Distinct Roles of Glutamine Metabolism in Benign and Malignant Cartilage Tumors With IDH Mutations.

Enchondromas and chondrosarcomas are common cartilage neoplasms that are either benign or malignant, respectively. The majority of these tumors harbor mutations in either IDH1 or IDH2. Glutamine metabolism has been implicated as a critical regulator of tumors with IDH mutations. Using genetic and pharmacological approaches, we demonstrated that glutaminase-mediated glutamine metabolism played distinct roles in enchondromas and chondrosarcomas with IDH1 or IDH2 mutations. Glutamine affected cell differentiation and viability in these tumors differently through different downstream metabolites. During murine enchondroma-like lesion development, glutamine-derived α-ketoglutarate promoted hypertrophic chondrocyte differentiation and regulated chondrocyte proliferation. Deletion of glutaminase in chondrocytes with Idh1 mutation increased the number and size of enchondroma-like lesions. In contrast, pharmacological inhibition of glutaminase in chondrosarcoma xenografts reduced overall tumor burden partially because glutamine-derived non-essential amino acids played an important role in preventing cell apoptosis. This study demonstrates that glutamine metabolism plays different roles in tumor initiation and cancer maintenance. Supplementation of α-ketoglutarate and inhibiting GLS may provide a therapeutic approach to suppress enchondroma and chondrosarcoma tumor growth, respectively.

Introduction

Enchondroma is a common benign cartilaginous neoplasm and is estimated to be present in 3% of the total population.( , ) These tumors develop from dysregulated chondrocyte differentiation in the growth plate and are mostly present in the metaphysis of long bones.( ) In patients with multiple enchondromatosis (more than one enchondroma lesion) such as Maffucci's syndrome and Ollier's disease, the risk of malignant transformation is reported to be up to 60%.( ) Chondrosarcoma is the second most common primary malignancy of the bone.( ) They arise de novo or develop from preexisting benign tumors including enchondromas.( ) High‐grade chondrosarcomas have high metastatic potential and poor prognosis.( ) Currently there are no universally effective pharmacologic therapies for enchondromas or chondrosarcomas.

Somatic mutations of isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) are the most frequent genetic variations in enchondromas and chondrosarcomas.( , , , , ) They are present in 56% to 90% of enchondroma tumors( , ) and in about 50% of chondrosarcoma tumors.( , ) Although IDH1 and IDH2 mutations are present in enchondromas and chondrosarcomas, it is not known whether these tumors share similar metabolic requirements for tumor development and cell viability. Wild‐type IDH1 and IDH2 enzymes catalyze the reversible conversion between isocitrate and α‐ketoglutarate (α‐KG) in the cytoplasm or the mitochondria, respectively. Mutant IDH1 and IDH2 enzymes lose their original function and gain a neomorphic function that converts α‐KG to D‐2‐hydroxyglutarate (D‐2HG).( , , ) D‐2HG is considered as a putative “oncometabolite” in various cancers with mutations in IDH1/2.( , , , , , , ) Interestingly, pharmacological inhibition of mutant IDH1 enzyme did not alter chondrosarcoma tumorigenesis despite effective reduction of D‐2HG synthesis.( ) Several clinical trials of mutant IDH inhibitors have been conducted in patients with chondrosarcoma.( ) The results to date have been variable, showing at best stabilization of the disease.( ) Recently, an investigation of the metabolomes in chondrosarcomas showed global alterations in cellular metabolism in IDH1 or IDH2 mutant chondrosarcomas, suggesting such alteration might drive the neoplastic phenotype or be possible therapeutic targets for cancers with IDH1 or IDH2 mutations.( )

Glutamine metabolism is an important metabolic pathway that is critical for the survival of various cancers as well as proper proliferation and differentiation of different cell types.( , , , , , , ) Glutamine metabolism starts when glutaminase (GLS) deaminates glutamine to glutamate. Glutamate could be further used to generate α‐ketoglutarate (α‐KG), glutathione, other non‐essential amino acids, and nucleotides, etc. Through different downstream metabolites, glutamine regulates cancer cell behaviors by modulating bioenergetics, biosynthesis, redox homeostasis, etc.( ) In cancers with IDH1 or IDH2 mutations, glutamine is utilized as the primary source for D‐2HG production.( , , ) In addition, some tumors with IDH1 or IDH2 mutations are reported to be dependent on glutamine metabolism for tumor growth or cell viability.( , , , ) In non‐cancerous cells, glutamine metabolism regulates cell differentiation mainly through downstream metabolites, including amino acids, α‐KG, and acetyl‐CoA.( , , , )

In the context of cartilage tumors, it is unknown how glutamine regulates tumor development in enchondroma and cancer cell survival in chondrosarcoma. We therefore used a genetically engineered mouse model of enchondroma and primary human chondrosarcoma samples to address these questions. Here we identified that GLS was upregulated in both human patient chondrosarcoma samples with mutations in IDH1 or IDH2 and murine chondrocytes with Idh1 mutation. However, deleting Gls in the mouse led to increased number and size of benign tumor‐like lesions and affected hypertrophic chondrocyte differentiation, likely due to a reduction of α‐KG; whereas inhibiting GLS in IDH1 or IDH2 mutant chondrosarcomas led to a smaller tumor size and a reduction in cell viability, associated with the compromised production of non‐essential amino acids. Collectively, these data highlight a previously unknown stage‐dependent role of glutamine metabolism in cartilage tumors and may provide a therapeutic approach for the malignant cartilage tumor chondrosarcoma.

Materials and Methods

Animals

Idh1 , Gls (JAX: #017894), Col2a1Cre, Col2a1CreERT2, and NOD scid gamma (NSG) mice are as previously described.( , , , , , ) Congenic animals were used in every experiment. Idh1 mice bear a conditional knock‐in of the point mutation IDH1‐R132Q as previously clarified.( , ) Mice were housed on a 12‐hour light and dark cycle with free access to water and PicoLab Rodent Diet 20 (LabDiet #5053, St. Louis, MO, USA) at 23°C. All animal procedures were approved by the institutional animal care and use committee at Duke University.

Isolation of primary chondrocytes

Primary sternal chondrocyte isolation was as described before.( ) In brief, mouse sterna and ribs from P3 neonates were cleaned and digested by pronase (Roche, Indianapolis, IN, USA; 11459643001) at 2 mg/mL PBS at 37°C with constant agitation for 1 hour, Collagenase IV (Worthington, Lakewood, NJ, USA; LS004189) at 3 mg/mL DMEM at 37°C for 1 hour, Collagenase IV at 0.5 mg/mL DMEM at 37°C for 3 hours, and filtered using 45 μm cell strainer.

