New therapeutic agents in gastrointestinal stromal tumours
Johanna Falkenhorst, Rainer Hamacher, and Sebastian Bauer
INTRODUCTION
Gastrointestinal stromal tumours (GISTs) comprise the most common mesenchymal malignancy of the gastrointestinal tract. Although GISTs are rare can- cers with an incidence of 10– 15 new cases per million per year, the prevalence of patients with active disease has constantly risen [1]. The majority of patients is characterized by activating mutations in the c-kit (80– 85%) or pdgf-receptor alpha genes (PDGFRA; 5–7%) [2,3] and the subgroup of patients lacking mutations in these genes have mostly been further categorized by the leading oncogenic event. The treatment of KIT/PDGFR-mutant GIST was rev- olutionized with the introduction of small molecule inhibitors against the respective gene products. Although responses are often long-lasting, resis- tance is mostly inevitable [median PFS for imatinib (Gleevec; Novartis International AG, Basel, Switzerland): 20 months] but notably KIT/PDGFRA seem to remain dominant oncogenic drivers in patients with progressive disease [4–6]. Both suni- tinib (Sutent, Pfizer Inc., New York City, New York, USA) and regorafenib (Stivarga; Bayer AG, Leverku- sen, Germany), the approved second and third-line treatments, show a broader inhibitory spectrum than imatinib with mostly complimentary inhibi- tion of either the ATP-binding pocket (Exon 13/14) or various activation loop (a-loop, Exon 17) muta- tions of KIT [7&]. The higher potency against KIT comes with less kinase specificity; both drugs had been primarily developed as VEGF-receptor inhib- itors, explaining the less favourable toxicity profile compared with imatinib. The genomic heterogene- ity of secondary KIT resistance mutations represents the greatest obstacle for salvage drugs, as yet direct KIT inhibitors have failed to inhibit KIT regardless of the type of mutation. This illuminates why in sec- ond and third-line the median PFS is comparably lower (sunitinib: 6.2 months; regorafenib: 4.8 months [8,9]). Even if the majority of GIST patients carry imatinib-sensitive mutations, some tumours are primarily resistant by containing KIT exon 17 primary mutations or the PDGFRA Exon 18 D842V mutation [5]. Approved systemic treatment options are not available, but recently several new tyrosine kinase inhibitors have been clinically evaluated that show promising activity.
Within the KIT/PDGFRA-wild-type group of GIST, the molecular spectrum of driving oncogenes has been further differentiated leaving the term ‘wild-type GIST’ unhelpful. Apart from neurofibro- matosis 1 associated GIST [10], BRAF/KRAS-mutant GIST [11] and SDH-deficient GIST, recently very rare cases of GIST-like tumors gave also been shown to have NTRK as an oncogenic driver [12,13].
TARGETING-RESISTANT KIT: NEW MECHANISMS, NEW THERAPEUTICS
After approval of regorafenib in 2014, several kinase inhibitors have been tested in pretreated GIST, all of those not particularly developed or comprehen- sively validated for a KIT/PDGFRA-mutant back- ground [14–17]. Several strategies are currently being pursued to improve disease control such as improving single-agent activity, combination or rapid cycling of complementary inhibitors.
Two novel compounds, avapritinib (BLU-285; Blueprint Medicines, Cambridge, Massachusetts, USA) and ripretinib (DCC-2618; Deciphera Pharma- ceuticals, Waltham, Massachusetts, USA), stand out in recent years as inhibitors specifically targeting the unmet need of a higher specificity combined with a broader inhibitory profile. Both inhibitors have passed the early clinical phase already with a strong focus on GIST and are currently undergoing phase III testing in different treatment lines [18]. Although the clinical experience from phase I has yet not been fully published, the preliminary data have been presented at several conferences that will be further discussed and is summarized in the context of other trials in Table 1 [4,8,9,17,19–26,27].
