Futibatinib Is a Novel Irreversible FGFR 1–4 Inhibitor That Shows Selective Antitumor Activity against FGFR-Deregulated Tumors

The FGF/FGFR signaling axis is involved in many cellular processes that include proliferation, differentiation, migration, and survival (1, 2). Deregulated FGFR signaling is associated with developmental disorders and several cancers (2–4). FGFR genomic aberrations, such as gene amplifications, fusions/rearrangements, activating mutations, and altered splicing, have been reported in a range of tumor types, including cholangiocarcinoma, breast, lung, gastric, bladder, hematologic, and other malignancies; these aberrations have been tied to oncogenesis and to tyrosine kinase inhibitor (TKI) resistance in these cancers (5–10). As a result of TKI resistance, cancers characterized by FGF/FGFR aberrations are typically difficult to treat and have a poor prognosis (11, 12). FGFR has emerged as a promising target in these FGFR-deregulated cancers that have a substantial unmet need for novel therapies (1, 13, 14).

Several selective small-molecule FGFR inhibitors entered clinical development in the past few years and have shown responses in patients with FGFR signaling pathway aberrations in phase I/II clinical studies (14–20). However, the majority of these agents bind reversibly to the ATP-binding pocket of FGFRs. As has been observed previously with EGFR TKIs, the occurrence of drug resistance caused by mutations within the drug-binding site is fairly common with reversible ATP-competitive kinase inhibitors (12, 21). Consistent with this, reports of drug-resistant mutations with reversible FGFR inhibitors are emerging (22, 23).

Covalently binding kinase inhibitors inhibit kinase activity irreversibly and may achieve superior potency along with a longer duration of action compared with conventional reversible ATP-competitive inhibitors (21, 24, 25). Despite these potential advantages, the development of covalently binding selective small-molecule pan-FGFR inhibitors has been slow, with PRN1371 among the few currently under clinical investigation (26).

The objective of this study was to describe the preclinical profile of the investigational anticancer agent futibatinib (TAS-120; FBN). These data demonstrate that futibatinib is a structurally novel, irreversible, highly selective FGFR1–4 inhibitor that potently inhibits the growth of FGFR-deregulated cell lines and tumor xenografts. Consequently, these findings support the clinical investigation of futibatinib in FGFR-driven tumors.

Futibatinib (1-[(3S)-[4-amino-3-[(3,5-dimethoxyphenyl)ethynyl]-1H-pyrazolo[3,4-d] pyrimidin-1-yl]-1-pyrrolidinyl]-2-propen-1-one) was synthesized using the procedure described in International Patent Application WO 2013/108809 (27), in which futibatinib is described as example 2 (Fig. 1A). The covalent binding of futibatinib to FGFR was assessed using LC/MS analysis of the kinase domain of recombinant FGFR2, which was incubated in the presence or absence of futibatinib (described in Supplementary Methods).

The cell lines used in this study and their sources are listed in Supplementary Table S1. Cell lines were authenticated by short tandem repeat profiling (Bio-Synthesis Inc.) or were used within 5 passages of original purchased stocks. The cells were confirmed to be free of mycoplasma by PCR (CLEA Japan, Inc.). AZD4547 was synthesized at Taiho Pharmaceutical Co., Ltd., infigratinib and pemigatinib were purchased from MedChemexpress Co., Ltd., and erdafitinib was purchased from Namiki Shoji Co., Ltd.

The kinase-inhibitory properties of futibatinib were investigated in a FGFR1–4 enzymatic assay. As futibatinib binds to the FGFR ATP–binding pocket in the cytoplasmic kinase domain, recombinant FGFR cytoplasmic domains were used in this assay. Phosphorylation of peptide substrate was quantified by the off-chip mobility shift assay (MSA) using the LabChip3000 System (Caliper Life Science). The ATP concentration used in each assay was similar to the Km (Michaelis constant) for the respective kinase. Futibatinib kinase selectivity was assessed against a panel of 296 kinases (Carna Biosciences, Inc.); phosphorylation of each peptide was quantified using the MSA or immobilized metal ion affinity–based fluorescence polarization technologies.

Futibatinib-mediated FGFR kinase inhibition in FGFR-deregulated cell lines was examined using immunoblotting and ELISA with antibodies against phosphorylated FGFR2 (detailed in Supplementary Methods).

