Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
eTOXsys – Highly Relevant Proprietary Data and Predictive Models for Drug Development Scott McDonald Director of Member Services [email protected] Summary • Background – eTOX • eTOXsys overview • • • • • Interface Data Ontology Models SEND • Use cases • Sustainability Who is involved 4 Project aims by 31st December 2016 Predict more complex toxicity endpoints by: • Sharing high quality proprietary toxicity data previously accessible for modelling only by the owning company • Development of several models representing the different components in the mechanism leading to a toxic effect • Integrated tool into an overall decision-making tool Data Models Integrated System Value proposition • Improved early drug candidate safety assessment • Results can be applied to: • Targeted in vitro screening • Triggering the inclusion of specific target organs in early in • • • • vivo studies Extrapolation from short term to long term effects Raising hypotheses with regards to on-target and off-target toxicities Interaction with MedChem and Pharmacology with regards to hypothesis on structural moieties responsible for effects A facilitated target safety review Data Confidentiality levels Confidential Shared Data In-House Data Only accessible via secrecy agreement Only accessible to EFPIA data owner Non-Confidential Shared Data Public Data Accessible to eTOX consortium Accessible outside eTOX consortium Data Database contents = Latest release 2016.2 Non-confidential Structures 1,449 Other routes; 53 Other injection; 251 Dermal; 58 Confidential Structures 466 Non-confidential Studies 7,509 Baboon; 4 Inhalation; 259 Dog; 2035 Guinea pig; 9 Intravenous; 1068 Hamster; 6 Rat; 4515 Marmoset; 38 Monkey; 449 Oral; 5788 82-100 days; 593 101-364 days; 400 Mouse; 349 Pig; 16 >365; 86 Rabbit; 81 36-81 days; 147 <20 days; 3944 20-35 days; 2333 Data Top 10 organs affected by histopathology Liver 806 658 Thymus 852 426 Kidney 990 391 Spleen 884 373 Adrenal gland 867 291 Stomach 265 Lung 265 Mesenteric lymph node 799 1005 727 249 Bone marrow 202 Testis 199 0 200 523 873 400 Negative structures 600 800 Positive structures 1000 1200 Ontobrowser & the eTOX ontology Data Predictive Models • eTOX computational modellers have developed a range of models • ADME, organ toxicity and target safety pharmacology • Built on chemical space highly relevant to drug development • Verified and well documented Modular Approach • Integrated but modular package • Access to other databases will be technically enabled • Predictive models will be available and users can choose the ones they want eTOXsys – the integrated system eTOXsys – data query eTOXsys – data query eTOXsys – data query eTOXsys – data query eTOXsys eTOXsys eTOXsys Predictive models Predictive models Predictive Models Predictive models Predictive models SEND • Standard for Exchange of Nonclinical Data, captured at individual animal level • FDA require the data for a NDA to be submitted in SEND format for studies which start after the 18th of December 2016 • eTOX donated data captured at dose group level • Lhasa has developed a tool that will permit the import of data in SEND format into eTOXsys Use case 1 – Framework for early drug development candidate assessment • Framework of four key approaches, each accessible through eTOXsys, was introduced to early small molecule drug development • 60 projects have been assessed to May 2016 at the transition point from hit to lead against this framework • Results for target search • Identical targets found for 10 projects (17%) • Similar/related targets (down-stream targets, antagonists agonists) for 10 projects (17%) • Results for chemical similarity search: • >50% similarity in 34 projects (57%) • >60% similarity in 8 projects (13%) • 70% similarity in 3 projects (5%) Use case 1 – Framework for early drug development candidate assessment • The eTOX database represents a valuable tool for early assessment of drug candidate and target assessment • Results of the database search can be used: • To identify tool compounds for comparison of the safety profile • To optimise the design of early in vivo screening studies (inclusion of non-routine organs and tissues) • To initiate discussions between Medicinal Chemistry, Pharmacology and Toxicology for compound optimisation Use Case 2 – Relevance of histopathology findings observed in a rat toxicity study • Renal papillary necrosis was observed in a rat toxicity study • eTOXsys was searched for compounds with the same effect • 49 compounds (81 studies) were identified (all 49 affected rats, only 3 for dogs) • 30 compounds with annotated pharmacology: 53% were EGF or VEGF inhibitors • Further literature research established a link between EGF receptors and renal papillary necrosis Use Case 2 – Relevance of histopathology findings observed in a rat toxicity study • Inhibiting epidermal/vascular growth factors (on- or offtarget) could be related with the induction of renal papillary necrosis • Rats are more susceptible to chemically induced renal papillary necrosis compared to dogs/other species • Both conclusions used for assessment of the relevance of renal papillary necrosis observed in rats for the human situation Use Case 2 – Relevance of histopathology findings observed in a rat toxicity study Use Case 3 – Consideration of potential target related effects in preclinical screening • Investigation of antagonists for specific receptors within the P2Xn receptor family due to their prominent role in nociception • Searching eTOXsys for structurally similar compounds led to the retrieval of Fluopyram, a fungicide • Reported treatment-related effects included: • Liver (hypertrophy) in the mid-dose • Thyroid (hypertrophy) in the high dose • These effects are mechanistically well characterised and considered to be species-specific without relevance for human safety (Rouquié et al. 2014) Use Case 3 – Consideration of potential target related effects in preclinical screening • The study director was advised to include thyroid in the screening panel (not normally included) • If comparable liver or thyroid effects are observed during screening of pivotal 4-week toxicity studies, these can be de-risked based on the knowledge obtained from eTOXsys and the literature Summary - eTOXsys • eTOXsys delivers: • Access to proprietary data that is not available elsewhere • An ontology and mining platform built specifically for this type of data • Well documented and verified models tested over a chemical space that is highly relevant for drug development • SEND compatibility Sustainability • eTOX project ends in December 2016 • Moves to sustainability phase in January 2017 • Molecular Networks and Lhasa Limited will continue to work closely together • Lhasa Limited will act as the ‘business broker’ • Framework is being put in place to allow non-eTOX participants to gain access to eTOXsys Sustainability • Product development • Promotion • Licensing • Distribution • Training • Support Summary – eTOXsys and Sustainability • eTOXsys delivers: • Access to proprietary data that is not available elsewhere • An ontology and mining platform built specifically for this type of data • Well documented and verified models tested over a chemical space that is highly relevant for drug development • SEND compatibility • Sustainability allows: • This valuable resource to be used more extensively • Provides ongoing updates • Users will be well supported and trained Next Steps • Please let us know if you would like further information on eTOXsys • Join us for our Virtual ICGM: • Wednesday 16th November at 4pm (UK time) • https://www.lhasalimited.org/events/webinar-etoxsys-highlyrelevant-proprietary-data-and-predictive-models-for-drugdevelopment/4075 Acknowledgements • All those involved in the eTOX project Questions?