Robust Activity of Avapritinib, Potent and Highly Selective Inhibitor of Mutated KIT, in Patient-Derived Xenograft Models of Gastrointestinal Stromal Tumors

Home , Avapritinib

Author Manuscript Published OnlineFirst on October 1, 2018; DOI: 10.1158/1078-0432.CCR-18-1858 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Robust activity of avapritinib, potent and highly selective inhibitor of mutated KIT, in patient-derived xenograft models of gastrointestinal stromal tumors

Yemarshet K. Gebreyohannes1*, Agnieszka Wozniak1*#, Madalina-Elena Zhai1, Jasmien Wellens1, Jasmien Cornillie1, Ulla Vanleeuw1, Erica Evans2, Alexandra K. Gardino2, Christoph Lengauer2, Maria Debiec-Rychter3, Raf Sciot4, Patrick Schöffski1

1Laboratory of Experimental Oncology, Department of Oncology, KU Leuven, and Department of General Medical Oncology, University Hospitals Leuven, Leuven Cancer Institute, Leuven, Belgium; 2Blueprint Medicines Corporation, Cambridge, MA, USA; 3Department of Human Genetics, KU Leuven and University Hospitals Leuven, Leuven, Belgium; 4Department of Pathology, KU Leuven and University Hospitals Leuven, Leuven, Belgium.

*equal contribution of authors

Running title: Activity of avapritinib in GIST xenografts

Key Words: gastrointestinal stromal tumor, avapritinib, resistance, tyrosine kinase inhibitors

Disclosure of Potential Conflicts of Interest: Blueprint Medicines provided avapritinib and financial support for the study. EE, AKG, and CL are/were employees of Blueprint Medicines. PS received institutional support from Blueprint Medicines for consulting/advisory role, research funding, and travel/ accommodation/ expenses. Other authors declare no potential conflicts of interest.

#Corresponding author: Agnieszka Wozniak, Laboratory of Experimental Oncology, Department of Oncology, KU Leuven, Herestraat 49, 3000 Leuven, Belgium. Phone +32 16 341 669; Fax: +32 16 346 901; email: [email protected]

Prior presentation: Part of the results of this study have been presented at the American Association for Cancer Research Annual Meeting 2017 (April 1-5, 2017 Washington DC) (abstract 2081).

Abstract: 250 words

Main text: 4676 words

Number of figures and tables: 4 figures and 2 tables

Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2018 American Association for Cancer Research. Author Manuscript Published OnlineFirst on October 1, 2018; DOI: 10.1158/1078-0432.CCR-18-1858 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Supplemental material: 4 figures and 1 table

Translational relevance

Advanced gastrointestinal stromal tumors (GIST) are routinely treated with tyrosine kinase inhibitors (TKI). However, over time the vast majority of patients develop resistance to TKI, mainly due to the acquisition of a secondary mutation in the activation loop of KIT. Both imatinib and sunitinib are ineffective in treating GIST with such mutations. Although regorafenib is active against some of these activation loop mutants, in the clinic it achieves a median progression-free survival of only 4.5 months. Avapritinib, a novel, potent and selective inhibitor of KIT and PDGFRA activation loop mutations showed robust in vivo anti- tumor activity in patient-derived GIST xenografts. Our preclinical findings indicate that avapritinib could be a relevant treatment for GIST patients with primary or secondary resistance to approved TKI and support investigation in ongoing clinical trials in patients with GIST.

Abstract

Purpose: Gastrointestinal stromal tumors (GIST) are commonly treated with tyrosine kinase inhibitors (TKI). The majority of patients with advanced GIST ultimately become resistant to TKI due to acquisition of secondary KIT mutations, while primary resistance is mainly caused by PDGFRA p.D842V mutation. We tested the activity of avapritinib, potent and highly selective inhibitor of mutated KIT and PDGFRA, in three patient-derived xenograft (PDX) GIST models carrying different KIT mutations, with differential sensitivity to standard TKI.

Experimental design: NMRI nu/nu mice (n=93) were transplanted with human GIST xenografts with KIT exon 11+17 (UZLX-GIST9KIT11+17), exon 11 (UZLX-GIST3KIT11) or exon 9 (UZLX-GIST2BKIT9) mutations, respectively. We compared avapritinib (10 and 30 mg/kg/qd) vs. vehicle, imatinib (50 mg/kg/bid) or regorafenib (30 mg/kg/qd) [UZLX- GIST9KIT11+17]; avapritinib (10, 30, 100 mg/kg/qd) vs. vehicle or imatinib [UZLX- GIST3KIT11]; and avapritinib (10, 30, 60 mg/kg/qd) vs. vehicle, imatinib (50, 100 mg/kg/bid) or sunitinib (40 mg/kg/qd) [UZLX-GIST2BKIT9].

Results: In all models avapritinib resulted in reduction of tumor volume, significant inhibition of proliferation and reduced KIT signaling. In two models, avapritinib led to remarkable histologic responses, increase in apoptosis and inhibition of MAPK-phosphorylation. Avapritinib showed superior (UZLX-GIST9KIT 11+17 and -GIST2BKIT9) or equal (UZLX- GIST3KIT11) anti-tumor activity to standard dose of imatinib. In UZLX-GIST9KIT11+17, the anti-tumor effects of avapritinib were significantly better than with imatinib or regorafenib.

Conclusions: Avapritinib has significant anti-tumor activity in GIST PDX models characterized by different KIT mutations and sensitivity to established TKI. These data provide strong support for the ongoing clinical trials with avapritinib in patients with GIST (NCT02508532, NCT03465722).

