Experimental determination and computational design of antiaggregatory effect of polyphenolics

Među prirodnim spojevima prisutnim u svakodnevnoj prehrani, flavonoidi su pokazali povoljan učinak u prevenciji kardiovaskularnih bolesti koji se, barem djelomično, može pripisati antiagregacijskom učinku. S obzirom na farmakološki interes, u potrazi za antitrombocitnim lijekovima, potrebna je sustavna eksperimentalna procjena antiagregacijskog učinka flavonoida. Takvi podaci mogli bi služiti za QSAR modeliranje antiagregacijskog učinka, istraživanje signalnih putova i procjenu utjecaja na invitro testove agregacije trombocita. Skup od trideset flavonoida, odabran je za procjenu antiagregacijskog učinka, na uzorcima pune krvi pomoću Multiplate® funkcionalnog analizatora (Dynabyte, Njemačka) i ADP-a (ADPtest) kao slabog agonistaagregacije. Pet učinkovitih flavonoida iz ADPtesta je analizirano s četiri dodatna agonistaagregacije (arahidonska kiselina, kolagen, ristocetin i TRAP-6). Za računalno simuliranje antiagregacijskog učinka, generirano je 155 molekulskih deskriptora koji opisuju fizikalno-kemijska, odnosno globalna svojstva molekula i supstituenata. Od računalnih metoda za procjenu mehanizama djelovanja primijenjeno je hijerarhijsko formiranje klasteravišedimenzijskim ponovnim uzorkovanjem. Slučajna šuma, metoda statističkog učenja, korištena je za QSAR modeliranje. Laboratorijski rezultati su iskazani kao minimalna koncentracija flavonoida koja dovodi do statistički značajnog smanjenja agregacije trombocita u odnosu na netretirani uzorak (minimalna antiagregacijska koncentracija ‒ MINaAC). MINaACflavonoida, u pojedinim testovima agregacijekoja je potaknuta ADP-om, kolagenom, TRAP-6 i ristocetinom, bila je u sljedećim rasponima: 0,12‒122,07 μM; 15,26‒244,14 μM; 15,26‒122,07 μM i 0,06‒15,26 μM. U testu agregacije trombocita koja je potaknuta arahidonskom kiselinom, proagregacijski učinak je zapažen kod pinocembrin-7-metiletera, epikatehina, hesperetina i 3,6- dihidroksiflavona. Literaturni su podaci bili nedostatni za interpretaciju rezultata dobivenih tvorbom hijerarhijskih klastera uslijed raznolikih mehanizama djelovanja pojedinih flavonoida. Validacija predviđanja učinka koja se temelji na metodi slučajne šume, rezultirala je niskom točnošćupredviđanja od 40,67%. Mjerljivantiagregacijski učinak, na submikromolarnoj razini koncentracija flavonoida, sugerira da i svakodnevna konzumacija flavonoida prehranom može utjecati na in vivoagregaciju trombocita, što ukazuje i na njihovu moguću terapijsku primjenu. Temeljem antiagregacijskih testova s različitim induktorima agregacije, može se zaključiti da flavonoidi interferiraju s invitroagregacijom trombocita, ili antiagregacijski ili proagregacijski, što može utjecati na interpretaciju testova agregacije trombocita na punoj krvi. Razvoj pouzdanog QSAR modela onemogućuju raznovrsni mehanizmi djelovanja flavonoida kojima se ostvaruje antiagregacijski učinak. Stoga se daljnja istraživanja trebaju usmjeriti na pojedinačne mete i na povećanje broja analiziranih supstancija.Among natural compounds, present in every day diet, flavonoids have shown beneficial effect in prevention of cardiovascular diseases that can be attributed, at least partially, to their antiaggregatory activity. Due to the ever increasing pharmacological interest in antiplatelet agents, a systematic experimental evaluation of large flavonoid series is needed. This will serve as possible data set for QSAR modeling of antiaggregatory activity, assessment of signaling pathways and evaluation of the in vitro effects of flavonoids on platelet aggregation in whole blood. A set of thirty flavonoid aglycones was selected for the evaluation. Aggregation measurements were performed on the whole blood samples with multiple platelet functional analyzer (Dynabyte, Germany) and adenosine diphosphate (ADPtest) as a weak agonist of aggregation. Five potent flavonoids from the ADPtest were further analyzed using the four additional aggregation inducers (arachidonic acid, collagen, ristocetin and TRAP-6). Computational design of antiaggregatory effect was based on 155 molecular descriptors of physical and chemical properties; global properties of molecule and substituents. Method for the assessment of the possible mechanisms of action used was hierarchical clustering with multiscale bootstrap resampling. Random forest, a statistical learning method, was used for QSAR modeling. Laboratory results were expressed as minimal concentration of flavonoid that can significantly lower the platelet aggregation compared to the corresponding untreated sample (minimal antiaggregatory concentration ‒ MINaAC). MINaAC of flavonoids in individual tests were reported in the following ranges: 0.12‒122.07 μM; 15.26‒244.14 μM; 15.26‒122.07 μM; and 0.06‒15.26 μM for ADP, collagen, TRAP-6 and ristocetin aggregation-inducers, respectively. When arachidonic acid was used for induction of platelet aggregation, a proaggregatory effect was observed for pinocembrin-7-methylether, epicatechin, hesperetin and 3,6- dihydroxyflavone. Literature data was inconclusive for proper interpretation of hierarchical clustering due to different mechanism by which flavonoids achieve antiaggregatory effect. Validation of random forest prediction model resulted in 40.67% accuracy. Measurable antiplatelet activity established at submicromolar flavonoid concentrations suggests that even a dietary consumption of some flavonoids can make an impact on in vivo aggregation of platelets. These findings also point out a therapeutic potential of some flavonoids. Based on the test with different agonists of aggregation it can be concluded that flavonoids interfere with in vitro platelet aggregation assays exhibiting either anti- or pro-aggregatory influencing the interpretation of the results of platelet aggregation. Development of reliable QSAR model under described settings is not possible due to different mechanisms responsible for antiaggregatory effect. Further studies should focus on specific targets; number of the analyzed substances should be increased