Cell culture

Chondrocytes were cultured in high‐glucose DMEM with 10% FBS and 1% penicillin/streptomycin. High‐glucose DMEM was used to maintain the growth and identity of primary costal chondrocytes. Primary chondrosarcoma cells were cultured in α‐MEM with 10% FBS and 1% penicillin/streptomycin. In some experiments, chondrocytes and chondrosarcoma cells were treated with 10 μM Bis‐2‐(5‐phenylacetamido‐1,2,4‐thiadiazol‐2‐yl)ethyl sulfide (BPTES), 100 μM epigallocatechin gallate (EGCG), 500 μM aminooxyacetate (AOA), 2X non‐essential amino acids, or DMSO as indicated for indicated 48 hours. Cells were cultured at 37°C in a humidified 5% CO2 incubator.

Transfection

In indicated experiment, chondrocytes were transfected with adenovirus‐GFP and adenovirus‐Cre at 400 MOI.

RNA isolation and qPCR

Total RNA isolation was isolated using Trizol and chloroform. An amount of 1 mL Trizol was added to cell pellet. After resuspension, the tubes were incubated at room temperature for 10 minutes. Then 200 μL of chloroform was added to the tube and mixed well. After centrifugation at maximum speed for 10 minutes, the solution was separated into three layers. The top layer of aqueous phase solution was transferred to a clean tube and mixed with 200 μL of isopropanol to precipitate RNA. After incubation at room temperature for 10 minutes, the tube was centrifuged at maximum speed for 10 minutes, and supernatant was discarded. RNA pellet was washed with 70% ethanol two times. After that, RNA was dissolved in molecular‐grade water.

An amount of 1 μg RNA was used to synthesize cDNA through reverse transcription using iScript cDNA synthesis kit. Quantitative real‐time PCR (qPCR) was then performed using the Powerup SYBR green reagent (Invitrogen, Carlsbad, CA, USA; A25777) on a QuantStudio 3 real‐time PCR system (Thermo Fisher Scientific, Waltham, MA, USA). Relative gene expression was calculated and normalized to a housekeeping gene (beta‐actin) using the 2ΔΔCt method.

Western blotting

Cultured chondrocytes were lysed in RIPA buffer containing protease and phosphatase inhibitor tablet (Roche, 11836170001, and Sigma [St. Louis, MO, USA], 4906845001, respectively). Protein concentration was quantified using the BCA method (Pierce, Rockford, IL, USA). Proteins were resolved on 8% polyacrylamide gels and transferred onto Immuno‐Blot PVDF membrane (Bio‐Rad, Hercules, CA, USA). After the wet transfer, the membranes were blocked for 1 hour at room temperature in 5% milk powder in TBS with 0.1% Tween (TBST) and then incubated at 4°C with the primary antibody (Anti‐Glutaminase antibody: Abcam [Cambridge, MA, USA] AB93434; Anti‐β‐actin: Invitrogen MA5‐15739) overnight. Membranes were washed three times with TBST and further incubated with anti‐rabbit IgG, HRP‐linked antibody (RRID: AB_2099233) in 5% milk (in TBST) for 1 hour at room temperature. All blots were developed using either enhanced or normal chemiluminescence (Clarity Substrate Kit, Bio‐Rad).

Metatarsal organ culture

The second, third, and fourth metatarsal bones were dissected from the hindlimbs of embryos at E16.5. They were transferred to 24‐well non‐adherent plates with 1 mL α‐MEM supplemented with 50 μg/mL ascorbic acid, 1 mM β‐glycerophosphate, and 0.2% bovine serum albumin. The media was changed every other day. Explants were cultured at 37°C in a humidified 5% CO2 incubator for 5 days and then fixed with 10% neutral buffered formalin (NBF).

EdU assay

EdU assay (Thermo Fisher Scientific, C10337) was performed according to the manufacturer's instructions. In brief, cells were cultured with 10 mM EdU for 12 hours before fixation with 4% PFA/PBS for 15 minutes and permeabilization with 0.5% Triton X‐100 for 20 minutes at room temperature. Cells were then incubated with Click‐iT reaction cocktail for 30 minutes in the dark at room temperature and stained with DAPI.

TUNEL assay

TUNEL assay was performed according to the manufacturer's instructions (Roche, 11684795910). In brief, cells were fixed with 4% PFA/PBS at room temperature for 1 hour and permeabilized with 0.1% Triton X‐100 in 0.1% sodium citrate for 2 minutes on ice. After rinse with PBS, cells were incubated with TUNEL reaction mixture at 37°C for 1 hour and stained with DAPI.

Cell viability assay

Cell viability was determined by CellTiter‐Glo Assay according to the manufacturer's instructions (Promega, Madison, WI; G7570). In brief, CellTiter‐Glo Buffer was thawed at room temperature and transferred to CellTiter‐Glo Substrate. Cell culture plate was equilibrated to room temperature for 30 minutes. An amount of 100 μL CellTiter‐Glo Reagent was added to the cell culture media in each well. The plate was mixed for 2 minutes on an orbital shaker and incubated for 10 minutes at room temperature. Luminescence was then recorded. Cell viability of each primary chondrosarcoma patient sample was normalized to the average cell viability in the vehicle group of the same sample.

Annexin V/PI staining

Annexin V and Propidium Iodide staining was performed according to the manufacturer's instructions (Invitrogen R37176 and P1304MP). In brief, 1 drop of Annexin V APC Ready Flow Conjugate was added to 0.5 mL of annexin‐binding buffer with 2.5 mM calcium. Then 1 mg/mL PI was added to the APC binding buffer with 1/1000 dilution. The cells were incubated for 15 minutes at room temperature. Fluorescence was detected by flow cytometry.

GLS activity assay

Primary chondrocytes were cultured before the experiment. After washing cells with Hanks' Buffered Saline Solution (HBSS) three times, cells were cultured in α‐MEM‐containing media containing 2 μM Glutamine and 4 mCi/mL L‐[3,4‐3H(N)]‐Glutamine (PerkinElmer, Waltham, MA, USA; NET551250UC). GLS activity was terminated by washing cells with ice‐cold HBSS three times followed by scraping cells with 1 mL ice‐cold milliQ water. Cells were lysed by sonication for 1 minute with 1‐second pulse at 20% amplitude. Cell lysates were bound onto AG 1‐X8 polyprep anion exchange column. Uncharged glutamine was eluted with 2 mL of milliQ water three times. Negatively charged glutamate and downstream metabolites were eluted with 2 mL of 0.1 M HCl three times. After adding 4 mL of scintillation cocktail to the eluent, DPM of the solution was measured by a Beckman Coulter (Brea, CA, USA) LS6500 Scintillation counter.

