FDA Science Forum: A Comparison Between Classical AD Laboratory Mass Spectrometry and Next-Generation Emerging Analytical Technology (Rebel)

Abstract

Analytical instruments are critical for assessing product quality and controlling the development process of biopharmaceutical products. In the pharmaceutical industry, they are employed in a wide range of applications, such as research related to product quality—e.g., glycan analysis in monoclonal antibodies and determination of amino acid and metabolite concentrations in cell culture media. Although liquid chromatography-mass spectrometry (LC-MS) offers flexible capabilities for addressing complex analytical challenges, its method development typically takes more time than other analytical approaches. Recently, novel analytical devices like REBEL from 908 Devices have adopted user-friendly, simplified operational procedures to replace traditional, labor-intensive LC-MS workflows. REBEL is an instrument capable of rapidly quantifying amino acids and other important metabolites, enabling comprehensive analysis and high-throughput sample testing. Our goal is to compare this next-generation analytical platform (REBEL by 908 Devices) with the conventional LC-ToF MS approach and evaluate the quality of the resulting data.

Introduction

Biopharmaceutical companies are increasingly focused on understanding the composition of culture media and ensuring product quality. To enhance production efficiency, they must address the time-consuming nature of traditional methods. Cell line development and optimization are critical components of process strategy. Traditionally, liquid chromatography-mass spectrometry (LC-MS) has been used for studying the nutritional and metabolomic profiles of culture media, a technique that requires highly specialized technical expertise and is both labor-intensive and time-consuming. However, optimizing experiments necessitates real-time monitoring of bioreactor parameters, process analytics, and results related to product quality attributes. The fully automated culture medium component analyzer system from 908 Devices—known as REBEL—is a novel analytical instrument that has garnered significant attention within the biopharmaceutical industry due to its rapid analysis capabilities and user-friendly operation.

This article introduces the latest advancements in using REBEL as an alternative to traditional LC-MS methods for monitoring bioreactor media. The key advantages of REBEL for media analysis include direct analysis without derivatization, rapid analysis times, and simple operation. However, since this instrument is a newcomer to the market, its analytical techniques remain insufficiently described. Therefore, we conducted a comparative study of media analysis using Waters’ non-derivatization and normal-phase chromatography methods with BioAccord RDa time-of-flight (ToF) LC-MS. Our goal is to demonstrate the comparability of REBEL and RDa results for media analysis and to explore the cost-effectiveness of these two analytical instrument approaches. Furthermore, we will establish both analytical methods in our laboratory to evaluate process changes, feeding strategies, and their impact on product quality in future metabolic studies, thereby deepening our understanding of critical process parameters and their influence on the quality attributes of pharmaceutical products.

Results and Discussion

Figure 1. Trends in amino acid concentrations from 68 to 227 hours for Bioreactor 1 (top) and Bioreactor 2 (bottom), with pH set at 7.1. Samples were analyzed using the Waters BioAccord RDa LC-MS (blue) and the REBEL Media Analyzer (red).

Figure 2. Trends in amino acid concentrations from 68 to 227 hours for bioreactors 3 (top) and 4 (bottom), with pH set at 7.3. Samples were analyzed using the Waters BioAccord RDa LC-MS (blue) and the REBEL Media Analyzer (red).

Table 1: Comparison between REBEL and LC-MS, where “+” indicates the required technical expertise or the quantity needed.

