
,
99
103
4
Bonvento
,
G.
,
Herard
,
A.S.
and
Voutsinos-Porche
,
B.
(
2005
)
The astrocyte–neuron lactate shuttle: a debated but still valuable hypothesis for brain imaging
.
J. Cereb. Blood Flow Metab.
25
,
1394
1399
5
Jones
,
R.G.
and
Thompson
,
C.B.
(
2007
)
Revving the engine: signal transduction fuels T cell activation
.
Immunity
27
,
173
178
6
Borle
,
A.B.
,
Nichols
,
N.
and
Nichols
, Jr,
G.
(
1960
)
Metabolic studies of bone in vitro. I. Normal bone
.
J. Biol. Chem.
235
,
1206
1210
PMID:
[PubMed]
7
Cohn
,
D.V.
and
Forscher
,
B.K.
(
1962
)
Aerobic metabolism of glucose by bone
.
J. Biol. Chem.
237
,
615
618
PMID:
[PubMed]
8
Guntur
,
A.R.
,
Le
,
P.T.
,
Farber
,
C.R.
and
Rosen
,
C.J.
(
2014
)
Bioenergetics during calvarial osteoblast differentiation reflect strain differences in bone mass
.
Endocrinology
155
,
1589
1595
9
Lee
,
W.C.
,
Guntur
,
A.R.
,
Long
,
F.
and
Rosen
,
C.J.
(
2017
)
Energy metabolism of the osteoblast: implications for osteoporosis
.
Endocr. Rev.
38
,
255
266
10
Vaupel
,
P.
(
2004
)
The role of hypoxia-induced factors in tumor progression
.
Oncologist
9
(
Suppl 5
),
10
17
11
Jones
,
N.P.
and
Schulze
,
A.
(
2012
)
Targeting cancer metabolism–aiming at a tumour's sweet-spot
.
Drug Discov. Today
17
,
232
241
12
Vousden
,
K.H.
and
Ryan
,
K.M.
(
2009
)
P53 and metabolism
.
Nat. Rev. Cancer
9
,
691
700
13
Kubbutat
,
M.H.
and
Vousden
,
K.H.
(
1998
)
Keeping an old friend under control: regulation of p53 stability
.
Mol. Med. Today
4
,
250
256
14
Liu
,
H.
and
Li
,
B.
(
2010
)
P53 control of bone remodeling
.
J. Cell. Biochem.
111
,
529
534
15
Lengner
,
C.J.
,
Steinman
,
H.A.
,
Gagnon
,
J.
,
Smith
,
T.W.
,
Henderson
,
J.E.
,
Kream
,
B.E.
et al (
2006
)
Osteoblast differentiation and skeletal development are regulated by Mdm2-p53 signaling
.
J. Cell Biol.
172
,
909
921
16
Xian
,
L.
,
Wu
,
X.
,
Pang
,
L.
,
Lou
,
M.
,
Rosen
,
C.J.
,
Qiu
,
T.
et al (
2012
)
Matrix IGF-1 maintains bone mass by activation of mTOR in mesenchymal stem cells
.
Nat. Med.
18
,
1095
1101
17
Zhang
,
W.
,
Shen
,
X.
,
Wan
,
C.
,
Zhao
,
Q.
,
Zhang
,
L.
,
Zhou
,
Q.
et al (
2012
)
Effects of insulin and insulin-like growth factor 1 on osteoblast proliferation and differentiation: differential signalling via Akt and ERK
.
Cell Biochem. Funct.
30
,
297
302
18
Zoidis
,
E.
,
Ghirlanda-Keller
,
C.
and
Schmid
,
C.
(
2011
)
Stimulation of glucose transport in osteoblastic cells by parathyroid hormone and insulin-like growth factor I
.
Mol. Cell. Biochem.
348
,
33
42
19
Wei
,
J.
,
Shimazu
,
J.
,
Makinistoglu
,
M.P.
,
Maurizi
,
A.
,
Kajimura
,
D.
,
Zong
,
H.
et al (
2015
)
Glucose uptake and Runx2 synergize to orchestrate osteoblast differentiation and bone formation
.
