LAUSR

It has become clear that cancer associated fibroblasts (CAFs), the mRNA/protein levels with pathway fluxes and phenotypes, are not almost abundant stromal cells in the tumor microenvironment, through the secretion of cytokines and other growth factors, and by increasing the estradiol levels, interact with cancer cells to help promoting tumor progression, metastasis and therapeutic resistances [1]. Like tumor cells, CAFs also undergo energy metabolism reprogramming in a process induced by hypoxia and adjacent tumor cells [2,3]. Due to their enhanced glycolysis, CAFs release large amounts of lactate to the extracellular milieu. This metabolite is rapidly internalized and transformed by adjacent tumor cells to pyruvate, which is then presumably taken up and fully oxidized by mitochondria for ATP generation via oxidative phosphorylation (OxPhos) to support the accelerated cellular proliferation and other ATP-demandingprocesses [4]. Thismetabolic interaction between cancer cells and CAFs has been called “Reverse Warburg Effect” [2], which requires functional mitochondria in cancer cells [5]. Thus, the identification of the activationmechanisms of glycolysis in CAFs is needed. Although it has been demonstrated that hypoxia and oxidative stress through HIF1-α [3] potently activate the glycolytic flux in CAFs, other factors may also regulate CAFs glycolysis under hypoxia or normoxia. Recently, in this article of EBioMedicine, Sun et al. [6] found a strong relationship between the levels of an oxidized ATM (ataxia-telangiectasia mutated) protein kinase and the glycolytic activity in CAFs. Oxidation of ATM protein kinase was triggered by hypoxiainduced oxidative stress. There are not well-defined markers that may allow distinguishing between CAFs and non-cancer associated fibroblasts (NAFs). The Sun et al. study clearly demonstrates that oxidized ATMmay be used as a specific CAFs marker. Two strengths of the Sun et al. study have to bementioned. First, the integral analysis of the CAFs energy metabolism by assessing the transcription, protein contents and pathway fluxes of both glycolysis and OxPhos. Such an approach avoids mechanistic explanations of the observed phenotype based solely on transcriptomic and proteomic data, as the actual biological function (i.e., pathway activities) is also assessed. One should be aware that strict correlations between either mRNA levels with protein levels, protein levels with enzyme activities, or