Slide 1: De-Risking Clinical Hepatotoxicity in Early Drug Discovery
De-Risking Clinical Hepatotoxicity in Early Drug Discovery Webinar December 3, 2024 1 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 2: Webinar Outline
Webinar Outline • Simulations Plus Overview • NEW! Liver Safety+ Package: De-Risking Clinical Hepatotoxicity in Early Drug Discovery – Introduction to the ADMET Predictor DILIsym (APD) module – Development of machine learning models to predict DILIsym toxicity inputs – How to use the APD module and
Slide 3: Who We Are
Who We Are NASDAQ: SLP Regulatory Agencies Using our Technology Pharmaceutical, biotechnology, chemicals, cosmetics, & consumer >280 goods companies in the U.S., Europe, Asia, and South America 200+ Employees >25 Established Worldwide yrs. In 1996 3 © Copyright 2024, Simulations Plus, Inc. All Right
Slide 4: GastroPlus/ADMET Predictor: By the Numbers
GastroPlus®/ADMET Predictor®: By the Numbers… 200+ 15+ Regulatory agencies Companies licensing with reviewers worldwide trained on software 350+ 120+ Universities in 50+ Services projects countries with executed over past access to software 12 months 700+ 90+ Scientists received Peer-reviewed softwa
Slide 5: Confidence Level on Clinical Performance
Complementary Solutions Clinical Pharmacology & Pharmacometrics Medical Communications Decrease development uncertainty, cost, time, and failure rates. Regulatory/ Exposure NCA Response Post Market PBPK / PBBM QSP: Inflammatory, Immune-Oncology, Confidence Level on Clinical Performance Adaptive Appr
Slide 6: World's Largest Library of Platform QSP and QST Models
Simulations Plus Has the World’s Largest Library of Platform QSP and QST Models to Predict Disease and Injury Outcomes QST: Liver and Kidney Safety QSP: Inflammation and Immunology (including fibrotic diseases) • Drug induced acute kidney injury • Asthma/COPD (in development) • Drug induced liver in
Slide 7: Highlights of DILIsym Version X (DSX)
Highlights of DILIsym Version X (DSX) • Completely NEW software platform! – Much faster and more user-friendly design – Command line and GUI options – No reliance on MATLAB base or runtime – Server/cloud computing capability (HPGL) • 4 NEW exemplar compounds included with varying clinical presentati
Slide 8: Webinar Outline
Webinar Outline • Simulations Plus Overview • NEW! Liver Safety+ Package: De-Risking Clinical Hepatotoxicity in Early Drug Discovery – Introduction to the ADMET Predictor DILIsym (APD) module – Development of machine learning models to predict DILIsym toxicity inputs – How to use the APD module and
Slide 9: The ADMET Predictor DILIsym (APD) Module Adds Liver Safety Insights
The ADMET Predictor 12 DILIsym (APD) Module Adds Liver Safety Insights to Evidence Informing Compound Selection • Liver safety liabilities are commonly identified when a compound enters the clinic, sometimes as late as phase 3 clinical trials, imparting considerable expense and potential delays to d
Slide 10: Roadmap to Derive an Early Assessment of Hepatotoxic Risk
Simulations Plus Has Developed a Roadmap to Derive an Early Assessment of Hepatotoxic Risk • New DILIsym module in ADMET Predictor 12 generates outputs that can be used to inform inputs within the quantitative systems toxicology (QST) modeling platform DILIsym – Permissive of liver safety assessment
Slide 11: DILIsym QST Model Predicts Liver Toxicity
DILIsym QST Model Predicts Liver Toxicity by Integrating Exposure, Mechanisms, and Inter-Patient Variability Tox Exposure Mechanisms • Mitochondrial dysfunction • Oxidative stress • Bile acid transporter inhibition Tox • Phospholipid transporter inhibition Relevant Biochemistry 11 © Copyright 2024,
Slide 12: Predicted Liver Exposure Interacts with Mechanisms of Toxicity in DILIsym
Predicted Liver Exposure Interacts with Data-Defined Mechanisms of Toxicity in the DILIsym In Vivo Environment Mechanisms of toxicity in DILIsym 1• Reactive oxygen species (ROS) 2• Mitochondrial dysfunction 3• Bile acid transporter inhibition – Bile salt export pump (BSEP) 4 1 – Multidrug resistance
Slide 13: Liver Safety+ Prediction Package Tailored for Early Discovery Data
Liver Safety+ Prediction Package Tailored for Early Discovery Data Toxicity Exposure Mechanisms • Constant liver exposure • Machine learning informed • High-throughput PK (HTPK) structure-based outputs DILI Relevant Biochemistry 13 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASD
