- Email: [email protected]
Cell culture media analysis using rapid spectroscopic methods
Available online at www.sciencedirect.com
ScienceDirect Cell culture media analysis using rapid spectroscopic methods Alan G Ryder Cell culture media (CCM) used in industrial biotechnology are complex mixtures of molecules and elements that are inherently difficult to analyze comprehensively. CCM quality analysis is of utmost importance for efficient production of protein-based therapeutics with the correct Critical Quality Attributes (CQAs). Here we discuss the use of rapid spectroscopic methods for routine screening of CCM molecular variance which include electronic (UV–vis absorption and fluorescence) and vibrational (Raman, FT-IR, and NearInfra-red) spectroscopies. CCM analysis needs to provide: identity testing, compositional variance analysis (i.e. lot-to-lot variation), validation of media preparation protocols, and correlations with Critical Process Parameters (CPPs) or product CQAs. Rapid spectroscopic methods can fulfil some of these requirements but only with correct sample handling and preparation. Accurate analysis requires the use of chemometrics combined with a detailed knowledge of sample behavior such as water absorption and chemical stability. Address Nanoscale Biophotonics Laboratory, School of Chemistry, National University of Ireland-Galway, University Road, Galway, H91 CF50, Ireland Corresponding author: Ryder, Alan G ([email protected])
Current Opinion in Chemical Engineering 2018, 22:11–17 This review comes from a themed issue on Biotechnology and bioprocess engineering: biomanufacturing Edited by Michael Betenbaugh and Nigel Titchener-Hooker
https://doi.org/10.1016/j.coche.2018.08.008
therapeutics CQAs include factors such as the glycosylation pattern, stability, aggregation, impurity profiles, and so on. Many CQAs are directly affected by CCM [1] and unknown changes in composition can have very serious adverse consequences for the process. CCM are carefully designed and optimized for each process (and often for different stages in the process) and cell line, which results in a very wide range of different CCM formulations. The resulting media are highly complex, molecular and elemental mixtures, containing amino acids, carbohydrates, vitamins, growth factors, trace and bulk minerals. The most complex of the raw material components used in media manufacture are the various hydrolysates of protein sources (like yeast or soy). These hydrolysates are a complex mixture of carbohydrates, amino acids, peptides, and a host of other often unknown molecules and minerals [2–5]. These contain many more components than chemically defined media and often display much greater molecular and elemental variance but are often critical for antibody manufacturing processes [6]. This variance can arise from changes in source material, manufacturing issues, or improper handling/storage. For these materials, both molecular [7] and elemental variance are important [8] in the context of their effect on cell culture. This molecular and elemental variance impacts on CCM quality in terms of its intended purpose, namely sustaining a specific bioprocess. Therefore, when we use the term media quality it refers to a specific media and a specific process. It should be noted that the same media could be used in two different processes and for each process, the effect of media variance may be different.
2211-3398/ã 2018 Published by Elsevier Ltd.
Introduction Cell culture media (CCM) are a critical element in biopharmaceutical manufacturing which directly affect process yield and product quality. Therefore, the analysis of media variance and prediction of performance is of high importance. The essential purpose of CCM is to produce and maintain an optimum physiological environment to allow large-scale culturing of cells where they remain healthy for an appropriate amount of time and express the right product with the correct Critical Quality Attributes (CQAs). For protein-based www.sciencedirect.com
Another major issue is the fact that components are present in widely varying concentrations, from trace (ppm) levels for some elements to >5% w/w for carbohydrates. In liquid media water is the major component (>90% w/w) diluting components (millimolar, micromolar, and nanomolar concentrations) and making analyte detection more difficult for spectroscopic methods. Because of these compositional and concentration factors, media analysis using traditional analytical methods, is challenging, very time consuming, requiring multiple methods and techniques. This makes comprehensive analysis too expensive for routine use. As the industrial cell culture production of therapeutic proteins continues to grow so too does the need for rapid, inexpensive, reliable, robust, and non-destructive analytical methods that can be integrated into process control to improve end Current Opinion in Chemical Engineering 2018, 22:11–17
12 Biotechnology and bioprocess engineering: biomanufacturing
product yield and quality. A recent review [9] provides a more in-depth assessment of these raw materials issues. Whereas here we focus on the use of spectroscopic measurements on the media and take a critical look into the quality of the data being generated. The quality of the spectral data that these rapid methods can provide is very sensitive to sample handling and failure to recognize this can lead to the generation of erroneous data and ultimately misleading conclusions.
Fast spectroscopic methods
There are many potential spectroscopic methods that can be applied to media analysis, however, here we limit ourselves to commercially available, tried and tested methods. Relatively recent reviews describe the needs for Process Analytical Technologies (PAT) in biopharmaceutical manufacturing [10–13] and provide more background detail about how rapid spectroscopic techniques could be incorporated into manufacturing processes and are directly relevant to CCM analysis. The most important methods are the electronic spectroscopies (UV–vis absorption and fluorescence) and vibrational spectroscopies (MIR and NIR absorption and Raman scattering). The correct application of all methods requires an indepth appreciation of the practical difficulties of dealing with chemically complex mixtures. Sample handling therefore plays a critical role in the collection of accurate, correct, and reproducible data which is required for the development of robust analytical methods.
Sample handling
A critical issue with CCM analysis relates to sample handling and stability both in powdered and solution forms (Table 1). CCM have complex physicochemical properties which mean that the samples are neither chemically nor physically stable for varying time periods of minutes to hours depending on the type. It is very important to recognize here that stability in the context of measurements is different from stability associated with CCM performance in a bioprocess. For example, CCM powders are usually hygroscopic, and will quickly absorb atmospheric water rapidly (within a couple of minutes) and are thus unsuitable for vibrational spectroscopy analysis of molecular variance (as opposed to just measuring the effects of water absorption) unless properly dried. Whereas small amounts of absorbed water are unlikely to have any impact on cell culture performance. Water adsorption from the atmosphere induces variations in baselines and erratic signals which are not always clearly evident to the analyst. One answer is to use specialist powder cells [14] to prevent water adsorption. Another approach is to use Spatially Offset Raman Spectroscopy (SORS) to analyse CCM powder in sealed packages and thus avoid the water adsorption problem [15], while this works well with simple media components it has not been proven with CCM powders, where fluorescence and the large number of components with varying concentrations are also major issues. MIR and EEM fluorescence have been used to measure media aging effects [16] and the authors here concluded that
Table 1 Some general sample handling and storage requirements for the spectroscopic methods used for cell culture media analysis Powders
Solutions
Hydrolysates
Chemically Defined
Hydrolysates
Chemically Defined
Hygroscopic
Yes, very [42]
Yes, very [16]
Drying required before NIR/ IR/Raman measurements Light sensitivity Protection measures
Yes
Yes
n/a, unless very concentrated n/a
n/a, unless very concentrated n/a
Some Avoid intense or prolonged light exposure Some 20 C, dry Need to minimize potential microbial growth as these materials are usually not sterilized Yes, if sufficient water absorbed Store dry
Some Avoid intense or prolonged light exposure Some 20 C, dry Need to minimize potential microbial growth as these materials are usually not sterilized Yes, if sufficient water absorbed Store dry
Yes Store in dark containers
Yes [23] Store in dark containers
Yes 70 C [20]
Yes [17] 70 C
Yes, very sensitive [22].
