Skip to main content
Environmental Sciences Europe
Download PDF

Introducing oxygen transfer rate measurements as a novel method for time-resolved cytotoxicity assessment in shake flasks

Environmental Sciences Europe volume 34, Article number: 97 (2022) Cite this article

Abstract

Background

Determining the cytotoxicity of test substances is essential for the safety assessment of chemicals. To quantify the cytotoxicity, dose–response curves are determined and the half-maximum inhibitory concentration (IC50) is subsequently calculated. Results often rely on a single endpoint evaluation which typically requires manual sampling and subsequent sample analysis to determine the IC50. Hence, no information on culture behavior are available during treatment. Here, measurement of the oxygen transfer rate (OTR) was applied as a method for time-resolved assessment of the cytotoxicity to (a) provide information on culture behavior during treatment and (b) quantitatively assess the cytotoxic effect of a test substance.

Results

To investigate the applicability of the method, different types and concentrations of test substances were added and the OTR was measured for two different CHO suspension cell lines in shake flasks: (a) For CHO cell line one, culture behavior was not affected when 0.1 mM ZnCl2 was added. However, for both cell lines, adding up to 0.2 mM CoCl2 led to a decrease in the OTR that was concentration dependent. For CHO cell line two, adding up to 5% DMSO led to a concentration-dependent decrease in the OTR and in the slope of the OTR; (b) From the measured OTR, dose–response curves were established and used to calculate the IC50. For cell line one, the IC50 for CoCl2 was 0.03 mM while the IC50 from a comparative measurement based on trypan blue exclusion was 0.06 mM. Cell line two was less susceptible to CoCl2 as the IC50 was not reached at the concentrations tested. For DMSO, single OTR values and determined slopes were used for evaluation. In both cases, the IC50 was calculated to 2.3%.

Conclusions

Our approach provides a method for analyzing the cytotoxicity of a test substance based on OTR measurements. This method provides (i) insights on culture behavior, (ii) information on the progression of cytotoxicity, (iii) dose–response curves, and (iv) a first indication of IC50 values. Quantitative assessment of cytotoxicity is possible non-invasively and in real-time during treatment. Compared to traditional endpoint evaluation, our method simplifies experimentations and enhances the comparison of assay results.

Background

A cytotoxicity assay aims to quantitatively describe the effect of a test substance on a cell culture. Hence, cytotoxicity tests play an important role for the safety assessment of chemicals. Different cell lines, including mammalian CHO cell lines, may be used in cytotoxicity assays. To quantitatively evaluate the cytotoxicity, dose–response curves are determined. The test substance concentration is plotted (either in linear or logarithmic form) against the measured cytotoxicity. The relationship between dose and effect varies for different substances, resulting in various shapes for dose–response curves [49, 55]. If the test substance concentration is plotted in logarithmic form, a sigmoid curve is usually obtained [40]. However, biphasic relationships between dose and response have been reported as well. In hormesis, a stimulation at low dose and an inhibition at high dose results in a U- or J-shaped dose–response curve [8, 12]. A characteristic value for evaluating and comparing the cytotoxicity of different test substances is the IC50 value. IC50 is the concentration at which the half-maximal inhibition is observed, i.e., the test substance concentration at which the cytotoxicity reaches a value of 50%. This value is obtained from the fitting of the dose–response curve. For curves with a sigmoidal shape, 4-parameter logistic regression (4PL) is usually applied [56].

Despite a recent trend toward a more quantitative risk assessment, most of the currently established assays, including those recommended by the OECD guidelines for the testing of chemicals, rely on endpoint determinations [25]. Established assays can be grouped into different categories. In dye exclusion tests, cell viability is evaluated directly after treatment with the test substance by counting cells that took up the dye (and become colored) and cells that excluded the dye (and remain uncolored) [3]. Examples are the trypan blue exclusion assay or the erythrosine B assay. Among the colorimetric assays, the MTT (3-(4,5-dimethylthiazol-2-yl)-2–5-diphenyltetrazolium bromide) assay is most commonly used. Other assays may be based on fluorometric or luminometric evaluation [3]. It needs to be noted that differences were found in the calculated cytotoxicity based on the method used to determine it [18].

