Antitumor Efficacy of the Novel KIT Inhibitor IDRX-42 (Formerly M4205) in Patient- and Cell Line–Derived Xenograft Models of Gastrointestinal Stromal Tumor (GIST)

This article is featured in Highlights of This Issue, p. 2739

Gastrointestinal stromal tumors (GIST) are the most common mesenchymal malignancies of the gastrointestinal tract (1, 2). The annual incidence of GIST in European countries is 12/1.000.000/year (1). Most GIST harbor gain-of-function mutations in KIT or platelet-derived growth factor alpha (PDGFRA) oncogenes, encoding for the respective receptor tyrosine kinases. Activation of the receptors stimulates downstream signaling pathways, which in turn promote tumor proliferation and survival of the tumor cells (3–6).

Localized GIST is treated with surgery, while locally advanced, inoperable, or metastatic GIST requires systemic therapy (2). Imatinib (Glivec, Novartis), a tyrosine kinase inhibitor (TKI), is very active in the majority of GIST cases and associated with a high rate of objective responses. It prolongs progression-free survival and increases overall survival in patients with advanced or metastatic GIST (7–9). Treatment of inoperable GIST with imatinib is considered palliative, as most patients will develop disease progression over time, usually caused by acquired secondary resistance mutations in KIT or PDGFRA (10–13). Between the first regulatory approval of imatinib for the treatment of GIST and today, a number of alternative TKI have been introduced and are currently used as consecutive lines of treatment of advanced GIST after imatinib failure or in the rare case of imatinib intolerance. Sunitinib (Sutent, Pfizer) is the current standard of care in the second line. Regorafenib (Stivarga, Bayer) is used after failure of both imatinib and sunitinib, whereas ripretinib (Qinlock, Deciphera Pharmaceuticals) is routinely given after tumor progression after three or more lines of treatment (2, 14). Avapritinib (Ayvakit, Blueprint Medicines) is exclusively used in GIST with mutations in PDGFRA exon 18, preferably in cases with the imatinib-resistant mutation p.D842V (14). Even after treatment of advanced GIST with all approved oral agents, many patients still qualify for further lines of therapy, which usually involves the off-label use of alternative KIT- or PDGFRA inhibitors typically used for other indications, illustrating the high unmet medical need for further treatment options in this setting.

New drugs for the treatment of GIST can be tested in preclinical models of this rare malignancy, but the number of reliable tools for GIST research is still limited. Since 2004, our Laboratory of Experimental Oncology has developed a number of very well-characterized, patient-derived GIST models growing in immunodeficient mice, which are ready for experimental use. In the current study, we explored the in vivo efficacy of IDRX-42, formerly M4205 (Merck KGaA), an oral, selective KIT inhibitor with broad activity against mutations in KIT exons 11, 13, and 17 (15, 16), in patient-derived xenografts (PDX) with diverse KIT mutations. We benchmarked this novel kinase inhibitor against established agents for the treatment of GIST. IDRX-42 was designed as a potent inhibitor of clinically relevant mutations of KIT. The compound has an excellent kinome selectivity profile sparing critical off-targets such as fms-like tyrosine kinase 3 (FLT3) and vascular endothelial growth factor receptor (VEGFR; refs. 15, 16). Reduction of cellular KIT signaling and tumor cell viability was demonstrated in GIST cell lines expressing oncogenic driver and resistance mutations of KIT upon treatment with IDRX-42. Dose-dependent inhibition of KIT activation and tumor growth was confirmed in GIST in vivo models driven by disease-relevant mutations in KIT (15, 16).

Imatinib, sunitinib, avapritinib, and the experimental compound IDRX-42 were provided by Merck KGaA. Imatinib was dissolved in sterile water, sunitinib in 0.5% carboxymethyl cellulose (Colorcon Inc.)/0.25% Tween 20 (VWR) in 100 mmol/L Na-citric buffer (pH 4.5). Avapritinib was dissolved in 0.5% carboxymethyl cellulose/0.25% Tween 20 in sterile water (pH 4.0), and IDRX-42 was dissolved in 0.1% carboxymethyl cellulose/0.25% Tween 20 in 50 mmol/L Na-citric buffer (pH 3.0). The IDRX-42 vehicle was used as the nonactive treatment of control mice. The resulting suspensions were kept at room temperature and protected from light for a maximum of 7 days.