Yes, very sensitive
Aseptic sample preparation and 0.2 mm filtering recommended [16]
Aseptic sample preparation and 0.2 mm filtering recommended [16,23]
Temperature sensitivity Long-term storage conditions
Microbial growth Protection measures
Current Opinion in Chemical Engineering 2018, 22:11–17
www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 13
NIR was better. However, the sample handling controls described would suggest that what was observed and measured were the effects of water absorption. For solution analysis, one needs to be aware that the chemical and physical properties are sensitive to microbial growth, chemical reactions in solution, and environmental conditions (temperature, light, etc.). CCM if not sterilized will change very quickly due to microbial growth. The increased bacterial load will make solutions more turbid and thus increase light scatter leading to increased baselines and noise in spectra which may invalidate the measurements. Even when sterilized and stored at low temperature (2–4 C) chemical changes will occur that can have an effect on the spectral measurements and adversely affect the interpretation of those results. These subtle chemical changes in the liquid solution have been detected and monitored using Surface Enhanced Raman Scattering (SERS) spectroscopy, with most of the changes being associated with cysteine/cystine concentration changes [17]. A similar situation exists with environmental water samples where changes in EEM signal [18] were measured during storage at low temperature (2–4 C). These authors investigated the use of preservatives to minimize
microbial growth on their non-sterilized samples and found, as one might expect, that these adversely affected the fluorescence measurements. One approach to minimize and control all these issues which yields very reproducible results is to store all biogenic samples (media or broths) at 70 C [19–22] and then test after defrosting. Most media also contain riboflavin and are thus photochemically unstable in solution and this degradation can easily be measured and monitored using fluorescence spectroscopy [23,24]. High temperature treatment (e.g. pasteurization) of CCM also causes compositional changes [25] and this also needs to be accounted for during sample handling and chemometric modelling. Another aspect which is important are the spurious signals which can originate from the container used to hold solutions. In particular for Raman analysis the use of plastic multi-well plates or glass holders are not recommended because they contribute signals to the already weak and complex sample spectra. The use of electro-polished stainless-steel multi-well plates produced better quality, more reproducible spectra and these holders were also impervious to sample induced corrosion in cases of acidic media components [19,26]. The most common spectroscopic measurement artefacts in CCM analysis are listed in Table 2.
Table 2 Summary of the key strengths, weaknesses, and common measurement issues encountered during spectroscopic analysis of cell culture media. In some published cases, the effect has been mentioned but not explicitly dealt with during measurement
Raman spectroscopy
FT-IR
NIR
Fluorescence (2 or 3D)
www.sciencedirect.com
Key strengths
Weakness
Common measurement issues
Best for powder samples. Minimal sample handling required. High molecular specificity for component identification. Can quantify components at 0.5 w/w% but dependent on the number of components present [37]. Commonly used technology in small molecule API manufacture and bioreactor monitoring in biopharma. Immersion probes available for in-situ measurements. Best for powder samples. High molecular specificity for component identification. Widely used technology in small molecule API manufacture. Best for powder samples. Minimal sample handling required. Widely used technology in small molecule API manufacture. Immersion probes available for in-situ measurements. Milli- and micromolar detection of some analytes like Trp, Tyr, Riboflavin, Pyridoxine possible which are intrinsically fluorescent [29]. Immersion probes available for in-situ measurements.
Liquid formulations, less useful: Good only for > millimolar concentrations, not always possible at that level for complex mixtures [37]. Cannot be used with strongly fluorescent samples.
Weak analyte signals [19,26,38]. Signals from fiber optic probe (usually sapphire) [45]. Signals from sample containers (e.g. glass slides or plastic multi-well plates) [26]. Fluorescence from powder and liquid hydrolysates [22,38,45]. Water contamination in complex powder samples not always obvious.
Less useful for liquid media because of very strong water signal.
Water contamination in powder samples not always obvious. Sub sampling an issue with small sample areas available with ATR systems used for FT-IR. Water contaminated signals, change baselines & bands in solid media [42].
Less useful for liquid media because of very strong water signal. Less good for component identification due to broad and overlapping spectral bands. Liquid samples only, not suitable for powdered media.
Increased scatter from suspended particles and/or microbes [28,36]. Unknown measurement variance because of the reporting of single or unknown replicate measurements [36,38,48].
Current Opinion in Chemical Engineering 2018, 22:11–17
14 Biotechnology and bioprocess engineering: biomanufacturing
UV–vis absorption
Vibrational spectroscopy
The method is quick and can produce data which can enable the identification of gross changes in media quality such as relatively large concentration changes or microbial contamination. For specific media component quantification there are many colorimetric assays based on UV–vis measurements available which could be applied once interferents and matrix effects have been considered. For example, the use of a hexokinase reagent was recently demonstrated for glucose quantification in media [27]. Unfortunately, this approach is too time-consuming and impractical for the analysis of more than a few analytes. However, it may be suitable for some analytes which have a validated high impact on the process of product CQAs. Overall, UV–vis spectroscopy lacks the sensitivity and specificity required to measure molecular variance to a level that is required for routine CCM analysis.
One driver for the use of vibrational spectroscopy is the provision of much more informative molecular fingerprints for each component in a mixture. However, for CCM analysis, the large variety of different analytes and concentration variations severely limit the application of vibrational spectroscopy techniques. A recent review has discussed in detail the use of Raman spectroscopy in biopharma [37] and covered most of the important issues involved in CCM analysis many of which are also relevant to MIR and NIR spectroscopic analysis.
Fluorescence
Most CCM contain multiple fluorophores (Trp, Tyr, Riboflavin, Pyridoxine, etc.) and other photophysically active molecules which either emit or affect fluorescence emission. This means that there are multiple processes such as energy transfer, static quenching, and dynamic quenching which affect the wavelength and intensity of the emission [28]. The complete fluorescence emission space can be recorded with either an Excitation Emission Matrix (EEM) or Total Synchronous Scan Fluorescence (TSFS) measurement and these represent a unique media ‘molecular fingerprint’. EEM is a very effective and useful tool for the qualitative characterization of nearly all the different solutions that are encountered in biopharmaceutical manufacturing: CCM [20,21,23,29], biogenic raw materials [22], bioprocess broths [30–32], and proteins [33]. To achieve this, once needs to correctly use multivariate analysis combined with chemical knowledge, ‘chemometrics’, to best interpret, understand, and use EEM. A recent IUPAC technical document [34] provides more information and guidance about the correct implementation of these methods and the chemometric data analysis procedures. One of the key goals in media analysis is to establish correlations between the measured variance and Critical Process Parameters (CPPs) such as product yield [20] or quality. In one case, EEM fluorescence proved to be much better than Raman spectroscopy, which indicated that the critical components related to process performance were present at too low a concentration to be detected by Raman [19,20,26,35]. A more recent study [36] suggested that EEM fluorescence measurements could be correlated with antioxidant behavior for a soybased hydrolysate and also looked at the filtration process. More recently, the combination of anisotropy with the multidimensional EEM measurement has been used to quantify protein content in a model cell culture media [33] and had similar accuracy to the Bradford assay. Current Opinion in Chemical Engineering 2018, 22:11–17
In 2010, Li et al. showed that Raman spectroscopy could characterize variance and discriminate different CCM components in solution from a commercial recombinant-protein manufacturing process [35]. However, this study and others [38] highlighted some of the difficulties associated with Raman spectroscopy of media solutions: the water signal is stronger than the weak analyte spectra, thus only relatively large composition changes are measurable. SERS [39] could be used to increase sensitivity, however, the compositional complexity of media means that it is difficult to implement. In 2011, Wen et al. suggested SERS as a potential quality-control method for certain cell-culture-media components. They looked at melamine adulteration of several media raw materials and in a commercially-available CCM [40]. They showed reasonable results for simple components (with minimal SERS response) but that for media with multiple components it was not possible. In 2012, SERS was evaluated as a method for assessing variance in yeastolate, a complex hydrolysate [22], which could not be analyzed by Raman because of its intense fluorescence. SERS could detect small variations in composition (mainly associated with adenine) that were associated with different manufacture/source and could be used as a rapid screening tool for lot-to-lot consistency [22]. Unfortunately, a variety of issues including competing surface adsorption processes, the efforts required to development robust measurement protocols, and high numbers of replicate measurements required make it an impractical method for routine CCM analysis [17,22]. MIR & NIR absorption
MIR and NIR because they are absorption based and operate under different selection rules are much more efficient processes than Raman spectroscopy, but they are also much more sensitive to the presence of polar compounds and water in particular. Furthermore, the wavelengths of light used in each technique have significant impacts on the material light interaction which has practical consequences for media analysis. For example, infrared light (approx. 2.5–50 mm) results in very short, sample path lengths in samples (mm) which can lead to sub sampling problems. NIR utilizes shorter-wavelength light (800–2,500 nm) which provides for many practical benefits in terms of sampling and instrument design. As NIR www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 15
spectra comprise combinations and overtones of molecular vibrations and this produces spectra (for solids and liquids) with broad, overlapping bands. This means that NIR does not generate the same level of molecular specificity compared to MIR or Raman spectroscopy. However, the light wavelengths involved enable the use of simple lamp-based light sources and relatively inexpensive glass-based optical components, fibers, and sampling systems. Secondly, the lower molar absorptivity in the NIR allows for longer sample path-lengths/more homogeneous sampling and facilitate sampling through relatively thick container walls. NIR has been applied to media [41] and hydrolysate powder analysis [42] and to assess the quality of complex media such as corn steep liquor used in bioprocesses [43]. However, experimental details for the sample handling and/or measurement processes were insufficiently detailed to enable one to determine the quality/reliability of the NIR data. Some authors have mentioned that the materials are hygroscopic [42], however, no control measures were implemented. An interesting application of FT-IR microscopy and elemental composition analysis to the study of filter fouling by media was demonstrated by Cao et al. [44]. FT-IR microscopy was used to show that most fouling arose from cellular debris arising from microbial contamination rather than the precipitation of specific inorganic components of the media.
