• 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br Results and discussion br The fibroblast component increa


    3. Results and discussion
    3.1. The fibroblast component increases the glycolytic phenotype of tumors
    As a first approach to study the role of the fibroblast component of tumors, we studied the correlation of PGC-1α (master regulator of mitochondrial metabolism) and the GAPDH/MT-CO1 ratio mRNA le-vels with the fibrosis grade in a cohort of 12 stage IIIA lung adeno-carcinoma patients (Fig. 1A). The results showed significant differences in PGC-1α (p=0.048) and GAPDH/MT-CO1 levels (p=0.006) between high and low fibrosis grade in the tumor biopsies. Interestingly, an increase in fibroblast component present in the tumor leads to higher levels of the ratio GAPDH/MT-CO1 and lower levels of PGC-1α, which are typical of an increased glycolytic metabolism and increased PET 
    To confirm these results, expression data from 73 patients with stage IIIA pulmonary adenocarcinoma present in the TCGA database were analyzed using the bioinformatic tool “R2: Genomics Analysis and Visualization Platform”. The correlations between vimentin levels (as marker of the fibroblast component) with the levels of genes related to glucose metabolism (Glycolysis, TCA and OXPHOS system) were ana-lyzed. The results show a similar change to that described in our cohort of patients (Fig. 1B and Supplementary Table 1). Overall, the results show a positive correlation of vimentin levels with genes related to an increase in glycolytic metabolism (GLUT3, HIF1A, HK3, ENO, PFKFB4…). On the other hand, genes related to the Krebs Marizomib (FH, MPC2, IDH3A, CS, SUCLG1…) or related to OXPHOS complexes (ATP5G3, NUDFA8,9,10, SDHA,C, COX5A…) and their assembly fac-tors (COA5, COA6, NDUFAF5, UQCC2…) negatively correlate with vi-mentin levels in the tumor. Interestingly, some of these genes have been described in CAFs reprograming. The downregulation of IDH3A (Iso-citrate dehydrogenase 3α) has been pointed out as responsible for the metabolic switch in CAFs trough HIF1α stabilization [26]. Remarkably, the only OXPHOS subunits that correlate positively with the vimentin levels are NDUFA4L2, COX7A1 y COX7B2, being the first two subunits related to hypoxia [27–29] and the third of unknown function. r> It seems clear that an increase in fibroblast component in the tumor produces an increase in glycolytic genes with a decrease in OXPHOS genes. However, in literature, the glucose metabolism of differentiated cells such as normal fibroblasts is more oxidative than the metabolism of proliferative tumor cells [2].
    Our results evaluating the metabolism of monocultured tumor cells (A549 and H1299) and normal fibroblasts confirm this. The mi-tochondrial inner membrane potential (MIMP) and the galactose to glucose growth ratio were higher for the fibroblast cell lines compared to the tumor cell lines A549 and H1299 (Fig. 1C).
    Thus, despite the behavior of the cells in monoculture, the results on the tumor biopsies showed another role for the fibroblasts. Both results would indicate a possible effect of metabolic reprogramming towards glycolytic metabolism in the CAFs of NSCLC, that would modify the PET images [30]. However, we cannot rule out whether this relationship is produced by a specific metabolic change of tumor cells that would be more glycolytic in response to the fibroblasts present in the micro-environment, if it is the presence of metabolically reprogramed fibro-blasts in the sample with a more glycolytic metabolism, or even a combination of both theories.
    3.2. The in vitro co-culture of fibroblasts and NSCLC cells induce a metabolic reprogramming in both cell types
    To clarify these issues and to deepen the mechanisms of this me-tabolic relationship in NSCLC we decided to move on to an in vitro model of co-cultured tumor and fibroblasts cell lines (Fig. 2A).
    To identify the A549 cells in the co-culture analysis, we exploited their positivity for the epithelial membrane marker EpCAM, which ef-ficiently discriminates these cells from the accompanying fibroblasts through antibody labeling. On the contrary, since the H1299 line has lost many of the epithelial markers it is not possible to distinguish it from the co-cultured fibroblasts. Thus, we transfected the H1299 cell line to stably express the far-red fluorescent protein Katushka2S [22]. This strategy allowed us to identify the different cell lines when co-cultured with fibroblasts and measure other parameters, both by flow cytometry and confocal microscopy (Fig. 2A).
    After 4 days of co-culture with CF1 (Control Fibroblast 1), the MIMP was measured. MIMP is one of the most representative parameters modified in the metabolic reprogramming of co-cultured cells [10]. Surprisingly, we only saw changes in MIMP for the H1299 cell line, but no changes were observed for A549 cells when co-cultured (Fig. 2B). Furthermore, when CF1 was co-cultured with the H1299 cell line a decrease in the MIMP was observed, effect that does not occur when