Avapritinib is a highly selective TKI (PDGFRA, KIT, Abl; Lck, Src; Flt, FGFR) with strong inhibitory potency against KIT, particularly activation loop mutations, as well as the highly resistant PDGFRA D842V mutation as shown in various preclinical models [28&&,29]. Within the phase I NAVIGATOR trial, so far, 167 mostly metastatic GIST patients have received treatment with a median of four prior treatment lines. A daily (q.d.) dose of 400 mg was determined as the maximum tolerated dose (MTD) [19]. Preliminary analyses presented at the 2018 CTOS meeting showed a clinical benefit rate (CBR) of 40% for the 4 months time point in a heavily pretreated cohort with evidence of tumour reduction in 60% of patients. In the third/fourth line cohort, a CBR of 70% was observed, serving as a rationale for the currently ongoing phase III clinical trial (VOYAGER), which compares avapritinib with the standard regorafenib (NCT03465722, [30]). Both preclinical and clinical data suggest that avapritinib inhibits ATP-binding domain mutations less potently and thereby narrowing the therapeutic window. Notably, detection of those mutations (KIT V654A and T670I) by plasma sequencing at baseline appeared to be a negative predictor of tumour shrinkage within the NAVIGATOR trial. Avapritinib was well tolerated with periorbital oedema and anaemia (25%) as most common grade effects. Other toxicities ( grade 3) were diarrhoea (4%), decreased appetite (3%) and fatigue. Of note, reversible grade 1/2 cognitive impairment, particularly short-term memory loss, was reported for 26% of all patients (60/230 pts, 45 grade 1, 15 grade 2) indicating that avapritinib crosses the blood– brain barrier.
Skin-toxicity, particularly hand-foot syndrome, which is frequently seen with regorafenib treatment, was much less common [19]. Ripretinib (DCC-2618) is a type III KIT-inhibitor that was developed for the treatment of GIST [31]. Unlike ATP-competitive drugs, it prevents the active conformation by blocking the so-called switch pocket, which appears to provide a broader efficacy against various resistance mutations in vitro, as shown in KIT-mutant mast cell models [32&&]. This is further underscored by an almost universal reduc- tion of circulating tumour DNA for a broad spec- trum of resistance mutations found in the plasma of patients treated within the trial. Preliminary results reported for this trial showed a median progression- free survival (mPFS) of 42 weeks in the second line, 40 weeks in the third line and 24 weeks in the fourth line along with cohort (n 38, 29, 111, respectively) using a dose of 100–150 mg q.d. Grade 3/4 adverse events were rare with hypertension, anaemia, diar- rhoea, electrolyte changes, abdominal and back pain as well as asymptomatic increase in lipaseand bilirubin among the most common side effects [20,33]. Results of the phase III INVICTUS trial comparing Ripretinib and placebo in fourth-line treatment are expected for 2019 (NCT03353753). Currently, the INTRIGUE trial compares ripretinib with sunitinib in patients progressing on imatinib (second line).
The multikinase inhibitor cabozantinib (XL184/ Cabometyx, Exelixis Inc., Alameda, California, USA) targets MET, AXL, VEGFR2, RET, KIT and ROS1, and is approved for treatment of hepatocellular and renal cell carcinoma. A phase I study with this compound had included four patients with pre- treated GIST in whom a surprising clinical benefit was observed (disease stabilization of 6–20 months). Common adverse effects were palmar-plantar eryth- rodysesthesia, hypertension, diarrhoea, stomatitis, rash and headache [16]. A comprehensive validation of the inhibitor in preclinical models of GIST further confirmed its tumour-specific effects on growth, proliferation and angiogenesis, in both imatinib- sensitive and imatinib-resistant models [15]. A phase II study by the EORTC currently investigates the clinical activity in a larger cohort of patients (NCT02216578). Ponatinib (Iclusig; Takeda, Osaka, Japan) is a potent inhibitor of BCR-ABL, including the gate- keeper mutation at position 315 (T315I) [34]. It has also shown strong preclinical activity against a broad spectrum of resistance mutations, including the notorious activation loop mutations at codon 816, which confers resistance to all the approved inhibitors for GIST [35]. Those findings were sup- ported by incidental observations of tumour shrinkage in heavily pretreated patients. A phase II study testing ponatinib at 45 mg in 45 heavily pre- treated GIST had resulted in a promising CBR of 37% (10/27) at 10 weeks in patients with a primary KIT exon 11 mutation. Among the notable side effects were several cardiovascular events (one myocardial ischemia, one cerebrovascular accident, one periph- eral artery stenosis, one deep vein thrombosis and one pulmonary embolism, four patients with ven- tricular dysfunction) [36]. These cardiovascular side effects were similar as those seen in CML patients and are considered in part dose dependent. Against the background of the clinical benefit, a follow-up phase II study (POETIG) was initiated albeit using a lower q.d. dose of 30 mg of ponatinib in second-line and last-line patients (NCT03171389). Clinical activity had also been observed in two of three heavily pretreated patients with a tolerable toxicity profile [35].
