Clinical advances and challenges in targeting FGF/FGFR signaling in lung cancer

Abstract

Fibroblast growth factors (FGFs) and their receptors regulate numerous cellular processes, such as metabolism and signal transduction, but can also drive tumorigenesis. Specifically, in lung cancer, the overexpression of FGFs, as well as the amplification, mutation and fusion of FGFR genes, are closely linked to the initiation, progression and resistance of the disease, suggesting that targeting FGF/FGFR is an attractive therapeutic strategy for lung cancer treatment. Nintedanib, a multitarget tyrosine kinase inhibitor (TKI) used in combination with docetaxel, has shown some success as a second-line therapy for lung cancer. However, clinical trials evaluating other FGFR inhibitors have yielded mixed results, indicating substantial complexity in targeting aberrant FGF/FGFR signaling. In this review, we describe the aberrations in FGF/FGFR signaling in lung cancer and summarize the clinical efficacy of FGFR inhibitors, such as multitarget TKIs, selective FGFR-TKIs and biological agents. We also discuss various challenges associated with FGFR targeting in lung cancer, including precision patient selection, toxicity and resistance. Finally, we provide perspectives on future directions, namely, developing novel FGFR-targeting drugs, such as FGFR degraders and more specific FGFR-TKIs, adopting combination therapy and targeting FGFs.

Introduction

Fibroblast growth factor receptors (FGFRs), similar to epidermal growth factor receptor (EGFR), belong to the tyrosine kinase receptor family. The FGFR family consists of five members [1]. Among them, FGFR1-4 have similar structures and can conduct signal transmission after binding with ligands; however, while FGFR5 can bind with ligands, it lacks the intracellular catalytic domain and is unable to conduct signal transduction [2]. A series of intracellular biological processes are triggered when FGFR1-4 receptors are activated by fibroblast growth factors (FGFs). For example, FGFRs promote embryonic development [3].

FGFR signaling pathway dysregulation, however, can also play a role in various pathological processes, including lung cancer [4]. Preclinical studies have suggested that abnormal FGFR signaling could be a potential therapeutic target, leading to the development of FGFR-targeted drugs [5]. Among these are multitargeting TKIs, selective inhibitors, and monoclonal antibodies. To date, the FDA has approved four anti-FGFR drugs (erdafitinib, futibatinib, infigratinib, and pemigatinib) because of their excellent efficacy and safety. Currently, erdafitinib is used for FGFR2/3-altered urothelial cancer [6]. Futibatinib, pemigatinib and infigratinib are used for the treatment of FGFR2-altered cholangiocarcinoma [7,8,9]. Additionally, pemigatinib has also received approval in patients with FGFR1-altered myeloid/lymphoid neoplasms (MLNs) [10]. To target FGFR in lung cancer, nintedanib has received approval from the European Medicines Agency’s Committee for the treatment of non-small cell lung cancer (NSCLC) in combination with docetaxel as a second-line treatment [11].

In this review, we describe the important mechanisms of oncogenic FGF and FGFR signaling in lung cancer. Additionally, we offer a comprehensive overview of the most recent advancements in FGFR-targeted therapies, with a focus on both their successes and limitations in clinical settings. Furthermore, challenges, including the precise selection of patients with FGFR alterations, toxicity and resistance to FGFR-targeting therapies, are summarized. Finally, further directions for the development of FGFR-targeting therapies for lung cancer are discussed.

FGF ligands and receptor signaling in lung cancer

FGFs are bioactive proteins that play essential roles in critical physiological processes such as tissue repair, metabolic homeostasis and responses to injuries [12,13,14]. Furthermore, accumulating evidence suggests that dysregulation of FGFs contributes to tumorigenesis [15, 16]. Several canonical FGFs, including FGF1, FGF2, FGF3, FGF4, and FGF9, and endocrine FGFs, such as FGF19 and FGF21, have been implicated in the development and progression of lung cancer (Fig. 1).

Analysis of FGF1 expression in primary human NSCLC tissues has identified FGF1 as a prognostic biomarker and a potential target for NSCLC patients [17]. High expression of FGF1 is positively correlated with primary tumor size and vascular invasion [18]. Additionally, research has shown that FGF1 promotes the amplification and cancer stemness of lung cancer cells in a manner dependent on the MAPK signaling pathway [18]. The FGF2 mRNA and protein are overexpressed in both NSCLC and small cell lung cancer (SCLC) tissues [19, 20]. FGF2 increases ERK1/2 phosphorylation, downregulates Beclin-1, and inhibits autophagy [21], thereby facilitating cell survival in FGFR1-amplified NSCLC. Wang K et al. reported FGF2 promotes cell proliferation, epithelial-mesenchymal transition (EMT), and metastasis by activating the FGFR1-ERK1/2-SOX2 axis [22]. FGF2 also significantly increases the proliferation of SCLC cells via activation of the PI3K/AKT and MAPK pathways [23]. Pardo OE reported that FGF2 prevents SCLC cells form undergoing apoptosis by increasing the expression of antiapoptotic proteins, such as XIAP and Bcl-XL [24]. FGF3 has been detected in 61% of NSCLC cases and plays a crucial role in the pathogenesis of lung carcinoma [25]. Interestingly, amplification of FGF3/4/19 was found to be five times more common in smokers than in nonsmokers with lung squamous cell carcinoma (LSCC) [26]. Qi et al. reported that FGF4 induces EMT through inducing store-operated calcium entry in lung adenocarcinoma [27]. Zhang et al. reported that FGF19 activated ERK1/2 signaling to increase CCND1 activity, resulting in cell proliferation [28]. They also reported that FGF19 accompanied by GLI family zinc finger 2 (GLI2) induces EMT to promote LSCC metastasis [29]. Co-clinical trials have identified FGF3 and FGF19 as predictive biomarkers for the response to dovitinib in patients with LSCC [30]. Upregulation of FGF3/4/19 also significantly confers gefitinib resistance in NSCLC cells [26].