Cell
161
,
1576
1591
20
Ohnishi
,
T.
,
Okamoto
,
A.
,
Kakimoto
,
K.
,
Bandow
,
K.
,
Chiba
,
N.
and
Matsuguchi
,
T.
(
2010
)
Involvement of Cot/Tp12 in bone loss during periodontitis
.
J. Dent. Res.
89
,
192
197
21
Kusuyama
,
J.
,
Bandow
,
K.
,
Shamoto
,
M.
,
Kakimoto
,
K.
,
Ohnishi
,
T.
and
Matsuguchi
,
T.
(
2014
)
Low intensity pulsed ultrasound (LIPUS) influences the multilineage differentiation of mesenchymal stem and progenitor cell lines through ROCK-Cot/Tpl2-MEK-ERK signaling pathway
.
J. Biol. Chem.
289
,
10330
10344
22
Ohnishi
,
T.
,
Bandow
,
K.
,
Kakimoto
,
K.
,
Kusuyama
,
J.
and
Matsuguchi
,
T.
(
2014
)
Long-time treatment by low-dose N-acetyl-L-cysteine enhances proinflammatory cytokine expressions in LPS-stimulated macrophages
.
PLoS One
9
,
e87229
23
Bandow
,
K.
,
Maeda
,
A.
,
Kakimoto
,
K.
,
Kusuyama
,
J.
,
Shamoto
,
M.
,
Ohnishi
,
T.
et al (
2010
)
Molecular mechanisms of the inhibitory effect of lipopolysaccharide (LPS) on osteoblast differentiation
.
Biochem. Biophys. Res. Commun.
402
,
755
761
24
Ohnishi
,
T.
,
Kakimoto
,
K.
,
Hashida
,
S.
,
Fujii
,
M.
,
Hirono
,
S.
,
Nishiyama
,
K.
et al (
2000
)
Development of highly sensitive enzyme-linked immunosorbent assays for hepatocyte growth factor/scatter factor (HGF/SF): determination of HGF/SF in serum and urine from normal human subjects
.
J. Immunol. Methods
244
,
163
173
25
Vaupel
,
P.
,
Kallinowski
,
F.
and
Okunieff
,
P.
(
1989
)
Blood flow, oxygen and nutrient supply, and metabolic microenvironment of human tumors: a review
.
Cancer Res.
49
,
6449
6465
PMID:
[PubMed]
26
Wu
,
W.
and
Zhao
,
S.
(
2013
)
Metabolic changes in cancer: beyond the Warburg effect
.
Acta Biochim. Biophys. Sin. (Shanghai)
45
,
18
26
27
Mazurek
,
S.
,
Boschek
,
C.B.
,
Hugo
,
F.
and
Eigenbrodt
,
E.
(
2005
)
Pyruvate kinase type M2 and its role in tumor growth and spreading
.
Semin. Cancer Biol.
15
,
300
308
28
Chen
,
M.
,
Zhang
,
J.
and
Manley
,
J.L.
(
2010
)
Turning on a fuel switch of cancer: hnRNP proteins regulate alternative splicing of pyruvate kinase mRNA
.
Cancer Res.
70
,
8977
8980
29
David
,
C.J.
,
Chen
,
M.
,
Assanah
,
M.
,
Canoll
,
P.
and
Manley
,
J.L.
(
2010
)
HnRNP proteins controlled by c-Myc deregulate pyruvate kinase mRNA splicing in cancer
.
Nature
463
,
364
368
30
Wang
,
X.
,
Kua
,
H.Y.
,
Hu
,
Y.
,
Guo
,
K.
,
Zeng
,
Q.
,
Wu
,
Q.
et al (
2006
)
P53 functions as a negative regulator of osteoblastogenesis, osteoblast-dependent osteoclastogenesis, and bone remodeling
.
J. Cell Biol.
172
,
115
125
31
Artigas
,
N.
,
Gamez
,
B.
,
Cubillos-Rojas
,
M.
,
Sanchez-de Diego
,
C.
,
Valer
,
J.A.