Xenograft

Before the experiment, chondrosarcoma cells from each patient's tumor were maintained subcutaneously in vivo in NSG mice. For the xenograft experiment, tumors were surgically removed from each mouse and divided into explants of 5 × 5 × 5 mm each, and implanted into the subcutaneous tissue on the pack of NSG mice. BPTES and vehicle control (10% DMSO/PBS) treatment started 10 days after implantation. Mice were treated with BPTES at 0.2 g/kg or vehicle via intraperitoneal injection daily for 14 days. Tumor weights were recorded upon harvest. For each patient‐derived xenograft experiment, relative tumor weight was determined by normalizing the tumor weight of each tumor to the average tumor weight of the vehicle control group.

Histological analysis

Bone histomorphometry for adult mice was performed on hindlimbs fixed in 10% neutral buffered formalin for 3 days followed by decalcification with 14% EDTA for 2 weeks at room temperature. Histomorphometry for embryonic skeletons was performed on tibias fixed in 10% neutral buffered formalin overnight followed by decalcification with 14% EDTA overnight at room temperature. After decalcification, skeletons were embedded in paraffin and sectioned at 5 μm thickness. Safranin O staining was performed following standard protocol. Animals of both sexes were used for analysis. Hindlimbs from the left side were used for histological analyses. For quantification, each data point represents one individual animal.

Immunohistochemistry

Immunohistochemistry was performed on 5‐μm paraffin‐sectioned limbs. For type X collagen, antigen retrieval was performed by citrate buffer incubation at 85°C for 15 minutes and hyaluronidase digestion at 10 mg/mL at 37°C for 30 minutes. For BrdU staining, BrdU labeling reagent (Invitrogen, 000103) was injected into pregnant female mice at 1 mL/100 g body weight 2 hours before euthanasia. Antigen retrieval was performed by proteinase K digestion at 10 μg/mL at room temperature for 10 minutes. For MMP13, antigen retrieval was performed by hyaluronidase digestion at 5 mg/mL at 37°C for 30 minutes. For all immunohistochemistry, endogenous peroxidase activity was blocked by incubation with 3% H2O2/Methanol for 10 minutes followed by incubation with Dako Dual Endogenous Enzyme Block reagent (Agilent Dako, Santa Clara, CA; S2003) for 30 minutes at room temperature. The specimen was blocked with 2% horse serum at room temperature for 30 minutes followed by incubation with antibodies for Col X (1:500, Thermo Fisher Scientific, 14‐9771‐82), BrdU (1:1000, Thermo Fisher Scientific, MA3‐071), MMP13 (1:100, MilliporeSigma, Burlington, MA, USA; MAB13424) overnight at 4°C. TUNEL assay was performed according to the manufacturer's instructions (Roche, 11684795910). In brief, the tissue was incubated with proteinase K at 10 μg/mL at room temperature for 20 minutes and then in TUNEL labeling reagent 37°C for 1 hour. Animals with both sexes were used for analysis. Hindlimbs from the left side were used for histological analyses. For quantification, each data point represents one individual animal. Quantification of the length of hypertrophic zone and bone elements was done using the image processing software Fiji Image J.

In situ hybridization

In situ hybridization was performed on 5‐μm paraffin‐sectioned limbs. Paraffin sections were deparaffinized and rehydrated, followed by fixation with 4% PFA at room temperature for 15 minutes. Sections were then treated with 20 μg/mL proteinase K for 15 minutes at room temperature, fixed with 4% PFA at room temperature for 10 minutes, and acetylation solution [1.33% triethanolamine, 0.175% HCl (37%), 0.375% acetic anhydride in MilliQ water] for 10 minutes. Sections were then incubated with hybridization buffer at 58°C for 3 hours, and then incubated with Digoxigenin‐labeled RNA probe (Col2a1, Pth1r, Co10a1) at 58°C overnight. Sections were then washed with 5x SSC for one time at 65°C, followed by RNAse A treatment at 37°C for 30 minutes. Sections were then washed 2x SSC for one time, 0.2x SSC for two times at 65 °C, and then blocked with 2% Boehringer Blocking Reagent/20% Heat Inactivated Sheep Serum for 1 hour. Sections were then blocked with Anti‐Digoxigenin antibody (1:4000 in blocking solution) at 4°C overnight. Sections were developed with BM Purple at room temperature until color developed. Animals of both sexes were used for analysis. Hindlimbs from the left side were used for histological analyses. For quantification, each data point represents one individual animal.

Analysis of enchondroma‐like lesions and trabecular bone volume

Tamoxifen was administered daily for 10 days at 100 mg/kg body weight/day via intraperitoneal injection starting at 4 weeks of age. Mice were euthanized at 6 months of age. Hindlimbs were harvested for analyzing growth‐plate and enchondroma‐like phenotype. Hip‐joint cartilage was used for confirming DNA recombination. Quantification of enchondroma‐like lesions was performed as previously described.( ) In brief, enchondroma‐like lesions were first identified by Safranin O staining, which was performed on one slide (2 sections, 5 μm/section) in every 10 consecutive slides (10 μm). We then examined every slide consecutively under the light microscope to identify lesions that did not span to the Safranin O‐stained slide. The number of lesions were then recorded. For every Safranin O‐stained slide, we manually outlined each lesion and measured the area of each lesion using the image processing software Fiji Image J. We estimated the lesion size of each animal by adding up the areas of all the lesions in that animal. The lesion size for each animal was normalized to the average lesion size of Col2CreERT2;Idh1 animals. Quantification of trabecular bone volume was performed by adding up the trabecular bone surface of each Safranin O‐stained slide. For every Safranin O‐stained slide, we manually outlined the trabecular bone 1200 μm below the growth plates of each tibia and femur and measured the area of trabecular bones using the image processing software Fiji Image J. We estimated the lesion size of each animal by adding up the areas of all the trabecular bones in our region of interest in that animal. Animals of both sexes were used for analysis. Hindlimbs from the left side were used for histological analyses. For quantification, each data point represents one individual animal.

D‐2HG measurement

D‐2HG is analyzed by liquid chromatography–tandem mass spectrometry (LC/MS/MS). Cell culture media was collected from each sample. After adding 2‐HG‐2H4 (internal standard), the sample was dried under nitrogen and derivatized by (+)‐O,O′‐diacetyl‐L‐tartaric anhydride (DATAN) for measurement.

