FGFR inhibitor resistance in cholangiocarcinoma: current understanding and future directions

KEY WORDS:

• biliary tract cancer
• biomarkers
• precision medicine
• resistance mechanisms
• targeted therapy
• FGFR

1. Introduction

The options for a molecular targeted therapy of biliary tract cancer (BTC) have dramatically changed in recent years [1]: Coming from the ‘middle of nowhere’ with only standard gemcitabine treatment available, possible therapeutic strategies broadened from combination chemotherapy like FOLFOX and most recently with durvalumab in combination with gemcitabine and cisplatin to more sophisticated targeted therapy approaches, especially in cases of BTCs with detected and druggable alterations like mutations in isocitrate dehydrogenase (IDH) or fibroblast growth factor receptors (FGFR). Especially, FGFR inhibitor therapy in BTC with overexpression, amplification, or mutations seems a promising therapeutical approach [2]. Interestingly, these alterations are nearly exclusively found in intrahepatic cholangiocarcinoma (iCCA), while extrahepatic cholangiocarcinoma is rather related to PRKACA and PRKACB fusions or to ELF3 and ARID1B mutation. Gallbladder cancer is associated to EGFR, ERBB3, PTEN, ARID2, MLL2, MLL3, and TERT promoter mutations [3].

Patients with BTCs are mostly diagnosed in late stages of disease with only limited therapeutic success. Efficacy of standard chemotherapy and of targeted agents is impaired by different mechanism of resistance before and during therapy which still gives a very poor outcome to BTC [4].

We will here focus on FGFR-based therapies where initial results with a biomarker-based therapeutic approach were encouraging although intrinsic and acquired resistance mechanisms dampen the overall success rate. Our increasing understanding of ‘who is the right patient’ for this treatment may improve the efficacy of these novel drugs.

2. Resistance to FGFR inhibitors

Several small and large molecule inhibitors of FGFR signaling were recently explored in clinical trials and lead to approvals in solid tumors. For BTC, especially iCCA, futibatinib, infigratinib, and pemigatinib received FDA approval as second-line treatment of iCCA. These pan-FGFR inhibitors achieved objective response rates (ORR) of up to approximately 40%, mean progression free survival (mPFS) of approximately 7 months and mean overall survival mOS of approximately 21 months when selecting patients based on FGFR2 fusions or rearrangements [1,5]. Although these initial response rates and survival times are encouraging for iCCA patients, they indicate that most patients do not derive long-term or even no benefit from this treatment, despite being FGFR2-fusion-positive.

This poses two questions:

1. Do we have the best patient selection biomarker?

2. Do we understand mechanisms of acquired resistance to FGFR inhibitors?

The focus to answer ‘Who is the right patient?’ was so far laid more on the first question as patients amenable to FGFR inhibitor treatment are usually selected based on genomic alterations in FGFR genes (e.g. rearrangements, point mutations or copy number variations). Whereas in many other solid tumors all of the aforementioned FGFR DNA alterations confer drug sensitivity toward FGFR inhibitor treatment, sometimes even irrespective of the FGFR subtype or the FGFR inhibitor investigated [6–8], the benefit in iCCA seems to be mostly limited to patients with an FGFR2 fusion or rearrangement [9] as so far only one patient with an FGFR2 mutation achieved a partial response [10]. Although FDA approval as a companion diagnostic test was received for the detection of FGFR2 fusions in iCCA patients, the response rate and the duration of response in FGFR2-fusion bearing patients clearly lags behind as compared to what is achieved in patients with ALK1 or NTRK fusions upon treatment with respective inhibitors [5]. As a clinical proof, data from a clinical trial with pemigatinib in second-line treatment of iCCA indicate that less than half of the patients derived clinical benefit, implying primary resistance mechanisms. A retrospective analysis of genomic profiling data of patients from the FIGHT-202 trial suggests that co-occurring alterations may explain the lack of response. Here, patients with BAP1 alterations or with alterations in CDKN2A/B or PBRM1 had shorter median progression-free survival (mPFS). Patients with TP53 alterations showed significantly inferior mPFS and no tumor responses. Furthermore, the presence of tumor suppressor gene loss (BAP1, CDKN2A/B, TP53, PBRM1, ARID1A, or PTEN) predicted a significantly shorter mPFS. Future analyses of clinically annotated genomic data from clinical trials and real-world patients will likely reveal more granular underpinning of primary resistance mechanisms and help develop treatment regimens targeting multiple signaling pathways leading to improved tumor response and patient outcome [11].

