Targeting Wild-Type and Mutationally Activated FGFR4 in Rhabdomyosarcoma with the Inhibitor Ponatinib (AP24534)

Rhabdomyosarcoma (RMS) is the most common childhood soft tissue sarcoma. Despite advances in modern therapy, patients with relapsed or metastatic disease have a very poor clinical prognosis. Fibroblast Growth Factor Receptor 4 (FGFR4) is a cell surface tyrosine kinase receptor that is involved in normal myogenesis and muscle regeneration, but not commonly expressed in differentiated muscle tissues. Amplification and mutational activation of FGFR4 has been reported in RMS and promotes tumor progression. Therefore, FGFR4 is a tractable therapeutic target for patients with RMS. In this study, we used a chimeric Ba/F3 TEL-FGFR4 construct to test five tyrosine kinase inhibitors reported to specifically inhibit FGFRs in the nanomolar range. We found ponatinib (AP24534) to be the most potent FGFR4 inhibitor with an IC50 in the nanomolar range. Ponatinib inhibited the growth of RMS cells expressing wild-type or mutated FGFR4 through increased apoptosis. Phosphorylation of wild-type and mutated FGFR4 as well as its downstream target STAT3 was also suppressed by ponatinib. Finally, ponatinib treatment inhibited tumor growth in a RMS mouse model expressing mutated FGFR4. Therefore, our data suggests that ponatinib is a potentially effective therapeutic agent for RMS tumors that are driven by a dysregulated FGFR4 signaling pathway.


Introduction
Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma in childhood, accounting for about 3% of all childhood tumors [1]. Treatment of RMS includes the use of intensive chemotherapeutic regimens in combination with surgical and radiation therapy. This strategy has improved the survival rate for patients with localized disease to 70% albeit with significant toxicity [2]. Despite aggressive multimodal therapy, high risk patients continue to have a poor prognosis with overall survival rates of 20-30% [3]. Therefore, there remains a great need for new therapies targeting the molecular pathways which are found to be altered in RMS.
RMS tumors typically arise from skeletal muscle and are categorized as either of the alveolar (ARMS) or embryonal (ERMS) subtype based on their histology. ARMS tumors are driven by a translocation involving chromosome 2 or 1 with chromosome 13, resulting in the production of the fusion oncogene PAX3or PAX7-FOXO1, respectively [4]. In contrast, ERMS tumors commonly harbor loss of heterozygosity at 11p15.5 [5] as well as point mutations in TP53 [6], NRAS, KRAS, HRAS [7], PIK3CA [8] and FGFR4 [9] genes.
Fibroblast Growth Factor Receptor 4 (FGFR4), a FGF receptor family member, is a receptor tyrosine kinase that is implicated in the differentiation of myoblasts into skeletal muscle [10] and muscle regeneration after injury [11]. Highlighting a potential role in RMS, early microarray studies of RMS cell lines and tumors showed massive overexpression of FGFR4 [12] and subsequent work showed that FGFR4 is a direct transcriptional target of the PAX3-FOXO1 fusion protein [13]. Of note, recent sequencing studies identified activating mutations specific to FGFR4 in 7.5% of RMS tumors. These mutations occur at amino acid 535 and 550 of the kinase domain and promote tumor growth and metastasis in vivo by constitutively activating FGFR4 [9]. These reports emphasize the importance of FGFR4 in RMS and establish this cell surface tyrosine kinase receptor as a candidate target for RMS therapy.
Ponatinib is an orally administered tyrosine kinase inhibitor that was initially developed as an inhibitor for native and mutant forms of BCR-ABL [14]. Recently, this therapy received accelerated FDA approval for the treatment of adult patients with Philadelphia chromosome positive acute lymphoblastic leukemia (Ph+ ALL) and chronic phase, accelerated phase, or blast phase chronic myeloid leukemia (CML) who are resistant or intolerant to prior tyrosine kinase inhibitor therapy. The inhibition profile of ponatinib includes several other tyrosine kinases, including FLT3, SRC, KIT, PDGFR, and FGFR [14,15]. Of note, ponatinib has been shown to inhibit all four members of the FGFR family with an IC 50 of less than 40 nM [16]. Inhibition of FGFR family members by ponatinib has been demonstrated in preclinical models of endometrial cancers with FGFR2 mutations, bladder cancers with FGFR3 mutations, as well as breast, lung, and colon cancer cell lines harboring amplification of the FGFR1 or FGFR2 gene [16]. In this study, a panel of RMS cell lines as well as a Ba/F3 cell line engineered to overexpress FGFR4 were tested for sensitivity to five FGFR tyrosine kinase inhibitors, including AP24534 (ponatinib), AZD2171 (Cediranib), BIBF1120 (Nintedanib), TKI258 (Dovitinib), and PHA739358 (Danusertib). Of these, ponatinib was found to be the most potent FGFR4 inhibitor, inhibiting both wild-type and mutated FGFR4 phosphorylation and cell growth. Ponatinib also inhibited growth of tumors expressing mutated FGFR4 in vivo. Therefore, our results indicate that ponatinib is an effective FDA-approved drug which has the potential to treat RMS with overexpressed or mutated FGFR4.

