Coupling Size Exclusion Chromatography and Capillary-Enhanced Raman Spectroscopy for Analysis of Hemolyzed Serum Samples

A joint study between the Institute for Environment and Energy, Technology and Analytics (IUTA e. V., Duisburg, Germany) and Heinrich Heine University Düsseldorf (Germany) explored the combination of size exclusion chromatography (SEC) with diode array detector (DAD) and capillary-enhanced Raman spectroscopy (CERS) to directly analyze hemolyzed serum samples. While previously doing so was a challenge in clinical diagnostics due to interference, the researchers found that their approach efficiently enabled detailed characterization and differentiation of key blood serum proteins, including immunoglobulin G (IgG), serum albumin, and hemoglobin, as well as allowed effective investigation of the oxidation and binding states of the heme group. We spoke to Jana Thissen, first author of the paper resulting from this study, about her team’s work.

Your study (1) focuses on the application of the online coupling of size exclusion chromatography (SEC) to a diode array detector (DAD) and capillary-enhanced Raman spectroscopy (CERS) for direct analysis of biomarkers in hemolyzed serum samples. What did you hypothesize are the advantages of these combined techniques as opposed to vibrational spectroscopic techniques such as Fourier transform infrared (FT-IR) and Raman spectroscopy (RS), which you state in the paper have been increasingly investigated for their potential applicability in clinical diagnosis?

The online coupling of size exclusion chromatography, diode array detection and capillary-enhanced Raman spectroscopy (SEC-DAD-CERS) requires no time-consuming sample preparation and offers a short analysis time. A single measurement can be performed in less than 15 minutes. For example, serum samples may need to be fractionated using centrifugal filters or chromatographic separation with a fraction collector prior to detection using vibrational spectroscopic techniques alone. For hemolyzed samples, it is particularly important to separate multiple fractions where the heme group is detectable, as its resonance-enhanced Raman bands could obscure other protein Raman bands. In contrast, the continuous acquisition of data points by SEC-DAD-CERS allows to selectively average the scans acquired over a defined time window. In an offline setup – where fractionation is combined with vibrational spectroscopic techniques for detection – adjusting the time window requires rerunning the experiment with a different set of selected fractions. This makes SEC-DAD-CERS significantly more flexible in terms of data processing and evaluation.

In addition, Raman spectroscopic techniques such as CERS are compatible with aqueous samples such as serum, whereas water causes significant interference in FT-IR spectroscopy. This may require the fractions to be dried before FT-IR analysis. However, the drying process can alter the samples or cause inhomogeneity, which must be accounted for in the analysis.

In summary, the advantages of SEC-DAD-CERS over offline vibrational spectroscopy are the combination of short analysis times, flexible data processing and evaluation without time-consuming sample-preparation before the analysis.

What are the protein classes of biomarkers that you are specifically seeking to identify and analyze, and can you please describe what makes these biomarkers important?

The study focused on the identification of serum albumin and immunoglobulin G (IgG). These two protein classes play essential roles in the human body and are important biomarkers of a patient's overall health condition. Serum albumin maintains oncotic pressure, transports a wide range of endogenous and exogenous molecules (including hormones, vitamins, minerals, and drugs), acts as an antioxidant, and plays a role in pH regulation. It is mainly used as a biomarker to diagnose specific medical conditions, such as liver and malabsorption syndromes. IgG provides long-term immunity by neutralizing pathogens, promoting phagocytosis, activating the complement system, and facilitating antibody-dependent cellular cytotoxicity (ADCC). IgG, for example, is used as a biomarker for immune response, autoimmune diseases, and chronic inflammation.

We have also investigated the determination of different oxidation and binding states of the heme prosthetic group of hemoglobin. Hemoglobin is responsible for the transport and transfer of oxygen in the human body and can be measured as oxyhemoglobin (oxygenated form, active form) and deoxyhemoglobin (deoxygenated form, active form). However, poisoning with oxidative substances, for example, can lead to an increased formation of methemoglobin (oxidized form, inactivated form), resulting in oxygen deficiency in tissues. This is why we investigated the different forms of hemoglobin in our study.

Briefly summarize the findings that you discuss in your article, and whether the hypotheses you supposed as to the advantages of the coupling of techniques were correct.

The online coupling of SEC-DAD-CERS allows the identification of protein classes in serum, in particular immunoglobulin G, serum albumin, and hemoglobin. While the method provides a shorter analysis time compared to offline fractionation, its sensitivity is limited to the most abundant proteins due to chromatographic resolution and detector limitations. The study demonstrates that the hit quality index (HQI) value allows for clear differentiation between protein classes, although variations in spectral quality and fluorescence interference affect accuracy. The likely detection of a haptoglobin-hemoglobin complex suggests further applications in studying hemolysis-related biomarkers. Methemoglobin can be detected accurately. However, deoxyhemoglobin and oxygenated hemoglobin approach an equilibrium state of oxygenation under the conditions of the SEC. Therefore, the level of oxygen binding is not accessible with SEC-DAD-CERS. Overall, SEC-DAD-CERS shows promise for clinical diagnostics and biological research, with potential improvements in data pre-processing and system optimization.

