FGFR blockade inhibits targeted therapy-tolerant persister cells in basal FGFR1 and FGF2 high expressing cancers with driver oncogenes

doi:10.21203/rs.3.rs-2357127/v1

Download PDF

Article

FGFR blockade inhibits targeted therapy-tolerant persister cells in basal FGFR1 and FGF2 high expressing cancers with driver oncogenes

https://doi.org/10.21203/rs.3.rs-2357127/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 25 Oct, 2023

Read the published version in npj Precision Oncology →

You are reading this latest preprint version

Cancer cell resistance arises when tyrosine kinase inhibitor (TKI)-targeted therapies induce a drug-tolerant persister (DTP) state with growth via genetic aberrations, making DTP cells potential therapeutic targets. We screened an anti-cancer compound library and identified fibroblast growth factor receptor 1 (FGFR1) promoting alectinib-induced ALK fusion-positive DTP cell's survival. FGFR1 signaling promoted DTP cell survival generated from basal FGFR1- and FGF2-high expressing cells, following alectinib treatment, which is blocked by FGFR inhibition. The hazard ratio for progression-free survival of ALK-TKIs tended to increase in ALK fusion-positive non-small cell lung cancer patients with FGFR1- and FGF2-high expression. Combination of FGFR and targeted TKIs enhanced cell growth inhibition in FGFR1- and FGF2-high expressing cells with ALK fusion, HER2 amplification, and EGFR or BRAF mutations. Initial dual blockade of FGFR and various driver oncogenes based on FGFR1 and FGF2 expression levels before starting treatment would be a potent treatment strategy to prevent intrinsic resistance to targeted TKIs through DTP cells.

Biological sciences/Cancer/Lung cancer

Biological sciences/Drug discovery/Target identification

Oncogene addiction is a well-established paradigm in which a single gain-of-function oncogene sustains the growth and survival of cancer cells; targeting oncogene addiction by protein or pathway inhibition has great potential for use in cancer therapies¹. Clinically actionable targets and active inhibitors against targets in cancer cells have been demonstrated², and small molecule inhibitors have been approved by the US Food and Drug Administration (FDA) for use against tumor cells with addicted oncogenes, including HER2, ALK, BRAF, and EGFR³. However, despite the response to these inhibitors, resistant tumor cells develop due to on-target alterations, bypass alterations, or alterations that cause phenotypic changes, loss of target, and target dependency⁴.

One process by which these three resistance mechanisms are acquired is entry into a reversible slow proliferation state known as the drug-tolerant persister (DTP) state by a small population of cancer cells⁵; DTP cells evade cell death induced by targeted therapy long enough to acquire additional multiple resistance mechanisms⁶. A stepwise accumulation of genetic mutations has been observed in resistant tumors during sequential therapy with several ALK-TKIs^7,8. Furthermore, both on- and off-target mechanisms have been identified after acquired resistance even in the same patient⁹.

Multiple resistance mechanisms may coexist in tumors, making it difficult to devise second-line therapies and increase treatment response. Thus, more efficacious targeted therapies should be implemented up-front to prevent the emergence of additional resistance mechanisms rather than individually selecting drugs corresponding to each mechanism. However, even at the early stage of targeted TKI monotherapy, potentially complex mechanisms exist conferring a DTP state⁶. Thus, combinatorial approaches may be effective in initial treatments, and it is essential to identify the mechanisms by which a targeted TKI-induced DTP state is acquired and select a suitable combination treatment that suppresses the generation of DTP cells to maximize expected outcomes. Despite accumulating evidence for the presence of DTP cells in vitro⁶, the mechanisms by which cancer cells modulate targeted TKI tolerance in DTP states remain poorly understood.

Therefore, we explored novel candidates that act synergistically with ALK-TKI by screening an anti-cancer compound library. Additionally, we tested oncogene-targeted TKIs with EGFR, HER2, and BRAF inhibitors using EGFR-mutant (EGFR+) NSCLC, HER2-amplified (HER2+) breast cancer (BC), and BRAF-mutant (BRAF+) melanoma cells¹⁰. To increase the applicability to clinical practice, we selected FDA approved agents for use in this study³.

Compound screening identifies FGFR1 as a candidate promoting alectinib-induced DTP cells

To find novel targets that promote cell survival against ALK inhibition in ALK + NSCLC cells, we generated DTP cells that survive targeted therapy through reversible and non-mutational mechanisms from ALK + NCI-H2228 cells following exposure to alectinib¹¹. An anti-cancer compound library was screened in NCI-H2228 parental cells or DTP cells. We calculated the ratio of anti-proliferative effect on DTP cells to parental cells to identify novel candidates specifically required for cell survival against alectinib treatment, and nine compounds showed ratios of less than 0.7 with no suppression of parental cell growth (Appendix 1 and SFig. 1). To confirm the results of this screen, a cell proliferation assay was performed using these nine compounds to determine the IC₅₀ values. Erdafitinib showed the strongest inhibitory effect on DTP cells as compared to parental cells, with IC₅₀ values of 6.7 nM and 6136.4 nM in DTP and parental cells, respectively (SFig. 2A and 2B). Erdafitinib is a pan-FGFR-TKI that inhibits all members of the FGFR family, including FGFR1, FGFR2, FGFR3, and FGFR4^12,13. An analysis of mRNA expressions of these FGFR members in parental cells revealed the highest expression of FGFR1, which was greater than 1.00 transcript per million (TPM) (SFig. 2C). Therefore, we examined whether FGFR1 contributed to cell survival in the presence of alectinib in NCI-H2228 cells.

DTP cells escape ALK-TKI-induced cell death through activation of FGFR signaling

To determine whether pharmacologic inhibition of FGFR1 sensitizes ALK + NCI-H2228 DTP cells, we assessed sensitivity to pan-FGFR-TKI (BGJ398). Parental cells were sensitive to ALK-TKIs (alectinib and lorlatinib) and insensitive to BGJ398 (IC₅₀ values of 219.6 nM for alectinib and > 1000 nM for BGJ398; IC₃₀ value of 3.8 nM for lorlatinib), whereas DTP cells were insensitive to ALK-TKIs and sensitive to BGJ398 (IC₅₀ values of > 1000 nM for alectinib and 57.8 nM for BGJ398; IC₃₀ value of > 1000 nM for lorlatinib) (Fig. 1A, SFig. 3A). Additionally, although alectinib inhibited phosphorylation of ALK, as well as STAT3, AKT, ERK, and S6, which are involved in ALK or FGFR1 downstream signaling, in parental cells, it did not inhibit STAT3, AKT, ERK, and S6 phosphorylation in DTP cells regardless of complete suppression of ALK phosphorylation (Fig. 1B). BGJ398 did not inhibit phosphorylation of STAT3, AKT, ERK, and ALK in parental cells, whereas BGJ398 significantly inhibited FGFR1 phosphorylation and markedly inhibited ERK phosphorylation in DTP cells (Fig. 1B and 1C).

NCI-H2228 regrown cells, generated from DTP cells by culturing in the alectinib-free medium for 37 days, restored sensitivity to alectinib and insensitivity to BGJ398 compared with DTP cells (Fig. 1A, SFig. 3A). The susceptibility of regrown cells was similar to that of parental cells, indicating that their reversibility is identical to that of DTP cells reported by Mikubo et al.⁶. Therefore, FGFR1 signaling may be essential for the survival of DTP cells upon ALK blockade.