Ethics Statement
This study is compliant with the animal care and experimental procedures which were approved by the National Institutes of Health Animal Care and Use Committee (Proposal Number: PB-038).

Drug-dose Response Assay and IC 50 Calculations
We measured relative cell number with the CellTiter-Glo assay (Promega) or relative percent confluency with the IncuCyte (Essen Bioscience). In brief, cells were seeded overnight in 180 mL of culture medium per well in 96-well plates so that they would reach 80% confluency by the end of the assay. After overnight incubation, 20 mL of culture medium containing the various inhibitors was added. Relative cell number was measured at 24 hour intervals after the addition of drug. Prism (GraphPad Software) was used for curve fitting and calculating IC50s.

Quantitative RT-PCR
RNA was extracted using the AllPrep DNA/RNA Mini Kit (Qiagen). RNA integrity number (RIN) was calculated using the RNA 6000 Nano Kit (Agilent) and all were greater than 9.0. Quantitative RT-PCR using Taqman assays (FGFR4: Hs00242558_m1 and GAPDH: Hs99999905_m1) on a Fluidigm system was previously described [21]. Briefly, cDNA was generated from 200 ng of RNA using reverse transcription. Then, PCR was carried out on a 48.48 Dynamic Array using the BioMark HD real-time PCR system (Fluidigm). Twelve replications were performed for each gene and sample, and average threshold cycle numbers were calculated. FGFR4 gene expression levels were represented by normalizing against GAPDH.

Cell Cycle Assay
The FITC BrdU Flow Kit (BD Biosciences) was used for cell cycle analysis. In a T25 flask, 1 million cells were seeded and incubated overnight. Cells were treated with ponatinib at a final concentration of 0, 0.625, 1.25, or 2.5 mM for 24 hours. Then the cells were pulsed with 1 mM BrdU for 30 minutes and stained with an anti-BrdU antibody, followed by 7-AAD staining per the manufacturer's guidelines. FACS data was analyzed using CellQuest software (BD Biosciences).

Caspase-3/7 Assay
In an opaque, flat-bottom 96-well plate, 5,000 cells were seeded in each well with 80 mL of culture medium. After overnight incubation, 20 mL of culture medium containing ponatinib was added to reach final concentrations of 0, 1.25, 2.5, and 5 mM. After 6 hours, ApoLive-Glo (Promega) was used to measure caspase-3/7 activity per the manufacturer's protocol.

In vivo Tumor Growth Assay
Animal studies were conducted with 6-to 8-week-old nude female, athymic NCr-nu/nu mice (Animal Production Program, SAIC-Frederick, MD). RMS772 transductants were used to assess in vivo tumor growth. Approximately 1 million cells were injected subcutaneously into the right flank of each mouse. Mice were monitored every other day. Tumor volume measurements were also performed every other day by caliper and the following formula was used to calculate tumor size: (long axis x short axis 2 )/ 2. Daily oral administration by gavage feeding of ponatinib at 30 mg/kg started when the tumor volume exceeded 100 mm 3 . Mice were euthanized when tumors reached 1,500 mm 3 .

