An Efficient Synthesis of PARP Inhibitors Containing a 4-Trifluoromethyl Substituted 3,6,7,7a-Tetrahydro-1 H -pyrrolo[3,4-d ]pyrimidine-2,5-dione Scaffold

Poly(ADP-ribose) polymerases (PARPs) are key enzymes in the DNA repair pathway. Inhibitors of these enzymes belong to a new type of anticancer drugs that selectively kill cancer cells by targeting the homologous recombination genetic defects. This study presents a new synthetic approach to PARP inhibitors containing a 4-trifluoromethyl substituted 3,6,7,7a-tet-rahydro-1 H -pyrrolo[3,4-d ]pyrimidine-2,5-dione scaffold. The method is based on a practical one-step cyclocondensation of 2-(2-oxo-1,2,3,4-tetrahydropyrimidin-4-yl)acetic acid derivatives via the Curtius rearrangement of the corresponding acyl azides formed in situ upon the treatment with diphenylphosphoryl azide. The resulting products have been found to possess a potent inhibitory effect on PARP-1 and PARP-2 isoforms of poly(ADP-ribose) polymerases. The structure–activity analysis has revealed that the N1-aryl substituent is crucial to the selectivity and high potency towards PARP-2, and that the p -fluorobenzyl group is the optimal group for the non-selective and potent PARP-1 and PARP-2 inhibition.


■ Introduction
Poly(ADP-ribose) polymerase (PARP) has evolved as a promising molecular target in the cancer-targeted chemotherapy due to its pivotal role in restoring the genomic integrity [1].The knowledge of its functions has led to the development of PARP inhibitors, which exert their anticancer activity by interrupting DNA repair mechanisms [2].Several compounds (olaparib, niraparib, talazoparib, and rucaparib) have already been marketed for treating advanced ovarian cancer and breast cancer patients [3].As a monotherapy, PARP inhibitors have been shown to selectively kill tumors harboring mutations or deletions in DNA repair genes related to homologous recombination pathways, such as BRCA-1 and BRCA-2.This phenomenon referred to as "synthetic lethality" has been successfully translated to clinical practice and now exemplifies a "personalized" approach to the cancer therapy [4].
A poly(ADP-ribose) polymerase family has 18 nuclear and cytoplasmic enzymes that cleave NAD+ to nicotinamide and ADP-ribose to form long and branched ADP-ribose polymers on target proteins, including topoisomerases, histones and PARP itself, and thus impact diverse cellular processes (replication, transcription, differentiation, gene regulation, protein degradation).In the family, PARP1 and PARP2 are the most attractive therapeutic targets for the anticancer drug development.In fact, the vast majority of PARP inhibitors developed to date demonstrate a lack of specificity for any given PARP isoform.This raises the challenging problem of developing small molecule PARP inhibitors with a high isoform selectivity as better-tolerated drugs [5].PARP-1 accounts for more than 90 % of the cellular PARP activity, thus representing a primarily targeted PARP isozyme for the therapeutic intervention.However, none of the currently FDA-approved PARP inhibitors selectively inhibits PARP-1.Moreover, olaparib (Figure ), the most clinically successful drug, demonstrates superior (5-fold) inhibition of PARP-2 [6].Compound NMS-P118 is one of the most PARP-1 selective clinical candidates (~150-fold selectivity over PARP-2) endowed with a pronounced anticancer activity in preclinical studies [7].Zhao et al. identified compound 11a with a nearly 40-fold PARP-2/ PARP-1 selectivity [8].
Recently, we have reported that the 4-trifluoromethyl-substituted 3,6,7,7a-tetrahydro-1H-pyrrolo [3,4-d]pyrimidine-2,5-dione core may serve as a scaffold for designing a novel class of PARP inhibitors, due primarily to the presence of the endocyclic amide fragment in the pyrrolidone ring, which resembles a key structural motif of the majority of PARP's nicotinamide site ligands [9].This suggestion was supported by a molecular docking study, which additionally revealed that the trifluoromethyl group in position 4 of the novel scaffold may significantly contribute to the binding with the PARP enzyme NAD+ binding site due to the C•••F orthogonal dipolar interaction with the backbone peptide fragment.As a result,