Carbon isotope labeling

Chondrocytes were cultured in DMEM with 4500 mg/L Glucose and 4 mM 15C5‐Glutamine in 6‐cm cell culture plates for specified times. Primary chondrosarcoma cells were cultured in α‐MEM with 1000 mg/L Glucose and 2 mM 15C5‐Glutamine in 6‐cm cell culture plates for specified times. Methanol (500 μL) was used to extract metabolites from each plate. After centrifugation at 13,800 g for 15 minutes, supernatant was dried at 37°C. The dried residues were resuspended in 25 μL of methoxylamine hydrochloride (2% [w/v] in pyridine) and incubated at 40°C for 1.5 hours in a heating block. After brief centrifugation, 35 μL of MTBSTFA + 1% TBDMS was added, and the samples were incubated at 60°C for 30 minutes. The derivatized samples were centrifuged for 5 minutes at 20,000g, and the supernatants were transferred to GC vials for GC–MS analysis. A modified GC–MS method was employed.( ) The injection volume was 1 μL, and samples were injected in splitless mode. GC oven temperature was held at 80°C for 2 minutes, increased to 280°C at 7°C/minute, and held at 280°C for a total run time of 40 minutes. GC–MS analysis was performed on an Agilent 7890B GC system equipped with a HP‐5MS capillary column (30 m, 0.25 mm i.d., 0.25 μm‐phase thickness; Agilent J&W Scientific, Santa Clara, CA, USA), connected to an Agilent 5977A Mass Spectrometer operating under ionization by electron impact (EI) at 70 eV. Helium flow was maintained at 1 mL/min. The source temperature was maintained at 230°C, the MS quad temperature at 150°C, the interface temperature at 280°C, and the inlet temperature at 250°C. Mass spectra were recorded in mass scan mode with m/z from 50 to 700.

‐based stable isotope analysis

M0, M1, …, Mn refers to the isotopologues containing n heavy atoms in a molecule. The stable isotope distribution of individual metabolites was measured by GC–MS as described above. The isotopologue enrichment or labeling in this work refers to the corrected isotope distribution.( , )

Gene expression of Gls from chondrosarcoma patient samples

We used retrospective data from RESOS INCA network of bone, collected from 102 cartilage tumors in different French hospitals. Multiple samples may be taken from each tumor and sequenced. More details about the experiment, such as RNA isolation method and profiling, can be found the original article.( ) For our purposes, we utilized IDH mutation information to categorize samples into two groups: IDH1 or IDH2 mutation (n = 46) and IDH1 and IDH2 wild type (n = 98). We examined several gene expression levels across these two groups and reported adjusted p value for GLS result (with Benjamini–Hochberg correction). All the data processing and calculation were performed using Bioconductor docker (devel), with R version 4.0.3 and Bioconductor version 3.13.

Quantification and statistical analysis

Statistical analyses were performed using GraphPad (La Jolla, CA, USA) Prism 9 software. Data were presented as mean ± SEM, mean ± SD, or mean ± 95% as specified in each figure. Statistical significance was determined by two‐tailed Student's t test or one‐way or two‐way ANOVA with multiple comparisons test as specified in each figure.

Results

or mutant chondrosarcomas cells exhibited increased glutamine contribution to anaplerosis and non‐essential amino acids production

From a published data set showing mRNA profiling (E‐MTAB‐7264) of chondrosarcoma tumors,( ) we observed that expression of GLS was upregulated in chondrosarcomas with IDH1 or IDH2 mutations (Supplemental Fig. S1), indicating glutamine metabolism might be important for cartilage tumor with IDH1 or IDH2 mutation. To understand how glutamine was utilized in chondrosarcoma cells, we performed a carbon tracing experiment with 13C5‐glutamine in chondrosarcoma cells derived from 3 patients, each with wild‐type IDH1 and IDH2, mutant IDH1, and mutant IDH2. 13C contribution to downstream metabolites was determined by measuring the isotope‐labeling pattern. To be utilized by a cell, glutamine is first deaminated to glutamate through GLS. After that, glutamate can be converted to α‐KG, a key metabolite in the tricarboxylic acid (TCA) cycle (Fig. 1). There was a significant amount of glutamate labeled with 13C in chondrosarcomas of different genotypes (Fig. 1). Importantly, the labeling of 13C in glutamate was significantly higher in IDH1 or IDH2 mutant chondrosarcomas when compared with chondrosarcoma cells with wild‐type IDH1 and IDH2 (Fig. 1). The 13C labeling pattern was examined in the TCA cycle intermediates and non‐essential amino acids. Chondrosarcomas with IDH1 or IDH2 mutations had significantly higher carbon contribution to all TCA cycle intermediates (Fig. 1) and some non‐essential amino acids such as alanine (Fig. 1). Thus, IDH1 or IDH2 mutant chondrosarcomas were more efficient than IDH1 and IDH2 wild‐type chondrosarcomas in converting glutamine to its downstream metabolites.

Fig. 1

Chondrosarcomas with IDH1 or IDH2 mutations had significantly increased contribution from glutamine carbon to downstream metabolites. (A) Graphical depiction of tracing glutamine metabolism using 13C5‐glutamine. Filled circles indicate 13C and open circles indicate 12C. Green dashed line with arrowhead indicates the direction of oxidative decarboxylation. Blue dashed line with arrowhead indicates the direction of reductive carboxylation. (B–H) Percentage of 13C5‐glutamine contribution to glutamate (B), α‐ketoglutarate (C), fumarate (D), malate (E), oxaloacetate/aspartate (F), citrate (G), and alanine (H). The data represent three primary human chondrosarcoma samples each with wild‐type IDH1/2, IDH1 mutation, or IDH2 mutation. Each data point represents technical replicates. Data are presented as box plots with indication of median; whiskers represent min to max values. The p values on the graph were determined by one‐way ANOVA, Dunnett's multiple comparison tests. Results for ANOVA: (B) F = 722.7, p < 0.0001; (C) F = 1099, p < 0.0001; (D) F = 433.6, p < 0.0001; (E) F = 2535, p < 0.0001; (F) F = 647.5, p < 0.0001; (G) F = 213.4, p < 0.001; (H) F = 208.0, p < 0.0001.