Oligonucleotides and Plasmids
pDonr253 is a Gateway Donor vector modified from pDonr201 (Life Technologies). pDonr253 replaces the kanamycin resistance gene with a gene encoding spectinomycin resistance, and contains several sequencing primer sites to aid in sequence verification of Entry clones.

Cloning of TEL-FGFR4
The TEL-FGFR4 gene was constructed by fusing the kinase domain of FGFR4 in frame with the extracellular domain of TEL ( Figure S1A).The TEL-FGFR4 chimera was cloned using overlap extension PCR from cDNA constructs for FGFR4 (Accession # BC011847) and ETV6/TEL (human ORFeome clone). Initial PCRs (left and right) were carried out using Phusion DNA polymerase (New England Biolabs) under standard conditions using a 30 second (TEL) or 60 second (FGFR4) extension time and 200 nM of flanking primers for 20 cycles. PCR products from these reactions were cleaned using the QiaQuick PCR purification kit (Qiagen), and equal amounts of each product pair were combined in a second 20 cycle PCR reaction using the flanking primers. These primers contain Gateway recombination signal sequences, attB1 at the 59 end and attB2 at the 39 end. The final PCR products were cleaned using the QiaQuick PCR purification kit (Qiagen), and recombined into pDonr253 using the Gateway BP recombination reaction (Life Technologies) per the manufacturer's protocols. BP reactions were transformed into E. coli DH10B cells, and colonies were isolated on LB plates containing 50 mg/mL spectinomycin. Plasmid DNA was prepared and sequenced using a variety of internal and external sequencing primers to verify the sequence.

Subcloning of TEL-FGFR4 into a Retroviral Expression Vector
pMSCV-hyg (Clontech) was digested with HpaI and a Gateway reading frame cassette (Life Technologies) was introduced. Proper clones were selected for using ampicillin and chloramphenicol and sequence validated for directionality of the insert and proper sequence at the junctions. The modified Gateway Destination vector was called pDest-450 and was used for subcloning of the TEL-FGFR4 Entry clone via Gateway LR recombination (Life Technologies). Final clones were then transformed into E. coli STBL3 cells (Life Technologies) and final expression constructs were validated by agarose gel electrophoresis and restriction mapping.

Retroviral Transfection of Ba/F3 Cells
Retrovirus containing the TEL-FGFR4 fusion gene was transfected into the PT67 packaging cells (Invitrogen) using Fugene 6 Transfection Reagent (Roche). Two days later, the cell culture supernatant containing the virus was collected, centrifuged, and filtered. Polybrene (Sigma-Aldrich) was added to the viral supernatant at a final concentration of 5 mg/mL. The resulting supernatant was then used to infect the Ba/F3 cells for eight hours.
The stable, transfected cells were established by selecting cells in full media containing 1 mg/ml of hygromycin.
RT-PCR of the PAX3/7-FOXO1 Fusion Gene PAX3/7-FOXO1 fusion gene status was determined by RT-PCR. The same cDNA used in the quantitative RT-PCR was also used for this PCR (35 cycles of 95uC for 30 seconds, 60uC for 30 seconds, and 72uC for 45 seconds). We used the forward primer CCGACAGCAGCTCTGCCTAC and reverse primer AT-GAACTTGCTGTGTAGGGACAG to amplify the fusion gene. PCR product was analyzed using the DNA 1000 kit (Agilent) to check for a 172 bp band.

DNA Sequencing
We used the following PCR Primers for FGFR4 protein-coding exons 12 and 13-exon 12 forward: GATTCAGCCCTAGACC-TACG; exon 12 reverse: CACTCCACGATCACGTAC; exon 13 forward: CAACCTGCTTGGTGTCTG; and exon 13 reverse: GGAAAGCGTGAATGCCTG. Cell line DNA was amplified by PCR and the product was confirmed by gel electrophoresis. PCR DNA was purified using the AMPure XP kit (Agencourt) and then sequenced by Sanger sequencing. Sequencing results were analyzed with Sequencher (Gene Codes).

