General research focus
Hypoxia is a cancer hallmark that imposes a selection pressure on tumor cells. It promotes (i) the uncoupling between glycolysis and oxidative phosphorylation (OXPHOS) together with an increased glycolytic flux in order to optimize oxygen-independent energy production and cell proliferation (the glycolytic phenotype), (ii) the emergence of tumor cells able to directly or indirectly stimulate the formation of new blood vessels from pre-existing ones (the angiogenic phenotype), and (iii) tumor cell dissemination (the metastatic phenotype). While these adaptations are usually studied independently of each other, our research strategy is to consider a necessary coordination between them during tumor development. The general hypothesis is that the glycolytic switch, initially proceeding through inhibition of the Pasteur effect, precedes and favors angiogenesis and tumor metastasis, therefore potentially making of antimetabolic drugs bona fide anti-angiogenic and antimetastatic treatment options.
Main recent achievements
Hypoxic tumor cells use glycolysis and release lactic acid, creating a gradient of lactate that mirrors the oxygen gradient in tumors. Although lactate is generally considered as a waste product, we found that it is a prominent substrate that fuels the oxidative metabolism of oxygenated tumor cells. In the tumor symbiont, preference for lactate by oxidative tumor cells renders glucose available for glycolytic tumor cells. We showed that this exquisite cooperation can be disrupted therapeutically: inhibition of monocarboxylate transporter 1 (MCT1) that we identified as the prominent path for lactate uptake induces a switch from lactate-fueled respiration to glycolysis in oxygenated tumor cells. As a consequence, hypoxic/glycolytic tumor cells die from glucose starvation. We further exploited tumor reoxygenation upon MCT1 inhibition to target the remaining oxidative tumor cells with radiotherapy. MCT1 inhibition and X-ray radiotherapy exerted more than additive effects resulting in sustained tumor shrinkage. Validation of this new therapeutic strategy using three different tumor models and the robust expression of MCT1 that we found in an array of primary human tumors provide clinical significance to anticancer MCT1 inhibition.
In addition of being a metabolic fuel for oxidative tumor cells, we found that lactate stimulates paracrine vascular endothelial growth factor (VEGF) signaling as a consequence of the activation of the transcription factor hypoxia-inducible factor-1 (HIF-1) in tumor cells (increased VEGF production) and in endothelial cells (stimulation of VEGF-receptor 2 expression). It also triggers autocrine pro-angiogenic signaling in endothelial cells through basic fibroblast growth factor (bFGF, indirectly controlled by HIF-1) and interleukin-8 (IL-8, a NF-kB target gene product). Importantly, we found that the in vitro and in vivo pro-angiogenic activities of lactate can be blocked with MCT1 inhibitors in a new anticancer treatment modality, whereas they can be exploited to accelerate wound healing and to prevent muscular atrophy in injured animals. While MCT1 inhibition combines antimetabolic and anti-angiogenic activities within a same drug, we conversely showed that the FDA-approved polymer poly-D,L-lactide-co-glycolide (PLGA) can be used as a 10-mg implant in mice to accelerate wound healing. Our data collectively support the pro-angiogenic use of PLGA to treat wounds and pathologies characterized by a deficient vasculature, and the anti-angiogenic use of MCT1 inhibitors to treat cancer, with particular caution for patients with vascular disorders or at risk for heart infarction or stroke.
Targeting MCT1 or its chaperon protein CD147/basigin, itself involved in the aggressive malignant phenotype, is an attractive therapeutic option for cancer, but basic information is still lacking regarding the regulation of the expression, interaction and activities of both proteins in tumors. While inspecting changes of MCT1 and CD147 expression in response to parameters typical of the tumor microenvironment, we found that glucose withdrawal dose-dependently upregulates MCT1 and CD147 protein expression and their interaction in oxidative tumor cells. We further identified that the stabilization of MCT1 and CD147 proteins is posttranscriptional and depends on mitochondrial impairment and the associated generation of reactive oxygen species. When glucose was a limited resource (a situation occurring naturally or during the treatment of many tumors), MCT1-CD147 heterocomplexes formed preferentially in cell protrusions. It endowed oxidative tumor cells with increased migratory capacities towards glucose. Migration was inhibited by providing an alternative oxidative fuel to glucose-starved cells or by targeting MCT1 using RNA interference or pharmacological inhibition. While our study identifies the mitochondrion as a glucose sensor promoting tumor cell migration, MCT1 is also revealed as a transducer of this response, providing an additional rationale for the use of MCT1 inhibitors in cancer.
Cancers evolve a subpopulation of tumor cells that metabolically rely on glycolysis uncoupled from oxidative phosphorylation irrespectively of oxygen availability (aerobic glycolysis). Given that most metastases are abnormally avid for glucose (which is the rationale for their clinical detection using FDG-PET) and because clinical data show a positive correlation between lactate production and tumor metastasis, we reasoned that cells performing aerobic glycolysis could constitute a population of metastatic progenitor cells that would remain glycolytic in the blood stream. We found a different metabolic phenotype, though. Indeed, using serial rounds of in vitro selection of highly invasive tumor cells (starting from wild-type SiHa human cervix adenocarcinoma cells) and in vivo selection of supermetastatic tumor cells (starting from B16-F10 mouse melanoma cells), we identified a mitochondrial switch corresponding to an overload of the TCA cycle with preserved mitochondrial functions (including ATP production) but increased mitochondrial superoxide production. The switch provided a metastatic advantage which was phenocopied by moderate OXPHOS inhibition associated with mild mitochondrial superoxide increase. Both conditions involved as downstream pathways Src activation in mitochondria and protein tyrosine kinase PTK2B/Pyk2 induction. Thus, two different events, OXPHOS overload or moderate OXPHOS inhibition, promote superoxide-dependent tumor cell migration, invasion, clonogenicity, and metastasis; demonstrating the central role of mitochondrial superoxide generation in the pathogenesis of metastasis. Consequently, we report that mitochondria-specific superoxide scavenging (using mitoTEMPO or mitoQ) inhibits metastatic dissemination from primary mouse and human tumors, which opens a new avenue for the therapeutic prevention of tumor metastasis.