Rapidly emerging clonal selection of mutations that mediate resistance have been described for other small molecule tyrosine kinase inhibitors, e.g. T790M for first-generation EGFR inhibitors, or G120R for ALK inhibitors and have led to the development of second- or third-generation inhibitors with improved properties (including brain penetration) [12]. Surprisingly, little data are available on mechanisms of acquired resistance upon FGFR inhibitor treatment. Next-generation sequencing-based serial analysis of circulating free tumor DNA (cfDNA) to elucidate the molecular basis of the acquired resistance mechanisms reported widespread development of polyclonal resistance mutations primarily converging on the intracellular kinase domain of FGFR2. Most of the currently available FGFR inhibitors suppress the FGFR2 fusion protein-driven oncogenic signaling by competing with ATP at the binding pocket of the kinase domain. Consequently, mutations affecting the ATP binding site (gatekeeper mutations) prevent the binding of FGFR inhibitors due to steric hindrance [11]. To overcome this acquired resistance, compounds that bypass these gatekeeper mutations would be required. Futibatinib, an irreversible covalent FGFR 1–4 inhibitor, retained activity against several gatekeeper mutations by altering the conformational dynamics of FGFR2. In the analysis of the most common mutations in cell lines, futibatinib was active against all except the FGFR2 p.V565F gatekeeper mutation. Furthermore, a clinical trial conducted in 2018 (NCT02052778) in 45 patients previously treated with chemotherapy or prior FGFR inhibitors showed definitive clinical activity of futibatinib against tumors resistant to primary therapy and proved that out of 28 patients with FGFR2 gene fusions, 20 (71%) experienced tumor shrinkage and 7 had confirmed partial responses [13]. So far, an isotype switch from e.g. FGFR2 to FGFR3 is currently not an issue due to pan-FGFR activity of available compounds.

Development of resistance as a result of an alternative pathway upregulation opens opportunities for combination approaches. In vitro, inhibition of FGFR2 enhances chemosensitivity to gemcitabine in cholangiocarcinoma through the AKT/mTOR and EMT signaling pathways [14]. PI3K pathway mutations detected in post-progression and autopsy biopsies confirmed the heterogeneity of these mutations and the complexity of treating affected patients. Early clinical data with a combination of mTOR and FGFR inhibitors appear encouraging in iCCA resistance to FGFR inhibitors [11]. It is intriguing to see that FGFR signaling has been implicated as a resistance mechanism to EGFR inhibitor treatment [15,16] and the reverse finding was observed in an experimental setting in FGFR3-dependent cell lines after incubation with infigratinib [17]. Preclinically, activation or alteration of various downstream signaling molecules like c-MET, c-Myc, AKT, or GSK3 have been described after FGFR inhibitor treatment [18–21].

Immune checkpoint inhibitors have become a mainstay of modern oncology. Recent data from FGFR3-altered urothelial cancers indicate an inverse relationship between expression FGFR and PD-L1 which also correlates with response to checkpoint inhibitor treatment in these patients [22]. Other studies indicate a potential synergistic effect in combining FGFR inhibitors with checkpoint inhibitors in this disease [23]. So far, only limited data on this potential mechanisms of resistance to checkpoint inhibitors via FGFR signaling are available for iCCA, whereby different interactions of FGFR with the tumor microenvironment are known leading to an immunosuppressive milieu overall (see Figure 1).

Figure 1. Different interactions of FGFR with the tumor microenvironment (cellular/acellular) resulting in an immunosuppressive milieu and promoting tumor progression, overall (original Figure, information taken from own review [4]).

A Chinese study found an enrichment of high PD-L1 expression in FGFR2-rearranged iCCA, but this result could be impacted by an also increased rate of HBV infected patients in this population [24]. In contrast, Sridharan et al. found low PD-L1 expression in FGFR2-fusion positive cholangiocarcinomas [25]. Several clinical trials are currently ongoing to further investigate the effect of combined FGFR and immune checkpoint inhibitor treatment in FGFR-related tumors like iCCA but no results have been reported so far (see Table 1).

Table 1. Clinical trials dealing with combinatory treatment strategies with FGFR-inhibitors (with exception of clinical trial NCT05859477 (indicate by *including such cases with positive FGFR2 overexpression status by immunohistochemistry and/or amplification of FGFR2) and checkpoint inhibitors from a query of ClinicalTrials.Gov (https://clinicaltrials.gov/ with the query terms ‘FGFR’ and ‘checkpoint inhibitor,’ latest access on 2023.06.11). Used abbreviations: CCC = cholangiocellular cancer, GC = gastric cancer, GJC = gastroesophageal junction cancer, HCC = hepatocellular cancer, NSCLC = non-small-cell lung cancer, UC = urothelial cancer.