RMS cells with overexpressed wild-type FGFR4 are more sensitive to ponatinib
Using our Ba/F3 model system ( Figure S1A-C), we confirmed previous reports that ponatinib is the most sensitive FGFR4 inhibitor with an IC 50 of 72.2 nM ( Figure S2A) and that it inhibits wild-type FGFR4 phosphorylation ( Figure S2B) [16]. To demonstrate ponatinib's efficacy in RMS, we tested the molecule on a panel of RMS cell lines. We used six fusion-positive and eight fusion-negative RMS cell lines (fusion status verified by RT-PCR; Figure S3A). All tested cell lines showed nanomolar sensitivity to the drug and the sensitivity was dependent on the level of FGFR4 mRNA expression ( Figure 1A, p = 0.0261, Spearman correlation). Of note, fusion-positive cell lines had significantly lower IC 50 s compared to fusion-negative cell lines, which have a wide range of sensitivity to ponatinib ( Figure 1B, p = 0.0125, F-test). This is consistent with the fact that fusion-positive RMS cell lines typically express higher levels of FGFR4 ( Figure S3B, p = 0.0005), because it is directly induced by the PAX3-FOXO1 fusion gene [13]. Indeed, cell lines expressing the highest levels of FGFR4 (above a relative level of six) were the most sensitive to ponatinib ( Figure 1C, p = 0.0344). These results, therefore, suggest that increased expression of FGFR4 confers sensitivity to ponatinib.

RMS cells expressing activating FGFR4 mutations have increased sensitivity to ponatinib
The effect of ponatinib on RMS cells with mutationally activated FGFR4 was also tested in RMS772 cells engineered to express empty vector, wild-type FGFR4, FGFR4 N535K, and FGFR4 V550E [9,17]. We found that the cells with the FGFR4 N535K and V550E mutation showed a significantly lower IC 50 (215 and 204 nM, respectively) compared to wild-type FGFR4 (960 nM; Figure 2, * , **p,0.0001), suggesting that the activating FGFR4 mutations make cells more sensitive to ponatinib than wild-type FGFR4. Furthermore, we found in a kinetic study that the effect of ponatinib on cell confluency was rapid, within 6 hours, in RMS772 cell lines harboring the N535K or V550E mutations ( Figure S4A). However, no significant difference in IC 50 was seen between wild-type FGFR4 and the empty vector ( Figure  S4B).

Ponatinib treatment results in decrease in S-phase fraction of cell cycle and augmented apoptosis in RMS cells
To investigate the effects of ponatinib on cell cycling, we tested two RMS cell lines, RH4 and RH5, which displayed the highest sensitivity to ponatinib and the two RMS772 cell lines which express mutationally activated FGFR4s. After three hour exposure to ponatinib at final concentrations of 0, 0.625, 1.25, and 2.5 mM, BrdU incorporation showed a decrease in S phase and an increase in sub G 1 across all four cell lines, indicating reduced cell growth and increased apoptosis, respectively ( Figure 3A, Figures S5 and  S6). Furthermore, a caspase 3/7 assay demonstrated that similar concentrations of ponatinib induced apoptosis in all four cell lines ( Figure 3B, *p = 0.0029, **p = 0.0027, ***p = 0.0017, ****p = 0.0001; Figure S7).

Ponatinib inhibits the phosphorylation of wild-type and mutationally activated FGFR4 and its downstream target STAT3 in a dose-dependent manner
To determine whether ponatinib inhibits the phosphorylation of wild-type and mutationally activated FGFR4, we performed western blot analysis on cells treated with ponatinib for eight hours. Wild-type FGFR4 phosphorylation was inhibited by ponatinib in a dose-dependent manner for two fusion-positive cell lines, RH4 and RH5 ( Figure 4A), as well as the mutationally activated FGFR4 cell lines, RMS772/FGFR4 N535K and FGFR4 V550E ( Figure 4B).
We have previously reported that STAT3 is a downstream target of FGFR4 [9]. Therefore, we investigated the effect of ponatinib on the phosphorylation of STAT3 and found a dosedependent reduction of STAT3 phosphorylation in three fusionpositive (RH4, RH5, and RH41) and a fusion-negative cell lines (CTR; Figure 4C), as well as the mutationally activated FGFR4 cell lines ( Figure 4D).