We were able to successfully run all samples using the preliminary method on REBEL and BioAccord RDa LC-MS, enabling us to develop a processing method on UNIFI for the identification, mass confirmation, and quantification of all amino acids from standard calibration curves in our RDa LC-MS analyses. REBEL exported calculated concentrations for all amino acids and generated manual pivot tables to illustrate the results shown on the left. Bioreactor replicates 1 and 2 were operated at a set pH of 7.1 and exhibited similar amino acid trends throughout the culture (Figure 1). Similarly, bioreactor replicates 3 and 4, when operated at a set pH of 7.3, shared comparable trends with each other (Figure 2). The amino acid trends for bioreactors 1–4 are shown in blue for BioAccord RDa LC-MS analysis and in red for REBEL analysis. Both REBEL and RDa LC-MS data indicate that the amino acid trends across all reactors are quite consistent. For sample preparation, the REBEL method uses a sample dilution ratio of 1:100, compared to 1:500 for RDa LC-MS, suggesting that RDa LC-MS has at least five times greater sensitivity than REBEL. However, we observed that at the later time points—hours 204 and 227—the amino acid concentrations measured by RDa LC-MS were significantly higher than those measured by REBEL. This discrepancy is likely due to the fact that these values were detected beyond the range of the standard curve; thus, extrapolating concentrations introduced calculation errors. REBEL can analyze only 33 analytes, including 22 amino acids, 5 vitamins, and 6 biogenic amines, whereas RDa, being a TOF LC-MS, can analyze these and many other unknown compounds. Table 1 provides a comparison between REBEL and conventional LC-MS. REBEL is easy to use and requires relatively little time and effort to operate, but it sacrifices the ability to analyze compounds outside its predefined scope and exhibits slightly lower sensitivity compared to TOF mass spectrometry. Conventional LC-MS methods can be specifically optimized for enhanced sensitivity, yet this requires a higher level of technical expertise. REBEL offers a comprehensive reagent kit for automated analysis, though the cost per sample for kit-based analysis is considerably higher than the materials required for LC-MS methods. Overall, while REBEL performs relatively well in terms of cost per run versus performance within its analytical scope, its ease of use and rapid analysis capabilities represent key advantages for advancing real-time process optimization.

Materials and Methods

We collected daily samples from Sartorius’ AMBR250 bioreactor (VRC01 CHO cell culture) and conducted a 10-day batch experiment. The pH of bioreactors 1 and 2, which were repeated, was 7.1, while the pH of bioreactors 3 and 4 was 7.3; all other conditions remained consistent. The samples were processed and filtered for analysis of the free amino acid concentrations in the cell culture medium.

The REBEL culture medium analysis kit (908 Devices) provides BGE solution, diluent, and pre-prepared standards. Samples are diluted 100-fold with the diluent. All concentration data are exported in .csv file format. The LC-MS method for BioAccord RDa (Waters) was adapted from Waters’ application guidelines to suit the specific configuration of the Accuilty Patrols as the LC system. The amino acid cell culture standards (Waters) were freshly prepared according to the instructions; each amino acid’s calibration curve consisted of six points ranging from 5 µM to 0.5 µM, except for cystine, whose calibration range was from 2.5 µM to 0.25 µM. Freshly prepared samples were diluted 1:500 with 0.1% formic acid. The LC-MS data obtained were processed using UNIFI (Waters) to generate calibration curves and to identify and quantify the amino acid concentrations. All samples (REBEL and BioAccord RDa) were analyzed in triplicate.

Conclusion

Our preliminary RDa LC-MS method indicates that the data collected by the REBEL cell culture analyzer are comparable to those obtained from CHO cell culture bioreactors. Currently, based on the REBEL-based experimental results, we are optimizing the RDa LC-MS method by using an 8-point calibration curve (ranging from 0.1 to 30 μM) to accurately capture the concentrations of all amino acids. Once all samples have been analyzed using the finalized RDa LC-MS method, we will conduct further method comparisons and statistical analyses. We will test the null hypothesis that measurements obtained using REBEL differ from those obtained using RDa LC-MS, and then seek data-driven evidence to demonstrate that the measurements are indeed equivalent. If the data support the rejection of the null hypothesis, we can conclude that these instruments exhibit comparable performance in measuring metabolites in each comparison. Future studies will further assess REBEL’s ability to evaluate the impact of metabolism on product quality—for example, how temperature changes affect the utilization of metabolic products and their influence on the glycan profiles in microbial reactors. Given that cellular metabolism can introduce collinearity in both the utilization and concentration of metabolic products, we will employ multivariate data analysis (MVDA) to characterize cellular metabolism and its effects on glycan profiles. The ease of use and rapid analysis capabilities of REBEL allow for higher sample throughput while maintaining data quality comparable to that achieved with conventional LC-MS methods, thereby enabling us to rapidly advance our understanding of the impact on critical quality attributes and enhance our comprehension of key process parameters.