A novel LED‐based 2D‐fluorescence spectroscopy system for in‐line monitoring of Chinese hamster ovary cell cultivations – Part I

A new two‐dimensional fluorescence sensor system was developed for in‐line monitoring of mammalian cell cultures. Fluorescence spectroscopy allows for the detection and quantification of naturally occurring intra‐ and extracellular fluorophores in the cell broth. The fluorescence signals correlate to the cells’ current redox state and other relevant process parameters. Cell culture pretests with twelve different excitation wavelengths showed that only three wavelengths account for a vast majority of spectral variation. Accordingly, the newly developed device utilizes three high‐power LEDs as excitation sources in combination with a back‐thinned CCD‐spectrometer for fluorescence detection. This setup was first tested in a lab design of experiments study with process relevant fluorophores proving its suitability for cell culture monitoring with LOD in the μg/L range. The sensor was then integrated into a CHO‐K1 cell culture process. The acquired fluorescence spectra of several batches were evaluated using multivariate methods. The resulting batch evolution models were challenged in deviating and “golden batch” validation runs. These first tests showed that the new sensor can trace the cells’ metabolic state in a fast and reliable manner. Cellular distress is quickly detected as a deviation from the “golden batch”.     

Advanced online monitoring of cell culture off‐gas using proton transfer reaction mass spectrometry

Mass spectrometry has been frequently applied to monitor the O₂ and CO₂ content in the off‐gas of animal cell culture fermentations. In contrast to classical mass spectrometry the proton transfer reaction mass spectrometry (PTR‐MS) provides additional information of volatile organic compounds by application of a soft ionization technology. Hence, the spectra show less fragments and can more accurately assigned to particular compounds. In order to discriminate between compounds of non‐metabolic and metabolic origin cell free experiments and fed‐batch cultivations with a recombinant CHO cell line were conducted. As a result, in total eight volatiles showing high relevance to individual cultivation or cultivation conditions could be identified. Among the detected compounds methanethiol, with a mass‐to‐charge ratio of 49, qualifies as a key candidate in process monitoring due to its strong connectivity to lactate formation. Moreover, the versatile and complex data sets acquired by PTR MS provide a valuable resource for statistical modeling to predict non direct measurable parameters. Hence, partial least square regression was applied to the complete spectra of volatiles measured and important cell culture parameters such as viable cell density estimated (R² = 0.86). As a whole, the results of this study clearly show that PTR‐MS provides a powerful tool to improve bioprocess‐monitoring for mammalian cell culture. Thus, specific volatiles emitted by cells and measured online by the PTR‐MS and complex variables gained through statistical modeling will contribute to a deeper process understanding in the future and open promising perspectives to bioprocess control. © 2013 American Institute of Chemical Engineers Biotechnol. Prog., 30:496–504, 2014     

Soft sensor based on 2D‐fluorescence and process data enabling real‐time estimation of biomass in Escherichia coli cultivations

In bioprocesses, specific process responses such as the biomass cannot typically be measured directly on‐line, since analytical sampling is associated with unavoidable time delays. Accessing those responses in real‐time is essential for Quality by Design and process analytical technology concepts. Soft sensors overcome these limitations by indirectly measuring the variables of interest using a previously derived model and actual process data in real time. In this study, a biomass soft sensor based on 2D‐fluorescence data and process data, was developed for a comprehensive study with a 20‐L experimental design, for Escherichia coli fed‐batch cultivations. A multivariate adaptive regression splines algorithm was applied to 2D‐fluorescence spectra and process data, to estimate the biomass concentration at any time during the process. Prediction errors of 4.9% (0.99 g/L) for validation and 3.8% (0.69 g/L) for new data (external validation), were obtained. Using principal component and parallel factor analyses on the 2D‐fluorescence data, two potential chemical compounds were identified and directly linked to cell metabolism. The same wavelength pairs were also important predictors for the regression‐model performance. Overall, the proposed soft sensor is a valuable tool for monitoring the process performance on‐line, enabling Quality by Design.     

