Title Razvoj 3D-QSAR modela za tip I inhibitore Src kinaze Title (english) Development of 3D-QSAR models for Type I Src kinase inhibitors Author Lidija Stošić Mentor Sanja Koštrun (mentor) Committee member Željko Svedružić (predsjednik povjerenstva) Committee member Jelena Ban (član povjerenstva) Granter University of Rijeka Defense date and country 2021-10-25, Croatia Scientific / art field, BIOTECHNICAL SCIENCES Abstract ZNAČAJ: Src kinaza je tirozin kinaza koja sudjeluje u prijenosu signala povezanih s brojnim staničnim procesima, kao što su proliferacija, diferencijacija, ili adhezija. Src kinaza poželjna je terapijska meta jer ima prekomjernu ekspresiju, visoku aktivnost i jak utjecaj na razvoj metastaza u raznovrsnim tipovima karcinomima.
METODE: Razvili smo kvantitativne i kvalitativne 3D-QSAR modele koji se temelje na molekulskim poljima koristeći literaturno dostupne Src inhibitore tipa I.
REZULTATI: Razvijena su 4 odvojena modela za 4 klase strukturno sličnih analoga; PP2 analozi, Dasatinib analozi, Saracatinib analozi, Zajednički skup, od kojih se model za skup Saracatinib analoga pokazao kao najbolji model visoke točnosti i stabilnosti s prihvatljivom prediktivnom moći (PLS: q2=0.63; SVM: q2=0.60) i dobrim koeficijentom korelacije (PLS: r2=0.92, rtes2=0.85; SVM: r2=0.95, rtest2=0.28). Dodatno, cilj je bio usporediti pristup temeljen na ligandima (engl. LBDD, ligandbased drug design) s pristupom temeljenim na strukturi (engl. SBDD, structure-based drug design). Spojevi korišteni za izgradnju 3D-QSAR modela uklopljeni su u DFG-"in" konformaciju Src kinaze koristeći literaturno dostupne kristalne strukture, te su uklopljene komplekse minimizirani MM-GBSA metodom. Niska korelacija (r=0.48) eksperimentalnih aktivnosti i Gibbsove energije interakcije dobivene MMGBSA metodom, ukazuje kako pristup temeljen na strukturi ne poboljšava predviđanje aktivnosti Src inhibitora u odnosu na pristup temeljen na ligandu. U posljednjem koraku dobiveni rezultati su uspoređeni s prethodno objavljenim studijama te je pokazano da je predstavljeni model za saracatinib skup podataka bolji od do sada objavljenih studija te da MM-GBSA metoda nije korisna za efikasno predviđanje aktivnosti.
ZAKLJUČAK: Opisali smo odnose između strukture i aktivnosti za tip I inhibitore Src kinaza koristeći robusnije i pouzdanije modele s boljom prediktivnom moći za dizajn, filtriranje i prioritiziranje virtualnih biblioteka potencijalnih SRC inhibitora. Predstavljeni rezultati mogu se korsititi za planiranje sinteze novih i aktivnijih inhibitora.
Abstract (english) BACKGROUND: Src kinase is a tyrosine kinase that participates in the transmission of signals associated with numerous cellular processes such as proliferation, differentiation, or adhesion. Src is a desirable therapeutic target because it is overexpressed and highly active in various cancers, and it influences the development of metastasis as well.
METHODS: Field-based quantitative and qualitative 3D-QSAR models were developed for literature-available Type I Src inhibitors.
RESULTS: Four separate models have been developed for 4 classes of structurally similar analogues; PP2 analogues, Dasatinib analogues, Saracatinib, analogues, Common set, of which the model for the Saracatinib analogue set proved to be the best model of high accuracy and stability with acceptable predictive power (PLS: q2=0.63; SVM: q2=0.60), and good correlation coefficient (PLS: r2=0.92, rtes2=0.85; SVM: r2=0.95, rtest2=0.28). Additionally, the aim was to compare the ligand based approach (LBDD) with the structure-based drug design (SBDD). Compounds used to build the 3D-QSAR model were incorporated into the DFG-"in" conformation of Src kinase literature-available crystal structures, and the incorporated complexes were minimized by the MM-GBSA method. The low correlation (r=0.48) of the experimental activities and the Gibbs interaction energy obtained by the MM-GBSA method indicates that the structure-based approach does not improve the prediction of Src inhibitor activity compared to the ligand-based approach. Finally, the obtained results were compared with previously published studies and we concluded that the saracatinib data set is better than the previously published studies, and that the MM-GBSA method is not useful for an efficient prediction of activity.
CONCLUSION: We described the structure-activity relationships for type I Src kinase inhibitors using more robust and reliable models with better predictive capability; design; filtering and prioritization of virtual libraries of potential Src inhibitors. The presented results can be used to plan the synthesis of new and more active inhibitors.
Keywords
Kinazni inhibitori
SRC kinaza
CADD
3-D QSAR
Molekulsko uklapanje
Keywords (english)
Kinase inhibitors
Src kinase
CADD
3D QSAR
Molecular docking
Language croatian URN:NBN urn:nbn:hr:193:795263 Study programme Title: Drug research and development Type of resource Text File origin Born digital Access conditions Open access Terms of use Repository Repository of the University of Rijeka, Faculty of Biotechnology and Drug Development Created on 2021-10-26 11:35:12