Introduction

Gastrointestinal stromal tumors (GIST) are the most common soft tissue sarcomas of the gastrointestinal tract with an annual incidence of 10-15 cases per million people (1). The discovery that the vast majority of GIST are driven by activating mutations in KIT or platelet derived growth factor receptor alpha (PDGFRA) has improved our understanding of the molecular pathogenesis of GIST and led to the successful development of targeted therapies for this malignancy (2). Mutations in KIT or PDGFRAlead to a constitutive, ligand- independent activation of kinase activity and their downstream signaling cascades, resulting in increased tumor cell proliferation and survival (2). Surgical resection is the only available curative treatment for primary, localized, resectable GIST, yet 40%–50% of patients will experience recurrent or metastatic disease during follow-up (3). Furthermore, a subset of patients is not eligible for surgical treatment due to anatomic limitations, general condition or the presence of synchronous metastatic disease (1). The dependence of GIST on mutated RTK led to the exploration of tyrosine kinase inhibitors (TKI) for the systemic treatment of this rare but well characterized malignancy. Imatinib, a small molecule with ATP-mimetic properties, has become the standard first-line line treatment for patients with locally advanced, recurrent, inoperable or metastatic disease (4). Imatinib has tremendously improved the survival of GIST patients with advanced disease and achieves disease control in ~85% of cases (5). Furthermore, the drug was found to extend relapse-free survival and overall survival when used in the adjuvant setting, after surgery in patients with high risk of relapse (1). In most patients with metastatic GIST however, the duration of response to imatinib is limited. The occurrence of resistance, which is mainly caused by the acquisition of secondary mutations, leads to progression during treatment with imatinib or related compounds. Other small molecule TKI, such as sunitinib and regorafenib, are used in patients who are progressing on or are intolerant to imatinib (6). Despite their well-documented clinical activity in imatinib- refractory GIST and their broader activity against a variety of molecular targets, progression on these agents typically occurs after a median treatment duration of less than a year (6). To date, there are no established standard treatment alternatives for GIST patients after failure on all three approved lines of TKI treatment, but a number of compounds are currently being tested in early clinical trials in patients with refractory tumors. Nevertheless, there is still an unmet medical need for novel treatment approaches, which should be tested first in preclinical settings.

Secondary mutation in KIT or PDGFRA is likely the most important event leading to TKI resistance. These mutations can occur in KIT exon 13 and 14, encoding the ATP binding pocket of the receptor, or in exon 17 and 18, in the kinase activation loop. The latter stabilize the receptor in its active conformation, and the majority of these mutations are known to cause resistance to both imatinib and sunitinib which are widely used as first and second line agents in the clinic (4). Although regorafenib is active against some of these mutated forms, a typical patient receiving this third line treatment progresses after a median period of only 4-5 months (7,8). Apart from the unsatisfactory efficacy of second- and third-line agents, their broader activity against multiple molecular targets leads to off- target toxicity, and many patients do not tolerate sunitinib and regorafenib as well as the first-line standard of care.

Avapritinib (BLU-285, Blueprint Medicines) is an oral, highly-selective, potent investigational inhibitor with activity against KIT exon 17 activation loop mutants, including KIT p.D816V. This mutation is a known driver mutation in systemic mastocytosis (SM) (9). In vitro, avapritinib disrupts KIT signaling as assessed by inhibition of both KIT phosphorylation and activation of downstream proteins such as AKT and signal transducer and activator of transcription 3 (STAT3) in human mast cell and leukemia cell lines (10). In vivo, avapritinib achieves dose-dependent tumor growth inhibition in a mouse model of SM (10). Moreover, avapritinib also inhibits PDGFRA p.D842V (11), the mutation responsible for one out of five primary gastric GIST, for which there is no effective treatment available (12). Avapritinib is currently under investigation in patients with unresectable, treatment-resistant solid tumors including GIST (ClinicalTrials.gov: NCT02508532 and NCT03465722) and in advanced SM (NCT02561988).

In the present work we assessed the preclinical activity of avapritinib in vivo, using three patient derived xenograft (PDX) models of GIST, characterized by diverse KIT mutations and varying sensitivity to the available standard TKI therapies.

Materials and Methods

Xenograft models

For the current project, we transplanted three PDX models, established and fully characterized in the Laboratory of Experimental Oncology, KU Leuven, Belgium. A total of 93 NMRI (nu/nu) mice (Janvier Laboratories) were bilaterally engrafted with models UZLX-GIST9KIT 11+17 (KIT: p.P577del;W557LfsX5;D820G), -GIST3KIT 11 (KIT: p.W557_V559delinsF) and - GIST2BKIT 9 (KIT: p.A502_Y503dup), which are known to retain morphological and

molecular features of the original tumor during passaging. The model characteristics and experimental set up are presented in Table 1. Xenografting of human tumors from consenting GIST patients was approved by the Medical Ethics Committee of the University Hospitals Leuven and the animal experiments using PDX were approved by the Ethics Committee for Animal Research, KU Leuven, and performed according to its guidelines and Belgian regulations.

Drugs, reagents and experimental design

The dosing solutions of imatinib, sunitinib and regorafenib (all from Sequoia Research) were prepared as described earlier (13). Avapritinib, provided by Blueprint Medicines, was dissolved in 0.5% carboxymethyl cellulose supplemented with 1% Tween 80. The resulting suspension was kept at 4°C protected from light; a fresh suspension was prepared every three days. Chemical structures of all drugs used in the study are presented in Supplemental Figure S1, the structure of avapritinib has been previously published by Evans et al. (11). When tumor growth had reached a threshold of 500 mm3, mice were treated with the oral compounds for 16 days by gavaging. The doses of 10 and 30 mg/kg avapritinib, tested in all three models, were chosen based on previous preclinical in vivo work in a KIT exon 17 mutant mastocytoma model demonstrating dose responsive activity that was well tolerated in vivo (11). The increased doses of avapritinib, used in models with primary mutations (i.e. UZLX- GIST3KIT 11 and UZLX-GIST2BKIT 9), were chosen based on the biochemical evaluation of avapritinib which demonstrated IC50 of approximately 0.2-1 nM on KIT exon 17 mutants, 1-2 nM against KIT exon 11 mutants, and >50 nM IC50 against KIT wildtype kinase domain (11) and suggested that potentially higher exposure of avapritinib might be required for inhibition of GIST tumors driven by primary KIT exon 9 and 11 mutants. The detailed information about treatment groups and doses are presented in Table 1. During the dosing period, tumor volume was measured three times per week using a digital caliper and the body weight and general well-being of the animals was followed up daily. At the end of the experiment, mice were sacrificed two hours after the last dose, and tumors were divided with one half snap frozen in liquid nitrogen and one half fixed in 4% buffered formaldehyde for further histopathological and molecular assessments. Anti-tumor activity was assessed based on the evolution of tumor volume expressed as the percentage of the normalized baseline value. Furthermore, histopathology and Western blotting were performed. For each mouse, the bilateral tumors were counted as independent events.