GLS was upregulated in IDH mutant chondrocytes

To understand the role of GLS in murine chondrocytes with Idh1 mutation, we first examined whether expression of mutant Idh1 gene could lead to altered GLS function. Expression of a mutant IDH1R132Q enzyme in chondrocytes was shown to be sufficient to initiate enchondroma‐like lesion formation in mice.( ) Chondrocytes were isolated from mice expressing the conditional IDH1R132Q knock‐in allele and transduced with adenovirus GFP or adenovirus Cre.( ) GLS activity in these chondrocytes was determined by measuring the conversion from radioactive 3H‐glutamine to radioactive 3H‐glutamate. In primary chondrocytes expressing mutant IDH1R132Q, GLS activity was significantly upregulated by 2.5‐fold (Fig. 2).

Fig. 2

Chondrocytes with Idh1 mutation had increased GLS activity and glutamine contribution to downstream metabolites. (A) GLS activity of AdGFP;Idh1 and AdCre;Idh1 chondrocytes. (B–H) Percentage of 13C5‐glutamine contribution to glutamate (B), α‐ketoglutarate (C), fumarate (D), malate (E), oxaloacetate/aspartate (F), citrate (G), and alanine (H). Each data point represents individually transfected cell populations. Data are presented as box plots with indication of median; whiskers represent min to max values. The p values were determined by unpaired Student's t test.

To investigate how chondrocytes with wild‐type and mutant IDH1 enzymes utilize glutamine, we conducted 13C tracing studies. In chondrocytes expressing the mutant Idh1, the percentage of 13C labeling in glutamate (Fig. 2), α‐KG (Fig. 2), other TCA cycle intermediates (Fig. 2), and the non‐essential amino acid alanine (Fig. 2) was significantly increased.

Glutamine is a primary source for the production of D‐2HG in IDH1 or IDH2 mutant cancers including chondrosarcoma.( , , ) To examine whether glutamine is also the primary source for D‐2HG production in the murine chondrocytes expressing mutant Idh1, we cultured these chondrocytes with 13C5‐glutamine and examined the percentage of 13C‐labeled D‐2HG at different time points. After 10 hours, more than 80% of the D‐2HG was labeled with 13C, confirming glutamine is the primary source for D‐2HG in IDH1R132Q cells (Supplemental Fig. S2B). D‐2HG was not detected in chondrocytes without Idh1 mutation (Supplemental Fig. S2C). Thus, chondrocytes with an Idh1 mutation showed increased GLS activity and more efficient conversion of glutamine to downstream metabolites.

GLS was important for chondrocyte differentiation and proliferation in chondrocytes with Idh1 mutation

There are two isoforms of GLS, kidney‐type GLS (encoded by Gls) and liver‐type GLS (encoded by Gls2). Expression of Gls2 was significantly lower than Gls in primary chondrocytes of both Col2a1Cre;Idh1 and Col2a1Cre animals (Fig. 3, GSE123130), consistent with the notion that GLS activity in chondrocytes was mainly catalyzed by Gls. We thus focused on Gls in the murine chondrocytes. To examine the role of Gls in vivo, we conditionally deleted Gls in these cells by crossing mice carrying a conditional null allele of Gls (Gls ) to the Col2a1Cre deleter mouse, which expresses Cre‐recombinase in chondrocytes (Fig. 3).( , ) Col2a1Cre;Gls chondrocytes had reduced glutamine uptake and glutamate production (Fig. 3), confirming the function of GLS was efficiently deleted in these mice.

Fig. 3

Gls is the primary glutaminase isoform in chondrocytes. (A) Expression of Gls and Gls2 in chondrocytes of Col2a1Cre and Col2a1Cre;Idh1 animals. (B) Expression of Gls in chondrocytes of Gls and Col2a1Cre;Gls animals. (C) Western blot of GLS and β‐Actin in chondrocytes of Gls and Col2a1Cre;Gls animals. (D) 3H‐glutamate production (i) and 3H‐glutamine uptake (ii) in chondrocytes of Gls and Col2a1Cre;Gls animals. For A, p values on the graph were determined by two‐way ANOVA, Šídák's multiple comparisons test. F (1, 8) = 243.7, p < 0.001. For B and D, p values were determined by unpaired Student's t test.

We first examined the phenotype of tibias in mice lacking Gls in chondrocytes using H&E staining (Fig. 3). At embryonic (E) day 14.5, deleting Gls in Idh1 wild‐type chondrocytes or expressing mutant Idh1 gene in chondrocytes did not cause a significant change in bone length (Fig. 4). Deleting Gls in Idh1 mutant chondrocytes significantly reduced the bone length (Fig. 4). At E16.5 and E18.5, we observed reduced bone length and defects in chondrocyte hypertrophic differentiation in Idh1 mutant mice (Supplemental Fig. S3A–F), consistent with our previous report.( ) Deleting Gls did not cause an obvious skeletal phenotype in Idh1 wild‐type and Idh1 mutant mice (Supplemental Fig. S3A–F).

Fig. 4

Deleting GLS in Col2a1Cre;Idh1 chondrocytes affected chondrocyte differentiation. (A) H&E staining and quantification of the length of tibias. (B) In situ hybridization of Col2a1 and quantification of the length of Col2a1‐negative zone. (C) In situ hybridization of Pth1r and quantification of the length of Pth1r‐negative zone. (D) In situ hybridization of Col10a1 and quantification of the length of Col10a1‐positive zone. (E) Immunohistochemistry of Col X and quantification of the length of Col X‐positive area. (F) Immunohistochemistry of BrdU and quantification of percentage of BrdU‐positive cells. Each data point represents an individual animal. For A–E, p values were determined by one‐way ANOVA, Tukey's multiple comparisons test. Results for ANOVA: (A) F = 4.149, p = 0.0223; (B) F = 43.37, p < 0.0001; (C) F = 31.03, p < 0.001; (D) F = 16.92, p < 0.0001; (E) F = 10.83, p = 0.0001. For F, the p value was determined by unpaired Student's t test. Data are presented as box plots with indication of median; whiskers represent min to max values. Scale bar = 100 μm.

We then used in situ hybridization to examine markers important for chondrocyte differentiation. Early chondrocyte marker Col2a1 and pre‐hypertrophic chondrocyte marker Pth1r were expressed by tibial chondrocytes at E14.5 (Fig. 4). However, the regions between Col2a1‐expressing cells and Pth1r‐expressing cells were reduced in Idh1 mutant animals. In Gls;Idh1 double mutant mice, Col2a1‐expressing cells and Pth1r‐expressing cells became single regions in the middle of the bone with no separation (Fig. 4). Col10a1 is expressed by hypertrophic chondrocytes. At E14.5, the expression pattern of Col10a1 was comparable among control, Gls mutant, and Idh1 mutant mice, but in Gls;Idh1 double mutant mice, there was reduced zone of staining (Fig. 4). We also examined proliferation and apoptosis in these animals. No difference in proliferation was observed between control and Gls mutant animals in Idh1 wild‐type background, but proliferation was increased when Gls was deleted in Idh1 mutant mice (Fig. 4). Apoptosis was not detected in all animals at this stage based on immunohistochemistry of cleaved caspase 3 (Supplemental Fig. S3G).