Analysis of FGFR1 and related ligand expression revealed positive correlations between FGFR1 mRNA expression and FGF9 mRNA levels, as well as its protein level [20], suggesting that FGF9 may serve as a novel unfavorable prognostic indicator and a potential therapeutic target [31]. A study by Arai D et al. reported that prolonged overexpression of FGF9 in the lung epithelium resulted in rapid adenocarcinosis in mice [32, 33]. Mechanistic studies demonstrated that FGF9-induced lung adenocarcinoma recruited and activated M2-biased tumor-associated macrophages (TAMs), which suppressed the immune functions of T cells and resulted in high expression of TGF-β, VEGF, FGF2, FGF10, and FGFR2, thereby supporting tumor growth [34]. Hegab et al. reported the contribution of cancer-associated fibroblasts (CAFs) in a mouse model. CAFs increase the number of M2-biased TAMs and secrete TFG-β, MMP7, FGF9, and FGF2 and inflammatory cell-recruiting cytokines, leading to lung adenocarcinoma progression [35]. In addition to exerting its role in the tumor microenvironment, FGF9 was found to activate FAK, AKT, and ERK signaling through FGFR1, inducing EMT to stimulate tumorigenesis and hepatic metastasis [36]. Reports indicate that the transdifferentiation of lung adenocarcinoma to SCLC occurs in a subset of lung cancer patients. Recently, Ishioka K et al. identified upregulation of FGF9 through an integrative omics analysis of paired tumor samples from a patient with transdifferentiated SCLC and validated this finding in four out of six paired samples [37]. FGF21 is overexpressed in NSCLC and facilitates cell growth and migration by activating the SIRT1/PI3K/AKT signaling pathway [38]. Collectively, these studies suggest that deregulation of FGF ligand signaling is associated with oncogenesis, tumor progression and resistance to anticancer therapies in lung cancer.

FGFRs belong to the RTK family in the human genome. The structure of FGFR is mainly composed of an extracellular ligand-binding domain, a transmembrane domain, and an intracellular tyrosine kinase domain [39]. The basic structure of FGF-FGFR complexes consists of two receptor molecules, two FGFs, and one partner heparan sulfate proteoglycan (HSPG) [40]. After FGFR binds to a ligand, it can induce the formation of FGFR dimers and catalyze its own phosphorylation, thereby activating key downstream signaling pathways [41]. FGFRs facilitate cell migration by recruiting and phosphorylating PLCγ, activating the MAPK and AKT pathways through PKC. Upon activation, FGFRs phosphorylate FRS2a, which binds to the adapter GRB2 [42]. GRB2 subsequently interacts with SOS and GAB1 via its SH3 domain, leading to the activation of the RAS/RAF/MAPK signaling pathway and Casitas B-lineage lymphoma (CBL) [43, 44]. Additionally, activated FGFRs stimulate phosphatidylinositol-3 kinase (PI3K) and STAT. Multiple studies have proposed the aberrant FGFR signaling pathway as a potential therapeutic target in lung cancer [45].

Fig. 1

Dysregulation of FGF signaling in lung cancer. (A) FGF1 overexpression enriches cancer stem cells and increases stem cell proliferation. It also promotes cancer vascular invasion. (B) FGF2 promotes cancer cell proliferation, EMT and metastasis through various signaling pathways. (C) FGF4 facilitates cell migration and invasion by inducing Ca²⁺ influx. (D) FGF19 increases CCND1 activity to promote cell proliferation and interacts with GLI2 to induce cell metastasis. (E) FGF9 supports tumor growth through regulating the tumor immune microenvironment and signal transduction pathways. (F) FGF21 facilitates cell growth and metastasis by activating the SIRT1/PI3K/AKT axis

FGFR genomic alterations in lung cancer

In recent years, significant advancements in sequencing technologies have led to the discovery of a wide range of FGFR genomic alterations [46]. These alterations include FGFR amplifications, activating mutations, and oncogenic fusions, which have been identified in various cancers at different frequencies. Helsten T et al. conducted an analysis of FGFR aberration frequencies in 4,853 solid tumors, revealing that 7.1% of cancers with FGFR aberrations comprised 66% amplifications, 26% mutations, and 8% oncogenic fusions [47]. Similarly, Gu W et al. examined FGFR alterations in 5557 solid tumors and reported that 9.2% of cancer cases presented with FGFR1-4 alterations, with gene amplifications (51.5%) and mutations (40.7%) being common, whereas gene rearrangements were less prevalent (10.0%) [48]. These studies suggest that FGFR inhibition could serve as a crucial therapeutic strategy and that molecular profiling holds significant promise for predicting the response to FGFR inhibition.

The most prevalent genomic alteration within the FGFR family is gene amplification, leading to the overexpression of FGFR and hyperactivation of downstream signaling pathways [49]. FGFR1, located on 8p11.23, regulates tumor cell proliferation. FGFR1 amplification has been detected in approximately 17% of patients with squamous cell NSCLC and serves as an independent poor prognostic marker [50]. Both in vitro and in vivo studies have demonstrated a response to FGFR inhibition in FGFR1-amplified NSCLC [51, 52]. FGFR3 amplification was also detected in NSCLC samples at a relatively low frequency [53].