The potential of random forest and neural networks for biomass and recombinant protein modeling in Escherichia coli fed‐batch fermentations

Product quality assurance strategies in production of biopharmaceuticals currently undergo a transformation from empirical “quality by testing” to rational, knowledge‐based “quality by design” approaches. The major challenges in this context are the fragmentary understanding of bioprocesses and the severely limited real‐time access to process variables related to product quality and quantity. Data driven modeling of process variables in combination with model predictive process control concepts represent a potential solution to these problems. The selection of statistical techniques best qualified for bioprocess data analysis and modeling is a key criterion. In this work a series of recombinant Escherichia coli fed‐batch production processes with varying cultivation conditions employing a comprehensive on‐ and offline process monitoring platform was conducted. The applicability of two machine learning methods, random forest and neural networks, for the prediction of cell dry mass and recombinant protein based on online available process parameters and two‐dimensional multi‐wavelength fluorescence spectroscopy is investigated. Models solely based on routinely measured process variables give a satisfying prediction accuracy of about ± 4% for the cell dry mass, while additional spectroscopic information allows for an estimation of the protein concentration within ± 12%. The results clearly argue for a combined approach: neural networks as modeling technique and random forest as variable selection tool.     

Advancing 2D fluorescence online monitoring in microtiter plates by separating scattered light and fluorescence measurement,using a tunable emission monochromator

Online fluorescence monitoring has become a key technology in modern bioprocess development, as it provides in-depth process knowledge at comparably low costs. In particular, the technology is widely established for high-throughput microbioreactor cultivation systems, due to its noninvasive character. For microtiter plates, previously also multi-wavelength 2D fluorescence monitoring was developed. To overcome an observed limitation of fluorescence sensitivity, this study presents a modified spectroscopic setup, including a tunable emission monochromator. The new optical component enables the separation of the scattered and fluorescent light measurements, which allows for the adjustment of integration times of the charge-coupled device detector. The resulting increased fluorescence sensitivity positively affected the performance of principal component analysis for spectral data of Escherichia coli batch cultivation experiments with varying sorbitol concentration supplementation. In direct comparison with spectral data recorded at short integration times, more biologically consistent signal dynamics were calculated. Furthermore, during partial least square regression for E. coli cultivation experiments with varying glucose concentrations, improved modeling performance was observed. Especially, for the growth-uncoupled acetate concentration, a considerable improvement of the root-mean-square error from 0.25 to 0.17 g/L was achieved. In conclusion, the modified setup represents another important step in advancing 2D fluorescence monitoring in microtiter plates.     

Generic Raman‐based calibration models enabling real‐time monitoring of cell culture bioreactors

Raman‐based multivariate calibration models have been developed for real‐time in situ monitoring of multiple process parameters within cell culture bioreactors. Developed models are generic, in the sense that they are applicable to various products, media, and cell lines based on Chinese Hamster Ovarian (CHO) host cells, and are scalable to large pilot and manufacturing scales. Several batches using different CHO‐based cell lines and corresponding proprietary media and process conditions have been used to generate calibration datasets, and models have been validated using independent datasets from separate batch runs. All models have been validated to be generic and capable of predicting process parameters with acceptable accuracy. The developed models allow monitoring multiple key bioprocess metabolic variables, and hence can be utilized as an important enabling tool for Quality by Design approaches which are strongly supported by the U.S. Food and Drug Administration. © 2015 American Institute of Chemical Engineers Biotechnol. Prog., 31:1004–1013, 2015     

miRNA engineering of CHO cells facilitates production of difficult‐to‐express proteins and increases success in cell line development

In recent years, coherent with growing biologics portfolios also the number of complex and thus difficult‐to‐express (DTE) therapeutic proteins has increased considerably. DTE proteins challenge bioprocess development and can include various therapeutic protein formats such as monoclonal antibodies (mAbs), multi‐specific affinity scaffolds (e.g., bispecific antibodies), cytokines, or fusion proteins. Hence, the availability of robust and versatile Chinese hamster ovary (CHO) host cell factories is fundamental for high‐yielding bioprocesses. MicroRNAs (miRNAs) have emerged as potent cell engineering tools to improve process performance of CHO manufacturing cell lines. However, there has not been any report demonstrating the impact of beneficial miRNAs on industrial cell line development (CLD) yet. To address this question, we established novel CHO host cells constitutively expressing a pro‐productive miRNA: miR‐557. Novel host cells were tested in two independent CLD campaigns using two different mAb candidates including a normal as well as a DTE antibody. Presence of miR‐557 significantly enhanced each process step during CLD in a product independent manner. Stable expression of miR‐557 increased the probability to identify high‐producing cell clones. Furthermore, production cell lines derived from miR‐557 expressing host cells exhibited significantly increased final product yields in fed‐batch cultivation processes without compromising product quality. Strikingly, cells co‐expressing miR‐557 and a DTE antibody achieved a twofold increase in product titer compared to clones co‐expressing a negative control miRNA. Thus, host cell engineering using miRNAs represents a promising tool to overcome limitations in industrial CLD especially with regard to DTE proteins. Biotechnol. Bioeng. 2017;114: 1495–1510. © 2017 The Authors. Biotechnology and Bioengineering Published by Wiley Periodicals, Inc.     