Combined analysis
The reliance on a single analytical technique for CCM characterization is risky because cell culture performance is affected by both molecular and elemental components. Thus, one needs a combination of techniques to properly identify and characterize relevant media variance. NIR, fluorescence, X-ray fluorescence, and Raman spectroscopy were used to investigate hydrolysate compositional variability with a goal of predicting CPPs such as integrated viable cell density (IVCD) and product yield from spectral measurements [45]. Techniques were compared individually and then in combination however, details on sample handling and replicate measurements of some techniques (to assess measurement variability) were not provided. Furthermore, the Raman solution spectra presented showed sharp peaks associated with the immersion probe sapphire window. The authors concluded that this data-fusion approach offers significant benefits in assessing media variance. A similar study [38] compared several methods to evaluate wheat hydrolysate variability. Unfortunately, sample preparation and handling details were vague and thus one would be unsure as to the quality/reliability of the solid-state measurements. Likewise, the solution measurements were only single measurements for the fluorescence. Furthermore, the Raman spectra were recorded from solution which does not provide very good quality data because of several factors [19]. www.sciencedirect.com
Conclusions Rapid spectroscopic methods if correctly applied can be used for the rapid and effective screening of cell culture media to identify molecular variance and potential issues with media manufacture. However, the effective application of these methods is very dependent on correct sample handing which in turn is based on an in-depth appreciation of the physical and chemical stability of these complex materials. In many publications, critical information about sample storage, handling, preparation, measurement variance, and robustness was missing which makes an objective assessment of the data quality impossible. Furthermore, it makes repeating these studies difficult. Some of this is due to publisher page/word limits and confidentiality, however the lack of recognized standard measurement methods also plays a part. Chemometrics can deal with some measurement error, however, for CCM one cannot often easily identify measurement and sample induced errors. The quality (robustness, reliability, and repeatability) of the output chemometric models is highly dependent on input spectra quality. For each spectroscopic technique different sample factors (stability, hygroscopicity, etc.) will have greater or lesser impact on the measurements and each should be risk assessed and controlled appropriately (Tables 1 and 2). Once data has been collected it needs to be validated before analysis/interpretation and the utilization of the results for process control. Media analysis requires a set of recognized standard methods and technical guidance documents for spectroscopic measurements similar, for example to those produced by the ASTM [46] and by IUPAC [47].
Conflicts of interest statement Nothing declared.
Acknowledgements This publication was supported in part by a research grant from Science Foundation Ireland (SFI) and is co-funded under the European Regional Development Fund under Grant number (14/IA/2282, Advanced Analytics for Biological Therapeutic Manufacture, to AGR).
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Radhakrishnan D, Robinson AS, Ogunnaike BA: Controlling the glycosylation profile in mAbs using time-dependent media supplementation. Antibodies 2018, 7:21.
2.
Sun XD: Enzymatic hydrolysis of soy proteins and the hydrolysates utilisation. Int J Food Sci Technol 2011, 46:24472459.
3.
Luo Y, Pierce KM: Development toward rapid and efficient screening for high performance hydrolysate lots in a recombinant monoclonal antibody manufacturing process. Biotechnol Progr 2012, 28:1061-1068. Current Opinion in Chemical Engineering 2018, 22:11–17
16 Biotechnology and bioprocess engineering: biomanufacturing
4.
Gupta AJ, Gruppen H, Maes D, Boots JW, Wierenga PA: Factors causing compositional changes in soy protein hydrolysates and effects on cell culture functionality. J Agric Food Chem 2013, 61:10613-10625.
21. Li B, Ryan PW, Shanahan M, Leister KJ, Ryder AG: Fluorescence excitation-emission matrix (EEM) spectroscopy for rapid identification and quality evaluation of cell culture media components. Appl Spectrosc 2011, 65:1240-1249.
5.
Gupta AJ, Hageman JA, Wierenga PA, Boots JW, Gruppen H: Chemometric analysis of soy protein hydrolysates used in animal cell culture for IgG production - an untargeted metabolomics approach. Process Biochem 2014, 49:309-317.
22. Li B, Sirimuthu NMS, Ray BH, Ryder AG: Using surface enhanced Raman scattering (SERS) and fluorescence spectroscopy for screening yeast extracts, a complex component of cell culture media. J Raman Spectrosc 2012, 43:1074-1082.
6.
Ho SCL, Nian R, Woen S, Chng J, Zhang PQ, Yang YS: Impact of hydrolysates on monoclonal antibody productivity, purification and quality in Chinese hamster ovary cells. J Biosci Bioeng 2016, 122:499-506.
23. Calvet A, Li B, Ryder AG: A rapid fluorescence based method for the quantitative analysis of cell culture media photodegradation. Anal Chim Acta 2014, 807:111-119.
7.
Spearman M, Chan S, Jung V, Kowbel V, Mendoza M, Miranda V, Butler M: Components of yeast (Sacchromyces cervisiae) extract as defined media additives that support the growth and productivity of CHO cells. J Biotechnol 2016, 233:129-142.
24. Neutsch L, Kroll P, Brunner M, Pansy A, Kovar M, Herwig C, Klein T: Media photo-degradation in pharmaceutical biotechnology – impact of ambient light on media quality, cell physiology, and IgG production in CHO cultures. J Chem Technol Biotechnol 2018, 93:2141-2151.
8.
Mitchelson FG, Mondia JP, Hughes EH: Effect of copper variation in yeast hydrolysate on C-terminal lysine levels of a monoclonal antibody. Biotechnol Progr 2017, 33:463-468.
Rathore AS, Kumar D, Kateja N: Role of raw materials in biopharmaceutical manufacturing: risk analysis and fingerprinting. Curr Opin Biotechnol 2018, 53:99-105. A good overview of many of the critical issues involved with the raw materials and cell culture media used in biopharmaceutical manufacturing.
9.