Rank	NCT number	Starting date	Status	Conditions	FGRF-drug	FGFR-selectivity	Checkpoint-inhibitor	Checkpoint-selectivity	Phases
1	NCT02325739	December, 2014	Completed	Solid tumors (including HCC)	FGF401	4	Spartalizumab	Anti-PD-1	1/2
2	NCT02393248	February, 2015	Terminated	Solid tumors (including GC, UC, CCC), hematological neoplasms	Pemigatinib	1–3	Pembrolizumab	Anti-PD-1	1/2
3	NCT03123055	April, 2017	Terminated	Locally advanced or metastatic UC	Vofatmab (B-701)	3	Pembrolizumab	Anti-PD-1	1/2
4	NCT03390504	March, 2018	Recruiting	UC	Erdafitinib	1–4	Pembrolizumab	Anti-PD-1	3
5	NCT03473756	May, 2018	Active, not recruiting	UC	Rogaratinib	1–4	Atezolizumab	Anti-PD-L1	1
6	NCT04003610	May, 2020	Terminated	Metastatic or unresectable UC	Pemigatinib	1–3	Pembrolizumab	Anti-PD-1	2
7	NCT04045613	July, 2019	Completed	UC	Derazantinib	1–3	Atezolizumab	Anti-PD-L1	1/2
8	NCT04601857	January, 2021	Recruiting	Advanced and metastatic UC	Futibatinib	1–4	Pembrolizumab	Anti-PD-1	2
9	NCT04604132	October, 2020	Terminated	GC	Derazantinib	1–3	Atezolizumab	Anti-PD-L1	1/2
10	NCT04699643	July, 2020	Unknown status	Advanced solid tumors	EVER4010001	4	Pembrolizumab	Anti-PD-1	1/2
11	NCT04828486	May, 2021	Recruiting	Advanced and metastatic HCC	Futibatinib	1–4	Pembrolizumab	Anti-PD-1	2
12	NCT05004974	September, 2021	Not yet recruiting	Advanced NSCLC	Pemigatinib	1–3	Sintilimab	Anti-PD-1	2
13	NCT05111626	March, 2022	Recruiting	GC, GJC	Bemarituzumab	2b	Nivolumab	Anti-PD-1	3
14	NCT05174650	April, 2022	Recruiting	Intrahepatic CCC	Derazantinib	1–3	Atezolizumab	Anti-PD-L1	2
15	NCT05267470	March, 2022	Recruiting	NSCLC	Bemarituzumab	2b	Pembrolizumab	Anti-PD-1	1
16	NCT05322577	May, 2022	Recruiting	GC, GJC	Bemarituzumab	2b	Nivolumab	Anti-PD-1	1
17	NCT05510427	February, 2024	Not yet recruiting	CCC, HCC	Infigratinib	1–4	Atezolizumab	Anti-PD-L1	1
18	NCT05614739	January, 2023	Recruiting	UC	LOXO-435	3	Pembrolizumab	Anti-PD-1	1
19	NCT05775874	September, 2022	Recruiting	UC	AZD4547	1–3	Tislelizumab	Anti-PD-1	2
20	NCT05859477	June, 2022	Recruiting	Metastatic GC	*		Nivolumab	Anti-PD-1	2

3. Conclusion

FGFR inhibitors demonstrated good clinical efficacy in patients with iCCA. While much emphasis was laid on identifying the right patient (FGFR2 mutated/rearranged/amplified, mRNA overexpressing, etc.), the understanding of concomitant primary or acquired secondary resistance mechanisms is still in its beginning. While dissecting the mutational landscape of iCCA further may limit the number of eligible patients, these patients will derive a larger clinical benefit from an even better tailored treatment option by using appropriate combination approaches. The next steps in optimizing therapy of iCCA must be to obtain a deeper and a more thorough understanding of the interplay of signaling pathways like FGFR with modulators of the tumor microenvironment and the immune control of tumors like PD-L1.

4. Expert opinion

The development and approval of several small molecule kinase inhibitors that target FGFR isoforms has significantly improved the treatment options for iCCA. Preclinical experiments and clinical data have shown that patients with alterations in (predominantly) FGFR2 (like gene fusions, gene amplification or activating point mutations) show elevated levels of FGFR2 mRNA expression and have a better response to FGFR inhibitor treatment. This understanding of which patient might derive the biggest benefit from an FGFR inhibitor treatment focused the clinical development on identifying patients with these alterations, e.g. by FISH/CISH techniques for gene copy number variations or more recently on next-generation sequencing (NGS) to detect mutations or gene overexpression.

Initially, only little efforts were made to understand primary and acquired resistance to FGFR inhibitors. With the broader use of NGS panels for patient identification, additional data on alterations of downstream mediators within the FGFR signaling pathway or of other potential oncogenic drivers are now becoming available. While this may lead to further segmentation of patients eligible for FGFR monotherapy, it brings interesting opportunities for rational combination therapies.

More recently, immune checkpoint inhibitors were explored in iCCA and interestingly, FGFR pathway activation may lead to an immune ‘cold’ (T cell excluded) phenotype which may explain the disappointing overall response rate of checkpoint inhibitors in iCCA. Several studies are ongoing to investigate if FGFR inhibition would revert this phenotype and lead to immune ‘hot’ (T cell inflamed) tumors. This complex interplay between growth factor signaling and the tumor microenvironment underlines the need for more research in this area to overcome the still existing hurdles for better tolerated and more effective treatment of iCCA patients. Such studies would then also broaden the eligible patient population again, so that the commercial attractiveness for the pharmaceutical industry to invest in targeted agents and associated diagnostic tests is again given.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.