FAK modulates glioblastoma stem cell energetics via regulation of glycolysis and glutamine oxidation

Abstract

Glycolysis and the tricarboxylic acid cycle (TCA) cycle are reprogrammed in cancer cells to meet bioenergetic and biosynthetic demands, including by engagement with the extracellular matrix (ECM). However, the mechanisms by which the ECM engagement reprograms core energy metabolism is still unknown. We showed that the canonical cell-ECM adhesion protein focal adhesion kinase (FAK, also known as PTK2) and, specifically, its kinase activity, is driving cellular energetics. Using a mouse stem cell model of glioblastoma, we showed that deletion of the FAK gene simultaneously inhibits glycolysis and glutamine oxidation, increases mitochondrial fragmentation, elevates phosphorylation of the mitochondrial protein MTFR1L at serine residue 235 (S235) and triggers a mesenchymal-to-epithelial transition. These metabolic and structural changes arise through altered contractility of actomyosin, as shown by myosin light chain type II (MYL2, also known as MLC2) phosphorylated (p) at S19. This process can be reversed by Rho-kinase (ROCK) inhibitors revealing mechanotransduction pathway control of both mitochondrial dynamics and glutamine oxidation. FAK-dependent metabolic programming is associated with regulation of cell migration, invasive capacity and tumour growth in vivo. Our work describes a previously unrecognised FAK-ROCK axis that couples mechanical cues to the rewiring of energy metabolism, linking cell shape, mitochondrial function and malignant behaviour.

Keywords: Adhesion proteins; Extracellular matrix; Glutamine oxidation; Glycolysis; Mechanical forces; Mitochondria.

Fig. 1.

Genetic deletion of FAK drives phenotypic and transcriptional changes in a GBM mouse stem cell model. (A) Regularised-logarithm transformation (rlog) normalised gene expression counts of Ptk2 in NS and NPE cells. Gene expression counts obtained from Gangoso et al., (2021) from (PMID: 33857425). Error bars indicate the mean ±s.d. Statistics: unpaired two-tailed t-test (n=3). ****P<0.001. (B) Western blot showing FAK, NF1, PTEN, EGFR and EGFRvIII protein levels in NS and NPE cells. COX IV was used as a loading control. The bands for EGFR and EGFRvIII are saturated in this exposure and are shown for qualitative purposes only. (C) Western blot analysis showing protein expression of pFAK(Y397) and FAK in NPE, FAK−/− and FAK Rx cells. Cofilin was used as a loading control. (D) Cell metabolic activity measured by alamarBlue assay of FAK Rx and FAK−/− cells cultured for 3 days at normal culture conditions. Error bars indicate the mean ±s.e.m. Statistics: unpaired two-tailed t-test (n=7). ***P<0.005. (E) Representative super-resolution microscopy images of FAK Rx and FAK−/− cells showing the F-actin cytoskeleton labelled with fluorophore-conjugated phalloidin (green) and α-tubulin (orange); nuclei were stained with DAPI (blue). Image brightness and contrast were autoscaled individually to enable visualisation. Scale bars: 20 µm. (F) Gene set enrichment analysis (GSEA) plot of the ‘Hallmark epithelial-mesenchymal transition’ gene set for proteins differentially regulated in FAK Rx cells compared to FAK−/− cells (n=3). Normalised enrichment score (NES) and false discovery rate (FDR) are shown. (G) Migration speed (in μm/min) of NPE, FAK−/− and FAK Rx cells plated on laminin I and imaged every 10 min for 48 h. Each dot represents one cell, data from two independent cultures. Error bars indicate the mean ±s.d. Statistics: Kruskal–Wallis test followed by Dunn's multiple comparison test. **P<0.01, ****P<0.001. (H) Left: Representative optical sections of NPE, FAK−/− and FAK Rx cells stained with Calcein AM invading through Matrigel. Scale bars: 200 μm. Right: Plotted is the percentage of cells present in each displayed optical section of total cell area throughout the Matrigel (n=3 technical repeats). Error bars indicate the mean +s.d. Statistics: Kruskal–Wallis test followed by non-adjusted Dunn's test.

Fig. 2.

