Central nervous system penetration and enhancement of temozolomide activity in childhood medulloblastoma models by poly(ADP-ribose) polymerase inhibitor AG-014699

Abstract

Background: Temozolomide shows activity against medulloblastoma, the most common malignant paediatric brain tumour. Poly(ADP-ribose) polymerase (PARP) inhibitors enhance temozolomide activity in extracranial adult and paediatric human malignancies.

Methods: We assessed the effect of AG-014699, a clinically active PARP inhibitor, on temozolomide-induced growth inhibition in human medulloblastoma models. Pharmacokinetic, pharmacodynamic and toxicity assays were performed in tumour-bearing mice.

Results: Sensitivity to temozolomide in vitro was consistent with methylguanine methyltransferase (MGMT) and DNA mismatch repair (MMR) status; MGMT(+) MMR(+) D384Med cells (temozolomide GI(50)=220 μM), representative of most primary medulloblastomas, were sensitised fourfold by AG-014699; MGMT⁻ MMR(+) D425Med cells were hypersensitive (GI(50)=9 μM) and not sensitised by AG-014699, whereas MGMT(+) MMR⁻ temozolomide-resistant D283Med cells (GI₅₀=807 μM) were sensitised 20-fold. In xenograft models, co-administration of AG-014699 produced an increase in temozolomide-induced tumour growth delay in D384Med xenografts. Consistent with the in vitro data, temozolomide caused complete tumour regressions of D425Med xenografts, whereas D283Med xenografts were relatively resistant. AG-014699 was not toxic, accumulated and reduced PARP activity ≥75% in xenograft and brain tissues.

Conclusion: We show for the first time central nervous system penetration and inhibition of brain PARP activity by AG-014699. Taken together with our in vitro chemosensitisation and toxicity data, these findings support further evaluation of the clinical potential of AG-014699-temozolomide combinations in intra-cranial malignancies.

Figure 1

Protein expression levels of PARP-1 and proteins related to temozolomide sensitivity/resistance in medulloblastoma cell lines. Expression levels of α-tubulin (control) are also shown. Protein expression was determined for 50 μg of whole-cell lysate by western blot analysis and visualised by chemoluminescence. D425 cells have no detectable MGMT and D283 cells have no detectable MLH-1 and minimal PMS2 consistent with MGMT and mismatch repair deficiency, respectively, whereas D384 cells are proficient for both.

Figure 2

Chemosensitisation to TMZ by AG-014699 in vitro in medulloblastoma cell lines. (A–C) Growth inhibition of D425Med, D283Med and D384Med cells, determined by XTT assay following a 6-day exposure to TMZ either alone (solid triangles) or with 0.4 μM AG-014699 (open triangles). Data are shown normalised to 0.5% DMSO or 0.4 μM AG-014699 controls, respectively. A single representative replicate for TMZ±0.4 μM AG-014699 in D425Med (A), D283Med (B) and D384Med (C) cells (±95% confidence intervals (CI)) is shown.

Figure 3

AG-014699 pharmacokinetics and PARP-1 inhibition in mouse brain and medulloblastoma xenograft. Following a single (i) or the last of four daily doses (ii) of AG-014699 (1 mg kg⁻¹, i.p.), samples were taken and tissues harvested at the indicated times post-administration: (A) concentration of AG-014447 in plasma (solid triangle), brain (open circle) and D283Med xenograft homogenates (solid circle) from tumour-bearing mice. Corresponding PARP-1 activity in D283Med xenograft homogenates (B) and the brain (C) is shown. Concentrations of AG-014447 are given as the mean (±95% CI) from three tumour-bearing mice per time point; PARP activity measurements are given as individual values with the horizontal lines indicating the mean value.

Figure 4

Enhancement of TMZ efficacy by AG-014699 in in vivo models of medulloblastoma. Growth of D425Med (A), D283Med (B) or D384Med (C) tumour xenografts over a 100-day period, following daily treatment for 5 days with vehicle control alone (solid circles), AG-014699 alone (1 mg kg⁻¹, open circles), TMZ alone (68 mg kg⁻¹, solid triangles) or TMZ (68 mg kg⁻¹)+AG-014699 (1 mg kg⁻¹) (open triangles). Dashed horizontal line corresponds to RTV4. Data are median of five mice per group until first humane killing owing to tumour burden.