Case Study
Predicting 3D Culture Success from 2D Data
Minaris Advanced Therapies
Condition Screening
Cell Therapy
Summary
This work aimed to predict, at an early stage, which Mesenchymal Stem Cell lines could successfully scale using bead-to-bead (BtB) transfer, identify optimal transfer timing, and reduce time-consuming feasibility studies in 3D bioreactor scale-up.
Auxilart applied mechanistic modeling to translate limited 2D passage data into a predictive 3D model, explicitly capturing cell growth dynamics and BtB transfer behavior using minimal experimental input.
The simulation enabled forward-looking forecasts of transfer timing and culture readiness, indicating the potential for an estimated 38% reduction in labor, tighter scheduling, and more efficient MSC scale-up operations.
The Challenge
Minaris Advanced Therapies, a leading CDMO in cell and gene therapy, needed to scale up Mesenchymal Stem Cell (MSC) production using 3D microcarrier bioreactors.
Although BtB cell transfer* offers a practical path to scale, not all MSC lines adapt well, forcing long feasibility studies.
The conventional workflow, cell banks to 2D culture, then small-scale 3D trials before moving into production bioreactors, stretched over weeks to months and consumed substantial resources. The critical need was to predict, at an early stage, which cell lines could succeed, determine the optimal BtB transfer timing, and reduce wasted effort.
* BtB cell transfer: A scale-up method where MSCs migrate and adhere from confluent microcarriers onto fresh ones in the bioreactor.
Rather than relying on extensive experimental screening, modeling was used to evaluate a wide range of operating variables virtually, including temperature, mixing, and media conditions. This reduced experimental burden and allowed resources to be focused on the most relevant operating windows.
Approach & Technology Deployed
To cut down uncertainty and reduce the need for trial-and-error, Minaris and Auxilart applied mechanistic modeling, a physics- and biology-based approach that predicts outcomes with minimal data.
Principle-based, not black-box: Unlike purely data-driven models, mechanistic models embed biological knowledge, enabling reliable predictions even with limited experimental input.
Built from real experiments: A 2D cultivation model was calibrated using passage data, capturing cell growth dynamics such as mass balance, adaptation (lag) phase, and specific growth rate.
Extended to 3D scale-up: These parameters were then translated into a 3D mechanistic model, explicitly incorporating BtB transfer steps to simulate bioreactor scale-up.
This approach provided predictive power early in development, cutting back unnecessary feasibility trials and focusing effort on MSC lines most likely to succeed.
Results
The mechanistic modeling enabled the creation of a 3D predictive model capable of transforming simple 2D passage data into accurate forecasts of cell growth. By incorporating biological principles rather than relying solely on historical data, the model successfully reproduced experimental outcomes and precisely captured BtB transfer timing and its downstream effects.
In plots below, the red dots represent experimental measurements of cell density over time, while the blue curves show the model predictions. The close alignment between the two demonstrates the model’s reliability. Importantly, it accurately predicted cell density at transfer points, confirming its practical value for process planning and scheduling.
Impact
Previously, unpredictable lot-to-lot variability meant passaging days could not be planned with confidence, forcing broad resource allocation and standby time. With Auxilart’s predictive simulation, Minaris gained the ability to forecast when each culture was expected to be ready for transfer, indicating the potential for:
An estimated 38% reduction in labor requirements
Tighter scheduling, less downtime
Improved operational efficiency and resource use
By turning uncertainty into forward-looking insight, mechanistic modeling demonstrated a clear path to faster, more precise, and more cost-effective MSC scale-up.
