Identification and Quantification of Components of In-Vitro Transcription (IVT) Reaction in a Mixture Using Process Raman

Significance of the topic

Ribonucleic acid (mRNA) therapeutics and vaccines rely on in vitro transcription (IVT) reactions that consume multiple costly ribonucleoside triphosphates (rNTPs), modified nucleotides, and capping agents. Real-time monitoring of these critical process parameters (CPPs) can improve yield, reduce by-products, and lower reagent costs. Process Raman spectroscopy offers molecular specificity, minimal interference from water, and non‑destructive in‑line measurement capabilities, making it a promising process analytical technology (PAT) for IVT monitoring and control.

Objectives and overview of the study

This feasibility (proof‑of‑concept) study aimed to determine whether process Raman spectroscopy can (1) identify individual IVT components within mixtures and (2) quantify those components accurately across a typical IVT concentration range (0–10 mM). Components studied included ATP, GTP, CTP, UTP, two modified nucleotides (2‑Thio‑UTP and 5‑Methyl‑CTP), and an anti‑reverse capping analogue (ARCA). The study also evaluated the detectability of other IVT constituents (enzymes, template DNA, buffer additives) and assessed the unique spectral signature of synthesized mRNA.

Methodology and experimental details

• Sample preparation: Individual analytes and designed mixtures were prepared in 25 mM tris‑acetate buffer (pH 7.9). Single‑component spectra were measured at 3 mM; DoE mixtures covered 0–10 mM for each component.
• Design of experiments (DoE): A uniform design algorithm generated 35 mixture samples that uniformly span the seven‑component concentration space to train chemometric models.
• Raman acquisition settings: Spectra recorded with probe at 450 mW laser power, 3000 ms integration, and 3 accumulations. Spectral region used for quantification: 480–3240 cm⁻¹.
• Spectral preprocessing: Baseline/baseline removal (Automatic Whittaker Smoothing), normalization to water band (3080–3230 cm⁻¹) or SNV, Savitzky–Golay derivative filtering (1st or 2nd derivative, window 13), and mean centering.
• Chemometrics: Principal component analysis (PCA) for visualization and discrimination; partial least squares (PLS) regression for quantification. Data split by Kennard–Stone (70% training, 30% test). Internal cross‑validation via Venetian blind CV. Model selection based on RMSEC/RMSECV convergence and RMSEC:RMSECV ratio, avoiding overfitting.

Used instrumentation

• Thermo Scientific MarqMetrix All‑In‑One Process Raman Analyzer with a micro‑immersible probe for microvolume sampling.
• Software and chemometrics: SOLO 9.3.1 (Eigenvector Research) used for PCA and PLS model development and validation.

Main results and discussion

• Identification: Distinct Raman features were observed for ATP, GTP, CTP, UTP, ARCA, 2‑Me‑CTP, and 5‑Thio‑UTP enabling clear clustering in PCA score space. DTT and spermidine also produced detectable signatures; several protein/enzymatic constituents (T7 polymerase, pyrophosphatase, RNase inhibitor, DNA template) produced Raman signals below detection limits at the tested concentrations.
• Quantification: Individual PLS models (4–5 latent variables) were developed for each target analyte using the DoE training set and validated against independent test samples. Key performance metrics across analytes:
  • RMSEP range approximately 0.16–0.44 mM.
  • RMSECV and RMSEC values were low and comparable, indicating robust calibration and limited overfitting.
  • Prediction R² values between ~0.98 and 0.996, demonstrating high linearity and predictive power over 0–10 mM.
• Variable importance: VIP score analysis identified analyte‑specific Raman shifts (VIP > 1) that drove each PLS model, supporting model specificity rather than nonspecific correlation among components.
• mRNA detection: Raman spectrum of synthesized mRNA (2 mg/mL) showed a distinguishing O–P–O vibration band near ~814 cm⁻¹ that is not present in monomer spectra, indicating feasibility for monitoring net mRNA accumulation during IVT.
• Limitations observed: Enzymes and low‑abundance components produced signals below instrument detection at process concentrations; the presented models are built on static mixture samples rather than dynamic in‑process data, so model robustness under reaction conditions (temperature changes, turbidity, reaction kinetics) remains to be demonstrated.

Benefits and practical applications

• Real‑time, in‑line monitoring: Raman offers rapid, reagent‑free measurements compatible with aqueous IVT buffers, enabling continuous tracking of CPPs (rNTPs, modified nucleotides, capping reagents) to inform process control decisions.
• Process optimization and cost savings: Accurate quantification of reagent depletion and mRNA formation can guide feed strategies, improve yield, and limit waste of expensive modified nucleotides and caps.
• Quality by design and PAT integration: Raman‑based chemometrics can be integrated into manufacturing control strategies to maintain consistent product quality and reduce risks of by‑product formation such as dsRNA.