10. Read EK, Park JT, Shah RB, Riley BS, Brorson KA, Rathore AS: Process Analytical Technology (PAT) for biopharmaceutical products: part I. Concepts and applications. Biotechnol Bioeng 2010, 105:276-284. 11. Read EK, Shah RB, Riley BS, Park JT, Brorson KA, Rathore AS: Process Analytical Technology (PAT) for biopharmaceutical products: part II. Concepts and applications. Biotechnol Bioeng 2010, 105:285-295. 12. Rathore AS, Bhambure R, Ghare V: Process analytical technology (PAT) for biopharmaceutical products. Anal Bioanal Chem 2010, 398:137-154. 13. Rathore AS, Bhambure R, Shirke A, Rajashekhar S, Muthukumar S: Enabling Technologies for Implementation of Quality by Design for Biopharmaceuticals. Abstracts of Papers of the American Chemical Society; 2011:241. 14. Lekgoathi MDS, le Roux JP: Attenuated total reflectance powder cell for infrared analysis of hygroscopic samples. Spectrochim Acta Part A-Mol Biomol Spectr 2011, 82:529-531. 15. Bloomfield M, Andrews D, Loeffen P, Tombling C, York T, Matousek P: Non-invasive identification of incoming raw pharmaceutical materials using spatially offset Raman spectroscopy. J Pharm Biomed Anal 2013, 76:65-69. 16. Hakemeyer C, Strauss U, Werz S, Folque F, Menezes JC: Nearinfrared and two-dimensional fluorescence spectroscopy monitoring of monoclonal antibody fermentation media quality: aged media decreases cell growth. Biotechnol J 2013, 8 835-+. 17. Calvet A, Ryder AG: Monitoring cell culture media degradation using surface enhanced Raman scattering (SERS) spectroscopy. Anal Chim Acta 2014, 840:58-67. 18. Park M, Snyder SA: Sample handling and data processing for fluorescent excitation-emission matrix (EEM) of dissolved organic matter (DOM). Chemosphere 2018, 193:530-537. A good example of some of the issues involved in the measurement of complex samples via fluorescence spectroscopy. This case is for water analysis and is also relevant for media analysis. However, the use of preservatives is not recommended. 19. Li B, Ryan PW, Ray BH, Leister KJ, Sirimuthu NMS, Ryder AG: Rapid characterisation and quality control of complex cell culture media using Raman spectroscopy and chemometrics. Biotechnol Bioeng 2010, 107:290-301. 20. Ryan PW, Li B, Shanahan M, Leister KJ, Ryder AG: Prediction of cell culture media performance using fluorescence spectroscopy. Anal Chem 2010, 82:1311-1317. Current Opinion in Chemical Engineering 2018, 22:11–17
25. Floris P, McGillicuddy N, Albrecht S, Morrissey B, Kaisermayer C, Lindeberg A, Bones J: Untargeted LC-MS/MS profiling of cell culture media formulations for evaluation of high temperature short time treatment effects. Anal Chem 2017, 89:9953-9960. 26. Ryder AG, De Vincentis J, Li BY, Ryan PW, Sirimuthu NMS, Leister KJ: A stainless steel multi-well plate (SS-MWP) for highthroughput Raman analysis of dilute solutions. J Raman Spectry 2010, 41:1266-1275. 27. Mallya S, Lay B, McAleer L, Emory A, Afonina N: A platform approach for the identity testing of multi-component cellculture media. BioPharm Int 2016, 29:30-39. 28. Lakowicz JR: Principles of Fluorescence Spectroscopy. 3rd edn. New York: Springer; 2006. 29. Calvet C, Li B, Ryder AG: Rapid quantification of tryptophan and tyrosine in chemically defined cell culture media using fluorescence spectroscopy. J Pharm Biomed Anal 2012, 71:8998. 30. Li B, Shanahan M, Calvet A, Leister KJ, Ryder AG: Comprehensive, quantitative bioprocess productivity monitoring using fluorescence EEM spectroscopy and chemometrics. Analyst 2014, 139:1661-1671. 31. Ohadi K, Legge RL, Budman HM: Development of a soft-sensor based on multi-wavelength fluorescence spectroscopy and a dynamic metabolic model for monitoring mammalian cell cultures. Biotechnol Bioeng 2015, 112:197-208. 32. Rowland-Jones RC, van den Berg F, Racher AJ, Martin EB, Jaques C: Comparison of spectroscopy technologies for improved monitoring of cell culture processes in miniature bioreactors. Biotechnol Progr 2017, 33:337-346. 33. Groza RC, Calvet A, Ryder AG: A fluorescence anisotropy method for measuring protein concentration in complex cell culture media. Anal Chim Acta 2014, 821:54-61. 34. Ryder AG, Stedmon CA, Harrit N, Bro R: Calibration, standardization, and quantitative analysis of multidimensional fluorescence (MDF) measurements on complex mixtures (IUPAC Technical Report). Pure Appl Chem 2017, 89:1849-1870. A recent report which covers all of the major issues involved in the use of multidimensional fluorescence spectroscopy measurements of complex samples such as CCM. Provides a series of recommendations that should facilitate the collection of more reproducible data and correct implementation of chemometric methods of analysis. 35. Li B, Ray BH, Leister KJ, Ryder AG: Performance monitoring of a mammalian cell based bioprocess using Raman spectroscopy. Anal Chim Acta 2013, 796:84-91. 36. Ranamukhaarachchi SA, Peiris RH, Moresoli C: Fluorescence spectroscopy and principal component analysis of soy protein hydrolysate fractions and the potential to assess their antioxidant capacity characteristics. Food Chem 2017, 217:469-475. 37. Buckley K, Ryder AG: Applications of Raman spectroscopy in biopharmaceutical manufacturing: a short review. Appl Spectrosc 2017, 71:1085-1116.
www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 17
Recent overview of the application of Raman spectroscopy to applications in the biopharma sector. Includes assessments of the major issues with a variety of different samples included solid and solution media. 38. Trunfio N, Lee H, Starkey J, Agarabi C, Liu J, Yoon S: Characterization of mammalian cell culture raw materials by combining spectroscopy and chemometrics. Biotechnol Prog 2017, 33:1127-1138. 39. Ryder AG: Surface enhanced Raman scattering for narcotic detection and applications to chemical biology. Curr Opin Chem Biol 2005, 9:489-493. 40. Wen ZQ, Li GY, Ren D: Detection of trace melamine in raw materials used for protein pharmaceutical manufacturing using surface-enhanced Raman spectroscopy (SERS) with gold nanoparticles. Appl Spectrosc 2011, 65:514-521. 41. Kirdar AO, Chen GX, Weidner J, Rathore AS: Application of nearInfrared (NIR) spectroscopy for screening of raw materials used in the cell culture medium for the production of a recombinant therapeutic protein. Biotechnol Progr 2010, 26:527-531. 42. Li G, Wen Z-q: Screening soy hydrolysates for the production of a recombinant therapeutic protein in commercial cell line by combined approach of near-infrared spectroscopy and chemometrics. Appl Microbiol Biotechnol 2013, 97:2653-2660. 43. Hofer A, Kamravamanesh D, Bona-Lovasz J, Limbeck A, Lendl B, Herwig C, Fricke J: Prediction of filamentous process
www.sciencedirect.com
performance attributes by CSL quality assessment using midinfrared spectroscopy and chemometrics. J Biotechnol 2018, 265:93-100. 44. Cao XL, Loussaert JA, Wen ZQ: Microspectroscopic investigation of the membrane clogging during the sterile filtration of the growth media for mammalian cell culture. J. Pharm. Biomed. Anal. 2016, 119:10-15. A nice example of the application of FT-IR spectroscopy to the analysis of filter membrane clogging. Experimental details are comprehensive and the integration with EDAX results for elemental screening is very useful. 45. Lee HW, Christie A, Xu J, Yoon S: Data fusion-based assessment of raw materials in mammalian cell culture. Biotechnol Bioeng 2012, 109:2819-2828. 46. ASTM E578-01: Standard Test Method for Linearity of Fluorescence Measuring Systems. 100 Barr Harbor Drive, West Conshohocken, PA: ASTM International; 2001. 47. Ameloot M, vandeVen M, Acuna AU, Valeur B: Fluorescence anisotropy measurements in solution: methods and reference materials (IUPAC technical report). Pure Appl Chem 2013, 85:589-608. 48. Jose GE, Folque F, Menezes JC, Werz S, Strauss U, Hakemeyer C: Predicting mAb product yields from cultivation media components, using near-infrared and 2D-fluorescence spectroscopies. Biotechnol Progr 2011, 27:1339-1346.
