Hybrid AI for modeling Prognosis of Inflammation process in AGING – HAPI-AGING

Aging is associated with a progressive decline in physiological functions that may lead to age-related disease (ARD), frailty and eventually dependence. This project is based on the hypothesis that tissue aging is linked to a series of cellular and tissue challenges that are not properly addressed physiologically. Among the biological processes supporting physiological decline, gerosciences recently identified an alteration of the inflammatory response, and especially the resolution phase, as a key process. To date, current aging bioclinical markers used by physicians are ineffective to predict and potentially prevent age-dependent functional decline.
Neutrophils, macrophages (Mph) and mesenchymal stromal cells (MSC) are major actors involved in the resolution of inflammation. They all play critical but specific roles in maintaining tissue homeostasis and supporting tissue repair through resolving inflammation. . However, the precise role of MSC and Mph/MSC dynamic interactions in the inflammatory process, including inflammation resolution in the context of tissue repair and aging, is still poorly understood. We hypothesize here that aging induces an alteration of the immunomodulatory role of MSC and/or MSC/Mph dynamic interactions that in turn contributes to the alteration of the resolution of inflammation.
Computational approaches are thus fundamental for testing in silico the effect of transient and repeated challenges, of various intensities, to model long term effects from a series of short-term events and to identify new functional decline markers. Obtaining such a model paves the way for tissular digital twin, and in silico optimization of new personalized therapies. Indeed, HAPI-AGING in silico model of inflammation may provide inflammatory process insights, including the sequential cause-and-effect relationships between the different cellular actors. Both partners developed an Agent-Based Model (ABM) of inflammation integrating Mph and their preliminary data suggest a missing agent in the model. We hypothesized that MSC could be these “missing agents” and we propose to implement the current Mph-based ABM of inflammation by adding MSCs as agents and by attributing them an immunomodulatory role. Moreover, given the high number of parameters involved in the inflammatory process and the computationally cost of ABM in comparison to other approaches such as differential equations, we aim to hybrid ABM to artificial intelligence (AI) Machine-Learning (ML) modeling. To achieve this goal, we envision to bring the proof of concept of this new hybrid AI to create a model for inflammation process prognosis in aging with MSC and Mph. Using our Web-oriented modeling platform to interactively build ABM and biological data from our relevant models of inflammation, our proposal will aim (i) to improve the current ABM by adding MSC with an immunomodulatory role and surrogate this model using ML to create an innovative Hybrid ABM-ML (Hybrid AI) model, (ii) to simulate the impact of aging on the inflammatory process and demonstrate the new hybrid model predictability.
HAPI-AGING addresses one of the major challenges of our societies by proposing an innovative tool to fight against frailty, dependence, ARD and prevent their associated consequences. Based on the high-level competencies of our consortium members, our project combines multidisciplinary approaches to develop new predictive approaches of the physiological age of individuals and their aging trajectory, defining early biomarkers of tissue dysfunctions. This project will open new perspectives in the definition of the operational process for a future translational approach for cell therapy in human.