Western blotting and immunohistochemistry (IHC) were conducted using the following antibodies and reagents: KIT from Dako/Agilent; Discovered on GIST 1 (DOG1) from Novocastra; phospho-KITY719 (pKITY719), phospho-KITY703 (pKITY703), phospho- AKTS473 (pAKTS473), AKT, alpha-tubulin, p42/44 mitogen-activated protein kinase (MAPK), phospho-MAPK (pMAPK), 4-E binding protein 1 (4EBP1), phospho-4EBP1 (p4EBP1), histone H3 (HH3), phospho-HH3 (pHH3), and cleaved-PARP (poly ADP ribose polymerase) all from Cell Signaling Technology; Ki67 from Thermo Scientific; EnVision+ System-HRP and 3’diaminobenzidine-tetrahydrochloride (DAB) both from Dako/Agilent. For Western blotting the secondary antibodies, conjugated with horseradish peroxidase, were from Cell Signaling Technology and Western Lightning® Plus-ECL from PerkinElmer was used for band visualization.

Histopathology

Hematoxylin and eosin (H&E) staining was performed to evaluate the general tumor morphology, the histologic response (HR) to treatment, as well as to count mitotic and apoptotic cells. Stained tissue sections were analyzed using an Olympus CH-30M microscope (Olympus). Representative pictures were captured using the Olympus Color View digital camera and analyzed with Olympus Cell D imaging software. The HR was graded by assessing the magnitude of necrosis, myxoid degeneration and/or fibrosis using a previously described grading system: grade 1 (0-10% of tumor area), grade 2 (> 10% and ≤ 50%), grade 3 (> 50% and ≤ 90%), and grade 4 (> 90%) (14,15). Moreover, IHC was performed for KIT and DOG1, Ki67 and pHH3 staining was used to assess proliferation, cleaved-PARP to quantify apoptosis and pMAPK to evaluate KIT pathway activity. Proliferation and apoptosis were assessed by counting the number of mitotic and apoptotic cells on H&E-stained slides and the IHC analysis was also based on counting positive cells. Both evaluations were performed in 10 high power fields (HPF) at 400-fold magnification. The Ki67 labeling index was calculated as an average percentage of Ki67-stained nuclei in 5 digital images taken at 400-fold magnification. KIT pathway inhibition was evaluated by grading the intensity of the pMAPK staining as well as the percentage of tumor area showing positivity, as described in Supplemental Table S1.

Western blotting analysis

To determine the effect of the different treatments on the KIT signaling pathway, Western blotting was performed as described previously (16). Densitometry was done using the AIDA software (Raytest) to do semi-quantitation of the phospho-protein levels (17).

Statistical analysis

The comparison between tumor volumes on day 1 vs. day 16 was done using the Wilcoxon matched pair test (WMP). Comparisons between different treatment groups were done using the Mann–Whitney U test (MWU). A value of p < 0.05 was defined as statistically significant. STATISTICA version 13 (Dell Inc.) was used for all calculations.

Results

Tumor volume assessment

After the 16-day treatment period, vehicle-treated tumors from all models showed a steady and statistically significant increase in relative tumor volume (216% of the baseline volume for UZLX-GIST9KIT 11+17, 293% for -GIST3KIT 11 and 172% for -GIST2BKIT 9; p< 0.05 for all, WMP) (Fig.1). Consistent with previous reports, no significant difference was observed in the UZLX-GIST9KIT 11+17 model between the relative tumor volume of the vehicle- and imatinib treated tumors (13), while regorafenib resulted in modest tumor regression (82% of baseline, p = 0.02 WMP). As expected, treatment with imatinib led to a regression in tumor volume (to 32% of baseline) in UZLX-GIST3KIT 11, confirming the imatinib sensitivity previously described (18). In contrast, UZLX-GIST2BKIT 9 tumors treated with the standard dose of imatinib (50 mg/kg) grew significantly. In this model, doubling the dose of imatinib led to tumor stabilization owing to the dose-dependent sensitivity to imatinib previously observed in this model and in line with the known behavior of KIT exon 9 mutated GIST in the clinic (1). Sunitinib caused significant tumor shrinkage in this model (Fig. 1).

In all three xenograft models, avapritinib (10 mg/kg) resulted in tumor volume stabilization compared to the baseline value. This effect was comparable to the effects induced by the higher dose of imatinib in UZLX-GIST2BKIT 9 (100 mg/kg) or to regorafenib in UZLX- GIST9KIT 11+17 (Fig.1). Remarkably, at the dose of 30 mg/kg, avapritinib treatment resulted in substantial tumor regression as compared to baseline in two of the tested models (to 27% in UZLX-GIST9KIT 11+17 (p=0.005) and to 26% in -GIST3KIT 11 (p=0.008, both WMP), and tumor volume stabilization (90%) in UZLX-GIST2BKIT 9 (p=0.08, WMP). Similarly, in UZLX- GIST3KIT 11 higher dose of avapritinib (100 mg/kg), led to a significant tumor regression to

26% of baseline value (p = 0.005, WMP), which was similar to the effect of imatinib in this model (Fig.1B). In addition, in UZLX-GIST2BKIT 9, avapritinib at a dose of 60 mg/kg led to tumor shrinkage, which was significantly better than imatinib (at both doses) and comparable to sunitinib (Fig. 1C). Taken together, avapritinib induced remarkable and dose-dependent effects on tumor volume in all three models.

During the course of this study, the treatment with avapritinib was well tolerated, and mice had a stable body weight within ethically acceptable limits (Supplemental Fig. S2). We did observe a yellowish skin discoloration in all mice treated with 100 mg/kg of avapritinib, although it did not have any impact on the well-being of animals.

