Functional mitochondrial respiration is essential for glioblastoma tumour growth

Abstract

Horizontal transfer of mitochondria from the tumour microenvironment to cancer cells to support proliferation and enhance tumour progression has been shown for various types of cancer in recent years. Glioblastoma, the most aggressive adult brain tumour, has proven to be no exception when it comes to dynamic intercellular mitochondrial movement, as shown in this study using an orthotopic tumour model of respiration-deficient glioblastoma cells. Although confirmed mitochondrial transfer was shown to facilitate tumour progression in glioblastoma, we decided to investigate whether the related electron transport chain recovery is necessary for tumour formation in the brain. Based on experiments using time-resolved analysis of tumour formation by glioblastoma cells depleted of their mitochondrial DNA, we conclude that functional mitochondrial respiration is essential for glioblastoma growth in vivo, because it is needed to support coenzyme Q redox cycling for de novo pyrimidine biosynthesis controlled by respiration-linked dihydroorotate dehydrogenase enzyme activity. We also demonstrate here that astrocytes are key mitochondrial donors in this model.

Fig. 1. Glioblastoma cells without mtDNA form tumours with delay.

Parental (par) and ρ⁰ GL261 cells were assessed for the level of mtDNA by qPCR (A), for the level of mtDNA-coded mtCO1 protein by WB (B), and for routine mitochondrial respiration using the Oxygraph (C). Dependency of ρ⁰ cells for exogenous uridine was confirmed by cultivation of parental and ρ⁰ cells in media with ( + uri) or without (-uri) added uridine (D). 5 × 10⁴ parental or ρ⁰ cells were implanted into the right hemisphere (coordinates: bregma 2,2) of syngeneic C57Bl/6 mice as indicated (E). Mice grafted with parental and ρ⁰ cells (n ≥ 9) were assessed for survival (F). Tumours formed from parental cells (par T) and from ρ⁰ cells (ρ⁰ T), and contralateral brain tissue were excised at the endpoint of the experiment (after detection of 15% animal weight loss and behavioural changes) and their CI- and CII-dependent respiration (CI R, CII R, respectively) was assessed by the Oxygraph (G), protein expression of subunits of respiratory complexes (CI, CII, CIII and CV) was assessed by WB (H), and gene expression of mtDNA-encoded genes by qRT-PCR (I).

Fig. 2. ρ⁰ cells acquire mitochondria with mtDNA from the host resulting in restored tumour growth.

Tumour-derived cell lines D30, D75 and D106 were isolated after 30, 75 and 106 days post grafting, respectively, by sorting GFP-positive cells from tumours formed from GFP-tagged ρ⁰ cells implanted orthotopically into C57Bl/6 mice (A). The presence of mtDNA was assessed by qPCR (B) and visualised by stimulation emission depletion (STED) microscopy using anti-Tomm20 IgG to label mitochondrial membranes and anti-DNA IgG to label mitochondrial nucleoids; images were deconvolved and contrast-adjusted (D). Level of mtCO1 was probed for in tumour-derived cell lines by WB (C). The host identity of the mtDNA was confirmed by Sanger sequencing of two mtDNA polymorphic sites in Nd3 and tRNA^Arg genes, with differences in polymorphic sites indicated by the black arrowhead (E). Tumour-derived cell lines were orthotopically injected into C57Bl/6 mice (n ≥ 8) to assess the animal survival, the ethical endpoint of the experiment being 15% animal weight loss and behavioural changes (F). EtBr – ethidium bromide.

Fig. 3. Mitochondria acquired by ρ⁰ cells in vivo recover mitochondrial cristae structure, assembly of respiratory complexes and respiration.

Parental (par) and ρ⁰ cells, and cells derived from tumours grown from ρ⁰ cells at different times post grafting were assessed for mitochondrial morphology and cristae structure by transmission electron microscopy (TEM; A) and for assembly of respiratory complexes by blue native gel electrophoresis (BNGE) and their activity by high resolution clear native electrophoresis (hrCNE; B). Mitochondrial oxygen consumption rate (OCR; C) and extracellular acidification rate (ECAR; D) were evaluated using the Seahorse Extracellular Flux Analyzer. ATP level was evaluated using the CellTiter-Glo kit with 2-deoxyglucose (2-DG; 50 mM) used to inhibit glycolysis (E). FCCP - Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone, CTRL – control.