,
Pons
,
G.
et al (
2017
)
P53 inhibits SP7/osterix activity in the transcriptional program of osteoblast differentiation
.
Cell Death Differ.
24
,
2022
2031
32
Wan
,
C.
,
Shao
,
J.
,
Gilbert
,
S.R.
,
Riddle
,
R.C.
,
Long
,
F.
,
Johnson
,
R.S.
et al (
2010
)
Role of HIF-1α in skeletal development
.
Ann. NY Acad. Sci.
1192
,
322
326
33
Kaelin
, Jr,
W.G.
and
Ratcliffe
,
P.J.
(
2008
)
Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway
.
Mol. Cell
30
,
393
402
34
Karner
,
C.M.
and
Long
,
F.
(
2017
)
Wnt signaling and cellular metabolism in osteoblasts
.
Cell. Mol. Life Sci.
74
,
1649
1657
35
Kel
,
A.E.
,
Gossling
,
E.
,
Reuter
,
I.
,
Cheremushkin
,
E.
,
Kel-Margoulis
,
O.V.
and
Wingender
,
E.
(
2003
)
MATCH: a tool for searching transcription factor binding sites in DNA sequences
.
Nucleic Acids Res.
31
,
3576
3579
36
Pham
,
N.L.
,
Franzen
,
A.
and
Levin
,
E.G.
(
2004
)
NF1 regulatory element functions as a repressor of tissue plasminogen activator expression
.
Arterioscler Thromb. Vasc. Biol.
24
,
982
987
37
Barath
,
P.
,
Poliakova
,
D.
,
Luciakova
,
K.
and
Nelson
,
B.D.
(
2004
)
Identification of NF1 as a silencer protein of the human adenine nucleotide translocase-2 gene
.
Eur. J. Biochem.
271
,
1781
1788
38
Olianas
,
M.C.
,
Dedoni
,
S.
and
Onali
,
P.
(
2011
)
δ-Opioid receptors stimulate GLUT1-mediated glucose uptake through Src- and IGF-1 receptor-dependent activation of PI3-kinase signalling in CHO cells
.
Br. J. Pharmacol.
163
,
624
637
39
Basu
,
S.
,
Hubbard
,
B.
and
Shevach
,
E.M.
(
2015
)
Foxp3-mediated inhibition of Akt inhibits Glut1 (glucose transporter 1) expression in human T regulatory cells
.
J. Leukoc. Biol.
97
,
279
283
40
Mayo
,
L.D.
and
Donner
,
D.B.
(
2002
)
The PTEN, Mdm2, p53 tumor suppressor-oncoprotein network
.
Trends Biochem. Sci.
27
,
462
467
41
Nowacki
,
T.M.
,
Kuerten
,
S.
,
Zhang
,
W.
,
Shive
,
C.L.
,
Kreher
,
C.R.
,
Boehm
,
B.O.
et al (
2007
)
Granzyme B production distinguishes recently activated CD8(+) memory cells from resting memory cells
.
Cell Immunol.
247
,
36
48
42
Sanchez-Arago
,
M.
,
Garcia-Bermudez
,
J.
,
Martinez-Reyes
,
I.
,
Santacatterina
,
F.
and
Cuezva
,
J.M.
(
2013
)
Degradation of IF1 controls energy metabolism during osteogenic differentiation of stem cells
.
EMBO Rep.
14
,
638
644
43
Ghosh-Choudhury
,
N.
,
Abboud
,
S.L.
,
Nishimura
,
R.
,
Celeste
,
A.
,
Mahimainathan
,
L.
and
Choudhury
,
G.G.
(
2002
)
Requirement of BMP-2-induced phosphatidylinositol 3-kinase and Akt serine/threonine kinase in osteoblast differentiation and Smad-dependent BMP-2 gene transcription
.
J. Biol. Chem.
277
,
33361
33368
44
Vinals
,
F.
,
Lopez-Rovira
,
T.
,
Rosa
,
J.L.
and
Ventura
,
F.
(
2002
)
Inhibition of PI3K/p70 S6K and p38 MAPK cascades increases osteoblastic differentiation induced by BMP-2
.