To examine the mechanism by which dependence on FGFR1 kinase is acquired in DTP cells, we analyzed the time course of ALK and FGFR1 signal transduction activation under alectinib treatment. FGF-1, 2, 3, 4, 5, 6, 10, 19, 20, 21, and 22 specifically bind to FGFR1 and activate signaling pathways involving AKT and ERK^12,14. Analysis of mRNA expression levels of these 11 FGFs in parental cells revealed that only FGF2 was expressed at greater than 1.00 TPM (SFig. 3B). Thus, we examined FGFR1 and FGF2 protein expression levels. We found that FGF2 protein levels and FGFR1 phosphorylation levels were significantly increased 13 days after treatment with alectinib compared to those before treatment, whereas almost no difference in FGFR1 protein levels was observed (Fig. 1D and 1E). Additionally, phosphorylation levels of STAT3, AKT, and ERK were elevated 13 days after treatment, despite the complete suppression of ALK phosphorylation during alectinib exposure (Fig. 1D). DTP cells reportedly present morphological alterations, stemness, or epithelial-mesenchymal transition (EMT), including vimentin (VIM) upregulation and cadherin 1 (CDH1) downregulation^6,15. NCI-H2228-DTP cells also acquired these features (Fig. 1D, SFig. 3C). BIM, a pro-apoptotic BCL2-family protein that mediates ALK inhibitor-induced apoptosis in ALK + lung cancer cells¹⁶, was increased from 24 hours to 13 days after treatment, whereas cleaved PARP protein, which indicates apoptosis, was temporarily increased 3 hours after treatment but disappeared thereafter (Fig. 1D). These findings suggest that apoptosis was immediately induced in NCI-H2228 cells via suppression of ALK signaling 3 hours after alectinib treatment, but cells survived by promoting a DTP state through the acquisition of stemness, EMT features, and activation of FGFR1 signaling, particularly downstream ERK reactivation, via increased FGF2 expression.

FGFR1 and FGF2 expressions promote cell survival against ALK-TKI

To validate whether activation of FGFR1 signaling by elevated expression of FGF2 protein is essential for cell survival against ALK-TKIs, we conducted knockdowns of FGFR1 and FGF2 using small interfering RNA (siRNA) and CRISPR/Cas9 in NCI-H2228 cells. Alectinib sensitivity was increased in NCI-H2228 cells transfected with FGFR1 or FGF2 siRNA relative to control siRNAs, with IC₄₀ values of > 1000 nM and 246.4 nM for control siRNAs, 47.0 nM and 14.7 nM for FGFR1 siRNAs, and 41.5 nM and 14.9 nM for FGF2 siRNAs (Fig. 2A, SFig. 4A). Additionally, alectinib sensitivity increased in both FGFR1- or FGF2-knockout clone transfected with FGFR1 or FGF2 crRNA relative to parental cells, with IC₅₀ values of 5.4 nM and 7.3 nM for FGFR1-knockouts, 5.6 nM and 6.1 nM for FGF2-knockouts, and 219.6 nM for parental cells (Fig. 2B, SFig. 4B). Alectinib-induced apoptosis and ERK inhibition were enhanced in both FGFR1- or FGF2-knockouts as compared to those in parental cells (Fig. 2C, SFig. 4C). FGFR1- or FGF2-knockout also enhanced sensitivity to lorlatinib, with IC₅₀ values of 0.7 nM in both FGFR1-knockouts, 0.7 nM and 0.9 nM for FGF2-knockouts, and > 1000 nM for parental cells (SFig. 4D).

To further verify that both FGFR1 and FGF2 are required for cell survival against ALK-TKIs, we established FGFR1- and FGF2-overexpressing SNU2535 cells with ALK + NSCLC. SNU2535 cells expressed extremely low levels of FGFR1 and FGF2 proteins (FGFR1^low and FGF2^low) compared to NCI-H2228 cells (FGFR1^high and FGF2^high) (Fig. 2D). Alectinib-induced inhibition of cell growth and AKT and ERK phosphorylation were restored in both FGFR1- and FGF2-overexpressing cells but not in control or FGF2-overexpressing cells (Fig. 2E and 2F). These findings suggest that cell survival against ALK-TKIs is activated through FGFR1 signaling involving AKT or ERK phosphorylation in FGFR^high and FGF2^high ALK + NSCLC cells, and both FGFR1 and FGF2 proteins promote escape from ALK-TKI-induced cell death.

FGFR1- and FGF2-expressing patients with ALK + NSCLC show poor response to ALK-TKIs

We retrospectively evaluated the association between clinical efficacy of ALK-TKIs and FGFR1 or FGF2 expression levels using data from the J-ALEX phase III study of patients with ALK + NSCLC treated with alectinib or crizotinib¹⁷. Since the number of FFPE tumor samples collected before treatment with alectinib (n = 38) or crizotinib (n = 31) was limited, we assessed the relationship between both drugs and estimated hazard ratios (HR) using the various cutoff values for basal FGFR or FGF2 mRNA expression levels in tumors before starting the treatment determined by RNA sequencing, and estimated the HR of high to low expression levels in each subset (SFig. 5B). The number of patients and PFS events for each subset are shown in SFig. 5C. We found that PFS against both FGFR1 and FGF2 tended to be short in an mRNA level-dependent manner (Fig. 3A). The group with the shortest PFS (log HR value over 1.0) had higher FGF2 expression levels corresponding to subsets 18 to 26 (Fig. 3A). Although the results were unclear, FGF2^high patients tended to express relatively high levels of FGFR1 mRNA compared with FGF2^low patients (Fig. 3B). When patients were categorized by the magnitude of FGF2 mRNA expression in subset 20, which showed the maximum statistics for the multivariate Cox model, the PFS for FGF2^high patients was shorter than that for FGF2^low patients (Fig. 3C). In contrast, when patients were categorized in subset 9, which showed the minimum statistics for the multivariate Cox model, PFS was similar between these patients. However, when we evaluated the relationship between FGFR1 or FGF2 expression and patient prognosis using the Gene Expression Profiling Interactive Analysis 2 database, high expression levels of FGFR1 or FGF2 were not significantly associated with poor prognosis in any cancers (SFig. 5D). Collectively, the decreased clinical efficacy of ALK-TKIs for FGFR1^high and FGF2^high patients were consistent with our nonclinical results in FGFR1^high and FGF2^high NCI-H2228 and FGFR1^low and FGF2^low SNU-2535 cells, suggesting that basal FGFR1^high and FGF2^high ALK + NSCLC tumors are associated with worse PFS in patients who received ALK-TKIs as a first-line treatment.

Combined ALK- and FGFR-TKI suppresses the growth of FGFR1^high and FGF2^high ALK + cells

To enhance the efficacy of ALK-TKIs in FGFR1^high and FGF2^high patients, we examined whether combined ALK- and FGFR-TKI suppresses proliferation of NCI-H2228 cells (Fig. 2D). Although NCI-H2228 cells were insensitive to FGFR-TKIs alone (BGJ398 and AZD4547) (IC₅₀ values > 1000 nM), ALK-TKI (alectinib and lorlatinib)-induced cell growth inhibition and apoptosis were enhanced upon the combination with FGFR-TKIs (Fig. 4A, 4B, and SFig. 6A). Co-treatment with alectinib and BGJ398 suppressed AKT and ERK phosphorylation compared to single agent treatment in NCI-H2228 cells (Fig. 4C). In contrast, there were no combinatorial effects on FGFR1^low and FGF2^low SNU-2535 cells. Therefore, the addition of FGFR-TKIs to ALK-TKIs may suppress the reactivation of cell survival signaling molecules through activation of FGFR1 kinase by binding FGFR1 and FGF2 proteins in FGFR1^high and FGF2^high ALK + NSCLC cells.