Slide 14: APD Module Applies Machine Learning to Bridge from Compound Structure to DILIsym
APD Module Applies Machine Learning to Bridge from Compound Structure to DILIsym Compound Library In Vitro Assay Data Filtering, Automated Fitting, Translation Mitochondrial respiration Databases (Seahorse assay) Oxidative stress (High content screening) Bile acid efflux transporter inhibition (Ins
Slide 15: APD Module Outputs Include Values for Four Key Mechanisms of Hepatotoxicity
APD Module Outputs Include Values for Four Key Mechanisms of Hepatotoxicity • APD module provides classifications (yes/no) and key parameter values for each of the four main mechanisms of toxicity represented in DILIsym • Outputs are evaluated for potential toxicity • If outputs suggest toxicity, us
Slide 16: Webinar Outline
Webinar Outline • Simulations Plus Overview • NEW! Liver Safety+ Package: De-Risking Clinical Hepatotoxicity in Early Drug Discovery – Introduction to the ADMET Predictor DILIsym (APD) module – Development of machine learning models to predict DILIsym toxicity inputs – How to use the APD module and
Slide 17: The ADP Module Contains Three Mitochondrial Dysfunction Models
The ADP Module Contains Three Mitochondrial Dysfunction Models • Mito_Tox – Classification model that predicts Yes or No for mitochondrial toxicity based on the The single mispredicted negative from the test set is fenclozic acid, a Seahorse assay compound that was withdrawn from the market due to j
Slide 18: Mitochondrial Dysfunction Models: Mito_MEC
Mitochondrial Dysfunction Models With 2D and 2D+3D Descriptors Were Created: Mito_MEC • The Mito_MEC dataset contains 127 compounds with 13 (~10%) in the test set • The most active compound is rotenone, with an observed MEC value of 0.001 μM Plots show the log of the experimental Mito_MEC value in m
Slide 19: Mitochondrial Dysfunction Models: Mito_AC50
Mitochondrial Dysfunction Models With 2D and 2D+3D Descriptors Were Created: Mito_AC50 • The Mito_AC50 dataset contains 129 compounds with 22 (~17%) in the test set • The two most active compounds are antimycin A (Mito_AC50=0.01 μM) and rotenone (Mito_AC50=0.013 μM) Plots show the log of the experim
Slide 20: The ADP Module Contains Three Reactive Oxygen Species Models
The ADP Module Contains Three Reactive Oxygen Species Models • ROS_Tox – Classification model that predicts Yes or No for reactive oxygen species formation – Based on dataset containing 243 molecules with 25 (~10%) in the test set • ROS_MEC – Predict the minimum effective concentration (MEC) that si
Slide 21: BSEP Models and New MRP3 Model for Bile Acid Transporter Inhibition
The ADP Module Utilizes the Existing BSEP Models in ADMET Predictor and Contains a New MRP3 Model for Bile Acid Transporter Inhibition • BSEP_Inh – Classification model that predicts Yes or No for inhibition of the bile salt export pump (BSEP), a bile acid transporter on the canalicular membrane of
Slide 22: Webinar Outline
Webinar Outline • Simulations Plus Overview • NEW! Liver Safety+ Package: De-Risking Clinical Hepatotoxicity in Early Drug Discovery – Introduction to the ADMET Predictor DILIsym (APD) module – Development of machine learning models to predict DILIsym toxicity inputs – How to use the APD module and
Slide 23: Import/Load a Chemical Structure in ADMET Predictor
Import/Load a Chemical Structure (e.g., SMILES) in ADMET Predictor 23 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 24: Generate 3D Structure and Calculate ADMET: DILIsym Properties
Generate 3D Structure/Coordinates and Calculate ADMET: DILIsym Properties 24 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 25: Review and Export APD Module Results
Review and Export APD Module Results 25 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 26: Use and Interpretation of APD Module Classifications and Parameter Values
Use and Interpretation of APD Module Classifications and Parameter Values • Model provides yes/no classification predictions for active toxicity mechanisms based on compound structure Does model predict toxicity for – Mitochondrial dysfunction (solithromycin: yes) given mechanism (i.e., yes/no class
Slide 27: Model-Predicted MEC and AC50 for Mitochondrial Toxicity