Problems of assays relying on endpoint determination

The utilization of endpoint measurement is disadvantageous because cytotoxicity is usually dose and time dependent, and the progression of cytotoxicity varies among test substances [17]. Moreover, the determination of cytotoxicity usually requires several steps after exposure of the cells to the test substance (e.g., staining and cell counting or additional incubation steps). Thus, analysis is often time consuming and requires additional materials. Further, cytotoxicity might be underestimated, depending on the method used. For trypan blue exclusion, it was demonstrated that cytotoxicity is vastly underestimated if performed directly after treatment. The reason is that only cells that are dead at the time that the assay is performed are evaluated [18]. Cells that die after the assay has been conducted, however, are not considered. Moreover, no time-resolved information is provided, as processing and analysis often take considerable time, and only the endpoint is used for evaluation. In particular, no quantitative information of culture behavior is available during treatment with the test substance. However, the treatment time with the test substance and the sampling time for analysis were crucial steps during the determination of cytotoxicity [32, 67]. Differences in these parameters were most likely responsible for different genotoxicity results between CHO and Chinese Hamster Lung (CHL) cultivations [58]. Time-resolved monitoring also offers the advantage to determine changes in cellular response (e.g., changes in growth rate, progression of viability). This way, differences in the mode of action of different test substances and different cell lines might be detected. In addition, the consistency of the assay is monitored, as abnormalities in culture behavior (e.g., insufficient growth of the control culture) become visible in real time.

Real-time measurement of cytotoxicity

Despite the advantages that real-time monitoring of culture progression during treatment offers, real-time monitoring and evaluation of cytotoxicity are yet to be used routinely. Real-time assessment of cytotoxicity was demonstrated by impedance measurement either in 96-well plates [10] and in the commercialized xCELLigence device (ACEA Biosciences/Agilent) [7, 15, 26, 42, 45, 59] with the last publication providing a very comprehensive overview of applications and comparison to endpoint-based assessment In addition, a real-time assay for the determination of cytotoxicity based on the fluorescent luciferase assay has been reported [17]. Furthermore, the Seahorse technology, offered by Agilent, enables measurement of mitochondrial toxicity in microtiter plates based on measurement of the oxygen consumption rate [54].

Online monitoring of the oxygen uptake rate (OUR)

The oxygen uptake rate (OUR) provides a wide variety of information during mammalian cell culture cultivation [23, 28, 29, 41]. As oxygen consumption is directly linked to cell metabolism, the OUR provides information on culture viability. At a constant cell-specific oxygen uptake rate (qO2), the OUR is directly proportional to the viable cell concentration (VCC). Additionally, the OUR is very close to the oxygen transfer rate (OTR), if the change in the oxygen concentration in the liquid over time is negligibly small. In slow mammalian cell cultures this is always the case, and the OUR may be determined from measurement of the OTR. Measurement of the OTR of CHO suspension cells has been realized in shake flasks using a prototype device of the Respiration Activity Monitoring System (RAMOS) [28, 29] as well as a commercially available Transfer-Rate Online Measurement (TOM) device [29]. Both systems utilize electrochemical sensors to determine the OTR. For mammalian cells, the overall breathing activities are low and the culture dynamics are slow compared to the measurement times [28]. Additionally, the change of the oxygen concentration in the liquid phase is much smaller than OTR and OUR, it can be assumed that the OUR equals the OTR [28, 44]. For measurement using RAMOS and TOM, a measurement and a gas flow phase were alternately applied. During the gas flow phase, the shake flask is aerated with gas. During the measurement phase, the inlet and outlet valves are closed. Thus, cell respiration results in a linear decrease in the headspace's oxygen partial pressure. The slope of the partial pressure decrease over time is used to calculate the OTR [1, 2]. Further details on the measurement principle are given in Additional file 1: Additional data (Sect. 1.1). Measurement is non-invasive with no manual intervention as the electrochemical sensor is placed outside of the sterile area of the flask [1, 2]. Additionally, measurements are usually carried out in up to eight shake flasks in parallel. However, parallelization is only limited by the number of flasks that can be operated in the incubator simultaneously. Even at the comparably low breathing activity of CHO cells, up to one OTR value per hour can be obtained [28, 29]. For CHO suspension cells, information on the VCC and differences in culture behavior at different initial cell concentrations were determined using RAMOS [29].

The RAMOS technology has already been successfully applied to bacterial systems to determine toxicity [34, 43]. Furthermore, in a miniaturized format in microtiter plates, it successfully has been applied to determine mutagenicity via an improved version of the bacterial Ames fluctuation test [30, 31].

This study aims to evaluate, if time-resolved and non-invasive measurement of the OTR generally allows to evaluate the cytotoxicity of test substances on CHO suspension cells during treatment in real time. Furthermore, it is investigated how the gathered OTR data may be utilized to quantitatively draw conclusions about the cytotoxicity of a test substance. By measuring the OTR, time-resolved information on culture behavior during treatment with the test substance is obtained. For this initial assessment and to determine a possible workflow for interpreting the OTR data, experiments are conducted in shake flasks. To investigate, if the method may be applied in a more general way, different test substances, CHO cell lines and cultivation conditions, including culture media and initial seeding densities, are investigated (Table 1). It should be noted that it is not within the scope of this study to provide a statistically reliable analysis of the cytotoxicity of a certain test substance, but rather to investigate if monitoring of the OTR may in general be suited to gather such data. The advanced and required statistical assessment is left to those more experienced in the field.