Tumor cells were seeded into 96-well plates and treated with futibatinib or vehicle (DMSO) for 72 hours. Tumor cell viability was determined by using the CellTiter-Glo Luminescent Cell Viability Assay Reagent (Promega), and a multimode plate reader was used to detect luminescence. The cell count at time zero (time at which the compound was added) was used to determine the concentration of futibatinib that elicited 50% growth inhibition (GI₅₀).

The antiproliferative activity of futibatinib was further evaluated in murine and rat xenograft models. Animal studies were performed according to prescribed guidelines with the approval of the Institutional Animal Care and Use Committee of Taiho Pharmaceutical Co., Ltd. Tumor xenografts were established in 6- to 12-week-old athymic nude (nu/nu) mice, NOD-SCID mice, and nude rats (CLEA Japan, Inc.) by subcutaneous implantation of tumor cells (1 × 10⁷ each of OCUM-2MD3 gastric cancer, KMS-11 multiple myeloma, or RT-112/84 bladder cancer cells; 3.8 × 10⁶ of MFM-223 breast cancer cells; or 5 × 10⁶ of H1581 lung cancer cells) mixed 1:1 with Matrigel (BD Biosciences) into the side flank. Treatment with futibatinib or vehicle control (0.5 w/v% hydroxypropyl methylcellulose) was initiated when the transplanted tumor reached a predetermined size of more than 0.2 cm³. Futibatinib or vehicle control was administrated orally by gavage; each dose was administered to multiple animals (mice, n = 6; rats, n = 5). Futibatinib was administered for 14 days, except for the KMS-11 multiple myeloma model, in which futibatinib was dosed for 27 days. Futibatinib was dosed once daily in OCUM-2M, KMS-11, and MFM-223 models; administration every other day or twice per week was also tested. Animals were euthanized immediately at the end of treatment. Tumor volume (measured by caliper) and animal body weight were monitored over the course of the treatment. Statistical analysis was performed using one-way ANOVA. Differences in mean tumor volume between each dose group and control group were analyzed using the Dunnett test.

Tumor samples were collected at 4, 8, 12, 18, and 24 hours after a single dose of futibatinib or vehicle. Frozen tumors from xenograft models were lysed in standard cell lysis buffer or reagents according to the manufacturer's instructions (DuoSet IC Kit, R&D Systems Inc.) and phosphorylated FGFR quantified by ELISA or Western blotting (described in Supplementary Methods).

Resistant clones were generated by continuously exposing OCUM-2MD3 cells to futibatinib or AZD4547. The futibatinib dose was increased incrementally to 20 nmol/L over 10 weeks, and the AZD4547 dose was increased to 400 nmol/L over 11 weeks. Resistant cells were maintained at these final doses.

Mutant FGFR2 expression vectors were constructed using site-directed mutagenesis; mutant and wild-type constructs were transfected into HEK293T cells using Lipofectamine 2000 according to the manufacturer's protocol. Transfected cells were treated with futibatinib, stimulated with FGF-7 and heparin, and lysates were analyzed for FGFR2 phosphorylation. Additional details are provided in Supplementary Methods.

Futibatinib potently inhibited the kinase activities of recombinant FGFR1, 2, 3, and 4 in a dose-dependent manner, with IC₅₀ (±SD) values of 1.8 ± 0.4, 1.4 ± 0.3, 1.6 ± 0.1, and 3.7 ± 0.4 nmol/L, respectively (Fig. 1A). The kinase selectivity of futibatinib was assessed against a panel of 296 human kinases using 100 nmol/L futibatinib, which is 50 times the IC₅₀ for FGFR2/3 (Supplementary Table S2). Futibatinib showed strong selectivity for FGFR1–4 in this panel; only 3 non-FGFR kinases showed more than 50% inhibition by futibatinib: mutant RET (S891A; 85.7%), MAPKAPK2 (54.3%), and CK1α (50.7%). These results highlighted the high selectivity of futibatinib inhibition for FGFR1–4, with very low off-target activity against other kinases.

The covalent binding of futibatinib to FGFR was assessed by LC/MS analysis of the recombinant FGFR2 kinase domain incubated in the presence or absence of futibatinib. In the presence of futibatinib, the observed mass of the FGFR2 kinase domain was 418.5 ± 0.2 Da higher than that in the absence of futibatinib (Fig. 1B). This mass difference almost exactly corresponded to the molecular weight of futibatinib (418.45 Da), which indicated covalent binding of futibatinib to the FGFR2 kinase domain.

