Osteosarcoma and chondrosarcoma are sarcomas of the bone and the cartilage that are primarily treated by surgical intervention combined with high toxicity chemotherapy. In search of alternative metabolic approaches to address the challenges in treating bone sarcomas, we assessed the growth dependence of these cancers on leucine, one of the branched-chain amino acids (BCAAs), and BCAA metabolism. Tumor biopsies from bone sarcoma patients revealed differential expression of BCAA metabolic enzymes. The cytosolic branched-chain aminotransferase (BCATc) that is commonly overexpressed in cancer cells, was down-regulated in chondrosarcoma (SW1353) in contrast with osteosarcoma (143B) cells that expressed both BCATc and its mitochondrial isoform BCATm. Treating SW1353 cells with gabapentin, a selective inhibitor of BCATc, further revealed that these cells failed to respond to gabapentin. Application of the structural analog of leucine, N-acetyl-leucine amide (NALA) to disrupt leucine uptake, indicated that all bone sarcoma cells used leucine to support their energy metabolism and biosynthetic demands. This was evident from the increased activity of the energy sensor AMP-activated protein kinase (AMPK), down-regulation of complex 1 of the mammalian target of rapamycin (mTORC1), and reduced cell viability in response to NALA. The observed changes were most profound in the 143B cells, which appeared highly dependent on cytosolic and mitochondrial BCAA metabolism. This study thus demonstrates that bone sarcomas rely on leucine and BCAA metabolism for energy and growth; however, the differential expression of BCAA enzymes and the presence of other carbon sources may dictate how efficiently these cancer cells take advantage of BCAA metabolism.

Introduction

Leucine, and the other branched-chain amino acids (BCAAs), isoleucine, and valine, have received substantial scientific interest in recent years with respect to their role in tumor progression [1]. BCAAs are considered important sources of carbon, nitrogen, and energy for cancer growth [1]. How cancer cells utilize BCAAs, however, depends on the tumor type, aggressiveness, and tumor burden [2]. BCAAs may either be metabolized to produce energy, or may participate in intracellular signaling events, such as mTOR pathway to co-ordinate protein synthesis and cellular growth [2–4]. Under conditions of extreme catabolism, such as cancer cachexia, BCAA degradation is significantly increased. Degradation of BCAAs begins with two common steps, after which each individual BCAA follows its unique catabolic pathway [2]. The first step in BCAA degradation is a reversible transamination reaction catalyzed by the cytosolic and mitochondrial branched-chain aminotransferases (BCATc and BCATm) [5,6]. The transfer of the α-amino group to α-ketoglutarate produces glutamate and the corresponding branched-chain α-keto acids (BCKAs). The second step in BCAA degradation is the irreversible oxidative decarboxylation of BCKAs catalyzed by the branched-chain α-keto acid dehydrogenase complex (BCKDC), an intramitochondrial multienzyme complex organized around a core scaffold of dihydrolipoamide acyltransferase (E2) subunits to which multiple copies of the branched-chain α-keto acid dehydrogenase (E1) and the dihydrolipoamide dehydrogenase (E3) are attached noncovalently [7]. The products of the BCKDC reaction are ultimately converted into intermediates, such as acetyl-CoA, propionyl-CoA, and succinyl-CoA, that are utilized in the TCA cycle, glycolysis, and other metabolic pathways [1] (Figure 1). BCATm and BCKDC account for the mitochondrial transamination and oxidation of BCAAs and they are ubiquitously expressed in most adult human tissues. In contrast, BCATc is responsible for the cytosolic transamination of BCAAs and is mostly expressed in the central nervous system and proliferating cells [7,8]. However, recent research has revealed that BCATc is up-regulated in many different cancers, including cancers of the brain, liver, breast, stomach, pancreas, and ovaries, which have implicated BCATc as a prognostic marker for cancer [1,9–13].

Schematic illustration of leucine metabolism and signaling.

Figure 1.
Schematic illustration of leucine metabolism and signaling.

Leucine, isoleucine, and valine share the first two steps (BCAT and BCKDC) in their catabolism followed by unique degradation pathways each leading to important intermediates for energy, biosynthesis, and glycolysis. The schematic follows the leucine degradation pathway where α-ketoisocaproate (KIC) is metabolized into isovaleryl-CoA and ultimately converted into acetyl-CoA. Apart from being metabolized, leucine stimulates protein synthesis by activating mTORC1 pathway. Inhibitors of leucine uptake (NALA), BCATc (gabapentin), mTORC1 pathway (rapamycin), glycolysis (2-DG), and an activator of AMPK (metformin) are shown.

Figure 1.
Schematic illustration of leucine metabolism and signaling.

Leucine, isoleucine, and valine share the first two steps (BCAT and BCKDC) in their catabolism followed by unique degradation pathways each leading to important intermediates for energy, biosynthesis, and glycolysis. The schematic follows the leucine degradation pathway where α-ketoisocaproate (KIC) is metabolized into isovaleryl-CoA and ultimately converted into acetyl-CoA. Apart from being metabolized, leucine stimulates protein synthesis by activating mTORC1 pathway. Inhibitors of leucine uptake (NALA), BCATc (gabapentin), mTORC1 pathway (rapamycin), glycolysis (2-DG), and an activator of AMPK (metformin) are shown.

Although BCATc and oncogenic BCAA metabolism received substantial attention in the recent years, their role in primary bone cancers, such as chondrosarcoma and osteosarcoma, has never been studied. Osteosarcomas and chondrosarcomas are the two most common and challenging to treat primary bone cancers [14]. They are known as difficult-to-treat because they are largely chemo- and radiotherapy resistant [15–17]. Moreover, the chemotherapy regimens used for primary bone cancers are associated with significant toxicity including long-term complications [18,19]. The 5-year survival rate of many cancer types has increased in the last decade, however, the clinical outcomes for patients with osteosarcoma and chondrosarcoma have not shown comparable improvement [14,20]. A powerful approach to identify novel targets for bone cancer treatment is to explore the role of cellular metabolism on the growth and development of cancer cells. This study aimed to assess how different bone sarcomas metabolize leucine followed by investigation on the impact of disrupted leucine uptake and inhibited BCAA degradation (at the BCATc step) on the metabolism and growth of these cancers.

Experimental

Antibodies and reagents

Antibodies against the ribosomal protein S6 (S6), phosphorylated (P)-S6 (Ser240/244), β-tubulin, AMPKα, P-AMPKα (Thr172), p27Kip1, retinoblastoma (Rb), and P-Rb (Ser807/811) were purchased from Cell Signaling Technologies (Danvers, MA). Antibodies against BCATm, BCATc, E1α, and E2 were previously described [7,21]. Horseradish peroxidase-conjugated donkey anti-rabbit antibody was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Rapamycin, NALA, gabapentin, 2-deoxy-d-glucose (2-DG), metformin, propidium iodide (PI), and RNase A were purchased from LC Laboratories (Woburn, MA), BACHEM (Bubendorf, Switzerland), Chem-Impex Int'l Inc. (Wood Dale, IL), Acros Organics (NJ, U.S.A.), Alfa Aesar Chemicals (Tewksbury, MA), and Thermo Scientific (Waltham, MA), respectively.

DNA methylation and RNA expression data analysis

All analyses and visualizations of published chondrosarcoma DNA methylation (Illumina 450k bead arrays) and RNA expression (Affymetrix hugene20t) data of chondrosarcomas [22,23] and osteosarcomas [24,25] were performed using the ‘R2: Genomics Analysis and Visualization Platform’ (http://r2.amc.nlhttp://r2platform.com).

Tumor biopsies

Tumor biopsies from bone sarcoma patients were generously provided by Dr. Charles Keller (Oregon Health and Science University). All tumor biopsy donors provided informed consent. The pathological diagnosis, age, gender, and the source of biopsies are described in Table 1. The use of human material was approved by the Institutional Review Board (IRB ID 02-15-01) and the Institutional Biosafety Committee (IBC ID 9-14-3) of Des Moines University.

Table 1.
Human tumor biopsy characteristics
BiopsyPathological diagnosisGradeAgeGenderChemotherapy/radiationSource
Chondrosarcoma Grade 1 52 Male naive Surgical 
Chondrosarcoma Grade 2 53 Male naive Surgical 
3* Osteosarcoma Grade 2 Female n/a Surgical 
4** Osteosarcoma Grade 2 48 Female naive Surgical 
BiopsyPathological diagnosisGradeAgeGenderChemotherapy/radiationSource
Chondrosarcoma Grade 1 52 Male naive Surgical 
Chondrosarcoma Grade 2 53 Male naive Surgical 
3* Osteosarcoma Grade 2 Female n/a Surgical 
4** Osteosarcoma Grade 2 48 Female naive Surgical 

*tumor biopsy that contains small chondroblastic areas; **tumor biopsy that contains osteoblastic, chondroblastic, and telangiectatic areas.

Cell cultures and treatments

The human chondroblastoma cell line PCB0023 was kindly provided by Dr. Charles Keller (Oregon Health and Science University), while the human osteosarcoma 143B, chondrosarcoma SW1353, and osteoblast hFOB1.19 cell lines were purchased from ATCC (Manassas, VA) and maintained in a Dulbecco's modified Eagle's medium (DMEM, Mediatech), formulated with 25 mM glucose and 0.8 mM of each BCAA, and supplemented with 10% fetal bovine serum (FBS) (HyClone, UT), 100 IU/ml of penicillin and 100 µg/ml of streptomycin (HyClone, UT), in a tissue culture incubator at 37°C and 5% CO2. The human osteoblast cell line, hFOB1.19, was maintained as described above with the following modifications: the growth medium was 1 : 1 mixture of Ham's F12 Medium:DMEM, with 2.5 mM l-glutamine (without phenol red) and the tissue culture incubator temperature was 34°C to stimulate proliferation or 39°C to stimulate differentiation. Proliferation growth curves for each bone sarcoma cell line were generated using cell counts collected every 24 h for a 96 h period by the trypan blue dye exclusion method [26]. For NALA and gabapentin concentration-gradient growth curves, between 0.2 and 2 × 106cells from each cell line were treated with different concentrations of NALA (0, 12.5, 25, and 50 mM) or gabapentin (0, 5, 10, 20, and 50 mM) for 24 h (osteosarcoma) or 24 and 48 h (chondrosarcoma, chondroblastoma), respectively. For the rest of the treatments, each cell line was incubated in DMEM, 10% FBS, and antibiotics (unless otherwise specified) and left untreated or treated with NALA (25 mM), gabapentin (5 or 10 mM), rapamycin (100 nM), 2-DG (25 mM), or metformin (5 mM) for 24 h. At the end of each treatment, cells were either prepared for metabolic assays as described below or washed twice with an ice-cold PBS buffer, pelleted at 18 500g for 5–10 min at 4°C, and stored at −80°C until Western Blotting as described below.

