What question did this study set out to answer?

The aim is to develop a mechanistic IVT model to optimize mRNA production while maintaining quality and reducing costs.

March 21, 2026Open Access

Enhancing mRNA Therapeutics Production: A Platform Technology Approach Through IVT Modeling Insights

Key Points

The aim is to develop a mechanistic IVT model to optimize mRNA production while maintaining quality and reducing costs.
Utilized a Latin Hypercube design with over 100 experiments to examine IVT conditions.
Measured the effects of NTP concentration, magnesium concentration, T7 RNA polymerase, and template DNA on RNA yield.
Validated the model against empirical data for two mRNA templates: CSP and eGFP.
Established a Quality-by-Design (QbD) framework linking process parameters to quality metrics.
The model predicts RNA yield with R2 values between 0.85 and 0.87.
Key critical process parameters (CPPs) driving RNA yield variability were identified.
The QbD framework proposed an operating window that reduces reagent costs while adhering to quality standards.

Abstract

• Generalizable tool enabling cost-efficient, large-scale mRNA manufacturing • In-house LHS-validated IVT model (100 experiments) for CSP & eGFP (R 2 = 0.87) • Integrates Mg–PPi precipitation, NTP consumption, spermidine & template effects • QbD framework mapping CPPs/CMAs to CQAs for scalable mRNA design spaces • Global Sobol sensitivity reveals key CPPs driving RNA yield variability Rapid, scalable production of high-quality mRNA underpins vaccines, gene therapies, and protein replacement medicines, yet optimisation of the in vitro transcription (IVT) step is often driven by template-specific trial-and-error. We present a first-principles mechanistic IVT bioprocess model—constructed from thermodynamic equilibria, enzyme kinetics, and stoichiometric mass balances—to support Quality-by-Design (QbD) manufacturing. The model is calibrated and validated against > 100 experiments from a Latin Hypercube design that varies nucleotide triphosphate (NTP) concentration, magnesium concentration, T7 RNA polymerase loading, template DNA, spermidine, and reaction time, using two contrasting templates: a SARS-CoV-2 spike construct (CSP) and enhanced green fluorescent protein (eGFP). The model predicts RNA yield with R 2 = 0.85 –0.87 while explicitly tracking individual NTP consumption and magnesium pyrophosphate precipitation. Within a QbD context, the framework links critical process parameters (CPPs) and critical material attributes (CMAs) to product-oriented metrics and manufacturing key performance indicators (KPIs), and identifies a model-derived operating window that reduces reagent cost subject to specified quality constraints. Mechanistic refinements, including inorganic pyrophosphatase kinetics and sequence-aware per-base NTP accounting, support transfer between the two templates studied here and provide a foundation for broader multi-construct validation. Overall, the approach offers a data-efficient route to accelerate IVT process development and reduce consumable spend for emerging mRNA therapeutics.

Read Full Paperexternally

AIに質問

Bookmark

View Full Paper