Do you anticipate similar results in using your technique for detecting other proteins, or investigating other bodily fluids?

In my first publication (2), we successfully identified and elucidated the structure of other proteins, such as hen egg white lysozyme, β-lactoglobulin, and the monoclonal antibody rituximab. In general, the online coupling of SEC-DAD-CERS can be applied to various sample types, including biopharmaceuticals and biological fluids such as serum.

However, the detection of proteins depends on their concentration in the sample solution. We have used the proteins in typical concentrations of biopharmaceutical products, which are approximately between 10 and 25 mg/mL. Trace level concentrations of proteins cannot be analyzed by SEC-DAD-CERS. A preconcentration step is then required prior to analysis.

What difficulties did you encounter in your work, specifically sampling and data interpretation challenges?

A common challenge for innovative techniques is the lack of dedicated software for data analysis. The design of the CERS detector enables the continuous recording of individual spectra. This enables precise identification of target fractions in the chromatogram and targeted processing of their Raman spectra by selective averaging of the scans acquired within the time window of the individual fractions. However, manual processing of this data is very time-consuming.

In addition, using a laser line that matches the absorption bands of hemoglobin induces resonance excitation of the heme prosthetic group, allowing sensitive detection of its various oxidation and binding states. However, this also results in fluorescence and absorption in the Raman spectral region, posing an additional challenge for data pre-processing. Despite these issues, we were able to process the data.

How did you process the chemometric data to obtain the results you were looking for?

In the study, target fractions were first selected based on the DAD chromatogram. Multiple Raman spectra of the sample fraction and a background spectrum of the pure mobile phase were then averaged. To ensure reproducibility, both the sample and the background spectra were normalized, followed by background correction (subtraction of the background spectrum), smoothing, and baseline correction. The pre-processed spectra were then scaled and an HQI was calculated based on Pearson’s coefficient of determination (R²).

Were there any factors that might affect the accuracy of your findings?

We found that variability in HQI values between runs due to spectral noise, fluorescence background, and data processing interference can affect the reliability of spectral matching. Therefore, establishing reliable and consistent data pre-processing workflows within the study is essential.

How do you imagine the results of your study can/will be more broadly applied?

I believe that SEC-DAD-CERS has potential in biopharmaceutical characterization. In my first publication (2), we were able to show that SEC-DAD-CERS allows the determination of the molecular weight, an impurity analysis regarding the formation of aggregates and fragments, and the assessment of the higher-order structure of the therapeutic protein. In addition, SEC-DAD-CERS can be used to quantify functional excipients such as disaccharides. However, I believe that the automated and reproducible pre-processing and evaluation of SEC-DAD-CERS data is essential for the practical application of this technique. The data analysis must be able to reliably detect and evaluate significant differences between various spectra.

Another application could be the structural identification of yet unknown proteins in various biological fluids. However, this will most likely require pre-concentration of the target protein prior to analysis.

Have you received any feedback on your study from analysts doing similar work?

Every time I present my work at conferences, I receive helpful feedback that I use to improve my work.

Are there any next steps in this research?

In the next step, we will focus on the automated analysis of the SEC-CERS data. This will be done using the open-source analysis platform StreamFind. We are currently working on a manuscript that will be submitted to a peer-reviewed journal in the coming weeks.

Since February 2021, Jana Thissen has been working as Scientific Associate at the Institut für Umwelt & Energie, Technik & Analytik (IUTA) e. V. in Duisburg, Germany. Her work at IUTA focuses on the development and optimization of methods for coupled analytical systems consisting of one- and two-dimensional HPLC, Raman spectroscopy and mass spectrometry for the analysis of pharmaceuticals and biopharmaceutical drugs. Thissen is pursuing her doctoral degree, focusing on the online coupling of size exclusion chromatography and capillary-enhanced Raman spectroscopy for the structural identification of proteins and their prosthetic groups under the supervision of Prof. Dr. Jörg Breitkreutz at the Institute of Pharmaceutics and Biopharmaceutics, Faculty of Mathematics and Natural Sciences, Heinrich Heine University Düsseldorf, Germany. She holds a Master of Science in Applied Chemistry with a specialization in Instrumental Analysis and Laboratory Management from the Niederrhein University of Applied Sciences in Krefeld, Germany.Photo courtesy of Thissen.

References

1. Thissen, J.; Klassen, M. D.; Hacker, M. C.; Breitkreutz, J.; Teutenberg, T.; Fischer, B.; Online Coupling of Size Exclusion Chromatography to Capillary-Enhanced Raman Spectroscopy for the Identification of Protein Classes in Hemolyzed Blood Serum. Anal. Bioanal. Chem. 2025, 417 (2), 335–344. DOI: 10.1007/s00216-024-05649-3

2. Thissen, J.; Klassen, M. D.; Constantinidis, P.; Hacker, M. C.; Breitkreutz, J.; Teutenberg, T.; Fischer, B.; Online Coupling of Size Exclusion Chromatography to Capillary Enhanced Raman Spectroscopy for the Analysis of Proteins and Biopharmaceutical Drug Products. Anal. Chem. 2023, 95, 17868−17877. DOI: 10.1021/acs.analchem.3c03991