Mutations in different regions of FGFR activate the FGFR signaling pathway through distinct mechanisms [54]. When FGFR mutations occur in the extracellular region or transmembrane domain, they alter FGFR specificity and/or affinity to FGF ligands or induce ligand-independent dimerization. However, when mutations occur in the tyrosine kinase domain, they constitutively activate FGFRs by disrupting autoinhibitory mechanisms [54]. Somatic variants of FGFR1 characterized by exon 1–8 deletions have been found in SCLC patients through deep genomic analysis [55]. FGFR2 mutations, which are detected in approximately 4–5% of NSCLC patients, exhibit different functional effects on the basis of the mutation location [56]. For example, the extracellular domain mutations W290C and S320C in FGFR2 increase receptor-ligand binding affinity, whereas the kinase domain mutations K660E and K660N lead to constitutive receptor dimerization [57]. In contrast, mutations E471Q and T787K in FGFR2 do not drive cell colony formation above wild-type levels. Conditional expression of an extracellular (W290C) or kinase domain (K660N)-activating mutation of FGFR2 induced high-grade lung adenocarcinoma and increased sensitivity to a pan-FGFR inhibitor in a genetically engineered mouse model [58]. Through targeted next-generation sequencing (NGS), a novel somatic insertion mutation, FGFR2 A266_S267ins, was identified in the extracellular domain of FGFR2 in a 37-year-old male nonsmoker with advanced NSCLC [59]. Additionally, the FGFR T730S and V755I mutations located within the kinase domain of FGFR2 were detected in two patients with NSCLC [59]. The FGFR3 extracellular domain mutations R248C and S249C have been identified in human LSCC and display oncogenic properties [60]. Conversely, mutations S435C and K717M in FGFR3 do not result in enhanced colony formation compared with that of wild-type FGFR3. Additionally, FGFR3 mutations have been reported in lung adenocarcinoma cases of Indian origin, including the previously described S294C mutation and the novel G691R mutation, both of which exhibit oncogenic activity [60]. For FGFR4, a single-nucleotide polymorphism (SNP) in FGFR4 (rs351855) causes a G388R substitution, leading to a conformational change in FGFR4, thus increasing STAT3 binding and significantly enhancing STAT3 signaling [61]. Through whole-exome and whole-genome sequencing, high frequencies of single nucleotide variants and copy number variants in FGFR2 have been identified in primary tumors of squamous cell lung carcinoma [62]. Gene fusions involving FGFRs have been found to activate the FGFR signaling pathway in NSCLC, albeit at a low incidence. In a study by Qin A et al., hybrid capture-based comprehensive genomic profiling was conducted on 26,054 consecutive formalin-fixed, paraffin-embedded NSCLC samples, revealing that 0.2% (52) of the samples were positive for FGFR fusions [63]. These included 37 cases of FGFR3-transforming acidic coiled-coil containing protein 3 gene (TACC3) fusion, two cases of FGFR2-shootin 1 gene (KIAA1598) fusion, and one case of BCL2 associated with the pseudogene 4 gene (BAG4)- FGFR1 gene fusion. Additionally, 11 novel FGFR fusions were identified in 12 patients, with seven involving FGFR2 fusions as the 5’ partner and one involving an FGFR3 fusion (FGFR3–pleckstrin homology-like domain family B member 3 gene [PHLDB3]) as the 5’ partner. Three novel fusions with FGFR as the 3’ partner and retention of the kinase domain were also identified (as shown in Table 1). Notably, a patient with an FGFR2-LZTFL1 fusion exhibited a partial response to the pan-FGFR inhibitor erdafitinib, with 60% tumor shrinkage after 2 months of therapy, and continued treatment with erdafitinib for 11 months. These findings suggest that FGFR fusions may serve as predictive markers for the response to FGFR inhibition and warrant further investigation. FGFR3-TACC3 fusions were also detected in tumor tissues following treatment with EGFR-TKIs, indicating that FGFR fusions may act not only as primary driver mutations but also as potential mechanisms of resistance to other targeted agents [64-66]. Consequently, clinical trials exploring the combination of FGFR inhibitors with EGFR-TKIs are currently underway.

Table 1 FGFR genomic alterations in lung cancer

Targeting FGFR in clinical studies of lung cancer

The involvement of aberrant FGFR signaling in tumorigenesis has spurred advancements in therapeutic strategies aimed at the FGFR pathway [67]. These approaches include small-molecule FGFR inhibitors that target the ATP-binding site of the tyrosine kinase domain of various growth factor receptors (multitarget TKIs), selective TKIs that specifically target the kinase domain of FGFRs, and biological agents (Fig. 2). We summarize the most recent advancements in the use of these drugs in the treatment of lung cancer.

Multitarget TKIs

Owing to the phylogenetic relevance of the FGFR, vascular endothelial growth factor receptor (VEGFR), and platelet-derived growth factor receptor (PDGFR) families in kinase domains, several nonselective TKIs initially designed to target VEGFRs have also been demonstrated to inhibit FGFR in vitro. Dovitinib, lucitanib, nintedanib, pazopanib, lenvatinib, and ponatinib are all multi-kinase inhibitors that target FGFR and exhibit activity against various RTKs beyond FGFRs. FGFRs as well as other RTKs regulate angiogenesis, immunity, and tumorigenesis [68]. Therefore, multi-kinase FGFR inhibitors have the potential to target both cancer cells and their microenvironment [69]. However, their broad spectrum of activity complicates the understanding of their mechanisms of action and potential adverse effects. Nintedanib, an oral triple angiokinase inhibitor, inhibits VEGFR 1–3, PDGFR-alpha, PDGFR-beta, and FGFR 1–3 [70]. The LUME-Lung-1 trial demonstrated a significant increase in progression-free survival (PFS) and overall survival (OS) in the docetaxel plus nintedanib group compared with the docetaxel plus placebo group [71]. On the basis of these promising clinical results, nintedanib has been approved by the European Medicines Agency for use in combination with docetaxel for treating adult patients with locally advanced, metastatic, or locally recurrent NSCLC of adenocarcinoma histology following first-line chemotherapy. The LUME-Lung-2 trial was subsequently initiated to assess the efficacy and safety of nintedanib plus pemetrexed in patients with pretreated non-squamous NSCLC [72]. Unfortunately, recruitment was halted after enrolling 713 out of the planned 1300 patients on the basis of a preplanned futility analysis of investigator-assessed PFS. Subsequent analysis revealed that nintedanib plus pemetrexed improved PFS. The treatment options for advanced squamous NSCLC are limited, and these patients have poor prognoses [73]. Thus, the LUME-Lung-3 trial was conducted to evaluate the safety and pharmacokinetics of nintedanib plus cisplatin/gemcitabine as a first-line treatment for these patients [74]. The results indicated that five patients achieved partial response, eight had stable disease, and three patients survived for 17–35 months. The safety profile of nintedanib (200 mg bid plus cisplatin/gemcitabine) was manageable, with adverse events consistent with previous observations. These preliminary positive outcomes warrant further studies involving more patients across multiple clinical centers.