To evaluate the in vivo efficacy of these combinations, we treated mice xenografts of NCI-H2228 cells with BGJ398, alectinib, or a combination. NCI-H2228 tumors showed no response to BGJ398 alone, whereas the combination treatment resulted in significant tumor regression and decrease in ERK phosphorylation compared with alectinib alone (Fig. 4D and 4E, SFig. 6B and 6C). These in vivo combinatorial effects coincided with the in vitro effects on NCI-H2228 cells (Fig. 4A and 4B). Combination treatment was well tolerated with no weight loss over the course of treatment (SFig. 6D).

We further evaluated the effects of sequential treatment of ALK- and FGFR-TKIs with the later withdrawal of FGFR-TKI on NCI-H2228 cells in vitro (Fig. 4F). Concurrent combination treatment (No. 6) markedly inhibited cell growth compared with sequential treatments (Nos. 4 and 7) after 5 weeks of treatment (Fig. 4G). The inhibitory effects of concurrent combination treatment (No. 7) and 1-week concurrent combination followed by 4 weeks of alectinib treatment (No. 5) were similar up to 5 weeks, whereas significant cell regrowth was observed in treatment No. 5, but not in treatment No. 7 (Fig. 4G). Thus, it is important to continue concurrent combination therapies to suppress the growth of FGFR1^high and FGF2^high ALK + NSCLC cells.

Combined FGFR and EGFR, HER2, or BRAF inhibition enhances response in FGFR1^high and FGF2^high cells

We next determined whether FGFR-TKIs combined with targeted agents other than ALK-TKIs enhanced responses. Among all cell lines tested in this study, FGFR1 was highly expressed in EGFR + NSCLC NCI-H1650, HCC827, NCI-H1975, and HER2 + BC HCC1569 cells and BRAF + melanoma RPMI-7951, IGR-39, and SK-MEL-3 cells, and FGF2 was highly expressed in NCI-H1650, HCC827, HCC1569, RPMI-7951, and IGR-39 cells (Fig. 5A). We treated each cell line with targeted agents against EGFR, HER2, BRAF, or MEK alone or in combination with FGFR-TKIs. All cells were insensitive to FGFR-TKIs alone (IC₅₀s > 1000 nM), but each combination strongly inhibited cell growth and induced apoptosis compared with targeted agents alone and enhanced suppression of ERK phosphorylation in all FGFR1^high and FGF2^high cells but not in FGFR1^low and FGF2^low cells, regardless of the type of driver oncogenes (Fig. 5B, SFig. 7A-7D). Additionally, there was no combinatorial effect of FGFR-TKIs with osimertinib in FGFR1^high and FGF2^low NCI-H1975 cells (Fig. 5B and SFig. 7B).

In FGFR1^high and FGF2^high NCI-H1650 and HCC1569 xenograft tumors, co-treatment with FGFR-TKIs and EGFR-TKI osimertinib or HER2-TKI neratinib strongly inhibited tumor growth and ERK phosphorylation compared with osimertinib or neratinib alone, whereas FGFR1^low and FGF2^low II-18 xenograft tumors did not affect tumor growth (Fig. 5C and SFig. 7E-7G). Combination treatments were well tolerated with no weight loss over the course of treatment (SFig. 7H).

To verify that both FGFR1 and FGF2 proteins are required for cell survival against these TKIs, we conducted knockdown of FGFR1 and FGF2 using siRNAs. Sensitivity to osimertinib or BRAF-TKI dabrafenib tended to increase in FGFR1^high and FGF2^high NCI-H1650, HCC827, and RPMI-7951 cells transfected with siRNA against FGFR1 or FGF2 at doses ranging from 250 to 1000 nM compared to cells transfected with control siRNA (SFig. 7I). In FGFR1^high and FGF2^high HCC1569 cells, sensitivity to neratinib tended to increase at 1000 nM of neratinib following double knockdown of FGFR1 and FGF2 (SFig. 7I).

These findings are consistent with the results observed in ALK + NCI-H2228 cells and suggest that combination treatment with FGFR-TKIs and targeted agents is effective in FGFR1^high and FGF2^high cells having any driver oncogenes, indicating that FGFR1 protein expression alone is insufficient for the activation of FGFR1-induced survival against targeted agents, and coexistence with both FGFR1 and FGF2 may be essential for FGFR1 activation to escape targeted agents-induced cell death.

Combined EGFR- and FGFR-TKIs enhance response in cells with specific FGFR and FGF proteins

Co-treatment with BGJ398 and dabrafenib was partially effective in FGFR1^high BRAF + SK-MEL-3 cells regardless of FGF2^low expression (Fig. 5B). We found that SK-MEL-3 cells highly expressed FGF12 protein (SFig. 7J), suggesting that FGF12 partially activates FGFR1 survival against dabrafenib. To determine which of FGFRs and FGFs contribute to survival in cancer cells, we developed four FGFR-overexpressing cells using FGFR1^low and FGF2^low II18 cells and treated them with osimertinib in the presence of exogenous seven FGF proteins including FGF1, FGF2, FGF7, FGF9, FGF11, FGF12 and FGF18 that were expressed in NSCLC tumors in the J-ALEX study (data not shown). The pairs of FGFR1 and FGF2 or FGF9; FGFR2 and FGF2, FGF7, or FGF9; FGFR3 and FGF2 recovered more strongly from osimertinib-induced cell death than other pairs (viability values > 2.0), and these effects were strongly abrogated following addition to BGJ398 (Fig. 6B). Consistent with these results, the coexistence of FGFR1 or FGFR2 with FGF2 protein completely reactivated the osimertinib-induced decrease in phosphorylation of AKT, ERK, and S6, whereas the coexistence of FGFR3 or FGFR4 with FGF2 only partially reactivated the decrease in phosphorylation of ERK (SFig. 7K).

To confirm these observations, we further assessed FGFR1^high and FGF2^low NCI-H1975 cells in the presence of seven FGFs. FGF2 or FGF9 protein more strongly rescued osimertinib-induced cell death than other five FGFs (SFig. 7L), which was consistent with the results of FGFRs-overexpressing II-18 cells. Therefore, FGF12-mediated FGFR1 activation was much lower than FGF2-mediated FGFR1 activation, explaining the partial combinatorial effect of FGFR-TKIs observed in FGFR1^high, FGF2^low−and FGF12^high SK-MEL-3 cells.

Collectively, these findings suggest that cancer cells and patients with high protein expression levels of FGFR1 with FGF2 or FGF9; FGFR2 with FGF2, FGF7, or FGF9; FGFR3 with FGF2 may show a weak response to targeted therapy due to the activation of survival signaling from FGFR kinase, and the addition of FGFR-TKIs could overcome cell survival against targeted therapy in such FGFR- and FGF-positive cases.

DTP cells escape EGFR- and HER2-TKI-induced cell death through activation of FGFR1 signaling

We also assessed whether EGFR- or HER2-TKI induce a DTP state in FGFR1^high and FGF2^high cells by activation of FGFR1 survival signaling. Although these TKI-induced DTP cells showed the same features as H2228 DTP cells: upregulation of CD133 and a decrease in sensitivity to each targeted agent regardless of FGFR1 and FGF2 expression level, sensitivity to BGJ398 was increased in FGFR1^high and FGF2^high DTP cells but not in FGFR1^low and FGF2^low DTP cells including BRAF + COLO 679 cells (Fig. 7 and SFig. 8A). BGJ398 markedly inhibited AKT and ERK phosphorylation in FGFR1^high and FGF2^high DTP cells but not in FGFR1^low and FGF2^low DTP cells (SFig. 8B). The increased sensitivity to BGJ398 in FGFR1^high and FGF2^high DTP cells were lost in regrown cells. Therefore, the dependency of DTP cells on FGFR1 signaling may increase in FGFR1^high with FGF2^high cancer cells, regardless of tissue origins and driver oncogenes.

