Interaction of leachable model compounds and their impact on Chinese hamster ovary cell cultivation

The presence of leachables in biopharmaceutical processes using single‐use technologies (SUT) is well known. For the detection and quantification of the latter, extractable studies of SUT are very common nowadays. Although a mixture of compounds is regularly found in extractable studies, research has only been carried out regarding the effect of individual compounds on cell culture and the cumulative effect of a mix of leachables has not been investigated yet. In this study, a set of leachable model compounds (LMCs) was chosen and the effect of the LMCs on a Chinese hamster ovary DG44 cell line producing an IgG antibody was investigated concerning cell growth, cell cycle distribution and productivity. It was shown that even if worst‐case concentrations were used, the LMCs solely impact cell growth. Additionally, interaction studies revealed that the inhibiting effect of the mix is lower than the expected cumulative effect. A strong antagonism between the antioxidant butylated hydroxytoluene and the plasticizer Tris(2‐ethylhexyl)trimellitate was found using an isobologram analysis.


| INTRODUCTION
The fast growing market of biopharmaceuticals has increased the pressure to develop cost-effective and flexible production processes.
Single-use (SU) technologies meet these demands and have become well established in the biopharmaceutical industry. However, concerns have arisen that potentially toxic substances migrate out of SU material into the process fluid and perturb the tightly regulated manufacturing process. To assess a possible impact of these compounds on the process, risk analysis are carried out and the migrated substances are identified and quantified. 1,2 Furthermore, in 2019 a tool was developed that simplifies the decision between the use of disposables versus conventional multi-use equipment. 3 One of the most prominent leachables with growth inhibiting effects on cell cultures is bis(2,4-di-tert-butylphenyl)phosphate (bDtBPP), a degradation product from the antioxidant Irgafos ® 168, found in gammairradiated polyethylene bioreactor bags. 4 Since then, many efforts have been made to develop test systems for the early identification of potential critical leachables from SU material. Consequently, biocompatibility testing recommendations for SU material have been published, 5,6 which enabled effective optimization of polymeric film formulations without any negative impact on cell growth 7,8 and productivity. 9 Apart from potential leachable induced negative effects on upstream processes, a possible harmful impact on the drug product has to be examined to ensure patient safety. A recent study by Hauk et al 10 demonstrates that SU components used in downstream processes can be sinks of leachables. Moreover, Paudel et al 11 describe that leachables can be removed from the process fluid upon contact with Chinese hamster ovary (CHO) cells, too, indicating an interaction mechanism between the latter and the leachables. Even if the impact of well-known leachables on cell cultures is described, 12,13 little is known about the effect of leachables on the cells' constitution and productivity. The leachable 3,5-Dinitro-bisphenol A, found in extracts of polycarbonate flasks, is known to cause a cell cycle arrest in CHO-S cells. 14 Since the cell cycle distribution of a culture can impact the productivity, 15 cell cycle arresting substances can have an impact on the production performance. Interestingly, Kelly et al 16 describe that even if the cell cycle distribution of CHO cells is not altered upon treatment with the leachable bDtBPP, a reduced IgG productivity was observed. However, based on the data presented in the publication, it can be assumed that the reduction in productivity may be attributed to a decreased viable cell density (VCD) and the cell-specific productivity appears to be not impaired.
A bioreactor assembly usually consists of different polymeric materials. Thus, it can be anticipated that a mixture of different leachables can be found in the culture broth. For this reason, it is of utmost importance to study the combined effect of a set of leachables on production processes. In this study, we have examined the impact of a set of leachables on an IgG production process that potentially migrate out of different SU materials. The leachables used in this study are called leachable model compounds (LMCs), each of them representing an own class of organic additive (e.g., antioxidants, plasticizers). First, literature research was carried out to define concentrations of the LMCs from extractable analysis, representing a realistic worst-case approach with exaggerated extraction conditions. The toxicity of the LMC-mixture (LMC-mix) was then assessed in order to estimate whether these concentrations are suited for a long-term study. Thereafter, a fed-batch cultivation was performed in shake flasks and the effect of the LMC-mix on cell growth, viability and productivity was examined. Furthermore, the characterization of apoptosis and cell cycle distribution was carried out. Additionally, toxicity of each individual compound was assessed in batch experiments. To gain a deeper understanding of how leachables interfere with each other, possibly resulting in synergistic or antagonistic effects, toxicity studies with different leachable combinations were performed.

| Selection of LMCs
For this study, a subset of leachables was chosen as model compounds that represent typical substances commonly detected in extractable analysis of polymeric material and are commercially available. These selected LMCs (Table 1) are different in terms of chemical characteristics (i.e., molecule size and lipophilicity expressed as the log of the partition in an octanol/water system, log K O/W ). The concentrations of the LMCs were found under exaggerated extraction conditions and thus, represent a worst-case scenario in this study. bDtBPP is built during gamma-irradiation of polyethylene-based films of SU bioprocessing materials. 4   Shaking speed was adjusted to 120 rpm with an orbital diameter of 50 mm in 80% humidified atmosphere. Cultures were seeded at 0.2 × 10 6 cells/ml unless stated otherwise.  After spiking, cells were cultivated for 3 days. Samples for analysis were taken every day. The specific growth rate μ was calculated according to formula (2) for each condition in the exponential growth phase and normalized to the reference.