Western Blotting

Protein was extracted from ground tumor biopsies or bone cell pellets followed by determination of total protein and Western Blotting as described [8]. ImageJ software [27] was used to quantify protein bands (BCATc, BCATm, E1α, E2, p27Kip1) and normalize to a loading control (β-tubulin) or to calculate the ratio between phosphorylated and total concentrations of S6, AMPKα, and Rb.

Lactate assay

Each bone tumorigenic cell line was treated with gabapentin, NALA, rapamycin, and 2-DG as indicated above and grown in phenol red- and glucose-free DMEM media supplemented with Cell-Ess serum replacement (Essential Pharmaceuticals, LLC). Untreated and treated cells were next separated in three equal portions, and supplemented with 0, 5, or 25 mM glucose, respectively. After 24 h, cell pellets and supernatants were collected. The cell pellets were used in Western Blotting, while the supernatants were used to determine the amount of extracellular l-lactate. Each supernatant was diluted between 5- and 10-fold and incubated in l-lactate assay solution (Eton Biosciences, INC) for 30 min at 37°C followed by addition of 0.5 M acetic acid to stop the reaction, and the absorbance was measured at 490 nm using a microplate reader. The concentration of extracellular l-lactate (mM lactate/mg protein) was calculated by using the equation obtained from the linear regression of a standard lactate curve.

Leucine oxidation and transamination assay

Between 1 and 5 × 106 cells from each tumorigenic cell line, were either untreated or treated with 10 mM gabapentin or 25 mM NALA for 24 h followed by a radioactive metabolic assay of each cell suspension in the presence of 1 mM [1–14C] leucine (specific activity 200 dpm/nmol) to measure leucine transamination and oxidation, which was described previously [8]. Cell suspensions from each cell line were run in triplicate and counts of radioactivity in dpm were recorded. Next, dpms of captured 14CO2 were converted into pmol14CO2/min/mg protein.

MTT assay

Measurement of cell viability was performed by feeding the cells with yellow tetrazolium MTT (0.5 mg/1 ml, Tocris Bioscience), which is reduced to insoluble formazan by metabolically active cells. Following a 24 h of treatment with 5 mM gabapentin, 25 mM NALA, 100 nM rapamycin, or 5 mM metformin, all cells were incubated with MTT for 1 h in a tissue culture incubator at 37°C. The cells were then observed under a light microscope for the presence of purple formazan crystals. The formazan was solubilized with DMSO and the absorbance was measured at 570 and 620 nm using a microplate reader. Results were obtained by calculating the difference between the two absorbances and are presented as a percentage of the values obtained with untreated cells.

ATP assay

ATP was extracted from cells using the boiling water method [28]. Following a 24 h treatment with 5 mM gabapentin or 25 mM NALA, cell pellets were collected and resuspended in ice-cold H2O, followed by boiling of cell lysates for 10 min. Cell debris were collected by centrifugation at 18 500g for 5 min at 4°C. The extracted ATP was found in the supernatant. The concentration of ATP was determined using an ATP Assay Kit (Cat #: A-125; Biomedical Research Service Center, University at Buffalo, State University of New York). The bioluminescence was measured using a microplate reader. The concentration of cellular ATP (µM) was calculated by reference to a standard curve of ATP.

Cell cycle analysis

Between 1 and 2 × 106 cells from each tumorigenic cell line were either untreated or treated with 5 mM gabapentin or 25 mM NALA for 24 h followed by pelleting the cells at 500 g for 5 min at 4°C. The pellets were then washed with ice-cold PBS and after another centrifugation, they were resuspended in 400 µl ice-cold PBS to which 800 µl of ice cold 100% ethanol (final concentration 66%) was slowly added. The cells were incubated in ethanol for at least 2 h at 4°C followed by pelleting of the cells as above. Next, the cells were washed with PBS and incubated in a 50 µg/ml PI and 100 µg/ml RNase A staining solution for 30 min, in the dark, at 37°C. Fluorescence activated cell sorter (FACS) analysis was performed to monitor the progression of the cells through the cell cycle.

Statistical analysis

One-way ANOVA was used to determine statistically significant differences between the three bone cell lines and between untreated and treated with NALA, gabapentin, rapamycin, metformin, or 2-DG cells from each bone cell line. Values are means ± SEM and P ≤ 0.05 was considered statistically significant.

Results

Osteosarcoma and chondrosarcoma differ in the protein expression of BCAA metabolic enzymes

Initial screening of a small set of tumor biopsies, obtained from patients with chondrosarcoma and osteosarcoma, has revealed differential expression pattern between the two subtypes of bone sarcomas where the chondrosarcoma tissues expressed relatively low BCATc protein (Table 1 and Figure 2A, compare BCATc in 1,2 with 3,4). In contrast, the protein expression of BCATc was easily detectable in the osteosarcoma biopsies (Figure 2A, tumor biopsies 3,4). Comparison of other BCAA metabolic enzymes, represented by the mitochondrial BCATm, the α-subunit of enzyme E1 (E1α), and enzyme E2 of the BCKDC complex, demonstrated that they had similar protein expression patterns, but varied in intensity among the studied biopsies (Figure 2A).

Differential expression of BCAA metabolic enzymes in human bone sarcomas.

Figure 2.
Differential expression of BCAA metabolic enzymes in human bone sarcomas.

(A) Protein expression of BCATc, BCATm, BCKDC-E1α, and BCKDC-E2 assessed by Western Blotting in tumor biopsies from patients with chondrosarcoma (1,2) and osteosarcoma (3,4). The loading control was a 25 kDa band from Coomassie Brilliant Blue staining of the PVDF membrane. Samples were run on a SDS–protein gel between 2 and 4 times and representative blot images are shown. (B–D) DNA methylation and RNA seq data analysis. (B) DNA methylation analysis of the BCAT1 locus in IDH1 mutant (IDH1mut), IDH2 mutant (IDH2mut), or WT chondrosarcomas. (C,D) RNA seq analysis of BCAT1, BCAT2, BCKDHA, and DBT genes in chondrosarcomas (n = 82) (C) and osteosarcomas (n = 53) (D). *P ≤ 0.05 as compared with IDH1mut chondrosarcomas. **P ≤ 0.05 as compared with IDH2mut chondrosarcomas.

Figure 2.
Differential expression of BCAA metabolic enzymes in human bone sarcomas.

(A) Protein expression of BCATc, BCATm, BCKDC-E1α, and BCKDC-E2 assessed by Western Blotting in tumor biopsies from patients with chondrosarcoma (1,2) and osteosarcoma (3,4). The loading control was a 25 kDa band from Coomassie Brilliant Blue staining of the PVDF membrane. Samples were run on a SDS–protein gel between 2 and 4 times and representative blot images are shown. (B–D) DNA methylation and RNA seq data analysis. (B) DNA methylation analysis of the BCAT1 locus in IDH1 mutant (IDH1mut), IDH2 mutant (IDH2mut), or WT chondrosarcomas. (C,D) RNA seq analysis of BCAT1, BCAT2, BCKDHA, and DBT genes in chondrosarcomas (n = 82) (C) and osteosarcomas (n = 53) (D). *P ≤ 0.05 as compared with IDH1mut chondrosarcomas. **P ≤ 0.05 as compared with IDH2mut chondrosarcomas.

The BCAT1 locus is hypermethylated in chondrosarcomas carrying IDH1 and IDH2 mutations

Guided by the initial observation that the chondrosarcoma biopsies showed much lower BCATc expression than the osteosarcoma biopsies, we analyzed publicly available datasets to explore whether the BCATc gene, BCAT1, is subject to epigenetic silencing in human chondrosarcomas that carry IDH mutations, as previously was demonstrated for glioblastoma [13]. We, therefore, analyzed published DNA methylation data generated using Illumina 450k bead arrays and matched RNA expression data of 82 chondrosarcomas with known IDH mutation status [22]. Visualization of DNA methylation clearly showed IDH mutation-dependent hypermethylation of the CpG Island of one of the two known BCAT1 promoters (Figure 2B and Supplementary Figure S1) [13]. None of the other genes, BCAT2, BCKDHA, or DBT that encode BCATm, BCKDC-E1α, or BCKDC-E2, respectively, showed IDH-mutant dependent hypermethylation of their promoters (data not shown). Surprisingly, this hypermethylation did not seem to significantly suppress BCAT1 RNA expression (Figure 2C) suggesting that in chondrosarcomas, BCAT1 might also be expressed from the first promoter, which is located ∼40 kb upstream of the second promoter and is not hypermethylated (Supplementary Figure S1). On the other hand, all genes (BCAT1, BCAT2, BCKDHA, and DBT) were expressed in osteosarcomas with BCAT1 being particularly abundantly expressed (Figure 2D) [24]. Further RNA seq data analysis [23,25] of each of the studied genes revealed that BCAT1 and DBT are significantly overexpressed in osteosarcomas compared with normal bone tissues (Figure 3A). In contrast, none of the genes was overexpressed in chondrosarcomas when these cancers were compared with normal cartilage tissues (Figure 3B).

BCAT1 and DBT are overexpressed in osteosarcomas.

Figure 3.
BCAT1 and DBT are overexpressed in osteosarcomas.

RNA seq data analysis of BCAT1, BCAT2, BCKDHA, and DBT in normal bone tissue (n = 4) and osteosarcomas (n = 14) is shown in (A), or in normal cartilage (n = 6) and chondrosarcomas (n = 21) is shown in (B). The datasets were analyzed by using the R2: Genomics Analysis and Visualization Platform as described in ‘the Experimental section’. *P ≤ 0.05 as compared with normal bone tissue, **P ≤ 0.05 as compared with normal cartilage tissue.

Figure 3.
BCAT1 and DBT are overexpressed in osteosarcomas.

RNA seq data analysis of BCAT1, BCAT2, BCKDHA, and DBT in normal bone tissue (n = 4) and osteosarcomas (n = 14) is shown in (A), or in normal cartilage (n = 6) and chondrosarcomas (n = 21) is shown in (B). The datasets were analyzed by using the R2: Genomics Analysis and Visualization Platform as described in ‘the Experimental section’. *P ≤ 0.05 as compared with normal bone tissue, **P ≤ 0.05 as compared with normal cartilage tissue.