Deleting GLS in chondrocytes increased the number and size of enchondroma‐like lesions

We next examined how deleting Gls in chondrocyte postnatally affected chondrocyte homeostasis in growth plates and enchondroma‐like lesion formation in adult animals. We induced deletion of Gls at 4 weeks of age by tamoxifen and examined the phenotype at 6 months of age, a time point when growth plates were completely remodeled in control animals and enchondroma‐like lesions were stable in Idh1 mutant mice. Enchondroma‐like lesions were identified in Idh1 mutant animals as we previously reported and these animals showed less trabecular bone volume (Fig. 5). No obvious growth plate phenotype was observed in Gls mutant mice, and histologic examination showed less trabecular bone (Fig. 5). In contrast, the number and size of enchondroma‐like lesions were significantly increased when Gls was deleted in Idh1 mutant animals, and the trabecular bone volume was further reduced (Fig. 5).

Fig. 5

Deleting Gls in Col2a1CreERT2;Idh1 chondrocytes increased the number and size of enchondroma‐like lesions. (A) Representative Safranin O staining of mice of specified genotypes. Enchondroma‐like lesions are highlighted in yellow circles. (B) Quantification of the volume of trabecular bone based on Safranin O staining. (C, D) Quantification of the number (B) and size (C) of enchondroma‐like lesions in animals of specified genotypes. Each data point represents an individual animal. Data are presented as box plots with indication of median; whiskers represent min to max values. For B, p values shown on the graph were determined by one‐way ANOVA, Tukey's multiple comparisons test. F = 12.57, p < 0.001. For C and D, the p value was determined by unpaired Student's t test. Scale bar = 100 μm.

GLS regulated chondrocyte differentiation through α‐ketoglutarate

Glutamine metabolism regulates chondrocyte differentiation through its downstream metabolites α‐KG in Idh1 wild‐type background.( ) We examined whether GLS could regulate the differentiation of chondrocytes containing a Idh1 mutation through α‐KG. Our previous 13C tracing experiments showed α‐KG was mainly derived from glutamine, and Idh1 mutant chondrocytes were more efficient in converting glutamine to α‐KG (Fig. 2). We found the level of α‐KG was reduced by more than 70% in Idh1 mutant chondrocytes (Fig. 6). Because chondrocytes with Idh1 mutation were more efficient in converting glutamine to α‐KG, the lower intracellular concentration was likely because α‐KG was converted to D‐2HG. When GLS was inhibited in Idh1 mutant chondrocytes by Bis‐2‐(5‐phenylacetamido‐1,3,4‐thiadiazol‐2‐yl)ethyl sulfide (BPTES), a small molecular inhibitor for GLS,( ) α‐KG level was further reduced by about 60% (Fig. 6).

Fig. 6

GLS regulated Col2a1Cre;Idh1 chondrocyte differentiation through the downstream metabolite α‐KG. (A) Relative intracellular α‐KG concentration in AdGFP;Idh1 and AdCre;Idh1 chondrocytes. (B) Relative intracellular α‐KG concentration in AdCre;Idh1 chondrocytes treated with vehicle control or 10 μM BPTES. (C) In situ hybridization of Col10a1. (D) Immunohistochemistry of Col X. (E) Quantification of the length of Col X‐positive zone. (F) Immunohistochemistry of BrdU. (G) Quantification of the percentage of BrdU‐positive cells. (H) Immunohistochemistry of cleaved caspase 3. For A and B, each data point represents individually transfected cell populations. The p values were determined by unpaired Student's t test. For E and G, each data point represents an individual animal. The p values were determined by one‐way ANOVA, Tukey's multiple comparisons test. Results for ANOVA: (E) F = 10.54, p = 0.0001; (F) F = 2.774, p = 0.0592. Data are presented as box plots with indication of median; whiskers represent min to max values. Scale bar = 100 μm.

Next, we investigated whether exogenous α‐KG could rescue the defects of chondrocyte differentiation in Col2a1Cre;Gls ;Idh1 animals. We injected cell‐permeable dimethyl‐α‐KG (DM‐α‐KG) to pregnant dams every day from E11.5 to E13.5 and harvested the embryos at E14.5. Treatment with DM‐α‐KG rescued chondrocyte hypertrophy as Col10a1 mRNA and COLX protein expression were both observed in Gls;Idh1 double mutant mice (Fig. 6). Finally, DM‐α‐KG treatment led to a modest decrease in proliferation (Fig. 6) without inducing apoptosis (Fig. 6).

Inhibiting GLS in or mutant chondrosarcoma xenografts reduced tumor weight

Previously it was reported that inhibiting GLS reduced viability in chondrosarcoma cell lines.( ) However, our murine data in enchondroma‐like lesions suggests that GLS inhibition would have an opposite effect. We thus established patient‐derived xenograft models as previously described to examine the effects of GLS inhibition in chondrosarcoma.( ) We established and maintained chondrosarcoma xenografts by injecting primary patient chondrosarcoma cells into immune‐deficient Nod‐scid‐gamma mice. We used five patient samples with IDH1 mutation and one patient sample with wild‐type IDH1 and IDH2. All of these chondrosarcomas were histological grade 2/3. Chondrosarcoma tumors were divided into pieces at 5 mm × 5 mm × 5 mm and implanted to immune‐deficient Nod‐scid‐gamma mice subcutaneously. We treated these animals with BPTES or vehicle control for 14 days and measured the tumor weight at the time of death. No adverse effects were observed in these animals after BPTES or vehicle treatment. For each patient sample, the tumor weight of each mouse was normalized to the average tumor weight in the vehicle control group of the same patient sample. BPTES treatment significantly reduced tumor weight of chondrosarcoma tumors with IDH1 mutations (Fig. 7). We examined proliferation and apoptosis in each of these tumors using immunohistochemistry of Ki67 and cleaved caspase 3, respectively. BPTES treatment significantly increased apoptosis in chondrosarcoma xenografts, but we did not observe a substantial difference in proliferation (Fig. 7). We then examined chondrosarcoma cells treated with BEPTS in vitro. Consistent with our findings in vivo, BPTES treatment reduced cell viability, increased cell apoptosis, and did not alter proliferation in IDH1 or IDH2 mutant chondrosarcomas (Fig. 7). Interestingly, BPTES treatment did not alter tumor burden of IDH1 and IDH2 wild‐type chondrosarcoma (Supplemental Fig. S4). These data support previously published studies and show a different role of GLS in human chondrosarcoma than in murine enchondroma‐like lesions.

