A Generic Growth Test Method for Improving Quality Control of Disposables in Industrial Cell Culture

ABSTRACT

Disposable bags are widely used in the biotechnology industry. The two main purposes are to store cell-culture media and to grow cells for inoculum or production. Several groups have reported growth inhibition resulting from the use of such products. This report shows independent data from four companies, using several different cell lines and growth media, and suggests a method that can be implemented for quality control at disposable-bag vendors.

Photo Credit: Nicholas Rigg/Getty ImagesRecently, a group of engineers and scientists within the cell-culture process-development community convened the first inaugural Cell-Culture Industry Forum (CCIF), which was held in Longmont, Colorado, in February 2011. More than 30 individuals from nine companies met over three days to present topics related to the common challenges and problems within the large-scale cell-culture industry. A highlighted theme at the first forum was the observation of cell growth inhibition, presumably resulting from the use of disposable bags. Disposables are widely used in the biotechnology industry for numerous bioprocessing functions. Extractables and leachables from disposables are frequently discussed (1). Two of the functions of disposable bags in the industry are to store cell-culture media and to grow cells within the bags at all steps of cell culture-based manufacturing. At the meeting, as well as described in other forums (2–6), it was clear that cell growth was sometimes affected by the use of disposables for these applications. A subteam was formed after the CCIF adjourned to further discuss and act upon these observations.

This paper describes collaboration in which four of the participating CCIF companies jointly assessed cell growth inhibition resulting from the use of disposable cell-culture bags (hereafter referred to as bags). The assessments are meant to uncover subtle effects on cell growth when medium is warmed in bags prior to use in cell-culture steps. Developing a widely applicable growth test is not a trivial exercise—it should be cell line- and media-independent so that vendors can apply it with confidence and avoid unwanted false-positive results. To that end, data are shown from these four companies, encompassing many different cell lines grown in four independent and proprietary media in which cell growth effects are apparent when water used for media preparation is warmed in the bags prior to use. The method described has leveraged strategies developed at each company as part of their ongoing work in disposable-bag selection. It is implementable across bag vendors and can potentially uncover problematic bag lots as well as identify when bag-film changes lead to inhibition of cell growth.

MATERIALS AND METHODS

Bags used for cell-growth testing

The data shown here are derived from three bag vendors. For all graphs, Vendors A, B, and C are the same vendors among each company's figures. Each company used the bag(s) in use at their facility at the time of testing. The bags are used for either media storage (and warming) or cell-culture steps (seed train, scale up, and production) in the overall bioprocess.

Preparation of media used in cell-growth testing

The water used to prepare media was from the same source as used to prepare cell-culture media for normal process development and/or pilot-plant operations.

Companies 1 and 2 employed the following method. Sterilized water was introduced into bags at the lowest volume-to-surface area ratio normally employed for cell bag (growth in bags) applications. If any company didn't normally use bags for growth steps (i.e., they only use bags for media storage), they added water at a low volume-to-surface area ratio, similar to that of a company who would use bags for such an application. Water was incubated in bags for four days at 37°C. As a control, water was incubated in glass bottles for four days at 37°C. After incubation, both water types were used to prepare cell-culture media at Companies 1 and 2 according to their normal operating procedures. Following preparation, the media were sterile-filtered (0.1 μm pore size rating filters of either hydrophilic polyvinylidene fluoride [PVDF] or polyethersulfone [PES] filter media) into and stored in glass or Nalgene-type bottles. Media were then held at 2–8°C until used for cell culture-growth testing, up to 30 days. Company 2 also held prepared media derived from water incubation in bags at 2–8°C in glass bottles for 16 weeks. This case is labeled "aged medium" for the data of Company 2. Company 3 used a similar method to Company 1 and 2, except that the water incubation temperature was 36.5°C, and the control medium was prepared using the same cell culture-grade water according to the standard preparation method at the company without four days' incubation at 36.5°C.

Company 4 modified the approach of Companies 1–3 slightly. Medium was sterile-filtered at a low media-to-bag area ratio into either a glass flask (control) or into a 10-L bag from three vendors. The containers were incubated at 37°C for three days prior to the media being used in the growth assessments.

The data shown here represent at least four independent media. Because of the proprietary nature of these media, it is unknown among the authors how similar or unique they are to one another.

The second step consisted of passaging cells using the incubated media. This was done by transferring incubated media into separate shake flasks, inoculating with cells, and passaging three times, each passage lasting 3–4 days. This procedure was used to test films from three different vendors, each run in at least duplicate cultures.

Cell lines

All companies used at least one in-house Chinese hamster ovary (CHO) cell line that had been transfected to produce a protein of interest. CHO lines used dihydrofolate reductase (DHFR [-]) or glutamine synthetase (GS) selection systems. Company 2 also performed the work with an NS0 cell line. The cell lines in some cases were already known to be affected by this bag application as discovered by those companies previously. In some cases, the cell lines were tested for growth inhibition for the first time. All cell lines used for these studies are believed to represent typical production lines within each company.