FEBS Lett.
510
,
99
104
45
Sakakura
,
Y.
,
Shibui
,
T.
,
Irie
,
K.
and
Yajima
,
T.
(
2008
)
Metabolic mode peculiar to Meckel's cartilage: immunohistochemical comparisons of hypoxia-inducible factor-1α and glucose transporters in developing endochondral bones in mice
.
Eur. J. Oral Sci.
116
,
341
352
46
Schwartzenberg-Bar-Yoseph
,
F.
,
Armoni
,
M.
and
Karnieli
,
E.
(
2004
)
The tumor suppressor p53 down-regulates glucose transporters GLUT1 and GLUT4 gene expression
.
Cancer Res.
64
,
2627
2633
47
Inman
,
C.K.
and
Shore
,
P.
(
2003
)
The osteoblast transcription factor Runx2 is expressed in mammary epithelial cells and mediates osteopontin expression
.
J. Biol. Chem.
278
,
48684
48689
48
Regan
,
J.N.
,
Lim
,
J.
,
Shi
,
Y.
,
Joeng
,
K.S.
,
Arbeit
,
J.M.
,
Shohet
,
R.V.
et al (
2014
)
Up-regulation of glycolytic metabolism is required for HIF1α-driven bone formation
.
Proc. Natl. Acad. Sci. U.S.A.
111
,
8673
8678
49
Dang
,
C.V.
,
Kim
,
J.W.
,
Gao
,
P.
and
Yustein
,
J.
(
2008
)
The interplay between MYC and HIF in cancer
.
Nat. Rev. Cancer
8
,
51
56
50
Wieman
,
H.L.
,
Wofford
,
J.A.
and
Rathmell
,
J.C.
(
2007
)
Cytokine stimulation promotes glucose uptake via phosphatidylinositol-3 kinase/Akt regulation of Glut1 activity and trafficking
.
Mol. Biol. Cell
18
,
1437
1446
51
Zhou
,
Z.Q.
,
Shung
,
C.Y.
,
Ota
,
S.
,
Akiyama
,
H.
,
Keene
,
D.R.
and
Hurlin
,
P.J.
(
2011
)
Sequential and coordinated actions of c-Myc and N-Myc control appendicular skeletal development
.
PLoS One
6
,
e18795
52
Kim
,
N.H.
,
Kim
,
H.S.
,
Kim
,
N.G.
,
Lee
,
I.
,
Choi
,
H.S.
,
Li
,
X.Y.
et al (
2011
)
P53 and microRNA-34 are suppressors of canonical Wnt signaling
.
Sci. Signal.
4
,
ra71
53
Esen
,
E.
,
Chen
,
J.
,
Karner
,
C.M.
,
Okunade
,
A.L.
,
Patterson
,
B.W.
and
Long
,
F.
(
2013
)
WNT-LRP5 signaling induces Warburg effect through mTORC2 activation during osteoblast differentiation
.
Cell Metab.
17
,
745
755
54
Melo
,
R.F.
,
Stevan
,
F.R.
,
Campello
,
A.P.
,
Carnieri
,
E.G.
and
de Oliveira
,
M.B.
(
1998
)
Occurrence of the Crabtree effect in HeLa cells
.
Cell Biochem. Funct.
16
,
99
105
55
Mazurek
,
S.
,
Grimm
,
H.
,
Boschek
,
C.B.
,
Vaupel
,
P.
and
Eigenbrodt
,
E.
(
2002
)
Pyruvate kinase type M2: a crossroad in the tumor metabolome
.
Br. J. Nutr.
87
(
Suppl 1
),
S23
S29
56
Christofk
,
H.R.
,
Vander Heiden
,
M.G.
,
Wu
,
N.
,
Asara
,
J.M.
and
Cantley
,
L.C.
(
2008
)
Pyruvate kinase M2 is a phosphotyrosine-binding protein
.
Nature
452
,
181
186
57
Mazurek
,
S.
(
2011
)
Pyruvate kinase type M2: a key regulator of the metabolic budget system in tumor cells
.
Int. J. Biochem. Cell Biol.
43
,
969
980