The osteosarcoma 143B cells demonstrate the highest rate of leucine transamination compared with malignant and non-malignant tumorigenic bone cells

Following the initial protein expression of BCAA enzymes in the human tumor biopsies and the IDH-mutant dependent BCAT1 hypermethylation, we hypothesized that chondrosarcomas, which show low BCATc expression, may preferentially utilize mitochondrial BCAA metabolism, while osteosarcomas may rely on both cytosolic (via BCATc) and mitochondrial BCAA metabolism. To test this hypothesis, we selected two malignant tumorigenic bone sarcoma cell lines, the chondrosarcoma SW1353 cells that carry IDH2 mutation [29] and the osteosarcoma 143B cell line. The two bone sarcoma cell lines were compared with slowly growing benign tumorigenic chondroblastoma PCB0023 cells that showed profound expression of BCATc (Figure 4A). The growth rate characteristics of these cell lines were determined prior to the assays and are shown in Supplementary Figure S2.

Osteosarcoma cells use both BCATc and BCATm to transaminate leucine.

Figure 4.
Osteosarcoma cells use both BCATc and BCATm to transaminate leucine.

Chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were maintained in DMEM media at 37°C and 5% CO2 prior to the assays. (A) Protein expression of BCATc, BCATm, BCKDC-E1α, and BCKDC-E2 was assessed by Western Blotting. ImageJ was used to quantify the relative band intensity of each BCAA enzyme as normalized to β-tubulin. (B, C) Leucine metabolic assay. The DMEM media was replaced by Krebs buffer and cells were fed radioactive leucine (14C-leucine) to determine the rates of leucine transamination (B) and oxidation (C) of each cell line. In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with PCB0023, P ≤ 0.05 as compared with SW1353 cells.

Figure 4.
Osteosarcoma cells use both BCATc and BCATm to transaminate leucine.

Chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were maintained in DMEM media at 37°C and 5% CO2 prior to the assays. (A) Protein expression of BCATc, BCATm, BCKDC-E1α, and BCKDC-E2 was assessed by Western Blotting. ImageJ was used to quantify the relative band intensity of each BCAA enzyme as normalized to β-tubulin. (B, C) Leucine metabolic assay. The DMEM media was replaced by Krebs buffer and cells were fed radioactive leucine (14C-leucine) to determine the rates of leucine transamination (B) and oxidation (C) of each cell line. In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with PCB0023, P ≤ 0.05 as compared with SW1353 cells.

The chondrosarcoma SW1353, but not the osteosarcoma 143B, cells demonstrated a 90% reduction in the protein expression of BCATc when compared with the chondroblastoma PCB0023 cells (Figure 4A). In contrast, both SW1353 and 143B cells had significantly increased protein expression of BCATm by 17.6- and 9.6-folds as compared with the PCB0023 cells (Figure 4A). BCKDC-E1α and BCKDC-E2 were expressed in all cell lines without statistically significant differences in their protein expression levels (Figure 4A). The differential expression of BCATc and BCATm suggested that the bone sarcoma cells may differ in their ability to metabolize BCAAs. As a readout of BCAA metabolism, radioactive leucine (14C-leucine) was added to all cell suspensions and leucine transamination and oxidation were measured. As shown in Figure 4B, leucine transamination was significantly higher by 1.4- and 2.5-folds in the chondrosarcoma SW1353 and osteosarcoma 143B cells as compared with the chondroblastoma PCB0023 cells. Notably, the osteosarcoma 143B cells demonstrated the highest rate of leucine transamination. However, the three cell lines did not show a significant difference in leucine oxidation (Figure 4C). In summary, the high rate of leucine transamination in the osteosarcoma 143B cells correlated with the expression of BCATc and BCATm in these cells, while the rates of leucine oxidation and the expression of BCKDC enzymes remained similar in all cell lines suggesting that non-malignant and malignant bone cells may not differ in their ability to oxidize BCAAs.

Pharmacological inhibition of leucine uptake by NALA, or BCATc by gabapentin, reduces the ability of bone sarcoma cells to metabolize leucine

To assess whether the bone sarcoma cell lines would respond to modulations in leucine availability or metabolism, two pharmacological inhibitors were used: the leucine structural analog NALA, previously shown to reduce leucine availability in lymphoma and immune cells [8,30,31], and the anticonvulsant drug gabapentin, a well described inhibitor of the BCATc enzyme activity [32] (see Figure 1). A treatment with 10 mM gabapentin significantly decreased the rates of leucine transamination in the non-malignant chondroblastoma PCB0023 and the osteosarcoma 143B cells when compared with their corresponding untreated controls. Gabapentin did not significantly impact the same process in the chondrosarcoma SW1353 cells (Figure 5A). A treatment with 25 mM NALA reduced the rates of leucine transamination in all cell lines but the reduction was significant only for the malignant cells (40.8% in the SW1353 and 44.2% in the 143B cells when compared with untreated cells, Figure 5A). The rates of leucine oxidation were significantly reduced by gabapentin in the chondrosarcoma SW1353 cells only, while NALA caused significant reduction in the rates of leucine oxidation in both types of malignant cells (Figure 5B). Taken together, the results demonstrated that limiting leucine uptake by NALA reduced the ability of the malignant cells to transaminate and oxidize leucine. On the other hand, applying gabapentin to inhibit the activity of BCATc was efficient in the osteosarcoma 143B cells but not in the chondrosarcoma SW1353 cells. This was consistent with the observed down-regulation of the BCATc protein expression in these cells.

The leucine structural analog NALA and the BCATc inhibitor gabapentin inhibit leucine metabolism in tumorigenic bone cell lines.

Figure 5.
The leucine structural analog NALA and the BCATc inhibitor gabapentin inhibit leucine metabolism in tumorigenic bone cell lines.

Chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were left untreated or treated with 10 mM gabapentin or 25 mM NALA for 24 h. At the end of each treatment, cells were fed 14C-leucine dissolved in Krebs buffer to determine the rates of leucine transamination (A) and oxidation (B). In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

Figure 5.
The leucine structural analog NALA and the BCATc inhibitor gabapentin inhibit leucine metabolism in tumorigenic bone cell lines.

Chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were left untreated or treated with 10 mM gabapentin or 25 mM NALA for 24 h. At the end of each treatment, cells were fed 14C-leucine dissolved in Krebs buffer to determine the rates of leucine transamination (A) and oxidation (B). In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

Modulations in leucine supply and metabolism by NALA and gabapentin impact the energy charge and cell viability of bone sarcoma cells

Complete oxidation of leucine delivers metabolic intermediates, such as acetyl-CoA, that can be oxidized in the TCA cycle and used as a source of energy [1] (Figure 1). Inhibition of the first two steps in leucine degradation by gabapentin or NALA in the malignant bone sarcoma cells (Figure 5A,B) suggested that the energy charge of these cells might be reduced in response to the applied treatments. The falling energy status of the cells can trigger phosphorylation and subsequent activation of the cellular energy sensor AMPK [33]. According to Figure 6A–C, changes in the phosphorylation state of AMPKα, the ATP levels, and the cell viability of the non-malignant chondroblastoma PCB0023 and the chondrosarcoma SW1353 cells were not detected in response to 5 mM gabapentin when each cell line was compared with their corresponding untreated controls. In contrast, the phosphorylation state of AMPKα was significantly increased by 69% along with a 32% reduction in the total intracellular concentrations of ATP and a 20% decrease in the cell viability of the osteosarcoma 143B cells in response to gabapentin (Figure 6A–C, light gray bars). Interestingly, treatment with 25 mM NALA negatively impacted the energy state and the cell viability of all cell lines with the effect of NALA being most profound in the osteosarcoma 143B cells (a 50% reduction in cell viability as compared with untreated cells) (Figure 6A–C).

Blocking leucine uptake or metabolism impacts the energy status of the tumorigenic bone cell lines.

Figure 6.
Blocking leucine uptake or metabolism impacts the energy status of the tumorigenic bone cell lines.

The chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were either left untreated or treated with 5 mM gabapentin or 25 mM NALA for 24 h. (A) Western Blotting of total and phosphorylated (P) AMPKα. The relative ratio of P-AMPKα to AMPKα was determined by using ImageJ. β-tubulin was used as a loading control. (B) Endogenous ATP concentrations. (C) Cell viability determined by the MTT assay. In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

Figure 6.
Blocking leucine uptake or metabolism impacts the energy status of the tumorigenic bone cell lines.

The chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were either left untreated or treated with 5 mM gabapentin or 25 mM NALA for 24 h. (A) Western Blotting of total and phosphorylated (P) AMPKα. The relative ratio of P-AMPKα to AMPKα was determined by using ImageJ. β-tubulin was used as a loading control. (B) Endogenous ATP concentrations. (C) Cell viability determined by the MTT assay. In all graphs, data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

The differential response of the bone sarcoma cells to gabapentin and NALA in respect to their energy status and cell viability prompted us to compare the activity state of AMPKα in the absence of the pharmacological inhibitors. As shown in Figure 7A, the chondrosarcoma SW1353 cells demonstrated 2.1- and 5.1-folds higher phosphorylation states of AMPKα when compared with chondroblastoma PCB0023 and osteosarcoma 143B cells, respectively. Moreover, there was an inverse correlation between the protein expression of BCATc and the activity state of AMPKα in the SW1353 cells as compared with the PCB0023 cells (compare Figures 4A and 7A). Because one of the downstream effects of AMPK is to inhibit mTORC1 pathway [34], we measured the phosphorylation state of the ribosomal S6 protein (a downstream target of mTORC1 pathway). A significant reduction in the phosphorylation state of S6 was noted in the SW1353 cells when compared with the PCB0023 cells (Figure 7A) These findings suggested that the suppression of the BCATc protein expression may have contributed to the activation of AMPK and the subsequent down-regulation of mTORC1 pathway in the SW1353 cells. To further validate these findings, we used a human osteoblast hFOB1.19 line, grown under two different temperatures, 34°C that induces proliferation and 39°C for 5 days that stimulates cell differentiation. Because the gene expression of BCATc is suppressed in differentiated tissues except for the brain [1,7], we hypothesized that the BCATc protein expression in hFOB1.19 cells would decrease upon differentiation. Indeed, as the hFOB1.19 cells underwent differentiation, the BCATc protein expression became reduced (Figure 7B). The observed 51% reduction in the BCATc protein expression inversely correlated with a 1.5-fold increase in the phosphorylation state of AMPKα in these cells. In opposite, the activity of mTORC1 pathway as judged by the 80% decrease in the phosphorylation state of S6 was inhibited in the hFOB1.19 cells upon differentiation (Figure 7B). These findings were reminiscent of those determined for the SW1353 cells in Figure 7A. Taken together the results validated that the impact of gabapentin on the SW1353 cells, in contrast with 143B cells, was limited by the significantly reduced expression of its target BCATc in the SW1353 cells.