Histopathological evaluation

In all three models, vehicle-treated tumors showed spindle cell morphology and diffuse KIT and DOG1 immunopositivity (Supplemental Fig. S3). These characteristics resembled features observed in the original patient samples used for xenografting, as well as those found in previous passages, proving stable morphology of the models. In addition, KIT mutational analysis of ex-mouse tumor samples confirmed the presence of mutations as seen in the patient biopsy.

HR was assessed on H&E stained slides by evaluating the magnitude of the necrosis, fibrosis and myxoid degeneration in the tumor tissue, induced by different treatments (Fig. 2A). Only minimal (grade 1) HR was observed UZLX-GIST9KIT 11+17 tumors treated with imatinib. Regorafenib induced grade 2 HR in 36% of these tumors mainly through the induction of necrosis. As expected, and in line with our prior observations, imatinib caused grade 2 or higher HR in all treated UZLX-GIST3KIT 11 tumors, with 67% of the xenografts showing grade 3 HR. In the UZLX-GIST2BKIT 9 model, all tumors treated with imatinib (both at standard and higher dose) showed grade 1 HR.

In two out of three models, avapritinib resulted in remarkable HR. In UZLX-GIST9KIT 11+17, 30 mg/kg avapritinib induced grade 2 and grade 3 HR in 60% of the tumors. In UZLX- GIST3KIT 11, 10 mg/kg avapritinib induced grade 2 HR in the vast majority (80%) of the tumors. Moreover, the higher doses (30 and 100 mg/kg) led to grade 3 and grade 4 HR in this model. This effect was slightly more pronounced than in tumors treated with imatinib. Interestingly, in both UZLX-GIST9KIT 11+17 and -GIST3KIT 11, the HR observed with avapritinib was characterized mainly by the induction of myxoid degeneration, which is a

typical response pattern observed in tumors collected from GIST patients who responded to the treatment with imatinib in the clinic (19) (Fig. 2A).

In all models, tumors in the vehicle-treated group showed high mitotic activity with an average of ≥ 45 mitotic figures in 10 HPF. Compared to vehicle, all doses of avapritinib led to a significant reduction of proliferation in all three xenograft models (Fig. 2B, Table 2). Remarkably, in the imatinib-resistant model UZLX-GIST9KIT 11+17, both 10 and 30 mg/kg avapritinib inhibited tumor proliferation significantly better than the other tested agents. The anti-proliferative effect of avapritinib was comparable to imatinib in the imatinib-sensitive model, UZLX-GIST3KIT 11. In UZLX-GIST2BKIT 9 higher doses (30 and 60 mg/kg) of avapritinib inhibited the proliferation significantly better than imatinib. Moreover, avapritinib at 60 mg/kg had the same anti-proliferative effect as sunitinib. These findings were confirmed by Western blotting of pHH3 and by IHC for pHH3 and Ki67 (Fig. 2B, Supplemental Fig. S4A, Table 2). The expression level of pHH3 showed a near-complete absence in avapritinib treated tumors at both 10 and 30 mg/kg doses in UZLX-GIST9KIT 11+17, a marked decrease in - GIST3KIT 11, and a dose-dependent inhibition in -GIST2BKIT 9 (Fig. 2B).

In two out of three models, avapritinib had significant pro-apoptotic activity. In UZLX- GIST9KIT 11+17, avapritinib (30 mg/kg) induced a significant increase in apoptosis (3.4 fold increase compared to vehicle-treated tumors, p <0.001, MWU). This was comparable to the effects of regorafenib (3 fold increase) (Table 2). In UZLX-GIST3KIT 11 all doses of avapritinib led to a significant and dose-dependent increase of apoptotic activity compared to the vehicle-treated. However, the difference in the induction of apoptosis between the 30 and 100 mg/kg avapritinib or in comparison to imatinib were not statistically significant. Of note, in this model the majority of the tumor cells in actively treated tumors were replaced by myxoid matrix, therefore counting apoptotic cells could only be done in areas with remaining viable cells, which may have had an impact on the reliability of the analysis.

Evaluation of KIT signaling

Western blot analysis showed that KIT and its downstream signaling proteins were expressed and activated in vehicle-treated tumors from all three models, as expected (Fig. 3A). In UZLX-GIST9KIT 11+17, avapritinib (30 mg/kg) inhibited phosphorylation of KITY703 as well as its downstream components, AKT and MAPK, and to a lesser extent phosphorylation of 4EBP1. Moreover, in UZLX-GIST3KIT 11, all treatments inhibited the phosphorylation of KIT as well as the downstream signaling proteins. In addition, the expression of total forms of the proteins were lower in imatinib and avapritinib treated tumors (30 and 100 mg/kg) in 11

comparison with vehicle-treated, which was most likely related to the substantial decrease of the cellularity in the response to treatment. Similarly, expression of KIT was found to be lower in UZLX-GIST9KIT 11+17 tumors treated with 30 mg/kg avapritinib as compared to the vehicle-treated group (Fig. 3A). In UZLX-GIST2BKIT 9, phosphorylation of KITY703 and downstream proteins was inhibited by all treatments, with remarkable inhibition resulting from treatment with sunitinib and avapritinib at dose of 60 mg/kg (Fig. 3 A and B). Even though avapritinib did not inhibit phospho-KIT completely in this model (Fig. 3A), the phospho-KIT expression, normalized against the total form of KIT, is significantly lower when compared to the vehicle-treated tumors (Fig. 3B). In UZLX-GIST2BKIT 9sunitinib showed increased ability to decrease phospho-KIT, however its downstream inhibitory effect was mainly thought the inhibition of phospho-AKT and not phospho-MAPK pathway, as is seen with other effective GIST agents (Fig. 3).

Subsequently, we also performed histopathological assessment of KIT signaling using pMAPK immunostaining. The evaluation was based on both the staining intensity and percentage of tumor area showing positivity (Supplemental Table S1). In two models (UZLX- GIST9KIT 11+17 and -GIST3KIT 11) we observed a strong to very strong MAPK phosphorylation in the majority of vehicle-treated tumors, while in UZLX-GIST2BKIT 9 staining was variable with focal positivity. In UZLX-GIST9KIT 11+17, treatment with imatinib or regorafenib did not affect MAPK phosphorylation, but in the majority of tumors treated with avapritinib there was almost complete inhibition of phosphorylation, independent of a dose (Fig. 4). In UZLX- GIST3KIT 11, all active treatments reduced MAPK phosphorylation substantially. In UZLX- GIST2BKIT 9 however, only a slight inhibition of phosphorylation was observed in avapritinib treated tumors in comparison to the vehicle-treated group.