Future trends and potential applications

• Model reinforcement with process data: Acquiring real‑time Raman measurements across multiple IVT runs (varying templates, polymerase loads, temperatures) will improve model robustness and allow dynamic calibration to reaction kinetics.
• Extension to impurities and by‑products: With richer data, models can be expanded to detect and quantify undesirable species (e.g., dsRNA) and correlate their formation to process conditions for proactive mitigation.
• Hybrid PAT strategies: Combining Raman with orthogonal sensors (UV absorbance, HPLC offline fingerprinting, real‑time pyrophosphate monitoring) will enhance confidence in control decisions and regulatory acceptance.
• Digital control loops: Implementing Raman‑driven feedback loops (e.g., reagent feed, reaction quench) to maintain optimal rNTP ratios and capping efficiency in continuous or semi‑continuous IVT formats.

Conclusion

This feasibility study demonstrates that process Raman spectroscopy, paired with appropriate chemometric models, can identify and quantify key IVT reagents (rNTPs, modified nucleotides, and ARCA) in multi‑component mixtures with good accuracy across the 0–10 mM range. The unique Raman band of mRNA (~814 cm⁻¹) further supports feasibility for direct monitoring of product formation. To transition from proof‑of‑concept to a validated in‑line PAT, the next steps are collection of dynamic in‑process Raman data, model reinforcement, assessment of matrix and process interferences, and demonstration of real‑time control benefits in production‑relevant settings.

References

1. Kwon H., Kim M., Seo Y., Moon Y. S., Lee H. J., Lee K., Lee H. Emergence of Synthetic mRNA: In Vitro Synthesis of mRNA and Its Applications in Regenerative Medicine. Biomaterials. 2018;156:172–193. doi:10.1016/j.biomaterials.2017.11.034.
2. Mullard A. FDA Approves mRNA‑Based RSV Vaccine. Nat Rev Drug Discov. 2024. doi:10.1038/d41573-024-00095-3.
3. Beck J. D., Reidenbach D., Salomon N., Sahin U., Türeci Ö., Vormehr M., Kranz L. M. mRNA Therapeutics in Cancer Immunotherapy. Mol Cancer. 2021;20(1):69. doi:10.1186/s12943-021-01348-0.
4. Liu C., Shi Q., Huang X., Koo S., Kong N., Tao W. mRNA‑Based Cancer Therapeutics. Nat Rev Cancer. 2023;23(8):526–543. doi:10.1038/s41568-023-00586-2.
5. Yamamoto A., Kormann M., Rosenecker J., Rudolph C. Current Prospects for mRNA Gene Delivery. Eur J Pharm Biopharm. 2009;71(3):484–489. doi:10.1016/j.ejpb.2008.09.016.
6. Abu‑Absi N. R., Kenty B. M., Cuellar M. E., Borys M. C., Sakhamuri S., Strachan D. J., Hausladen M. C., Li Z. J. Real‑Time Monitoring of Multiple Parameters in Mammalian Cell Culture Bioreactors Using an In‑Line Raman Spectroscopy Probe. Biotechnol Bioeng. 2011;108(5):1215–1221. doi:10.1002/bit.23023.
7. Esmonde‑White K. A., Cuellar M., Lewis I. R. The Role of Raman Spectroscopy in Biopharmaceuticals from Development to Manufacturing. Anal Bioanal Chem. 2022;414(2):969–991. doi:10.1007/s00216-021-03727-4.
8. Esmonde‑White K. A., Cuellar M., Uerpmann C., Lenain B., Lewis I. R. Raman Spectroscopy as a Process Analytical Technology for Pharmaceutical Manufacturing and Bioprocessing. Anal Bioanal Chem. 2017;409(3):637–649. doi:10.1007/s00216-016-9824-1.
9. Buckley K., Ryder A. G. Applications of Raman Spectroscopy in Biopharmaceutical Manufacturing: A Short Review. Appl Spectrosc. 2017;71(6):1085–1116. doi:10.1177/0003702817703270.
10. Lee K. H., Song J., Kim S., Han S. R., Lee S.‑W. Real‑Time Monitoring Strategies for Optimization of in Vitro Transcription and Quality Control of RNA. Front Mol Biosci. 2023;10. doi:10.3389/fmolb.2023.1229246.
11. Fang K.‑T., Lin D. K. J., Winker P., Zhang Y. Uniform Design: Theory and Application. Technometrics. 2000;42(3):237–248. doi:10.1080/00401706.2000.10486045.
12. Zhang L., Liang Y.‑Z., Jiang J.‑H., Yu R.‑Q., Fang K.‑T. Uniform Design Applied to Nonlinear Multivariate Calibration by ANN. Anal Chim Acta. 1998;370(1):65–77. doi:10.1016/S0003-2670(98)00256-6.
13. Stover N. M., Ganko K., Braatz R. D. Mechanistic Modeling of in Vitro Transcription Incorporating Effects of Magnesium Pyrophosphate Crystallization. Biotechnol Bioeng. (in press). doi:10.1002/bit.28699.