Synthesis
The method previously developed for the preparation of 4-bromophenyl substituted 3,6,7,7atetrahydro-1H-pyrrolo [3,4-d]pyrimidine-2,5-dione 1c relied on regioselective nitromethane addition to ethyl carboxylate 2c, chemoselective nitro group reduction in the adduct 3c and subsequent intramolecular cyclocondensation of the resulting amine intermediate 4c (Scheme 1, A) [10].The main drawbacks of this approach were the low thermal stability of nitro compound 3c, which easily underwent a retro nitro-Michael reaction to the starting reactants, and facile saturation of the C=C double bond in the 1H-pyrrolo [3,4d]pyrimidine-2,5-dione system under a number of reducing conditions tested for step b.It was found that reduction of the nitro group with zinc dust in acetic acid furnished amino ester hydrochloride 4c in an acceptable 60 % yield.Finally, compound 1c was obtained in 83 % yield after the treatment with aqueous NaOH at room temperature.This method is unsuitable for synthesizing N1-alkyl substituted derivatives due to an insufficient reactivity of N1-alkyl pyrimidones 2 in the nitromethane addition step.
With the aim of developing a practical general approach to the desired heterocyclic system 1, lacking the above-mentioned limitations, we used stable (2-oxo-1,2,3,4-tetrahydropyrimidin-4-yl)acetic acid derivatives 6a -j (Scheme 2) as key intermediates in this work.Latter compounds were obtained in high yields from readily available pyrimidin-2(1H)-ones 7a -j according to our reported procedure based on the regioselective decarboxylative malonic acid addition [11].As found out, acids 6 can be easily converted into target compounds 1 in a single step by the interaction with diphenylphosphoryl azide and triethylamine under heating in toluene.The reaction proceeds via the Curtius rearrangement of the corresponding

Biological activity
The results of studying the inhibitory activity for compounds 1a -i,k against PARP-1 and PARP-2 enzymes are presented in Table (represented as the residual enzyme activity after the incubation with the inhibitors in the concentration of 10 μM).The results of the radiometric PARP inhibitory assay revealed that the N1-aryl-substitution within the 3,6,7,7a-tetrahydro-1H-pyrrolo-[3,4-d]pyrimidine-2,5-dione system (compounds 1a,c -f) led to the complete PARP-2 inhibition in this concentration, while the activity of PARP-1 enzyme was in the range of 23.85 -50.21 %.The substitution by a N1-p-fluorobenzyl group gave rise to a potent and merely isoform non-selective inhibitor (compound 1b).When R is a N1-p-fluorophenethyl group, compound 1g exhibited a moderate selectivity and a good potency (up to 1.22 % and 17.48 % of the remaining activity of PARP-2 and PARP-1, respectively).The introduction of simple N1-alkyl groups, such as cyclobutyl or allyl substituents, in compounds 1h,i, or a piperazine amide fragment in compound 1k resulted in a weaker non-selective inhibition with 8.59 -35.2 % of the remaining enzyme activity.

■ Conclusions
A series of 3,6,7,7a-tetrahydro-1H-pyrrolo [3,4d]pyrimidine-2,5-diones has been synthesized using a novel practical approach, and their inhibitory activity against PARP-1 and PARP-2 isoforms has been evaluated.Among all the compounds tested, derivatives 1a -g have exhibited a potent enzyme inhibitory activity with the complete inhibition at a concentration of 10 μM, and 1a,d-f have demonstrated specificity to PARP-2 (up to 41.64 of the remaining PARP-1 activity at the same concentration).These findings can provide promising structural chemotypes for the development of novel PARP inhibitors with a high potency and specificity.

Chemistry, general experimental information
All chemicals were obtained from Enamine LTD and used without further purification.Solvents were purified according to the standard procedures.Melting points were uncorrected. 19F NMR, 1 H NMR and 13 C NMR spectra were recorded on Varian Mercury-400 ( 13 C: 101 MHz; 19   19 F: 188 MHz) spectrometers with TMS or CCl 3 F as an internal standard.Multiplets were assigned as s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet) and br.s (broad singlet).Mass spectrometric detection of samples was performed with an Infinity 1260 UHPLC system (Agilent Technologies, Waldbronn, Germany) coupled to a 6224 Accurate Mass TOF LC/MS system (Agilent Technologies, Singapore).Compounds 6b,i and 7a -f,i were prepared according to the literature procedures [11].
The substrate was prepared in the Reaction Buffer freshly obtained.The indicated PARP was delivered into the substrate solution and gently mixed.Compounds were delivered in DMSO into the reaction mixture using an Acoustic Technology (Echo 550, LabCyte Inc. Sunnyvale, CA) in the nanoliter range and incubated 20 min at room temperature. 32P-NAD + was delivered into the reaction mixture to initiate the reaction.The mixture was incubated for 2 h room temperature and then delivered to filter-paper and washed with 0.75 % phosphoric acid for detection.
The data were analyzed using the Excel and GraphPad Prism software.The PARP activity data were expressed as the percent of the remaining PARP activity in test samples (data 1 and data 2) compared to vehicle (dimethyl sulfoxide) reactions.IC 50 values and curve fits were obtained using the Prism4 Software (GraphPad).

The synthesis of compounds 1a-j
To a solution of compounds 6a -j (1 mmol) and diphenylphosphoryl azide (0.33 g, 1.2 mmol) in toluene (10 mL), triethylamine (0.15 g, 1.5 mmol) was added.The mixture was stirred at 80 °C for 6 h.After completion of the reaction, the mixture was cooled to room temperature and evaporated to dryness under reduced pressure.The residue obtained was treated with saturated solution of sodium hydrogen carbonate (20 mL).The solid was filtered, washed with water (20 mL), methyl tert-butyl ether (10 mL), dried on air and purified by recrystallization from acetonitrile to obtain the corresponding product as a white solid.