Design and modelling in early drug development in targeting HIV-1 reverse transcriptase and Malaria
Kuupäev
2017-07-10
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Abstrakt
Käesolev uurimus keskendub kahele ohtlikule infektsioonhaigusele: inimese immuunpuudulikkuse viirus tüüp 1 (HIV-1) ja malaaria. Uue ravimi väljatöötamine algusest lõppuni on aega nõudev ning kulukas protsess, mis jaotatakse viieks etapiks: baas uurimistöö, põhi sihtmärgi ja baas ühendi(te) leidmine, eelkliiniline arendus, kliiniline arendus ja vajalike dokumentide esitamine ravimiametisse. Antud väitekirjas keskendutakse kahele esimesele etappidele, mida tuntakse ka varajase ravimiarenduse faasina.
HIV-1 uurimisel oli kaks põhisuunda. Esmalt tuginedes eelnevalt tehtud virtuaalsõelumise tulemustele teostati uudsete s-triasiini derivaatide avastamine, disainimine, ja süntees, mille tulemused valideeriti eksperimentaalselt ning analüüsiti valk-ligand interaktsioonimudelite abil. Kõige tõhusam HIV-1 mitte-nukleosiidne pöördtranskriptaasi inhibiitor oli madala molekulmassiga, heade ligandi efektiivsust näitavate parameetritega, ja madala toksilisusega, võimaldades edasist modifitseerimist ja arendamist. Tehtud aktiivse keemilise struktuuri avastus motiveeris HIV-1 inhibiitorite keemilise struktuuriruumi laiemat uurimist, et kindlaks teha kas uudsed s-triasiinid moodustavad ka unikaalsed keemiliste ühendite grupi HIV-1 mitte-nukleosiidsete pöördtranskriptaasi inhibiitorite maastikul. Selle läbiviimiseks koostati, korrastati ja kureeriti ChEMBL-i andmebaasist saadud andmetest fokusseeritud andmeseeriad HIV-1 mitte-nukleosiidne ja nukelosiidsete pöördtranskriptaasi inhibiitorite jaoks, kuhu lisati ka avastatud s-triasiini derivaadid. Andmeseeriate struktuuride analüüs hierarhilise klassifitseerimise meetodil grupeeris ühendid keemiliste struktuuritüüpide (nn. vanematüüp) järgi. Selgus, et avastatud s-triasiinid moodustasid eraldiseisva struktuuritüübi grupi. Leitud struktuuritüüpe analüüsiti, lisades juurde ka vastavad mõõdetud seondumise afiinsuse tasakaalukonstandid (Ki). Selle analüüsi käigus toodi välja struktuurifragmendid, mis omavad olulist rolli afiinsuse ning stabiilsuse seisukohast. Lisaks võimaldasid struktuurselt mitmekesised ja unikaalsed HIV-1 mitte-nukleosiidne ja nukelosiidsete pöördtranskriptaasi inhibiitorite andmeseeriad esmakordselt arendada kirjeldavaid kvantitatiivsete struktuur-aktiivsus sõltuvuste prognoosmudeleid, mida on võimalik kasutada järgnevas uurimustöös uute aktiivsete keemiliste ühendite avastamisel.
Selleks et leida uudseid malaaria ravimikanditaate koostati ja kureeriti süsteemselt andmebaas eksperimentaalsete anti-Plasmodium andmetega kasutades nii asutusesisesed, kui ka ChEMBL-i andmebaasis olevad andmed. Saadud andmete ulatusliku kureerimise, filtreerimise ning ühendamise tulemusena saadi kolmkümmend modelleeritavat andmeseeriat, millele koostati klassifitseerimise mudelid, eesmärgiga eristada aktiivsed ja mitteaktiivsed ühendid. Nendest seitsmeteistkümnele andmeseeriale saadi ennustusvõimelised nn. üksmeele (inglise keeles consensus) mudelid. Loodud mudelitega teostati ennustusi asutusesiseselt olemasolevatele curcuminoidide seerjale ning nende analoogidele, millest parima ennustusvõimega ühenditele teostati eksperimentaalne valideerimine in vitro katsetega, kus aktiivseks osutusid seitseteist ühendit, mida saab edasistes uuringutes täpsemini uurida. Samuti tehti kindaks, et arvutuslikult tuvastatud mitteaktiivsed ühendid jäid mitteaktiivseks ka eksperimentaalse valideerimise käigus, mis näitas süsteemselt kureeritud ja koostatud andmeseeriate ning prognoosmudelite jätkusuutlikust.
