Inhibitory effects of midostaurin and avapritinib on myeloid progenitors derived from patients with KIT D816V positive advanced systemic mastocytosis

Advanced systemic mastocytosis (advSM) is characterized by the presence of an acquired KIT D816V mutation in >90% of patients. In the majority of patients, KIT D816V is not only detected in mast cells but also in other hematopoietic lineages. We sought to investigate the effects of the KIT-inhibitors midostaurin and avapritinib on single-cell-derived myeloid progenitor cells using granulocyte-macrophage colony-forming-units of patients with KIT D816V positive advSM. Colonies obtained prior to treatment were incubated in vitro with midostaurin (n = 10) or avapritinib (n = 11) and showed a marked reduction (≥50%) of KIT D816V positive colonies in 3/10 (30%) and 7/11 (64%) patient samples, respectively. Three of those 7 (43%) avapritinib responders were resistant to midostaurin in both, in vitro and in vivo. Colonies from four patients with high-risk molecular profile and aggressive clinical course were resistant to both drugs. The in vitro activity of midostaurin strongly correlated with clinical and molecular responses, e.g., relative reduction of KIT D816V allele burden and the proportion of KIT D816V positive colonies obtained after six months midostaurin-treatment in vivo. We conclude that the colony inhibition assay provides useful information for prediction of responses on midostaurin and that avapritinib has a superior in vitro activity compared to midostaurin.

Because of the significance of KIT D816V in the pathogenesis of advSM, targeted drugs against the oncogenic mutation have been developed. Assessing the safety and efficacy of midostaurin (PKC-412) in a multicenter, openlabel, single-arm phase 2 study (NCT00233454), the multikinase/KIT-inhibitor (IC 50 of 2.9 nM) has demonstrated an overall response rate (ORR; major + partial response) of 60% per Valent criteria (28% in a separate post hoc analysis by the European medicines Agency [EMA] [20,21]. However, validated biomarkers for prediction of response in advSM patients treated with midostaurin are still lacking. Avapritinib (BLU-285), a potent and highly selective KIT D816V inhibitor (IC 50 of 0.27 nM), has shown preclinical activity as well as encouraging results in an open-label, dose-escalation in phase I trial evaluating the safety and antineoplastic activity (NCT02561988) [22][23][24].
The aim of the present study was to establish an amenable in vitro assay to investigate the inhibitory effects of midostaurin and avapritinib on single-cell-derived myeloid progenitor cells using granulocyte-macrophage colonyforming-units (CFU-GM) of patients with KIT D816V advSM and to correlate in vitro colony data with clinical and molecular characteristics at baseline, and response parameters of midostaurin-treated advSM patients in vivo.
The clinical response to treatment was evaluated by measurable C-findings (excluding ascites and osteolytic lesions) according to modified Valent response criteria as previously described [3,20].
Reference pathologists of the ECNM evaluated all bone marrow biopsies. The study design adhered to the tenets of the Declaration of Helsinki and was approved by the relevant institutional review board of the Medical Faculty of Mannheim, Heidelberg University, as part of the 'German Registry on Disorders of Eosinophils and Mast Cells'. All patients provided written informed consent.

Quantitative assessment of KIT D816V
Quantitative assessments of the KIT D816V expressed alele burden (EAB) were performed using allele-specific quantitative real-time reverse-transcriptase polymerase chain reaction (qRT-PCR) analysis on RNA/complementary DNA as previously described [14].

Targeted next-generation sequencing (NGS) analysis
Next-Generation Deep Amplicon Sequencing by 454 FLX amplicon chemistry (Roche, Penzberg, Germany) with consistent detection sensitivity of EAB down to 3% was performed in all patients to investigate 18 candidate genes as previously described [15]. The customized sequencing panel targeted the hotspot or complete coding regions of the following 18 genes: ASXL1, CBL, ETV6, EZH2, IDH1, IDH2, JAK2, KRAS, NPM1, NRAS, RUNX1, SETBP1, SF3B1, SRSF2, TET2, TP53, U2AF1, and ZRSR2. The sequential NGS approach is based on library preparation by the Access Array Technology (Fluidigm, San Francisco, CA) and sequencing on the MiSeq Instrument (Illumina, San Diego, CA). Gene mutations were annotated using the reference sequence of the Ensembl Transcript ID (Ensembl release 85: July 2016).