Down-regulation of the BCATc protein correlates with higher AMPKα but lower mTORC1 activities.

Figure 7.
Down-regulation of the BCATc protein correlates with higher AMPKα but lower mTORC1 activities.

(AC) The protein expression of BCATc, AMPKα, or S6, and their phosphorylated states was determined in (A) untreated PCB0023, SW1353, and 143B cells or (B) normal human osteoblast hFOB1.19 cells that were kept in a proliferating stage at 34°C, or differentiated at 39°C for 5 days, or (C) PCB0023, SW1353, and 143B cells left untreated or treated with 5 mM metformin for 24 h. The relative ratio of P-AMPKα to AMPKα and P-S6 to S6 was determined by using ImageJ. The activation state of S6 was used as a measurement of active mTORC1 pathway. β-tubulin was used as a loading control. (D) Cell viability was determined by the MTT assay in cells left untreated or treated with 5 mM metformin. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with PCB0023, P ≤ 0.05 as compared with SW1353, P ≤ 0.05 as compared with proliferating osteoblast hFOB1.19, **P ≤ 0.05 as compared with untreated cells. Representative images are shown.

Figure 7.
Down-regulation of the BCATc protein correlates with higher AMPKα but lower mTORC1 activities.

(AC) The protein expression of BCATc, AMPKα, or S6, and their phosphorylated states was determined in (A) untreated PCB0023, SW1353, and 143B cells or (B) normal human osteoblast hFOB1.19 cells that were kept in a proliferating stage at 34°C, or differentiated at 39°C for 5 days, or (C) PCB0023, SW1353, and 143B cells left untreated or treated with 5 mM metformin for 24 h. The relative ratio of P-AMPKα to AMPKα and P-S6 to S6 was determined by using ImageJ. The activation state of S6 was used as a measurement of active mTORC1 pathway. β-tubulin was used as a loading control. (D) Cell viability was determined by the MTT assay in cells left untreated or treated with 5 mM metformin. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with PCB0023, P ≤ 0.05 as compared with SW1353, P ≤ 0.05 as compared with proliferating osteoblast hFOB1.19, **P ≤ 0.05 as compared with untreated cells. Representative images are shown.

Lastly, we compared the response of the three tumorigenic bone cell lines to metformin, a selective activator of AMPK [35]. A 24 h treatment with 5 mM metformin resulted in activation of AMPKα in all cell lines as seen by the increased phosphorylation state of AMPKα in Figure 7C. However, activated AMPKα had no negative impact on the cell viability or the activity of mTORC1 pathway in PCB0023 or SW1353 cells as compared with their corresponding untreated controls (Figure 7CD). In contrast, metformin not only activated AMPKα in the osteosarcoma 143B cells, but it also caused decreases in the phosphorylation state of S6 and cell viability by ∼39% and 31%, respectively (Figure 7CD, light gray bars). The effects of metformin were reminiscent of the effects of NALA observed in the 143B cell line (compare with Figure 6). The metformin experiment implies that different bone sarcomas vary in how they exploit the energy sensor AMPK, which did not appear essential for the down-regulation of mTORC1 pathway in the PCB0023 and SW1353 cells.

Leucine regulates complex 1 of mTOR signaling in bone sarcoma cells in a glucose-dependent manner

Leucine is one of the most potent activators of mTORC1 signaling and previous investigations by our group, as well as, some of the findings from this study have demonstrated that a limited supply of leucine or disruption in BCAA metabolism may cause down-regulation of mTORC1 [36–38]. The response of mTORC1 pathway to NALA and gabapentin in the non-malignant and malignant bone cells was tested next and compared with the known effect of rapamycin, a selective inhibitor of mTORC1 [37]. The phosphorylation of S6 was significantly reduced in the PCB0023 cells following a 24 h treatment with 5 mM gabapentin, 25 mM NALA, or 100 nM rapamycin, respectively (Figure 8A, 25 mM glucose). Rapamycin demonstrated a strong inhibitory effect on P-S6 in all malignant cell lines and its effect on cell viability resembled that of NALA (compare Figure 8D–F with Figure 6C); however, gabapentin had no effect and NALA only slightly impacted P-S6 (28.4% and 21.3% decreases in P-S6) in the chondrosarcoma SW1353 and the osteosarcoma 143B cells compared with untreated cells, Figure 8B,C, 25 mM glucose). These findings were unexpected in regard of previous studies [30,38]. Because the cell lines were grown in the presence of high glucose (25 mM) in the DMEM medium, which exceeded by several folds the normal blood glucose concentrations, we speculated that the high glucose supply made the malignant cells less dependent on leucine or BCAA metabolism for mTORC1 signaling. To test this, we reduced the concentrations of glucose in the growth medium to 5 and 0 mM and treated the cells with gabapentin or NALA for 24 h. NALA had either no effect or slight stimulatory effect on P-S6 of the chondrosarcoma SW1353 cells in the presence of 5 or 0 mM glucose (Figure 8B). In contrast, the inhibitory effect of NALA on P-S6 became stronger as the glucose concentrations decreased leading to 49% (5 mM glucose) and 76.3% (0 mM glucose) reductions in P-S6 in the osteosarcoma 143B cells (Figure 8C). Similarly, the osteosarcoma cells responded to gabapentin with a significant 32.5% reduction in P-S6 when grown for 24 h in DMEM medium with 0 mM glucose (Figure 8C). In summary, disrupted leucine supply and BCAA metabolism in bone sarcoma cells had an impact on the activation of mTORC1 pathway (as measured by changes in P-S6), but was dependent on: (1) the concentration of glucose in the DMEM medium and (2) the heterogeneity of the bone sarcomas. The prior observation suggested that the reduction or complete removal of glucose from the growth medium may have impacted differently the ability of the cells to switch to other carbon sources.

The impact of leucine on mTORC1 signaling in bone sarcoma cells is dependent on the presence of glucose in the growth medium.

Figure 8.
The impact of leucine on mTORC1 signaling in bone sarcoma cells is dependent on the presence of glucose in the growth medium.

(A–F) Chondroblastoma PCB0023 (A,D), chondrosarcoma SW1353 (B,E), and osteosarcoma 143B (C,F) cells were left untreated or treated with 5 mM gabapentin, 25 mM NALA, or 100 nM rapamycin in the presence of 25, 5, or 0 mM glucose (Glc) in the growth medium for 24 h. In (A–C), the ratio between phosphorylated and total ribosomal S6 protein was used as a measurement of active mTORC1 pathway. ImageJ was used to quantify the resulting protein bands. β-tubulin was used as a loading control. In (D–F), cell viability of the three tumorigenic bone cell lines in the presence of 100 nM rapamycin is shown. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells. Representative images are shown.

Figure 8.
The impact of leucine on mTORC1 signaling in bone sarcoma cells is dependent on the presence of glucose in the growth medium.

(A–F) Chondroblastoma PCB0023 (A,D), chondrosarcoma SW1353 (B,E), and osteosarcoma 143B (C,F) cells were left untreated or treated with 5 mM gabapentin, 25 mM NALA, or 100 nM rapamycin in the presence of 25, 5, or 0 mM glucose (Glc) in the growth medium for 24 h. In (A–C), the ratio between phosphorylated and total ribosomal S6 protein was used as a measurement of active mTORC1 pathway. ImageJ was used to quantify the resulting protein bands. β-tubulin was used as a loading control. In (D–F), cell viability of the three tumorigenic bone cell lines in the presence of 100 nM rapamycin is shown. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells. Representative images are shown.

The ability of bone sarcoma cells to release lactate depends on glucose and is differently impacted by disruption of leucine uptake and BCAA metabolism

Intermediates of BCAA metabolism, such as glutamate, acetyl-CoA, and succinyl-CoA, can feed into TCA cycle or the gluconeogenic pathway, as well as provide carbon for lactate production [39]. Since cancer cells are highly glycolytic and release substantial amounts of lactate [40], we tested whether the pharmacological disruption of leucine uptake and BCAA metabolism by NALA and gabapentin would impact the ability of the bone sarcoma cell lines to produce lactate.

In contrast with 2-DG, a selective inhibitor of glycolysis, none of the treatments (gabapentin, NALA, rapamycin) significantly affected the release of lactate from any of the cell lines when grown in high (25 mM) glucose DMEM medium for 24 h (Figure 9A–C, 25 mM glucose). When the chondroblastoma PCB0023 or the osteosarcoma 143B cells were supplied with 5 mM glucose in the growth medium, gabapentin and NALA significantly reduced the lactate secretion from these cells as compared with untreated cells (Figure 9A,C, 5 mM glucose). Interestingly, the chondrosarcoma SW1353 cells responded to NALA with a significant 89.4% increase in lactate release under low (5 mM) glucose conditions, while gabapentin had no significant impact on lactate production; the latter agreed with the low expression level of BCATc in this cell line (Figure 9B).

Lactate release from bone sarcoma cells is altered in response to gabapentin and NALA in a glucose-dependent manner.

Figure 9.
Lactate release from bone sarcoma cells is altered in response to gabapentin and NALA in a glucose-dependent manner.

(A–C) Chondroblastoma PCB0023 (A), chondrosarcoma SW1353 (B), and osteosarcoma 143B (C) cells were left untreated or treated with 5 mM gabapentin, 25 mM NALA, 100 nM rapamycin, or 25 mM 2-DG in the presence of 25 mM or 5 mM glucose (Glc) for 24 h followed by measurement of lactate release in the growth medium. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

Figure 9.
Lactate release from bone sarcoma cells is altered in response to gabapentin and NALA in a glucose-dependent manner.

(A–C) Chondroblastoma PCB0023 (A), chondrosarcoma SW1353 (B), and osteosarcoma 143B (C) cells were left untreated or treated with 5 mM gabapentin, 25 mM NALA, 100 nM rapamycin, or 25 mM 2-DG in the presence of 25 mM or 5 mM glucose (Glc) for 24 h followed by measurement of lactate release in the growth medium. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells.