As futibatinib bound to the FGFR2 kinase domain in vitro, the effect of futibatinib on FGFR phosphorylation and signaling was examined in FGFR-deregulated cell lines. Treatment of the FGFR2-amplified gastric cancer cell lines OCUM-2MD3 and OCUM-2M with increasing doses of futibatinib for 30 minutes resulted in dose-dependent inhibition of FGFR phosphorylation as demonstrated by Western blotting (Fig. 2A) and ELISA (Fig. 2B), with an IC₅₀ value of 4.9 ± 0.1 nmol/L obtained by ELISA. Futibatinib treatment also resulted in similar reductions in Akt and ERK phosphorylation levels in proportion to that of FGFR phosphorylation (Fig 2A). Similar results were observed with SNU-16, another gastric cancer cell line with FGFR2 amplification, and OPM-2 and KMS-11, both multiple myeloma cell lines with FGFR3 translocations (Supplementary Fig. S1A–S1C). These results suggest that futibatinib potently inhibits FGFR phosphorylation and downstream signaling via MAPKs and PI3K/Akt pathways.

The antiproliferative activity of futibatinib was assessed in cancer cell lines from diverse tissue origins (gastric, breast, lung, endometrial, and bladder cancer, as well as multiple myeloma) with or without FGFR genomic aberrations by examining cell growth after 3 days of exposure to futibatinib. Across all tumor types studied, futibatinib inhibited the growth of cell lines with various FGFR genomic aberrations, but not of cell lines that did not harbor such aberrations (Table 1; Fig. 3A). These FGFR aberrations included FGFR1/2 amplifications (breast cancer), FGFR1 amplifications (lung cancer), FGFR2 amplifications (gastric cancer), FGFR2 point mutations (endometrial cancer), FGFR3 fusions (bladder cancer), and FGFR3 translocations (multiple myeloma; Table 1; refs. 5, 6, 28–30). Among cell lines with FGFR genomic aberrations, the potency of futibatinib inhibition ranged between GI₅₀ values of approximately 1 and 50 nmol/L for most cell lines. At least three cell lines with FGFR aberrations did not show sensitivity to futibatinib; these included the FGFR3 fusion–expressing SW780 bladder cancer cell line (GI₅₀, 4,800 nmol/L) and multiple myeloma cell lines H929 (GI₅₀, 1,000 nmol/L) and LP-1 (GI₅₀ > 3,000 nmol/L), both of which harbor FGFR3-t(4;14) translocations. Together, these results indicate that futibatinib demonstrated antiproliferative activity against FGFR-deregulated cancer cell lines of diverse tissue origins. Antiproliferative activity was specific to tumor cell lines harboring FGFR genomic aberrations, and inhibition was seen regardless of FGFR subtype or the nature of the genetic alteration (amplification, fusion, or point mutation).

The antitumor activity of futibatinib was further evaluated in murine or nude rat xenograft models having tumors expressing FGFR genomic aberrations. Mice were administered oral futibatinib once daily for 2 weeks (except for KMS-11, as explained below). In the FGFR2-amplified OCUM-2MD3 gastric cancer model, once-daily administration of futibatinib at 0.15 mg/mL resulted in statistically significant tumor growth inhibition, and dosing at 0.5, 1.5, and 5 mg/kg resulted in dose-dependent tumor reduction (Fig. 3B). There was no early evidence of resistance during the treatment period, even in mice treated with lower futibatinib doses (e.g., 0.5 mg/kg). In the FGFR3-fusion–positive KMS-11 multiple myeloma model, significant tumor regression was observed at 5 mg/kg futibatinib (Fig. 3C). Because tumor regression in this model was observed beginning on day 12 with definite tumor regression on day 15, dosing was continued until day 27 in this model. In the FGFR1/2-amplified MFM-223 breast cancer xenograft model, robust growth inhibition was noted with daily futibatinib doses ranging from 12.5 to 50 mg/kg (Fig. 3D).