Dasatinib (Sprycel; Brystol-Myers Squibb, New York City, New York, USA) is not only a highly potent inhibitor of BCR-ABL as well but also inhibits various KIT mutants. In addition, SRC kinases and ephrin are among the kinases that are targeted. Recently, the long-term outcome of a first-line phase II study of dasatinib in patients with GIST has been published. At a dose of 70 mg twice a day, dasatinib had strong clinical activity with a median PFS of 13.6 months. In another study (second line), 48 evaluable patients were stratified by Src kinase activation (phosphorylation). The 6-month-PFS was 29% in the overall population and 50% in 14 patients harbouring Src activation suggesting a potential biomarker. Given the high rate of adverse events including higher grade pleural effusion, nau- sea and vomiting, muscle weakness and haemor- rhage, dasatinib is unlikely to go into further clinical development, at least in the dose regimen used.
Preliminary results have also recently been pre- sented for a trial testing the feasibility of combining two drugs that show complementary activity against ATPB-domain and A-loop mutations. The com- bination consisted of pexidartinib (Plexxikon Inc., Berkeley, California, USA), a highly potent CSF1- receptor inhibitor, which is also active against KIT primary and Exon 13/14 mutations, and PLX9486 (Plexxikon Inc.), a strong inhibitor of A-loop muta- tions. PLX9486 was tested as a single agent and in combination (500 mg) with sunitinib 25 mg
q.d. or pexidartinib 600 mg q.d. Clinical activity was observed in both single and combinational treatment with a tolerable toxicity profile [37].
Against the background of a complementary inhibitory profile of sunitinib and regorafenib, Ser- rano et al. [7&] have tested the concept of rapidly alternating drugs to avoid the expected toxicity of a continuous combination. Interestingly, GIST cell lines require several days to restart the cell cycle machinery as measured by Ki-67 or mitotic count after having full biochemical inhibition of KIT phos- phorylation. A corresponding clinical trial based on these observations unfortunately did not hold the promise of the preclinical findings presumably because of a small therapeutic window for this par- ticular drug combination. However, the concept may still deserve further clinical evaluation given the toxicity profiles of next-generation KIT inhibitors. A recent retrospective analysis summarized the data on 71 patients who had been treated with imatinib after having received the three approved TKIs. In this cohort, the median PFS was 5.4 months with an overall survival (OS) of 10.6 months. KIT Exon 11 deletions negatively correlated with PFS and OS [38]. The relatively long median PFS con- trasts with the results of the RIGHT trial, a random- ized controlled phase III trial, wherein imatinib was associated with median PFS of 1.8 months and placebo with 0.9 months. Although retrospective trials usually overestimate the usefulness of drugs, it nonetheless underscores the expert consensus that patients not qualifying for a clinical trial who progress after having received all approved drugs should receive a rechallenge with imatinib or use other TKIs beyond progression [39].
Patients harbouring the PDGFRA D842V mutation have yet been untreatable; none of the approved drugs has shown any meaningful evidence of activ- ity and the prognosis of patients is considered as poor as for metastatic GIST with KIT mutations in the pre-imatinib era [40]. Within the NAVIGATOR trial, 56 patients with D842V mutations have been treated with avapritinib suggesting a universal acti- vity with an overall response rate of 84% and tumour reduction of any degree in 98% of the population [19]. Given the lack of any approved drug for these patients, this represents a long- awaited breakthrough, given the long-lasting knowledge of this oncogenic mutation.