Pharmacological inhibition by NALA and gabapentin is most impactful on the growth of the osteosarcoma 143B cells

The impact that NALA or gabapentin exerted on the metabolism and cell viability of the studied bone sarcoma cells implied that the apoptotic program and cell proliferation might be affected by the applied pharmacological treatments. Analysis of the cell death and apoptotic markers, such as BCL2 and BAX, led to little or no differences between treatments and across the cell lines (data not shown). To test the impact on cell growth, a concentration gradient of NALA and gabapentin was applied, and cell proliferation was measured. As shown in Figure 10A,B, low concentrations of NALA (12.5 mM) had a stimulatory effect on the proliferation of the chondroblastoma PCB0023 and chondrosarcoma SW1353 cells; however, concentrations of NALA ≥ 25 mM had an inhibitory effect by the 48-h time point. Gabapentin showed no inhibitory effect on the growth of these cells (Figure 10A,B). In contrast, the proliferation of the osteosarcoma 143B cells was inhibited by both NALA and gabapentin in a concentration-dependent manner within the first 24 h (Figure 10C). In summary, the effects of gabapentin and NALA were most impactful on the growth of the osteosarcoma 143B cells.

The proliferation of osteosarcoma 143B cells, but not chondroblastoma PCB0023 or chondrosarcoma SW1353 cells, is inhibited by NALA or gabapentin in a concentration-dependent manner.

Figure 10.
The proliferation of osteosarcoma 143B cells, but not chondroblastoma PCB0023 or chondrosarcoma SW1353 cells, is inhibited by NALA or gabapentin in a concentration-dependent manner.

(A–C) Chondroblastoma PCB0023, chondrosarcoma SW1353, or osteosarcoma 143B cells were treated with 0–50 mM NALA or 0–20 mM gabapentin for 0, 24, and 48 h followed by cell counting by using the trypan blue exclusion method. In all graphs, data represent mean ± SEM, n = 3 independent cultures. *P ≤ 0.05 as compared with untreated cells, P ≤ 0.05 as compared with 12.5 mM NALA-treated cells.

Figure 10.
The proliferation of osteosarcoma 143B cells, but not chondroblastoma PCB0023 or chondrosarcoma SW1353 cells, is inhibited by NALA or gabapentin in a concentration-dependent manner.

(A–C) Chondroblastoma PCB0023, chondrosarcoma SW1353, or osteosarcoma 143B cells were treated with 0–50 mM NALA or 0–20 mM gabapentin for 0, 24, and 48 h followed by cell counting by using the trypan blue exclusion method. In all graphs, data represent mean ± SEM, n = 3 independent cultures. *P ≤ 0.05 as compared with untreated cells, P ≤ 0.05 as compared with 12.5 mM NALA-treated cells.

Disruption in leucine supply and metabolism impacts the phosphorylation of retinoblastoma (Rb) and p27Kip1 degradation and inhibits the cell cycle progression of the bone sarcoma cells

Low expression levels of the cyclin dependent kinase inhibitor p27Kip1 along with high phosphorylation of Rb are essential for the transition of cells from G1 to S phase and are thus important checkpoint regulators of the ability of the cells to continuously grow and proliferate [30]. To explore whether these molecular factors were impacted by gabapentin or NALA, the phosphorylation state and the protein expression levels of Rb and p27Kip1 were determined in all cells as indicated in Figure 11. Gabapentin had a significant impact on Rb and p27Kip1 expression in the osteosarcoma 143B cells only; these cells responded to a 24 h treatment with 5 mM gabapentin with a 70.7% reduction in the phosphorylation of Rb and a 1.8-fold increase in the protein expression of p27Kip1 (Figure 11A,B, light gray bars).

Alterations in Rb phosphorylation and p27Kip1 protein expression are most profound in osteosarcoma 143B cells treated with gabapentin and NALA.

Figure 11.
Alterations in Rb phosphorylation and p27Kip1 protein expression are most profound in osteosarcoma 143B cells treated with gabapentin and NALA.

(A,B) Western Blotting of total and phosphorylated (P) Rb (A) and total p27Kip1 (B) was determined after a 24 h treatment with 5 mM gabapentin or 25 mM NALA of chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells. The ratio between phosphorylated and total Rb is shown in (A), while β-tubulin was used to normalize the protein expression levels of p27Kip1 in B. ImageJ was used to quantify the protein bands and the results are expressed as the relative ratio between the compared bands. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells. Representative protein images are shown.

Figure 11.
Alterations in Rb phosphorylation and p27Kip1 protein expression are most profound in osteosarcoma 143B cells treated with gabapentin and NALA.

(A,B) Western Blotting of total and phosphorylated (P) Rb (A) and total p27Kip1 (B) was determined after a 24 h treatment with 5 mM gabapentin or 25 mM NALA of chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells. The ratio between phosphorylated and total Rb is shown in (A), while β-tubulin was used to normalize the protein expression levels of p27Kip1 in B. ImageJ was used to quantify the protein bands and the results are expressed as the relative ratio between the compared bands. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells. Representative protein images are shown.

All cell lines showed significant reduction in the phosphorylation of Rb in response to a 24 h treatment with 25 mM NALA (Figure 11A), however, NALA had a significant impact on the protein expression of p27Kip1 in the osteosarcoma 143B cells only (a 1.24-fold increase, Figure 11B, light gray bars).

To test whether the pharmacological treatments with gabapentin or NALA would directly impact the progression of the bone cells through the cell cycle, untreated and treated cells were stained with a PI-RNase A solution and the distribution of the DNA content among the different cell cycle phases was examined by FACS. As shown in Figure 12A, ∼70% of the untreated chondroblastoma PCB0023 cells were arrested in the G1 phase of the cell cycle. Treatment with gabapentin did not significantly impact the cell cycle of this cell line, while a treatment with NALA led to a small but significant increase in the number of cells in the G1 phase while reducing the number of cells in the S phase (Figure 12A). In contrast with the PCB0023 cells, most of the untreated chondrosarcoma SW1353 cells were in the S phase (∼65%) (Figure 12B). None of the SW1353 cells were found in G2/M phase, however, ∼30% of the cells were found to have DNA content >G2 suggesting that this small population of cells might have been a part of polyploid fraction [41]. Treatment with gabapentin did not have significant impact on the cell distribution among the cell cycle phases, while NALA arrested the SW1353 cells in the G1 phase (48% of the cells) thus reducing the number of cells in the S phase to 42% (Figure 12B).

The malignant SW1353 and 143B cells responded to NALA with delayed progression through the cell cycle.

Figure 12.
The malignant SW1353 and 143B cells responded to NALA with delayed progression through the cell cycle.

The chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were either left untreated or treated with 5 mM gabapentin or 25 mM NALA for 24 h followed by ethanol fixation and staining with PI and RNAase to determine the DNA content in each phase of the cell cycle. Averaged cell distribution among G1, S, G2/M phases of the cell cycle and representative histograms of chondroblastoma PBC0023 (A,D), chondrosarcoma SW1353 (B,E), and osteosarcoma 143B (C,F) cells in response to the indicated treatments are shown. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells in G1 phase. P ≤ 0.05 as compared with untreated cells in S phase. P ≤ 0.05 as compared with untreated cells in G2/M phase.

Figure 12.
The malignant SW1353 and 143B cells responded to NALA with delayed progression through the cell cycle.

The chondroblastoma PCB0023, chondrosarcoma SW1353, and osteosarcoma 143B cells were either left untreated or treated with 5 mM gabapentin or 25 mM NALA for 24 h followed by ethanol fixation and staining with PI and RNAase to determine the DNA content in each phase of the cell cycle. Averaged cell distribution among G1, S, G2/M phases of the cell cycle and representative histograms of chondroblastoma PBC0023 (A,D), chondrosarcoma SW1353 (B,E), and osteosarcoma 143B (C,F) cells in response to the indicated treatments are shown. Data represent mean ± SEM, n = 3 independent experiments. *P ≤ 0.05 as compared with untreated cells in G1 phase. P ≤ 0.05 as compared with untreated cells in S phase. P ≤ 0.05 as compared with untreated cells in G2/M phase.

The distribution of the untreated osteosarcoma 143B cells among G1, S, and G2/M phases was 30%, 47%, and 20% of cells, respectively (Figure 12C). Treatment with gabapentin had mild but significant effect on the 143B cells in the G1 and S phases arresting more cells in the G1 phase and reducing the number of cells in the S phase. The most profound and statistically significant effect, however, was that of NALA. Upon NALA treatment ∼48% of the cells remained arrested in the G1 phase leaving ∼39% and 11% of the cells in S and G2/M phases as compared with the untreated cells, respectively (Figure 12C). Taken together, these results indicate that disruption in leucine uptake by NALA was most impactful on the cell cycle progression of the malignant tumorigenic (SW1353 and 143B) cells where one mechanism of cell cycle suppression may involve stabilization of the checkpoint inhibitors Rb and p27Kip1 in response to NALA.

Discussion

Despite the development of new therapeutic approaches, bone sarcomas continue to be difficult-to-treat cancers that lead to morbidity and mortality in children and adolescents [20,42]. This is further complicated by the high genetic diversity of these cancers [43]. Thus, not surprisingly, the different types of tumorigenic bone cells investigated in this study varied in their preference for leucine and BCAA metabolism.

The central role of BCAA metabolism in tumor progression is largely linked to the overexpression of the gene encoding BCATc in cancer cells. Indeed, multiple reports have suggested important roles for BCATc in cancers, such as gliomas, acute myeloid leukemia, ovarian cancer, and non-small lung carcinoma (reviewed in [1]). While BCATc has received significant scientific interest in the field of oncology, the role of this protein, and its mitochondrial isoform BCATm, in bone sarcomas has not been explored. In this report, we found a profound difference in the expression pattern of BCATc and BCATm in different bone sarcomas. While human osteosarcoma biopsies expressed both BCATc and BCATm, chondrosarcoma tissues had barely detectable BCATc protein. Consistent with this finding, the selected chondrosarcoma SW1353 cell line, that showed over a 90% reduction in the protein expression of BCATc, was largely unresponsive to the effects of gabapentin, a well-established competitive inhibitor of the enzyme activity of BCATc [44]. In an attempt to identify a plausible mechanism for the suppression of BCATc expression in chondrosarcomas, we found that the second promoter that drives the expression of BCAT1 was subjected to epigenetic hypermethylation in human chondrosarcomas that carry IDH1 and IDH2 mutations. This finding was consistent with several other studies that reported epigenetic silencing of BCATc via DNA methylation of the BCAT1 promoter in glioblastomas or AML [13,45–47]. About 90% of glioblastomas are wild type IDH1wt [13]. BCATc is overexpressed in IDH1wt glioblastomas, but is silent in glioblastomas with IDH1mut, demonstrating a connection between the IDH1 mutation and BCATc. Similar to glioblastomas, chondrosarcomas frequently harbor gene mutations in IDH1 and are used to distinguish chondrosarcomas from chondroblastic osteosarcomas [48]. Although our findings suggest that the expression of BCAT1 may be altered by the IDH1 mutation status in chondrosarcomas, we did not find statistically significant reduction in the BCAT1 RNA expression in chondrosarcomas of IDHmut types when compared with IDHwt types in the public datasets. While this might be partially explained by the activity of the first promoter of BCAT1, which appeared unimpacted by DNA hypermethylation in our analysis, future studies are needed to provide evidence regarding the precise mechanism/s of gene regulation of BCAT1 in IDHmut and IDHwt chondrosarcomas.

