Anin vitro- agent-based modelling approach to optimisation of culture medium for generating muscle cells

Abstract

Many of the protocols in contemporary tissue engineering remain insufficiently optimised. Methodologies for culturing the complex structures of muscle tissue are particularly lacking, both in terms of quality and quantity of mature cells. Here, we analyse images from in vitro experimentation to quantify the effects of the composition of culture media on mouse-derived myoblast behaviour and myotube cell quality. We then apply computational modelling to predict the optimum range of media compositions for culturing. We define metrics of uniformity of myonuclei distribution as an early indicator of cell quality and difference in myonuclei density over time as an indicator of cell quantity. Analysis of live and static images of muscle cell differentiation revealed that changes in culture media result in significant changes in indicators of cell quantity and quality as well as changes in myoblast migratory behaviour. By describing media composition as a set of functions of cell behaviour we designed a model for predicting cell quality. Cell behaviours were taken directly from experimental images or inferred using Approximate Bayesian Computation and applied as inputs to an agent-based model of cell differentiation with cell quality indicators as outputs. Our results suggest that culturing muscle cells in a neural cell differentiation medium does not diminish cell quality. We show that, while high concentrations of serum are detrimental to cell development, increasing serum concentration raises the total amount of myoblast fusion, leading to a trade-off between the quantity and quality of cells produced when choosing a culture medium. Our numerical results provided a good prediction of experimental results for media with 5% serum provided the background cell proliferation rate was known.

Generating fast-twitch myotubes in vitro with an optogenetic-based, quantitative contractility assay

“…Because contraction depends on muscle architecture, we sought to understand if changes in cellular structures underlie the faster contraction cycles observed after long-term in vitro training. We performed a quantitative structural analysis by extracting parameters of myotube maturation from immunofluorescent images at day 4 of differentiation ( Fig 3A and B ; Hardman et al, 2021 Preprint ). Trained myotubes display an increased cell width ( Fig 3C ), suggesting an addition of sarcomeres because of training.…”

Generating fast-twitch myotubes in vitro using an optogenetic-based, quantitative contractility assay

“…Since contraction depends on muscle architecture, we sought to understand if changes in cellular structures underlie the faster contraction cycles observed after long-term in vitro training. We performed a quantitative structural analysis by extracting parameters of myotube maturation from immunofluorescent images at day 4 of differentiation ( figure 3A, B ; Hardman et al, 2021). Trained myotubes display an increased cell width ( figure 3C ), suggesting an addition of sarcomeres due to training.…”

Aninvitroagent-based modelling approach to optimization of culture medium for generating muscle cells

“…Data accessibility. Imaging data used in this study is available and can be accessed from the datashare repository: https://datashare.ed.ac.uk/ handle/10283/4404 [25]. Metrics extracted from image data are provided either in the main text or in the electronic supplementary material [26].…”

“…Imaging data used in this study is available and can be accessed from the datashare repository: https://datashare.ed.ac.uk/ handle/10283/4404 [25]. Metrics extracted from image data are provided either in the main text or in the electronic supplementary material [26]. The ImageJ macros for pre-processing images, Matlab code for image analysis, the Netlogo code for the ABM and Python code for linking the ABM with pyABC are available from the Zenodo repository: https://zenodo.org/records/10300311 [27].…”