A-TEEM: A Novel, Powerful Tool for Evaluating and Monitoring Cell Culture Media

Abstract

Maintaining the consistency of cell culture media is essential to ensuring reliable cell growth and high-quality biopharmaceutical products. These media are complex mixtures of nutrients, growth factors, and other essential components that support cell development in bioreactors. Even minor variations in composition can influence cell growth and impact final product quality. Accurate characterization and real-time monitoring of cell culture media are crucial for process control. Traditional analytical techniques, such as high-performance liquid chromatography and liquid chromatography–mass spectrometry, provide detailed compositional analysis but are time-consuming, expensive, and unsuitable for high-throughput applications. In contrast, fluorescence spectroscopy has emerged as a rapid, cost-effective alternative. The simultaneous acquisition of absorbance, transmittance, and excitation-emission matrices (A-TEEM) enables precise molecular fingerprinting, facilitating the detection of subtle compositional changes. This study demonstrates the effectiveness of A-TEEM in assessing lot-to-lot variability in commercially available GMP-grade cell culture media. By applying parallel factor analysis, the author and her team successfully monitor compositional variations across different batches. The findings highlight the potential of fluorescence spectroscopy, combined with multivariate analysis, as a powerful tool for evaluating and monitoring cell culture media in both research and manufacturing settings.

Introduction

Cell culture media play a critical role in biotherapeutic production by providing essential nutrients and growth factors necessary for cell proliferation and product formation. However, the certificate of analysis for commercially available media typically includes only basic specifications, such as pH and osmolality. This limited characterization makes it essential for end users to implement their own qualification systems to assess media variability and its potential impact on product yield.

The chemical complexity and variability of cell culture media pose significant challenges in ensuring consistent performance and product quality. These formulations often contain hundreds of components, including amino acids, vitamins, sugars, and trace elements. Several analytical challenges arise when characterizing media composition:

• The heterogeneous nature of media formulation complicates analysis using conventional techniques.
• Routine quality control metrics, such as pH and osmolality, provide only general assessments and fail to capture detailed compositional differences.
• High-performance liquid chromatography (HPLC) and liquid chromatography–mass spectrometry (LC–MS) offer detailed compositional data but require extensive sample preparation, specialized instrumentation, and lengthy processing times, making them impractical for high-throughput screening in biomanufacturing.

To address these challenges, researchers have explored alternative techniques for monitoring cell culture media variability. Fluorescence spectroscopy has emerged as a promising approach due to its rapid, non-invasive, and cost-effective nature (1). This technique provides a unique molecular fingerprint of the media, enabling the detection of subtle compositional changes that may affect cell growth and product yield.

The simultaneous acquisition of absorbance, transmittance, and excitation-emission matrices (A-TEEM) enables comprehensive optical characterization of cell culture media, allowing for the identification of a wide range of fluorophores and chromophores. When combined with multivariate analysis techniques, such as parallel factor analysis (PARAFAC) and principal component analysis (PCA), A-TEEM can effectively classify different media types, monitor compositional changes over time, and track variations in media stored under different conditions (2). In this study, the author and her team extend this application to tracking lot-to-lot variability within the same type of media, providing a practical approach for qualifying media in bioprocesses.

Fluorescence excitation-emission spectroscopy for media characterization was pioneered by Alan G. Ryder’s research group at the University of Galway in Ireland (1,3-5). The group’s work demonstrated that fluorescence-based techniques offer real-time, non-invasive molecular insights, overcoming key limitations of traditional analytical methods. This advancement suggests that fluorescence spectroscopy could revolutionize quality control in biopharmaceutical manufacturing.