In the first two experiments, in UZLX-GIST9^KIT11+17 and UZLX-GIST25^KIT13, IDRX-42 was administered at two different doses: 10 mg/kg (IDRX-42_10) and 25 mg/kg (IDRX-42_25). In the remaining two experiments, in GIST882^KIT13 and UZLX-GIST2B^KIT9, only the higher dose of IDRX-42 was tested based on the results of the first two experiments where a dose-dependent effect was seen.

For the in vivo evaluation of IDRX-42, four GIST xenograft models were selected, based on their molecular profile and known TKI sensitivity. We used three PDX models: UZLX-GIST9^KIT11+17 (with KIT: p.P577del;W557LfsX5;D820G mutations), UZLX-GIST25^KIT13 (KIT: p.K642E), and UZLX-GIST2B^KIT9 (KIT: p.A502_Y503dup), and one cell line–derived model GIST882^KIT13 (KIT: p.K642E; cell line courtesy of Jonathan Fletcher, Brigham and Women's Hospital, Boston, MA; ref. 17). The PDX models have been established at the Laboratory of Experimental Oncology, KU Leuven (Leuven, Belgium) from material collected from consenting GIST patients diagnosed and treated at the University Hospitals Leuven (UZ Leuven, Leuven, Belgium). All models are well characterized and are known to retain stable histologic and molecular features, similar to those observed in the original donor tumor (18). The model characteristics and experimental setup are presented in Table 1 and Supplementary Table S1. Initial xenografting of human tumors from GIST patients was performed after written informed consent from the patient and approved by the Ethics Committee UZ/KU Leuven (Leuven, Belgium; project number S53483) and performed to the ethical guidelines set forth in the Declaration of Helsinki. The animal experiments were approved by the Ethics Committee for Animal Research, KU Leuven (projects P175-2015 and P196-2020) and performed according to local guidelines and Belgian/European Union regulations.

A total of 146 adult female NMRI (nu/nu) mice (Janvier Laboratories) were engrafted unilaterally under ketamine (Dechra)/xylazine [Inovet; V.M.D. nv/sa) anesthesia. Animals were randomized into predefined treatment groups once the average tumor volume was in the range of 150–300 mm³, with 9–11 mice in each group (Table 1). Randomization was done by arranging all tumors by their volumes and assigning the corresponding mouse to control group (A), reference group (B), and IDRX-42 group (C) in the following order: A, B, C, C, B, A, A, B, C, starting from the lowest value to the highest. The tumor volume was measured by a digital caliper three-dimensionally, and the tumor volume was calculated as V (mm³) = x × y × z. The treatment was given by daily oral gavage, which lasted for 15–29 days depending on the tumor growth of the control groups (Table 1). During treatment, the tumor volume was measured three times per week, whereas the body weight and general well-being of the animals were followed daily. At the end of the experiment, when control tumors had reached on average at least 200%–350% of the starting volume and/or a statistically significant difference was seen between the vehicle and the IDRX-42_25 group, mice were sacrificed by cervical dislocation after an intraperitoneal overdose of pentobarbital (Vetoquinol): half of the animals were sacrificed 6 hours after the last dose of the TKI, whereas the remaining ones were euthanized 24 hours after last dose. All tumors were collected, partly snap-frozen in liquid nitrogen in Precellys tubes (Bertin Instruments), and partly fixed in 4% buffered formaldehyde for formalin-fixed, paraffin-embedded (FFPE) blocks and subsequent histopathologic and molecular assessment. Antitumor activity was assessed based on the evolution of tumor volume expressed as the percentage of the normalized baseline value. Differences in absolute tumor volume were compared between the last and the first day of the experiment. All tumors were included in the relative tumor volume evaluation graphs, but the tumors from mice dying prematurely were excluded from the statistical analysis of tumor volume.

