Fig 2: Glycogen utilization fuels glycolysis and oxidative phosphorylation in mutant IDH chondrosarcomas and promotes tumor cell viability, proliferation, and survival(A) Glycogenolysis enzymic activity (PYGL) was inhibited with CP-91149 drug. Increasing concentrations of CP-91149 induced sensitivity to primary patient mutant IDH chondrosarcoma cells as displayed by the dose-response curve (n = 3).(B) 82.5 µM CP-91149 reduced cell viability in 4 patient cell lines (n = 3).(C) Glycogenolysis blockade reduced BrdU incorporation (cell proliferation) in 4 out of 5 cell lines (n = 3).(D) Glycogenolysis blockade induced apoptosis in 3 out of 4 cell lines (n = 3). Arrows indicate TUNEL-positive stained cells. 103 magnification.(E and F) PYGL blockade reduced secreted (n = 3) (E) and intercellular (F) lactate levels, suggesting that glycogen utilization fuels glycolysis (n = 3).(G) PYGL blockade reduced glycolytic (n = 3) (G) and oxidative (H) capacities of chondrosarcoma cells (n = 3). p value = Student’s t test p < 0.05, an asterisk (*) indicates that significant p values are shown. Means and error bars representing standard deviations are shown. Scale bars: 100 µm in white.

Fig 3: HIF1a transcriptionally regulates glycogen metabolism in mutant IDH chondrosarcomas(A) Analysis of published gene expression data shows that HIF1a gene expression levels are elevated in mutant IDH1 chondrosarcomas (n = 116) compared with non-mutant tumors (n = 28).(B) Mutant IDH1 JJ012 cells treated with 0.5 mM DMOG and 0.25 µM digoxin stabilized and depleted HIF1a, respectively. DMOG treatment displayed elevated protein levels of HIF1a and total PYGL, and digoxin treatment displayed depleted protein levels of HIF1a, total PYGL, pPYGL, GYS1, and IDH1. Quantitative densitometry analysis of relative protein expression levels are displayed beneath respective protein bands. Protein density measurements are displayed as relative measurements by normalization to ß-actin protein density measurements compared with fold change to DMSO control treatment.(C) Primary patient chondrosarcoma cells treated with 0.5 µM digoxin display a reduction in PDK1 and VEGF levels, confirming that HIF1a is depleted in cells (n = 4). Relative gene expression for four patient-derived cell cultures compared to DMSO-treated control cultures (normalized to ß-actin).(D) HIF1a depletion is sufficient to knock down gene expression levels of glycogen metabolism genes, GYS1, PYGL, PPP1R3C, and IDH1, suggesting that HIF1a is a regulator of glycogen metabolism (n = 4). Relative gene expression for four patient-derived cell cultures compared to DMSO-treated control cultures (normalized to ß-actin). p value = Student’s t test p < 0.05, an asterisk (*) indicates that significant p values are shown. Means and error bars representing standard deviations are shown.