Table 1 Overview of the combination of CHO cell lines, cultivation conditions, and test substances investigated within this study
Full size table

Methods

Cell lines and culture media

Two different CHO suspension cell lines were used for cultivation. Cell line one is an industrial CHO suspension cell line producing an IgG1 antibody developed by Rentschler Biopharma SE (see [28, 51]). Cell line two is a CHO suspension cell line obtained from Cell Lines Service (CLS) GmbH, Germany. It is not modified to produce an antibody. Two different cell lines were chosen, to demonstrate that the obtained results are of general nature and not due to a unique characteristic of a single cell line.

In general, the CHO cell lines used in this study were expected to behave differently, because the variability in genotypes and phenotypes is quite substantial among CHO cells [37, 52, 65, 66]. In addition, as mentioned above, CHO cell line one was genetically modified to express a gene of interest. Native and mAb producing CHO cell lines were reported to differ in growth and substrate consumption [39]. This difference might also influence the sensitivity of the cells toward certain test substances.

When it comes to growth behavior, utilization of suspension cultures has been recommended in the past [33]. Advantages of using suspension cell lines in comparison to adherent CHO cells include shorter cultivation times and no need for trypsinization for detachment of cells [38, 46]. Both cell lines were cultured in a serum-free, chemically defined medium. Different culture media were used for cultivation to explore, if the method is robust and, thus, may be applied to different cell lines cultured in different media. Media were stored at 4 °C until use. Stock solutions for media supplements were stored at − 20 °C in aliquots and thawed as needed. Before use, media and supplements were pre-heated to the cultivation temperature for about 30 min in a water bath (VWB2 12, VWR, USA). Cell line one was cultured in PowerCHO™ 2 serum-free chemically defined culture medium, containing HEPES buffer and Pluronic® F68 (Lonza AG, Switzerland). Here, it should be noted that the cells may uptake Pluronic® [22], which might influence cytotoxicity results. However, as the primary focus of the study was not to provide improved culture media for cytotoxicity assessment, no change to the culture medium was made. This way a comparison to previously published studies [28, 29] is still possible. Directly before use, the medium was supplemented with 6 mM l-glutamine (Gibco Life Sciences, Thermo Fisher Scientific, USA) and 1% (v/v) PenStrep (stock with 10,000 Units mL−1 penicillin and 10 g L−1 streptomycin) (Sigma-Aldrich, USA). Cell line two was cultured in sciNX medium (CLS GmbH, Germany) and supplemented with 5 mM l-glutamine and 1% (v/v) PenStrep.

Cryopreservation, thawing, and cell passaging

Cryopreservation was performed as described in detail by Ihling et al. [28] (cell line one) and (cell line two) [29] and according to the basic principles described, for example, by Freshney [20]. After thawing, one washing step was included to remove DMSO.

All passages were carried out in 250 mL Corning® polycarbonate flasks closed with a vent cap (Sigma-Aldrich, USA) at a temperature of 36.5 ± 0.25 °C. A filling volume of 20 mL with a shaking frequency of 140 rpm and a shaking diameter of 50 mm was used for cultivation. An orbital shaking incubator with CO2 and humidity control (ISF1-X, Kuhner AG, Switzerland) set to 5% and 70%, respectively, was used for passaging. Cells were passaged every 3–4 days. The seed cell concentration for subsequent cultures was set to 0.2∙106 mL−1 (cell line one) or 0.25∙106 mL−1 (cell line two), respectively.

Main culture cultivation conditions

After an appropriate number of passages (see Additional file 1: Table S1), cytotoxicity experiments were started. Cells were cultured at a temperature of 36.5 °C (cell line one) or 37 °C (cell line two). A shaking diameter of 50 mm, a filling volume of 20 mL (cell line one) or 50 mL (cell line two), and a shaking frequency of 140 rpm were used in all cases. Cultivation was carried out in 250 mL glass flasks. Flasks were modified in the upper part to allow active aeration but remained unmodified in the lower part that was in contact with the liquid (see Additional file 1: Additional Data Section 1.1 for details). Active aeration of flasks in the TOM device was carried out using incubator air (70% relative humidity and 5% CO2) (ISF1-X, Kuhner AG, Switzerland). Cells cultured in the RAMOS device were incubated in a temperature-controlled incubator (ISF1-X, Kuhner AG, Switzerland) and aerated using gas (5% CO2 in synthetic air) supplied by a gas bottle. The initial viable cell concentration (VCC) was set to 0.2∙106 mL−1 (cell line one) and 0.5∙106 mL−1 (cell line two), respectively. The initial VCCs were different, because cultivation protocols were previously independently established for each of the cell lines in the respective culture media.