Fixed tumors were embedded in paraffin, and 4-μm sections were cut for hematoxylin and eosin (H&E) and IHC stainings. H&E staining was performed to evaluate the general tumor morphology and the mitotic and apoptotic activity, which was evaluated by counting the number of mitotic figures and apoptotic cells in 10 high-power fields (HPF, 400-fold magnification, field of view diameter 450 μm), respectively. Furthermore, the histologic response (HR) was graded as previously described in GIST by assessing the magnitude of necrosis, myxoid degeneration and/or fibrosis with the following grading system: grade 1 (0%–10% responding tumor area), grade 2 (10%–50%), grade 3 (50%–90%), and grade 4 (>90%) (19). Additionally, IHC was performed for KIT and Discovered on GIST 1 (DOG-1) staining to confirm the GIST diagnosis of the PDX models. IHC was also done for phospho-histone H3 (pHH3) and Ki-67 as markers of proliferation and for cleaved poly-ADP-ribose-polymerase (cPARP) as a marker of apoptosis. Mitosis and apoptosis were assessed by counting the average number of immunohistochemically positive pHH3 and cPARP cells, respectively, in 10 HPF at 400-fold magnification. The Ki67-labeling index was calculated as the average percentage of Ki67-positive cells in 5 digital images taken at 400-fold magnification, as described previously (20). The following primary antibodies were used: KIT (A450229-2, Agilent), DOG-1 (clone K9, DOG-1-L-CE, Leica Biosystems), pHH3 (9701L, Cell Signaling Technologies), Ki67 (clone SP6, MA5–14520, Thermo Fisher Scientific), and cPARP (clone 5E1, ab32064, Abcam). Antigen–antibody complexes were visualized using diaminobenzidine (Agilent), incubated for 10 minutes, and slides were counterstained with Gill's hematoxylin (VWR). Stained tissue sections were analyzed using an Olympus BX43 microscope (Olympus). Representative pictures were captured using the Olympus UC30 digital camera (Olympus) and analyzed with Olympus cellSens Dimension imaging software (Olympus). Furthermore, all control- and IDRX-42-treated tumors were evaluated for the presence of KIT mutations (in exons 9, 11, 13, and 17) using bidirectional Sanger sequencing as previously described (11).

Tumors from mice that were found dead during the experiment could not be collected and were therefore excluded from histologic and molecular analyses. Tumors from mice that were sacrificed before the final day of the experiment were included in the histologic assessment and corresponding statistical analysis. Priority was given to collection of snap-frozen tumor material for phosphorylation pathway analysis. Therefore, from some tumors, there was no material for FFPE blocks and, as a consequence, these tumors were also excluded from histologic analysis.

Levels of total and phosphorylated KIT (Y703) were detected in tumor samples at the end of the efficacy studies. About 50–100 mg of frozen tumor tissue was lysed with a Precellys 24 homogenizer (Bertin Instruments). Total protein concentrations were determined using the BCA Protein Assay Kit (#23225, Thermo Fisher Scientific). Per sample, 50 μg of lysate was analyzed in technical triplicates in a Luminex-based assay using a mouse-total KIT antibody (#ab111033, Abcam) coupled to Luminex microsphere beads for capturing, and a respective rabbit total KIT (#3074, Cell Signaling Technologies) or P-Y703-KIT (#3073, Cell Signaling Technologies) antibody as well as a PE-labeled donkey anti-rabbit antibody (#711 116 152, Dianova) for detection. Levels of total and phosphorylated ERK1/2 (T202/Y204 and T185/Y187) were determined using the MSD MULTI-SPOT Biomarker Detection Base Kit Phospho (T202/Y204; T185/Y187)/Total ERK1/2 according to the manufacturer's instruction (#15107A-3, Meso Scale Discovery). Levels of phosphorylated AKT (pAKT S473) were detected in a Luminex-based assay using a mouse total AKT antibody (#05-591, EMD Millipore) coupled to Luminex microsphere beads for capturing and a P-S473-AKT antibody (#4058, Cell Signaling Technologies) as well as a PE-labeled donkey anti-rabbit antibody (#711-116-152, Dianova). Counts for phosphorylated proteins were normalized to counts for respective total proteins for each technical sample. The average of the normalized phosphor-protein levels from the vehicle-treated probes was set as 100% control, and counts of drug-treated samples were calculated as a percentage of control.

GIST882p.6 was cultured in DMEM/F12 supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin (all Thermo Fisher Scientific) until 70% confluent. Cells were seeded in 96-well plates at a cellular density of 30,000 cells per well. After 24 hours, cells were treated with imatinib or IDRX-42 in 5-fold dilutions, starting from 5 μmol/L. There were three independent experiments, with two technical replicates, including 0.3% DMSO as a negative control. Cells were exposed to drugs for 45′, afterward the medium was removed and cells were kept at −20°C until the phosphorylation pathway analysis, which was performed as described above.