Ponatinib inhibits mutant FGFR4-driven RMS tumor growth in vivo
We then used a previously reported mouse xenograft model [9] to test the efficacy of ponatinib in vivo against RMS cells harboring the constitutively activating FGFR4 mutations. Mice were injected subcutaneously with the RMS772 cell lines stably expressing the empty vector, FGFR4 WT, FGFR4 N535K, or FGFR4 V550E. After 10 days of treatment, the tumor sizes of mice bearing the two mutant FGFR4 RMS772 cell lines were significantly smaller compared to their untreated counterpart ( Figure 5A and 5B). However, there was no difference in tumor volume for mice injected with the RMS772 FGFR4 WT or empty vector cell line when treated with or without ponatinib ( Figure 5C and 5D), indicating that RMS tumors with activating FGFR4 mutations at their tyrosine kinase domain may be more sensitive to the inhibition of ponatinib in vivo.

Discussion
Alteration of FGFR4 signaling is a common mechanism of oncogenesis in both fusion positive and fusion negative rhabdomyosarcoma (RMS). Thus far, at least three mechanisms have been reported to result in the gain of function of FGFR4 in RMS.  First, elevated FGFR4 expression in RMS tumors can be a direct result of the PAX3-FOXO1 fusion oncogene, since FGFR4 was reported to be one of the direct targets of the transcription factor [13]. Secondly, up-regulation of FGFR4 expression in RMS can be achieved through localized gene amplification [22]. Thirdly, 7.5% of primary RMS tumors harbor a damaging missense mutation in the tyrosine kinase domain of FGFR4 which results in a constitutively active signaling molecule [9]. The first two mechanisms result in elevated expression of wild-type FGFR4 in RMS, which is both common and associated with poor outcome [9,12,23]. Previous studies have also shown that knockdown of FGFR4 in RMS cell lines results in inhibition of cell proliferation in vitro and metastasis in vivo [9,24]. The third mechanism of somatic mutation of FGFR4 results in the constitutive, ligand-independent activation of FGFR4 [25]. Given these findings, we hypothesized that inhibition of FGFR4 signaling would be an effective strategy for the treatment of RMS.
In our previous study, we have shown that RMS772 cell lines expressing the FGFR4 V550E or N535K mutation were sensitive to PD173074, a FGFR inhibitor, in the micromolar range [9]. To find FGFR inhibitors that are more effective, we searched for small molecule tyrosine kinase inhibitors that inhibit a FGFR member in clinically achievable concentrations and found five candidate compounds: AP24534 (ponatinib), AZD2171 (Cediranib), BIBF1120 (Vargatef), TKI258 (Dovitunib), and PHA739358 (Danusertib) ( Table S1). It is noteworthy to mention that ponatinib has received accelerated FDA approval for treatment of adult patients with Ph+ ALL or chronic phase, accelerated phase, or blast phase CML who are resistant or intolerant to prior tyrosine kinase inhibitor therapy. Using Ba/F3 cells, we found that transduction of TEL-FGFR4 results in activation of FGFR4 via autophosphorylation and IL-3 independent survival and growth ( Figure S1A-C). Furthermore, we screened the five FGFR inhibitors with this Ba/F3 TEL-FGFR4 model and verified that ponatinib is the most potent FGFR4 inhibitor among those tested [16].
Consistent with these findings, ponatinib inhibited the growth of multiple fusion-positive and fusion-negative RMS cell lines, all with IC 50 values in the nanomolar range. Our data confirmed that FGFR4 mRNA expression was significantly higher in fusionpositive cell lines than fusion-negative cell lines ( Figure S3B, p = 0.0005) [9,24]. In addition, the sensitivity to ponatinib correlated with FGFR4 mRNA expression levels and that fusion-positive cell lines with the higher FGFR4-expressing levels were consistently sensitive to ponatinib. However, cell lines that were fusion-negative or expressed FGFR4 at low levels (less than a relative level of 6) had a wider variation of ponatinib sensitivity. Therefore, it is possible that a certain threshold of FGFR4 expression is needed for consistent nanomolar sensitivity to ponatinib. In addition, no mutations were found in any of the cell lines by Sanger sequencing of exon 12 and 13 of FGFR4 (data not shown), eliminating the possibility of a mutation conferring sensitivity in any of the studied cell lines. However, since ponatinib is a multikinase inhibitor, which includes inhibition of RET, LYN, LCK, FYN, and ABL at subnanomolar concentrations (Table S2), it is possible that the activity is related to the inhibition of other kinases. Indeed, even some cell lines with lower levels of FGFR4 expression continue to demonstrate sensitivity to ponatinib and it is possible that this effect may be the result of inhibition of targets other than FGFR4. This is demonstrated in normal skin fibroblast, osteosarcoma, and Ewing's sarcoma cell lines ( Figure S8).