Idiopathic pulmonary fibrosis (IPF) is a chronic and progressive lung disease characterized by scarring of the lung tissue [75]. It is a fatal disease with a poor prognosis [76]. The scarring of lung tissue in IPF can lead to a decline in lung function, resulting in symptoms such as shortness of breath, persistent cough, fatigue, and eventually respiratory failure [77]. Acute exacerbation is the primary cause of mortality in those affected with IPF [78]. In replicated, randomized, placebo-controlled phase III trials, nintedanib significantly reduced the annual decline in forced vital capacity and delayed the onset of acute exacerbation in individuals with IPF [79]. On the basis of these findings, nintedanib has received approval for treating IPF in the USA, European Union, and Japan [80]. Otsubo K et al. conducted a phase III trial to assess nintedanib plus chemotherapy versus chemotherapy alone for advanced NSCLC patients with IPF [81]. Although there was no significant difference in exacerbation-free survival between the two groups, nintedanib plus chemotherapy improved overall survival in patients with non-squamous histology. Ikeda S et al. evaluated the safety and efficacy of nintedanib plus chemotherapy for unresectable SCLC patients with comorbid IPF [82]. The incidence of IPF acute exacerbation 28 days after the last chemotherapy dose was 3.0%, meeting the primary endpoint. Therefore, carboplatin, etoposide, and nintedanib combination therapy could become a standard treatment for SCLC patients with comorbid IPF.

Dovitinib (TKI258) has demonstrated modest efficacy and a manageable safety profile in various cancers, including NSCLC[83–85]. In a single-arm phase II trial (NCT01861197), the efficacy and safety of dovitinib were evaluated in 26 previously treated patients with advanced squamous NSCLC featuring FGFR1 amplification [86]. The results revealed an overall response rate of 11.5% and a disease control rate of 50%, with 3 patients achieving partial response. The most common grade 3 or higher adverse events included fatigue (19.2%), anorexia (11.5%), and hyponatremia (11.5%). Additionally, co-clinical trials have been conducted to identify predictive biomarkers for dovitinib efficacy [30]. Five patient-derived xenograft models were established from LSCC patients who were enrolled in a phase II trial of dovitinib (NCT01861197) and underwent surgery. The results revealed that FGFR1-3 alterations are not associated with sensitivity to dovitinib. Notably, differentially expressed genes, such as FGF3 and FGF19, were identified between the dovitinib-sensitive and dovitinib-resistant models. In addition, two FGFR-related signaling pathways were enriched, and FGFR ligand binding/activation and the SHC-mediated cascade pathway were substantially upregulated in dovitinib-sensitive models compared with dovitinib-resistant models. Another phase II trial investigated the efficacy of dovitinib in advanced NSCLC patients who had progressed on anti-VEGF therapy [87]. However, the ten patients included in this study showed no benefit, and significant toxicity was observed. All patients experienced at least one grade ≥ 3 treatment-related adverse event.

Pazopanib (votrient) is a multitarget drug currently approved for treating metastatic renal cell cancer and soft-tissue sarcoma [88, 89]. In a single-arm phase II trial, its efficacy was evaluated in patients with advanced NSCLC of non-squamous histology who had experienced disease progression on bevacizumab-containing therapy [90]. The observed disease control rate was 13% (2/15), leading to the trial being halted owing to futility. The efficacy of pazopanib as maintenance treatment following standard first-line platinum-based chemotherapy in advanced NSCLC patients was subsequently investigated [91]. The EORTC 08092 study, a double-blind randomized phase III trial, was terminated because of a lack of efficacy after the median overall survival and progression-free survival rates were analyzed. Moreover, pazopanib did not improve relapse-free survival in completely resected stage I NSCLC patients [92]. Interestingly, second-line treatment with pazopanib in platinum-sensitive SCLC was well tolerated and resulted in promising objective responses and disease control [93]. Additionally, pazopanib maintenance significantly prolonged progression-free survival in SCLC patients with extensive disease, as evidenced by the results from the KCSG-LU12-07 phase II study [94]. Dynamic monitoring of changes in different subpopulations of circulating tumor cells during pazopanib treatment revealed its potential to eliminate various CTC subpopulations in patients with relapsed SCLC [95]. Furthermore, the detection of VEGFR2⁺CTCs during treatment could serve as a surrogate marker associated with resistance to pazopanib [96].

Ponatinib (AP24534) has received approval for treating chronic myelogenous leukemia and acute lymphoblastic leukemia [97]. A phase II trial (NCT01761747) was initiated to investigate FGFR1 amplification and mRNA expression as predictive markers for ponatinib sensitivity in patients with NSCLC and SCLC [98]. However, owing to the poor tolerability and safety concerns associated with ponatinib, only a small number of patients (n = 4) could be treated with the drug, leading to early termination of the trial without reaching its patient accrual target. Among the four patients, two exhibited stable disease, while the remaining two experienced disease progression.

On the basis of the abovementioned clinical trial outcomes, the combination of nintedanib and chemotherapy has therapeutic potential for patients with refractory squamous cell carcinoma of NSCLC. Moreover, this treatment regimen may be effective for patients with NSCLC, regardless of whether it is squamous cell carcinoma or adenocarcinoma accompanied by IPF. However, although pazopanib has an unsatisfactory clinical effect on patients with NSCLC, it has favorable efficacy and safety in patients with SCLC and may constitute a potential second-line treatment option (Table 2).