Preliminary experiments were also performed to examine intermittent futibatinib dosing schedules (every other day or twice weekly) in FGFR-deregulated xenograft mouse (OCUM-2MD3) or nude-rat (RT-112 and H1581) models. Tumor reduction was observed with intermittent dosing in the OCUM-2MD3 gastric cancer and RT-112 bladder carcinoma models (Supplementary Fig. S2A–S2C), but not in the H1581 non–small cell lung cancer (NSCLC) model (Supplementary Fig. S2D). Tumor growth curves for individual mice or rats are shown in Supplementary Fig. S3A–S3C (once-daily dosing) and Supplementary Fig. S4A–S4D (intermittent dosing). Collectively, these results highlight potent and broad-spectrum antitumor efficacy of futibatinib against mouse or rat xenograft models harboring various FGFR aberrations.

All once-daily dosing regimens of futibatinib were well tolerated in mice and did not induce body weight reduction during the experiments. In our experiments in the RT-112 and H1581 nude-rat xenograft models, hyperphosphatemia was observed in rats after futibatinib dosing, an effect that has been previously reported with other FGFR inhibitors and was expected given the mechanism of action of futibatinib (31). The hyperphosphatemia observed was transient, and phosphate levels returned to normal 48 hours after futibatinib dosing.

Because futibatinib binds covalently (irreversibly) to a unique cysteine residue in the ATP-binding pocket of FGFRs (32), the inhibitory effect of futibatinib on FGFR kinase activity was expected to persist after even short drug exposures. The association between futibatinib FGFR kinase inhibition and antitumor activity was assessed both in cell lines and in xenograft models.

To assess the duration of FGFR kinase inhibition, OCUM-2M and SNU-16 cells were treated with futibatinib for 1 hour, followed by drug washout and quantification of FGFR phosphorylation. Without washout, phosphorylation of FGFR2 in both OCUM-2M and SNU-16 cell lines remained inhibited with 10 nmol/L futibatinib; however, with washout of futibatinib, FGFR phosphorylation levels reversed to that of control within 6 hours (Fig. 2B; Supplementary Fig. S1A). Similar results were obtained in the OPM-2 and KMS-11 cell lines (Supplementary Fig. S1B–S1C). In these cell lines, similar recovery of FGFR phosphorylation levels was observed after washout experiments with AZD4547 (Supplementary Fig. S5A–S5B). Experiments with brefeldin A (which suppresses new protein synthesis) indicated that only 20% of phosphorylated FGFR protein remained after 4 to 8 hours of treatment (detailed in Supplementary Fig. S5C–S5E), which suggested a high turnover of FGFR in the cell lines studied. This high turnover explains why a short exposure of futibatinib did not cause sustained inhibition of FGFR in these cell lines.

To determine the optimal exposure time of futibatinib for antitumor activity in FGFR-deregulated cell lines, the FGFR2-amplified OCUM-2M gastric cancer cell line was treated with futibatinib for 4, 8, 24, 48, or 72 hours, followed by drug washout. Cell viability was then monitored at 72 hours. Maximal growth inhibition was observed with 24 hours of treatment proportional to the dose administered, and no additional antiproliferative effect was observed when treatment was extended beyond 24 hours (Fig. 4A). These results indicated that continuous futibatinib exposure up to 24 hours was essential for maximal efficacy in futibatinib-sensitive FGFR-deregulated cells. This was consistent with the high turnover of FGFR in these cells as observed above. However, futibatinib treatment for longer than 24 hours did not result in additional antiproliferative activity.

The correlation between tumor regression and FGFR kinase inhibition with futibatinib treatment was investigated in xenograft models. Futibatinib was orally administered to OCUM-2MD3-tumor–bearing nude mice at doses of 0.5, 1.5, and 5 mg/kg; xenografts were harvested 4 hours after a single dose of futibatinib, and phosphorylation of FGFR2, ERK, and Akt was measured by Western blotting. As shown in Fig. 4B, futibatinib exhibited dose-dependent inhibition of FGFR2 phosphorylation in the OCUM-2MD3 tumors. In addition, dose-dependent inhibition of ERK phosphorylation (which is downstream of FGFR) was also observed, although Akt phosphorylation was only marginally inhibited. This indicated that FGFR phosphorylation and downstream signaling were inhibited in xenograft models.

