(P 203) PRECLINICAL EVALUATION AMG 650, A SMALL MOLECULE INHIBITOR OF KIF18A, IN PEDIATRIC SOLID TUMORS BY THE PEDIATRIC PRECLINICAL IN VIVO TESTING (PIVOT) PROGRAM

Objective: KIF18A is a mitotic kinesin that plays a role in maintaining bipolar spindle integrity during mitosis. KIF18A knockdown selectively inhibits growth and induces apoptosis for cancer cell lines with chromosomal instability (CIN). AMG 650 is a selective inhibitor of KIF18A that phenocopies the effects observed for KIF18A knockdown. Osteosarcoma is characterized by a chaotic genome with chromothripsis and a large number of chromosomal structural variants. Anaplastic Wilms tumor is associated with TP53 mutation as are some Ewing sarcoma cases, and alveolar rhabdomyosarcoma often exhibits tetraploidy. In this study the pre-clinical efficacy of AMG 650 was tested using PDX models for these cancers.

Methods: AMG 650 was tested in 7 osteosarcoma, 2 Wilms tumor, 3 rhabdoid tumor, 15 rhabdomyosarcoma, and 7 Ewing sarcoma PDX models. A conventional testing design was used for osteosarcoma models with ten AMG 650-treated and ten vehicle control animals for each model. A single mouse trial design was used for the other cancer types, with a single vehicle control and AMG 650-treated mouse for each model. AMG 650 was administered orally at the dose of 100 mg/kg for 21 days. Mice were monitored for 12 weeks. Objective response (partial and complete) was assessed based on serial tumor volume measurements, and for osteosarcoma models median time to event (event defined as 4-fold increase in tumor volume) was evaluated using previously described PIVOT program testing methods. Plasma was collected for pharmacokinetic studies at different time points following the final administered dose (2h, 4h, 6h, 12h, and 24h; n=2 per time point).

Results: For 6 evaluable osteosarcoma models, no objective responses were observed, and a single model showed > 2-fold increase in median time to event. Among 15 rhabdomyosarcoma models, one achieved a complete response with tumor regrowth beginning approximately 3 weeks after treatment completion. Among 8 Ewing sarcoma models studied, 2 achieved partial responses with tumor regrowth occurring on-treatment or within several weeks of treatment completion. No objective responses were observed for the 2 Wilms tumor models and 3 rhabdoid tumor models tested. Drug levels were in line with previous results and will be presented.

Conclusion: In osteosarcoma, Ewing sarcoma, and rhabdomyosarcoma models, antitumor activity of AMG 650 was variable, ranging from no objective responses to one model with a complete response. Identification of specific determinants of sensitivity and the evaluation of activity in combination will aid in potential use of AMG 650 in these pediatric tumor settings.