Table 2 Clinical trials of multitarget TKIs in lung cancer

Selective FGFR inhibitors

To enhance on-target FGFR inhibition in patients with FGFR alterations and minimize off-target toxicity, researchers have developed selective inhibitors that target FGFRs. However, owing to the high structural similarity in the kinase domain of FGFR1-3, some selective inhibitors effectively target al.l three FGFRs to varying degrees [39]. These selective FGFR inhibitors can be categorized as pan-FGFR, FGFR1/2/3, and FGFR2/3 inhibitors, as well as selective FGFR2, FGFR3, or FGFR4 inhibitors. Examples of pan-FGFR inhibitors include erdafitinib [99], futibatinib [100], rogaratinib [101], and resigratinib [102]. On the other hand, FGFR1/2/3 inhibitors include pemigatinib [103], infigratinib [104], fexagratinib [105], and zoligratinib [106]. Compound 19 [107] is an FGFR2/3 inhibitor, whereas lirafugratinib [108] is a selective FGFR2 inhibitor. LOXO-435 is a selective FGFR3 inhibitor [109], and roblitinib [110] and fisogatinib [111] are selective FGFR4 inhibitors. Notably, erdafitinib, futibatinib, infigratinib, and pemigatinib have received accelerated approval in second-line or later-line settings for several cancers, although not for lung cancer. Ongoing clinical trials are evaluating the efficacy and safety of selective FGFR inhibitors for the treatment of lung cancer patients with FGFR abnormalities (Table 3). Erdafitinib received approval in the USA in 2019 for treating locally advanced or metastatic FGFR3 or FGFR2 urothelial carcinoma [6]. Interestingly, sporadic reports have shown that patients with FGFR2 or FGFR3 alterations in NSCLC respond to erdaftinib and demonstrate positive therapeutic effects. Pham C et al. reported successful erdaftinib treatment and retreatment[112]. In detail, this patient with squamous NSCLC experienced three previous failed treatments. A biopsy from the left upper lobe of the tumor was sent for NGS, which revealed an FGFR3-TACC3 fusion. Thus, FGFR inhibitors were recommended. The CT imaging results revealed that treatment with erdafitinib controlled the tumor for 11 months. Gemcitabine was subsequently administered for 4 months, and vinorelbine was administered for 1 month to control the condition. At this time, a biopsy was performed on the lung mass in the upper right lobe. The NGS results revealed an FGFR3–TACC3 fusion, the FGFR inhibitor erdafitinib was readministered, and the disease control time reached 8 months. Urrutia Argueta SA et al. reported a case of FGFR2-BICC1 fusion adenocarcinoma NSCLC in a patient who demonstrated stable disease for 8 months on initial therapies, followed by an additional 5 months of stability with chemotherapy [113]. The patient then achieved 12 weeks of disease control with erdafitinib treatment. Further exploration of erdafitinib treatment for patients with FGFR2 or 3 fusion NSCLC could broaden therapeutic options for these individuals. Notably, Haura EB et al. reported that erdafitinib can overcome FGFR3-TACC3-mediated resistance to osimertinib [114]. Therefore, the addition of erdafitinib in the treatment of patients with FGFR gene fusions who are progressing on osimertinib is crucial, and prospective clinical trials are warranted to validate this approach.

Preclinical studies have shown that the efficacy of rogaratinib is closely associated with FGFR mRNA expression levels [101]. In a phase I study conducted by Schuler M et al. in patients with advanced cancers selected on the basis of FGFR mRNA expression, a phase 2 dose of rogaratinib at 800 mg twice daily was recommended [115]. Among the 20 FGFR mRNA-positive NSCLC patients, one patient achieved a partial response, 15 patients had stable disease, and four experienced progressive disease. The median progression-free survival was 98 days. A serious treatment-related adverse event of acute kidney injury was observed in one patient, and hyperphosphatemia was noted as the most common adverse event [115]. The clinical data indicate that rogaratinib has both a favorable safety profile and clinical efficacy in NSCLC patients on the basis of FGFR1-3 mRNA expression levels. As a result, there is a strong rationale for conducting further investigations into the potential of rogaratinib in this unique patient population.

Infigratinib (BGJ398) is a selective FGFR1-3 kinase inhibitor with half-maximal inhibitory concentrations ranging from 0.9 to 1.4 nM for FGFR1-3 and 60 nM for FGFR4. It has received approval for the treatment of FGFR2 fusion cholangiocarcinoma [9]. A global phase I study revealed that out of 36 patients with FGFR1-amplified squamous NSCLC, 4 patients achieved a partial response, and 14 patients had stable disease. Common adverse events include hyperphosphatemia, constipation, decreased appetite, and stomatitis [116]. Pemigatinib, a selective inhibitor of FGFR 1–3, has been approved for adults with previously treated, unresectable, locally advanced, or metastatic cholangiocarcinoma with FGFR2 fusions or other rearrangements [8]. The phase I/II FIGHT-101 study reported that 1 patient achieved a partial response out of 10 enrolled NSCLC patients [117]. A phase I study of an FGFR1-3 inhibitor, fexagratinib (AZD4547), revealed that only 1 patient achieved a confirmed partial response, and four patients had stable disease out of 13 FGFR1-amplified patients [118]. The SWOG S1400D phase II study is evaluating AZD4547 in FGFR-altered squamous NSCLC, showing an acceptable safety profile, but only 1 patient with FGFR1 amplification had a confirmed partial response [119]. Several other selective FGFR inhibitors, such as ASP5878 and LY2874455, are undergoing phase I studies to evaluate safety and determine the recommended dose for phase II clinical trials [120, 121].

Existing clinical trial studies have demonstrated that selective FGFR inhibitors display therapeutic efficacy for patients with various FGFR alterations (Table 3). In particular, erdafitinib can overcome osimertinib resistance mediated by the FGFR3-TACC3 fusion. Thus, precise detection of genetic alterations in the tumor tissues of patients with lung cancer is important. Further studies are needed to fully evaluate the efficacy and safety of selective FGFR inhibition in patients with lung cancer.