Trypan blue exclusion method

To determine the viable cell concentration (VCC), a sample was taken from each RAMOS flask after 161 h (experiment 1; Additional file 1: Table S1) and 164 h (experiment 2; Additional file 1: Table S1), respectively. If required, the sample was appropriately diluted with pre-warmed culture medium. The sample was then mixed with trypan blue (1:1) and 10 µL of the mixture was added to a hemocytometer (Counting chamber C-Chip Neubauer improved, Carl Roth, Germany). The number of viable (unstained) cells was manually counted in four quadrants, averaged, and corrected by the dilution factor given by the manufacturer to determine the viable cell concentration.

Addition of test substances

The addition of test substances was either performed directly at the beginning of the experiment or after 24 h (see Additional file 1: Table S1). A 20 mM stock solution of CoCl2∙6H2O (purity ≥ 99%, Carl Roth, Germany) and a 10 mM stock solution of ZnCl2 (Sigma-Aldrich, USA) were each prepared in the culture medium used for CHO cell line 1(experiments 1 + 2). For experiments 3–5, performed with CHO cell line 2, a stock solution of 50 mM CoCl2∙6H2O was prepared using cell culture grade water (HyClone HyPure Cell culture grade water, GE Healthcare, Germany). In experiments 6 + 7, dimethyl sulfoxide (DMSO, ≥ 99.7%, Merck KgA Germany) was added without dilution to reach the final concentration in the culture medium. DMSO concentrations are given as (v/v).

Logistic fit of the dose–response curve

Obtained dose–response curves were fitted using OriginPro® 2020 (OriginLab Corporation, USA). A 4-parameter logistic fit function (Eq. 1) was chosen to determine the IC50. In the equation, A1 is the minimum cytotoxicity if the concentration of the test substance is 0 and A2 is the maximum cytotoxicity reached. The value x0 in the equation represents the IC50. P represents the slope at the steepest part of the curve.

$$\mathrm{Cytotoxicity }\left[\mathrm{\%}\right]= \frac{{A}_{1}-{A}_{2}}{1+{(\frac{x}{{x}_{0}})}^{p}}+{A}_{2.}$$
(1)

Monitoring of the oxygen transfer rate

The oxygen transfer rate (OTR) was monitored in shake flasks using electrochemical sensors. Two different devices, an in-house built Respiration Activity Monitoring System (RAMOS) and a commercially available Transfer-Rate Online Monitoring (TOM) system (Kuhner AG, Birsfelden, Switzerland), were used to monitor the OTR. Single measurement values that deviated more than 30% from the next measurement value were considered outliers and excluded from further evaluation. Further details may be taken from Additional file 1: Additional Data Section 1.1.

Results and discussion

Qualitative determination of cytotoxicity with and without medium exchange

In a first step, monitoring of the OTR was carried out for CHO suspension cell line one in culture medium one (Fig. 1A). Two well-known test substances, ZnCl2 and CoCl2, were used to evaluate culture behavior during treatment. Reduced growth of CHO cells upon exposure to metal salts was previously demonstrated to correlate with strand breaks in the DNA [53]. Exposure with 0.1 mM ZnCl2 for 4 h did not cause DNA strand breaks in CHO cells, and consequently, growth was not affected [53]. On the other hand, an IC50 of 86 µM was determined for CHO cells after 16 h of incubation in the presence of ZnCl2 [60]. Following this literature information, a ZnCl2 concentration of 0.1 mM was investigated (Fig. 1A, yellow triangles).

Fig. 1
figure 1

Oxygen transfer rate (OTR) of CHO cell line one upon addition of ZnCl2 or CoCl2 with and without medium exchange. A Cells were cultured without addition of test substance (closed black triangles), with addition of 0.1 mM ZnCl2 (closed yellow squares) or with addition of 0.2 mM CoCl2 (blue circles). Test substances were added at the beginning of the cultivation (t = 0 h). B Cells were cultured without addition of test substance (closed black triangles), with addition of 0.1 mM ZnCl2 (closed orange squares) or with addition of 0.2 mM CoCl2 (open cyan circles). After 24 h (black arrow), medium was exchanged by fresh culture medium without a test substance added. Data from experiment 1 (see Additional file 1: Table S1)

Full size image

A strong influence of CoCl2 on cell growth was described for adherently grown CHO cells with the IC50 determined to be about 140 µM [60]. CoCl2 has also been shown to induce apoptosis by excessive generation of reactive oxygen species (ROS) in PC12 cells [68]. Exposure to ROS and increasing levels of apoptosis can lead to a decrease in cell growth and viable cell number [35]. As a result, a CoCl2 concentration of 0.2 mM (Fig. 1A, blue circles) was chosen in our study.

A control culture without the addition of test substances (Fig. 1A, closed black triangles) was run for comparison. Culture behavior, represented by the course of the OTR over the cultivation time, was not affected when 0.1 mM ZnCl2 was added to the culture medium (Fig. 1A, closed yellow squares). When 0.2 mM CoCl2 was added (Fig. 1A, blue circles), the OTR did not increase but stayed relatively constant at a value of about 0.15 mmol L−1 h−1 until the end of the experiment. Here, it can be concluded that ZnCl2 in the chosen concentration did not harm cell growth and is not cytotoxic under the conditions applied in this study. CoCl2, on the other hand, completely suppressed cell growth at the investigated concentration and for the cell line and cultivation conditions investigated. The influence of CoCl2 on culture behavior will be discussed in detail in Section “Quantitative determination of CoCl2 cytotoxicity.”