Fig. 7

Inhibiting GLS reduced tumor weight and induced apoptosis in chondrosarcomas with IDH mutations. (A) Experimental design for the xenograft experiment. (B) Representative pictures of xenografted tumor at the time of harvest and relative tumor weight of xenografted chondrosarcoma tumors at the time of harvest. (C) Representative picture of immunohistochemistry of Ki67 of xenografted tumors and quantification of relative proliferation of each tumor determined by percentage of Ki67‐positive cells. (D) Representative picture of immunohistochemistry of cleaved caspase 3 of xenografted tumors and relative apoptotic of each tumor determined by percentage of cleaved caspase 3‐positive cells. (E) Relative cell viability of IDH1/2Mut chondrosarcoma cells treated with 10 μM BPTES determined by CellTiter Glo cell viability assay. (F) Relative proliferation rate of IDH1 or IDH2 mutant chondrosarcoma cells treated with 10 μM BPTES in vitro determined by EdU staining. (G) Relative apoptosis of IDH1 or IDH2 mutant chondrosarcoma cells treated with 10 μM BPTES in vitro determined by TUNEL staining. For B–D, each data point represents one tumor. The p value was determined by unpaired Student's t test. For E–G, each data point represents the average of three technical triplicates of one patient sample. Data are presented as box plots with indication of median; whiskers represent min to max values.

Inhibition of GLS reduced the production of non‐essential amino acids in chondrosarcoma

We next sought to understand how glutamine metabolism regulated cell survival of IDH1 or IDH2 mutant chondrosarcomas. Similar to in the mouse, α‐KG levels were lower with BPTES treatment (Fig. 8). However, unlike our murine findings, adding DM‐α‐KG did not rescue the apoptosis changes found with BEPTS treatment (Fig. 8).

Fig. 8

GLS regulated cell apoptosis of chondrosarcomas with IDH1 or IDH2 mutations through production of non‐essential amino acids. (A) Relative concentration of α‐KG in IDH1 or IDH2 mutant chondrosarcomas treated with 10 μM BPTES. (B) Relative apoptosis of IDH1 or IDH2 mutant chondrosarcoma cells treated with 10 μM BPTES, or 10 μM BPTES + 1 mM DM‐α‐ketoglutarate determined by Annexin V staining. (C) Relative cell viability of IDH1/2Mut chondrosarcoma cells treated with 100 μM AOA and 500 μM EGCG. Each dot represents one patient sample. (D) Relative concentration of different amino acids in IDH1 or IDH2 mutant chondrosarcomas treated with 10 μM BPTES. (E) Relative apoptosis of IDH1 or IDH2 mutant chondrosarcoma cells treated 10 μM BPTES, or 10 μM BPTES + 2X NEAA determined by Annexin V staining. For B, C, and E, each dot represents one patient sample. For A, unpaired Student's t test was used to determine p value. For B and E, paired Student's t test was used to determine p values. For D, p values are determined by two‐way ANOVA, Šídák's multiple comparisons test. F (1, 50) = 65.22, p < 0.0001. Data are presented as box plots with indication of median; whiskers represent min to max values.

Downstream of GLS, glutamate could be further metabolized to α‐KG through glutamate dehydrogenase and transaminases. Inhibiting transaminases by aminooxyacetate (AOA) caused a reduction of cell viability in IDH1 or IDH2 mutant chondrosarcomas similar to the effects of BPTES, but inhibiting glutamate dehydrogenase by epigallocatechin gallate (EGCG) did not affect cell survival (Fig. 8). These data suggested transaminases might be more critical for chondrosarcoma cell survival downstream of GLS.

Non‐essential amino acids can be produced from glutamine through GLS and then transaminases. They are known to play a role in regulating cancer apoptosis, raising the possibility that they may be playing this role in chondrosarcoma. We found lower levels of multiple non‐essential amino acids when GLS was inhibited by BPTES (Fig. 8). We treated these chondrosarcoma cells with non‐essential amino acids, and in contrast to our data from DM‐α‐KG, we found a modest rescue of apoptosis changes (Fig. 8). However, supplementing non‐essential amino acids to Col2a1Cre;Gls ;Idh1 metatarsal organ culture did not lead to rescue of chondrocyte differentiation defects (Supplemental Fig. S5), suggesting they might not be the key mediator of the murine enchondroma‐like lesions.

Discussion

In this study, we found that glutamine metabolism plays distinct roles in the tumor initiation of enchondroma and cancer maintenance of chondrosarcoma. Glutamine contribution to downstream metabolites was upregulated in chondrosarcomas and chondrocytes with mutations in IDH1 or IDH2 enzymes. During development of enchondroma, deleting Gls in murine chondrocytes with Idh1 mutation interrupted hypertrophic differentiation during embryonic development and led to increased number and size of enchondroma‐like lesions in adult animals. In malignant IDH1 or IDH2 mutant chondrosarcomas isolated from human patients, pharmacological inhibition of GLS led to reduced tumor weight.

The distinct roles of glutamine metabolism in enchondroma and chondrosarcoma were possibly due to different metabolic needs at different stages of tumor development. We observed that supplementation of α‐KG and non‐essential amino acids rescued the defects in chondrocyte differentiation and chondrosarcoma apoptosis, respectively. As enchondroma rises from dysregulated chondrocyte differentiation, α‐KG—the regulator for cell differentiation—plays an important role in the development of the benign tumor. For malignant chondrosarcomas, changes in differentiation might not be critical for tumor growth. However, these IDH1 or IDH2 mutant cancer cells might require some amino acids to support cell viability.

