BioPharmaceutical approach with spectroscopy

Significance of the topic

Biopharmaceutical manufacturing increasingly relies on rapid, non‑destructive spectroscopic techniques (FTIR, NIR, Raman, UV‑Vis) coupled with chemometrics to enable process analytical technology (PAT), raw material control, in‑process monitoring and real‑time release strategies. These spectroscopies provide complementary molecular information (secondary structure, concentrations of proteins/excipients, aggregation state, impurity detection and surface modification of nanoparticles) while reducing laboratory turnaround, sample volume and operator dependence. The compendium documents practical implementations, analytical figures of merit and workflow integration examples that are directly relevant to upstream, downstream and fill‑finish stages of biologics production.

Objectives and study overview

The collected application notes present multiple goals: (i) demonstrate FTIR (transmission and ATR) for protein secondary‑structure analysis and deconvolution; (ii) show in‑line/process Raman for automated glucose feeding and for downstream buffer/excipient monitoring; (iii) describe method transfer approaches between Raman instruments; (iv) validate FT‑NIR for predicting target protein concentration in cell culture; (v) use UV‑Vis and integrating‑sphere measurements to detect and quantify protein aggregation; (vi) illustrate NanoDrop UV‑Vis workflows for DNA purity/QC, protein (A205) quantitation and nanoparticle SPR monitoring; and (vii) present Raman‑based multi‑attribute testing enabling potential real‑time release testing of biologics.

Methods and instrumentation

The studies combine bench and process spectrometers with chemometric model building (primarily PLS) and validation approaches (LOOCV, RMSECV/RMSEP):

• FTIR: Transmission FTIR (BioCell CaF2) and ATR (ConcentratIR2 diamond ATR) on Nicolet iS10/iS50 spectrometers; data processed with OMNIC and PROTA‑3S for amide‑band deconvolution and secondary‑structure prediction.
• NIR: Thermo Scientific Antaris MX FT‑NIR Process Analyzer with transflectance probe for in‑process protein concentration prediction.
• Raman (upstream/downstream/PAT): Thermo Scientific MarqMetrix All‑In‑One Process Raman Analyzer with BallProbe or FlowCell sampling optics; Thermo Scientific DXR3 SmartRaman+ for lab/final product testing; chemometrics using PLS (SOLO, Python, TQ Analyst, MarqMetrix software).
• UV‑Vis: Thermo Scientific NanoDrop One/Ultra/ Eight and Evolution UV‑Vis spectrophotometers; Evolution ISA‑220 integrating sphere used to collect scatter‑free Kubelka–Munk signals for aggregate quantitation.
• Auxiliary: LC‑MS/HPLC and biochemical reference methods (e.g., BioProfile) used as orthogonal references in model development and validation.

Main results and discussion

FTIR protein secondary structure

• Transmission and ATR FTIR allowed identification of amide I/II features and spectral deconvolution (second derivative + peak fitting). Examples: cytochrome C, concanavalin A and BSA showed FTIR‑predicted secondary structure broadly consistent with X‑ray data (e.g., cytochrome C ~45% α‑helix by FTIR).

NIR protein concentration in cell culture (Antaris MX)

• PLS model for target protein (albumin as surrogate) across 0.16–5.0 g/L delivered calibration R = 0.977, RMSEC ≈ 0.33 g/L, RMSEP ≈ 0.31 g/L and RMSECV ≈ 0.51 g/L — practical prediction error ≈ 0.5 g/L or better for live cultures.

Process Raman for automated glucose feed (MarqMetrix)

• PLS glucose model (regions: 1065–1232, 1595–1863, 2704–3078 cm−1) used five latent variables; RMSEC ≈ 0.43 g/L, RMSECV ≈ 0.49 g/L and RMSEP ≈ 0.45 g/L, R²CV ≈ 0.94. Lactate model RMSEP ≈ 0.24–0.31 g/L, R² ≈ 0.92–0.96. Real‑time Raman predictions enabled automated once‑per‑day bolus feed control, maintaining glucose within target bounds and producing comparable titers and lactate profiles to manual control.