This first CHO cell line investigated in our study clearly exhibited a lesser sensitivity for ZnCl2 than the CHO cell line used in the study by Tan et al. [60], as cell growth was not affected in the presence of 0.1 mM ZnCl2 in our study, but an IC50 of 86 µM was reported by Tan et al. [60]. In general, ZnCl2, more particularly the availability of Zn2+, has to be considered with caution as various elements influencing the availability, and consequently also the cytotoxicity of Zn2+ have been proposed [47]. This includes the presence of fetal calf serum (FCS) in the culture medium, as will be discussed further below. Also, the protein metallothionein may play an important role in intracellular zinc metabolism. Its metal-unsaturated form may act as a chelator for intracellular Zn2+ [50], which might help reduce the cytotoxicity of Zn2+.

A more general reason for the observed differences in ZnCl2 sensitivity might be that adherent CHO-K1 cells, which were more sensitive in a cell-mediated cytotoxicity assay [16], were formerly used by Tan et al. [60] instead of the currently tested suspension cells. Growth adaptation from adherent to suspension may in general affect cellular responses [57]. A fundamental difference between adherent and suspension cultures is that suspension cells may interact with a test substance on their complete surface, while the surface of adherently grown cells is not completely accessible. On the other hand, no difference in assay outcome has been observed between cells growing in suspension cultures or adherently [61]. Another reason for the different sensitivity toward ZnCl2 might be the different culture media used. Even though the presence of FCS might reduce the cytotoxicity compared to media with less or no FCS [5, 27], this effect was not observed for all test substances [5]. Further, FCS provides essential Zn2+ in a culture medium [47]. The addition of FCS may, thus, result in an unintended and undefined amount of Zn2+ being additionally added to the culture medium. Further, and as stated above, the use of Pluronic® F68 might have also affected culture behavior.

To evaluate if the cytotoxic effect was persistent, even when the test substance was removed from the medium, and to simulate a change in cultivation conditions during the performance of a cytotoxicity test, cultivation with an exchange of the medium was carried out (Fig. 1B). Cells were cultured in the presence of ZnCl2 (Fig. 1B, closed orange squares) or CoCl2 (Fig. 1B, open cyan circles) for 24 h. After 24 h (Fig. 1B, arrow), the medium was replaced with freshly prepared culture medium where neither ZnCl2 nor CoCl2 was added. After medium exchange, both cultures grew quite similar compared to the control culture (Fig. 1B, closed black triangles) that was performed in the original culture medium from the beginning on. For ZnCl2, no cytotoxic effect was observed, indeed. For CoCl2, cells could again proliferate once CoCl2 was removed from the medium. Consequently, growth seems to have been suppressed in the presence of CoCl2 but resumed after CoCl2 was removed from the medium.

As demonstrated by this experiment, monitoring of the OTR provided insights on culture behavior during treatment and enabled a direct qualitative assessment of the cytotoxic potential of a test substance. Monitoring of the OTR is also suited to evaluate if cells can recover after test substance exposure and subsequent cultivation in the absence of the test substance. It may further be applied also if changes to the culture conditions are made during the assay.

Quantitative determination of CoCl2 cytotoxicity

As cell growth was strongly affected during treatment with 0.2 mM CoCl2, different concentrations of CoCl2 were tested in a subsequent experiment (Fig. 2A). Besides the IC50 of 140 µM reported in the literature for adherently grown CHO cells [60], addition of 0.3 mM CoCl2 was found to quickly result in cell death [48]. Hence, the CoCl2 concentration was varied between 0.05 and 0.3 mM.

Fig. 2
figure 2

Oxygen transfer rate (OTR) and cytotoxicity for different CoCl2 concentrations using CHO cell line one. The OTR over time is depicted in duplicate for CHO cells with addition of 0 mM CoCl2 (= control) (black triangles), 0.05 mM CoCl2 (orange squares), 0.1 mM CoCl2 (green stars), or 0.3 mM CoCl2 (pink diamonds). CoCl2 was added at the beginning of the cultivation (t = 0 h). Data were smoothed by a moving average of 3 points. Original OTR data are shown in Additional file 1: Fig. S1. B Cytotoxicity over time calculated from OTR for different CoCl2 concentrations. Cytotoxicity was calculated by dividing the OTR of the treated culture by the OTR of the control culture at the respective point in time. Data for 0.05 mM CoCl2 (orange squares), 0.1 mM CoCl2 (green stars), and 0.3 mM CoCl2 (pink diamonds) and corresponding control from experiment 2 (Additional file 1: Table S1). OTR data are depicted in A. Data for 0.2 mM CoCl2 (blue circles) and corresponding control from experiment 1 (Additional file 1: Table S1). OTR data are depicted in Fig. 1A

Full size image

As the data density over time was comparably high, but the overall signal was rather noisy (Additional file 1: Fig. S1), data were smoothed by a moving average of 3 data points.