Similar to the RMS cell lines with high expression of wild-type FGFR4, we have shown ponatinib to be effective against a RMS model system with constitutively activating FGFR4 mutations N535K and V550E. After treatment of cells expressing the mutated FGFR4 with ponatinib, IC 50 values were achieved in the nanomolar range within 24 hours ( Figure S4A-B). We found there was G 1 /S arrest of cell cycling with an increase in the sub G 1 phase fraction indicating cell death, which was confirmed by caspase 3/7 induction.
It is interesting to note that the in vitro data shows ponatinib to be effective against wild-type and mutant FGFR4, whereas our in vivo results show that ponatinib only inhibits tumor growth of cells harboring the FGFR4 mutations but not the wild-type FGFR4. One possible reason for this may come from our observation that the murine RMS cells expressing wild-type FGFR4 have a higher IC 50 than the cells expressing the two mutant FGFR4s. Therefore a higher inhibitory dosage than what was used may be necessary for the treatment of wild-type FGFR4 in order to observe an effect on tumor xenograft growth. Another possible reason for this may be due to the model system we use: our murine RMS772 cell line which artificially expresses human wild-type FGFR4. Although this models human embryonal rhabdomyosarcoma most closely, expressing human wild-type FGFR4 in a mouse cell or growing in an environment with murine stromal growth factors may alter its behavior differently. For example, we have previously shown that human wild-type FGFR4 does not increase growth or migration like mutated FGFR4 does in RMS772 cells [9]. Given our findings, we believe that targeting FGFR4 will be most effective in ERMS with high expression (due to amplification) or mutation of FGFR4 or in alveolar rhabdomyosarcoma (ARMS) where the PAX3/7-FOXO1 fusion gene found in ARMS directly increases expression of FGFR4. Future studies regarding this observation are being actively pursued using in vivo studies of human rhabdomyosarcoma cell lines.
Biochemically, we found that ponatinib effectively decreased phosphorylation of wild type and mutant FGFR4 in a dosedependent manner, indicating that at least one of its biological effects is through targeting FGFR4 kinase function. We also investigated STAT3 phosphorylation because the expression level of this gene is known to be high in RMS [12] and we have previously shown it to be activated by the FGFR4 mutations V550E and N535K as a downstream target of FGFR4 [9]. We found that STAT3 phosphorylation was also inhibited by ponatinib in a dose-dependent manner. Further dissection of the underlying molecular mechanism is underway to determine whether STAT3 is inhibited by ponatinib directly or via the FGFR4 pathway.
In summary, we find ponatinib as a multi-targeted tyrosine kinase inhibitor that displays potent pan-FGFR activity including nanomolar inhibition of FGFR4. In our FGFR inhibitor screen, ponatinib was identified as the most potent FGFR4 inhibitor, successfully inhibiting FGFR4 and its downstream STAT3 phosphorylation. Furthermore, ponatinib caused G 1 /S arrest of cell cycling and inhibition of in vivo tumor growth of RMS cells expressing the FGFR4 N535K and V550E mutations. Given that this molecule has demonstrated an acceptable efficacy and safety profile in ongoing clinical trials and that there is an urgent need to develop novel therapies for patients suffering from RMS, our presented pre-clinical findings strongly support further clinical investigation of ponatinib for RMS patients with overexpressed or mutationally activated FGFR4. Figure S1 The TEL-FGFR4 model system in Ba/F3 cells.