Current thesis focused on study of two highly prevalent infections affecting many regions in the world: alaria and human immunodeficiency virus 1 (HIV-1). Developing a new drug from scratch is time consuming and costly process. This could be divided into five stages: basic research, lead target and lead compound(s) discovery, preclinical development, clinical development and filing to drug administration agency. Present thesis focused on basic research and lead compound discovery stages, i.e. to the early drug discovery. For the HIV-1, the focus was two-fold. First, based on the earlier multi-objective in silico screening, novel s-triazine derivatives were designed, discovered, synthesized, and findings where supported by the modelling tasks and validated with biological evaluation. The most potent compound is with small molecular size, potent ligand efficiencies, and measured low toxicity permitting further exploration and modifications. Second, the discovered new bioactive s-triazines motivated to analyse the chemical landscape of HIV-1 RT inhibitors. For this the dataset was systematically created and curated for HIV-1 NNRT (non-nucleoside reverse transcriptase) and NRT (nucleoside reverse transcriptase) inhibitors based on data from ChEMBL database. The hierarchical classification of scaffold structures of curated datasets revealed common chemical parent types for the compounds, hierarchy in chemical structures and showed that discovered s-triazines formed a separate structural parent type group. Each group of compounds related to the parent type was analysed and examined together with corresponding binding affinity equilibrium constants (Ki). The structural fragments affecting the potency and stability of compounds were highlighted. The structurally diverse datasets for the HIV-1 NNRTIs and NRTIs with binding affinity equilibrium constants allowed development of novel descriptive and predictive QSAR models for log Ki, that in future will help in design of new compounds. In order to discover new promising antimalarial compounds, the experimental anti-Plasmodium data was gathered and systematically curated from in-house experimental studies and expanded with data from ChEMBL database. Extracted data was carefully extensively curated, fused, filtered, and grouped into thirty data sets for the modelling. The consensus models for each dataset for the classification of active/inactive compounds were established and seventeen models with promising prediction ability were used in consensus predictions and in identifying the series of curcuminoids and their structural analogues as potential inhibitors for the malaria. The selection of compounds was experimentally validated, i.e. tested in vitro, revealing seventeen potentially active compounds for further testing and modifications. The validation showed that computationally predicted inactive compounds were also inactive in experiment, being additional proof for the quality of data curation and dataset assembly process forming the ground for the modelling tasks
Current thesis focused on study of two highly prevalent infections affecting many regions in the world: alaria and human immunodeficiency virus 1 (HIV-1). Developing a new drug from scratch is time consuming and costly process. This could be divided into five stages: basic research, lead target and lead compound(s) discovery, preclinical development, clinical development and filing to drug administration agency. Present thesis focused on basic research and lead compound discovery stages, i.e. to the early drug discovery. For the HIV-1, the focus was two-fold. First, based on the earlier multi-objective in silico screening, novel s-triazine derivatives were designed, discovered, synthesized, and findings where supported by the modelling tasks and validated with biological evaluation. The most potent compound is with small molecular size, potent ligand efficiencies, and measured low toxicity permitting further exploration and modifications. Second, the discovered new bioactive s-triazines motivated to analyse the chemical landscape of HIV-1 RT inhibitors. For this the dataset was systematically created and curated for HIV-1 NNRT (non-nucleoside reverse transcriptase) and NRT (nucleoside reverse transcriptase) inhibitors based on data from ChEMBL database. The hierarchical classification of scaffold structures of curated datasets revealed common chemical parent types for the compounds, hierarchy in chemical structures and showed that discovered s-triazines formed a separate structural parent type group. Each group of compounds related to the parent type was analysed and examined together with corresponding binding affinity equilibrium constants (Ki). The structural fragments affecting the potency and stability of compounds were highlighted. The structurally diverse datasets for the HIV-1 NNRTIs and NRTIs with binding affinity equilibrium constants allowed development of novel descriptive and predictive QSAR models for log Ki, that in future will help in design of new compounds. In order to discover new promising antimalarial compounds, the experimental anti-Plasmodium data was gathered and systematically curated from in-house experimental studies and expanded with data from ChEMBL database. Extracted data was carefully extensively curated, fused, filtered, and grouped into thirty data sets for the modelling. The consensus models for each dataset for the classification of active/inactive compounds were established and seventeen models with promising prediction ability were used in consensus predictions and in identifying the series of curcuminoids and their structural analogues as potential inhibitors for the malaria. The selection of compounds was experimentally validated, i.e. tested in vitro, revealing seventeen potentially active compounds for further testing and modifications. The validation showed that computationally predicted inactive compounds were also inactive in experiment, being additional proof for the quality of data curation and dataset assembly process forming the ground for the modelling tasks
Kirjeldus
Väitekirja elektrooniline versioon ei sisalda publikatsioone
Märksõnad
malaaria, HIV-1, pöördtranskriptaas, malaariavastased ravimid, immuunsüsteemi mõjustavad ravimid, antiviiruslik aktiivsus, ravimidisain, malaria, reverse transcriptase, antimalarials, immunomodulators, antiviral activity, drug design