Raman method transfer between instruments

• Method‑transfer strategies compared: direct transfer, global (full calibration using spectra from all instruments) and correction/standardization approaches. Example with acetaminophen showed that choosing robust spectral regions (e.g., C–H stretch) and using correction spectra or global calibration reduces inter‑instrument bias; global methods gave best accuracy but require the most secondary‑instrument data.

UV‑Vis detection and quantitation of protein aggregates

• Aggregated BGG samples scatter light and produce baseline elevation and apparent absorption across UV region. For low scattering, a fitted scattering function (∝λ−4) subtracted from spectra can recover free protein signal. For turbid samples, integrating sphere (Kubelka–Munk F(R) representation) allows scatter‑robust quantitation; filtered samples and corresponding Kubelka–Munk spectra gave consistent free‑protein quantitation (example: filtered BGG concentration 0.20 mg/mL).

DNA purity QC for molecular cloning (NanoDrop Lite Plus)

• NanoDrop microvolume UV‑Vis quantification and purity ratios (A260/A280, A260/A230) rapidly identified phenol and EDTA contamination; contaminants affected restriction enzyme digestion (EDTA inhibited HindIII cleavage). Gel extraction followed by NanoDrop measurement assessed percent recovery and residual contaminants.

Protein quantitation at A205 (NanoDrop One/Ultra)

• A205 microvolume quantification offers higher sensitivity than A280. The platform offers three options for extinction coefficients: fixed ε205=31 mL·mg−1·cm−1 (good for peptides lacking aromatic residues), Scopes method (corrects via A280/A205 for Trp/Tyr content), and sequence‑specific ε205 (Anthis & Clore). NanoDrop One results matched conventional cuvette spectrophotometers across a wide dynamic range.

Gold nanoshell plasmon resonance (NanoDrop One)

• NanoDrop One (2 µL, auto pathlength) detected SPR red shifts after siRNA and mPEG‑SH conjugation to gold nanoshells (example: peak shift ~795 → ~804 nm at 100 pM), consistent with conventional UV‑Vis and orthogonal loading assays.

NanoDrop Eight with Acclaro sample intelligence

• Acclaro technology resolves low‑level contaminating nucleic acids by spectral deconvolution, reporting contaminant identity, absorbance contribution and corrected nucleic acid concentration. Tests with mouse and MCF‑7 DNA/RNA mixtures showed corrected concentrations close to theoretical values.

Raman for final product identity and multi‑attribute testing (DXR3 SmartRaman)

• Feasibility work demonstrated discriminant analysis for differentiating 15 biologic drug products in native vials and PLS quantitation of two preservatives (R² ≈ 0.998–0.999; RMSEP small), illustrating potential for at‑line identity screening and excipient quantitation to support RTRT.

Process Raman for downstream buffer and excipient workflow (MarqMetrix All‑In‑One)

• PLS models for histidine (RMSEC ≈ 0.20 mg/mL), arginine (RMSEC ≈ 0.90 mg/mL) and sucrose (RMSEC ≈ 0.44 mg/mL) were transferable across instruments. In a bench‑scale UF/DF run, Raman predictions matched HPLC references within ≲5% error. Raman also detected buffer degradation: sucrose hydrolysis to glucose/fructose over 15 days (concurrent pH drop and osmolarity increase) indicated buffer quality failure via elevated Q residuals in model space.

Benefits and practical applications

Key practical takeaways from the compendium:

• Reduced turnaround: in‑line/at‑line spectroscopies provide near real‑time data that reduce dependency on offline assays and enable faster decisions.
• Low sample volume and non‑destructive assays: microvolume UV‑Vis (NanoDrop) and fiber‑optic Raman probes conserve valuable samples and enable closed‑loop process monitoring.
• PAT enablement and automation: Raman models can automate routine operations (e.g., glucose feeding, UF/DF endpoint decisions), improving reproducibility and reducing operator errors.
• Complementary information: combining FTIR, NIR, Raman and UV‑Vis expands analytical coverage (structural, quantitative, stability and surface chemistry assessments) across bioprocess stages.
• Method transfer and robustness: practical guidance shows that careful spectral region selection, preprocessing and use of correction or global transfer strategies improve cross‑instrument model portability.

