o-Phenanthroline We next investigated whether pharmaceutical

We next investigated whether pharmaceutical inhibition of the PI3K pathway was synergistic with FGFR inhibition using BKM120, a pan-PI3K inhibitor with modest antitumor activity in cancer patients, as a single agent. We found that RT112 (- translocation) and JMSU1 ( amplification) were not significantly inhibited by BKM120 or AZD4547 monotherapy. However, synergy was observed with the combination in long-term colony formation assays ( A and B), incucyte proliferation assay (Supplementary Figs 1A and 1B), and synergy assays (Supplementary Fig. 1C). Moreover, biochemical analysis indicated that the combination resulted in the induction of cleaved poly (adenosine diphosphate-ribose) polymerase, a hallmark of apoptosis (Figs. 2C and 2D). To expand our results to another cancer type with alterations, we tested this drug combination in two independent amplified lung cancer cell line models: H520 (squamous cell carcinoma) and DMS114 (small cell carcinoma) . As observed in the UCC lines, these two lung cancer cell models showed apoptosis in response to the drug combination, but not to either drug alone (Supplementary Figs. 1D–G). Of note, similar results were recently obtained for ponatinib in combination with mechanistic target of rapamycin inhibition in NSCLC o-Phenanthroline . To address the molecular mechanism underlying the synergy between PI3K and FGFR inhibition, we further analyzed signaling pathways in AZD4547-treated RT112 cells using receptor tyrosine kinase phosphorylation blots (E). We found that FGFR inhibition induced a feedback mechanism to activate Erb-B2 receptor tyrosine kinase 3 and to a lesser extent EGFR. As a consequence, PI3K-AKT signaling was activated, thereby presumably blunting the effects of FGFRi (F). Feedback activation of receptor tyrosine kinases (EGFR and ligand-associated Erb-B2 receptor tyrosine kinase 2/Erb-B2 receptor tyrosine kinase 3) , has been previously reported. Our findings provide a strong rationale that feedback signaling converges on the PI3K pathway, but not on the BRAF-MEK-ERK pathway (F), and resistance could be counteracted by PI3K inhibitors, in order to enhance the FGFRi sensitivity. To test the combination treatment , we engrafted RT112 bladder cancer cells into NMRI-nu mice. When tumors reached approximately 100mm, mice were randomized into different cohorts and treated with vehicle, AZD4547, BKM120, or the drug combination. As shown in G, treatment with the single drugs AZD4547 or BKM120 resulted in limited tumor growth inhibition. However, treatment with the combination of AZD4547 and BKM20 resulted in persistent suppression of tumor growth throughout the duration of the experiment. Immunohistochemistry staining of the tumors at the end of the experiment (Supplementary Fig. 1I) shows the combination suppressed tumor proliferation (Ki67) and inducted apoptosis (cleaved caspase 3). In two -activated cell lines, MGHU3 and SW780, synergy was not observed (Supplementary Figs. 1H and 1J). MGHU3 cells carry an activating mutation in the (E17K) gene, causing pathway activation downstream of . Indeed, these cells were highly sensitive to mechanistic target of rapamycin inhibition (AZD8055). This finding suggests that comprehensive genetic profiling remains relevant before initiating combination treatment, in order to establish the best possible treatment. SW780 cells carry a translocation, . This translocation has not been found in other cancers and the mechanism of FGFR activation is unknown, though cells appear to be dependent on FGFR signaling . Other resistance pathways could therefore be active in this cell line. In conclusion, our data show that resistance pathways to FGFR inhibition often converge on the PI3K pathway and provide a strong rationale to treat -altered UCC with a combination of FGFR and PI3K inhibitors (H). Mutations in the gene, which can cooccur with FGFR mutations, would also be targeted by this strategy. These results may apply to other cancer types as well, for example squamous NSCLC.