The Wilcoxon matched-pairs test (WMP) was used to compare relative tumor volumes within treatment groups between the first day and last day of the experiment. Kruskal–Wallis test (KWt) with Dunn multiple comparisons test as post hoc test was used to compare relative tumor volumes on the last day of the experiment and histologic analysis between different treatment groups. The difference in phosphorylation pathway analysis one-way ANOVA and Dunnett posttest with a 95% confidence interval was applied, testing drug-treated groups against vehicle controls. GraphPad Prism 8 (GraphPad Software Inc.) was used for statistical analysis with P < 0.05 considered significant.

The data generated in this study are available upon request from the corresponding author.

The same histopathologic characteristics were seen in the original donor tumors or an early passage of the PDX model and the corresponding PDX models (Supplementary Fig. S1). All models showed spindle cell morphology, with an epithelioid component present in UZLX-GIST9^KIT11+17. Models UZLX-GIST9^KIT11+17 and UZLX-GIST25^KIT13 showed some degree of spontaneous necrosis. All models were KIT and DOG-1 positive, with confirmed presence of specific KIT mutations.

In UZLX-GIST9^KIT11+17, UZLX-GIST25^KIT13, and UZLX-GIST2B^KIT9, a steady increase in tumor volume in the control groups was observed (Fig. 1A, C, and E; Supplementary Fig. S2). In the imatinib-resistant model UZLX-GIST9^KIT11+17 and in the imatinib dose-dependent sensitive model UZLX-GIST2B^KIT9, the average relative tumor volume in the imatinib group reached 300.5% on day 15 and 160.6% on day 29, respectively. In UZLX-GIST25^KIT13, the average relative tumor volume in the sunitinib group reached 153.4% on day 20. This model is considered sunitinib-sensitive (Supplementary Table S1) based on previous, not published experiments; however, in this experiment, the administered sunitinib dose was lower than in our previous studies (20 mg/kg 5-day ON 2-day OFF in this study vs. 40 mg/kg/qd in the previous study), however, with the selected dose used in this study matching the human exposure.

In UZLX-GIST25^KIT13, IDRX-42_25 showed superior antitumor activity to sunitinib (P < 0.05, KWt). In the imatinib dose-dependent sensitive models GIST882^KIT13 and UZLX-GIST2B^KIT9, IDRX-42_25 showed superior antitumor activity as compared with imatinib at doses administered. There was a similar effect on tumor volumes of avapritinib compared with IDRX-42_25 in UZLX-GIST9^KIT11+17 (P > 0.9, KWt), and additionally in this model, no significant difference could be shown between relative tumor volumes of the imatinib- and IDRX-42_25-treated groups on day 14 of the experiment (P = 0.05, KWt).

In the first two experiments, two different doses of IDRX-42 (10 mg/kg/qd and 25 mg/kg/qd) were tested, and we observed a dose-dependent inhibitory effect on tumor volume. IDRX-42_25 showed superior antitumor activity as compared with the lower dose of the drug (P < 0.005, KWt, in UZLX-GIST9^KIT11+17 and P < 0.05, KWt, in UZLX-GIST25^KIT13), with statistically significant tumor growth compared with baseline for tumors treated with the lower dose (P < 0.05, WMP, in UZLX-GIST9^KIT11+17 and P < 0.005, WMP, in UZLX-GIST25^KIT13; Fig. 1A–C; Supplementary Fig. S2). Therefore, the lower dose (10 mg/kg/qd) was not explored further in the subsequent experiments.

IDRX-42_25 caused statistically significant tumor volume shrinkage (P < 0.05, WMP) in models UZLX-GIST25^KIT13, GIST882^KIT13, and UZLX-GIST2B^KIT9 with a relative decrease to 45.6%, 57.3%, and 35.1%, respectively on the last day of treatment as compared with baseline, and tumor growth delay (160.9%) compared with control in UZLX-GIST9^KIT11+17 (Fig. 1A–E; Supplementary Fig. S2). In UZLX-GIST9^KIT11+17, UZLX-GIST25^KIT13, and UZLX-GIST2B^KIT9, there was a statistically significant reduction of tumor volume in IDRX-42_25-treated tumors compared with control (P ≤ 0.0005, KWt; Fig. 1A, C, and E).

Overall, no significant body weight changes nor treatment-related toxicity were observed in any of the treated groups in any of the experiments. On average, the behavior of the mice did not change throughout the experiment. During the treatment period, some mice were found dead (n = 9) or were sacrificed (n = 8) before the final day of the experiments, but these findings were considered unrelated to drug toxicity because deaths occurred randomly among all treatment groups, and in most cases no prior symptoms were noticed (Supplementary Table S2). Two mice died because of aspiration after gavaging. From 9 mice that were found dead before the final day of the experiment and from one mouse that died because of aspiration no collection of the tumor could be done; therefore, they were not included in the histopathologic evaluation.