Synchronous fluorescence spectroscopy as a novel tool to enable PAT applications in bioprocesses

In this work, synchronous fluorescence spectroscopy (SFS) is evaluated as a new tool for real-time bioprocess monitoring of animal cell cultures. This technique presents several advantages over the traditional two-dimensional (2D) fluorometry since it provides data on various fluorescent compounds in a single spectrum, showing improved peak resolution and recording speed. Bioreactor cultures of three monoclonal antibody-producing CHO cell lines were followed in situ by both 2D and synchronous fluorometry techniques. The time profiles of the main spectral features in each data type present some differences, but principal component analysis indicated both as containing enough information to distinguish the cultures. Partial least squares regression models were then independently developed for viable cell density and antibody levels on the basis of the different fluorescence signals recorded, hiding half of the dataset for subsequent validation purposes. Regardless of the signal used, model predictions fit very well the off-line measurements; still, the synchronous spectra collected at a wavelength difference of 20 nm allowed comparable and superior performances for cell density and antibody titer, respectively, with validation accuracies higher than 91%. Therefore, SFS compares favorably with the traditional 2D approach, becoming an improved, faster option for real-time monitoring of cells and product titer over culture time. The readiness in data acquisition facilitates the design of process control strategies meeting the requirements of a PAT application.     

A New Evaluation Method for 2‐D Fluorescence Spectra Based on Theoretical Modeling

This article presents a new evaluation procedure of 2‐D fluorescence spectra obtained during a yeast cultivation without performing a calibration measurement. The 2‐D fluorescence spectra are used to predict the process variables biomass, glucose and ethanol. The new calibration procedure uses a theoretical model of these process variables, i.e., differential equations, to replace any calibration measurement. The theoretical model parameters are identified simultaneously during the calculation of the chemometric models. The root mean square error of prediction of the chemometric models with respect to off‐line measurements are 1.5 g/L, 0.40 g/L and 0.56 g/L for glucose, biomass and ethanol, respectively.     

Biosensor online control of citric acid production from glucose by Yarrowia lipolytica using semicontinuous fermentation

Our study aimed at the development of an effective method for citric acid production from glucose by use of the yeast Yarrowia lipolytica. The new method included an automated bioprocess control using a glucose biosensor. Several fermentation methodologies including batch, fed‐batch, repeated batch and repeated fed‐batch cultivation were tested. The best results were achieved during repeated fed‐batch cultivation: Within 3 days of cycle duration, approximately 100 g/L citric acid were produced. The yields reached values between 0.51 and 0.65 g/g and the selectivity of the bioprocess for citric acid was as high as 94%. Due to the elongation of the production phase of the bioprocess with growth‐decoupled citric acid production, and by operating the fermentation in cycles, an increase in citric acid production of 32% was achieved compared with simple batch fermentation.     