FGFR1 signaling activation was maintained in resistance cells

We investigated whether FGFR1 signaling is continuously activated after acquiring resistance to targeted treatment using alectinib-resistant NCI-H2228 and osimertinib-resistant HCC827 cells, established by exposing cells to 1000 nM of alectinib or osimertinib for 348 and 91 days, respectively. The IC₅₀ values of the resistant cells were > 4.6- and 277.8-fold higher than those of parental cells, respectively (Fig. 8A and SFig. 9A).

In alectinib-resistant cells, no ALK mutation was detected, whereas upregulation of FGFR1 and FGF2 proteins without an increase in FGFR1 and FGF2 copy number and induction of EMT and stemness characteristics were observed (Fig. 8B and 8C). Unlike alectinib-DTP cells, resistant cells were insensitive to BGJ398 alone (SFig. 9A). However, combined BGJ398 and alectinib strongly inhibited cell growth and ERK phosphorylation compared to single agents in resistant cells (Fig. 8A and SFig. 9A-9C).

In osimertinib-resistant HCC827 cells, no EGFR resistance mutation was detected (data not shown), whereas upregulation of FGFR1 proteins without an increase in FGFR1 copy number and MET protein with an approximately 4.4-fold increase in MET copy number was observed (Fig. 8B). EMT characteristics and small cell lung cancer (SCLC) transformation were induced, indicated by an increase in synaptophysin protein (Fig. 8C). MET amplification was reported in EGFR-TKI resistance cells¹⁸. Although a double combination of MET-TKI (capmatinib) plus osimertinib inhibited cell growth and inactivated ERK, a triple combination of capmatinib plus osimertinib plus BGJ398 was more effective in resistant cells (SFig. 9A-9C), indicating that FGFR1 survival signaling is partially retained when acquiring resistance to osimertinib, and MET-amplified resistant cells depend on EGFR, MET, and FGFR1 through MET amplification and upregulation of FGFR1 protein. Collectively, our findings suggest that activation of FGFR1 signaling is a mechanism underlying resistance to targeted therapy in FGFR1^high and FGF2^high cancer cells, and the addition of FGFR-TKIs to targeted therapy effectively suppress the development of not only DTP cells but also resistance cells from treatment-naive cancer in FGFR1^high and FGF2^high cancer, regardless of tissue origin or driver oncogenes.

Although genotype-directed target therapy is the standard of care and shows a dramatic response in many cancers, these therapies are rarely curative. In ALK + and EGFR + NSCLC, a complete response is observed in < 5% of patients who receive ALK- and EGFR-TKIs^19,20. Although small residual DTP cells can survive initial targeted TKI exposure, and develop drug resistance during treatment, predictive biomarkers underlying DTP mechanisms are unknown, despite target therapy-induced diverse DTP cells having been identified in vitro^6,21−26. Thus, it is essential to identify biomarkers before treatment to predict the weak response to targeted TKIs for establishing effective and less toxic DTP targeting therapies.

Here, we demonstrated that targeted TKIs for ALK, EGFR or HER2-induced DTP cells generated from FGFR1^high and FGF2^high cells acquired a dependency on FGFR signaling by activation of downstream molecules including AKT and ERK. Furthermore, knockdown of FGFR1, FGF2, or both increased targeted TKI-induced cell death, and co-treatment with FGFR-TKIs strongly enhanced the TKI-induced ERK inactivation, apoptosis, and cell death compared with single agents in FGFR1^high and FGF2^high cells. The FGFR and FGF ligand family consist of 4 and 23 members, and the binding of FGFs to FGFRs activates FGFR kinase followed by downstream signaling components^12,27. We found that, in addition to FGFR1 with FGF2, the sensitivity to EGFR-TKI was markedly decreased by the coexistence of FGFR1 with FGF2 or FGF9; FGFR2 with FGF2, FGF7, or FGF9; FGFR3 with FGF2 compared to FGFRs alone. Considering that acute responses to FGFR-TKIs combined with targeted TKIs were observed after 3 hours of ERK inhibition and apoptosis was induced several days after exposure in FGFR1^high and FGF2^high cells, both specific FGFRs and FGFs (FGFR1 with FGF2 or FGF9; FGFR2 with FGF2, FGF7, or FGF9; FGFR3 with FGF2)-high expressing cells at baseline may concurrently activate not only oncogenic signaling but also FGFR signaling by binding of steady state FGF. In line with this hypothesis, EGFR + NSCLC cells that co-express FGFR1 or FGFR2 with FGF2 or FGF9 tend to be insensitive to EGFR-TKIs²⁸. It was reported that both FGFR1- and FGF2-expressing EGFR + cells tended to show an EMT phenotype and upregulate FGF9 in FGFR1-expressing ALK + cells exhibiting characteristics of SCLC, and such features underly the acquired DTP state along with resistance to targeted TKIs^6,29−32. Collectively, the activation of FGFR with FGF autocrine signaling in such FGFR- and FGF-high expressing cells may contribute to the evasion of initial targeted TKI-induced acute apoptosis as an alternative pathway for cell survival and generate FGFR-dependent DTP cells immediately after targeted TKI treatment, regardless of driver oncogenes.

In this study, all FGF2-expressing cells highly expressed FGFR1 regardless of oncogenic mutation, and almost all FGF2^high ALK + patients had relatively high FGFR1 expression. Consistent with these observations, among the FGFRs and FGFs, FGFR1 and FGFR2 with FGF2 and FGF9 were frequently co-expressed in 33 NSCLC cell lines²⁸, and FGFR1 mRNA expression levels were also significantly correlated with FGF2 and FGF9 mRNA levels in patients with SCLC³³. Additionally, when we screened 135 lung cancer and 1,182 other cancer cell lines, many FGF2- or FGF9-high expressing cells tended to show relatively higher expression levels of FGFR1, FGFR2, or both (SFig. 10A). However, there were some FGF2- or FGF9-high expressing cells with low expression of FGFR1 or FGFR2 (SFig. 10A). This was also observed in FGF2- and FGF9-high expressing patients with bronchus and lung cancer (SFig. 10B). Collectively, high expression of both such FGFRs and FGFs may be a biomarker for the selection of patients who show low response to targeted TKIs but are expected to respond to combination therapy with FGFR-TKIs.

Although three FGFR-TKIs monotherapies have been approved by the FDA^34,35, no combination therapy with FGFR and targeted TKIs is approved. To perform clinical studies with these combination treatments, multiple FGFRs and FGFs expression levels and driver oncogenes must be detected. However, NGS-based tests to detect various oncogenic mutations including FoundationOne cannot measure mRNA levels, and immunohistochemistry is also impractical due to limited samples and cost. RNAscope is a commercially available and fully automated multiplex assay that simultaneously detects up to 4 different RNAs per slide, and is a highly sensitive and specific assay that can detect one RNA in fixed frozen, and FFPE tissue samples³⁶. In this study, FGFR1 and FGF2 RNA stained by RNAscope within an FFPE cell block of FGFR1^high and FGF2^high cells were higher than those of FGFR1^low and FGF2^low cells (SFig. 11), suggesting that FGFR1 and FGF2 expression levels observed in RNAscope and western blotting were comparable and RNAscope could be used to detect specific FGFRs and FGFs expression levels in cells in future clinical studies combining FGFR and targeted TKIs.