In the UZLX-GIST9^KIT11+17 and UZLX-GIST2B^KIT9 models, no HR grade 3–4 was seen in any group receiving active treatment, except for avapritinib in UZLX-GIST9^KIT11+17. Five of 7 avapritinib-treated tumors showed grade 2–3 HR and three out of 5 imatinib-treated tumors showed grade 2 HR in this model (Fig. 2A). At the same time, both KIT exon 13 mutated models (UZLX-GIST25^KIT13 and GIST882^KIT13) showed IDRX-42-treatment-related HR, characterized by diffuse myxoid degeneration, that is a characteristic morphologic change observed in GIST response to TKI treatment (21). In the UZLX-GIST25^KIT13 model, 33.3% of tumors treated with IDRX-42_25 showed grade 2 HR, 33.3% presented grade 3 HR, and in 33.3% grade 4 HR was observed, while in the GIST882^KIT13 model, 18.2% of tumors treated with IDRX-42_25 had grade 3 HR and 81.8% showed grade 4 HR (Fig. 2A). Representative images per treatment group of the experiment in GIST882^KIT13 are shown in Supplementary Fig. S3.

In all models, a significant reduction of mitotic activity on H&E was observed in the IDRX-42_25-treated tumors compared with untreated controls (Fig. 2B). This was confirmed on pHH3 stained slides, as well as for proliferative activity on Ki-67–stained slides for UZLX-GIST25^KIT13, GIST882^KIT13, and UZLX-GIST2B^KIT9 (Supplementary Fig. S4). Standard treatment showed variable effects on mitotic activity: A significant reduction of mitotic and proliferative activity compared with control was observed in UZLX-GIST25^KIT13 sunitinib-treated tumors, but only on pHH3 and Ki-67 stained slides, and in UZLX-GIST2B^KIT9 only on H&E for tumors treated with the higher dose of imatinib. A significant reduction of proliferative activity compared with control was seen in UZLX-GIST9^KIT11+17 on Ki-67 stained slides for avapritinib-treated tumors, but no significant changes were seen on H&E- or pHH3-stained slides. In GIST882^KIT13, a significantly greater decrease in mitotic activity as well as proliferative activity was observed in IDRX-42_25-treated tumors compared with imatinib. In UZLX-GIST25^KIT13, a significant decrease in mitotic activity as well as proliferative activity was seen in tumors treated with the higher dose of IDRX-42 compared with the lower dose.

In UZLX-GIST9^KIT11+17, both on H&E- and cPARP-stained slides, a significant reduction of apoptosis compared with control was observed in the IDRX-42_25-treated tumors. The decrease in apoptosis was also observed on H&E-stained slides of IDRX-42_25-treated tumors in GIST882^KIT13 and on cPARP-stained slides of IDRX-42_25-treated tumors in UZLX-GIST2B^KIT9 (Supplementary Figs. S5 and S6).

A number of 22 tumors were too small (range, 16–338 mm³) for additional tissue collection in order to perform histologic analysis (Supplementary Table S3). Two tumors (6%) in the experiment with GIST882^KIT13 showed grade 4 HR and were very small (40–85 mm³) which made analysis of 10 HPF for mitosis and apoptosis (H&E, pHH3, and cPARP) impossible.

Mutational analysis of ex-mouse tumors treated with IDRX-42 did not reveal any additional secondary mutations in KIT in comparison with the original donor samples and untreated controls.

The autophosphorylation of KIT was determined in available tumor samples 6 hours after the last dose at the study end to assess the treatment effect on KIT activity (Fig. 3A–C and E). In UZLX-GIST9^KIT11+17, KIT autophosphorylation at Y703 was strongly reduced in IDRX-42_25 and avapritinib-treated tumors, whereas there was only a weak effect in the IDRX-42_10 and imatinib groups. A pronounced inhibition of KIT autophosphorylation that led to a significant decrease of pAKT was observed in all actively treated groups of the UZLX-GIST25^KIT13 model. Moreover, a higher dose of IDRX-42 caused inhibition of pERK1/2 (Fig. 3D). Almost complete inhibition of KIT autophosphorylation was detected in IDRX-42_25-treated tumors in the UZLX-GIST2B^KIT9 model, whereas no effect was observed in the imatinib group. A significantly more pronounced inhibition of KIT autophosphorylation upon treatment with IDRX-42_25 compared with imatinib was shown in the cell line–derived GIST882^KIT13 model. Furthermore, an in vitro analysis performed with GIST882^KIT13 cells confirmed the superiority of IDRX-42 over imatinib on the Y703 level of phosphorylation (Fig. 3F).