Future trends and potential applications

Anticipated developments and opportunities include:

• Integration of multimodal PAT platforms (Raman + NIR + UV‑Vis + process sensors) with advanced chemometrics and machine learning for predictive control and RTRT workflows.
• Wider adoption of standardized method‑transfer protocols and standardized reference materials (NIST SRMs) to facilitate model portability across sites and instrument generations.
• Edge computing and closed‑loop control: embedding validated spectral models in process control systems (e.g., DCS, DeltaV, TruBio) to implement autonomous feed and buffer exchange strategies.
• Expansion of Raman/NIR for comprehensive downstream monitoring (protein aggregation, excipient retention, osmolality surrogates) and QC at fill‑finish (in‑vial identity testing without sample removal).
• Regulatory acceptance path: further validation case studies demonstrating equivalence to orthogonal assays (HPLC, LC‑MS, SEC) will accelerate regulatory confidence for RTRT use.

Conclusions

The compendium demonstrates that modern vibrational and UV‑Vis spectroscopies—combined with rigorous chemometric model building, validation and appropriate instrument standardization—are practical, high‑value tools for biopharmaceutical workflows. Key outcomes include robust protein/excipient quantitation in-process (NIR, Raman), structural insight (FTIR), aggregation detection (UV‑Vis with integrating sphere), nanoparticle conjugation monitoring (microvolume UV‑Vis), and rapid DNA/RNA purity assessments (NanoDrop + Acclaro). When integrated with automation and control systems, these technologies enable tighter process control, reduced labor and consumable costs, and support strategies for real‑time release testing.

Instrumentation used

Main instruments and accessories reported across the notes:

• FTIR: Thermo Scientific Nicolet iS10, iS50 (DTGS, MCT), ConcentratIR2 multiple‑reflection ATR, Smart OMNI‑Transmission accessory, BioCell CaF2 cell; OMNIC and PROTA‑3S software.
• NIR: Thermo Scientific Antaris MX FT‑NIR Process Analyzer with transflectance probe.
• Raman (process and lab): Thermo Scientific MarqMetrix All‑In‑One Process Raman Analyzer; MarqMetrix Performance BallProbe and FlowCell sampling optics; Thermo Scientific DXR3 SmartRaman+ Spectrometer; TQ Analyst, SOLO, MarqMetrix/TruBio and in‑house Python platforms for chemometrics.
• UV‑Vis and microvolume: Thermo Scientific NanoDrop One / NanoDrop Ultra / NanoDrop Eight; Evolution One Plus UV‑Vis; Evolution ISA‑220 integrating sphere accessory (Kubelka–Munk reporting).
• Orthogonal analytics: HPLC, LC‑MS and clinical analyzers (e.g., Nova BioProfile).

References

A representative (selected) set of references cited in the compendium (standard form):

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. Villa J. et al. Demonstrating chemometric model transferability for 5 mammalian cell lines and 5 media types using the Thermo Scientific MarqMetrix All‑In‑One Process Raman Analyzer. (Application note / conference report).
7. Zhang A. et al. Advanced process monitoring and feedback control to enhance cell culture process production and robustness. Biotechnol Bioeng. 2015;112(12):2495–2504.
8. Workman J. A review of calibration transfer practices and instrument differences in spectroscopy. Appl Spectrosc. 2018;73(3):340–365.
9. Guo S., Popp J., Bocklitz T. Chemometric analysis in Raman spectroscopy from experimental design to machine learning‑based modeling. Nat Protoc. 2021;16:5426–5459.
10. Hawe A., Friess W., Jiskoot W. Size‑exclusion chromatography and aggregation analysis of proteins. Anal Biochem. 2008;38:115–122.
11. Pignataro M.F. et al. Protein aggregation and food/functional implications. Molecules. 2020;25:4854.
12. 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.
13. Oldenburg S.J., Averitt R.D., Westcott S.L., Halas N.J. Nanoengineering of optical resonances. Chem Phys Lett. 1998;288:243–247.
14. Nolasco M., Pleitt K., Khadka N. Raman as in‑line tool for accurate protein quantification in downstream (application note).
15. Agrawal P. et al. A review of tangential flow filtration: process development and applications in the pharmaceutical industry. Org Process Res Dev. 2023;27(4):571–591.