Microfluidic biolector—microfluidic bioprocess control in microtiter plates

In industrial‐scale biotechnological processes, the active control of the pH‐value combined with the controlled feeding of substrate solutions (fed‐batch) is the standard strategy to cultivate both prokaryotic and eukaryotic cells. On the contrary, for small‐scale cultivations, much simpler batch experiments with no process control are performed. This lack of process control often hinders researchers to scale‐up and scale‐down fermentation experiments, because the microbial metabolism and thereby the growth and production kinetics drastically changes depending on the cultivation strategy applied. While small‐scale batches are typically performed highly parallel and in high throughput, large‐scale cultivations demand sophisticated equipment for process control which is in most cases costly and difficult to handle. Currently, there is no technical system on the market that realizes simple process control in high throughput. The novel concept of a microfermentation system described in this work combines a fiber‐optic online‐monitoring device for microtiter plates (MTPs)—the BioLector technology—together with microfluidic control of cultivation processes in volumes below 1 mL. In the microfluidic chip, a micropump is integrated to realize distinct substrate flow rates during fed‐batch cultivation in microscale. Hence, a cultivation system with several distinct advantages could be established: (1) high information output on a microscale; (2) many experiments can be performed in parallel and be automated using MTPs; (3) this system is user‐friendly and can easily be transferred to a disposable single‐use system. This article elucidates this new concept and illustrates applications in fermentations of Escherichia coli under pH‐controlled and fed‐batch conditions in shaken MTPs. Biotechnol. Bioeng. 2010;107: 497–505. © 2010 Wiley Periodicals, Inc.     

Metabolic analysis of antibody producing CHO cells in fed‐batch production

Chinese hamster ovary (CHO) cells are commonly used for industrial production of recombinant proteins in fed batch or alternative production systems. Cells progress through multiple metabolic stages during fed‐batch antibody (mAb) production, including an exponential growth phase accompanied by lactate production, a low growth, or stationary phase when specific mAb production increases, and a decline when cell viability declines. Although media composition and cell lineage have been shown to impact growth and productivity, little is known about the metabolic changes at a molecular level. Better understanding of cellular metabolism will aid in identifying targets for genetic and metabolic engineering to optimize bioprocess and cell engineering. We studied a high expressing recombinant CHO cell line, designated high performer (HP), in fed‐batch productions using stable isotope tracers and biochemical methods to determine changes in central metabolism that accompany growth and mAb production. We also compared and contrasted results from HP to a high lactate producing cell line that exhibits poor growth and productivity, designated low performer (LP), to determine intrinsic metabolic profiles linked to their respective phenotypes. Our results reveal alternative metabolic and regulatory pathways for lactate and TCA metabolite production to those reported in the literature. The distribution of key media components into glycolysis, TCA cycle, lactate production, and biosynthetic pathways was shown to shift dramatically between exponential growth and stationary (production) phases. We determined that glutamine is both utilized more efficiently than glucose for anaplerotic replenishment and contributes more significantly to lactate production during the exponential phase. Cells shifted to glucose utilization in the TCA cycle as growth rate decreased. The magnitude of this metabolic switch is important for attaining high viable cell mass and antibody titers. We also found that phosphoenolpyruvate carboxykinase (PEPCK1) and pyruvate kinase (PK) are subject to differential regulation during exponential and stationary phases. The concomitant shifts in enzyme expression and metabolite utilization profiles shed light on the regulatory links between cell metabolism, media metabolites, and cell growth. Biotechnol. Bioeng. 2013; 110: 1735–1747. © 2013 Wiley Periodicals, Inc.     

In situ Raman spectroscopy for simultaneous monitoring of multiple process parameters in mammalian cell culture bioreactors

In this study, the application of Raman spectroscopy to the simultaneous quantitative determination of glucose, glutamine, lactate, ammonia, glutamate, total cell density (TCD), and viable cell density (VCD) in a CHO fed‐batch process was demonstrated in situ in 3 L and 15 L bioreactors. Spectral preprocessing and partial least squares (PLS) regression were used to correlate spectral data with off‐line reference data. Separate PLS calibration models were developed for each analyte at the 3 L laboratory bioreactor scale before assessing its transferability to the same bioprocess conducted at the 15 L pilot scale. PLS calibration models were successfully developed for all analytes bar VCD and transferred to the 15 L scale. © 2012 American Institute of Chemical Engineers Biotechnol. Prog., 2012     

Implementation of a Fully Automated Microbial Cultivation Platform for Strain and Process Screening