Several preclinical studies have shown that FGFR signaling is upregulated by increased expression of several FGFRs and/or FGFs, including FGF2, FGF9, FGF13, FGFR1, FGFR2, and FGFR3, after acquired resistance to various targeted TKIs³⁷. Upregulation of FGF9 protein and transdifferentiation to SCLC was reported in patients with EGFR + NSCLC after acquired resistance to EGFR-TKIs³¹. Here, we found a close relationship between DTP cells and resistant cells generated from FGFR1^high and FGF2^high cells: both can use alternative FGFR signaling for cell survival and growth, namely alectinib-DTP cells increase FGF2 expression, and alectinib-resistant cells increase both FGFR1 and FGF2 expression compared with NCI-H2228 parental cells. In contrast, the dependency on FGFR signaling is maintained in osimertinib-resistant MET-amplified cells by increasing FGFR1 expression compared with HCC827 parental cells, whereas no increase in FGFR1 and FGF2 expression was observed in osimertinib-DTP cells despite their dependence on FGFR signaling. Since the osimertinib-resistant cells increased the expression of an SCLC marker (synaptophysin), we speculate that HCC827 cells may use FGFR1 with FGF2 as well as FGF9 as alternative autocrine signals for cell survival in a DTP state, and resistant cells acquire resistance from the DTP state by MET amplification and maintain FGFR signal activation through FGFR1 upregulation. Further experiments are required to verify this hypothesis. Nevertheless, activation of FGFR with FGF autocrine signaling in specific FGFR^high and FGF^high cells including FGFR1 with FGF2 or FGF9; FGFR2 with FGF2, FGF7, or FGF9; FGFR3 with FGF2 may have a compensatory role in promoting the survival and growth of DTP and resistant cells during the treatment with targeted TKIs alone, and combined FGFR and targeted TKIs may be a promising therapeutic strategy to prevent a targeted TKI-induced DTP state or drug resistance against these specific FGFR^high and FGF^high cancer at baseline with various driver oncogenes.

This study has some limitations. First, the criteria for FGFR and FGF expression levels that may predict the efficacy of combined FGFR and targeted TKIs remain unclear. Second, the efficacy of this combination treatment for cells with each pair of FGFR1-4 and FGF1-23 other than those examined is unknown. Third, we demonstrated the efficacy of combination treatment with ALK-, EGFR-, HER2-, and BRAF-TKIs and FGFR-TKIs, but did not assess other targeted TKIs such as ROS1- and RET-TKIs, since no FGFR- and FGF-positive cell lines harboring such mutations were available to us. Fourth, we retrospectively performed an integrated analysis of crizotinib and alectinib in ALK + NSCLC patients between high or low FGFR1 and FGF2 expression because of the small sample size. Since the efficacy of alectinib is much stronger than crizotinib¹⁷, we added the treatment regimen as a covariate in the multivariate analysis. However, other potential prognostic factors may have been unbalanced between high and low expression groups in some subsets. Furthermore, we did not assess cancer patients with other driver mutations since such clinical samples were not available. Finally, treatment regimens using more than one anti-cancer drug often result in increased toxic side effects that can compromise patient safety and treatment efficacy. In this mouse model, combining FGFR and targeted TKIs did not show obvious toxicity affecting body weight, whereas combination treatment with erlotinib and the unselective FGFR-TKI (dovitinib) was terminated for patients with NSCLC because of adverse effects³⁸. To overcome these limitations, further preclinical and prospective clinical studies are needed to assess the efficacy or tolerability of combination treatment with various targeted TKIs and highly selective FGFR inhibitors.

To our knowledge, this study is the first to report that high expression levels of specific FGFRs and FGFs (FGFR1 with FGF2 or FGF9; FGFR2 with FGF2, FGF7, or FGF9; FGFR3 with FGF2) at baseline rapidly promote tolerance and continuously maintain survival and growth of DTP and resistant cells during treatment with targeted TKIs in various driver oncogene-positive cancer cells including those with ALK, EGFR, HER2, or BRAF mutations. Therefore, it is plausible that an initial dual blockade of FGFR and various driver oncogenes may be a potent treatment strategy for patients with basal high expression of such FGFRs and FGFs to prevent the development of intrinsic resistance to targeted TKIs, and lead to tumor eradication.

Cell lines

The supplier, and culture medium of human cancer cell lines in this study are listed in SFig. 12. The oncogenic driver mutations were referred from the COSMIC cell database (https://cancer.sanger.ac.uk/cosmic) and previous reports ^39–42. Cells were maintained at 37°C under 5% CO₂.

Drugs and reagents

Alectinib was synthesized by Chugai Pharmaceutical Co., Ltd. (Tokyo, Japan). Erlotinib was provided by F. Hoffman-La Roche Ltd. (Basel, Switzerland). Infrigatinib (BGJ398), osimertinib, neratinib, and dabrafenib were obtained from LC Laboratories (Woburn, MA, USA). AZD4547 and trametinib were obtained from ChemScene (Monmouth Junction, NJ, USA). Lorlatinib, capmatinib, and lapatinib were obtained from Selleck Chemicals (Houston, TX, USA). The anti-cancer compound library was obtained from TargetMol (Boston, MA, USA). All agents were dissolved in dimethyl sulfoxide (DMSO: Sigma-Aldrich, St. Louis, MO, USA) for in vitro assays and in a 6% (w/v) solution of Captisol (ChemScene) for in vivo assays. DMSO and Captisol were used as vehicle controls. Recombinant human FGF1, FGF2, FGF7, and FGF12 proteins were obtained from R&D Systems (Minneapolis, MN, USA) and dissolved in phosphate buffered saline (PBS, Sigma-Aldrich). Recombinant human FGF9 and FGF18 proteins were obtained from Abcam (Cambridge, UK) and dissolved in distilled water. Recombinant human FGF11 protein was obtained from LifeSpan Biosciences (Seattle, WA, USA) and dissolved in Tris buffer (Horizon Discovery, Cambridge, UK).

Anti-cancer compound library screen of NCI-H2228 cells

NCI-H2228 parental and DTP cells were seeded in 384-well plates, and 3114 agents in the anti-cancer compound library were added at 100 nM on the following day. After 6 days, cell viability was determined using the CellTiter-Glo 3D Cell Viability Assay kit. Viability for agent- versus vehicle treatment was measured, and the ratio of antiproliferative effect for each agent on DTP cells was calculated as follows: viability of DTP cells/ viability of parental cells. Agents with an anti-proliferative effect less than 0.7 and with the viability of parental cells between 0.9 and 1.1 were selected as candidate compounds.

Cell proliferation assay

Cells were seeded in 384-well plates, and drugs were added at the indicated concentrations the following day. After 8 days, cell viability was determined by quantitation of cellular ATP, which indicates metabolically active cells, using the CellTiter-Glo 3D Cell Viability Assay (Promega, Madison, WI, USA). Viability with the agents relative to viability with the vehicle was measured. Each point represents the mean + standard deviation of triplicate experiments. IC₅₀, IC₄₀, and IC₃₀ values were calculated as described previously ⁴³.

Western blotting

Cells were seeded in 6-well plates and drugs were added at the indicated concentrations the following day and cultured for the indicated time. The same amount of protein lysate was loaded for each western blotting assay using the Sally Sue or Jess capillary electrophoresis-based protein analysis system (ProteinSimple, Santa Clara, CA, USA) according to the manufacturer’s protocol. Antibodies against ALK, phospho-ALK, FGFR1, FGFR2, FGFR3, FGFR4, ERK, phospho-ERK, AKT, phospho-AKT, EGFR, phospho-EGFR, MET, β-actin, phospho-MET, HER2, CD44, CD133, BIM, VIM, CDH1, S6, pS6 and STAT3, phospho-STAT3 (Cell Signaling Technology, Danvers, MA, USA), cleaved PARP, phospho-HER2, FGF2, and FGF12 (Abcam) were used. Uncropped immunoblots blots of each figure are included in the Source Data file 1.