This study aims to evaluate A-TEEM, in combination with multivariate analysis, as a tool for monitoring cell culture media variability. Specifically, the study seeks to:

• Assess the technique’s capability to track lot-to-lot variability in media composition
• Provide a robust, cost-effective, and efficient approach for ensuring media quality before use in bioreactors

Methodology

Sample preparation

For the lot-to-lot variability study, 12 lots of yeastolate good manufacturing practice (GMP)-ready batches were obtained from a commercial supplier. Sample concentrations were around 0.3 mg/mL: 9–10 mg of cell media in 30 mL 10 mM pH = 7.4 phosphate buffered saline buffer. The researchers prepared triplicates of each lot, repeated for two days, giving six samples of each lot in total.

Data acquisition

The yeastolate cell media samples were subjected to simultaneous acquisition of A-TEEMs using a spectrofluorometer (Aqualog UV-800, HORIBA). Absorbance and transmittance measurements, as well as excitation–emission matrices (EEMs), were recorded across a wavelength range of 250–800-nm increments, while emission wavelengths for EEMs were from 250–800 nm, with a step size of 5 nm. Two acquisition methods were applied for all samples: for amino acids region (excitation from 250–800 nm) data acquisition, integration time was set for 0.1s; while for vitamin region (excitation from 310–700 nm) data acquisition, integration time was set for 1s. National Institute of Standards and Technology traceable excitation and emission spectral correction factors, inner-filter correction, Rayleigh scatter masking, and Raman scatter unit normalization were applied. Blank reference was recorded using fluorometer cells (Starna 3Q-10 water, Starna Cells).

Multivariate analysis

The acquired optical data were analyzed using PARAFAC to classify the different media lots. PARAFAC is a multi-way decomposition method, a generalization of PCA to higher-order arrays, used to analyze data with multiple dimensions. It decomposes data into simpler components, revealing underlying patterns and structures.

In this study, PARAFAC was performed on the combined amino acid region and vitamin region A-TEEM dataset to extract and quantify the key spectral features that track their variabilities. PARAFAC modeling was performed with the aid of software (Solo, Eigenvector).

Results and discussion

Chemical insights of cell media from A-TEEM data and critical cell media components’ optical properties

The successful classification of media types and the quantification of components in cell media using A-TEEM, combined with PARAFAC or PCA, reveal intricate molecular distinctions. The spectral features driving this separation are likely associated with fluorophores and chromophores, including amino acid composition, vitamin concentrations, and growth factor profiles. These molecular signatures provide insights beyond traditional chemical analysis, offering a more holistic view of media composition.

The optical properties of five major fluorophores—tryptophan, tyrosine, pyridoxine, folic acid, and riboflavin—are listed in Table I. Their typical A-TEEM profiles, including absorbance and fluorescence fingerprint data, are shown in Figure 1. These components exhibit distinct spectral features with wide quantum efficiency.

Cell media lot-to-lot variations

The ability to track lot-to-lot variability in cell media was assessed using A-TEEM coupled with PARAFAC on 12 lots (Lot 1–12) of GMP-ready, commercially available yeastolate cell media. Six preparations of each lot (two days, triplicate each) were prepared and measured.

Examples of fluorescence fingerprints of the cell media, highlighting the aromatic amino acid and vitamin regions, are shown in Figure 2 for Lot 1 and Lot 12. All fingerprint colors are scaled to the intensity of the contour map, meaning the fingerprint represents media composition, while the color scale values reflect absolute concentration.

A visual inspection of the fluorescence fingerprints of Lot 1 and Lot 12 reveals that Lot 12 has a higher tyrosine content relative to tryptophan, as indicated by the appearance of a contour at excitation 280 nm and emission 300 nm. In contrast, Lot 1 shows a slightly higher riboflavin content relative to pyridoxine compared with Lot 12.

A calibrated five-component PARAFAC model, capturing 99.9% of the variance, successfully decomposed the fluorescence signals of 72 cell media sample data into five major fluorophores corresponding to tryptophan, riboflavin, pyridoxine, folic acid, and tyrosine, as shown in Figure 3.