Culture behavior was reproducible within one experiment, as shown by the duplicate measurements (Fig. 2A). Initial cell growth was comparable for all cultures except when 0.3 mM CoCl2 was added to the medium (Fig. 2A, pink diamonds). However, in all cultures treated with CoCl2, the OTR decreased earlier compared to the control culture. This early decrease indicates an earlier reduction of the VCC and consequently a decreasing viability of the cultures. The results indicate that the cell line might become more sensitive toward CoCl2, the longer the incubation lasts. As mentioned above, this agrees with data from the literature [35]. In addition, the achieved maximum OTR decreased with increasing CoCl2 concentrations, increasingly affecting cell growth as well. Concentrations of CoCl2 of 0.05 mM (Fig. 2A, orange squares) already impacted culture behavior visibly. For an IgG producing CHO suspension cell line, inoculated with an initial viable cell density of 3 ∙ 106 cells mL−1 and cultured for 72 h, addition of 0.05 mM CoCl2 did not affect cell growth, but 0.3 mM CoCl2 resulted in complete cell death [48]. In our study, cell growth was slightly affected for 0.05 mM CoCl2 (Fig. 2A, orange squares), whereas the OTR of the culture treated with 0.3 mM CoCl2 (Fig. 2A, pink diamonds) had largely decreased after 72 h. As differences in the behavior of even similar CHO cell lines are expected (see “Methods” section), the general trend for CoCl2 compares quite well with data from the literature.

For a quantitative assessment of the cytotoxic potential, it was investigated, if the cytotoxicity may be calculated from a comparison of the OTR values between the treated culture and the control culture. For this, Eq. 2 was used. The obtained absolute value of the OTR at a given CoCl2 concentration and at a given point in time (OTRtreated,t=x) was divided by the averaged OTR of the two control cultures (Fig. 1A, black triangles) at this point (OTRcontrol,t = x) (Eq. 2):

$${\mathrm{Cytotoxicity}}_{t = x}\left[\mathrm{\%}\right]=\left(1-\frac{{\mathrm{OTR}}_{\mathrm{treated}, t = {\text{x}}}}{{\mathrm{OTR}}_{\mathrm{control}, t = x}}\right)\cdot 100\mathrm{ \%}.$$
(2)

The presented approach enabled a time-resolved calculation of the cytotoxicity in real time and already during treatment instead of a single endpoint determination. The time-resolved cytotoxicity for different CoCl2 concentrations calculated from the OTR (Fig. 2A) is depicted in Fig. 2B. The calculated cytotoxicity for 0.2 mM CoCl2 (Fig. 2B, blue circles) from experiment 1 (see Additional file 1: Table S1) was also included. As the OTR was initially relatively low, a clear trend for cytotoxicity was only observed after about 40 h. Consequently, the calculated cytotoxicity is depicted from this time onwards.

The cytotoxic effect of CoCl2 increased with increasing cultivation time (Fig. 2B) which agrees with literature results reported for CHO suspension cells [48]. In addition, the time required to reach the maximum cytotoxicity decreased with increasing CoCl2 concentration which points at a dose-dependent effect that also agrees with the literature (see above). For 0.3 mM CoCl2 (Fig. 2B, pink diamonds), the maximum cytotoxicity was approached after around 72 h, while for 0.1 mM CoCl2 (Fig. 2B, green stars), maximum values were reached after around 120 h.

Traditionally, a dose–response curve is typically determined at the end of treatment or even after treatment has ended. As discussed above, dose–response curves are used to determine the IC50 as a characterizing parameter for a test substance. Here, we investigated, if cytotoxicity assessment based on OTR measurements could even be taken a step further by determining not only the course of the cytotoxicity from the OTR data but also a dose–response curve and the IC50 value. Thus, the maximum cytotoxicity of CoCl2 was determined by averaging the cytotoxicity values between 140 and 160 h of cultivation. Here, the cytotoxicity was relatively constant for all concentrations tested (Fig. 2B) and had reached its maximum. As CoCl2 was added from the beginning (t = 0 h), this timespan also corresponds to the treatment time. The determined cytotoxicity was plotted against the added CoCl2 concentration in a semi-logarithmic form to obtain the corresponding dose–response curve (Fig. 3A).