Current Opinion in Chemical Engineering 2018, 22:11–17
ScienceDirect Cell culture media analysis using rapid spectroscopic methods Alan G Ryder Cell culture media (CCM) used in industrial biotechnology are complex mixtures of molecules and elements that are inherently difficult to analyze comprehensively. CCM quality analysis is of utmost importance for efficient production of protein-based therapeutics with the correct Critical Quality Attributes (CQAs). Here we discuss the use of rapid spectroscopic methods for routine screening of CCM molecular variance which include electronic (UV–vis absorption and fluorescence) and vibrational (Raman, FT-IR, and NearInfra-red) spectroscopies. CCM analysis needs to provide: identity testing, compositional variance analysis (i.e. lot-to-lot variation), validation of media preparation protocols, and correlations with Critical Process Parameters (CPPs) or product CQAs. Rapid spectroscopic methods can fulfil some of these requirements but only with correct sample handling and preparation. Accurate analysis requires the use of chemometrics combined with a detailed knowledge of sample behavior such as water absorption and chemical stability. Address Nanoscale Biophotonics Laboratory, School of Chemistry, National University of Ireland-Galway, University Road, Galway, H91 CF50, Ireland Corresponding author: Ryder, Alan G ([email protected])
Current Opinion in Chemical Engineering 2018, 22:11–17 This review comes from a themed issue on Biotechnology and bioprocess engineering: biomanufacturing Edited by Michael Betenbaugh and Nigel Titchener-Hooker
https://doi.org/10.1016/j.coche.2018.08.008
therapeutics CQAs include factors such as the glycosylation pattern, stability, aggregation, impurity profiles, and so on. Many CQAs are directly affected by CCM [1] and unknown changes in composition can have very serious adverse consequences for the process. CCM are carefully designed and optimized for each process (and often for different stages in the process) and cell line, which results in a very wide range of different CCM formulations. The resulting media are highly complex, molecular and elemental mixtures, containing amino acids, carbohydrates, vitamins, growth factors, trace and bulk minerals. The most complex of the raw material components used in media manufacture are the various hydrolysates of protein sources (like yeast or soy). These hydrolysates are a complex mixture of carbohydrates, amino acids, peptides, and a host of other often unknown molecules and minerals [2–5]. These contain many more components than chemically defined media and often display much greater molecular and elemental variance but are often critical for antibody manufacturing processes [6]. This variance can arise from changes in source material, manufacturing issues, or improper handling/storage. For these materials, both molecular [7] and elemental variance are important [8] in the context of their effect on cell culture. This molecular and elemental variance impacts on CCM quality in terms of its intended purpose, namely sustaining a specific bioprocess. Therefore, when we use the term media quality it refers to a specific media and a specific process. It should be noted that the same media could be used in two different processes and for each process, the effect of media variance may be different.
2211-3398/ã 2018 Published by Elsevier Ltd.
Introduction Cell culture media (CCM) are a critical element in biopharmaceutical manufacturing which directly affect process yield and product quality. Therefore, the analysis of media variance and prediction of performance is of high importance. The essential purpose of CCM is to produce and maintain an optimum physiological environment to allow large-scale culturing of cells where they remain healthy for an appropriate amount of time and express the right product with the correct Critical Quality Attributes (CQAs). For protein-based www.sciencedirect.com
Another major issue is the fact that components are present in widely varying concentrations, from trace (ppm) levels for some elements to >5% w/w for carbohydrates. In liquid media water is the major component (>90% w/w) diluting components (millimolar, micromolar, and nanomolar concentrations) and making analyte detection more difficult for spectroscopic methods. Because of these compositional and concentration factors, media analysis using traditional analytical methods, is challenging, very time consuming, requiring multiple methods and techniques. This makes comprehensive analysis too expensive for routine use. As the industrial cell culture production of therapeutic proteins continues to grow so too does the need for rapid, inexpensive, reliable, robust, and non-destructive analytical methods that can be integrated into process control to improve end Current Opinion in Chemical Engineering 2018, 22:11–17
12 Biotechnology and bioprocess engineering: biomanufacturing
product yield and quality. A recent review [9] provides a more in-depth assessment of these raw materials issues. Whereas here we focus on the use of spectroscopic measurements on the media and take a critical look into the quality of the data being generated. The quality of the spectral data that these rapid methods can provide is very sensitive to sample handling and failure to recognize this can lead to the generation of erroneous data and ultimately misleading conclusions.
Fast spectroscopic methods
There are many potential spectroscopic methods that can be applied to media analysis, however, here we limit ourselves to commercially available, tried and tested methods. Relatively recent reviews describe the needs for Process Analytical Technologies (PAT) in biopharmaceutical manufacturing [10–13] and provide more background detail about how rapid spectroscopic techniques could be incorporated into manufacturing processes and are directly relevant to CCM analysis. The most important methods are the electronic spectroscopies (UV–vis absorption and fluorescence) and vibrational spectroscopies (MIR and NIR absorption and Raman scattering). The correct application of all methods requires an indepth appreciation of the practical difficulties of dealing with chemically complex mixtures. Sample handling therefore plays a critical role in the collection of accurate, correct, and reproducible data which is required for the development of robust analytical methods.
Sample handling
A critical issue with CCM analysis relates to sample handling and stability both in powdered and solution forms (Table 1). CCM have complex physicochemical properties which mean that the samples are neither chemically nor physically stable for varying time periods of minutes to hours depending on the type. It is very important to recognize here that stability in the context of measurements is different from stability associated with CCM performance in a bioprocess. For example, CCM powders are usually hygroscopic, and will quickly absorb atmospheric water rapidly (within a couple of minutes) and are thus unsuitable for vibrational spectroscopy analysis of molecular variance (as opposed to just measuring the effects of water absorption) unless properly dried. Whereas small amounts of absorbed water are unlikely to have any impact on cell culture performance. Water adsorption from the atmosphere induces variations in baselines and erratic signals which are not always clearly evident to the analyst. One answer is to use specialist powder cells [14] to prevent water adsorption. Another approach is to use Spatially Offset Raman Spectroscopy (SORS) to analyse CCM powder in sealed packages and thus avoid the water adsorption problem [15], while this works well with simple media components it has not been proven with CCM powders, where fluorescence and the large number of components with varying concentrations are also major issues. MIR and EEM fluorescence have been used to measure media aging effects [16] and the authors here concluded that
Table 1 Some general sample handling and storage requirements for the spectroscopic methods used for cell culture media analysis Powders
Solutions
Hydrolysates
Chemically Defined
Hydrolysates
Chemically Defined
Hygroscopic
Yes, very [42]
Yes, very [16]
Drying required before NIR/ IR/Raman measurements Light sensitivity Protection measures
Yes
Yes
n/a, unless very concentrated n/a
n/a, unless very concentrated n/a
Some Avoid intense or prolonged light exposure Some 20 C, dry Need to minimize potential microbial growth as these materials are usually not sterilized Yes, if sufficient water absorbed Store dry
Some Avoid intense or prolonged light exposure Some 20 C, dry Need to minimize potential microbial growth as these materials are usually not sterilized Yes, if sufficient water absorbed Store dry
Yes Store in dark containers
Yes [23] Store in dark containers
Yes 70 C [20]
Yes [17] 70 C
Yes, very sensitive [22].
