Efficacy and pharmacokinetic/pharmacodynamic evaluation of the Aurora kinase A inhibitor MLN8237 against preclinical models of pediatric cancer

Abstract

Purpose: To gain a greater understanding of the potential of the Aurora kinase A inhibitor MLN8237 in the treatment of pediatric malignancies.

Methods: The activity of MLN8237 was evaluated against 28 neuroblastoma and Ewing sarcoma cell lines, and its in vivo efficacy was studied over a range of doses against 12 pediatric tumor xenograft models. Pharmacokinetic, pharmacodynamic, and genomic studies were undertaken.

Results: In vitro neuroblastoma cell lines were generally more sensitive to MLN8237 than Ewing sarcoma lines. MLN8237 demonstrated significant activity in vivo against solid tumor models at the maximum tolerated dose (MTD); however, only 2 of 6 neuroblastoma models had objective responses at 0.25MTD. In contrast, MLN8237 induced objective responses at its MTD and at 0.5MTD in three ALL models and in two out of three at 0.25MTD. Pharmacokinetic studies at 0.5MTD demonstrated a T (max) of 0.5 h, C (max) of 24.8 μM, AUC((0-24)) of 60.3 μM h, and 12 h trough level of 1.2 μM. Mitotic indices increased 6-12 h after MLN8237 administration. AURKA copy number variation was frequent in xenografts, and expression was highly correlated with copy number.

Conclusions: Objective responses were more frequent in tumors with decreased AURKA copy number (5/8) compared to those with increased gene copy number (2/14). This report confirms the significant activity against both solid tumor and ALL xenografts at the MTD, with a steep dose response. These data support clinical development of MLN8237 in childhood cancer. Because of the steep dose-response relationship, such studies should target achieving trough levels of 1 μM or higher for sustained periods of treatment.

Fig. 1

MLN8237 in vivo activity against individual solid tumor xenografts (KT-10; NB-1643 xenograft panel a and b respectively) or ALL xenografts (ALL-2; ALL-8, panel c and d respectively). Results show growth of individual tumors in control, or mice treated with 2.6, 5.2, 10.4, or 20.8 mg/kg twice daily, 5-days per week for 6 weeks for solid tumors or 3 weeks for ALL

Fig. 2

Pharmacokinetic and pharmacodynamic activity of MLN8237. a MLN8237 (10.4 mg/kg (filled circle) or 20.8 mg/kg (open square)) was dosed orally in non-tumored scid mice, and blood was isolated at various times thereafter. MLN8237 concentrations were determined in plasma from 3 different animals per time point; means ± standard error of the means are shown; b Representative immunofluorescence images of tumor sections from NB-1771 xenografts stained with antibodies against MPM2 and pHistoH3 12 h after in vivo administration of vehicle control (upper panel) or MLN8237 (20.8 mg/kg, lower panel); c The percentage of cells positive for the mitotic markers MPM2 (dark bars) or pHistH3 (white bars) were determined from 3 different animals at multiple time points; means ± standard deviation are shown. Mice bearing the human neuroblastoma tumor NB-1771 were dosed once orally with MLN8237 at 20.8 mg/kg

Fig. 3

Gene expression, copy number analysis of the Aurora kinase genes, and drug sensitivity of the PPTP in vivo models. a Relative gene expression of Aurora kinases A, B, and C as determined by Affymetrix gene expression arrays; b Tumor sensitivity to MLN8237 administered at the MTD (data from ref [23]) presented as a categorical heat map. The colored heat map depicts group response scores: MCR (red), CR (orange), PR (yellow), SD (gray), PD2 (light green), PD1 (dark green), Not evaluated (black) (see Supplemental Fig. 2, and Median Group Response scoring in the Supplemental Response Definitions section); c Copy number assessment of Aurora kinase A (AURKA) from the Affymetrix SNP 6.0 array. The upper panel shows a continuous heat map representation of copy number log2 ratio, while the lower panel shows a categorical representation of copy gain (red), copy loss (blue), copy diploid (gray), or no data (white) (color figure online)

Fig. 4

Copy number analysis using the Affymetrix SNP 6.0 array. Copy number representation of the in vivo tested panel according to log2 ratio of segments identified showing copy number status across the Aurora kinase A locus. The location of the Aurora kinase A locus on chromosome 20 is indicated by a red bar across the top panel, and green vertical lines indicate the boundaries of the AURKA locus (color figure online)