Session Title: Rheumatoid Arthritis: Animal Models
Session Type: Abstract Submissions (ACR)
Background/Purpose: Rheumatoid Arthritis (RA) involves complex interactions of multiple cell systems, cytokines and mediators, and interlinked signaling. While molecularly-targeted therapies manipulate specific interactions, it is possible that other, previously redundant, paths are subsequently activated, thus causing drug resistance. Ideal therapy design requires simultaneous modulation of multiple targets to achieve eventual convergence and synergy. A predictive software algorithm would tremendously increase the ability to identify such therapies.
Methods: A predictive in-vivo equivalent simulation technology emulating RA pathophysiology at cellular and molecular level was designed by assimilating information from 8928 publications over the past 8 years. The simulation platform is a co-culture of 8 cell systems representing RA phenotypes, associated signaling and metabolic networks and includes nearly 100000 cross-talks and interactions among approximately 31366 molecules. The platform was extensively validated using more than 4000 studies and correlated with retrospective human clinical drug data and animal experiments. A digital library representing 100 targeted drugs from different indications was screened in combinations of two’s and three’s translating to more than one million in-vivo studies. Automated analytics engine using the assay data from the studies generated a therapy candidate which is a combination of 3 oral FDA-approved drugs (CWG952), based on criteria of efficacy, synergy and PK/PD compatibility. For in-vivo validations, mCIA was induced in male DBA/1 mice and those with paw scores between 1 and 5 on day 19 were randomly assigned to treatment arms (9/group): CWG952 and anti-TNF with treatment started at 12hrs of enrollment.
Results: The RA predictive model was dynamically simulated to disease condition followed by application of CWG952 therapy. The simulations predicted a percentage reduction in TNF, IL6, IL1B, IL17a, RANKL and CRP by 81.8, 90.5, 85.9, 86.5, 85.1 and 95.5 respectively from disease condition with respect to control state. The model also predicted a 83.1 % reduction in bone and cartilage degradation and 85.1% inhibition of osteoclast activity. Anticipated effect on phenotypic disease markers were close to those seen in clinical trials of TNF-alfa inhibitors. The figure shows results from animal experiments (mCIA model) demonstrating a comparable efficacy to etanercept.
Conclusion: In-vivo validations of mathematically-designed CWG952 show results comparable to etanercept. These results, along with other previously published validations , suggest a robust validation of a software-based mathematical modeling approach that incorporates the complex interaction of immunomodulating molecules, and can potentially lead to the development of innovative therapies for immunological diseases.
J. D. Cruz,
« Back to 2012 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/predictive-software-based-mathematical-modeling-a-novel-approach-to-development-of-oral-therapies-for-rheumatoid-arthritis-validation-in-a-murine-collagen-induced-arthritis-model/