Yes, very sensitive
Aseptic sample preparation and 0.2 mm filtering recommended [16]
Aseptic sample preparation and 0.2 mm filtering recommended [16,23]
Temperature sensitivity Long-term storage conditions
Microbial growth Protection measures
Current Opinion in Chemical Engineering 2018, 22:11–17
www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 13
NIR was better. However, the sample handling controls described would suggest that what was observed and measured were the effects of water absorption. For solution analysis, one needs to be aware that the chemical and physical properties are sensitive to microbial growth, chemical reactions in solution, and environmental conditions (temperature, light, etc.). CCM if not sterilized will change very quickly due to microbial growth. The increased bacterial load will make solutions more turbid and thus increase light scatter leading to increased baselines and noise in spectra which may invalidate the measurements. Even when sterilized and stored at low temperature (2–4 C) chemical changes will occur that can have an effect on the spectral measurements and adversely affect the interpretation of those results. These subtle chemical changes in the liquid solution have been detected and monitored using Surface Enhanced Raman Scattering (SERS) spectroscopy, with most of the changes being associated with cysteine/cystine concentration changes [17]. A similar situation exists with environmental water samples where changes in EEM signal [18] were measured during storage at low temperature (2–4 C). These authors investigated the use of preservatives to minimize
microbial growth on their non-sterilized samples and found, as one might expect, that these adversely affected the fluorescence measurements. One approach to minimize and control all these issues which yields very reproducible results is to store all biogenic samples (media or broths) at 70 C [19–22] and then test after defrosting. Most media also contain riboflavin and are thus photochemically unstable in solution and this degradation can easily be measured and monitored using fluorescence spectroscopy [23,24]. High temperature treatment (e.g. pasteurization) of CCM also causes compositional changes [25] and this also needs to be accounted for during sample handling and chemometric modelling. Another aspect which is important are the spurious signals which can originate from the container used to hold solutions. In particular for Raman analysis the use of plastic multi-well plates or glass holders are not recommended because they contribute signals to the already weak and complex sample spectra. The use of electro-polished stainless-steel multi-well plates produced better quality, more reproducible spectra and these holders were also impervious to sample induced corrosion in cases of acidic media components [19,26]. The most common spectroscopic measurement artefacts in CCM analysis are listed in Table 2.
Table 2 Summary of the key strengths, weaknesses, and common measurement issues encountered during spectroscopic analysis of cell culture media. In some published cases, the effect has been mentioned but not explicitly dealt with during measurement
Raman spectroscopy
FT-IR
NIR
Fluorescence (2 or 3D)
www.sciencedirect.com
Key strengths
Weakness
Common measurement issues
Best for powder samples. Minimal sample handling required. High molecular specificity for component identification. Can quantify components at 0.5 w/w% but dependent on the number of components present [37]. Commonly used technology in small molecule API manufacture and bioreactor monitoring in biopharma. Immersion probes available for in-situ measurements. Best for powder samples. High molecular specificity for component identification. Widely used technology in small molecule API manufacture. Best for powder samples. Minimal sample handling required. Widely used technology in small molecule API manufacture. Immersion probes available for in-situ measurements. Milli- and micromolar detection of some analytes like Trp, Tyr, Riboflavin, Pyridoxine possible which are intrinsically fluorescent [29]. Immersion probes available for in-situ measurements.
Liquid formulations, less useful: Good only for > millimolar concentrations, not always possible at that level for complex mixtures [37]. Cannot be used with strongly fluorescent samples.
Weak analyte signals [19,26,38]. Signals from fiber optic probe (usually sapphire) [45]. Signals from sample containers (e.g. glass slides or plastic multi-well plates) [26]. Fluorescence from powder and liquid hydrolysates [22,38,45]. Water contamination in complex powder samples not always obvious.
Less useful for liquid media because of very strong water signal.
Water contamination in powder samples not always obvious. Sub sampling an issue with small sample areas available with ATR systems used for FT-IR. Water contaminated signals, change baselines & bands in solid media [42].
Less useful for liquid media because of very strong water signal. Less good for component identification due to broad and overlapping spectral bands. Liquid samples only, not suitable for powdered media.
Increased scatter from suspended particles and/or microbes [28,36]. Unknown measurement variance because of the reporting of single or unknown replicate measurements [36,38,48].
Current Opinion in Chemical Engineering 2018, 22:11–17
14 Biotechnology and bioprocess engineering: biomanufacturing
UV–vis absorption
Vibrational spectroscopy
The method is quick and can produce data which can enable the identification of gross changes in media quality such as relatively large concentration changes or microbial contamination. For specific media component quantification there are many colorimetric assays based on UV–vis measurements available which could be applied once interferents and matrix effects have been considered. For example, the use of a hexokinase reagent was recently demonstrated for glucose quantification in media [27]. Unfortunately, this approach is too time-consuming and impractical for the analysis of more than a few analytes. However, it may be suitable for some analytes which have a validated high impact on the process of product CQAs. Overall, UV–vis spectroscopy lacks the sensitivity and specificity required to measure molecular variance to a level that is required for routine CCM analysis.
One driver for the use of vibrational spectroscopy is the provision of much more informative molecular fingerprints for each component in a mixture. However, for CCM analysis, the large variety of different analytes and concentration variations severely limit the application of vibrational spectroscopy techniques. A recent review has discussed in detail the use of Raman spectroscopy in biopharma [37] and covered most of the important issues involved in CCM analysis many of which are also relevant to MIR and NIR spectroscopic analysis.
Fluorescence
Most CCM contain multiple fluorophores (Trp, Tyr, Riboflavin, Pyridoxine, etc.) and other photophysically active molecules which either emit or affect fluorescence emission. This means that there are multiple processes such as energy transfer, static quenching, and dynamic quenching which affect the wavelength and intensity of the emission [28]. The complete fluorescence emission space can be recorded with either an Excitation Emission Matrix (EEM) or Total Synchronous Scan Fluorescence (TSFS) measurement and these represent a unique media ‘molecular fingerprint’. EEM is a very effective and useful tool for the qualitative characterization of nearly all the different solutions that are encountered in biopharmaceutical manufacturing: CCM [20,21,23,29], biogenic raw materials [22], bioprocess broths [30–32], and proteins [33]. To achieve this, once needs to correctly use multivariate analysis combined with chemical knowledge, ‘chemometrics’, to best interpret, understand, and use EEM. A recent IUPAC technical document [34] provides more information and guidance about the correct implementation of these methods and the chemometric data analysis procedures. One of the key goals in media analysis is to establish correlations between the measured variance and Critical Process Parameters (CPPs) such as product yield [20] or quality. In one case, EEM fluorescence proved to be much better than Raman spectroscopy, which indicated that the critical components related to process performance were present at too low a concentration to be detected by Raman [19,20,26,35]. A more recent study [36] suggested that EEM fluorescence measurements could be correlated with antioxidant behavior for a soybased hydrolysate and also looked at the filtration process. More recently, the combination of anisotropy with the multidimensional EEM measurement has been used to quantify protein content in a model cell culture media [33] and had similar accuracy to the Bradford assay. Current Opinion in Chemical Engineering 2018, 22:11–17
In 2010, Li et al. showed that Raman spectroscopy could characterize variance and discriminate different CCM components in solution from a commercial recombinant-protein manufacturing process [35]. However, this study and others [38] highlighted some of the difficulties associated with Raman spectroscopy of media solutions: the water signal is stronger than the weak analyte spectra, thus only relatively large composition changes are measurable. SERS [39] could be used to increase sensitivity, however, the compositional complexity of media means that it is difficult to implement. In 2011, Wen et al. suggested SERS as a potential quality-control method for certain cell-culture-media components. They looked at melamine adulteration of several media raw materials and in a commercially-available CCM [40]. They showed reasonable results for simple components (with minimal SERS response) but that for media with multiple components it was not possible. In 2012, SERS was evaluated as a method for assessing variance in yeastolate, a complex hydrolysate [22], which could not be analyzed by Raman because of its intense fluorescence. SERS could detect small variations in composition (mainly associated with adenine) that were associated with different manufacture/source and could be used as a rapid screening tool for lot-to-lot consistency [22]. Unfortunately, a variety of issues including competing surface adsorption processes, the efforts required to development robust measurement protocols, and high numbers of replicate measurements required make it an impractical method for routine CCM analysis [17,22]. MIR & NIR absorption
MIR and NIR because they are absorption based and operate under different selection rules are much more efficient processes than Raman spectroscopy, but they are also much more sensitive to the presence of polar compounds and water in particular. Furthermore, the wavelengths of light used in each technique have significant impacts on the material light interaction which has practical consequences for media analysis. For example, infrared light (approx. 2.5–50 mm) results in very short, sample path lengths in samples (mm) which can lead to sub sampling problems. NIR utilizes shorter-wavelength light (800–2,500 nm) which provides for many practical benefits in terms of sampling and instrument design. As NIR www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 15
spectra comprise combinations and overtones of molecular vibrations and this produces spectra (for solids and liquids) with broad, overlapping bands. This means that NIR does not generate the same level of molecular specificity compared to MIR or Raman spectroscopy. However, the light wavelengths involved enable the use of simple lamp-based light sources and relatively inexpensive glass-based optical components, fibers, and sampling systems. Secondly, the lower molar absorptivity in the NIR allows for longer sample path-lengths/more homogeneous sampling and facilitate sampling through relatively thick container walls. NIR has been applied to media [41] and hydrolysate powder analysis [42] and to assess the quality of complex media such as corn steep liquor used in bioprocesses [43]. However, experimental details for the sample handling and/or measurement processes were insufficiently detailed to enable one to determine the quality/reliability of the NIR data. Some authors have mentioned that the materials are hygroscopic [42], however, no control measures were implemented. An interesting application of FT-IR microscopy and elemental composition analysis to the study of filter fouling by media was demonstrated by Cao et al. [44]. FT-IR microscopy was used to show that most fouling arose from cellular debris arising from microbial contamination rather than the precipitation of specific inorganic components of the media.
