A mechanistic model of pure and lipid-mediated alpha-synuclein aggregation to advance therapeutic strategies against Parkinson’s disease
Elena Righetti (1,2), Luca Marchetti (1,2), Enrico Domenici (1,2), Federico Reali (1)
(1) Fondazione The Microsoft Research - University of Trento Centre for Computational and Systems Biology - COSBI, Italy; (2) University of Trento, Department of Cellular, Computational and Integrative Biology – CIBIO, Italy
Objectives:
The neuronal protein alpha-synuclein (aSyn) holds a crucial role in the intricate molecular landscape of Parkinson's disease. While aSyn unfolded form has essential physiological functions, mainly associated with the synaptic vesicle cycle, its aggregates have been suggested as key pathogenic triggers of neurodegeneration and considered potential therapeutic targets [1, 2]. Furthermore, recent advances in aSyn-based biomarkers have led to the definition of a new biological framework for the disease. However, many unclear aspects and mechanisms linked to aggregation hamper the quest for effective disease-modifying therapies.
The aggregation process relies on a complex chemical reaction network, including primary and secondary nucleation, oligomer interconversion, and fibril elongation. In addition, each reaction contributes to the whole process to a different extent depending on the pathophysiological conditions of the surrounding environment. Such a biological complexity comes with the limited availability of experimental measurements of aSyn intracellular dynamics in cell cultures and living systems. Data paucity results from technical difficulties in monitoring aggregates' time evolution and addressing oligomers' heterogeneous and elusive nature. As a result, a quantitative analysis of the aggregation kinetics in these settings must be improved [3].
Experimental research calls for a quantitative systems pharmacology (QSP) approach to address these knowledge gaps. We aim to develop a mechanistic model of aSyn aggregation to investigate in silico the disrupted proteostasis network underlying Parkinson's disease.
Methods:
Our model of protein aggregation relies on mechanistic insights from available chemical kinetic models tailored to ad-hoc experiments [4–7]. The system represents a nucleation-conversion polymerization process with toxic oligomers at the core and self-amplifying loops, including all the microscopic events explicitly related to aSyn in the literature [8]. We implemented the resulting ordinary differential equation system of mass action and catalytic reactions in MATLAB 2024a.
Upon global sensitivity analysis, we calibrated and validated the aggregation model on published data from in vitro aggregation assays obtained through smFRET and ThT assay techniques [4–6], adopting a stepwise approach. The fitting procedure relied on the covariance-based evolutionary strategy (CMA-ES) optimization algorithm to estimate unknown model parameters. To assess parameter reliability, we performed uncertainty quantification by generating virtual populations of the parameter estimates related to the best-in-fit. Furthermore, we carried out local-at-a-point and structural identifiability testing to analyze the robustness of the model structure and local sensitivity analysis to quantify the impact of each parameter on the system dynamics.
Results: Focusing on intraneuronal aSyn kinetics, we propose a mechanistic model of protein accumulation accounting for two distinct routes of aSyn aggregation, i.e., pure and lipid-mediated fibrillation, respectively. In agreement with the biological insights found in the literature, our model is flexible enough to capture multiple scenarios of aSyn accumulation that mimic physiologically relevant conditions, e.g., pH level variations and seeding and the effect of lipid vesicles [5, 6]. In addition, the model enables a broad spectrum of in silico experiments. For example, it allows us to explore the impact of risk factors such as aging and SNCA gene mutations through specific proxies in the model (e.g., lipid-to-protein ratio). Moreover, it provides insights into the interplay between pure aSyn and lipidic aggregation pathways. Finally, model predictions and results of local sensitivity analysis are compared and associated with reported effects of compounds currently under investigation, suggesting new candidate target mechanisms to counteract aggregation.
Conclusions:
Designed as an intracellular module of a more general QSP model of neurodegeneration, our model offers a robust mathematical framework for investigating disrupted aSyn homeostasis and can serve as a virtual lab for testing Parkinson's disease therapies.
References:
- C. R. Fields, N. Bengoa-Vergniory, R. Wade-Martins, Targeting alpha-synuclein as a therapy for Parkinson’s disease. Front Mol Neurosci 12 (2019).
- B. A. Hijaz, L. A. Volpicelli-Daley, Initiation and propagation of α-synuclein aggregation in the nervous system. Mol Neurodegener 15, 1–12 (2020).
- T. Sinnige, Molecular mechanisms of amyloid formation in living systems. Chem Sci 13, 7080–7097 (2022).
- M. Iljina, et al., Kinetic model of the aggregation of alpha-synuclein provides insights into prion-like spreading. Proc Natl Acad Sci U S A 113, E1206–E1215 (2016).
- R. Gaspar, et al., Secondary nucleation of monomers on fibril surface dominates α -synuclein aggregation and provides autocatalytic amyloid amplification. Q Rev Biophys 50, e6 (2017).
- C. Galvagnion, et al., Lipid vesicles trigger α-synuclein aggregation by stimulating primary nucleation. Nat Chem Biol 11, 229–234 (2015).
- A. J. Dear, et al., Molecular mechanism of α-synuclein aggregation on lipid membranes revealed. bioRxiv, 2023-10 (2023).
- E. Righetti, A. Antonello, L. Marchetti, E. Domenici, F. Reali, Mechanistic models of α-synuclein homeostasis for Parkinson’s disease: A blueprint for therapeutic intervention. Front Appl Math Stat 8 (2022).