In vitro activity: AG14361 is at least 1000-fold more potent than the benzamides. The IC50 for AG14361 is 29 nM in permeabilized SW620 cells and 14 nM in intact SW620 cells. Crystallographic analysis of AG14361 bound to the catalytic domain of chicken PARP-1 shows that the tricyclic ring system of AG14361 is located in a pocket composed of amino acid residues Trp861, His862, Gly863, Tyr896, Phe897, Ala898, Lys903, Ser904, Tyr907, and Glu988. AG14361 forms important hydrogen bonds with Ser904 and Gly863 and a water-mediated hydrogen bond with Glu988. AG14361-induced growth inhibition is not attributed to PARP-1-related effects because maximal PARP-1 inhibition is observed at much lower concentrations (≤1 μM) than the GI50. AG14361 at 0.4 μM does not affect cancer cell gene expression or growth, but it increases the antiproliferative activity of temozolomide and topotecan, and inhibits recovery from potentially lethal γ-radiation damage in LoVo cells by 73%. In addition, 0.4 μM AG14361 does not substantially alter gene expression as shown by microarray analysis. A 17-hour exposure of A549 cells to 0.4 μM AG14361 does not change the expression of the 6800 genes. Thus, although 0.4 μM AG14361 inhibits cellular PARP-1 activity by more than 85%, it essentially does not change gene expression and cell proliferation, indicating that the cellular effects of this low concentration of AG14361 are specific for PARP-1 inhibition. Higher, growth-inhibitory concentrations of AG14361 affects gene expression, but these effects are not likely to be related to PARP-1 inhibition because cell proliferation is affected equally in PARP-/- and PARP-1+/+ cells. AG14361 is rapidly absorbed into the bloodstream and distributed to the tumor and liver with lower concentrations detected in the brain. Tissue-to-plasma concentration ratio indicates that AG14361 is retained in tumor tissue over time in both xenograft models, with tumor concentrations (≥15 μM for 2 hours) in excess of that required to inhibit PARP-1 activity in vitro.

Kinase Assay: The activity of full-length recombinant human PARP-1 is measured in a reaction mixture containing 20 nM PARP-1, 500 μM NAD+ plus [32P]NAD+ (0.1–0.3 μCi per reaction mixture), and activated calf thymus DNA (10 μg/mL) at 25oC; the reaction is terminated after 4 minutes by adding ice-cold 10% (wt/vol) trichloroacetic acid. The reaction product [32P]ADP-ribose incorporated into acid-insoluble material is deposited onto Whatman GF/C glass fiber filters with a Bio-Dot microfiltration apparatus and quantified with a PhosphorImager. Inhibition of PARP-1 activity by AG14361 at 0–600 nM is measured, and the Ki for AG14361 is calculated by nonlinear regression analysis.

Cell Assay: LoVo and SW620 colorectal cancer cells and A549 non–small-cell lung carcinoma cells are maintained in RPMI-1640 medium containing 10% fetal calf serum. Cell growth inhibition is estimated in exponentially growing LoVo, A549, and SW620 cells in 96-well plates. Cells are exposed to AG14361 (0–20 μM) alone or in the presence of 400 μM temozolomide. After 5 days of culture, these cells are fixed with 10% trichloroacetic acid and stained with sulforhodamine B. The concentration of temozolomide, topotecan, and AG14361 alone or in combination that inhibits growth by 50% (GI50) is calculated from computer-generated curves. Recovery from potentially lethal damage is measured in confluent LoVo cell cultures arrested in G1 phase to mimic the radiation-resistant quiescent cell population in tumors. Such cells are exposed to 8 Gy of γ-irradiation and then harvested and plated for colony formation assay immediately or maintained as growth-arrested confluent cultures for a 4-hour or 24-hour recovery period before harvesting and plating for the colony formation assay. Where indicated, 0.4 μM AG14361 is added 30 minutes before irradiation and is present in the recovery incubation.

J Natl Cancer Inst. 2004 Jan 7;96(1):56-67; Clin Cancer Res. 2004 Feb 1;10(3):881-9; Clin Cancer Res. 2005 Dec 1;11(23):8449-57.