Advances in molecular biotechnology have resulted in the generation of numerous potential production strains. Because every strain can be screened under various process conditions, the number of potential cultivations is multiplied. Exploiting this potential without increasing the associated timelines requires a cultivation platform that offers increased throughput and flexibility to perform various bioprocess screening protocols. Currently, there is no commercially available fully automated cultivation platform that can operate multiple microbial fed‐batch processes, including at‐line sampling, deep freezer off‐line sample storage, and complete data handling. To enable scalable high‐throughput early‐stage microbial bioprocess development, a commercially available microbioreactor system and a laboratory robot are combined to develop a fully automated cultivation platform. By making numerous modifications, as well as supplementation with custom‐built hardware and software, fully automated milliliter‐scale microbial fed‐batch cultivation, sample handling, and data storage are realized. The initial results of cultivations with two different expression systems and three different process conditions are compared using 5 L scale benchmark cultivations, which provide identical rankings of expression systems and process conditions. Thus, fully automated high‐throughput cultivation, including automated centralized data storage to significantly accelerate the identification of the optimal expression systems and process conditions, offers the potential for automated early‐stage bioprocess development.     

Fed‐batch CHO cell t‐PA production and feed glutamine replacement to reduce ammonia production

Industrial therapeutic protein production has been greatly improved through fed‐batch development. In this study, improvement to the productivity of a tissue‐plasminogen activator (t‐PA) expressing Chinese hamster ovary (CHO) cell line was investigated in shake flask culture through the optimization of the fed‐batch feed and the reduction of ammonia generation by glutamine replacement. The t‐PA titer was increased from 33 mg/L under batch conditions to 250 mg/L with daily feeding starting after three days of culture. A commercially available fed‐batch feed was supplemented with cotton seed hydrolysate and the four depleted amino acids, aspartic acid, asparagine, cysteine, and tyrosine. The fed‐batch operation increased the generation of by‐products such as lactate and ammonia that can adversely affect the fed‐batch performance. To reduce the ammonia production, a glutamine‐containing dipeptide, pyruvate, glutamate, and wheat gluten hydrolysate, were investigated as glutamine substitutes. To minimize the lag phase as the cells adjusted to the new energy source, a feed glutamine replacement process was developed where the cells were initially cultured with a glutamine containing basal medium to establish cell growth followed by feeding with a feed containing the glutamine substitutes. This two‐step feed glutamine replacement process not only reduced the ammonia levels by over 45% but, in the case of using wheat gluten hydrolysate, almost doubled the t‐PA titer to over 420 mg/L without compromising the t‐PA product quality or glycosylation pattern. The feed glutamine replacement process combined with optimizing other feed medium components provided a simple, practical, and effective fed‐batch strategy that could be applied to the production of other recombinant therapeutic proteins. © 2012 American Institute of Chemical Engineers Biotechnol. Prog., 2013     

Combining mechanistic and data‐driven approaches to gain process knowledge on the control of the metabolic shift to lactate uptake in a fed‐batch CHO process

A growing body of knowledge is available on the cellular regulation of overflow metabolism in mammalian hosts of recombinant protein production. However, to develop strategies to control the regulation of overflow metabolism in cell culture processes, the effect of process parameters on metabolism has to be well understood. In this study, we investigated the effect of pH and temperature shift timing on lactate metabolism in a fed‐batch Chinese hamster ovary (CHO) process by using a Design of Experiments (DoE) approach. The metabolic switch to lactate consumption was controlled in a broad range by the proper timing of pH and temperature shifts. To extract process knowledge from the large experimental dataset, we proposed a novel methodological concept and demonstrated its usefulness with the analysis of lactate metabolism. Time‐resolved metabolic flux analysis and PLS‐R VIP were combined to assess the correlation of lactate metabolism and the activity of the major intracellular pathways. Whereas the switch to lactate uptake was mainly triggered by the decrease in the glycolytic flux, lactate uptake was correlated to TCA activity in the last days of the cultivation. These metabolic interactions were visualized on simple mechanistic plots to facilitate the interpretation of the results. Taken together, the combination of knowledge‐based mechanistic modeling and data‐driven multivariate analysis delivered valuable insights into the metabolic control of lactate production and has proven to be a powerful tool for the analysis of large metabolic datasets. © 2015 American Institute of Chemical Engineers Biotechnol. Prog., 31:1657–1668, 2015     

Feeding strategies enhance high cell density cultivation and protein expression in milliliter scale bioreactors