ELISA assay

Cells were seeded in 6-well plates, and the following day, drugs were added at the indicated concentrations and cultured for the indicated time. The same amount of protein lysate was loaded for each ELISA assay conducted using the PathScan phospho-FGFR1 Sandwich ELISA Kit (Cell Signaling Technology) according to the manufacturer’s protocol.

Apoptosis assay

Cells were seeded in 384-well plates, and the following day, drugs were added at the indicated concentrations and cultured for the indicated time. The activity caspases 3 and 7 was determined using the Caspase-Glo 3/7 Assay (Promega). The activity was corrected with respect to cell viability determined as described above, and relative activity with the agents against vehicle was calculated. Each point represents the mean + standard deviation of triplicate experiments.

Knockdown of FGFR1 or FGF2 by siRNA

Cells were transfected with ON-TARGETplus predesigned siRNA targeting FGFR1, FGF2, or non-targeting control (Horizon Discovery) using NEPA21 electroporation (Nepa Gene, Chiba, Japan) according to the manufacturer’s protocol, and the following day, drugs were added at the indicated concentrations and cultured for 6 days. Knockdown of FGFR1 or FGF2 was confirmed by observing the loss of each protein by western blotting after two days of siRNA transfection.

Establishment of FGFR1- and FGF2-knockout cells by CRISPR/Cas9

NCI-H2228 cells were transfected with Cas9 Nuclease protein NLS (Nippon Gene, Tokyo, Japan), Edit-R tracrRNA, and crRNA-targeting human FGFR1 (5’-GCATGGTTGACCGTTCTGGA-3’) or FGF2 (5’-ATGTGGCACTGAAACGAACT-3’) (Fasmac, Kanagawa, Japan) constructs by NEPA21 electroporation. Subsequently, each clone was isolated using a Smart Aliquotor (NT Science, Aichi, Japan), and the FGFR1- or FGF2-knockout cells were confirmed by western blotting.

Establishment of FGFR1- and FGF2-overexpressing SNU-2535 cells

The human CDS for the FGFR1, FGF2, or non-targeting control was synthesized, cloned, and inserted downstream of the EF1A promoter in a lentiviral vector by VectorBuilder (Chicago, IL, USA). SNU-2535 cells were infected with lentiviral vectors at a multiplicity of infection of 3 (MOI = 3) overnight according to the manufacturer’s protocol. On the following day, the culture medium was replaced with a puromycin-containing medium for 4 days to select the FGFR- and FGF2-overexpressing cells, and each stable clone was isolated using the Smart Aliquotor. Cells were maintained in a normal culture medium without puromycin, and FGFR1 or FGF2 protein expression was confirmed using western blotting.

Establishment of FGFR1, 2, 3, 4–overexpressing II-18 cells

The human CDS for the four FGFRs (FGFR1, FGFR2, FGFR3, and FGFR4) or non-targeting control was synthesized, cloned, and inserted downstream of the EF1A promoter in a lentiviral vector by VectorBuilder. II-18 cells were infected with lentiviral vectors at MOI = 10 overnight according to the manufacturer’s protocol. On the following day, the culture medium was replaced with puromycin- or G418-containing medium for 4 or 8 days, respectively, to select FGFR-overexpressing cells. Cells were maintained in a normal culture medium without antibiotics, and expression levels of four FGFRs were confirmed using western blotting.

Establishment of alectinib- and osimertinib-resistant cells

NCI-H2228 and HCC827 cells were exposed to 1000 nM of alectinib and osimertinib for 348 and 91 days, respectively. After washing, cells were cultured in a medium without drugs for more than 60 days. Resistance to each drug was confirmed using a cell proliferation assay.

Detection of EGFR mutation

Genomic DNA was obtained from cells using a Maxwell 16 Tissue DNA Purification Kit (Promega). EGFR mutations were measured using the AmoyDx EGFR 29 Mutations Detection kit and fluorescent PCR (Amoy Diagnostics, Xiamen, China) with a LightCycler 480 System (Roche Diagnostics, Basel, Switzerland), and positive mutations were defined based on the Ct value according to the manufacturer’s instructions.

Detection of ALK mutation

Briefly, DNA samples were quantified using a NanoDrop One Spectrophotometer (Thermo Fisher Scientific), and the DNA integrity number was determined using a Genomic DNA ScreenTape (Agilent, Santa Clara, CA, USA) on an Agilent 2200 TapeStation. Next, DNA sequencing libraries were prepared using SureSelect XT Low Input (Agilent) and SureSelect Human All Exon V7 (Agilent), and sequenced paired-end 100 bp with a NovaSeq6000 (Illumina). Reads were mapped to the human reference genome build hg38 using Burrows-Wheeler Aligner ver.0.7.17, and somatic SNVs and small indels were identified in the genomic data using Strelka ver.2.9.10.

Copy number analysis

Copy numbers of DNA were measured using the pre-designed TaqMan copy number probe (Thermo Fisher Scientific) for human FGFR1, FGF2, ALK, EGFR, and MET with quantitative real-time PCR using a LightCycler 480 System. Copy numbers normalized to the reference control gene of human RNase P were analyzed using CopyCaller ver. 2.1 software (Thermo Fisher Scientific) with human genomic DNA (Promega) as diploid control DNA.

Animals

Animal procedures were approved by the Institutional Animal Care and Use Committee at Chugai Pharmaceutical Co., Ltd. Male 5-week-old SCID mice (C.B-17/Icr-scid/scidJcl) were purchased from CLEA Japan, Inc (Tokyo, Japan). Male 4–5-week-old BALB/c-nu/nu mice (CAnN Cg-Foxn1 < nu>/CrlCrlj nu/nu) were purchased from Charles River Laboratories Japan, Inc (Yokohama, Japan). All animals were acclimatized for > 5 days prior to the study. Chlorinated water and irradiated food were provided ad libitum, and the animals were kept under a controlled 12/12 h light-dark cycle.

Mouse xenograft models

Mice were inoculated subcutaneously with 5×10⁶ cells per mouse into the right flank. Tumor volume and body weight were measured twice a week, and tumor volume was estimated as follows: tumor volume = ab²/2, where a and b are tumor length and width, respectively. After tumor establishment, mice were randomly allocated to 11-day treatment with the vehicle or indicated drug or drug combination in which each compound was orally administered daily at the same dose and schedule as the single agent.

Retrospective analysis of the J-ALEX study

Japanese patients who were ALK inhibitor-naïve and chemotherapy-naïve or those who had received one prior chemotherapy regimen were enrolled and randomized to receive alectinib (n = 103) or crizotinib (n = 104) until progressive disease, unacceptable toxicity, death, or withdrawal in the phase III J-ALEX study (JapicCTI-132316; JO28928) ¹⁷. In this study, we obtained 80 formalin-fixed, paraffin-embedded (FFPE) tumor specimens before the start of treatment with alectinib or crizotinib and stored them at 4°C until use.

RNA sequence

Briefly, RNA samples were quantified using a NanoDrop One Spectrophotometer, and the RNA integrity number was determined using an RNA 6000 Nano Kit (Agilent) on an Agilent 2100 Bioanalyzer. Next, RNA sequencing libraries were prepared using the SMART-Seq Stranded Kit (Takara, Shiga, Japan). Since the library was less than 2 nM, we excluded six samples. The libraries were then pooled and sequenced paired-end 150 bp with NovaSeq6000. Fastq sequence files were aligned to the human reference genome (hg38) using STAR 2.7.8a, and the reads were processed using StrandNGS 4.0 (Agilent). Since the mapping rate was less than 40%, we excluded five samples and analyzed the remaining 69 samples. The read counts for each gene were quantified as TPM.