To understand the relationship between FGFR kinase inhibition and antitumor activity in the preclinical models, the duration of FGFR inhibition was determined at once-daily doses of 1.5, 5, and 15 mg/kg. The serum half-life of futibatinib was between 0.25 and 1 hour in these models (Supplementary Fig. S6). FGFR2 phosphorylation was potently inhibited (≥80%) at 4 hours after single doses of either 1.5 or 5 mg/kg of futibatinib (Fig. 4C). The inhibition of FGFR2 phosphorylation was partially sustained at 8 hours at these doses (∼50%), with a return to control levels at 12 hours. In the 15-mg/kg dosing group, approximately 90% inhibition of FGFR phosphorylation was observed 8 hours after initial drug administration, and at 12 hours, FGFR inhibition remained at approximately 50%. These data showed sustained futibatinib-mediated inhibition of FGFR phosphorylation in the xenografted tumors and confirmed that the in vivo antitumor activity of futibatinib was associated with dose- and time-proportional pharmacodynamic modulation of FGFR2 phosphorylation, even though the futibatinib serum half-life was found to be short.

To assess the risk of drug resistance with futibatinib treatment, FGFR2-amplified OCUM-2MD3 gastric cancer cells were propagated in increasing concentrations of futibatinib or the reversible ATP-competitive FGFR inhibitor AZD4547 for comparison, until maximum concentrations for viability were reached (20 nmol/L for futibatinib and 400 nmol/L for AZD4547). A total of 12 resistant clones survived at 20 nmol/L futibatinib (a 13-fold higher concentration than the IC₅₀ value for wild-type FGFR2) compared with 170 resistant clones that were obtained through exposure to 400 nmol/L AZD4547 (45-fold higher than the IC₅₀ for wild-type FGFR2; Fig. 5A; Supplementary Table S3). This suggested that tumors with FGFR2 genomic aberrations exposed to futibatinib were less prone to the development of resistant clones than those exposed to a reversible ATP-competitive inhibitor such as AZD4547.

To characterize the mutations in the resistant clones, cDNA encoding the kinase domain of FGFR2 was sequenced. In 61.2% (104/170) of clones resistant to AZD4547, the K660N mutation was identified in the kinase domain (Fig. 5B), but this mutation was not present in futibatinib-resistant clones. Futibatinib potently inhibited FGFR phosphorylation and growth of AZD4547-resistant OCUM-2MD3 clones, with IC₅₀ values of 3.1 and 4.8 nmol/L, respectively (Supplementary Tables S4 and S5). On the other hand, the mean AZD4547 IC₅₀ for growth increased from 6.1 nmol/L in the parental cells to 158.6 nmol/L in resistant cells. Infigratinib (another reversible ATP-competitive FGFR inhibitor) also showed cross-resistance toward AZD4547-resistant clones, with the IC₅₀ for growth increasing from 7.6 nmol/L in parental cells to 55.6 nmol/L in resistant cells (Supplementary Table S5). These results indicated potent activity of futibatinib against tumor cells resistant to treatment with reversible ATP-competitive FGFR inhibitors.

We then determined whether futibatinib was effective against other mutations resistant to reversible ATP-competitive FGFR inhibitor treatment. The N550H and E566G mutations in the FGFR2 hinge region have been reported to cause resistance to dovitinib (33). K660M within the FGFR2 activation loop is an activating mutation reported in breast, endometrial, and cervical cancer (29), and V565I is a gatekeeper mutation, frequently seen in drug-resistant tumors (33). We generated 4 mutant FGFR2 constructs with these individual mutations and transfected them into HEK293T cells along with a wild-type FGFR2 construct as a control. The activity of the constructs was confirmed by detection of autophosphorylation activity, and the effect of futibatinib and reversible ATP-competitive inhibitors AZD4547 and infigratinib on FGFR2 autophosphorylation was determined using phosphorylated FGFR2 ELISA. The activity of AZD4547 and infigratinib is shown in Fig. 6A: robust inhibition of wild-type FGFR was observed as expected. Futibatinib showed potent inhibitory activity against all mutants, with IC₅₀ values ranging from 1.3 nmol/L (against V565I) to 5.2 nmol/L (against K660M), which was similar to its inhibitory potency against wild-type FGFR2 (0.9 nmol/L). In contrast, the activity of both reversible ATP-competitive FGFR inhibitors was significantly lower with the 4 mutants than with wild-type.