Crenolanib (AROG Pharmaceuticals, Dallas, Texas, USA), a highly selective PDGFRA and FLT3 inhibitor with nanomolar activity against the PDGFRa D842V mutation, has been tested in a dose-finding study (n 16). The CBR of 31% was encouraging, with several patients having a pro- longed disease control [21]. However, only a fraction of these patients has received the ideal dose and schedule that was later selected for the ongoing randomized phase III study [100 mg three times daily (t.i.d.), NCT02847429]. Olaratumab (Lartruvo; Eli Lilly and Company, Indianapolis, Indiana, USA) is a mAb against the PDGF-receptor. Recently, the results of a phase II study in 21 patients with GIST including those with PDGFRA D842V mutations (n 6) were reported. Treatment consisted of olaratumab 20 mg/kg every 14 days. The 12-week-CBR was 50% in the D842V cohort compared with 14.3% in the second cohort (KIT-mutant). The median PFS was 32 vs. 6 weeks [22]. Given the small number of patients, these results can only be descriptive; however, tumour shrinkage, as seen with small molecule inhibitors of PDGFRA D842V, was not observed.
TARGETS BEYOND KIT
GIST patients lacking mutations of KIT or PDGFRA are only poorly studied regarding most effective treatments. For those patients harbouring BRAF- mutations [41], BRAF inhibitors have in the past provided clinical benefit [42]. A clinical basket study evaluating a combination of dabrafenib (Tafinlar; Novartis International AG), a BRAF inhibitor, and trametinib (Mekinist; GlaxoSmithKline, London, UK), a MEK inhibitor, in patients with BRAF V600E mutations is expected to further broaden the clinical data on this subgroup (NCT02034110). A comprehensive genomic profiling of 186 GIST that encompassed 24 KIT/PDGFRA/RAS wildtype patients of whom 12 patients lacking SDHc alter- ations revealed, amongst other findings, patients harbouring ETV6-NTRK3 fusions. A heavily pre- treated patient responded to NTRK inhibitor laro- trectinib (Vitrakvi; Loxo Oncology Inc., Boulder, Colorado, USA) [13]. Corroborating this observa- tion, a phase I study testing larotrectinib in tumours harbouring NRTK fusions also included three GIST patients. All three patients showed a partial remis- sion. Across all entities (n 55), 71% of responses were ongoing after 1 year of treatment [43&]. Another series looked at a cohort of 88 KIT/PDGFRA wildtype GIST for EGF-receptor alterations. Three mutations were identified that could be addressed therapeutically. Taken together, these findings strongy recommend comprehensive genomic testing of patients with metastatic GIST. [44]. In KIT or PDGFRA mutant GIST, highly recur- rent, acquired additional receptor tyrosine kinase (RTK)-mutations seem to be rare events. However, FGFR has previously been shown to confer resis- tance. Targeting FGFR signalling in vitro has resen- sitized GIST cells to imatinib. A combination of imatinib and FGFR-inhibitor BGJ398 (Novartis International AG) showed synergistic effects in mouse GIST xenografts [45]. Another RTK has been investigated by Chen et al. [46] who observed agonistic effects of KIT and an insulin receptor dual inhibition in vitro. Imatinib treatment of GIST xenografts that were insulin receptor depleted (knock-down) resulted in greater tumour shrinkage [46].
Regarding intracellular pathways as potential therapeutic vulnerabilities, Zeng et al. found an activation of the Wnt-pathway by increased beta- catenin mRNA in GIST cell lines and mouse models following 2 weeks of imatinib treatment. The Wnt- pathway inhibitor PKF118–310 (purchased from Calbiochem, Merck Group, Darmstadt, Germany) was effective in imatinib-sensitive and resistant cell lines in vitro and in vivo. A tankyrase inhibitor (G007- LK, purchased from Calbiochem; Merck Group) was effective against GIST cells in vivo and Wnt pathway inhibition synergized with imatinib [47]. Interest- ingly, an analysis of 160 GIST tumour samples revealed an association of beta-catenin expression with KIT exon-11-mutations, high risk and tumour size [48]. Bosbach et al. [49&&] further investigated the MAP-kinase signalling pathway using KIT V558D mutant mice to induce further mutations at KIT protein binding sites. By altering the Src binding site, MAPK signalling and tumour growth was decreased. Double mutant mice lacking also the PI3K-binding site did not develop GIST tumours at all. Treating mice with both PI3K inhibitors (pilar- alisib; Exelixis Inc., Alameda, California, USA)/alpe- lisib (BYL719; Novartis International AG) and combined PI3K/mTOR inhibitors voxtalisib (Exe- lixis Inc.) underscored the oncogenic relevance of the pathway for GIST. Further inhibition of MEK [using binimetinib (Mektovi, Array Biopharma, Boulder, Colorado, USA)] enhanced this effect, again, supporting the essential role of these two pathways in KIT-mutant GIST [49&&].
