BioPharmaceutical approach with spectroscopy

Significance of the topic

This compendium reviews contemporary uses of optical spectroscopy (FTIR/ATR, NIR, Raman, and UV-Vis) as analytical tools across biopharmaceutical workflows. Spectroscopic PAT (process analytical technology) methods deliver rapid, non-destructive, and often reagent-free monitoring of raw materials, in-process streams, and final products. Their complementary strengths—molecular specificity (Raman), structural sensitivity (FTIR), bulk quantitation and high sensitivity in the deep UV (UV-Vis/A205), and deep-penetration process monitoring (NIR/FT-NIR)—address critical needs for real-time control, quality assurance, and method transferability in regulated biologics manufacturing.

Goals and overview of the studies

The collection comprises application notes and feasibility studies demonstrating how spectroscopy can be applied to:

• Elucidate protein secondary structure (FTIR/ATR).
• Automate glucose feeding in fed-batch bioreactors using in-line process Raman and chemometrics.
• Transfer Raman calibrations between instruments and strategies to standardize models.
• Predict protein concentration in cell cultures using FT‑NIR.
• Detect protein aggregation and quantify non‑aggregated protein by UV‑Vis with scattering correction and integrating sphere methods.
• Assess DNA purity for cloning QC with microvolume UV-Vis.
• Quantify peptides/proteins at A205 in microvolume spectrophotometers.
• Monitor plasmon resonance shifts after nanoparticle bioconjugation with microvolume UV-Vis.
• Enable real-time release and multi‑attribute testing with Raman for final product identity and excipient quantitation in downstream buffer workflows.

Methodology

The compendium presents a range of experimental approaches and chemometric strategies:

• FTIR/ATR for amide I/II analysis, spectral deconvolution, second-derivative enhancement, and database-based secondary-structure prediction (PROTA-3S/OMNIC).
• Process Raman with fiber/flow probes integrated in-line for continuous measurement; data preprocessing (Savitzky–Golay derivatives, SNV, mean centering); PLS regression for glucose and lactate models; LOOCV for model selection; software integration for automated feedback control of pumps.
• Method-transfer strategies between Raman instruments including direct transfer, global (multi-instrument) calibrations, and correction/standardization approaches; emphasis on careful spectral-region selection and preprocessing to mitigate peak shifts and intensity differences.
• FT‑NIR transflectance with adjustable pathlength for protein concentration prediction in cell culture matrices using PLS models built on first-derivative spectra; validation with RMSEP/RMSECV metrics.
• UV‑Vis absorption for aggregate detection via scattering artifacts; correction by baseline subtraction, parametric scattering models (λ⁻4 dependence), or integrating sphere (Kubelka–Munk transformation) to quantify free (non‑aggregated) protein using a reference standard and Beer’s law equivalence.
• Microvolume UV‑Vis A205 quantitation methods: fixed ε205=31, Scopes correction (using A280/A205 ratio), and sequence-specific ε205 (Anthis & Clore) for peptides/proteins with/without aromatic residues.
• Raman-based downstream UF/DF monitoring: Uniform Design for calibration mixture selection, spectral standardization for inter-instrument model transfer, PLS for excipient (histidine, arginine, sucrose) quantitation, and Q-residual/Hotelling T² statistics for buffer quality assessment (detecting sucrose hydrolysis to glucose/fructose).

Used instrumentation

Key instruments and accessories reported across the studies:

• Thermo Scientific Nicolet iS10 / iS50 / iS20 FTIR spectrometers with ConcentratIR2 diamond ATR, MCT or DTGS detectors, and BioCell CaF2 transmission cells.
• Thermo Scientific Antaris MX FT‑NIR Process Analyzer with transflectance probe for protein-in-culture monitoring.
• Thermo Scientific MarqMetrix All‑In‑One Process Raman Analyzer with BallProbe and FlowCell sampling optics for in‑line bioreactor and UF/DF monitoring; MarqMetrix software for data handling.
• Thermo Scientific DXR3 SmartRaman+ and SmartRaman spectrometers (method-transfer studies) with ASA accessory and TQ Analyst software.
• Thermo Scientific NanoDrop One / One Ultra / Eight microvolume UV‑Vis spectrophotometers and Evolution UV‑Vis with Evolution ISA integrating sphere for aggregation and nanoparticle SPR measurements.
• Evo/benchtop UV‑Vis and HPLC used as orthogonal references for model validation.

