Computer Aided Design And Anti-tumor Research Of ATR Inhibitors

Ataxia Telangiectasia and Rad3-related（ATR）protein is a key kinase involved in DNA damage repair,which functions not only to trigger cell repair in response to DNA damage,but also to stabilize replication forks and regulate cell cycle progression.Since ATR is critical for cell survival,in many tumor cells,ATR appears to be even more important than it’s in normal cells.Loss of alternative pathway for regulating cell cycles and tumorigenesis renders tumor cells more sensitive to disruption of ATR signaling,showing a synthetic lethality for ATR inhibition.Accordingly,considered as an ideal anti-tumor target,ATR inhibitor has not been fully exploited to date.To develop novel ATR inhibitors with high cytotoxicity and high selectivity,we have designed two ATR inhibitors using computer-aided drug design technology,followed by synthesis and validation.This dissertation includes five chapters to fully demonstrate our studies on the development of ATR inhibitors.Chapter 1.ReviewThis chapter summarized the ATR functions in tumor cells and anti-tumor applications of ATR inhibitors,followed by the research progress of common computer drug discovery techniques.First,the structure and functions of ATR protein and its functions in DNA damage repair pathways were briefly reviewed,indicating ATR as a potential biomarker as well as target in anti-tumor treatment.Second,some potent ATR inhibitors that have been developed in recent years were reviewed.Besides,computer-aided drug discovery approaches and their applications have been reviewed and compared,showing theoretical and technical supports for subsequent studies.Chapter 2.Design of ATR inhibitors based on pharmacophore model and molecular fragment methodThis chapter showed the development of novel ATR inhibitors using pharmacophore model and fragment-based drug design approach.First,28 ATR inhibitors with varied structures were employed to generate five structure-based pharmacophore models.To examine whether or not the established models were good enough to predict the active compounds,five models were validated by test set which was composed by 300 known ATR inhibitors,showing that the LB-2 models was the best model with the highest accuracy.Next,42 pharmacophoric features were extracted from 140 reported ATR inhibitors(IC₅₀<20 nM)using LB-2 model;then molecular docking technique was utilized to edit redundant features,indicating that only 29 pharmacophoric features were representative.After these operations,a structure-based pharmacophore model comprising the most import pharmacophoric features was built and subsequently,this LB-2 model was applied to design 125 novel ATR inhibitors with the help of fragment-based drug design approach.Chapter 3.Evaluation of designed ATR inhibitorsIn this chapter,these potential ATR inhibitors were virtually assessed using support vector machines（SVM）strategy and molecular docking models.First,a total of 700 compounds,which comprised of 400 negative non-inhibitors and 300 known positive inhibitors,was obtained from BindingDB database and ZINC database,and selected as training set.Additionally,a sample of 188 compounds,including 100non-inhibitors and 88 inhibitors,was also used as the test set.During the operation,94feature descriptors corresponding to inhibitor activities were selected to characterize structural properties of molecules.After examining based on training set,the best parameters to build the ATR inhibitor SVM predicational model（SVM-ATR）were identified,including the 7-Fold,c-SVC and RBF kernel functions.Results revealed that the evaluation index SE of the model was 99%,SP was 90.91%,ACC was95.21%,as well as the MCC was 90.59%.Using the training set to chose the model parameters,the 7-Fold,c-SVC Furthermore,the SVM-ATR model is further optimized with the grid search algorithm to determine the model parameters,indicating that parameters c,g were 1.4142,1.4142,respectively.The cross-validation accuracy rate is 95.5714%,and the final optimized prediction model（GS-SVM-ATR）SE was 100%,SP was 90.91%,ACC was 95.74%,and the MCC was 89.76%,respectively.Predication results of GS-SVM-ATR model showed that 66 compounds from the 125 compounds of designed might have ATR inhibitory activities.To further refine and remove the false positives,these 66 compounds were docked into binding site of ATR protein and scored compared with AZD6738 which is currently in clinical trial.Results showed that two compounds,which were labeled as 51,52,respectively,could stably bind with ATR protein,and the binding site and interact residues for two compounds were the same as that for AZD6738,showing that compounds 51,52might be potent ATR inhibitors.Chapter 4.Synthesis of ATR inhibitorsThe identified compounds,51 and 52,were subsequently synthesized for downstream study in this chapter.Firstly,the synthetic route of compounds 51,52 was established by inverse synthesis analysis.Next,compounds 51 and 52 were synthesized based on the synthetic route.The synthesized compounds were subsequently characterized by ¹H NMR,¹³C NMR and mass spectrometry,confirming that these compounds had correct structures.Chapter 5.In vitro inhibitory activity assay of ATR inhibitorsThe anti-tumor activities and the underlying mechanisms of compounds 51 and52 were studied in this chapter.Firstly,the anti-tumor activities of compounds 51 and52 were separately determined in breast cancer cell line MCF-7 at either single compound or a combination of one compound with doxorubicin.The inhibitory effects on the growth of MCF-7 cells of single compound was not obvious,showing that IC₅₀0 values were 496mM for compound 51,and 451mM for compound 52.However,the combination with a commercial anti-tumor agent doxorubicin demonstrated best anti-tumor effects,for example,IC₅₀0 values of doxorubicin mixed with compound 51 and 52 were 18.632mM and 12mM,respectively,which were significantly lower than that of doxorubicin alone.Moreover,the combination interaction（CI）value for compound 51 was 0.6511,showing moderate synergistic effect;the CI value for compound 52 was 0.4217,which was considered highly synergistic interaction.To be more specific,a variety of combination assay were determined to verify the potential anti-tumor mechanisms of two compounds,showing that these compounds could inhibit the proliferation of tumor cells during DNA damaged state and were similar to other ATR inhibitors in terms of anti-tumor mechanism.Besides,the optimal synergistic ratio for a combination of compounds 51and doxorubicin was 3.5:1,showing an IC₅₀0 value of 13mM;the best synergistic ratio was 3:1 for compound 52 and doxorubicin,resulting in a smaller IC₅₀0 value of 7mM.Taken together,these results indicated that a combination of designed ATR inhibitors with doxorubicin could more efficiently exert the anti-tumor activities.In summary,our studies have designed two potent ATR inhibitors based on a series of approaches,including ATR inhibitor pharmacophore model,molecular docking model,SVM prediction model and fragment based drug design method.The anti-tumor activities and its related anti-tumor mechanisms of two compounds were investigated,providing a strategy for the screening of novel ATR inhibitors.Also,our drug design strategy could be applied for the discovery of other small targeted drugs,showing alternative options for the clinical treatment of tumors.