The role of the immune system including immune- related profiles in GIST are still rather poorly under- stood compared with most epithelial cancers. Immune infiltrates appear to be prognostic [50] and particularly dense in PDGFRA-mutant GIST [51&]. Expression of check-point proteins seem rather low, but preclinical studies in KIT-mutant transgenic mice suggest that PD-1 blockade may improve antitumour activity of imatinib. However, monogenic transgenic mice models of GIST may not yet fully represent malignant human GIST. In addi- tion, very little evidence has yet been published that PD-1 or PD-L1 blockade alone leads to disease con- trol in GIST. Recently, the first empirical clinical trials have been published. The phase 2 PEMBROSARC study evaluated oral Cyclophosphamide and Pembrolizu- mab (Keytruda, Merck & Co, Whitehouse Station, New Jersey, USA) in various sarcoma subtypes. Ten patients with GIST that had progressed on at least imatinib and sunitinib were included. Although the median PFS of 1.3 months suggested no clinical activity, notably three patients (11%) were still sta- ble (SD) after the 6 months of treatment [52]. To this end, cyclophosphamide was chosen as an immuno- sensitizer and direct antitumoral effects against GIST seem unlikely. Another phase Ib study evaluated a combination of dasatinib and ipilimumab (Yervoy; Bristol Myers Squibb, New York City, New York, USA), a CTLA4 inhibitor, in 20 patients with GIST who had at least progressed on two pretreatments. The median PFS was 2.8 months and no objective remissions as per RECIST but seven out of 13 evaluable patients with a partial response by Choi criteria. Given the strong inhibitory potency of dasatinib against various KIT- mutations, the contribution of ipilumumab to this clinical activity is unclear [53&].
CONCLUSION
Novel inhibitors of KIT and PDGFRA are expected to change the treatment landscape in metastatic GIST due to improved specificity and toxicity profiles. When successful in later treatment lines, these next generation of inhibitors will not only be promising candidates for earlier treatment lines but also for combination studies. For KIT/PDGFRA-wildtype GIST, reference disease with comprehensive molec- ular testing may be needed to identify druggable oncogenic events.
Acknowledgements
None.
Financial support and sponsorship
Johanna Falkenhorst and Rainer Hamacher have received travel support from Lilly.
Conflicts of interest
Sebastian Bauer: Advisory Role: Blueprint Medicines, ADC Therapeutics, Lilly, Novartis, Daichii, Plexxikon, Nanobiotix, Deciphera; CME Honoraria: Novartis, Pfizer, Bayer, Lilly, Pharmamar; Research Support: Blue- print Medicines, Incyte, Novartis.
REFERENCES AND RECOMMENDED READING
1. Nilsson B, Bumming P, Meis-Kindblom JM, et al. Gastrointestinal stromal tumors: the incidence, prevalence, clinical course, and prognostication in the
preimatinib mesylate era: a population-based study in western Sweden. Cancer 2005; 103:821 –829.
2. Heinrich MC, Corless CL, Demetri GD, et al. Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin
Oncol 2003; 21:4342 –4349.
3. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 1998; 279.
4. Casali PG, Zalcberg J, Le Cesne A, et al. Ten-year progression-free and overall survival in patients with unresectable or metastatic GI stromal tumors:
long-term analysis of the European Organisation for Research and Treatment of Cancer, Italian Sarcoma Group, and Australasian Gastrointestinal Trials Group intergroup phase III randomized trial on imatinib at two dose levels. J Clin Oncol 2017; 35:1713 –1720.
5. Corless CL, Barnett CM, Heinrich MC. Gastrointestinal stromal tumours: origin and molecular oncology. Nat Rev Cancer 2011; 11:865–878.
6. Wardelmann E, Merkelbach-Bruse S, Pauls K, et al. Polyclonal evolution of multiple secondary KIT mutations in gastrointestinal stromal tumors under
treatment with imatinib mesylate. Clin Cancer Res 2006; 12:1743 –1749.