In this study, we tested the novel compound IDRX-42 in GIST xenograft models characterized by different KIT mutations and varying sensitivity to standard TKI. We have tested the new compound against standard treatments. Response to those treatments in our xenograft models depended on the known sensitivity to these standard TKI: imatinib showed comparable response as in our previous studies (22–25) with tumor growth in all imatinib-treated tumors, except for GIST882^KIT13 where we saw tumor volume stabilization. Unlike in our previous work, sunitinib led to tumor volume stabilization instead of shrinkage in UZLX-GIST25^KIT13, and avapritinib caused tumor volume growth in UZLX-GIST9^KIT11+17 (22, 24). However, in the present study, lower doses of respected TKI were administered, which may be insufficient to obtain volumetric changes as described in previous experiments. Nonetheless, the selected doses used in this study are matching the human exposure. In particular sunitinib was administered as 20 mg/kg 5 days ON / 2 days OFF, which according to the healthcare business of Merck KGaA is a regimen achieving the same exposure and target coverage as clinical dosing of sunitinib in patients (i.e., 50 mg 4 weeks ON/2 weeks OFF).

As shown in previous in vitro experiments, IDRX-42 is a potent and specific KIT inhibitor with activity against GIST with mutations in KIT exons 11, 13, and 17 (15, 16). In this study, we confirmed the in vitro results in in vivo experiments with IDRX-42 showing significant and dose-dependent antitumor activity in three patient- and one cell line–derived xenograft model of GIST. The most pronounced effect of IDRX-42 was seen in KIT exon 13–mutated GIST with tumor shrinkage, decrease of proliferative effect, and inhibition of the KIT phosphorylation pathway. Furthermore, extensive myxoid degeneration was observed in both models carrying this mutation. Myxoid degeneration is the replacement of tumor cells by connective tissue mucous matrix, leading to reduced cellularity and metabolically inactive tissue. It is typically seen in GIST when responding to TKI, thus considered as a desired, therapeutic effect of such treatment (21). KIT exon 13 mutations are seen in the clinic in 1%–2% of GIST patients (26, 27) and can be primary (p.K642E substitution) or secondary gene alterations (e.g., p.V654A; ref. 12). There is no clear evidence in the literature to make a link between primary KIT exon 13 mutations and more aggressive clinical behavior, due to small sample size (26, 28).

We saw profound antitumor activity of IDRX-42 in UZLX-GIST2B^KIT9. KIT exon 9 mutations are the second most common KIT mutation in GIST, occurring in about 7% to 15% of cases (26, 27). They typically originate from the small bowel, are characterized by aggressive clinical behavior, and require a higher imatinib dosage to reach the optimal progression-free survival when compared with KIT exon 11 mutants (29). For this reason, UZLX-GIST2B^KIT9 mice in the imatinib group were treated with a higher dose of imatinib (100 mg/kg 5d b.i.d. 2d qd). We did not perform an in vivo assessment in GIST with the most common and often imatinib-sensitive KIT mutation in exon 11, because model selection was primarily based on imatinib resistance, because there is a higher need for potent second-line treatment in GIST in these genotypes (15, 16). It is known that secondary mutations occur during treatment with imatinib and can cause resistance to the first-line treatment. In our TKI-resistant UZLX-GIST9^KIT11+17 model with secondary KIT exon 17 mutation, IDRX-42 delayed the tumor growth compared with control tumors and had similar antitumor activity as avapritinib in vivo. This is in line with earlier in vivo efficacy experiments by Blum and colleagues, where dose-dependent antitumor activity of IDRX-42 was seen in xenograft models expressing secondary resistance mutations in KIT exon 17 with shrinkage of tumors treated with IDRX-42 (20 mg/kg; refs. 15, 16). Additionally, prior in vitro experiments showed that there is biochemical and cellular activity on KIT autophosphorylation in imatinib-resistant cell lines with KIT exon 11 and secondary exon 13 mutations with an IC₅₀ < 60 nmol/L in GIST430/654 (15, 16).

