Glioblastoma (GBM) metabolism has traditionally been characterized by a primary dependence on aerobic glycolysis, prompting the use of the ketogenic diet (KD) as a potential therapy. In this study we evaluated the effectiveness of the KD in GBM and assessed the role of fatty acid oxidation (FAO) in promoting GBM propagation. In vitro assays revealed FA utilization throughout the GBM metabolome, and growth inhibition in nearly every cell line in a broad spectrum of patient-derived glioma cells treated with FAO inhibitors. In vivo assessments revealed that knockdown of carnitine palmitoyltransferase 1A (CPT1A), the rate limiting enzyme for FAO, reduced the rate of tumor growth and increased survival. However, the unrestricted ketogenic diet did not reduce tumor growth, and for some models significantly reduced survival. Altogether, these data highlight important roles for FA and ketone body metabolism that could serve to improve targeted therapies in GBM.

Gliomaspheres were dissociated into single cells with Accumax™ and 2x10⁵ cells were cultured for 48 hours in 2.5 mM glucose media in triplicate for each sample. Cells were then rinsed with PBS and re-plated in 2.5 mM glucose media, either unlabeled or containing 200 μM fully-labeled ¹³C-palmitic acid for an additional 48 hours. Cells were then centrifuged and rinsed with 1ml ice-cold 150 mM ammonium acetate (pH 7.3). Centrifugation was performed again and 1ml of ice-cold 80% methanol was added. Cells were transferred to an Eppendorf tube, and 10 nmol norvaline (Sigma-Aldrich, N7502) was added to each sample. Samples were centrifuged for 5 min at top speed and the supernatant was transferred into a glass vial. Samples were resuspended in 200 μL cold 80% methanol, followed again by centrifugation, after which the supernatant was added to the glass vial. Samples were dried in an EZ-2Elite evaporator. The remaining pellet was resuspended in RIPA buffer and a Bradford assay was performed to quantify total protein concentration for sample normalization. Dried metabolites were resuspended in 50% ACN and 5 μL loaded onto a Luna 3 μm NH2 100 A (150 × 2.0 mm) column (Phenomenex). The chromatographic separation was performed on an UltiMate 3000 RSLC (Thermo Scientific) with mobile phases A (5 mM NH4AcO pH 9.9) and B (ACN) and a flow rate of 200 μL/min. The gradient from 15% A to 95% A over 18 min was followed by 9 min isocratic flow at 95% A and re-equilibration. Metabolite detection was achieved with a Thermo Scientific Q Exactive mass spectrometer run in polarity switching mode (+3.5 kV/− 3.5 kV). TraceFinder 4.1 (Thermo Scientific) was used to quantify the area under the curve for metabolites by using accurate mass measurements (< 3 ppm) and the retention time of purchased reference standards. Relative amounts of metabolites were calculated by summing up all isotopologues of a given metabolite and normalized to cell number. Correction for naturally occurring ¹³C as well as calculation of fractional contributions and clustering analyses were done in R. Fractional contribution was calculated as , where n is the number of carbons in the metabolite,  is the iteration of each possible ¹³C-labeled carbon, and  is the relative abundance of the  isotopologue. The relative amount is calculated as the sum of all isotopologues of each metabolite normalized to total protein.

Description of Columns:

Name: the name of compound

KEGG.ID: KEGG ID

Condition: experimental condition (sgRNA/sgRNA CNTL = Control guide RNA; ETOM = 100uM etomoxir; CPT1A CRISPR = guide RNA targeting CPT1A)

Iso: isotopologue, the number of labeled carbons

Nr.C: number of carbons in compound

Exp1, Exp2, Exp3: techical replicate measurements

Norm_Av: normalized average value of Exp1, Exp2, Exp3

Norm_Std: normalized standard deviation of Exp1, Exp2, Exp3

CV: coefficient of variation (ie. SD / mean) of Exp1, Exp2, Exp3

Av: Average value of Exp1, Exp2, Exp3

MID1, MID2, MID3: mass isotopologue distribution (i.e. the percent contribution of each isotopologue to the pool of a given compound)

ANOVA: ANOVA p-value compared across experimental conditions (sgRNA CNTL, sgRNA + ETOM, CPT1A CRISPR)

Sig: degree of significance (" " = p>0.05; "*" = p<0.05; "**" = p<0.01; "***" = p<0.001).