Although the chondrosarcoma SW1353 cells did not respond to gabapentin, these cells took advantage of the mitochondrial BCAA catabolism as a source of energy, which was evident from the reduction in leucine transamination followed by decreased intracellular ATP concentrations and activation of the energy sensor AMPK in response to NALA. Likewise, the cell cycle progression of the chondrosarcoma SW1353 cells was partially arrested when leucine uptake was interrupted by NALA and the likely mechanism might have been the stabilization of the cell cycle inhibitor Rb in response to NALA. Similar response was recorded for the osteosarcoma 143B cells when treated with either NALA or gabapentin suggesting that leucine and BCAA metabolism may play important roles during the cell cycle progression of tumorigenic bone cells. This was consistent with another report in ERα-negative breast cancer, where BCATc controlled the cell cycle progression at the level of another cell cycle inhibitor (p27Kip1) [49].

In contrast with the SW1353 cells, the osteosarcoma 143B cells demonstrated high demand not only for mitochondrial but also for cytosolic catabolism of BCAAs, which when disrupted, negatively impacted the 143B cell energy status, downstream metabolic pathways, as well as cell viability and growth. These results were consistent with the role of BCAAs as nutrient signals supporting protein synthesis, but also as metabolic fuels and nitrogen donors for the synthesis of various intermediates [7,50]. The best described role of BCAAs, particularly leucine, is in muscle cells where leucine transamination is a source of nitrogen for the synthesis of glutamine and alanine, while complete leucine oxidation delivers acetyl-CoA for the TCA cycle and thus provides energy for the cells [50]. Treating myoblast cells with physiological concentrations of leucine was shown to increase the ATP content, reduce the AMP/ATP ratio, and subsequently suppress the energy sensor AMPK demonstrating that leucine is used to generate energy in the muscle. In the current study, both the chondrosarcoma SW1353 and the osteosarcoma 143B cells responded to NALA with activation of AMPK and reduction in the ATP content. Because NALA can reduce the leucine supply of growing cells [30,31], these findings suggested that leucine availability for metabolism was limited and ultimately led to lower energy status. In another study, NALA-treated or leucine-depleted mouse lymphoma cells also had highly active AMPK [38]. Alternatively, pancreatic tumors from mice on a leucine-supplemented diet demonstrated lower levels of phosphorylated AMPK indicative of higher energy status [51].

However, limiting leucine availability for metabolism was not the only factor that may have impacted the energy charge of the chondrosarcoma SW1353 cells as we observed an inverse correlation between the expression of BCATc and the activity state of AMPK in these cells. This implied that a loss of expression of one of the BCAT isoenzymes might be sufficient to perturb the energy charge of the SW1353 cells. Similar results were obtained by our research team in another study where the global deletion of BCATm correlated with an increased activity state of AMPK in lymphoma tumors [38].

Cancer cells rely on constitutively active mTORC1 signaling and high glycolytic flux to support increased biosynthetic demands [40,52]. Leucine is one of the most potent activators of mTORC1, but it is also an alternative fuel source for cancer cells [50]. As discussed above, limited leucine availability may result in activation of AMPK, which in turn may inhibit the mTORC1 pathway [34,53]. Comparison of the three tumorigenic bone cell lines provided unique opportunity to explore their metabolic signature in respect to mTORC1 and AMPK and how modulation in BCAA metabolism may impact those metabolic regulators. Not surprisingly, the genetic heterogeneity of the studied bone cells revealed complicated metabolic diversity in their response to NALA and gabapentin (Figure 13). The benign tumorigenic chondroblastoma PCB0023 cells responded with significant down-regulation of mTORC1 pathway to disruption in leucine uptake or BCAA metabolism that was independent from activation of AMPK and had only mild impact on the cell viability. In contrast, the activation of mTORC1 in the osteosarcoma 143B cells was dependent on both the activity state of AMPK and the concentration of glucose in the growth medium. Applying metformin to activate AMPK, led to downstream inhibition of mTORC1 and significantly reduced the cell viability of the 143B cells. Similar impact was observed upon NALA and gabapentin treatments. However, under high glucose concentrations (25 mM), treatment with NALA or gabapentin only mildly inhibited mTORC1 pathway in the 143B cells. Because the cells were grown in media with high glucose, we speculated that they may have preferably used glucose to support active glycolysis and to up-regulate mTORC1 signaling [54,55]. Indeed, when we reduced the glucose concentrations to physiological (5 mM), both mTORC1 and lactate secretion were significantly suppressed in the osteosarcoma 143B cells in the presence of NALA or gabapentin. This observation is consistent with reports stating that T cells can become overly dependent on glucose when grown in medium with 10–35 mM glucose [56].

Summary of the effects of NALA and gabapentin on leucine metabolism, mTORC1 signaling, AMPK, and the cell growth of the three tumorigenic bone cell lines.

Figure 13.
Summary of the effects of NALA and gabapentin on leucine metabolism, mTORC1 signaling, AMPK, and the cell growth of the three tumorigenic bone cell lines.

The mild effect refers to ≤20% impact, strong effect refers to ≥50% impact.

Figure 13.
Summary of the effects of NALA and gabapentin on leucine metabolism, mTORC1 signaling, AMPK, and the cell growth of the three tumorigenic bone cell lines.

The mild effect refers to ≤20% impact, strong effect refers to ≥50% impact.

In the chondrosarcoma SW1353 cells, both NALA and metformin caused activation of AMPK, however, activated AMPK did not appear essential for the down-regulation of mTORC1 pathway since a treatment with metformin did not affect mTORC1 pathway in these cells. In contrast, NALA caused moderate down-regulation of mTORC1 pathway under high glucose conditions (25 mM glucose) but when the concentration of glucose was reduced to 5 or 0 mM, the effect of NALA was reversed leading to activation of mTORC1 pathway and an increase in lactate release. While the mechanism of this finding is currently unknown, a future study may seek to address whether NALA has the potential to impact the bidirectional glutamine-leucine transport in the SW1353 cells in a glucose-dependent manner. The glutamine transporter, Slc3a2, uses intracellular glutamine as an efflux substrate to support the uptake of extracellular leucine by the Slc7a5 transporter [57]. Because glutamine is the second most important non-glycolytic metabolic fuel for cancer cells and an activator of mTORC1 pathway [58], reducing the concentrations of glucose may lead to higher demands on glutamine. This effect, combined with a blockage of leucine entry by NALA, may potentially contribute to the activation of mTORC1 by glutamine.

While the chondrosarcoma SW1353 cells failed to respond to gabapentin, this inhibitor led to a decrease in mTORC1 activity in both the PCB0023 and the 143B cells as discussed above. This finding was unexpected because gabapentin can inhibit the BCATc-dependent leucine transamination and potentially allow more leucine for mTORC1 signaling. Therefore, we speculated that gabapentin either competes with leucine for entry into cells or intermediates of leucine degradation activate mTORC1 signaling. Evidence from the literature supports both scenarios. Gabapentin can compete with leucine for entry into the cells [59] and the leucine metabolite acetyl-coenzyme A positively regulates mTORC1 activity by EP300-mediated acetylation of Raptor [60]. Use of mTORC1 inhibitors, such as rapamycin and everolimus, has been shown to decrease bone sarcoma tumor progression in rats [61]. In our study, the impact of rapamycin on the viability of the studied cells was similar to that of NALA where the most impactful reduction in cell viability, mTORC1 activity, and lactate secretion was recorded for the 143B cells. However, the use of rapamycin is associated with significant toxicity and large inter-individual variations in bioavailability and clearance [62]. Our findings that leucine and its metabolism regulate mTORC1 signaling in bone sarcomas may potentially offer less invasive metabolic interventions for future therapeutic applications.

In summary, this study revealed that human chondrosarcomas and osteosarcomas vary in their expression of the BCATc protein, which was found suppressed in the human chondrosarcoma biopsies and the SW1353 cell line selected in this study. One possible mechanism of BCATc down-regulation in chondrosarcomas represents the hypermethylation of the second BCAT1 promoter in chondrosarcomas that carry IDH1 and IDH2 mutations. This finding reveals a future potential to target BCAT1 for prognostic purposes in human chondrosarcomas.

Guided by the finding that the 143B cells required both cytosolic and mitochondrial BCAA metabolism, we concluded that disruption in the cytosolic BCAA transamination (by gabapentin) or leucine uptake (by NALA) had the most profound effect on the osteosarcoma 143B cells. As far as the 143B cell line represents an aggressive form of osteosarcoma [63], the findings from this study may suggest that modulating leucine uptake or the expression of leucine degrading enzymes in aggressive forms of osteosarcoma may serve as alternative metabolic approaches to suppress the growth of these cancers. In fact, the pubic datasets revealed that BCAT1 was significantly overexpressed in human osteosarcomas compared with normal bone tissues. This finding further signifies that BCATc may qualify as a potential target for therapeutic approaches. However, future studies aiming at targeting BCAT1 in normal bone tissues and osteosarcomas are needed to fully explore the role of BCATc in this type of cancer and the potential of BCATc for clinical applications.

Competing Interests

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

Funding

This work was funded by the Des Moines University, USA (IOER-112-3705 to EA) and the Iowa Science Foundation (ISF 17-09 to EA).

Author Contribution

E.A.A. and S.B.M. designed and directed the experiments, analyzed data, and wrote the manuscript. W.S.R., N.A.F., and A.A.M. conducted the Western Blotting and MTT analysis, A.J.S. conducted the ATP assay with E.A.A. M.P.B. conducted the cell proliferation analysis and assisted with the rest of the experiments. D.L.N. and A.A.M. conducted the metformin experiments, S.J.S. and E.A.A. conducted the cell cycle experiments, while B.R. and E.A.A. performed the DNA methylation and RNA expression data analysis of chondrosarcomas and osteosarcomas from the public dataset. All authors participated in data interpretation, reviewed, and approved the manuscript.

Acknowledgements

We thank Dr. Charles Keller (Oregon Health and Science University) for generously providing tumor biopsies from bone sarcoma patients and Dr. Susan Hutson (Virginia Tech) for assistance with the tumor biopsy analysis.