GIST with KIT primary exon 13 mutations are usually resistant or show dose-dependent sensitivity to imatinib. They tend to be sensitive to sunitinib (30). Although we did not test avapritinib nor ripretinib in our KIT exon 13–mutated models in these experiments, we know from previous work that avapritinib lacks activity against GIST with primary exon 13 mutations (31). Ripretinib, on the other hand, shows antitumor activity in cell lines of GIST with KIT exon 13 primary mutations, and is associated with a progression-free survival benefit in patients with this subtype (31, 32). Concerning KIT exon 9–mutated GIST, higher doses of imatinib are required to achieve the best progression-free survival in patients (29). These GIST are, however, quite sensitive to sunitinib and ripretinib (32). GIST with secondary mutations in exons 13 and 14 respond well to sunitinib, whereas regorafenib is effective against secondary mutated GIST in exons 17 and 18 (30, 33, 34). Almost all secondary mutated GIST respond well to ripretinib (33).

It is known that TKI-resistant GIST usually consists of different subclones resulting in inter- and intralesional heterogeneity in terms of TKI resistance and molecular profile (35). Ideally, one drug would target a relevant range of these clones and therefore attack the tumor as a whole, but as described, most of the existing TKI have a limited spectrum of activity against the different mutations, and thus resistant cells outgrow the targeted subclones during treatment (30). To tackle this problem, alternating TKI with complementary activity has been tested in vitro and in a phase I trial by Serrano and colleagues in order to minimize the overlapping toxic effects between some of the TKI (30). However, the approach showed difficulties in reaching correct plasma drug concentrations in order to suppress all imatinib-resistant clones and had a similar median progression-free survival to the imatinib rechallenge approach (30). Simultaneous combinations of TKI with different kinome specificity and acceptable safety profiles are currently also being explored, as well as combination therapy of TKI together with an inhibitor of KIT downstream pathways, so far without convincing results. These approaches have not yet reached the clinic (36) but are subject to ongoing clinical trials.

From our study and the prior in vitro study, we can conclude that IDRX-42 has activity against a broad spectrum of primary as well as secondary KIT mutations, seems to be superior to sunitinib, and has comparable activity to avapritinib. The main challenge concerning avapritinib in the clinic is still the cognitive side effects which seem to be limited for IDXR-42 because of a lower brain exposure in rodents compared with avapritinib (16). Additional (pre)clinical studies are required to compare IDRX-42 more precisely with standard-of-care treatments such as sunitinib, ripretinib, and avapritinib. Unfortunately, in vivo preclinical studies with ripretinib are challenging to perform due to the poor solubility of ripretinib. Therefore, a direct comparison with ripretinib was not performed in our experiments. In preclinical experiments, it was noted that the observed selectivity of IDRX-42 was significantly higher, compared with ripretinib profiled in the same panel (16). Based on its activity in the resistant xenograft model, IDRX-42 could be considered a promising future treatment of GIST after imatinib resistance occurs and should be explored further in this context. Additionally, the intentional combination of IDRX-42 with an inhibitor that shows activity in resistance mutations where IDRX-42 might (partially) lack efficacy could be a solution for patients with GIST with a complex mutational profile. Because IDRX-42 seems to have good antitumor activity in secondary KIT exon 17 mutations in vitro and in vivo, and partial activity in secondary KIT exon 13 mutations in vitro (15, 16), the combination with a drug that is directed against secondary KIT exon 13 mutations could be an ideal match for even better results in resistant disease.

In the meantime, IDRX-42 was licensed by IDRx, Inc., and a clinical phase I study of IDRX-42 is currently ongoing (NCT05489237) in patients with advanced (metastatic and/or surgically unresectable) GIST and will confirm the safety, tolerability, pharmacokinetics, and preliminary antitumor activity of this new compound. Hopefully, the early efficacy signal in this first-in-human trial will confirm the findings of our in vivo study with this very promising, novel KIT inhibitor.