CFU-GM colony assay
The CFU-GM colony assay is an in vitro assay based on primary bone marrow mononuclear cells using semi-solid methylcellulose (0.9%) matrix supplemented with 30% fetal bovine serum albumin (FBS), 1% BS albumin, 0.1 M 2mercaptoethanol and recombinant human GM-CSF (100 ng/ml; MethoCult, StemCell Technologies, Cologne, Germany) in 35 mm Petri-dishes. The cells (1 × 10 5 cells in 1 mL MethoCult) were incubated at 37°C in a humidified atmosphere with 5% CO 2 until colonies appeared after 10-14 days. Per colony, 100-300 cells were diluted in phosphate-buffered saline. Based on previous publications and for proof-of-principle, we incubated treatment-naive CFU-GM colonies with 100 nM to 1 µM of midostaurin and 22 nM to 90 nM of avapritinib, respectively. Based on the obtained data from these assays (maintenance of colony growth in combination with optimum decreasing of KIT D816V positive CFU-GM colonies), we performed our experiments with 600 nM midostaurin and 75 nM avapritinib, respectively [25][26][27]. Figure 1 outlines an overview on the various colony assays.

Genotyping of CFU-GM
Whole-genome amplification (REPLI-g, Qiagen, Hilden, Germany) was performed to determine the mutational status of single-cell-derived CFU-GM colonies (mean colonies per assay per patient, n = 15; range 10-30, at least 10 colonies were evaluated). Sanger sequencing for mutation validation of KIT D816V and additional mutations was performed after PCR amplification of the relevant region. CFU-GM colonies are expected to be either positive (50% in case of heterozygosity, 100% in case of homozygosity) or negative for any mutation since they are derived from a single myeloid progenitor cell.

Cytogenetic analysis
For cytogenetic analysis, at least 20 Giemsa-banded bone marrow metaphases cultured for 24 h and/or 48 h were prepared as previously described, analyzed by G-/Rbanding technique and interpreted according to the International System for Human Cytogenetic Nomenclature [28,29]. Table 1 Summarized clinical, laboratory, histological, and molecular characteristics of 13 KIT D816V positive advanced systemic mastocytosis patients prior to treatment based on response pattern in single-cell-derived myeloid progenitor cells (CFU-GM colonies, relative reduction of KIT D816V positive colonies), three cohorts were defined: midostaurin + avapritinib responder (cohort #1), midostaurin non-responder + avapritinibresponder (cohort #2), and midostaurin + avapritinib non-responder (cohort #3) Hemoglobin, g/dL; median (range) 9.9 ( ANC absolute neutrophil count, BM bone marrow, EAB expressed allele burden, MC mast cell, PB peripheral blood a Non-measurable C-findings (e.g., ascites and osteolytic lesions) were excluded b Organomegaly including hepatomegaly, splenomegaly and/or lymphadenopathy c Additional mutations were detected using targeted sequencing panel to investigate 18 candidate genes   Data not available, KIT D816V EAB in PB was 33%

Statistical analysis
All statistical analyses considered clinical and laboratory parameters as well as experimental data obtained at the time of midostaurin initiation and after six months treatment (in vivo). Pearson's correlation coefficient was used to compare the change of KIT D816V positive colonies in vitro after two weeks incubation with midostaurin and avapritinib and in vivo after six months midostaurintreatment. The phi coefficient was used to evaluate the association between response according to the mutational status and the KIT D816V EAB in peripheral blood and response to midostaurin in vitro/in vivo. A paired t-test was used to compare the relative reduction in the proportion of KIT D816V positive colonies from baseline to in vitrocolonies incubated with midostaurin and avapritinib. OS was defined as the time between diagnosis and the date of death or last contact. P values < 0.05 (two-sided) were considered significant. GraphPad Prism Software (version 5, GraphPad, La Jolla, CA, USA) and SPSS (version 21.0.0, IBM Cooperation, Armonk, NY) were used for statistical analysis.