Fig. 3
figure 3

Dose–response curves for CoCl2 at different points in time using CHO cell line one. A Dose–response curve for CoCl2from cultivation data between 140 and 160 h. The following workflow was applied: OTR values were measured time resolved for different CoCl2 concentrations (cf. Fig. 2A). From each measured OTR value, cytotoxicity was calculated using Eq. 2 (cf. Fig. 2B). Calculated cytotoxicity values were averaged in the time interval between 140 and 160 h (cf. Fig. 2B). Obtained cytotoxicity values were plotted against the respective CoCl2 concentration to obtain the depicted dose–response curve. Logistic curve fitting according to Eq. 1 was performed to determine the IC50. IC50 = 30 µM. B Dose–response curve for CoCl2 after 58 h of cultivation. The same workflow as in A was applied, but single OTR values measured after 58 h were used. IC50 = 111 µM. Data from experiments 1 and 2 (Additional file 1: Table S1)

Full size image

All concentrations tested led to comparably high cytotoxicity above 60% (Fig. 3A). Consequently, values needed to be extrapolated for the calculation of IC50. The values for A1 (minimum cytotoxicity) and A2 (maximum cytotoxicity) (Eq. 1) were set to 0% and 100%, respectively. Logistic fitting (Eq. 1) resulted in an IC50 of 30 µM. As mentioned above, Tan et al. [60] reported an IC50 of 140 µM for adherent CHO-K1 cells in a serum-containing medium. The higher cytotoxicity determined in our study likely resulted from the significantly longer incubation time after which the cytotoxicity was determined (Table 2). As shown in Fig. 2B, the cytotoxicity increased with time, decreasing the IC50 value. Thus, it appears very reasonable that differences in IC50 were observed at different incubation times. In addition, CHO cell lines are known to be genetically and physiologically diverse, which makes comparison of data between different cell lines difficult [64]. In addition, transcriptomic changes were observed for CHO cells upon adaption from adherent to suspension growth [57].

Table 2 Comparison of cultivation conditions for CHO cell line one from this study and the study by Tan et al. [60]
Full size table

Time-resolved determination of the cytotoxicity enabled the calculation of dose–response curves with high resolution of one dose–response curve per hour. The values at each point in time in Fig. 2B can be used to obtain a dose–response curve at this point leading to coverage of the whole range of cytotoxicity. Exemplarily and for better comparison with the data from Tan et al. [60], a dose–response curve was determined from the OTR after 58 h (Fig. 2B) and used to calculate the IC50 (Fig. 3B). In this case, an IC50 of 111 µM is obtained. This value is closer to the value reported by Tan et al. [60] (140 µM) and appears reasonable, as treatment times are more comparable in this case (Table 2). In addition, lower IC50 levels in real-time assays compared to conventional endpoint determination were also determined for several anticancer drugs [26]. Differences in CoCl2 sensitivity might have also been caused by different culture media utilized (Table 2). Here, partly the same reasons as discussed for ZnCl2 may have affected culture behavior. Particularly, using media with or without FCS as well as differences between adherently grown and suspension grown cells may have affected the culture behavior.

Again, it may be important to note that it was not the aim of this study to provide a statistically reliable analysis of the IC50 values, but rather to investigate if it is in general possible to obtain a dose–response curve using OTR data. Nevertheless, we compared the dose–response curve for CoCl2 that was obtained from OTR measurement (Fig. 3A, Additional file 1: Fig. S2A) to a measurement using trypan blue exclusion (Additional file 1: Fig. S2B). The IC50 determined from trypan blue exclusion measurement is around 60 µM. Even though this is about twofold higher than determined from OTR measurement, the values are in the same order of magnitude. Furthermore, the higher IC50 determined from trypan blue exclusion agrees well with the observation that trypan blue exclusion tends to underestimate the cytotoxicity of a test substance [18]. An underestimation of the cytotoxicity corresponds to a higher IC50. In addition, the IC50 from trypan blue exclusion was determined manually, from a single endpoint measurement and not averaged from multiple measurement values. To investigate, if the cell concentration at the time, when CoCl2 is added, affects the IC50, CoCl2 was added after 24 h in a subsequent experiment (Additional file 1: Fig. S3A). Time-resolved cytotoxicity was calculated from the OTR data (Additional file 1: Fig. S3B). The concentrations tested included CoCl2 concentrations as low as 0.01 mM, as the data obtained (Fig. 3 and Additional file 1: Fig. S2) indicated an IC50 for CoCl2 in the range of 30–60 µM. Cytotoxicity was again similar between 140 and 160 h of cultivation (Additional file 1: Fig. S3B). Thus, a dose–response curve was determined for the 20-h range between 140 and 160 h by dividing the averaged OTR values between treated and control cultures (Eq. 2) (Additional file 1: Fig. S3C). The IC50 was determined from 4-parameter logistic regression (Eq. 1) to 36 µM (Additional file 1: Fig. S3C). No extrapolation of the data was necessary. The results indicate that the cell concentration at the start of the treatment did not impact the maximum cytotoxicity and that IC50 values between consecutive experiments were comparable.