Main results and discussion

Highlights and quantitative outcomes:

• Protein secondary structure: FTIR (transmission and ATR) resolved amide I components via second-derivative and peak deconvolution to produce secondary-structure estimates (e.g., BSA deconvolution giving α‑helix/β‑sheet/random fractions consistent with X‑ray trends). ATR enabled analysis at low sample amounts by drying onto diamond crystal.
• Automated glucose feeding: Process Raman PLS models (selected spectral regions) predicted glucose with RMSECV ∼0.49 g/L and lactate RMSECV ∼0.31 g/L across 0–12 g/L ranges. Integrated PAT software controlled once‑daily bolus feeding; automated runs produced comparable lactate and titer profiles versus manual control, demonstrating feasible human‑free glucose control with RMSEP ~0.45 g/L for glucose.
• Method transfer: Direct transfer success depended on spectral region and preprocessing; global multi‑instrument calibration produced smallest prediction bias but required the most data; correction/standardization methods (limited secondary-instrument spectra) improved direct transfer with reduced effort. Peak shifts (few cm⁻1) and intensity differences can induce large quantitative errors, underlining need for standardization and robustness.
• FT‑NIR protein concentration in cell culture: Antaris MX PLS model over 0.16–5.0 g/L yielded calibration R² ≈0.977 and validation RMSEP ≈0.31 g/L (RMSECV ≈0.51 g/L), demonstrating reliable prediction of target protein concentrations in live culture matrices.
• Protein aggregation by UV‑Vis: Aggregation produced broadband scattering artifacts increasing apparent absorbance. For low scatter, parametric subtraction (λ⁻4 Rayleigh-like fit) can recover absorbance; for turbid samples, integrating-sphere Kubelka–Munk measurements allowed accurate estimation of free protein by comparison to non‑aggregated control and confirmed results after filtration. Example: heated BGG showed ~0.20 mg/mL free protein after aggregation, consistent between integrating sphere and filtrate A measurement.
• DNA purity for cloning QC: Microvolume UV‑Vis (NanoDrop Lite Plus) rapidly detected contaminants (phenol, EDTA) via A260/A280 and A260/A230; contaminated samples impaired restriction digestion (EDTA inhibited enzyme activity), supporting QC checkpoints pre/post digestion and showing microvolume spectrophotometers are practical tools for cloning workflows.
• A205 quantitation: Microvolume A205 methods (ε205=31 or Scopes correction) provided consistent concentration estimates versus benchtop UV‑Vis for peptides without aromatics (polymyxin) and for proteins, but choice of extinction coefficient/method matters when proteins contain Trp/Tyr; NanoDrop One showed high repeatability and agreement with cuvette instruments.
• Nanoshell SPR shifts: NanoDrop One microvolume UV‑Vis detected reproducible red shifts (∼9 nm in example) upon siRNA and PEG conjugation to gold nanoshells (150 nm), confirming surface modification with 1–2 μL samples and no dilution.
• Downstream multi‑attribute Raman: DXR3 Raman + chemometrics discriminated 15 drug-product classes (identity) and a PLS approach quantified two preservatives in drug vials with RMSE values comparable to HPLC; MarqMetrix Raman monitored excipient concentrations during UF/DF with <5% error versus HPLC and detected buffer instability (sucrose hydrolysis) early via Q-residual trending.

Benefits and practical applications

Practical advantages demonstrated:

• Real‑time, in‑line or at‑line monitoring enabling closed‑loop control (e.g., automated glucose feeding, UF/DF excipient control) and reduced reliance on offline laboratory assays.
• Conservation of precious samples through microvolume UV‑Vis and minimized sample prep for ATR/flow‑probe measurements.
• Fewer manual interventions → reduced human error, improved batch-to-batch reproducibility, faster decision-making and potential cost savings.
• Raman and FTIR provide complementary information: Raman excels for process PAT and multi‑attribute testing; FTIR (amide I) is a sensitive probe of secondary structure; UV‑Vis A205 offers sensitive microvolume protein quantitation; integrating sphere enables accurate aggregate analysis.