Model-Predicted MEC and AC50 for Mitochondrial Toxicity Used to Derive DILIsym Parameters for Mitochondrial Effects Example ETC inhibitor C alues provided a o of Control 5 by D module 05 0 Toxicity Data Point 1* Data Point 2† Data Point 3‡ Mechanism • MEC: minimum effective ETC inhibitor 1 (linear)
Slide 28: Model-Predicted MEC and AC50 for ROS Induction
Model-Predicted MEC and AC50 for ROS Induction Used to Derive DILIsym Parameters for Effect on ROS Production Example ROS Production S 5 alues provided by D module a o of Control 0 Toxicity Data Point 1* Data Point 2† Data Point 3‡ Mechanism • MEC: minimum effective ROS production 1 (linear) concent
Slide 29: Model-Predicted IC50 Values for BSEP and MDR3
Model-Predicted IC50 Values for BSEP and MDR3 Can Be Utilized Directly as DILIsym Parameters for Bile Acid and Phospholipid Transport Inhibition Effects • IC50: concentration at which 50% inhibition is observed 29 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 30: Webinar Outline
Webinar Outline • Simulations Plus Overview • NEW! Liver Safety+ Package: De-Risking Clinical Hepatotoxicity in Early Drug Discovery – Introduction to the ADMET Predictor DILIsym (APD) module – Development of machine learning models to predict DILIsym toxicity inputs – How to use the APD module and
Slide 31: Multiple Options for Liver Exposure in DILI Toxicity Ranking Process
Multiple Options for Liver Exposure in DILI Toxicity Ranking Process • APD module is designed to provide insight into DILI toxicity rankings at any stage in the drug development pipeline • Based on where a compound is in the drug development pipeline, different information about exposure in humans i
Slide 32: DILIsym Drug Parameter Set and Constant Liver Concentrations
DILIsym Drug Parameter Set and Use of Specified Data Feature to Define Constant Liver Concentrations 32 © Copyright 2024, Simulations Plus, Inc. All Rights Reserved. | NASDAQ: SLP
Slide 33: APD Module Predictions Are Used to Set Up Active Toxicity Mechanisms in DILIsym
APD Module Predictions Are Used to Set Up Active Toxicity Mechanisms in DILIsym Machine Learning Algorithms • Mitochondrial dysfunction • Oxidative stress • Bile acid efflux transporter inhibition • Phospholipid transporter inhibition ADMET Predictor 12 DILIsym module Novel Compound (Solithromycin)
Slide 34: APD Module Predictions Are Used to Set Up Active Toxicity Mechanisms in DILIsym
APD Module Predictions Are Used to Set Up Active Toxicity Mechanisms in DILIsym Liver concentration Liver concentration Maximum ALT (fold change from IC50)* (nM) (U/L) 4.3e-7 0.01 30 2.2e-6 0.05 30 4.3e-6 0.1 30 2.2e-5 0.5 30 4.3e-5 1 30 2.2e-4 5 30 4.3e-4 10 30 2.2e-3 50 35 4.3e-3 100 112 2.2e-2 50
Slide 35: Interpretation of Toxicity Ranking Results
Interpretation of Toxicity Ranking Results For drugs early on in development pipeline For drugs further along in development pipeline (using constant liver exposure method) (using known liver concentrations or predicted using ADMET Predictor HTPK module or PBPK model Compound 3 Lowest potential for
Slide 36: APD Module Reproduces Toxicity Rankings: Macrolide Antibiotics
APD Module Outputs Reproduce Clinical and Previous DILIsym Simulation Toxicity Rankings: Macrolide Antibiotics Clinical Data & Previous DILIsym ML Tox Model Predictions Simulation Results Azithromycin / Lowest Erythromycin potential for hepatotoxicity Telithromycin Solithromycin Highest Clarithromyc
Slide 37: APD Module Reproduces Toxicity Rankings: CGRPR Antagonists
APD Module Outputs Reproduce Clinical and Previous DILIsym Simulation Toxicity Rankings: CGRPR Antagonists Clinical Data & Previous DILIsym ML Tox Model Predictions Simulation Results Zavegepant Lowest potential for hepatotoxicity Ubrogepant Atogepant Rimegepant Highest Telcagepant potential for hep
Slide 38: Workflow: APD Module Enables Efficient Assessment of Hepatotoxic Rankings
Workflow: APD Module Enables Efficient Assessment of Hepatotoxic Rankings for In-Class Compounds at Any Stage of Drug Development! Determine relevant group Run APD module and Use assumptions Extract known EC50, IC50, of compounds to be export predictions for previously discussed to or Ki values for
Slide 39: How to Engage with SLP?
How to Engage with SLP? Business Development andrew.mueller@simulations-plus.com - Director, Business Dev. Today’s Presenters lisl.shoda@simulations-plus.com – Vice President and Director of Immunology, Quantitative Systems Pharmacology michael.lawless@simulations-plus.com – Senior Principal Scienti