Discussion

Primary and acquired resistance to treatment with established TKI represents the ultimate challenge in the clinical setting of GIST, as there is no approved therapy for those who progress on agents with regulatory approval for treatment of patients with GIST. In the present study we were able to document a robust anti-tumor activity of avapritinib, a potent and selective KIT inhibitor, in three GIST PDX models characterized by different KIT mutations and varying sensitivity to standard TKI. All three xenografts utilized in this study were previously validated for preclinical testing of novel anti-GIST therapies (17,20).

In the presented experiments, imatinib led to a significant tumor regression in UZLX- GIST3KIT 11 and a dose-dependent effect on tumor volume in UZLX-GIST2BKIT 9, as already observed in multiple studies, confirming once again the stable biological behavior of this model (13,17,18). These findings are also in line with the behavior of such tumors in the clinic; patients with KIT exon 9 mutations respond better to higher dose of imatinib, as confirmed by a clinical meta-analysis (21). Furthermore, in UZLX-GIST2BKIT 9, sunitinib caused tumor regression, which is consistent with others’ observations that KIT exon 9 mutations benefit more from the therapy with sunitinib than with other agents (22). Taken together, these data demonstrate that our xenograft models exactly mirror the clinical situation and can provide further useful guidance for clinical evaluation of novel therapeutic approaches for GIST.

Avapritinib was designed to potently and selectively inhibit of KIT exon 17 mutations (11). In vitro biochemical activity for the KIT p.D816V mutated receptor was achieved with an

IC50=0.27 nM (11). This mutation is known to be causative for SM, and is found in the vast majority of patients with this systemic disorder (10). Furthermore, avapritinib inhibits proliferation both in vitro and in vivo in leukemia models that harbor the KIT exon 17 p.N822K mutation (23). As expected, in our PDX model with KIT exon 11 and 17 mutations (UZLX-GIST9KIT 11+17 with KIT: p.P577del;W557LfsX5;D820G), avapritinib showed a beneficial effect on tumor volume as well as on proliferation, which most likely was caused by inhibition of KIT signaling. Of note the secondary p.D820G exon 17 KIT mutation, present KIT 11+17 in our UZLX-GIST9 model, has been reported in several TKI-resistant GIST cases with the incidence similar to other mutations affecting kinase activation loop domain of the receptor (24-27). In our UZLX-GIST9KIT 11+17 resistant model the dose of 10 mg/kg affected pathway activation, which resulted in a remarkable decrease of proliferation, however this dose led only to tumor stabilization and limited HR. On the other hand, the higher dose of the investigational agent (30 mg/kg) caused a striking tumor volume shrinkage (to 27% of baseline) and an impressive HR with complete absence of proliferative activity. This observation is similar to findings by Evans et al. where avapritinib potently inhibited the mouse equivalent of KIT p.D816Y driven tumor growth in vivo in a dose dependent manner (11). Moreover, in a GIST PDX model derived from a refractory GIST tumor which harbors KIT exon 11/17 (p.K557_K558del;Y823D), avapritinib again led to tumor regression from the baseline value (11). Interestingly, in our experiments the pattern of response observed in UZLX-GIST9KIT 11+17 tumors treated with 30mg/kg was characterized by myxoid

degeneration – a phenomenon where viable tumor cells are replaced by an amorphous collagenous matrix containing only a few scattered cells. Myxoid degeneration is described frequently as a feature characteristic for GIST responding to the treatment with imatinib, both in preclinical and clinical settings (18,19). Although in this resistant model the effect of 30 mg/kg avapritinib on the tumor volume and histological features was significantly better than what was achieved with 10 mg/kg, both treatment groups exposed to this novel TKI had better responses than imatinib or regorafenib treated tumors. The anti-tumor activity of avapritinib is currently being evaluated in GIST (NCT02508532 and NCT03465722) (28). Of note, the maximum tolerated dose of avapritinib in patients is 400 mg/day while the dose of 300 mg/day is being evaluated in the currently ongoing Phase 3 trial (NCT03465722). The animal equivalent dose, calculated based on the body surface (29), is respectively 82 and 61.5 mg/kg, which falls within the range of doses tested in the present study.

Subsequently, we evaluated avapritinib in two additional PDX models with a primary mutation in KIT exon 11 (UZLX-GIST3KIT 11, imatinib sensitive) or exon 9 (UZLX- GIST2BKIT 9, dose-dependent sensitivity to imatinib, sensitive to sunitinib). We observed a significant inhibition of proliferation in these models at all doses tested, suggesting broader inhibitory capacity of avapritinib against different KIT mutations outside of the activation loop region (11). In the UZLX-GIST3KIT 11 model, we found a pronounced effect on tumor volume and histological responses at 30 and 100 mg/kg, and this observation was similar to the effect caused by imatinib. On the other hand, avapritinib in UZLX-GIST2BKIT 9 was better than imatinib, but only the dose of 60 mg/kg led to a similar efficacy as achieved with sunitinib. This observation suggests that avapritinib could also be effective in GIST with primary mutations, however higher in vivo concentrations are required for exon 9 anti-tumor activity. Our studies in mice suggest avapritinib is well tolerated at these higher concentrations but clinical data will be required to fully assess the activity of avapritinib in patients with KIT exon 9 driven tumors The ongoing Phase 1 clinical trial with avapritinib accepts patients who have failed two or more agents and likely have accumulated secondary resistance mutations, and patients with PDGFRA p.D842 mutant-driven GIST independent of prior lines of treatment. Interestingly, it is already known that avapritinib inhibits the activity of PDGFRA p.D842V mutated receptor in vitro (11); the in vivo evaluation is hampered by the lack of relevant models with this mutation.