FAK promotes glycolysis, mitochondrial respiration and glutamine oxidation. (A,B) Extracellular acidification rate (ECAR) (A) and oxygen consumption rate (OCR) (B) of FAK Rx and FAK−/− cells (7000 cells per well, n=7 independent cultures, three replicate measures). Mitochondrial stress test conditions: basal, oligomycin (OLIGO), carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), a mixture of rotenone (ROT) and antimycin A (ANTI A). Error bars indicate the mean ±s.e.m. Dashed vertical lines indicate injections of oligomycin, FCCP and rotenone/antimycin A. (C) Basal ECAR vs OCR, calculated from data shown in A and B. Error bars indicate the mean ±s.e.m. (D) Bar graphs showing the quantification of mitochondrial parameters calculated from the normalized OCR data shown in B. Error bars indicate the mean ±s.e.m. Statistics: Mann–Whitney test. *P<0.05, **P<0.01. (E) Heatmap of intracellular metabolites as indicated. ATP+prec., ATP precursors. Shown is the normalised peak intensity in FAK Rx and FAK−/− cells. Each square represents one individual replicate (n=3 independent cultures per day). Statistics: unpaired two-tailed t-test. (F) Bubble plot showing glycolysis/glucogenesis and TCA cycle among the significantly enriched KEGG pathways in pathway enrichment analysis of intracellular metabolites in FAK Rx compared to FAK−/− cells. Circle sizes are proportional to the number of hits in that pathway. Darker colour represents more significance. (G,H) Atom fraction enrichment of glucose-derived ¹³C in glycolysis intermediates after incubation of FAK Rx and FAK−/− cells with ¹³C₆ D-glucose for 1 h (G) or ¹³C₅ glutamine for 3 h (H) (n=3 independent cultures on the same day). The main isotopologue of each metabolite is shown and plotted as the fraction of the sum of all isotopologues. Error bars indicate the mean±s.d. Statistics: unpaired two-tailed t-test. *P<0.05, **P<0.01, ***P<0.005.

Fig. 3.

FAK is associated with elongated mitochondria. (A) Representative super-resolution microscopy images of FAK Rx and FAK−/− cells showing mitochondrial morphology stained with MitoTracker Deep Red FM (magenta) and F-actin cytoskeleton labelled with fluorophore-conjugated phalloidin (cyan); nuclei were stained with DAPI (blue). Boxed areas in main images are shown magnified on the right. Scale bars: 20 µm (full field of view), 5 µm (insets). (B) Quantification of mitochondrial mean fragment (branch) length in FAK Rx and FAK−/− cells. n=2 independent experiments. Each dot represents a field of view. Error bars indicate the mean ±s.d. Statistics: unpaired two-tailed t-test. **P<0.01. (C) Volcano plot of post-translational modifications highlighting the differentially modified proteins in purple. p-MTFR1L S235 is labelled. n=3 independent cultures. (D) Abundance of p-MTFR1L S235 in FAK Rx and FAK−/− cells. n=7 (FAK Rx) and n=6 (FAK−/−) independent cultures. Data from two separate experiments conducted on different days are shown. Error bars indicate the mean and ±s.e.m. Statistics: unpaired two-tailed t-test. ***P<0.005.

Fig. 4.

FAK regulates mitochondrial morphology via the ROCK−pMLC2(S19) pathway. (A) Representative super-resolution microscopy images of FAK Rx, FAK−/− cells treated with or without 2 µM GSK269962A ROCK inhibitor for 24 h, showing pMLC2(S19) (yellow) and the F-actin cytoskeleton (labelled with fluorophore-conjugated phalloidin, cyan); nuclei were stained with DAPI (blue). Boxed areas in main images are shown magnified in insets on the right, showing cell−cell contacts. Scale bars: 20 µm (full field of view), 5 µm (insets). (B) Quantification of pMLC2(S19) mean fluorescence intensity at cell−cell contacts, cell edge and intracellularly. n=2 independent experiments. Each dot represents one field of view. Error bars indicate the mean ±s.d. Statistics: two-way ANOVA followed by Tukey's multiple comparison test. **P<0.01, ***P<0.005, ****P<0.001. (C) Representative super-resolution microscopy images of FAK Rx, FAK−/− cells treated with or without 2 µM GSK269962A for 24 h, showing mitochondrial morphology (stained with MitoTracker Deep Red FM, magenta) and F-actin cytoskeleton (labelled with fluorophore-conjugated phalloidin, cyan); nuclei were stained with DAPI (blue). Boxed areas in main images are shown magnified in insets on the right. Scale bars: 20 µm (full field of view), 5 µm (insets). (D) Quantification of mitochondrial mean fragment (branch) length in FAK Rx and FAK−/− cells with or without 2 µM GSK269962A. n=2 independent experiments. Each dot represents one field of view. Error bars indicate the mean ±s.d. Statistics: one-way ANOVA followed by Tukey's multiple comparison test. **P<0.01, ****P<0.001.