Cell growth and assessment

Prior to cell-growth testing, each cell line was grown in media identical to or similar to that of the actual test media composition. Cells were seeded into shake flasks for the first passage in test media following a low-g spin and resuspension step except for Company 3, where cultures were passaged without this step. The seeding densities used were in the low end of the range used at each company. Cultures were then maintained for at least 10 generations (doublings) over three passages for the control condition (water incubated in glass bottles), or for test conditions where there was cell growth. Each passage lasted 3 or 4 days. At each passage, the cells were diluted to the same seeding density as for the initial passage in test media (i.e., there were no fixed split ratios). Mixing and CO₂ gas levels followed each company's standard practice for the cell line being tested. Cell density and viability measurements were made using automated image analyzers that employ Trypan blue staining.

Cell growth levels and viabilities were compared between bag-incubated water or media, and glass bottle-incubated water or media preparations. Cell growth is shown as a multiple of the starting passage density and is labeled as normalized viable cell density. In all cases, actual viabilities are shown. Data points for cell density and viability in each figure are the average of at least two shake flask cultures.

RESULTS

Figure 1 shows the cell growth and viability for a CHO cell line previously observed at Company 1 to have growth inhibition when media was stored and warmed in disposable bags from Vendor A, as well as from another CHO line that did not exhibit growth inhibition. Cells grew to the highest densities when media were prepared from water incubated in glass bottles (Figure 1, Glass incubated water). One lot of bags from Vendor A resulted in limited growth inhibition for each CHO line (Lot 1) while another lot (Lot 2) affected growth significantly during passages one and two, where viability was also lower than other cases (Figure 1, Vendor A incubated water, CHO Line 2, Bag Lot 2).

Figure 1: Normalized cell growth and viability of Chinese hamster ovary (CHO) cell lines-results from Company 1. (ALL FIGURES COURTESY OF AUTHORS)

Figure 2 shows the results from Company 2 for an NS0 line and a CHO line. The NS0 cell line experienced a severe impact on growth and viability when cultured in medium derived from one lot of bag-incubated water (Figure 2, Vendor B incubated water, NS0 Line, Bag Lot 1). This same medium, when stored at 2–8°C for an additional duration (16 weeks in total) in a glass bottle, still maintained its growth inhibitive properties on this cell line (Figure 2, Vendor B incubated water, NS0 Line, Bag Lot 1, Aged Medium). However, this bag did not exhibit any growth inhibitive effects on the CHO cell line evaluated (Figure 2, Vendor B incubated water, CHO Line, Bag Lot 1). A second lot of bag-incubated water was able to support cell growth in the sensitive NS0 cell line (Figure 2, Vendor B incubated water, NS0 Line, Bag Lot 2).

Figure 2: Normalized cell growth and viability of NS0 and Chinese hamster ovary (CHO) cell lines-results from Company 2.

Company 3 assessed cell growth and viability on one CHO cell line (Figure 3) using two vendors and two lots of bags from each vendor. Growth was inhibited by nearly 50% when water incubated in Vendor A bags was used to make medium (Figure 3, Vendor A incubated water). Cell viabilities were not affected, however. For the cell line tested, there were no inhibition effects on cell growth and viability observed when using medium prepared from water incubated in Vendor B bags for both lots of bags tested.

Figure 3: Normalized cell growth and viability of Chinese hamster ovary (CHO) cell lines-results from Company 3.

Company 4 assessed cell growth and viability in media incubated in bags from three vendors (Figure 4). Vendor C bags led to the most pronounced growth and viability effects. In the case of media held in Vendor C bags, this CHO line did not grow at all (Figure 4, Vendor C incubated water). Vendor A bags, shown to inhibit CHO cell growth in a cell line from Company 1 (Figure 1) and Company 3 (Figure 3), led to 60–80% growth inhibition while viability was less affected. Vendor B bags did not lead to any pronounced growth or viability effects.

Figure 4: Normalized cell growth and viability of Chinese hamster ovary (CHO) cell lines-results from Company 4.

DISCUSSION

Disposable bags used in cell-culture applications have complex manufacturing processes. Polymeric films in many cases are sourced from secondary vendors where a similarly complex manufacturing environment may exist. Bag assemblers may not have detailed knowledge of film manufacturing methods or the authority to influence them if they wished to. Once bags are assembled, they are typically gamma irradiated, which along with the length of shelf-life post-irradiation, is another potential source of variability between manufacturing lots as well as between vendors' products with respect to this effect on cell growth (2).

The length of media or water warming used for the testing described here is slightly longer than would likely be used in practice. However, the method is useful in uncovering growth inhibition effects and reveals an underlying issue with these products. While industry may not be able to supply vendors with a suitable CHO or NS0 production line to apply this method, the general nature of the effect across multiple CHO lines suggests that a nonproprietary CHO line may be suitable for implementation. However, end-users would likely need some degree of bridging data prior to accepting results generated by a bag vendor using such a CHO cell line. The existence of such bridging data could, in a general sense, also reduce the total in-house effort required at a given end-user to accept bag vendor cell-culture test results in the future.