Very recently, 2 FGFR inhibitors, erdafitinib and pemigatinib, were approved for the treatment of patients with advanced/metastatic FGFR-aberrant urothelial cancer or cholangiocarcinoma, respectively (34, 35). The activity of futibatinib and these agents was evaluated with 3 representative FGFR2 mutants: N550K (hinge region), K660M (activation loop), and V565L (gate-keeper; Fig. 6B). All 3 inhibitors showed similar inhibitory activity against wild-type FGFR2. Futibatinib demonstrated the strongest inhibition of N550K and V565L, with IC₅₀ values only 3- and 14-fold higher than that with wild-type, respectively. In contrast, erdafitinib and pemigatinib demonstrated limited activity against these 2 mutations: relative to wild-type, IC₅₀ values for erdafitinib were 10- and 34-fold higher and those for pemigatinib were 83- and 236-fold higher with the N550K and V565L mutants, respectively.

Together, these results suggested that the risk of drug resistance with futibatinib was low and demonstrated potent inhibitory activity of futibatinib against FGFR mutants that are resistant to reversible ATP-competitive inhibitors.

In this article, we report the identification of futibatinib, a potent, selective, and covalent (irreversible) small-molecule inhibitor of FGFR1–4, and its first pharmacologic characterization as an orally bioavailable antitumor agent. Futibatinib demonstrated remarkable selectivity for FGFR among 296 kinases and was found to bind covalently to the FGFR kinase domain. The binding mode of futibatinib to FGFR was shown to be quite different from known reversible ATP-competitive FGFR inhibitors, such as AZD4547, infigratinib, erdafitinib, or pemigatinib, or irreversible FGFR inhibitors such as PRN1371 (26, 32, 36, 37). Recent structural analysis of the futibatinib–FGFR complex revealed that futibatinib targets the P-loop in the ATP-binding pocket of the FGFR tyrosine kinase domain, forming a rapid covalent adduct with a unique cysteine upon contact (32). Futibatinib is distinct in its ability to capture a number of conformations of the highly flexible FGFR P-loop. Futibatinib showed potent inhibition of all 4 FGFR isoforms (FGFR1–4) at nearly equivalent single-digit nanomolar IC₅₀ values, in contrast to both AZD4547 and infigratinib, which showed strong inhibition of FGFR1–3, but only marginal inhibition of FGFR4 in earlier studies (36, 38, 39). Similarly, PRN1371 has been reported to exhibit a 50-fold lower affinity for FGFR4 than for FGFR1–3 (26).

Futibatinib selectively suppressed cell growth of cancer cell lines harboring FGFR genomic aberrations, regardless of tumor type or type of FGFR aberration, which included activating mutations, gene amplifications, or translocations in FGFR1–3. The extent of growth inhibition was similar in cell lines with FGFR1, FGFR2, or FGFR3 genomic aberrations (GI₅₀ ranging from 1 to ∼50 nmol/L), which is consistent with the finding that futibatinib showed inhibition of all FGFR isoforms at nearly equivalent potency in biochemical experiments. Cell lines without FGFR genomic aberrations were insensitive to futibatinib. Although certain FGFR-deregulated cell lines [H929 and LP-1 (multiple myeloma) and SW780 (bladder cancer)] were also insensitive to futibatinib, further characterization of these cell lines provided possible explanations. H929 cells were found to have an activating NRAS mutation, the FGFR3 mutation in LP-1 cells was determined not to be an activating mutation (40), and the relative expression level of FGFR3 was low in the SW780 bladder cancer cell line compared with the other FGFR3-deregulated bladder cancer cell lines, RT4 and RT112/84 (30). Overall, the antiproliferative effect of futibatinib in cancer cells was consistent with its highly selective biochemical profile.

Futibatinib also demonstrated potent antitumor activity in multiple xenograft models when administered orally on a once-daily schedule. Dose-dependent tumor regression was observed across multiple tumor types, consistent with observations in the cancer cell lines.