Fig. 5.

Treating FAK−/− cells with ROCK inhibitor GSK269962A increases glutamine oxidation and cell metabolic activity. (A) The atom fraction enrichment of glutamine-derived ¹³C in TCA intermediates in FAK Rx, FAK−/− and FAK−/− cells treated with or without 2 µM GSK269962A for 24 h followed by incubation for 3 h with ¹³C₅ glutamine-supplemented medium. The main isotopologue of each metabolite is shown and plotted as the fraction of the sum of all isotopologues. α-KG, alpha-ketoglutarate. Error bars indicate the mean ±s.d. Statistics: one-way ANOVA with a Tukey multiple test correction (n=3 independent cultures on the same day). *P<0.05, **P<0.01, ***P<0.005, ****P<0.001. (B) Top: Representative phase-contrast images of FAK Rx, FAK−/− cells treated with and without ROCK inhibitors (+2 μM GSK269962A or +20 μM Y-27632) for 3 days. Bottom: Plotted is the metabolic activity of FAK Rx, FAK−/− cells treated with and without ROCK inhibitors (+2 μM GSK269962A or 20 μM Y-27632) for 3 days measured by alamarBlue assay. DMSO, vehicle control. Error bars indicate the mean ±s.e.m. Statistics: one-way ANOVA with a Dunnett multiple test correction (n≥3 independent experiments). **P<0.01.

Fig. 6.

FAK protein expression correlates with worse overall survival in vivo and in patients diagnosed with GBM. (A) Kaplan–Meier curve of CD-1 nude mice bearing intracranial tumours from injecting FAK Rx or FAK−/− cells. Statistics: Mantel-Cox test, n=5 for each group. ***P<0.005. (B) Total body flux of CD-1 nude mice bearing intracranial tumours after injection of luciferase-expressing FAK Rx (n=9 mice) or FAK−/− cells (n=8 mice) at 7 and 17 days after injection measured by IVIS imaging. Error bars indicate the mean±s.e.m. Statistics: two-way ANOVA with Tukey post hoc adjustment for multiple comparisons. Right panel: representative in vivo bioluminescent images of CD-1 nude mice bearing tumours after injection of FAK Rx and FAK−/− cells at day 17. The heatmap superimposed over the mouse heads represents the radiance (photon/cm²/s/steradian). ***P<0.005. (C) Percentage of Ki-67+cells in sections of intracranial tumours after injection of FAK Rx (n=9 mice) or FAK−/− cells (n=8 mice) at 17 days after injection. Error bars indicate the mean ±s.e.m. Statistics: unpaired two-tailed t-test. *P<0.05. (D) Representative images of intracranial tumours stained for GFP or Ki-67, or H&E-stained, after injection of FAK Rx (n=9 mice) or FAK−/− cells (n=8 mice) at 17 days after injection. (E) Distance of GFP+ tumour cells to tumour edge in sections of intracranial tumours after injection of FAK Rx (n=4) or FAK−/− cells (n=4) at day 17 after injection. Mean and s.e.m. are shown. Statistics: unpaired two-tailed t-test. *P<0.05. (F) Kaplan–Meier curve showing the overall survival of patients diagnosed with GBM (n=98) stratified into two groups representing higher (n=50) and lower quartile (n=48) FAK protein levels measured by RPPA. Data obtained from the TCGA database were accessed using cBioPortal. Mean overall survival decreased in patients that had tumours with higher FAK protein levels. Statistics: log rank test. **P<0.01.