In our study, we found only a moderate pro-apoptotic effect in UZLX-GIST3KIT 11 and - GIST9KIT 11+17 (at a dose of 30 mg/kg), in part explained by the presence of myxoid

degeneration as a response to avapritinib in these models. The reduced cellularity potentially leads to an underestimation of the pro-apoptotic effects of this treatment in our models. This speculation is supported by the observation of increased apoptosis in tumors harvested after 8 days of treatment with avapritinib, when the histological response was not as pronounced as on day 16 (11). The pro-apoptotic effect of avapritinib is likely induced by inhibition of KIT signaling. This observation matches previous findings reported by our group, where potent KIT signaling inhibition resulted in an increased apoptosis (16,17). Moreover, Evans et al. showed pro-apoptotic activity of avapritinib in a mouse mastocytoma cell line with KIT exon 17 substitution p.D814Y, the equivalent of human p.D816Y mutant (10,23). On the other hand, in the UZLX-GIST2BKIT 9, neither of the experimental treatments led to an increase of apoptotic activity after 16 days of treatment. A trend towards a decline of apoptosis was observed with increasing dose of avapritinib. Similar observations were previously reported in this model, where treatments that led to inhibition of KIT signaling resulted in a decrease of apoptotic activity (17,20). This effect could be a consequence of the variable expression of signaling molecules due to specific KIT genotype as reported by several groups (30-32).

Treatment with avapritinib was well tolerated in our nude mice. Although we observed yellow skin discoloration in mice treated with the highest avapritinib dose tested (100 mg/kg), it did not impact the animals’ well-being. We also did not find any macroscopic or microscopic changes in their organs, including the liver. The skin color change may be due to a stronger systemic inhibition of KIT, disrupting its physiological function in follicular melanocytes and or impairing melanogenesis (33). Skin depigmentation as a result of strong KIT inhibition was previously reported by Kim et al. in the evaluation of another potent KIT inhibitor, PLX3397 and was reported in metastatic renal cell carcinoma, treated with TKI sorafenib (34,35).

As a selective inhibitor of KIT and PDGFRA p.D842V mutations, avapritinib is less likely to cause significant off target treatment-related toxicity at efficacious doses, in contrast to multi- targeted agents currently being used in the clinical setting. Agents such as sunitinib, regorafenib, dasatinib or ponatinib show effectiveness in some genotypes of resistant GIST, but dose-limiting toxicities, arising from the simultaneous blockage of several kinases, translates into higher toxicity and limit the clinical usefulness of some of these agents (36,37). While very preliminary data from the ongoing Phase 1 clinical trial suggests a favorable safety profile (28), the high specificity of avapritinib can in theory increase the risk of rapid development of secondary resistance to this compound. GIST is known for its clinical inter- and intratumoral heterogeneity in terms of the mutational profile, and it is likely that some

clones of this disease may lead to further progression on treatment with the novel agent. In this context, it is noteworthy that we saw broad activity in a variety of GIST genotype in our mice.

In conclusion, we provide in vivo evidence that the novel TKI avapritinib has significant antitumor activity in GIST PDX models. Our results demonstrate that in KIT exon 11/17 double mutated GIST, this inhibitor is more active than established standard treatments. Moreover, in imatinib sensitive models with primary KIT mutations, avapritinib shows a similar or even higher level of activity in comparison with imatinib. In all models tested, the pharmacological anti-proliferative effect on tumor volume was mainly achieved through a marked inhibition of KIT signaling. Our data strongly support the scientific rationale of the ongoing exploration of avapritinib in GIST and provide arguments for exploration of the novel compound in the clinic in both imatinib sensitive and TKI resistant genotypes of this mesenchymal malignancy. The results seen in our patient derived mouse xenografts and the early findings reported from the ongoing clinical trial in GIST patients suggest that avapritinib has the potential to become an important treatment option for this orphan malignancy.

References

1. Joensuu H, Hohenberger P, Corless CL. Gastrointestinal stromal tumour. Lancet. 2013; 382:973-83. 2. Corless CL, Barnett CM, Heinrich MC. Gastrointestinal stromal tumours: origin and molecular oncology. Nat Rev Cancer. 2011; 11:865-78. 3. Rutkowski P, Hompes D. Combined Therapy of Gastrointestinal Stromal Tumors. Surg Oncol Clin N Am. 2016; 25:735-59. 4. Nishida T, Blay JY, Hirota S, Kitagawa Y, Kang YK. The standard diagnosis, treatment, and follow-up of gastrointestinal stromal tumors based on guidelines. Gastric Cancer. 2016; 19:3-14. 5. Demetri GD, von Mehren M, Blanke CD, Van den Abbeele AD, Eisenberg B, et al. Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors. N Engl J Med. 2002; 347:472-80. 6. Metaxas Y, Oikonomopoulos G, Pentheroudakis G. Update on clinical research and state of the art management of patients with advanced sarcomas and GIST. ESMO Open. 2016; 1:e000065.