Table 3 Clinical trials of selective FGFR inhibitors in lung cancer

Biological agents (targeting FGF signaling)

GSK3052230 (FP-1039) is a soluble fusion protein that functions as a ligand trap by sequestering FGFs [122]. In a phase IB study, GSK3052230 was assessed in combination with first-line chemotherapy (paclitaxel and carboplatin) or docetaxel in FGFR1-amplified NSCLC [123]. As anticipated, adverse events typically associated with small-molecule FGFR inhibitors were not observed, and the combination of GSK3052230 with chemotherapy was well tolerated. RBM-007 is an anti-FGF2 aptamer consisting of 37 single-stranded short oligonucleotides [124] that strongly and specifically binds to FGF2. It does not cross-react with other FGF family members, selectively blocking the interaction between human FGF2 and its receptors. Preclinical studies have shown that RBM-007 significantly inhibits PC9 gefitinib-resistant cell proliferation and induces apoptosis [125].

Fig. 2

Targeting FGF/FGFR signaling in lung cancer. Drugs targeting FGFR that are at the clinical stage include multitarget TKIs (nintedanib, dovitinib, pazopanib, and ponatinib) and selective TKIs (pan-FGFR inhibitors such as erdafitinib and rogaratinib; FGFR1/2/3 inhibitors such as pemigatinib, infigratinib and fexagratinib). FP-1039, which targets FGF, has also been investigated in clinical trials. New strategies for targeting FGF/FGFR include the use of FGFR degraders (DGY-09-192 and LC-MB12), specific FGFR inhibitors and anti-FGF2 aptamers

Challenges and opportunities

Targeting FGFR for lung cancer treatment faces several challenges. One of the main issues is identifying the right patient population who will benefit the most from FGFR-targeted therapy owing to the heterogeneity of FGFR alterations. Another challenge is the potential for off-target effects and toxicity associated with FGFR inhibitors, which can limit their therapeutic efficacy. Furthermore, the optimal dosing and duration of treatment with FGFR inhibitors are still being investigated to maximize their clinical benefits while minimizing adverse effects. Finally, the development of resistance to FGFR inhibitors over time can be challenging, leading to treatment failure.

Challenges of patient selection

Different FGFR aberrations respond differently to FGFR inhibitors. Furthermore, FGFR aberrations in individuals are complicated. Thus, the challenge in clinical trials is to select the best target patients with specific FGFR aberrations prospectively.

On the basis of preclinical data, FGFR1 amplification is recognized as a biomarker for predicting the response of advanced LSCC patients to FGFR inhibitors. As a result, FGFR1-amplified patients with advanced LSCC have often been enrolled in clinical trials. However, the results from clinical trials revealed that only 8–11% of patients with FISH-detected FGFR1 amplification respond to FGFR kinase inhibitors [86, 118]. The possible reason is that FISH-detected FGFR1 amplification in the samples was not consistently detected through deep genomic sequencing. In addition, discrepancies in the gene amplification and protein expression of FGFR1 have also been reported. Wynes MW et al. analyzed FGFR1 gene copy number (GCN) and mRNA levels in a tissue microarray composed of resected lung tumors. The results revealed that 22% of adenocarcinomas and 28% of SCCs presented high FGFR1 mRNA expression. Notably, only 46% of SCCs with FGFR1 GCN gains expressed high mRNA levels, implying that FGFR1 mRNA rather than GCN may serve as a superior biomarker for predicting the FGFR TKI response in lung cancer [126]. Bogatyrova O et al. investigated the molecular characteristics of FGFR1 expression in 635 NSCLC samples and demonstrated a weak correlation between FGFR1 copy number gain and FGFR1 gene and protein expression [127]. NSCLC PDX models with FGFR1 amplification and FGFR1 protein overexpression are highly sensitive to the FGFR inhibitor M6123 [127]. In addition, tumors overexpressing the FGFR1 protein are accompanied by an absence of driver alterations, such as EGFR and KRAS mutations, and an immunosuppressive microenvironment, such as reduced infiltration of T lymphocytes and lower PD-L1 expression. The distinctive molecular and immune characteristics of tumors with high FGFR1 expression offer a basis for stratifying patients in future clinical trials of drugs that target FGFR1.

Malchers F et al. performed an in-depth genomic and functional study of primary human SCLC, cancer cell lines, and patient-derived xenografts [55]. The results revealed that intragenic rearrangements of FGFR1 lead to ectodomain-deficient variants of FGFR1, increasing sensitivity to FGFR inhibitors. Furthermore, tail-to-tail rearrangements close to FGFR1 usually result in FGFR1-centered amplification, similarly linking sensitivity to FGFR inhibition. Taken together, these findings indicate that FGFR1 rearrangements, which are novel somatic genomic alterations, may define the sensitivity of squamous cell lung cancer to FGFR inhibitors and. may be novel predictive biomarkers for FGFR1-dependent therapies.

Owing to the inconsistent results of FGFR1 amplification detected by different methods, targeting DNA/RNA-based high-throughput sequencing platforms may be a better choice for detecting FGFR gene alterations. In addition, identifying new appropriate biomarkers for predicting patient sensitivity to FGFR-targeted inhibitors is crucial.

Toxicities

Multitarget TKIs have the potential to increase treatment efficacy through simultaneous blockade of redundant oncogenic pathways. However, they increase off-target toxicities and limit doses. Therefore, selective FGFR inhibitors have been developed. Clinical studies have revealed that hyperphosphatemia occurs in approximately 55–85% of enrolled patients who receive selective FGFR inhibitors [128]. Owing to the high probability of hyperphosphatemia, clinical monitoring is carried out by detecting the serum phospholipid level. Moreover, the phospholipid level in the body is regulated by changing the diet [129] or administering drugs [130]. If the hyperphosphorus state of a patient fails to be alleviated, the dosage of selective FGFR inhibitors should be decreased or even interrupted [128]. Studies have shown that the bone-derived hormone FGF23 plays a central role in regulating phosphate homeostasis [131, 132]. It lowers blood phosphate predominantly through binding to the FGFR1 receptor on the kidney, promoting phosphaturia. When the FGF23/FGFR1 pathway is blocked by FGFR inhibitors, reduced FGF23 bioactivity results in hyperphosphataemia [133]. Therefore, the inhibition of FGFR1 by selective FGFR inhibitors has led to inevitable hyperphosphatemia, limiting the clinical benefits of such drugs.