Abbreviations

     
  • 2-DG

    2-deoxy-d-glucose

  •  
  • 4E-BP1

    eukaryotic translation initiation factor 4E binding protein 1

  •  
  • AMPK

    AMP-activated protein kinase

  •  
  • BCAAs

    branched-chain amino acids

  •  
  • BCATm

    mitochondrial branched-chain aminotransferase

  •  
  • BCKAs

    branched-chain keto acids

  •  
  • BCKDC

    branched-chain α-keto acid dehydrogenase enzyme complex

  •  
  • Bcl2

    B-cell lymphoma

  •  
  • Bcr-Abl

    breakpoint cluster region-abelson hybrid gene

  •  
  • E1α

    α-subunit of branched-chain α-keto acid dehydrogenase

  •  
  • E2

    dihydrolipoamide acyltransferase

  •  
  • Glu

    glutamate

  •  
  • IDH1 and 2

    isocitrate dehydrogenase enzyme isoforms 1 and 2

  •  
  • KIC

    α-ketoisocaproate

  •  
  • mTOR

    mammalian target of rapamycin

  •  
  • mTORC1

    complex 1 of mTOR pathway

  •  
  • MTT

    methylthiazoletetrazolium

  •  
  • NALA

    N-acetyl-leucine amide

  •  
  • p27Kip1

    protein27, a cyclic dependent kinase inhibitor

  •  
  • Rb

    retinoblastoma

  •  
  • S6

    ribosomal protein S6

  •  
  • S6K

    70 kDa ribosomal protein S6 kinase

  •  
  • SDS

    Shwachman–Diamond syndrome

  •  
  • Src

    sarcoma virus type protein kinase

  •  
  • TCA cycle

    tricarboxylic acid cycle

  •  
  • α-KG

    α-ketoglutarate

References

References
1
Ananieva
,
E.A.
and
Wilkinson
,
A.C.
(
2018
)
Branched-chain amino acid metabolism in cancer
.
Curr. Opin. Clin. Nutr. Metab. Care
21
,
64
70
2
Baracos
,
V.E.
and
Mackenzie
,
M.L.
(
2006
)
Investigations of branched-chain amino acids and their metabolites in animal models of cancer
.
J. Nutr.
136
,
237S
242S
3
Neinast
,
M.
,
Murashige
,
D.
and
Arany
,
Z.
(
2019
)
Branched chain amino acids
.
Annu. Rev. Physiol.
81
,
139
164
4
Nie
,
C.
,
He
,
T.
,
Zhang
,
W.
,
Zhang
,
G.
and
Ma
,
X.
(
2018
)
Branched chain amino acids: beyond nutrition metabolism
.
Int. J. Mol. Sci.
19
,
954
5
Hutson
,
S.M.
,
Fenstermacher
,
D.
and
Mahar
,
C.
(
1988
)
Role of mitochondrial transamination in branched chain amino acid metabolism
.
J. Biol. Chem.
263
,
3618
3625
PMID:
[PubMed]
6
Ichihara
,
A.
and
Koyama
,
E.
(
1966
)
Transaminase of branched chain amino acids. I. Branched chain amino acids-alpha-ketoglutarate transaminase
.
J. Biochem.
59
,
160
169
7
Sweatt
,
A.J.
,
Wood
,
M.
,
Suryawan
,
A.
,
Wallin
,
R.
,
Willingham
,
M.C.
and
Hutson
,
S.M.
(
2004
)
Branched-chain amino acid catabolism: unique segregation of pathway enzymes in organ systems and peripheral nerves
.
Am. J. Physiol. Endocrinol. Metab.
286
,
E64
E76
8
Ananieva
,
E.A.
,
Patel
,
C.H.
,
Drake
,
C.H.
,
Powell
,
J.D.
and
Hutson
,
S.M.
(
2014
)
Cytosolic branched chain aminotransferase (BCATc) regulates mTORC1 signaling and glycolytic metabolism in CD4+ T cells
.
J. Biol. Chem.
289
,
18793
18804
9
Wang
,
Z.Q.
,
Faddaoui
,
A.
,
Bachvarova
,
M.
,
Plante
,
M.
,
Gregoire
,
J.
,
Renaud
,
M.C.
et al (
2015
)
BCAT1 expression associates with ovarian cancer progression: possible implications in altered disease metabolism
.
Oncotarget
6
,
31522
31543
10
Zhang
,
L.
and
Han
,
J.
(
2017
)
Branched-chain amino acid transaminase 1 (BCAT1) promotes the growth of breast cancer cells through improving mTOR-mediated mitochondrial biogenesis and function
.
Bioch. Biophys. Res. Commun.
486
,
224
231
11
Wang
,
H.G.
,
Xie
,
R.
,
Shen
,
P.
,
Huang
,
X.D.
,
Ji
,
G.Z.
and
Yang
,
X.Z.
(
2016
)
BCAT1 expression in hepatocellular carcinoma
.
Clin. Res. Hepatol. Gastroenterol.
40
,
e55
e56
12
Mayers
,
J.R.
,
Torrence
,
M.E.
,
Danai
,
L.V.
,
Papagiannakopoulos
,
T.
,
Davidson
,
S.M.
,
Bauer
,
M.R.
et al (
2016
)
Tissue of origin dictates branched-chain amino acid metabolism in mutant Kras-driven cancers
.
Science
353
,
1161
1165
13
Tonjes
,
M.
,
Barbus
,
S.
,
Park
,
Y.J.
,
Wang
,
W.
,
Schlotter
,
M.
,
Lindroth
,
A.M.
et al (
2013
)
BCAT1 promotes cell proliferation through amino acid catabolism in gliomas carrying wild-type IDH1
.
Nat. Med.
19
,
901
908
14
Siegel
,
R.L.
,
Miller
,
K.D.
and
Jemal
,
A.
(
2018
)
Cancer statistics
.
CA Cancer J. Clin.
68
,
7
30
15
Onishi
,
A.C.
,
Hincker
,
A.M.
and
Lee
,
F.Y.
(
2011
)
Surmounting chemotherapy and radioresistance in chondrosarcoma: molecular mechanisms and therapeutic targets
.
Sarcoma
2011
,
381564
16
David
,
E.
,
Blanchard
,
F.
,
Heymann
,
M.F.
,
De Pinieux
,
G.
,
Gouin
,
F.
,
Redini
,
F.
et al (
2011
)
The bone niche of chondrosarcoma: a sanctuary for drug resistance, tumour growth and also a source of new therapeutic targets
.
Sarcoma
2011
,
932451
17
He
,
H.
,
Ni
,
J.
and
Huang
,
J.
(
2014
)
Molecular mechanisms of chemoresistance in osteosarcoma (Review)
.
Oncol. Lett.
7
,
1352
1362
18
Hattinger
,
C.M.
,
Patrizio
,
M.P.
,
Luppi
,
S.
,
Magagnoli
,
F.
,
Picci
,
P.
and
Serra
,
M.
(
2019
)
Current understanding of pharmacogenetic implications of DNA damaging drugs used in osteosarcoma treatment
.
Expert Opin. Drug Metab. Toxicol.
15
,
299
311
19
Lipshultz
,
S.E.
,
Karnik
,
R.
,
Sambatakos
,
P.
,
Franco
,
V.I.
,
Ross
,
S.W.
and
Miller
,
T.L.
(
2014
)
Anthracycline-related cardiotoxicity in childhood cancer survivors
.
Curr. Opin. Cardiol.
29
,
103
112
20
Thanindratarn
,
P.
,
Dean
,
D.C.
,
Nelson
,
S.D.
,
Hornicek
,
F.J.
and
Duan
,
Z.
(
2019
)
Advances in immune checkpoint inhibitors for bone sarcoma therapy
.
J. Bone Oncol.
15
,
100221
21
Sweatt
,
A.J.
,
Garcia-Espinosa
,
M.A.
,
Wallin
,
R.
and
Hutson
,
S.M.
(
2004
)
Branched-chain amino acids and neurotransmitter metabolism: expression of cytosolic branched-chain aminotransferase (BCATc) in the cerebellum and hippocampus
.
J. Comp. Neurol.
477
,
360
370
22
Nicolle
,
R.
,
Ayadi
,
M.
,
Gomez-Brouchet
,
A.
,
Armenoult
,
L.
,
Banneau
,
G.
,
Elarouci
,
N.
et al (
2019
)
Integrated molecular characterization of chondrosarcoma reveals critical determinants of disease progression
.
Nat. Commun.
10
,
4622
23
Swanson
,
J.M.
,
Wood
,
G.C.
,
Xu
,
L.
,
Tang
,
L.E.
,
Meibohm
,
B.
,
Homayouni
,
R.
et al (
2012
)
Developing a gene expression model for predicting ventilator-associated pneumonia in trauma patients: a pilot study
.
PLoS One
7
,
e42065
24
Buddingh
,
E.P.
,
Kuijjer
,
M.L.
,
Duim
,
R.A.
,
Burger
,
H.
,
Agelopoulos
,
K.
,
Myklebost
,
O.
et al (
2011
)
Tumor-infiltrating macrophages are associated with metastasis suppression in high-grade osteosarcoma: a rationale for treatment with macrophage activating agents
.
Clin. Cancer Res.
17
,
2110
2119
25
Jones
,
K.B.
,
Salah
,
Z.
,
Del Mare
,
S.
,
Galasso
,
M.
,
Gaudio
,
E.
,
Nuovo
,
G.J.
et al (
2012
)
miRNA signatures associate with pathogenesis and progression of osteosarcoma
.
Cancer Res.
72
,
1865
1877
26
Strober
,
W.
(
2001
)
Trypan blue exclusion test of cell viability
.
Curr. Protoc. Immunol.
Appendix 3
,
Appendix 3B
27
Schneider
,
C.A.
,
Rasband
,
W.S.
and
Eliceiri
,
K.W.
(
2012
)
NIH image to ImageJ: 25 years of image analysis
.
Nat. Methods
9
,
671
675
28
Yang
,
N.C.
,
Ho
,
W.M.
,
Chen
,
Y.H.
and
Hu
,
M.L.
(
2002
)
A convenient one-step extraction of cellular ATP using boiling water for the luciferin-luciferase assay of ATP
.
Anal. Biochem.
306
,
323
327
29
van Oosterwijk
,
J.G.
,
de Jong
,
D.
,
van Ruler
,
M.A.
,
Hogendoorn
,
P.C.
,
Dijkstra
,