Combined analysis
The reliance on a single analytical technique for CCM characterization is risky because cell culture performance is affected by both molecular and elemental components. Thus, one needs a combination of techniques to properly identify and characterize relevant media variance. NIR, fluorescence, X-ray fluorescence, and Raman spectroscopy were used to investigate hydrolysate compositional variability with a goal of predicting CPPs such as integrated viable cell density (IVCD) and product yield from spectral measurements [45]. Techniques were compared individually and then in combination however, details on sample handling and replicate measurements of some techniques (to assess measurement variability) were not provided. Furthermore, the Raman solution spectra presented showed sharp peaks associated with the immersion probe sapphire window. The authors concluded that this data-fusion approach offers significant benefits in assessing media variance. A similar study [38] compared several methods to evaluate wheat hydrolysate variability. Unfortunately, sample preparation and handling details were vague and thus one would be unsure as to the quality/reliability of the solid-state measurements. Likewise, the solution measurements were only single measurements for the fluorescence. Furthermore, the Raman spectra were recorded from solution which does not provide very good quality data because of several factors [19]. www.sciencedirect.com
Conclusions Rapid spectroscopic methods if correctly applied can be used for the rapid and effective screening of cell culture media to identify molecular variance and potential issues with media manufacture. However, the effective application of these methods is very dependent on correct sample handing which in turn is based on an in-depth appreciation of the physical and chemical stability of these complex materials. In many publications, critical information about sample storage, handling, preparation, measurement variance, and robustness was missing which makes an objective assessment of the data quality impossible. Furthermore, it makes repeating these studies difficult. Some of this is due to publisher page/word limits and confidentiality, however the lack of recognized standard measurement methods also plays a part. Chemometrics can deal with some measurement error, however, for CCM one cannot often easily identify measurement and sample induced errors. The quality (robustness, reliability, and repeatability) of the output chemometric models is highly dependent on input spectra quality. For each spectroscopic technique different sample factors (stability, hygroscopicity, etc.) will have greater or lesser impact on the measurements and each should be risk assessed and controlled appropriately (Tables 1 and 2). Once data has been collected it needs to be validated before analysis/interpretation and the utilization of the results for process control. Media analysis requires a set of recognized standard methods and technical guidance documents for spectroscopic measurements similar, for example to those produced by the ASTM [46] and by IUPAC [47].
Conflicts of interest statement Nothing declared.
Acknowledgements This publication was supported in part by a research grant from Science Foundation Ireland (SFI) and is co-funded under the European Regional Development Fund under Grant number (14/IA/2282, Advanced Analytics for Biological Therapeutic Manufacture, to AGR).
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Radhakrishnan D, Robinson AS, Ogunnaike BA: Controlling the glycosylation profile in mAbs using time-dependent media supplementation. Antibodies 2018, 7:21.
2.
Sun XD: Enzymatic hydrolysis of soy proteins and the hydrolysates utilisation. Int J Food Sci Technol 2011, 46:24472459.
3.
Luo Y, Pierce KM: Development toward rapid and efficient screening for high performance hydrolysate lots in a recombinant monoclonal antibody manufacturing process. Biotechnol Progr 2012, 28:1061-1068. Current Opinion in Chemical Engineering 2018, 22:11–17
16 Biotechnology and bioprocess engineering: biomanufacturing
4.
Gupta AJ, Gruppen H, Maes D, Boots JW, Wierenga PA: Factors causing compositional changes in soy protein hydrolysates and effects on cell culture functionality. J Agric Food Chem 2013, 61:10613-10625.
21. Li B, Ryan PW, Shanahan M, Leister KJ, Ryder AG: Fluorescence excitation-emission matrix (EEM) spectroscopy for rapid identification and quality evaluation of cell culture media components. Appl Spectrosc 2011, 65:1240-1249.
5.
Gupta AJ, Hageman JA, Wierenga PA, Boots JW, Gruppen H: Chemometric analysis of soy protein hydrolysates used in animal cell culture for IgG production - an untargeted metabolomics approach. Process Biochem 2014, 49:309-317.
22. Li B, Sirimuthu NMS, Ray BH, Ryder AG: Using surface enhanced Raman scattering (SERS) and fluorescence spectroscopy for screening yeast extracts, a complex component of cell culture media. J Raman Spectrosc 2012, 43:1074-1082.
6.
Ho SCL, Nian R, Woen S, Chng J, Zhang PQ, Yang YS: Impact of hydrolysates on monoclonal antibody productivity, purification and quality in Chinese hamster ovary cells. J Biosci Bioeng 2016, 122:499-506.
23. Calvet A, Li B, Ryder AG: A rapid fluorescence based method for the quantitative analysis of cell culture media photodegradation. Anal Chim Acta 2014, 807:111-119.
7.
Spearman M, Chan S, Jung V, Kowbel V, Mendoza M, Miranda V, Butler M: Components of yeast (Sacchromyces cervisiae) extract as defined media additives that support the growth and productivity of CHO cells. J Biotechnol 2016, 233:129-142.
24. Neutsch L, Kroll P, Brunner M, Pansy A, Kovar M, Herwig C, Klein T: Media photo-degradation in pharmaceutical biotechnology – impact of ambient light on media quality, cell physiology, and IgG production in CHO cultures. J Chem Technol Biotechnol 2018, 93:2141-2151.
8.
Mitchelson FG, Mondia JP, Hughes EH: Effect of copper variation in yeast hydrolysate on C-terminal lysine levels of a monoclonal antibody. Biotechnol Progr 2017, 33:463-468.
Rathore AS, Kumar D, Kateja N: Role of raw materials in biopharmaceutical manufacturing: risk analysis and fingerprinting. Curr Opin Biotechnol 2018, 53:99-105. A good overview of many of the critical issues involved with the raw materials and cell culture media used in biopharmaceutical manufacturing.
9.