Using the PARAFAC component scores (Components 1, 2, and 3) plotted in a 3D format, we can clearly distinguish different lots of these cell media, as shown in Figure 4. The cluster plot indicates that Lot 4 and Lot 12 exhibit the most distinct differences in Component 3 composition (tyrosine). This also validates the visual observation that Lot 12 is richer in tyrosine compared with Lot 1.

Quantitative approach for cell media major components

However, how much variability is present? A quantitative approach would be helpful for scientists and chemists to evaluate the scale of the differences. By applying the same model to pure standard measurements, the researchers were able to calculate the concentration of each component in the cell media samples using PARAFAC scores derived from standard components, as shown in Table II. The variability of these components remained relatively consistent across all 12 GMP batches tested.

Tryptophan and tyrosine contents were close to the manufacturer’s product catalog value of 0.9%, while pyridoxine, folic acid, and riboflavin were not listed. It is important to note that the tryptophan and tyrosine standards used for calibration were in their free amino acid forms, whereas in cell media, they may exist in protein-bound forms. Therefore, a conversion factor may need to be considered.

Nevertheless, this method provides an effective approach for quantitatively tracking component composition across batches.

Implications for biopharmaceutical production

The results of this study demonstrate the effectiveness of the fluorescence spectroscopy and multivariate analysis approach in monitoring cell culture media variability. The ability to track lot-to-lot variability can have significant implications for the biotherapeutic industry.

This technique can be used as a quality control tool to ensure media consistency before use in bioreactors, potentially reducing the risk for unexpected cell culture performance issues and improving overall process robustness. Additionally, the simplicity and cost-effectiveness of this approach make it an attractive alternative to traditional, time-consuming and resource-intensive analysis methods.

The research demonstrates that fluorescence spectroscopy combined with multivariate analysis offers a transformative approach to quality control in cell-culture media management:

• rapid analysis: minutes-long assessment compared with hours-long traditional methods
• cost-effectiveness: minimal sample preparation and equipment requirements
• comprehensive insights: molecular-level understanding of media composition
• predictive capabilities: early detection of potential media degradation

Conclusion

This study has demonstrated the utility of A-TEEM and multivariate analysis approach for monitoring cell culture media variability. The simultaneous acquisition of A-TEEM, coupled with PARAFAC, allowed for effective monitoring of lot-to-lot variabilities in media composition.

The findings suggest that this technique offers a rapid, cost-effective, and robust method for ensuring media quality before use in bioreactors, potentially improving the overall reliability and efficiency of cell culture-based biotherapeutic production.

References

1. Li, B.; Ryan, P. W.; Shanahan, M.; Leister, K. J.; Ryder, A. G. Fluorescence Excitation–Emission Matrix (EEM) Spectroscopy for Rapid Identification and Quality Evaluation of Cell Culture Media Components. Appl. Spectrosc. 2011, 65 (11), 1240–1249.
2. HORIBA.Monitoring Cell Culture Media Variability Using the Veloci BioPharma Analyzer and A-TEEM Molecular FIngerprinting. Biopharma Application NoteFL-2025-01-28. Jan. 28, 2025.
3. Ryan, P. W.; Li, B.; Shanahan, M.; Leister, K. J.; Ryder, A. G. Prediction of Cell Culture Media Performance Using Fluorescence Spectroscopy. Anal. Chem. 2010, 82 (4), 1311–1317.
4. Calvet, A.; Li, B.; Ryder, A. G. Rapid Quantification of Tryptophan and Tyrosine in Chemically Defined Cell Culture Media Using Fluorescence Spectroscopy. J. Pharm. Biomed. Anal. 2012, 71, 89–98.
5. Calvet, A.; Li, B.; Ryder, A. G. A Rapid Fluorescence-Based Method for the Quantitative Analysis of Cell Culture Media Photo-Degradation. Anal. Chim. Acta 2014, 807, 111–119.

About the author

Lyufei Chen, PhD, Lyufei.chen@horiba.com, is an application scientist at HORIBA.