Discussion
In the vast majority of patients with advSM, the KIT D816V mutation is not only present in the mastcell lineage but also in multiple hematopoietic lineages (including the AHN compartment) [30][31][32]. The KIT D816V mutation can also be identified in CFU-GM colonies generated from myeloid progenitors [31] and recent data have highlighted the usefulness of these colonies for obtaining a more thorough insight into the clonal architecture of SM and other multimutated myeloid neoplasms [33][34][35][36][37][38][39].
In addition to improvement of C-findings, the assessment of responses is based on the relative reduction of mast cell burden, e.g., mast cell infiltration in bone marrow and serum tryptase [20,40]. However, this approach may not be sufficient to assess response in the non-mast cell (AHN) compartment of SM-AHN. In this respect, recent data have highlighted the importance and potential superiority of changes of the KIT D816V EAB as it represents in fact both compartments [41]. We therefore sought to assess the inhibitory effects of midostaurin and avapritinib on primary myeloid progenitor cells derived from KIT D816V positive advSM patients.
After two weeks incubation with midostaurin and avapritinib in vitro, the relative reduction of KIT D816V colonies was superior on avapritinib, including number of patients and depth of response. Of interest, three midostaurin non-responders had a significant response to avapritinib, while four avapritinib non-responders showed neither a response on midostaurin. These four patients were characterized by a relatively low mast cell burden with regard to mast cell infiltration in bone marrow histology and  EAB expressed allele burden, PB peripheral blood a Response defined as reduction of the KIT D816V EAB in PB ≥ 25% after six months [20] b Response defined as reduction of KIT D816V positive colonies ≥ 50% after two weeks in vitro c In three cases, in vivo data was used for statistical analysis because in vitro data was not available In three cases in vivo data was used for statistical analysis because in vitro data was not available serum tryptase level but a very high KIT D816V EAB (representing disease burden of both SM and AHN) and a poor-prognostic molecular risk profile with ≥ 2 mutations in the S/A/R gene panel. This data indicates that the KIT D816V EAB as marker for overall disease burden and the presence of additional somatic mutations in the S/A/R gene panel may be more important for prediction of response and resistance as the pure mast cell burden (Tables 1 and 3, Fig. 2a, b). The efficacy and safety of the highly selective KIT -inhibitor avapritinib in patients with advSM is currently being evaluated in an open-label, single-arm phase 2 study (NCT03580655). In an initial dose-escalation phase 1 study (NCT02561988), avapritinib demonstrated an ORR of 83% per IWG-MRT & ECNM consensus criteria in 29 evaluable patients. Consistent with our in vitro data, a therapeutic benefit of avapritinib was also observed in several patients with primary or secondary resistance on midostaurin [21,22,24,42].
On midostaurin, the relative reduction of KIT D816V positive colonies after two weeks incubation in vitro was fully paralleled by the relative reduction of KIT D816V positive colonies after 6 months therapeutic treatment (Fig. 3) and by the pattern of clinical response and resistance ( Table 3). The in vitro responses were strongly associated with absence of mutations in the S/A/R gene panel (p < 0.033) and reduction of the KIT D816V EAB ≥ 25% at month six (p < 0.003), parameters which were recently reported to be most predictive for response to treatment and favorable outcome (Tables 4a, b) [41]. This data therefore proves the hypothesis that midostaurin is not only able to target the mast cell compartment but also the KIT D816V positive AHN.
Disparate mechanisms may confer to resistance to midostaurin and avapritinib. We recently revealed the negative impact of mutations in the S/A/R gene panel on phenotype, response rates, resistance, early or late progression and consequently survival in midostaurin-treated patients suggesting primary resistance and/or outgrowth of a multimutated and clinically aggressive KIT D816V positive clone [9,15,41]. We now could also demonstrate that neither midostaurin nor avapritinib had an effect on the multimutated KIT D816V negative compartment, which may lead to KIT independent resistance and progression, e.g., secondary KIT D816 negative acute myeloid leukemia [43]. Other potential mechanisms of resistance to midostaurin and avapritinib may be unveiled in ongoing and upcoming clinical trials.
In conclusion, midostaurin is not only able to target the mast cell compartment but also the KIT D816V positive AHN while it may not overcome the adverse effect of high molecular risk mutations (S/A/R gene panel). The in vitro inhibition assay could be considered as a prognostic tool to predict the in vivo response to midostaurin (and potentially also to avapritinib) in patients with advSM. The highly selective KIT -inhibitor avapritinib has significant in vitro activity against KIT D816V, even in midostaurin non-responders. It will therefore be most interesting to extend this exploratory analysis to a larger cohort of midostaurin-treated patients but also to avapritinib-treated patients with or without prior midostaurin treatment. This assay may then help to determine the choice and sequence of available treatment options, e.g., in terms of the potential sequential use of KIT-inhibitors and alternative treatment options in non-responders including (intensive) chemotherapy and potentially early allogeneic stem cell transplantation [4,5,20,44].