Molecular mechanisms underlying the distinct roles of glutamine metabolism in cartilage tumors remain to be further investigated. α‐KG is reported to regulate cell fate mainly through epigenetic regulation by serving as a co‐factor for 2‐oxoglutarate‐dependent oxygenases as well as precursor for other downstream metabolites. In mouse embryonic stem cells, α‐KG suppressed cell differentiation by promoting demethylation of DNA, demethylation of repressive histone marks, and demethylation of active histone mark through TET‐family of DNA demethylase and Jumonji C‐domain containing histone demethylases.( , ) In the context of chondrocyte differentiation, Stegen and colleagues reported Gls regulated chondrocyte differentiation through α‐KG and its downstream metabolite acetyl‐CoA postnatally via regulating histone acetylation.( ) Because enchondroma develops from dysregulated chondrocyte differentiation, it is possible that α‐KG could affect expression of chondrogenic genes through altering histone modifications and DNA methylation. In malignant cancers, glutamine supported cancer cell proliferation and suppressed apoptosis and autophagy through producing other non‐essential amino acids.( , , , , ) Non‐essential amino acids mixture provides seven non‐essential amino acids: alanine, asparagine, aspartate, glycine, serine, proline, and glutamate. Many of them have been shown to be critical for cancer cell viability, especially under metabolic stress. In glioblastoma and neuroblastoma, glutamine deprivation induced apoptosis, which could be restored by exogenous asparagine.( ) In pancreatic cancer, aspartate becomes “essential” when glutamine availability or metabolism is limited.( ) Glutamine could also regulate cancer cell viability through sustaining cellular redox homeostasis.( ) For the next step, it will be important to understand which and how non‐essential amino acid(s) prevent apoptosis in IDH1 or IDH2 mutant chondrosarcomas.

One limitation of this study is that we studied the role of glutamine metabolism in enchondromas and chondrosarcomas using mouse models and human patient samples, respectively. Thus, it is possible the distinct roles of glutamine metabolism we observed were caused by different metabolic needs of murine and human cells. To address this concern, it will be important to establish mouse models to study chondrosarcoma and to determine the role of glutamine metabolism in human enchondroma samples.

In summary, our study showed that GLS‐mediated glutamine metabolism played distinct roles in IDH1 or IDH2 mutant enchondromas and chondrosarcomas through different downstream metabolites. In the context of enchondroma, deleting GLS in chondrocytes with Idh1 mutation reduced the intracellular levels of α‐KG, thus impairing hypertrophic differentiation and increasing cell proliferation and leading to increased number and size of enchondroma‐like lesions. In chondrosarcoma, GLS regulated cell apoptosis partially by producing non‐essential amino acids. This study demonstrates that glutamine metabolism plays different roles in the tumor initiation and cancer maintenance in cartilage tumors. Supplementation of α‐ketoglutarate and inhibiting GLS may provide a therapeutic approach to suppress enchondroma and chondrosarcoma tumor growth, respectively.

Disclosures

HZ is a current employee of Amgen Inc. All other authors state that they have no conflicts of interest.

Author Contributions

Hongyuan Zhang: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; resources; validation; visualization; writing – original draft; writing – review and editing. Vijitha Puviindran: Conceptualization; data curation; investigation; resources. Puviindran Nadesan: Data curation; investigation; resources. Xiruo Ding: Data curation; formal analysis; software. Leyao Shen: Investigation. Yuning J Tang: Conceptualization; writing – review and editing. Hidetoshi Tsushima: Conceptualization. Yasuhito Yahara: Conceptualization. Ga I Ban: Conceptualization. Guo‐Fang Zhang: Conceptualization; data curation; investigation; resources; writing – review and editing. Courtney M Karner: Conceptualization; methodology; resources; supervision; writing – review and editing. Benjamin A. Alman: Conceptualization; funding acquisition; project administration; supervision; writing – review and editing.

Peer Review

The peer review history for this article is available at https://publons.com/publon/10.1002/jbmr.4532.

Supplemental Fig. S1. GLS was upregulated in chondrosarcoma tumors with IDH1 or IDH2 mutations. Relative gene expression of GLS in human chondrosarcoma patient samples with wild‐type or mutant IDH1 or IDH2. Each dot represents one patient tumor. Data are presented in violin plot. Padj<0.0001.

Supplemental Fig. S2. Glutamine was the primary source for D‐2HG production in chondrocytes with Idh1 mutation. (A) Graphical depiction of tracing glutamine metabolism using 5C5‐glutamine. Filled circles indicate 13C and open circles indicate 12C. (B) AdCre;Idh1 chondrocytes used glutamine for D‐2HG production. Percentage of 13C labeling in D‐2HG at different time points. M0, M1, …, Mn refers to the isotopologues containing n heavy atoms in a molecule. (C) D‐2HG levels in chondrocytes transfected with adenovirus GFP and adenovirus Cre. For (C), data are presented as box plot with indication of median; whiskers represent min to max values. The p value was determined by unpaired Student's t test.

Supplemental Fig. S3. Col2a1Cre;Idh1 and Col2a1Cre;Gls ;Idh1 animals displayed defects in bone growth and hypertrophic chondrocyte differentiation. (A) H&E staining of murine tibias at E16.5. (B) Quantification of tibia length based on H&E staining at E16.5. (C) Immunohistochemistry of Col X of murine tibias at E16.5. (D) H&E staining of murine tibias at E18.5. (E) Quantification of tibia length based on H&E staining at E18.5. (F) Immunohistochemistry of Col X of murine tibias at E18.5. (G) Immunohistochemistry of cleaved caspase 3 of murine tibias at E14.5. Each data point represents an individual animal. For (B) and (E), p values were determined by one‐way ANOVA, Tukey's multiple comparisons test. Result for ANOVA: (B) F = 17.00, p < 0.0001; (E) F = 17.71, p < 0.0001. Data are presented as box plot with indication of median; whiskers represent min to max values.

Supplemental Fig. S4. BPTES treatment did not cause significant change in tumor weight and cell viability of IDH1 and IDH2‐wild‐type chondrosarcoma. (A) Relative tumor weight of xenografted IDH1 and IDH2‐wild‐type chondrosarcoma tumors at the time of harvest. (B) Relative proliferation rate of each IDH1 and IDH2‐wild‐type tumor determined by percentage of KI67‐positive cells. (C) Relative apoptotic rate of each IDH1/2WT tumor determined by percentage of cleaved caspase 3‐positive cells. Each data point represents an individual tumor. Data are presented as box plot with indication of median; whiskers represent min to max values. Significance was determined by unpaired Student's t test.

Supplemental Fig. S5. Supplementing non‐essential amino acids did not recue defects in hypertrophic chondrocyte differentiation in Col2a1Cre;Gls ;Idh1 metatarsal organ culture. (A) Immunohistochemistry of Col X of metatarsal organ culture at E16.5. (B) Quantification of the length of Col X‐positive zone in metatarsal organ culture. Each data point represents an individual metatarsal. Data are presented as box plot with indication of median; whiskers represent min to max values. One‐way ANOVA, Tukey's multiple comparisons test. F = 117.8, p < 0.0001.

Click here for additional data file.