Future trends and potential applications

Emerging directions and opportunities:

• Wider adoption of spectroscopic PAT for regulatory accepted real‑time release testing (RTRT) and Quality by Design (QbD) workflows as chemometric models mature and validation strategies standardize.
• Improved model transfer and standardization frameworks (instrument correction libraries, universal preprocessing pipelines, spectral standard materials) to enable multi-site deployment and method harmonization.
• Integration of multimodal sensors (Raman, NIR, UV‑Vis) and soft‑sensors (model ensembles) to increase robustness for complex matrices and to resolve confounding events (e.g., excipient degradation, aggregation).
• Automation of downstream unit operations (UF/DF, buffer exchange, fill‑finish) using spectroscopic feedback for tighter control of CQAs and reduced empirical over‑processing.
• Application of real‑time spectral quality metrics (Q‑residuals, Hotelling T²) for buffer and reagent QC to prevent downstream failures and enable on‑the‑fly corrective actions.
• Extension to nanoparticle therapeutics and conjugates where microvolume SPR/UV‑Vis measurements can speed conjugation QC and release testing.

Conclusion

The compendium demonstrates that vibrational spectroscopy (FTIR, NIR, Raman) together with UV‑Vis microvolume techniques provide effective, complementary analytical capabilities across biopharmaceutical development and manufacturing. When paired with robust chemometrics and appropriate validation, these tools support real‑time monitoring, automation, method transfer, and potential real‑time release strategies—improving process understanding, product quality, and operational efficiency in regulated biologics manufacturing.

References

1. Elliott A., Ambrose E.J. Structure of synthetic polypeptides. Nature. 1950;165:921–922.
2. Jackson M., Mantsch H.H. The use and misuse of FTIR spectroscopy in the determination of protein structure. Crit Rev Biochem Mol Biol. 1995;30:95–120.
3. Barth A. Infrared spectroscopy of proteins. Biochim Biophys Acta. 2007;1767:1073–1101.
4. Byler D.M., Susi H. Examination of the secondary structure of proteins by deconvolved FTIR spectra. Biopolymers. 1986;25:469–487.
5. Surewicz W.K., Mantsch H.H. New insight into protein secondary structure from resolution-enhanced infrared spectra. Biochim Biophys Acta. 1988;952:115–130.
6. Sukumaran S., Hauser K., Maier E., et al. Tracking unfolding and refolding pathways of outer membrane protein porin. Biochemistry. 2006;45:3972–3980.
7. Klose D., Janes R.W. 2Struc – the protein secondary structure analysis server. Biophys J. 2010;98:454–455.
8. Villa J., Zustiak M., Kuntz D., et al. Use of Lykos and TruBio software for automated feedback control to monitor and maintain glucose concentrations in real time (Thermo Fisher application notes and conference reports).
9. Yang J.-D., Lu C., Stasny B., et al. Fed-Batch Bioreactor Process Scale-up from 3-L to 2,500-L Scale for Monoclonal Antibody Production. Biotechnol Bioeng. 2007;98(1):141–154.
10. Bellon-Maurel V., Fernandez-Ahumada E., Palagos B., Roger J.-M., McBratney A. Critical review of chemometric indicators commonly used for assessing the quality of prediction by NIR spectroscopy. TrAC Trends Anal Chem. 2010;29(9):1073–1081.
11. Workman J. A review of calibration transfer practices and instrument differences in spectroscopy. Appl Spectrosc. 2018;73(3):340–365.
12. Kunz M.R., Kalivas J.H., Andries E. Model updating and calibration transfer using Tikhonov regularization variants. Anal Chem. 2010;82:3642–3649.
13. Nolasco M., Pleitt K., Khadka N. Using a Process Raman Analyzer as an In-Line Tool for Accurate Protein Quantification in Downstream Processes (Thermo Fisher application note).
14. Pignataro M.F., Herrera M.G., Dodero V.I. Protein aggregation monitoring and implications for formulation and processing. Molecules. 2020;25:4854.
15. Anthis N.J., Clore G.M. Sequence-specific determination of protein and peptide concentrations by absorbance at 205 nm. Protein Sci. 2013;22:851–858.
16. Oldenburg S.J., Averitt R.D., Westcott S.L., Halas N.J. Nanoengineering of optical resonances. Chem Phys Lett. 1998;288:243–247.
17. Agrawal P., Wilkstein K., Guinn E., et al. A review of tangential flow filtration: process development and pharmaceutical applications. Org Process Res Dev. 2023;27(4):571–591.