Resistance mechanisms to FGFR-targeting agents in lung cancer

Although targeting FGFR shows promise in the treatment of lung cancer, resistance to these inhibitors has become a major concern.

Gatekeeper residues are located within the Hinge region of the ATP-binding pocket of kinases, exerting a crucial impact on regulating the accessibility of TKIs to the hydrophobic ATP binding pocket and facilitating the active conformation of kinases by stabilizing the hydrophobic spine [134]. Typically, gatekeeper residues consist of relatively small, hydrophobic amino acids, frequently including valine. Nevertheless, under treatment with FGFR inhibitors, these residues may undergo mutation to larger residues (e.g., methionine, leucine, isoleucine), effectively impeding the access of inhibitors to the hydrophobic pocket. Sohl and colleagues investigated the structural and kinetic attributes of V561M FGFR1, an indicative gatekeeper mutation, and discovered that it confers a 38-fold increase in autophosphorylation. This enhancement is achieved at least in part by a network of interacting residues forming a stable hydrophobic spine that supports the active conformation. Furthermore, kinetic assays revealed that this mutation significantly augments resistance toward E3810 (a dual FGFR-VEGFR inhibitor) while retaining affinity for AZD4547 [135]. Structurally, E3810 exhibits limited flexibility, unlike AZD4547, which boasts greater conformational flexibility, allowing adjustment to accommodate mutations [136]. Moreover, Ryan et al. employed both in vivo and in vitro binding assays and demonstrated that cancer cells expressing FGFR1 V561M trigger strong activation of the STAT3 signaling pathway, resulting in substantial resistance to AZD4547 and ultimately promoting cancer progression [137]. However, knocking out STAT3 restored sensitivity to AZD4547 [137]. These findings underscore the profound importance of comprehending downstream signaling pathways as well as kinase-inhibitor complex structures. An exclusive reliance on Kit may not be sufficient to provide a comprehensive understanding of the mechanisms governing FGFR-TKI resistance.

However, no clinical studies have reported a correlation between FGFR inhibitor resistance and FGFR2 gatekeeper mutations in lung cancer. Nonetheless, a clinical study revealed the presence of the FGFR2 V564F mutation in all three cholangiocarcinoma patients who developed acquired resistance to BGJ398 therapy [138, 139]. Goyal et al. reported that TAS-120, an irreversible covalent FGFR inhibitor, can effectively overcome resistance to BGJ398 [140]. Importantly, FGFR3 is structurally similar to FGFR1. Specifically, the FGFR1 V561M gatekeeper mutation corresponds to the FGFR3 V555M gatekeeper mutation. Notably, a polykinase inhibitor, TKI258 (dovitinib), has shown potential for patients with both types of mutations; however, it is important to highlight that the V561M mutation significantly reduces sensitivity to PD173074 and AZD4547. Initially, patients harboring an FGFR3-TACC3 fusion demonstrated high sensitivity to AZD4547; however, this sensitivity diminished after the acquisition of a TKI-induced resistant phenotype due to the emergence of certain mutations. Moreover, structural modeling indicates that these mutations lead to increased resistance by generating steric clashes within specific molecular structures. Notably, conformational changes within the P-loop region may also contribute to increased levels of resistance [141]. Currently, studies of the selective FGFR4 inhibitor fisogatinib in hepatocellular carcinoma have been conducted. Hatlen et al. identified gatekeeper mutations (V550M/L) and Hinge-1 (C552) mutations in the FGFR4 kinase domain among patients receiving fisogatinib treatment, and these mutations prevent fisogatinib from covalently binding to FGFR4 [111]. Wu D et al., however, reported that LY2874455, a Pan-FGFR inhibitor, exceptionally overcomes the fisogatinib resistance induced by FGFR4 gatekeeper mutations [142]. Furthermore, crystallographic experiments confirmed that the binding site of LY2874455 is distant from the gatekeeper residue, thereby avoiding steric clashes with the ATP-binding pocket of FGFR4 [143].

Bypass and downstream pathway activation play important roles in resistance to FGFR inhibitors, similar to the resistance mechanism of EGFR-TKIs. These signaling pathways include other membrane RTK pathways and the downstream PI3K-AKT and RAS-MAPK pathways (Fig. 3).

Alternative activation of MET has been identified as a contributing factor to the resistance observed in FGFR-targeted therapies [144]. Kim et al. demonstrated that MET promoted FGFR-TKI resistance through activating the PI3K/AKT signaling pathway [145]. Furthermore, MET is capable of compensating for FGFR signals by activating downstream signaling pathways [146].

Activating the PI3K-AKT signaling pathway downstream of FGFR signaling is a well-established mechanism for inducing resistance to FGFR-TKIs [147]. Jharna et al. reported that the phosphorylation levels of both AKT and GSK3 were elevated in BGJ398-resistant DMS114 cells [148]. The regulation of the PI3K-AKT signaling pathway can be directly influenced by PHLDA1 and phosphatase and tensin homolog (PTEN), resulting in TKI resistance [149]. PHLDA1 competes with AKT for binding to PIP3, thereby inhibiting its activity [150].

Increased activation of the RAS-MAPK pathway has been identified as a critical factor contributing to resistance to FGFR-TKIs [151]. Kas et al. demonstrated that inactivation of Ras p21 protein activator 1 (RASA1) causes resistance to the FGFR inhibitor AZD4547 [152]. RASA1 knockdown activates the RAS-MAPK pathway, whereas restoring RASA1 inhibits both the MAPK and PI3K signaling pathways [153]. Furthermore, dysregulation of NRAS and DUSP6 has been identified in FGFR inhibitor-resistant lung cancer cells. The amplification of NRAS [154] and deletion of DUSP6 [155] result in reactivation of the MAPK pathway, thereby conferring resistance to FGFR inhibitors. Genomic analysis of cell lines sensitive and resistant to the selective FGFR inhibitor CPL304110 revealed increased p38 expression/phosphorylation with increased MAPK signaling in the resistant phenotype [156], suggesting that p38 MAPK serves as a pivotal driver of resistance to selective FGFR inhibition. Additionally, the HGF/MET-Pyk2 signaling axis confers resistance to CPL304110 in lung cancer cells [157].