Because futibatinib is a covalent inhibitor, it was expected that its inhibitory activity would persist even after short periods of exposure in the cells, as has been observed with other covalent inhibitors (26, 41). However, as FGFR turnover was determined to be rapid (half-life ∼4 hours) in the cell lines used in our study, the duration of futibatinib inhibition could not be accurately estimated. Nevertheless, evaluation of the duration of futibatinib exposure necessary for antiproliferative effects indicated that 24 hours was sufficient to achieve maximal growth inhibition in cancer cell lines. The sustained inhibitory effect of futibatinib was confirmed in xenograft models, even though the serum half-life of futibatinib was found to be short. Antitumor efficacy was noted beginning at 5 mg/kg in the OCUM-2MD3 model, and this correlated with inhibition of FGFR2 phosphorylation, which lasted for up to 8 hours after the first dose. At doses of 15 mg/kg, significant FGFR2 inhibition (up to 50%) was observed for up to 12 hours after initial dose administration. The short serum half-life of futibatinib could provide a possible explanation for the loss of FGFR2 inhibition beyond 18 hours of dosing.

Given the sustained inhibition of FGFR kinase activity with futibatinib, intermittent futibatinib dosing schedules (every other day or twice weekly) were examined in FGFR-deregulated xenograft models. Although robust antitumor efficacy was observed with intermittent dosing in some tumor types [gastric cancer (OCUM-2MD3) and bladder carcinoma (RT-112)], the antitumor effect was not optimal in other tumor types [NSCLC (H1581)]. On this basis, a continuous dosing regimen was chosen to maximize the efficacy of futibatinib in clinical studies. In initial results of a phase I dose-escalation study with futibatinib in patients with advanced solid tumors, responses were achieved in patients who received once-daily futibatinib (42–44). In these patients, maximum plasma concentrations of futibatinib were achieved 2.2 hours after dosing, and the plasma half-life was determined to be 3.3 hours (42). Efficacy has also been observed with once-daily futibatinib in patients with advanced/metastatic intrahepatic cholangiocarcinoma in an ongoing phase II study (objective response rate, 34.3%; ref. 45).

The appearance of resistance mutations is a major concern in the use of molecularly targeted kinase inhibitors. Consistent with this, secondary FGFR kinase domain mutations were reported to be an important mechanism of clinically acquired resistance to the reversible ATP-competitive FGFR inhibitor infigratinib (23). Although encouraging antitumor activity was observed in infigratinib-treated patients with FGFR2 fusion-positive cholangiocarcinoma in a phase II trial, acquired resistance inevitably developed, limiting the durability of response (15). Genomic analysis revealed a number of point mutations in the FGFR2 kinase domain (N549H, N549K, E565A, K641R, or K659M), as well as the gatekeeper mutations V565F and V565I (23). Similarly, AZD4547 has been shown to be susceptible to acquired resistance (46). Data on acquired resistance with the newer inhibitors erdafitinib and pemigatinib (which have shown efficacy in patients with advanced cancers; refs. 17, 20) have yet to be reported. The results reported here, however, revealed reduced in vitro activity of these two inhibitors against FGFR2 mutants associated with drug resistance.

Our experiments in FGFR-deregulated gastric cancer cell lines indicated that in contrast to the reversible ATP-competitive inhibitor AZD4547, very few resistant clones appeared with prolonged futibatinib treatment, and no mutations were observed in the FGFR2 kinase domain in futibatinib-resistant clones. Futibatinib also demonstrated robust activity against FGFR mutations known to be resistant to reversible ATP-competitive FGFR inhibitors. Specifically, potent inhibition of the V565I and V565L gatekeeper mutants of FGFR2 was observed, in addition to other drug-resistant mutants; IC₅₀ values with the mutants assayed in this study were generally comparable with that obtained with wild-type FGFR2.

There are at least two possible explanations for the observed activity of futibatinib against resistant FGFR mutations. The first is related to the covalent futibatinib–FGFR binding interaction: the rapid covalent bond formation between the cysteine in the ATP-binding pocket and futibatinib is less likely to be affected by mutations in the ATP-binding pocket that reduce the binding affinity of reversible ATP-competitive FGFR inhibitors. As covalent complexes are generally stable, this would allow effective target inhibition. Other irreversible FGFR inhibitors such as FINN-2, FINN-3, and PRN1371 have similarly shown activity against drug-resistant FGFR2 mutations, including gatekeeper mutations (26, 47). This provides support for the use of irreversible FGFR inhibitors in overcoming the resistance associated with reversible ATP-competitive FGFR inhibitors. Secondly, the amino acid residues within the FGFR ATP–binding pocket involved in binding interactions with futibatinib are different than those binding reversible ATP-competitive FGFR inhibitors. The binding of these inhibitors occurs primarily within the hinge region of the FGFR ATP–binding pocket (where drug-resistant mutations often occur), whereas futibatinib interacts with a reactive cysteine in the P-loop (32, 36). These results support the use of futibatinib in patients with resistance to prior TKI regimens. Indeed, in a preliminary report, clinical responses were observed with futibatinib in patients with cholangiocarcinoma that was resistant to reversible ATP-competitive inhibitor treatment (48). Future research with all of these agents is warranted to gain insight into resistance mechanisms with FGFR inhibitors and to clarify the role of futibatinib within the changing FGFR inhibitor landscape.