Statistical analysis

A multivariate Cox proportional hazard model was used to estimate the adjusted hazard ratio (HR) and 95% confidence interval of PFS against ALK TKIs associated with either FGFR1 or FGF2 mRNA expression levels in each subset. Subsets were derived by changing category definitions in accordance with the observed TPM value of FGFR1 or FGF2 mRNA as cut-off values ranging from 0 to the 90th percentile of the log₂(TPM + 1) scale. In each subset, the analysis involved adjustment for covariates known to be associated with PD or death including treatment drugs, ECOG performance status (PS) at baseline, treatment line, disease status, and brain metastases at baseline by IRF, which were used for the estimation (SFig. 5A). Kaplan–Meier methodology was used to estimate the distribution of PFS for specific cut-off values. These clinical analyses were conducted using R Studio version 1.0153 and R version 3.6.0.

In vitro, experimental data were analyzed by Student's t-test or Dunnett's test. In vivo, experimental data were analyzed using the Wilcoxon rank sum test followed by the Holm–Bonferroni method. All statistical analyses were conducted using JMP Ver. 15.0.0 (SAS Institute, Cary, NC, USA). Significance was set at a two-tailed P < 0.05.

Acknowledgements:

We are grateful to the investigators, and patients involved in this study. English language editing services was provided by Editage.

Funding: Chugai Pharmaceutical Co., Ltd.

Author contributions

K.F. designed the study, performed RNA sequencing, and in vitro and in vivo experiments; supervised the experiments; and wrote the manuscript. T.F. performed in vitro and in vivo experiments. T.A. and T.Y. analyzed clinical data of J-ALEX study. Y.Y. and S.Y. supervised the experiments.

Competing interests: All authors are employees of Chugai Pharmaceutical Co., Ltd., and declare no conflict of interest.

Data and materials availability:

Qualified researchers may request access to individual patient-level data through the clinical study data request platform (https://www.clinicalstudydatarequest.com/Default.aspx). For further details on Chugai's Data Sharing Policy and how to request access to related clinical study documents, see here (www.chugai-pharm.co.jp/english/profile/rd/ctds_request.html). RNA sequencing data of the J-ALEX study cannot be deposited with a third party involving any public database, because we did not obtain the patient's informed consent for depositing their RNA data into a third-party repository. Thus, we shall comply with the Act on the Protection of Personal Information in Japan and shall not disclose individual RNA sequencing data. Other all experimental data supporting the findings of this study are available within the article and its supplementary information files and from the corresponding author upon reasonable request.

Weinstein, I. B. & Joe, A. Oncogene addiction. Cancer Res 68, 3077-3080; discussion 3080, doi:10.1158/0008-5472.Can-07-3293 (2008).
Pagliarini, R., Shao, W. & Sellers, W. R. Oncogene addiction: pathways of therapeutic response, resistance, and road maps toward a cure. EMBO Rep 16, 280-296, doi:10.15252/embr.201439949 (2015).
Francies, H. E., McDermott, U. & Garnett, M. J. Genomics-guided pre-clinical development of cancer therapies. Nature Cancer 1, 482-492, doi:10.1038/s43018-020-0067-x (2020).
Aldea, M. et al. Overcoming Resistance to Tumor-Targeted and Immune-Targeted Therapies. Cancer Discovery 11, 874, doi:10.1158/2159-8290.CD-20-1638 (2021).
Sharma, S. V. et al. A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations. Cell 141, 69-80, doi:10.1016/j.cell.2010.02.027 (2010).
Mikubo, M., Inoue, Y., Liu, G. & Tsao, M. S. Mechanism of Drug Tolerant Persister Cancer Cells: The Landscape and Clinical Implication for Therapy. J Thorac Oncol 16, 1798-1809, doi:10.1016/j.jtho.2021.07.017 (2021).
Dagogo-Jack, I. et al. Treatment with Next-Generation ALK Inhibitors Fuels Plasma <em>ALK</em> Mutation Diversity. Clinical Cancer Research 25, 6662, doi:10.1158/1078-0432.CCR-19-1436 (2019).
Yoda, S. et al. Sequential ALK Inhibitors Can Select for Lorlatinib-Resistant Compound <em>ALK</em> Mutations in ALK-Positive Lung Cancer. Cancer Discovery 8, 714, doi:10.1158/2159-8290.CD-17-1256 (2018).
Recondo, G. et al. Molecular Mechanisms of Acquired Resistance to MET Tyrosine Kinase Inhibitors in Patients with MET Exon 14-Mutant NSCLC. Clin Cancer Res 26, 2615-2625, doi:10.1158/1078-0432.Ccr-19-3608 (2020).
Cohen, P., Cross, D. & Jänne, P. A. Kinase drug discovery 20 years after imatinib: progress and future directions. Nature Reviews Drug Discovery 20, 551-569, doi:10.1038/s41573-021-00195-4 (2021).
Tanimura, K. et al. Inhibition of c-Jun N-terminal kinase signaling increased apoptosis and prevented the emergence of ALK-TKI-tolerant cells in ALK-rearranged non-small cell lung cancer. Cancer Letters 522, 119-128, doi:https://doi.org/10.1016/j.canlet.2021.09.018 (2021).
Ornitz, D. M. & Itoh, N. The Fibroblast Growth Factor signaling pathway. Wiley Interdiscip Rev Dev Biol 4, 215-266, doi:10.1002/wdev.176 (2015).
Perera, T. P. S. et al. Discovery and Pharmacological Characterization of JNJ-42756493 (Erdafitinib), a Functionally Selective Small-Molecule FGFR Family Inhibitor. Mol Cancer Ther 16, 1010-1020, doi:10.1158/1535-7163.Mct-16-0589 (2017).
Tiong, K. H., Mah, L. Y. & Leong, C.-O. Functional roles of fibroblast growth factor receptors (FGFRs) signaling in human cancers. Apoptosis 18, 1447-1468, doi:10.1007/s10495-013-0886-7 (2013).
Hangauer, M. J. et al. Drug-tolerant persister cancer cells are vulnerable to GPX4 inhibition. Nature 551, 247-250, doi:10.1038/nature24297 (2017).
Takezawa, K., Okamoto, I., Nishio, K., Jänne, P. A. & Nakagawa, K. Role of ERK-BIM and STAT3-Survivin Signaling Pathways in ALK Inhibitor–Induced Apoptosis in EML4-ALK–Positive Lung Cancer. Clinical Cancer Research 17, 2140-2148, doi:10.1158/1078-0432.Ccr-10-2798 (2011).
Nakagawa, K. et al. Final progression-free survival results from the J-ALEX study of alectinib versus crizotinib in ALK-positive non-small-cell lung cancer. Lung Cancer 139, 195-199, doi:10.1016/j.lungcan.2019.11.025 (2020).
Jakobsen, K. R. et al. MET amplification and epithelial-to-mesenchymal transition exist as parallel resistance mechanisms in erlotinib-resistant, EGFR-mutated, NSCLC HCC827 cells. Oncogenesis 6, e307-e307, doi:10.1038/oncsis.2017.17 (2017).
Camidge, D. R. et al. Updated Efficacy and Safety Data and Impact of the EML4-ALK Fusion Variant on the Efficacy of Alectinib in Untreated ALK-Positive Advanced Non–Small Cell Lung Cancer in the Global Phase III ALEX Study. Journal of Thoracic Oncology 14, 1233-1243, doi:https://doi.org/10.1016/j.jtho.2019.03.007 (2019).
Soria, J.-C. et al. Osimertinib in Untreated EGFR-Mutated Advanced Non–Small-Cell Lung Cancer. New England Journal of Medicine 378, 113-125, doi:10.1056/NEJMoa1713137 (2017).
Tsuji, T. et al. YAP1 mediates survival of ALK-rearranged lung cancer cells treated with alectinib via pro-apoptotic protein regulation. Nature Communications 11, 74, doi:10.1038/s41467-019-13771-5 (2020).
Kurppa, K. J. et al. Treatment-Induced Tumor Dormancy through YAP-Mediated Transcriptional Reprogramming of the Apoptotic Pathway. Cancer Cell 37, 104-122.e112, doi:10.1016/j.ccell.2019.12.006 (2020).
Tanimura, K. et al. HER3 activation contributes toward the emergence of ALK inhibitor-tolerant cells in ALK-rearranged lung cancer with mesenchymal features. npj Precision Oncology 6, 5, doi:10.1038/s41698-021-00250-8 (2022).
Taniguchi, H. et al. AXL confers intrinsic resistance to osimertinib and advances the emergence of tolerant cells. Nature Communications 10, doi:10.1038/s41467-018-08074-0 (2019).
Yi, Y. et al. Cancer-associated fibroblasts promote epithelial-mesenchymal transition and EGFR-TKI resistance of non-small cell lung cancers via HGF/IGF-1/ANXA2 signaling. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1864, 793-803, doi:https://doi.org/10.1016/j.bbadis.2017.12.021 (2018).
Arasada, R. R. et al. Notch3-dependent β-catenin signaling mediates EGFR TKI drug persistence in EGFR mutant NSCLC. Nature Communications 9, doi:10.1038/s41467-018-05626-2 (2018).
Johnson, D. E. & Williams, L. T. Structural and functional diversity in the FGF receptor multigene family. Adv Cancer Res 60, 1-41, doi:10.1016/s0065-230x(08)60821-0 (1993).
Marek, L. et al. Fibroblast Growth Factor (FGF) and FGF Receptor-Mediated Autocrine Signaling in Non-Small-Cell Lung Cancer Cells. Molecular Pharmacology 75, 196-207, doi:10.1124/mol.108.049544 (2009).
Raoof, S. et al. Targeting FGFR overcomes EMT-mediated resistance in EGFR mutant non-small cell lung cancer. Oncogene 38, 6399-6413, doi:10.1038/s41388-019-0887-2 (2019).
Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603-607, doi:10.1038/nature11003 (2012).
Ishioka, K. et al. Upregulation of FGF9 in Lung Adenocarcinoma Transdifferentiation to Small Cell Lung Cancer. Cancer Research 81, 3916-3929, doi:10.1158/0008-5472.Can-20-4048 (2021).
Leonetti, A. et al. Resistance mechanisms to osimertinib in EGFR-mutated non-small cell lung cancer. British Journal of Cancer 121, 725-737, doi:10.1038/s41416-019-0573-8 (2019).
Zhang, L. et al. Fibroblast Growth Factor Receptor 1 and Related Ligands in Small-Cell Lung Cancer. Journal of Thoracic Oncology 10, 1083-1090, doi:10.1097/JTO.0000000000000562 (2015).
Jørgensen, J. T. Oncology drug-companion diagnostic combinations. Cancer Treatment and Research Communications 29, 100492, doi:https://doi.org/10.1016/j.ctarc.2021.100492 (2021).
Lee, P. C. et al. Targeting the Fibroblast Growth Factor Receptor (FGFR) in Advanced Cholangiocarcinoma: Clinical Trial Progress and Future Considerations. Cancers (Basel) 13, 1706, doi:10.3390/cancers13071706 (2021).
McGinnis, L. M., Ibarra-Lopez, V., Rost, S. & Ziai, J. Clinical and research applications of multiplexed immunohistochemistry and in situ hybridization. The Journal of Pathology 254, 405-417, doi:https://doi.org/10.1002/path.5663 (2021).
Zhou, Y. et al. FGF/FGFR signaling pathway involved resistance in various cancer types. J Cancer 11, 2000-2007, doi:10.7150/jca.40531 (2020).
Das, M. et al. Dovitinib and erlotinib in patients with metastatic non-small cell lung cancer: A drug–drug interaction. Lung Cancer 89, 280-286, doi:https://doi.org/10.1016/j.lungcan.2015.06.011 (2015).
Yoshimura, Y. et al. Antitumor activity of alectinib, a selective ALK inhibitor, in an ALK-positive NSCLC cell line harboring G1269A mutation: Efficacy of alectinib against ALK G1269A mutated cells. Cancer Chemother Pharmacol 77, 623-628, doi:10.1007/s00280-016-2977-y (2016).
Liu, Q. et al. A novel HER2 gene body enhancer contributes to HER2 expression. Oncogene 37, 687-694, doi:10.1038/onc.2017.382 (2018).
Xiao, J., Egger, M. E., McMasters, K. M. & Hao, H. Differential expression of ABCB5 in BRAF inhibitor-resistant melanoma cell lines. BMC Cancer 18, 675-675, doi:10.1186/s12885-018-4583-3 (2018).
Yuan, J. et al. The dimer-dependent catalytic activity of RAF family kinases is revealed through characterizing their oncogenic mutants. Oncogene 37, 5719-5734, doi:10.1038/s41388-018-0365-2 (2018).
Furugaki, K. et al. Impact of bevacizumab in combination with erlotinib on EGFR-mutated non-small cell lung cancer xenograft models with T790M mutation or MET amplification. Int J Cancer 138, 1024-1032, doi:10.1002/ijc.29848 (2016).