7. George S, Wang Q, Heinrich MC, Corless CL, Zhu M, Butrynski JE, et al. Efficacy and safety of regorafenib in patients with metastatic and/or unresectable GI stromal tumor after failure of imatinib and sunitinib: a multicenter phase II trial. J Clin Oncol. 2012; 30:2401-7. 8. Demetri GD, Reichardt P, Kang YK, Blay JY, Rutkowski P, Gelderblom H, et al. Efficacy and safety of regorafenib for advanced gastrointestinal stromal tumours after failure of imatinib and sunitinib (GRID): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet. 2013; 381:295-302. 9. Abid A, Malone MA, Curci K. Mastocytosis. Prim Care. 2016; 43:505-18. 10. Evans EK, Hodous BL, Gardino A, Zhu J, Shutes A, Davis A, et al. First selective KIT D816V inhibitor for patients with systemic mastocytosis. Blood. 2014; 124:3217 (#abstract 3217). 11. Evans EK, Gardino A, Kim JL, Hodous BL, Shutes A, Davis A, et al. A precision therapy against cancers driven by KIT/PDGFRA mutations. Sci Transl Med. 2017; 9. 12. Wozniak A, Rutkowski P, Schöffski P, Ray-Coquard I Hostein I, Schildhaus HU, et al. Tumor genotype is an independent prognostic factor in primary gastrointestinal stromal tumors of gastric origin: a European multicenter analysis based on ConticaGIST. Clin Cancer Res. 2014; 20:6105-16. 13. Van Looy T, Gebreyohannes YK, Wozniak A, Cornillie J, Wellens J, Li H, Vanleeuw U, et al. Characterization and assessment of the sensitivity and resistance of a newly established human gastrointestinal stromal tumour xenograft model to treatment with tyrosine kinase inhibitors. Clin Sarcoma Res. 2014; 4:10. 14. Antonescu CR, Besmer P, Guo T, Arkun K, Hom G, Koryotowski B, et al. Acquired resistance to imatinib in gastrointestinal stromal tumor occurs through secondary gene mutation. Clin Cancer Res. 2005; 11:4182-90. 15. Agaram NP, Besmer P, Wong GC, Guo T, Socci ND, Maki RG, et al. Pathologic and molecular heterogeneity in imatinib-stable or imatinib-responsive gastrointestinal stromal tumors. Clin Cancer Res. 2007; 13:170-81. 16. Floris G, Sciot R, Wozniak A, Van Looy T, Wellens J, Faa G, et al. The Novel HSP90 inhibitor, IPI-493, is highly effective in human gastrostrointestinal stromal tumor xenografts carrying heterogeneous KIT mutations. Clin Cancer Res 2011; 17:5604–14. 17. Floris G, Wozniak A, Sciot R, Li H, Friedman L, Van Looy T, et al. A potent combination of the novel PI3K Inhibitor, GDC-0941, with imatinib in gastrointestinal

stromal tumor xenografts: long-lasting responses after treatment withdrawal. Clin Cancer Res 2013; 19:620–30. 18. Van Looy T, Wozniak A, Floris G, Sciot R, Li H, Wellens J, et al. Phosphoinositide 3- kinase inhibitors combined with imatinib in patient-derived xenograft models of gastrointestinal stromal tumors: rationale and efficacy. Clin Cancer Res. 2014; 20:6071-82. 19. Sciot R, Debiec-Rychter M. GIST under imatinib therapy. Semin Diagn Pathol. 2006; 23:84-90. 20. Gebreyohannes YK, Schöffski P, Van Looy T, Wellens J, Vreys L, Cornillie J, et al. Cabozantinib Is Active against Human Gastrointestinal Stromal Tumor Xenografts Carrying Different KIT Mutations. Mol Cancer Ther. 2016; 15:2845-2852. 21. Gastrointestinal Stromal Tumor Meta-Analysis Group (MetaGIST). Comparison of two doses of imatinib for the treatment of unresectable or metastatic gastrointestinal stromal tumors: a meta-analysis of 1,640 patients. J Clin Oncol. 2010; 28:1247-53. 22. Reichardt P, Demetri GD, Gelderblom H, Rutkowski P, Im SA, Gupta S, et al. Correlation of KIT and PDGFRA mutational status with clinical benefit in patients with gastrointestinal stromal tumor treated with sunitinib in a worldwide treatment-use trial. BMC Cancer. 2016; 16:22. 23. Evans EK, Hodous BL, Gardino A, Davis A, Zhu J, et al. BLU-285, a potent and selective inhibitor for hematologic malignancies with KIT Exon 17 Mutations. Blood. 2015; 126 (23): 568 (#abstract 568). 24. Heinrich MC, Corless CL, Demetri GD, Blanke CD, von Mehren M, Joensuu H, et al. Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin Oncol. 2003; 21:4342-9. 25. Wardelmann E, Merkelbach-Bruse S, Pauls K, Thomas N, Schildhaus HU, Heinicke T, et al. Polyclonal evolution of multiple secondary KIT mutations in gastrointestinal stromal tumors under treatment with imatinib mesylate. Clin Cancer Res. 2006; 12:1743-9. 26. Liegl B, Kepten I, Le C, Zhu M, Demetri GD, Heinrich MC, et al. Heterogeneity of kinase inhibitor resistance mechanisms in GIST. J Pathol. 2008; 216:64-74. 27. Heinrich MC, Maki RG, Corless CL, Antonescu CR, Harlow A, Griffith D, et al. Primary and secondary kinase genotypes correlate with the biological and clinical activity of sunitinib in imatinib-resistant gastrointestinal stromal tumor. J Clin Oncol. 2008; 26:5352-9.

28. Heinrich MC, Jones R, Schoffski P, Bauer S, von Mehren M, Eskens F, et al. Preliminary safety and activity in a first-in-human phase 1 study of BLU-285, a potent, highly-selective inhibitor of KIT and PDGFR alpha activation loop mutants in advanced gastrointestinal stromal tumor (GIST). Eur J Cancer 2016; 68: abstract 6LBA. 29. Nair AB, Jacob S. A simple practice guide for dose conversion between animals and human. J Basic Clin Pharm. 2016; 7:27-31. 30. Landuyt B, Prenen H, Debiec-Rychter M, Sciot R, de Bruijn EA, van Oosterom AT. Differential protein expression profile in gastrointestinal stromal tumors. Amino Acids. 2004; 27:335-7. 31. Antonescu CR, Viale A, Sarran L, Tschernyavsky SJ, Gonen M, Segal NH, et al. Gene expression in gastrointestinal stromal tumors is distinguished by KIT genotype and anatomic site. Clin Cancer Res. 2004; 10:3282-90. 32. Duensing A, Medeiros F, McConarty B, Joseph NE, Panigrahy D, Singer S, et al. Mechanisms of oncogenic KIT signal transduction in primary gastrointestinal stromal tumors (GISTs). Oncogene 2004; 23: 3999–4006. 33. D’Mello SAN, Finlay GJ, Baguley BC, Askarian-Amiri ME. Signaling Pathways in Melanogenesis. Sugumaran M, ed. International Journal of Molecular Sciences. 2016;17(7):1144. doi:10.3390/ijms17071144. 34. Kim TS, Cavnar MJ, Cohen NA, Sorenson EC, Greer JB, Seifert AM, at al. Increased KIT inhibition enhances therapeutic efficacy in gastrointestinal stromal tumor. Clin Cancer Res. 2014; 20:2350-62. 35. Dasanu CA, Alexandrescu DT, Dutcher J. Yellow skin discoloration associated with sorafenib use for treatment of metastatic renal cell carcinoma. South Med J. 2007; 100: 328-30. 36. Ashman LK, Griffith R. Therapeutic targeting of c-KIT in cancer. Expert Opin Investig Drugs. 2013; 22:103-15. 37. Garner AP, Gozgit JM, Anjum R, Vodala S, Schrock A, Zhou T, et al. Ponatinib inhibits polyclonal drug-resistant KIT oncoproteins and shows therapeutic potential in heavily pretreated gastrointestinal stromal tumor (GIST) patients. Clin Cancer Res. 2014; 20:5745–55.