10. Read EK, Park JT, Shah RB, Riley BS, Brorson KA, Rathore AS: Process Analytical Technology (PAT) for biopharmaceutical products: part I. Concepts and applications. Biotechnol Bioeng 2010, 105:276-284. 11. Read EK, Shah RB, Riley BS, Park JT, Brorson KA, Rathore AS: Process Analytical Technology (PAT) for biopharmaceutical products: part II. Concepts and applications. Biotechnol Bioeng 2010, 105:285-295. 12. Rathore AS, Bhambure R, Ghare V: Process analytical technology (PAT) for biopharmaceutical products. Anal Bioanal Chem 2010, 398:137-154. 13. Rathore AS, Bhambure R, Shirke A, Rajashekhar S, Muthukumar S: Enabling Technologies for Implementation of Quality by Design for Biopharmaceuticals. Abstracts of Papers of the American Chemical Society; 2011:241. 14. Lekgoathi MDS, le Roux JP: Attenuated total reflectance powder cell for infrared analysis of hygroscopic samples. Spectrochim Acta Part A-Mol Biomol Spectr 2011, 82:529-531. 15. Bloomfield M, Andrews D, Loeffen P, Tombling C, York T, Matousek P: Non-invasive identification of incoming raw pharmaceutical materials using spatially offset Raman spectroscopy. J Pharm Biomed Anal 2013, 76:65-69. 16. Hakemeyer C, Strauss U, Werz S, Folque F, Menezes JC: Nearinfrared and two-dimensional fluorescence spectroscopy monitoring of monoclonal antibody fermentation media quality: aged media decreases cell growth. Biotechnol J 2013, 8 835-+. 17. Calvet A, Ryder AG: Monitoring cell culture media degradation using surface enhanced Raman scattering (SERS) spectroscopy. Anal Chim Acta 2014, 840:58-67. 18. Park M, Snyder SA: Sample handling and data processing for fluorescent excitation-emission matrix (EEM) of dissolved organic matter (DOM). Chemosphere 2018, 193:530-537. A good example of some of the issues involved in the measurement of complex samples via fluorescence spectroscopy. This case is for water analysis and is also relevant for media analysis. However, the use of preservatives is not recommended. 19. Li B, Ryan PW, Ray BH, Leister KJ, Sirimuthu NMS, Ryder AG: Rapid characterisation and quality control of complex cell culture media using Raman spectroscopy and chemometrics. Biotechnol Bioeng 2010, 107:290-301. 20. Ryan PW, Li B, Shanahan M, Leister KJ, Ryder AG: Prediction of cell culture media performance using fluorescence spectroscopy. Anal Chem 2010, 82:1311-1317. Current Opinion in Chemical Engineering 2018, 22:11–17
25. Floris P, McGillicuddy N, Albrecht S, Morrissey B, Kaisermayer C, Lindeberg A, Bones J: Untargeted LC-MS/MS profiling of cell culture media formulations for evaluation of high temperature short time treatment effects. Anal Chem 2017, 89:9953-9960. 26. Ryder AG, De Vincentis J, Li BY, Ryan PW, Sirimuthu NMS, Leister KJ: A stainless steel multi-well plate (SS-MWP) for highthroughput Raman analysis of dilute solutions. J Raman Spectry 2010, 41:1266-1275. 27. Mallya S, Lay B, McAleer L, Emory A, Afonina N: A platform approach for the identity testing of multi-component cellculture media. BioPharm Int 2016, 29:30-39. 28. Lakowicz JR: Principles of Fluorescence Spectroscopy. 3rd edn. New York: Springer; 2006. 29. Calvet C, Li B, Ryder AG: Rapid quantification of tryptophan and tyrosine in chemically defined cell culture media using fluorescence spectroscopy. J Pharm Biomed Anal 2012, 71:8998. 30. Li B, Shanahan M, Calvet A, Leister KJ, Ryder AG: Comprehensive, quantitative bioprocess productivity monitoring using fluorescence EEM spectroscopy and chemometrics. Analyst 2014, 139:1661-1671. 31. Ohadi K, Legge RL, Budman HM: Development of a soft-sensor based on multi-wavelength fluorescence spectroscopy and a dynamic metabolic model for monitoring mammalian cell cultures. Biotechnol Bioeng 2015, 112:197-208. 32. Rowland-Jones RC, van den Berg F, Racher AJ, Martin EB, Jaques C: Comparison of spectroscopy technologies for improved monitoring of cell culture processes in miniature bioreactors. Biotechnol Progr 2017, 33:337-346. 33. Groza RC, Calvet A, Ryder AG: A fluorescence anisotropy method for measuring protein concentration in complex cell culture media. Anal Chim Acta 2014, 821:54-61. 34. Ryder AG, Stedmon CA, Harrit N, Bro R: Calibration, standardization, and quantitative analysis of multidimensional fluorescence (MDF) measurements on complex mixtures (IUPAC Technical Report). Pure Appl Chem 2017, 89:1849-1870. A recent report which covers all of the major issues involved in the use of multidimensional fluorescence spectroscopy measurements of complex samples such as CCM. Provides a series of recommendations that should facilitate the collection of more reproducible data and correct implementation of chemometric methods of analysis. 35. Li B, Ray BH, Leister KJ, Ryder AG: Performance monitoring of a mammalian cell based bioprocess using Raman spectroscopy. Anal Chim Acta 2013, 796:84-91. 36. Ranamukhaarachchi SA, Peiris RH, Moresoli C: Fluorescence spectroscopy and principal component analysis of soy protein hydrolysate fractions and the potential to assess their antioxidant capacity characteristics. Food Chem 2017, 217:469-475. 37. Buckley K, Ryder AG: Applications of Raman spectroscopy in biopharmaceutical manufacturing: a short review. Appl Spectrosc 2017, 71:1085-1116.
www.sciencedirect.com
Cell culture media analysis using rapid spectroscopic methods Ryder 17
Recent overview of the application of Raman spectroscopy to applications in the biopharma sector. Includes assessments of the major issues with a variety of different samples included solid and solution media. 38. Trunfio N, Lee H, Starkey J, Agarabi C, Liu J, Yoon S: Characterization of mammalian cell culture raw materials by combining spectroscopy and chemometrics. Biotechnol Prog 2017, 33:1127-1138. 39. Ryder AG: Surface enhanced Raman scattering for narcotic detection and applications to chemical biology. Curr Opin Chem Biol 2005, 9:489-493. 40. Wen ZQ, Li GY, Ren D: Detection of trace melamine in raw materials used for protein pharmaceutical manufacturing using surface-enhanced Raman spectroscopy (SERS) with gold nanoparticles. Appl Spectrosc 2011, 65:514-521. 41. Kirdar AO, Chen GX, Weidner J, Rathore AS: Application of nearInfrared (NIR) spectroscopy for screening of raw materials used in the cell culture medium for the production of a recombinant therapeutic protein. Biotechnol Progr 2010, 26:527-531. 42. Li G, Wen Z-q: Screening soy hydrolysates for the production of a recombinant therapeutic protein in commercial cell line by combined approach of near-infrared spectroscopy and chemometrics. Appl Microbiol Biotechnol 2013, 97:2653-2660. 43. Hofer A, Kamravamanesh D, Bona-Lovasz J, Limbeck A, Lendl B, Herwig C, Fricke J: Prediction of filamentous process
www.sciencedirect.com
performance attributes by CSL quality assessment using midinfrared spectroscopy and chemometrics. J Biotechnol 2018, 265:93-100. 44. Cao XL, Loussaert JA, Wen ZQ: Microspectroscopic investigation of the membrane clogging during the sterile filtration of the growth media for mammalian cell culture. J. Pharm. Biomed. Anal. 2016, 119:10-15. A nice example of the application of FT-IR spectroscopy to the analysis of filter membrane clogging. Experimental details are comprehensive and the integration with EDAX results for elemental screening is very useful. 45. Lee HW, Christie A, Xu J, Yoon S: Data fusion-based assessment of raw materials in mammalian cell culture. Biotechnol Bioeng 2012, 109:2819-2828. 46. ASTM E578-01: Standard Test Method for Linearity of Fluorescence Measuring Systems. 100 Barr Harbor Drive, West Conshohocken, PA: ASTM International; 2001. 47. Ameloot M, vandeVen M, Acuna AU, Valeur B: Fluorescence anisotropy measurements in solution: methods and reference materials (IUPAC technical report). Pure Appl Chem 2013, 85:589-608. 48. Jose GE, Folque F, Menezes JC, Werz S, Strauss U, Hakemeyer C: Predicting mAb product yields from cultivation media components, using near-infrared and 2D-fluorescence spectroscopies. Biotechnol Progr 2011, 27:1339-1346.
Current Opinion in Chemical Engineering 2018, 22:11–17