(Not answered)

Appendix1.xlsx
Dataset 1
SupplementaryMaterialsNPG.docx

Download PDF

Journal Publication

published 25 Oct, 2023

Read the published version in npj Precision Oncology →

Editorial decision: revise
13 Dec, 2022
Editor assigned by journal
09 Dec, 2022
Submission checks completed at journal
09 Dec, 2022
First submitted to journal
08 Dec, 2022

You are reading this latest preprint version

FGFR blockade inhibits targeted therapy-tolerant persister cells in basal FGFR1 and FGF2 high expressing cancers with driver oncogenes

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Compound screening identifies FGFR1 as a candidate promoting alectinib-induced DTP cells

DTP cells escape ALK-TKI-induced cell death through activation of FGFR signaling

FGFR1 and FGF2 expressions promote cell survival against ALK-TKI

FGFR1- and FGF2-expressing patients with ALK + NSCLC show poor response to ALK-TKIs

Combined ALK- and FGFR-TKI suppresses the growth of FGFR1high and FGF2high ALK + cells

Combined FGFR and EGFR, HER2, or BRAF inhibition enhances response in FGFR1high and FGF2high cells

Combined EGFR- and FGFR-TKIs enhance response in cells with specific FGFR and FGF proteins

DTP cells escape EGFR- and HER2-TKI-induced cell death through activation of FGFR1 signaling

FGFR1 signaling activation was maintained in resistance cells

Discussion

Materials And Methods

Cell lines

Drugs and reagents

Anti-cancer compound library screen of NCI-H2228 cells

Cell proliferation assay

Western blotting

ELISA assay

Apoptosis assay

Knockdown of FGFR1 or FGF2 by siRNA

Establishment of FGFR1- and FGF2-knockout cells by CRISPR/Cas9

Establishment of FGFR1- and FGF2-overexpressing SNU-2535 cells

Establishment of FGFR1, 2, 3, 4–overexpressing II-18 cells

Establishment of alectinib- and osimertinib-resistant cells

Detection of EGFR mutation

Detection of ALK mutation

Copy number analysis

Animals

Mouse xenograft models

Retrospective analysis of the J-ALEX study

RNA sequence

Statistical analysis

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

Combined ALK- and FGFR-TKI suppresses the growth of FGFR1^high and FGF2^high ALK + cells

Combined FGFR and EGFR, HER2, or BRAF inhibition enhances response in FGFR1^high and FGF2^high cells