To date, the mechanisms that result in cell growth impacts by disposable bags are varied and not completely understood. Medium component adsorption has been identified as one mechanism that can inhibit cell growth in a cholesterol-dependant cell line (7). Cell growth impacts resulting from a specific leachable/extractable have been characterized (8, 9). The cell-based assay was able to identify which disposable bags impacted cell growth and eventually led to identification of the specific leachable inhibiting cell growth. The vendor was then able to optimize the film composition to ensure biocompatibility.

It cannot be assumed that these are the only mechanisms that can result in inhibition of cell growth. The use of a cell-based assay like those described here can be used to identify where such impacts exist. Once they have been determined, various factors such as raw materials, leachables and extractables, irradiation, storage, and other factors can be systematically investigated. This investigational approach underpinned by a cell-based assay can lead to the identification of the root cause of cell-growth impacts that can be readily addressed by the vendors once understood. The growth test method described is useful in troubleshooting and can be applied proactively as a harmonized first step in improving quality control of these widely used disposable products for both suppliers and customers.

The problem end-users face is not vendor-specific or fundamentally dependent on the end-users' cell lines or media. The growth test methods described can be applied at bag vendors' own sites. In doing so, vendors may improve the overall understanding and quality control of their disposable products. While some products may not lead to the same observed phenomena, as polymeric film manufacturing or gamma irradiation practices change around them, each vendor could use a test such as this to uncover if those changes have an effect on this particular application. As this issue is still prevalent years after it was first discussed, the authors and the cell culture-engineering community are eager to see the quality-control testing and practices for disposables manufacturers improve.

CONCLUSION

When cell-culture medium or the water used to make it are incubated at operating temperature (37°C) in disposable bags prior to use, growth inhibition effects can be observed. These effects are not cell-line, cell-type or cell culture-media specific. The problem is relevant for multiple bag manufacturers. The authors suggest that bag vendors implement the cell-growth test described here as part of their manufacturing quality control and as a means to understand critical film composition attributes and manufacturing controls. The question of what cell lines are best used for this testing remains, but it is believed to be solvable as the effect is observed in multiple CHO lines as well as an NS0 cell line. Having data from such a cell-growth test together with a deeper understanding and control of film composition and manufacture will facilitate adoption and continued use of disposable-bag technology. This knowledge will help ensure robust and predictable performance even through the inevitable changes in bag and bag film manufacturing processes.

ACKNOWLEDGEMENTS

The authors wish to thank Robert Kiss, Masaru Shiratori, Thomas Ryll, Reb Russell, and Richard Schicho for their discussion, review, and comments on this manuscript. Erica Graf and Richard Martel performed cell culture work at Bristol-Myers Squibb Company.

*Brian Horvath is a scientist at Process Technical Development, Late Stage Cell Culture, Genentech, a member of the Roche Group, South San Francisco, CA 94110; Valerie Liu Tsang is senior engineer III and Weimin Lin is a scientist, both at Cell Culture Development, Biogen Idec, RTP, NC 27709; Xiao-Ping Dai is manager at Cell Culture Science, Global Manufacturing and Supply, Bristol-Myers Squibb Company, Bloomsbury, NJ 08804; Kurt Kunas is a principal scientist and Greg Frank is a principal engineer, both at Amgen, Thousand Oaks, CA 91320.

* To whom all correspondence should be addressed, bhorvath@gene.com.

REFERENCES

1. A. Mire-Shis et al, BioProcess Int. 9 (2) 14-23 (2011).

2. P. Gammell, "The impact of Lot-to-lot Variability of Disposable Cell Culture Bags on Cell Growth During the Scale-Up of a Mammalian Production Cell Line: Root Cause Analysis and Lessons Learned from the Pipeline," presentation at the Cell Culture Engineering XIII Conference (Scottsdale, AZ, April 2012).

3. M. Shiratori, "Recent Experiences with Single-Use Containers Used for Media Storage, Media Handling and Cell Culture," presentation at the IBC 8th International Single-Use Applications for Biopharmaceutical Manufacturing Conference (Long Beach, CA, June 2011).

4. M. Shiratori, "Recent Experiences with Single-Use Containers Used for Media Storage, Media Handling and Cell Culture," presentation at the BioProcess International Conference & Exhibition Conference (Long Beach, CA, November 2011).

5. J. Wood, "Take II—More Findings and Updates on Our Experiences with Disposable Bags for Cell-culture media Storage," presentation at the IBC's 9th International Single-Use Applications for Biopharmaceutical Manufacturing Conference (San Francisco, CA, June 2012).

6. W. Lin and V.L. Tsang, "Risk Mitigation of Mammalian Cell Growth Issues when Using Disposable Cell Culture Bags," presentation at the 245th American Chemical Society National Meeting and Exposition (New Orleans, LA, April 2013).

7. J. Kadarusman et al, Biotechnol. Prog. 21 (4) 1341-1346 (2005).

8. D. Low, "Material Understanding and Supplier Collaboration for Single-Use Systems," presentation at the Bioprocess International Conference and Exhibitions (Providence, RI, October 2012).

9. M. Goodwin and S. Kline, "New Approaches to Managing Single-Use Systems," presentation at the 3rd Annual Single Use Bioprocess Technologies Symposium (Research Triangle Park, NC, February 2013).