br Fig The fibroblast component increases the glycolytic phe
Fig. 1. The fibroblast component increases the glycolytic phenotype of tumors. (A) Upper panels, representative images for High and Low desmoplastic stroma in NSCLC samples. Lower panels, PGC-1α and GAPDH/COX1 mRNA levels correlate with fibrosis grade. (B) Representative graphs showing the correlation of vimentin expression with diﬀerent glucose metabolism genes. (C) Monocultured normal fibroblasts (CF1, CF2, CF3) have higher OXPHOS function than tumor cell lines.
CF1 is co-cultured with the A549 cell line.
To test if this metabolic reprogramming on H1299 1851373-36-8 was fibro-blast-specific, or if it was just produced by diﬀerent cells that cohabit in
(caption on next page)
the same space, we co-cultured the H1299 K2S cell line with A549 cells (Fig. 2C). The results showed no significant diﬀerences in both cell lines when co-cultured together, reinforcing the specific role of the
Fig. 2. The Eﬀect of fibroblast co-culture in NSCLC cell lines. (A) Representative image of the co-cultures. Upper panel, confocal image of H1299 cells co-cultured with control fibroblasts (CF1). TMRE in red, H1299 K2S positive cells in blue. Lower panel, representative flow cytometry analysis of the co-cultures. Fibroblast and tumor cells were not distinguishable by size and complexity (forward and side scatter). K2S (for H1299 cells) or EpCAM (for A549 cells) positivity was used to gate and discriminate the tumor cells from the fibroblasts (B) MIMP was evaluated as TMRE fluorescence determined by flow cytometry. Data are means relative to monocultured tumor cells. Note that there are significant diﬀerences in co-cultured cells for the H1299 but not for A549 cell line. (C) No significant diﬀerences in MIMP were seen in the co-culture between tumor cell lines. (D) The eﬀect of fibroblast co-culture on LDHA and PGC-1α mRNA levels determinated by RT-qPCR (E) Mitochondrial mass was determinated by confocal quantification of TOM20 expression. Data are means from at least three diﬀerent experiments and are represented relative to monocultured cells. Error bars indicate standard deviation. Student's t-test p-value was considered to be statistically significant when was < 0.05 (*=p ≤0.05).
fibroblasts in this process.
We measured the LDHA and PGC-1α mRNA levels after the co-culture in both, H1299 and fibroblast cells (Fig. 2D). The results showed an increase of LDHA expression in the co-cultured fibroblasts and an increase of PGC-1α in the co-cultured tumor cells supporting a meta-bolic shift towards glycolysis in the fibroblasts and a metabolic shift towards OXPHOS in the tumor cells.
The OXPHOS function modification has been associated with an increase in mitochondrial mass in co-cultured tumor cells and a de-crease of it in the fibroblasts [10,12]. Beyond the levels of PGC-1α, we used the amount of TOM20 protein to directly asses the mitochondrial mass by confocal microscopy independently of the MIMP. The quanti-fication of this protein revealed a significant increase of the mi-tochondrial mass when the tumor cells are co-cultured with control fibroblasts, however no statistically significant diﬀerences were seen on co-cultured control fibroblasts (Fig. 2E). This point out diﬀerent me-chanisms for the MIMP changes in each cell type. Thus, the MIMP rise in tumor cells may be associated to PGC-1α and mitochondrial mass increase. However, the decrease of MIMP in fibroblast occurs without any significant modification in PGC-1α or mitochondrial mass levels.
These results support the existence of a metabolic reprogramming in NSCLC cells and fibroblasts in vitro, although this is restricted to some cell lines. Furthermore, these results may indicate that this change to-wards a more glycolytic metabolism in the whole tumor in patients with increased amount of fibroblasts is caused by the increased glycolytic metabolism of CAFs. This would also reinforce why fibroblasts parti-cipate in a worse prognosis of the patient , since they not only favor immunosuppression , metastasis  or drug resistance , but also would generate a tumor mass with a greater glucose uptake, which is usually associated with tumor's aggressiveness [34,35].
The fact that one tumor cell line induces this phenomenon and the other one does not on the same fibroblast cell line speaks of some de-termining characteristic of this process present in the tumor cell line. If these diﬀerences were also to be found in patients, it would be possible to identify diﬀerent groups of patients based on their ability to induce this eﬀect on the microenvironment, which would have implications for the diagnosis, staging or new drugs development.
3.3. Fibroblast metabolic reprogramming is not dependent on α-SMA expression