There are several irreversible FGFR inhibitors currently under investigation. PRN1371, which demonstrated robust FGFR1–3 inhibitory activity in preclinical experiments, has also been reported to potently inhibit several FGFR2 and FGFR3 drug-resistant mutants, but was shown to have reduced activity against the common drug-resistant gatekeeper mutant (V561M) of FGFR1 (26). A phase I trial (NCT02608125) of this drug is ongoing in patients with advanced tumors. FINN-2 and FINN-3 are irreversible FGFR inhibitors that have demonstrated in vitro inhibition of FGFR1 and FGFR2 gatekeeper mutants but have not yet been tested in the clinic (47). Other irreversible inhibitors in development include BLU9931 and fisogatinib, which target Cys522 in the hinge region of FGFR4 and selectively inhibit FGFR4, but have weak inhibitory activity against FGFR1–3. These compounds are only active in cell lines with FGFR4-pathway deregulation (49). Fisogatinib has shown antitumor activity in patients with advanced hepatocellular carcinoma with aberrant FGFR4-pathway signaling in a phase I study (NCT02508467; ref. 50). Among all irreversible inhibitors, futibatinib remains the most advanced in clinical development, with preliminary results of phase I and II trials reported and a phase III trial ongoing.

In summary, the results of this study demonstrate that futibatinib is a potent, irreversible, highly selective inhibitor of FGFR1–4 that exhibits broad antiproliferative activity in FGFR-deregulated cancer lines and animal tumor models. Sustained FGFR kinase inhibition was noted in preclinical models. Futibatinib was also associated with a low risk of drug resistance and exhibited activity against drug-resistant FGFR mutants. The preclinical and pharmacologic profiles of futibatinib provide strong support for clinical testing of futibatinib in patients with advanced FGFR-driven tumors and form the basis for ongoing phase I/II (NCT02052778), phase II (NCT04024436; NCT04189445), and phase III (NCT04093362) trials being conducted in patients with advanced tumors harboring FGFR aberrations.

K. Yonekura reports personal fees from TAIHO Pharmaceutical Co., Ltd., during the conduct of the study and personal fees from TAIHO Pharmaceutical Co., Ltd., outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

H. Sootome: Conceptualization, resources, formal analysis, supervision, investigation, visualization, methodology, writing-original draft, writing-review and editing. H. Fujita: Resources, formal analysis, investigation, writing-review and editing. Ke. Ito: Resources, investigation, writing-review and editing. H. Ochiiwa: Resources, investigation, writing-review and editing. Y. Fujioka: Resources, investigation, writing-review and editing. Ki. Ito: Resources, investigation, writing-review and editing. A. Miura: Resources, investigation, visualization, writing-original draft, writing-review and editing. T. Sagara: Conceptualization, formal analysis, visualization, writing-original draft, writing-review and editing. S. Ito: Conceptualization, formal analysis, writing-review and editing. H. Ohsawa: Resources, investigation, methodology, writing-review and editing. S. Otsuki: Conceptualization, formal analysis, visualization, writing-original draft, writing-review and editing. K. Funabashi: Resources, investigation, writing-review and editing. M. Yashiro: Resources, investigation, visualization, writing-original draft, writing-review and editing. K. Matsuo: Resources, investigation, writing-review and editing. K. Yonekura: Supervision, writing-review and editing. H. Hirai: Conceptualization, supervision, visualization, writing-original draft, writing-review and editing.

The authors thank TAS-120 project members in Tsukuba and Tokushima Research Institute and Taiho Pharmaceutical Co., Ltd., for the support they provided for this work.

This study was sponsored by Taiho Oncology, Inc., and Taiho Pharmaceutical Co., Ltd. Medical writing and editorial assistance were provided by Vasupradha Vethantham, PhD, and Anne Cooper of Ashfield Healthcare Communications.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.