Figures legend

Figure 1. Evolution of tumor volume during the treatment, presented as relative tumor volume (% change compared to normalized, baseline value) +/- standard deviation in UZLX- GIST9KIT 11+17 (A), -GIST3 KIT 11 (B) and –GIST2B KIT 9 (C).

Figure 2. Assessment of histologic response (HR) (A). Representative images of H&E staining after treatment (200x magnification) (B). Mean number of mitotic cells in 10 high power field (HPF) grouped by treatment (C). Assessment of phospho-histone H3 (pHH3) expression by Western blotting (D).

Figure 3. KIT signaling pathway. Assessment of the effect of treatments in different xenograft models (A). Densitometric assessment of phospho-protein forms in KIT signaling pathway (B).

Figure 4. Evaluation of pMAPK positivity based on the intensity and percent tumor area showing pMAPK positivity (A) Grading system is presented in Supplemental Table S1. Representative images of pMAPK immunostaining of the different treatment groups at 200x (B).

Table 1. Detailed description of xenograft models and experimental set-up.

Model name UZLX-GIST9KIT 11+17 UZLX-GIST3KIT 11 UZLX-GIST2BKIT 9

Model characteristics KIT mutation Exon 11: Exon 11: Exon 9: p.P577del;W557LfsX5; p.W557_V559delinsF p.A502_Y503dup exon 17: D820G

Sensitivity to imatinib Resistant Sensitive Dose dependent in vivo sensitivity

Treatment groups

control vehicle a (n=7) vehicle a (n=6) vehicle a (n=4)

imatinib 50 mg/kg/bid (n=7) 50 mg/kg/bid (n=5) 50 mg/kg/bid (n=4)

n/a n/a 100 mg/kg/bid (n=4)

sunitinib n/a n/a 40 mg/kg/qd (n=4)

regorafenib 30 mg/kg/qd (n=7) n/a n/a

avapritinib 10 mg/kg/qd (n=6) 10 mg/kg/qd (n=6) 10 mg/kg/qd (n=5)

30 mg/kg/qd (n=6) 30 mg/kg/qd (n=6) 30 mg/kg/qd (n=5)

n/a n/a 60 mg/kg/qd (n=5)

n/a 100 mg/kg/qd (n=6) n/a

a0.5% carbomethyl cellulose with 1% Tween 80; n-number of mice; n/a-not applicable, bid- twice daily; qd-once daily.

Table 2. Assessment of proliferation and apoptotic activity in GIST. Values are presented as fold change in comparison to the vehicle-treated tumors.

UZLX-GIST9KIT 11+17 UZLX-GIST3 KIT 11 UZLX-GIST2B KIT 9

pHH3 Ki67 pHH3 Ki67 pHH3 Ki67

imatinib (50 mg/kg) = = ↓↓↓* ↓↓↓* = =

imatinib (100 mg/kg) n/a n/a n/a n/a ↓2.5* ↓2.1*

sunitinib n/a n/a n/a n/a ↓↓↓* ↓↓↓* activity regorafenib = = n/a n/a n/a n/a

avapritinib (10 mg/kg) ↓14.4* ↓↓↓* ↓↓↓* ↓↓↓* ↓1.3* ↓1.9*

avapritinib (30 mg/kg) ↓↓↓* ↓↓↓* ↓↓↓* ↓↓↓* ↓3.4* ↓3.9* Proliferative Proliferative avapritinib (60 mg/kg) n/a n/a n/a n/a ↓13.9* ↓7.9*

avapritinib (100 mg/kg) n/a n/a ↓↓↓* ↓↓↓* n/a n/a

Cleaved Cleaved Cleaved H&E H&E H&E PARP PARP PARP imatinib (50mg/kg) ↑2.1* = ↑6.5* ↑6.8* = =

imatinib (100mg/kg) n/a n/a n/a n/a ↓6.8* ↓2.4*

sunitinib n/a n/a n/a n/a ↓3.7 ↓1.3*

regorafenib ↑3.0* ↑2.6* n/a n/a n/a n/a

avapritinib (10mg/kg) = ↓2.6* ↑4.5* ↑2.3* = =

avapritinib (30mg/kg) ↑3.4* ↑3.0* ↑5.7* ↑17.1* ↓2.3* ↓1.9* Apoptotic activity avapritinib (60mg/kg) n/a n/a n/a n/a ↓2.6* ↓1.9*

avapritinib (100mg/kg) n/a n/a ↑8.7* ↑18.1* n/a n/a H&E-hematoxylin and eosin, PARP - poly-ADP ribose polymerase, pHH3-phospho-histone H3, ↓ decease, ↓↓↓ > 50 fold decrease, ↑ increase, = no significant difference, *p < 0.05 compared to the vehicle-treated by Mann-Whitney U test.

Robust activity of avapritinib, potent and highly selective inhibitor of mutated KIT, in patient-derived xenograft models of gastrointestinal stromal tumors

Yemarshet K Gebreyohannes, Agnieszka Wozniak, Madalina-Elena Zhai, et al.

Clin Cancer Res Published OnlineFirst October 1, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-18-1858

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2018/09/29/1078-0432.CCR-18-1858.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet Manuscript been edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2018/09/29/1078-0432.CCR-18-1858. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.