P.D.
,
van Rijswijk
,
C.S.
et al (
2012
)
Three new chondrosarcoma cell lines: one grade III conventional central chondrosarcoma and two dedifferentiated chondrosarcomas of bone
.
BMC Cancer
12
,
375
30
Hidayat
,
S.
,
Yoshino
,
K.
,
Tokunaga
,
C.
,
Hara
,
K.
,
Matsuo
,
M.
and
Yonezawa
,
K.
(
2003
)
Inhibition of amino acid-mTOR signaling by a leucine derivative induces G1 arrest in Jurkat cells
.
Biochem. Biophys. Res. Commun.
301
,
417
423
31
Zheng
,
Y.
,
Delgoffe
,
G.M.
,
Meyer
,
C.F.
,
Chan
,
W.
and
Powell
,
J.D.
(
2009
)
Anergic T cells are metabolically anergic
.
J. Immunol.
183
,
6095
6101
32
Hutson
,
S.M.
,
Berkich
,
D.
,
Drown
,
P.
,
Xu
,
B.
,
Aschner
,
M.
and
LaNoue
,
K.F.
(
1998
)
Role of branched-chain aminotransferase isoenzymes and gabapentin in neurotransmitter metabolism
.
J. Neurochem.
71
,
863
874
33
Mihaylova
,
M.M.
and
Shaw
,
R.J.
(
2011
)
The AMPK signalling pathway coordinates cell growth, autophagy and metabolism
.
Nat. Cell Biol.
13
,
1016
1023
34
Gwinn
,
D.M.
,
Shackelford
,
D.B.
,
Egan
,
D.F.
,
Mihaylova
,
M.M.
,
Mery
,
A.
,
Vasquez
,
D.S.
et al (
2008
)
AMPK phosphorylation of raptor mediates a metabolic checkpoint
.
Mol. Cell
30
,
214
226
35
Hawley
,
S.A.
,
Gadalla
,
A.E.
,
Olsen
,
G.S.
and
Hardie
,
D.G.
(
2002
)
The antidiabetic drug metformin activates the AMP-activated protein kinase cascade via an adenine nucleotide-independent mechanism
.
Diabetes
51
,
2420
2425
36
She
,
P.
,
Reid
,
T.M.
,
Bronson
,
S.K.
,
Vary
,
T.C.
,
Hajnal
,
A.
,
Lynch
,
C.J.
et al (
2007
)
Disruption of BCATm in mice leads to increased energy expenditure associated with the activation of a futile protein turnover cycle
.
Cell Metab.
6
,
181
194
37
Neishabouri
,
S.H.
,
Hutson
,
S.M.
and
Davoodi
,
J.
(
2015
)
Chronic activation of mTOR complex 1 by branched chain amino acids and organ hypertrophy
.
Amino Acids
47
,
1167
1182
38
Ananieva
,
E.A.
,
Bostic
,
J.N.
,
Torres
,
A.A.
,
Glanz
,
H.R.
,
McNitt
,
S.M.
,
Brenner
,
M.K.
et al (
2018
)
Mice deficient in the mitochondrial branched-chain aminotransferase (BCATm) respond with delayed tumour growth to a challenge with EL-4 lymphoma
.
Br. J. Cancer
119
,
1009
1017
39
Bifari
,
F.
and
Nisoli
,
E.
(
2017
)
Branched-chain amino acids differently modulate catabolic and anabolic states in mammals: a pharmacological point of view
.
Br. J. Pharmacol.
174
,
1366
1377
40
Vazquez
,
A.
,
Kamphorst
,
J.J.
,
Markert
,
E.K.
,
Schug
,
Z.T.
,
Tardito
,
S.
and
Gottlieb
,
E.
(
2016
)
Cancer metabolism at a glance
.
J. Cell Sci.
129
,
3367
3373
41
Bovee
,
J.V.
,
van Royen
,
M.
,
Bardoel
,
A.F.
,
Rosenberg
,
C.
,
Cornelisse
,
C.J.
,
Cleton-Jansen
,
A.M.
et al (
2000
)
Near-haploidy and subsequent polyploidization characterize the progression of peripheral chondrosarcoma
.
Am. J. Pathol.
157
,
1587
1595
42
Heymann
,
D.
and
Redini
,
F.
(
2013
)
Targeted therapies for bone sarcomas
.
Bonekey Rep.
2
,
378
43
Brown
,
H.K.
,
Schiavone
,
K.
,
Gouin
,
F.
,
Heymann
,
M.F.
and
Heymann
,
D.
(
2018
)
Biology of bone sarcomas and new therapeutic developments
.
Calcif. Tissue Int.
102
,
174
195
44
Goto
,
M.
,
Miyahara
,
I.
,
Hirotsu
,
K.
,
Conway
,
M.
,
Yennawar
,
N.
,
Islam
,
M.M.
et al (
2005
)
Structural determinants for branched-chain aminotransferase isozyme-specific inhibition by the anticonvulsant drug gabapentin
.
J. Biol. Chem.
280
,
37246
37256
45
Cho
,
H.R.
,
Jeon
,
H.
,
Park
,
C.K.
,
Park
,
S.H.
,
Kang
,
K.M.
and
Choi
,
S.H.
(
2017
)
BCAT1 is a new MR imaging-related biomarker for prognosis prediction in IDH1-wildtype glioblastoma patients
.
Sci. Rep.
7
,
17740
46
Panosyan
,
E.H.
,
Lin
,
H.J.
,
Koster
,
J.
and
Lasky
, III,
J.L.
(
2017
)
In search of druggable targets for GBM amino acid metabolism
.
BMC Cancer
17
,
162
47
Raffel
,
S.
,
Falcone
,
M.
,
Kneisel
,
N.
,
Hansson
,
J.
,
Wang
,
W.
,
Lutz
,
C.
et al (
2017
)
BCAT1 restricts αKG levels in AML stem cells leading to IDHmut-like DNA hypermethylation
.
Nature
551
,
384
388
48
Kerr
,
D.A.
,
Lopez
,
H.U.
,
Deshpande
,
V.
,
Hornicek
,
F.J.
,
Duan
,
Z.
,
Zhang
,
Y.
et al (
2013
)
Molecular distinction of chondrosarcoma from chondroblastic osteosarcoma through IDH1/2 mutations
.
Am. J. Surg. Pathol.
37
,
787
795
49
Thewes
,
V.
,
Simon
,
R.
,
Hlevnjak
,
M.
,
Schlotter
,
M.
,
Schroeter
,
P.
,
Schmidt
,
K.
et al (
2017
)
The branched-chain amino acid transaminase 1 sustains growth of antiestrogen-resistant and ERα-negative breast cancer
.
Oncogene
36
,
4124
4134
50
Ananieva
,
E.A.
,
Powell
,
J.D.
and
Hutson
,
S.M.
(
2016
)
Leucine metabolism in T cell activation: mTOR signaling and beyond
.
Adv. Nutr.
7
,
798S
805S
51
Liu
,
K.A.
,
Lashinger
,
L.M.
,
Rasmussen
,
A.J.
and
Hursting
,
S.D.
(
2014
)
Leucine supplementation differentially enhances pancreatic cancer growth in lean and overweight mice
.
Cancer Metab.
2
,
6
52
Munoz-Pinedo
,
C.
,
El Mjiyad
,
N.
and
Ricci
,
J.E.
(
2012
)
Cancer metabolism: current perspectives and future directions
.
Cell Death Dis.
3
,
e248
53
Saha
,
A.K.
,
Xu
,
X.J.
,
Lawson
,
E.
,
Deoliveira
,
R.
,
Brandon
,
A.E.
,
Kraegen
,
E.W.
et al (
2010
)
Downregulation of AMPK accompanies leucine- and glucose-induced increases in protein synthesis and insulin resistance in rat skeletal muscle
.
Diabetes
59
,
2426
2434
54
Kwon
,
G.
,
Marshall
,
C.A.
,
Pappan
,
K.L.
,
Remedi
,
M.S.
and
McDaniel
,
M.L.
(
2004
)
Signaling elements involved in the metabolic regulation of mTOR by nutrients, incretins, and growth factors in islets
.
Diabetes
53
,
S225
S232
55
Sabatini
,
D.M.
(
2017
)
Twenty-five years of mTOR: uncovering the link from nutrients to growth
.
Proc. Natl. Acad. Sci. U.S.A.
114
,
11818
11825
56
O'Sullivan
,
D.
and
Pearce
,
E.L.
(
2015
)
Targeting T cell metabolism for therapy
.
Trends Immunol.
36
,
71
80
57
Nicklin
,
P.
,
Bergman
,
P.
,
Zhang
,
B.
,
Triantafellow
,
E.
,
Wang
,
H.
,
Nyfeler
,
B.
et al (
2009
)
Bidirectional transport of amino acids regulates mTOR and autophagy
.
Cell
136
,
521
534
58
Choi
,
Y.K.
and
Park
,
K.G.
(
2018
)
Targeting glutamine metabolism for cancer treatment
.
Biomol. Ther. (Seoul)
26
,
19
28
59
Su
,
T.Z.
,
Lunney
,
E.
,
Campbell
,
G.
and
Oxender
,
D.L.
(
1995
)
Transport of gabapentin, a gamma-amino acid drug, by system l alpha-amino acid transporters: a comparative study in astrocytes, synaptosomes, and CHO cells
.
J. Neurochem.
64
,
2125
2131
60
Son
,
S.M.
,
Park
,
S.J.
,
Lee
,
H.
,
Siddiqi
,
F.
,
Lee
,
J.E.
,
Menzies
,
F.M.
et al (
2019
)
Leucine signals to mTORC1 via its metabolite acetyl-coenzyme A
.
Cell Metab.
29
,
192
201.e7
61
Perez
,
J.
,
Decouvelaere
,
A.V.
,
Pointecouteau
,
T.
,
Pissaloux
,
D.
,
Michot
,
J.P.
,
Besse
,
A.
et al (
2012
)
Inhibition of chondrosarcoma growth by mTOR inhibitor in an in vivo syngeneic rat model
.
PLoS One
7
,
e32458
62
Taylor
,
P.J.
,
Franklin
,
M.E.
,
Graham
,
K.S.
and
Pillans
,
P.I.
(
2007
)
A HPLC-mass spectrometric method suitable for the therapeutic drug monitoring of everolimus
.
J. Chromatogr. B Anal. Technol. Biomed. Life Sci.
848
,
208
214
63
Garimella
,
R.
,
Eskew
,
J.
,
Bhamidi
,
P.
,
Vielhauer
,
G.
,
Hong
,
Y.
,
Anderson
,
H.C.
et al (
2013
)
Biological characterization of preclinical bioluminescent osteosarcoma orthotopic mouse (BOOM) model: a multi-modality approach
.
J. Bone Oncol.
2
,
11
21

Supplementary data