| IgG quantification
where N (cells/ml) is the VCD and t is the cultivation time, indices indicating different time points.

| Interaction study
For studying interaction effects of the leachables on cytotoxicity, an

| Fed-batch cultivation
In order to observe the effect of the LMC-mix on CHO cell cultivation over a longer cultivation period, an 8-day fed-batch cultivation was car-

| Isobologram analysis
For estimation of interaction effects between the LMCs based on graphical analysis, an isobologram evaluation was created. First, doseresponse experiments were carried out and different EC i -values were calculated after 3 days of incubation. The EC i value indicates at which concentration the cell growth is inhibited to ith extent (%). Then, an isobologram was created as described by Berenbaum. 31 This graph allows to study the combined effects of substance mixtures and characterize them as additive, synergistic or antagonistic, as is frequently used in the field of pharmacology. 32 If the graph is a straight line, there is additivity between the two substances. If the curve is concave, the effect is called synergistic and a convex curve is called antagonistic. 31 However, in order to classify the interactions in a more T A B L E 2 Fluorescence marker used for apoptosis assay, the related cellular signals detected and affected phases of apoptosis However, exposure time of cells to leachables in a bioreactor assembly is generally longer in industrial cultivation processes, for example, for the production of monoclonal antibodies. 34 (Figure 3b). Additionally, the binding affinity of the IgG does not change during the cultivation run and there is no effect of the LMC-mix on the binding affinity ( Figure S1b).

| Interactions of LMCs
Cell growth in the presence of a toxic single leachable can have a different impact on cell growth than in the presence of a mix of leachables. Combinations of leachables, as used in this study, may exhibit synergistic or antagonistic effects on cell growth caused by interactions as described for multiple drug treatment. 35 To elucidate whether there is any interaction between the leachables, CHO DG44 cultures were spiked with each leachable individually and an interaction study was conducted with concentrations provided in Table 1. As illustrated in Figure 4, 0.1 mg/L of the leachable bDtBPP caused a reduced cell growth of only 65% ± 12%, while the viability was not affected. This observation is in line with a study by Kelly et al, 16 who report that the EC 50 -value of bDtBPP has an impact on cell growth without apoptosis induction. It is assumed that bDtBPP causes oxidative stress. At higher concentrations this leads to apoptosis and consequently, lower viabilities. Until then, however, the cell can cope with the oxidative stress by overexpression of specific proteins, such as Hypoxia upregulated protein 1 (HYOU1). 16 Figure S2) and in combination. Error bars represent SD of n = 3 replicates spike with respect to the initial viability.  Figure S2). To study combination effects, 10 mg/L BHT and varying concentrations of TOTM were spiked to CHO DG44 cultures. Figure 5a shows the normalized VCD after 3 days of cultivation. Using these data, an isobologram (Figure 5b) was created as described in chapter 3.

| CONCLUSION
Former studies on leachables found in SU systems have focused on detrimental effects of these substances in biopharmaceutical processes. These studies, however, have only focused on one leachable individually, neglecting the fact that in a SU system a mix of different leachables migrates into the process fluid. Therefore, our study characterizes the effect of a mixture of selected LMCs, which are actually found in extractable analysis, on a CHO production process. Interestingly, even if we found growth inhibiting effects of the LMC mix in both the dose-response experiments and a fed-batch run, the cellspecific productivity was not impaired. These findings are contrary to the data presented by Kelly et al, who describe reduced productivity after bDtBPP treatment of CHO cultures. 16 Since worst-case approaches were applied for defining the LMC concentrations used in this study, it can be anticipated that under normal process conditions the impact of leachables on CHO cell cultures is even lower. Especially during a cultivation carried out in perfusion mode, the impact of leachables can be expected to be negligible due to lower leachable concentrations, as the cultivation medium is exchanged several times during a process. If any, we expect a possible impact of leachables at the beginning of a process, where the leachable concentration per cell is relatively high. Thus, the impact will likely decrease over time with growing cell numbers. Since no effect on cell cycle, apoptosis or cell viability was observed, long-term effects of LMCs on CHO cell cultivation are not anticipated. Surprisingly, we found that the inhibiting effect of LMC mix on a CHO cell culture can be lower than their expected cumulative effect and a strong antagonism is described for the combination of BHT and TOTM.
Although we have shown that one LMC can weaken the toxicity of another one, it must be stressed that this effect was only observed with one cell line and possible effects of cultivation medium were not considered. For a realistic assessment of the migration behavior of process-equipment related leachables, the solubility of the compound and the surface-to-volume ratio of the related SU material in the process must be taken into consideration. Therefore, the exposure estimation based on extractables and leachables data is of high importance for a thorough risk-analysis of leachables in a biopharmaceutical process.