Miniature bioreactors under parallel fed‐batch operations are not only useful screening tools for bioprocess development but also provide a suitable basis for eventual scale‐up. In this study, three feeding strategies were investigated: besides the established intermittent feeding by a liquid handler, an optimized microfluidic device and a new enzymatic release system were applied for parallel fed‐batch cultivation of Escherichia coli HMS174(DE3) and BL21(DE3) strains in stirred‐tank bioreactors on a 10 mL scale. Lower fluctuation in dissolved oxygen (DO) and higher optical densities were measured in fed‐batch processes applying the microfluidic device or the enzymatic glucose/fructose release system (conversion of intermittently added sucrose by an invertase), but no difference in dry cell weights (DCW) were observed. With all three feeding strategies high cell densities were realized on a milliliter scale with final optical density measured at 600 nm (OD₆₀₀) of 114–133 and final DCW concentrations of 69–70 g L^–1. The effect of feeding strategies on the expression of two heterologous proteins was investigated. Whereas no impact was observed on the expression of the spider silk protein eADF4(C16), the fluorescence of enhanced green fluorescence protein (eGFP) was reproducibly lower, if an intermittent glucose feed was applied. Thus, the impact of feeding strategy on expression is strongly dependent on the E. coli strain and/or expressed protein. As a completely continuous feed supply is difficult to realize in miniature bioreactors, the enzymatic release approach from this study can be easily applied in all microfluidic system to reduce fluctuations of glucose supply and DO concentrations.     

Feed development for fed‐batch CHO production process by semisteady state analysis

Semisteady state cultures are useful for studying cell physiology and facilitating media development. Two semisteady states with a viable cell density of 5.5 million cells/mL were obtained in CHO cell cultures and compared with a fed‐batch mode control. In the first semisteady state, the culture was maintained at 5 mM glucose and 0.5 mM glutamine. The second condition had threefold higher concentrations of both nutrients, which led to a 10% increase in lactate production, a 78% increase in ammonia production, and a 30% reduction in cell growth rate. The differences between the two semisteady states indicate that maintaining relatively low levels of glucose and glutamine can reduce the production of lactate and ammonia. Specific amino acid production and consumption indicated further metabolic differences between the two semisteady states and fed‐batch mode. The results from this experiment shed light in the feeding strategy for a fed‐batch process and feed medium enhancement. The fed‐batch process utilizes a feeding strategy whereby the feed added was based on glucose levels in the bioreactor. To evaluate if a fixed feed strategy would improve robustness and process consistency, two alternative feeding strategies were implemented. A constant volume feed of 30% or 40% of the initial culture volume fed over the course of cell culture was evaluated. The results indicate that a constant volumetric‐based feed can be more beneficial than a glucose‐based feeding strategy. This study demonstrated the applicability of analyzing CHO cultures in semisteady state for feed enhancement and continuous process improvement. © 2009 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

Perfusion seed cultures improve biopharmaceutical fed‐batch production capacity and product quality

Volumetric productivity and product quality are two key performance indicators for any biopharmaceutical cell culture process. In this work, we showed proof‐of‐concept for improving both through the use of alternating tangential flow perfusion seed cultures coupled with high‐seed fed‐batch production cultures. First, we optimized the perfusion N‐1 stage, the seed train bioreactor stage immediately prior to the production bioreactor stage, to minimize the consumption of perfusion media for one CHO cell line and then successfully applied the optimized perfusion process to a different CHO cell line. Exponential growth was observed throughout the N‐1 duration, reaching >40 × 10⁶ vc/mL at the end of the perfusion N‐1 stage. The cultures were subsequently split into high‐seed (10 × 10⁶ vc/mL) fed‐batch production cultures. This strategy significantly shortened the culture duration. The high‐seed fed‐batch production processes for cell lines A and B reached 5 g/L titer in 12 days, while their respective low‐seed processes reached the same titer in 17 days. The shortened production culture duration potentially generates a 30% increase in manufacturing capacity while yielding comparable product quality. When perfusion N‐1 and high‐seed fed‐batch production were applied to cell line C, higher levels of the active protein were obtained, compared to the low‐seed process. This, combined with correspondingly lower levels of the inactive species, can enhance the overall process yield for the active species. Using three different CHO cell lines, we showed that perfusion seed cultures can optimize capacity utilization and improve process efficiency by increasing volumetric productivity while maintaining or improving product quality. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:616–625, 2014