Metabolic reprogramming in cancer contributes to malignant transformation and progression, so the tracking of metabolic features can potentially find clinical application in patient diagnosis, stratification and monitoring. During surgical collection of tumour biopsy samples or tumour resection, tissue cauterization using the surgical device known as ‘iKnife’ leads to the generation of aerosol. Using a rapid mass-spectrometry based approach to analyse this tumour-derived aerosol (in near real-time), Koundouros et al. derived metabolic signatures that were associated with the tumour tissue and its genotype. Based on these signatures, they identified a new mechanism by which oncogenic PI3K signalling promotes tumour growth.
Rapid evaporative ionization mass spectrometry (REIMS), coupled to the iKnife, was used for lipidomic analysis of tumour tissue or cancer cell pellet-derived aerosol. Analysing 43 breast cancer cell lines, 18 patient-derived xenografts (PDXs) and 12 primary breast tumours the authors derived lipidomic signatures that in turn accurately classified tumours as oestrogen receptor positive, HER2+ or triple negative.
Unsupervised hierarchical clustering of the lipidomic profiles of the range of breast cancer cell lines led to identification of two metabolic subtypes, the signatures of which were enriched or depleted of certain lipid species. Gene mutation analysis of these cell lines showed that only oncogenic mutation in PIK3CA, which encodes the alpha subunit of PI3K, was significantly overrepresented in the lipid-enriched cluster. This correlated with overexpression of genes involved in de novo lipogenesis, and exogenous lipid uptake. Using a signature derived from the enriched lipid cluster as a classifier, PIK3CA mutation could be accurately predicted in breast cancer PDXs as well as primary tumours.
For analysis of the mechanisms underpinning the lipid-enriched subtype, isogenic MCF10A PIK3CA wild-type (WT) and mutant (E545K and H1047R) cell lines (the latter of which clustered in the lipid-enriched subtype) were treated with inhibitors of several components of the PI3K pathway, including AKT and mTOR. Lipid profiles were analysed, followed by shRNA-mediated knockdown of more specific pathway components. This analysis showed that oncogenic PI3K signalling via mTOR, specifically mTOR complex 2 (mTORC2), but independently of AKT, was driving the lipid-enriched profile of PIK3CA-mutant cells.
To look into the mutant PIK3CA-specific lipid profile in more detail, the authors used REIMS to analyse fatty acids (FAs), which are the main components of lipids, in a panel of isogenic PIK3CA-mutant cells or WT cells. Apart from FAs known to be associated with lipogenesis, the most abundantly upregulated FA in PIK3CA-mutant cells was arachidonic acid (AA), a bio-active lipid that is needed for the formation of eicosanoids, such as prostaglandin E2, and is involved in pro-inflammatory processes in cancer. Upregulation of AA was also found in PIK3CA-mutant PDXs and primary breast tumours as well as in tumour samples from other tissues, compared with PIK3CA-WT samples. Importantly, AA was secreted from PIK3CA-mutant cells, but not from PIK3CA-WT cells when grown in FA-free media in vitro. Conditioned media from PIK3CA-mutant cells induced proliferation in PIK3CA-WT cells, as well as in PIK3CA-mutant cells, when compared with the respective control cells treated with lipid-deprived conditioned media, suggesting paracrine and autocrine proliferative effects of secreted AA in the tumour microenvironment.
Further analysis identified the cytosolic calcium-dependent phospholipase (cPLA2), a previously unknown substrate of protein kinase C zeta (PKCζ), as a new downstream component of mutant PI3K signalling. After testing the effect of inhibiting cPLA2 using a specific inhibitor (ASB14780) or shRNA-mediated knockdown in 2D and 3D cell cultures, the authors tested the inhibitor in PIK3CA-mutant or WT triple negative PDX-bearing BALB/c nude mice lacking adaptive immunity. The mice received a normal or isocaloric fat-free diet, and only the combination of a fat-free diet with the inhibitor led to reduction in tumour weight in PIK3CA-mutant tumour bearing mice, whereas the growth of WT-tumours (which were smaller than PIK3CA-mutant tumours) was not affected by the inhibitor and/or fat-free diet. The combination treatment also restored intratumoural levels of natural killer (NK) cells in PIK3CA-mutant tumours, suggesting that reduced NK cell infiltration in PIK3CA-mutant tumours compared with WT tumours is linked to overproduction of AA-derived eicosanoids.
“only the combination of a fat-free diet with the inhibitor led to reduction in tumour weight in PIK3CA-mutant tumour bearing mice”
With oncogenic PIK3CA known to be associated with lack of response to immune checkpoint inhibition, this study underlines the preclinical and clinical prospects of ‘metabotyping’ and identifies oncogenic PI3K-driven AA production as a druggable pathway that directly links the tumour genotype, metabolic phenotype and nutrient availability in tumours.
Koundouros, N. et al. Metabolic fingerprinting links oncogenic PIK3CA with enhanced arachidonic acid-derived eicosanoids. Cell 181, 1596–1611 (2020)
Hoxhaj, G. & Manning, B. D. The PI3K–AKT network at the interface of oncogenic signalling and cancer metabolism. Nat. Rev. Cancer 20, 74–88 (2020)
About this article
Cite this article
Harjes, U. Metabolites going up in smoke. Nat Rev Cancer 20, 482 (2020). https://doi.org/10.1038/s41568-020-0289-3