Phenotypic plasticity and metabolic dysregulation are two well characterised hallmarks of cancer1, yet how these two phenomena interact to drive tumour progression remains less well known. A central question is whether changes in metabolism are simply by-products of the transcriptional changes associated with tumourigenesis, or, whether alterations in metabolism can also drive transcriptional variation and cell plasticity. Using novel machine learning techniques, we defined five cell states describing plasticity across primary and metastatic triple-negative breast cancer. Intriguingly, two states had highly divergent metabolism, one relying on glycolysis and another on oxidative phosphorylation.
We next aimed to investigate the temporal evolution of these cell states across primary tumour to metastatic evolution. To address this question, we leveraged metabolomic glucose tracing techniques in primary lesions, early metastases and late metastases to understand how metabolism supported early metastatic seeding. This data revealed a significant association between a number of key glycolytic metabolites and newly established metastatic tumours.
In order to target this tumour initiating, glycolytic phenotype, we investigated one of the most differentially expressed genes this cell state, glucose transporter-3 (GLUT3). Initial analysis validated an association between GLUT3 and the tumour-initiating cancer stem cell state. Functional knockdown assays were performed to identify the function of GLUT3, where knockdown resulted in decreased proliferation, migration and tumoursphere formation, suggesting reduced self-renewal capacity.
Mechanistically, GLUT3 knockdown led to a metabolic shift, reducing glucose uptake and glycolysis whilst also increasing lactate levels. Recently, a novel, lactate derived post-translational modification, lactylation has been described2,3. We hypothesised that the differential production of lactate may correlate with varied levels of lactylation and downstream changes in gene expression that may underlie cell state changes. Accordingly, we demonstrated that GLUT3 knockdown led to a reduction in overall lactylation levels. Pharmacological inhibition of lactylation led to a shift in key cancer stem cell markers, suggesting a potential mechanism where increased glycolysis assists in maintaining the aggressive tumour-initiating cancer stem cell state.
These data indicate metabolic regulation of cancer cell plasticity via a novel metabolic-epigenetic mechanism. This highlights the therapeutic potential of metabolic targets in limiting cancer cell plasticity, leading to reduced tumour progression.