Inhibition of KIT mainly induces apoptotic pathways (37). In our study, however, we could not find a proapoptotic effect of IDRX-42 in any of the models despite the shown inhibition of the KIT phosphorylation pathway in the in vivo tumors. It could be that we have missed the time point with the highest apoptotic activity. It was previously shown that apoptotic response is most prominent at 72 hours after imatinib dosing or other TKI treatment in the preclinical setting (37–39) or in the clinical setting on surgical biopsies after 7 days of preoperative imatinib treatment (40). We sacrificed all mice in our experiments at a later stage, between days 15 and 29. We collected material from three IDRX-42-treated tumors where the mice were sacrificed 1 week earlier than the end of their respective experiments (two IDRX-42_10-treated tumors on days 11 and 14 of the experiment because of aspiration after gavaging, and one IDRX-42_25-treated tumor on day 11 of the experiment was sacrificed because of weight loss). In these tumors, no increased apoptosis was seen compared with the other analyzed tumors. Furthermore, in aggressive models such as UZLX-GIST9^KIT11+17, the spontaneous necrosis in large zones of the control tumors made comparison of both HR and apoptosis to IDRX-42–treated tumors quite difficult. Compared with control tumors, we even saw a significant decrease in the number of apoptotic cells in IDRX-42_25–treated tumors in this model, as well as in the models GIST882^KIT13 (H&E) and UZLX-GIST2B^KIT9 (cPARP). In the latter model, this decrease in apoptosis in TKI-treated tumors has been reported previously (24, 41, 42). This stimulated a discussion about imatinib being rather autophagy-than apoptosis-inducing, reported first by Miselli and colleagues after examination of surgical specimens and later further studied by Gupta and colleagues (43, 44). This could also be a reason for lower apoptotic counts because only apoptotic cells and cPARP-stained cells were counted for the evaluation in this study.

During the experiment with GIST882^KIT13, we encountered some unexpected findings concerning tumor volume changes. Tumors in the control group did not seem to grow as expected. During histologic evaluation after the end of the experiment, we saw that those tumors had an intrinsic developing problem: They did not seem to grow as expected, showed low-grade myxoid degeneration and necrosis in all treatment groups, including in control tumors, and as a result, these tumors were not as proliferative as they were expected. These findings, already present to some extent in the donor tumor at an interim passage, were confirmed by a reference sarcoma pathologist. Despite these observations, we still noted a statistically significant tumor volume decrease in IDRX-42–treated tumors compared with control tumors. Additionally, on histologic evaluation, we noted evident response effects with generalized myxoid degeneration in IDRX-42–treated tumors: grade 4 HR in 9 out of 11 tumors, and grade 3 in the other two tumors. Therefore, we would not consider this tumor growth problem as a limitation of the study.

IDRX-42 has strong antitumor activity in patient- and cell line–derived GIST xenograft models with evidence for antitumor activity in primary as well as in secondary KIT mutations. Knowing the high medical need for further treatment options in pretreated GIST patients, IDRX-42 should be further investigated and developed in the clinic. A phase I first-in-human study is currently ongoing (NCT05489237) in patients with advanced, metastatic and/or surgically unresectable GIST to translate these intriguing preclinical findings into the potential to prevent or overcome resistance to currently available kinase inhibitor drugs by using IDRX-42.

L. De Sutter reports grants from Merck Healthcare KGaA, Germany, during the conduct of the study; grants from ONA Therapeutics, Spain outside the submitted work. A. Wozniak reports grants from Merck Healthcare KGaA, Germany during the conduct of the study; grants from Adcendo, Denmark, PharmaMar, Spain, and ONA, Spain outside the submitted work. P. Schöffski reports grants and personal fees from Merck KGaA during the conduct of the study; personal fees from Deciphera, Ellipses, Blueprint, Transgene, Exelixis, Medpace, Boehringer Ingelheim, and SQZ Biotechnology; grants and personal fees from Adcendo, Pharmamar, and Genmab; grants from CoBioRes, Eisai, Sartar Therapeutics, and ONA Therapeutics outside the submitted work. No disclosures were reported by the other authors.

L. De Sutter: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. A. Wozniak: Conceptualization, resources, supervision, validation, investigation, methodology, project administration, writing–review and editing. J. Verreet: Investigation, writing–review and editing. U. Vanleeuw: Methodology, writing–review and editing. L. De Cock: Investigation, methodology, writing–review and editing. N. Linde: Conceptualization, resources, funding acquisition, methodology, writing–review and editing. C. Drechsler: Conceptualization, resources, funding acquisition, methodology, writing–review and editing. C. Esdar: Conceptualization, resources, funding acquisition, methodology, writing–review and editing. R. Sciot: Validation, methodology, writing–review and editing. P. Schöffski: Conceptualization, resources, supervision, validation, methodology, project administration, writing–review and editing.

The work described in this article was performed by the Laboratory of Experimental Oncology at KU Leuven in collaboration with the healthcare business of Merck KGaA, Darmstadt, Germany. Both parties partially funded the research.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.