Fig. 4. Respiration recovery by mitochondrial transfer propels DHODH-linked respiration.

Graphical representation of de novo pyrimidine synthesis pathway converting glutamine to UMP and pyrimidines via the DHODH enzyme in cells with functional electron transport chain and CoQ redox-cycling. To drive forward the conversion od dihydroorotate to orotate, DHODH has to shuffle electrons to CoQ, whose redox-cycle is maintained by functional CIII and CIV (A). mRNA expression of the de novo pyrimidine synthesis pathway enzymes was measured by qRT-PCR (B) and protein expression was evaluated by WB (C) in parental and ρ⁰ cells, and in tumour-derived cell lines. DHODH-linked respiration was evaluated using the Oxygraph (D). The CoQ₉ redox state was assessed by HPLC with electrochemical detection as a relative ratio of reduced/total CoQ₉ (E). Parental and DHODH KO GL261 cells were assessed for the level of DHODH by WB (F), and routine and DHODH-linked respiration was evaluated by the Oxygraph (G). The CoQ₉ redox state of DHODH KO cells was assessed by HPLC as a relative ratio of reduced/total CoQ₉ (H). DHODH KO cells were orthotopically injected into mice and animal survival within the ethical endpoint was evaluated (I). Mice orthotopically injected with GL261 parental cells were treated with DHODH inhibitor BAY-2402234 daily by oral gavage until ethical endpoint (15% weight loss and behavioural changes) of the experiment was reached (J). CAD - carbamoyl-phosphate synthetase 2, aspartate transcarbamoylase, and dihydroorotase, DHODH – dihydroorotate dehydrogenase, UMPS - uridine monophosphate synthase, CoQ - ubiquinone, CoQH₂ – ubiquinol, Cyt c – cytochrome c, UMP – uridine monophosphate, NAD(H) – nicotinamide adenine dinucleotide, OMM – outer mitochondrial membrane, IMM – inner mitochondrial membrane, IMS – intermembrane space, H⁺ - proton, e^- - electron, KO – knock-out, R – respiration, BAY - BAY-240223 treatment, CTRL – control, d - days.

Fig. 5. DHODH plays an important role in tumour growth.

Levels of metabolites of de novo pyrimidine synthesis pathway were determined by LC-MS/MS (A). Total level of DHO was normalised to protein content and expressed relative to parental cells (B). Stable isotope tracing of labelled glutamine was used to determine the efficiency of the de novo pyrimidine synthesis pathway and the levels of labelled UMP were normalised to protien and expressed as total relative exchange rate in each cell line (C). Levels of de novo pyrimidine synthesis pathway metabolites in the brain and parental tumour (par T) tissues, visualised by H&E staining, were detected by MALDI (D) and mean intensities were plotted into a graph (n = 2; E). Sections of brains bearing par tumours (par T) were stained against proliferation marker Ki67 and counterstained with DAPI to visualise DNA (F). OMM – outer mitochondrial membrane, IMM – inner mitochondrial membrane, IMS – intermembrane space, CAD - carbamoyl-phosphate synthetase 2, aspartate transcarbamoylase, and dihydroorotase, DHODH - dihydroorotate dehydrogenase, UMPS - uridine monophosphate synthase, UMP – uridine monophosphate, UDP - uridine diphosphate, UTP - uridine triphosphate, H&E – hematoxylin and eosin.

Fig. 6. Astrocytes are the predominant mitochondrial donors for ρ⁰ GBM cells.

To determine which brain resident cell type is the donor of mitochondria for cancer cells, in vitro co-culture experiments with two cell types isolated from adult mouse brain were carried out (A). Co-cultures of neural cells and GL261 ρ⁰ GFP⁺ cells were assessed by fluorescent confocal microscopy after 48 h (B) and by flow cytometry after 24 h (C). mtDNA levels were detected in parental and ρ⁰ cells, and in ρ⁰ cells after their 72-h co-culture with potential neural donor cells (D). Horizontal transfer of mitochondria from astrocytes in vivo was confirmed by confocal microscopy (F) using the inducible MitoTag mice model as visualised (E). CC AST – co-culture with astrocytes, CC MIC - co-culture with microglia, TAM – tamoxifen, i.p. – intraperitoneal, P – postnatal day, Aldh1l1 – aldehyde dehydrogenase 1l1.