7. Serrano C, Marino-Enriquez A, Tao DL, et al. Complementary activity of
tyrosine kinase inhibitors against secondary kit mutations in imatinib-resistant
gastrointestinal stromal tumours. Br J Cancer 2019; 120:612 –620.
This is the first experimental approach to tackle genomic heterogeneity by rapidly alternating two complimentary drugs that could otherwise not been combined due to toxicity.
8. Demetri GD, van Oosterom AT, Garrett CR, et al. Efficacy and safety of sunitinib in patients with advanced gastrointestinal stromal tumour after failure
of imatinib: a randomised controlled trial. Lancet 2006; 368:1329– 1338.
9. Demetri GD, Reichardt P, Kang YK, et al. Efficacy and safety of regorafenib for advanced gastrointestinal stromal tumours after failure of imatinib and suni-
tinib (GRID): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet 2013; 381:295 –302.
10. Yantiss RK, Rosenberg AE, Sarran L, et al. Multiple gastrointestinal stromal tumors in type I neurofibromatosis: a pathologic and molecular study. Mod
Pathol 2005; 18:475– 484.
11. Miranda C, Nucifora M, Molinari F, et al. KRAS and BRAF mutations predict primary resistance to imatinib in gastrointestinal stromal tumors. Clin Cancer
Res 2012; 18:1769 –1776.
12. Janeway KA, Kim SY, Lodish M, et al. Defects in succinate dehydrogenase in gastrointestinal stromal tumors lacking KIT and PDGFRA mutations. Proc Natl
Acad Sci U S A 2011; 108:314 –318.
13. Shi E, Chmielecki J, Tang CM, et al. FGFR1 and NTRK3 actionable alterations in ‘Wild-Type’ gastrointestinal stromal tumors. J Transl Med 2016; 14:339.
14. Kang YK, Yoo C, Ryoo BY, et al. Phase II study of dovitinib in patients with metastatic and/or unresectable gastrointestinal stromal tumours after failure
of imatinib and sunitinib. Br J Cancer 2013; 109:2309 –2315.
15. Gebreyohannes YK, Schoffski P, Van Looy T, et al. Cabozantinib is active against human gastrointestinal stromal tumor xenografts carrying different KIT
mutations. Mol Cancer Ther 2016; 15:2845 –2852.
16. Nokihara H, Nishio M, Yamamoto N, et al. Phase 1 study of cabozantinib in Japanese patients with expansion cohorts in non-small-cell lung cancer. Clin
Lung Cancer 2018. [Epub ahead of print]
17. Mir O, Cropet C, Toulmonde M, et al. Pazopanib plus best supportive care versus best supportive care alone in advanced gastrointestinal stromal
tumours resistant to imatinib and sunitinib (PAZOGIST): a randomised, multicentre, open-label phase 2 trial. Lancet Oncol 2016; 17:632–641.
18. BLU-285, DCC-2618 show activity against GIST. Cancer Discov 2017; 7:121–122.
19. Heinrich M, von Mehren M, Jones RL, Bauer S, et al. Avapritinib is highly active and well tolerated in patients with advanced GIST driven by a diverse variety
of oncogenic mutations in KIT and PDGFRA. CTOS Annual Meeting; 2018; Rome, Italy. https://www.blueprintmedicines.com/wp-content/uploads/ 2019/01/CTOS-Avapritinib-Update-Nov-2018.pdf. [12 February 2019].
20. George S, Rodon Ahnert J, Wolf J, et al. 1603OInitial results of phase I study of DCC-2618, a broad-spectrum KIT and PDGFRa inhibitor, in patients (pts)
with gastrointestinal stromal tumor (GIST) by number of prior regimens. Ann Oncol 2018; 29(Suppl 8); mdy299.002, https://doi.org/10.1093/annonc/ mdy299.002.
21. Blay J-Y, Heinrich MC, Hohenberger P, et al. A randomized, double-blind, placebo-controlled, phase III study of crenolanib in advanced or metastatic
GIST patients bearing a D842 V mutation in PDGFRA: the Ripretinib CrenoGIST study. J Clin Oncol 2017; 35(Suppl 15):TS11080–TS11090.