Nishijima N et al. compared the miRNA profile changes between nintedanib-sensitive and resistant NSCLC cell lines and reported that miR-200b and miR-141 may predict the response to nintedanib [158]. Englinger B et al. investigated resistance mechanisms to nintedanib using a nintedanib-resistant SCLC cell line [159]. The results revealed that the multidrug resistance transporter ABCB1 was overexpressed and that the endothelin-A receptor (ETAR) signaling axis was hyperactivated after nintedanib resistance. Combining ETAR antagonists is a potential strategy to overcome nintedanib resistance. Interactions between cancer cells and the tumor microenvironment that mediate acquired resistance to FGFR inhibitors have been reported. Wang X et al. reported that cytokine reprogramming drives FGFR inhibitor resistance [160]. In detail, macrophages and fibroblasts interact with cancer cells to exaggerate secretome release. The secretome activates the transcription factor STAT3, enhancing resistance to FGFR inhibitors. Yang Z identified PLK1 as a synthetic lethal target in FGFR1-amplified lung cancer cells treated with AZD4547 via pooled single-guide RNA screening [161]. PLK1 overrides DNA damage and cell cycle arrest to mediate AZD4547 resistance. Antagonizing PLK1 in combination with an FGFR inhibitor synergistically induces cancer cell death by activating the γh2ax-CHK-E2F1 axis (Fig. 3).

Fig. 3

Resistance mechanisms of targeted FGFR therapies in lung cancer. In addition to FGFR mutations, changes in signaling pathways and the tumor microenvironment confer resistance to targeted FGFR therapies in lung cancer. (A) Bypass and downstream pathway activation. (B) Multidrug resistance pathway activation and DNA damage inhibition. (C) Cytokine reprogramming in the tumor microenvironment

Potential opportunities

Recently, degradation strategies have sought to utilize heterobifunctional molecules to recruit an E3 ubiquitin ligase complex to target proteins for ubiquitination and subsequent proteasome-mediated degradation. This approach aims at enhancing target selectivity compared with their parental inhibitors [162, 163]. For example, the FGFR degrader DGY-09-192 has shown dual degrading activity toward subtypes FGFR1 and 2 [164]. However, its efficacy is not reflected in its interaction with FGFR3 and 4. In addition, LC-MB12 stands out as a potent and orally bioavailable FGFR2 degrader that combines BGJ398 with a CRBN binder [165]. Notably, this compound exhibits superior potency in suppressing FGFR signaling while also displaying notable anti-proliferative activity compared with the parental inhibitor, indicating its potential for delivering greater clinical benefits. Overall, the advancements highlighted indicate that significant strides are being made in addressing resistance mechanisms associated with membrane RTK activation while also paving the way for more effective therapeutic interventions targeting such pathways.

Current therapies targeting the FGFR pathway by exploiting small-molecule kinase inhibitors have been linked to adverse events, specifically hyperphosphatemia and diarrhea, due to the undesirable inhibition of FGFR1 and FGFR4. Compared with current treatments, isoform-specific inhibitors for FGFR2 and FGFR3 that spare FGFR1 and FGFR4 could offer a more favorable toxicity profile and an improved therapeutic window. Nguyen MH et al. reported that Compound 19 effectively inhibited FGFR2 and FGFR3, as well as the gatekeeper mutant FGFR3-V555L, with high selectivity for both FGFR1 and FGFR4 [107]. Similarly, Subbiah V et al. designed RLY-4008 as an exceedingly selective inhibitor of FGFR2 that targets primary alterations and resistance mutations while inducing tumor regression without affecting other members of the FGFR family [108]. Combination therapeutic approaches may hold the key in overcoming the limitations associated with single-agent inhibition of FGFR in lung cancer. Blocking PI3K-AKT signaling through the AKT inhibitor GSK2141795 or siRNA intervention has shown promise in restoring the sensitivity of drug-resistant cell lines to BGJ398 [148]. Moreover, co-inhibition of the FGFR and MAPK pathways via the use of FGFR inhibitors in combination with the MEK inhibitor trametinib has demonstrated efficacy in inducing tumor degradation in xenografts derived from mesenchymal-like KRAS mutant cancer cell lines as well as patient-derived xenograft models [151].

Conclusions

The remarkable array of FGFR-activating mechanisms presents unique opportunities for the clinical application of FGFR inhibitors, with erdafitinib, futibatinib, pemigatinib and infigratinib already approved for patients harboring specific cancers with FGFR alterations. However, challenges encountered in preclinical and clinical studies of FGFR inhibitors in lung cancer have hindered their implementation in this context. Precision-based identification of patient populations with FGFR abnormalities is paramount. The occurrence of certain adverse events, such as hyperphosphatasemia, and the development of resistance pose significant hurdles to effective FGFR-targeted therapies. Accordingly, the development of novel FGFR-targeted drugs such as degraders or those specifically targeting FGFR2 or 3 holds promise for mitigating toxicity and overcoming resistance. Additionally, combination strategies involving FGFR-targeted drugs and other agents may represent a compelling therapeutic approach. Notably, FGFs play crucial roles in maintaining systemic homeostasis and possess therapeutic potential in diverse human diseases [166]. For example, recombinant FGFs are used for tissue and wound repair, and FGF analogs are under investigation in clinical trials for metabolic diseases [12, 167]. FGF traps and monoclonal antibodies show potential for cancer treatment. Therefore, actively exploring the role and mechanism of FGFs in lung cancer will provide support for future FGF drug therapy or adjuvant treatment of this cancer.