Fig 4: Glycogen levels are elevated in mutant IDH chondrosarcoma patient tissue and in the mutant Idh1-KI R132Q fetal growth plate(A) Transmission electron microscopy (TEM) images of IDH non-mutant chondrosarcoma cells (n = 4) in vitro display absent glycogen granules. Organelles distinct from glycogen granules shown in magnified insets labeled as follows: M, mitochondria; L, lysosome; ER, rough endoplasmic reticulum. TEM images: 7× magnification, inset images: 70× magnification.(B) TEM images of mutant IDH chondrosarcoma cells (n = 7) display glycogen pools, asterisks denote aggregates of glycogen pools in mutant IDH cells, and arrows in magnified insets indicate glycogen pools. Glycogen appears as closely packed circular granules in mutant IDH chondrosarcoma patient cells. Images: 7× magnification, inset images: 70× magnification.(C) Glycogen quantification from pulverized patient-derived xenograft chondrosarcoma tissues display an elevation of glycogen in mutant IDH1 (n = 17) and IDH2 (n = 8) tumors compared with non-mutant tumors (n = 8). One-way ANOVA confirms significant statistical difference of glycogen levels between tumor genotypes (F(2,30) = 6.150, p = 0.0058). Tukey’s multiple comparisons test indicates that the mean values of glycogen in mutant IDH1 (p = 0.0107) and mutant IDH2 groups (p = 0.0109) were significantly higher than in non-mutant tumors.(D) PAS staining identified glycogen deposits in mutant IDH1 (n = 4) and IDH2 (n = 4) patient chondrosarcomas compared with non-mutant tumors (n = 5). PAS-D staining displays dissolution of glycogen deposits in mutant IDH tumors, thus confirming the presence of glycogen deposits. Arrows in magnified insets indicate glycogen deposits in cytoplasm of cells. Images: 40× magnification, inset images: 60× magnification.(E) Quantification of PAS-stained area (µm2), normalized to total number of cells, and of PAS-D-stained area shows an elevation of glycogen deposits in mutant IDH1 (n = 4) and IDH2 (n = 4) chondrosarcomas compared with non-mutant tumors (n = 5). One-way ANOVA confirms significant statistical difference of glycogen levels between tumor genotypes (F(2,10) = 7.537, p = 0.0101). Tukey’s multiple comparisons test indicates mean values of glycogen in mutant IDH1 (p = 0.0380) and mutant IDH2 groups (p = 0.0123) were significantly higher than non-mutant tumors.(F) Glycogen deposits in Col2a1Cre; Idh1LSL/wt E18.5 growth plates (n = 9) shown by PAS staining. Arrows in magnified insets indicate glycogen in cell cytoplasm. Glycogen deposits minimally in Col2a1Cre; Idh1wt/wt control growth plates (n = 8).(G) Quantification of PAS-stained area (µm2), normalized to total number of cells, and PAS-D-stained area (n = 8).(H) GYS1 staining is elevated in Col2a1Cre; Idh1LSL/wt growth plates compared with Col2a1Cre; Idh1wt/wt growth plates (n = 5).(I) Quantification of GYS1 staining from hypertrophic to resting zones (n = 5).(J) PYGL staining is unchanged in Col2a1Cre; Idh1LSL/wt and Col2a1Cre; Idh1wt/wt growth plates (n = 6).(K) Quantification of PYGL staining from hypertrophic to resting zones (n = 6).(L) Gene expression levels of glycogen genes are elevated upon Idh1 mutation induced by adenovirus Cre recombinase transfection (n = 3). Relative gene expression compared with adenovirus GFP control group was calculated and normalized to ß-actin using the 2-??Ct method. One-way ANOVA with Tukey’s multiple comparisons test = p < 0.05. p value = Student’s t test p < 0.05, an asterisk (*) indicates that significant p values are shown. Means and error bars representing standard deviations are shown. Scale bars: 2 µm in white, 100 µm in black, and 10 µm in blue. Magnification: whole growth plate images: 13× magnification, inset images: 60× magnification.

Fig 5: HIF1? regulates glycogen metabolism in Idh1-KI chondrocytes and in the fetal growth plate(A) HIF1? staining is elevated in Col2a1Cre; Idh1LSL/wt E18.5 growth plates compared with Col2a1Cre; Idh1wt/wt growth plates (n = 3). Inset images taken at 20× magnification at the proliferative zone of the growth plate.(B) Quantification of HIF1? staining from hypertrophic to resting zones (n = 3).(C) GLUT1 staining is elevated in Col2a1Cre; Idh1LSL/wt growth plates compared with Col2a1Cre; Idh1wt/wt growth plates (n = 3). Inset images taken at 20× magnification at the proliferative zone of the growth plate.(D) Quantification of GLUT1 staining from hypertrophic to resting zones (n = 3).(E) Gene expression levels of HIF1? target genes are significantly reduced upon Idh1 mutation and HIF1? knockdown induced by adenovirus Cre recombinase transfection (n = 3).(F) Gene expression levels of glycogen metabolism genes are significantly reduced upon Idh1 mutation and HIF1? knockdown (n = 3).(G) Ex vivo explant cultures of E16.5 Col2a1Cre-ERT; Idh1LSL/wt metatarsals display reduced glycogen (PAS) and HIF1? levels upon pharmacological blockade of HIF1? by 0.5 µM digoxin treatment for 4 days (n = 6). Inset images are taken at 15× magnification at the central area of the growth plate. p value = Student’s t test p < 0.05, an asterisk (*) indicates that significant p values are shown. Means and error bars representing standard deviations are shown. Scale bars: 100 µm in black, 50 µm in blue, and 20 µm in white.