In summary, the IC50 concentration of CoCl2 determined from non-invasive, and time-resolved OTR measurement was in the same order of magnitude compared to conventional endpoint determination by trypan blue exclusion and to data previously reported in the literature. Considering different incubation times, cell lines, initial viable cell concentrations, and media used, as well as differences caused by the methods used to evaluate the cytotoxicity, the IC50s calculated from the OTR seem very reasonable. However, for a statistical reliable analysis of the IC50 values, further tests need to be performed, including a larger number of replicates and further comparison with established cytotoxicity assays. However, in contrast to single endpoint determination, these analyses may be based on a larger data basis.

In comparison to traditional endpoint determinations, where cells may only be cultured for 24–48 h, here evaluation was carried out after up to 160 h. Nevertheless, evaluation may also be performed after about 50 h (see Fig. 3B). For an even earlier evaluation, the noise of the OTR data would need to be reduced. This may, for example, be achieved by increasing the measurement time (see Additional file 1: Additional Data for details), which will result in a decreased temporal resolution. As currently one data point per hour is obtained, this may be acceptable. Further, the IC50 can be determined right after the respective OTR values have been measured with no time for sample analysis required.

Transfer of methodology to a different cell line

The effect of CoCl2 was investigated for a second CHO cell line cultured in another medium to further investigate the applicability of using OTR data to determine the cytotoxicity of a test substance. Furthermore, the negative influence of CoCl2 was exclusively determined for an IgG producing cell line [48] so that comparison with a non-producing cell line appeared interesting. General culture behavior in the utilized medium was described previously [29]. The overall course of the OTR of the control culture (Fig. 4, black squares) was different in shape and OTR values compared to cell line one (Fig. 1, black triangles). The differences can be attributed to different media with different glucose concentrations being used for cultivation. Consequently, the y-axis is seized differently compared to the previous experiments. For CHO cell line two, concentrations up to 0.2 mM CoCl2 were tested (Fig. 4).

Fig. 4
figure 4

Influence of CoCl2 on the oxygen transfer rate (OTR) of CHO cell line two. A The OTR over time is depicted in presence of 0 mM CoCl2 (= control) (black squares), 0.05 mM CoCl2 (closed orange stars), 0.075 mM CoCl2 (half-filled blue circles), 0.1 mM CoCl2 (closed red diamonds), 0.125 mM CoCl2 (open blue stars), 0.15 mM CoCl2 (closed purple triangles), or 0.2 mM CoCl2 (open yellow circles). CoCl2 was added at the beginning of the cultivation (t = 0 h). Data from experiment 4 (Additional file 1: Table S1)

Full size image

The effect of CoCl2 on the growth of the second CHO suspension cell line was much less detrimental compared to the first cell line. Even in the presence of 0.2 mM CoCl2, which entirely suppressed the growth of cell line one (Fig. 1B, blue circles), cells were initially able to grow (Fig. 4, open circles). With increasing cultivation time and CoCl2 concentration, lower overall OTR values compared to the control were reached. Maximum deviations between the cultures were reached after around 100 h (Fig. 4) when the OTR of the control had reached its maximum. After the OTR maximum, the VCC decreases and cell behavior is expected to change, as glucose is depleted from the culture medium and lactate is consumed [29]. Consequently, the maximum cytotoxicity was calculated by dividing the OTR values for a CoCl2 concentration of 0.2 mM after 100 h by the OTR of the control culture at the same time (Eq. 2). The calculation resulted in a cytotoxicity of about 40%. The culture behavior of the control and for two CoCl2 concentrations was tested in a subsequent experiment (Additional file 1: Table S1). It was found to be well comparable to the first experiment (Additional file 1: Fig. S4), even though a commercial TOM system was used for experiment 4 (Fig. 4) and an in-house built RAMOS system was used for experiment 5 (Additional file 1: Fig. S3). However, no differences in the OTR data were expected when using two different devices.

Differences in observed IC50 values for CoCl2 between the two different CHO suspension cell lines were expected for the reasons detailed in the “Methods” section and might be attributed to differences in the cell lines, culture media, or seeding densities. For cancer drug screening experiments, the cell line itself was shown to have a more considerable, significant impact on variations in viability compared to medium or seeding density, which had no significant effect [36]. Again, it was not the aim of this study to compare the response of different cell lines to the same test substance, but rather to explore if despite significant differences seen in the course of the OTR for different cell lines, the cytotoxicity may still be determined from the OTR using the presented, generalized workflow.

Influence of DMSO on CHO suspension cell line two

To further investigate the applicability of the method, DMSO was used as a third test substance. DMSO is usually used for the cryopreservation of cells. DMSO is also used to dissolve test substances that are not readily soluble in water, even though replacement of DMSO is recommended [21]. In general, the cytotoxic concentration of DMSO strongly varies among cell types [45]. A DMSO concentration of 10% did not alter membrane permeability in Caco2/TC7 cells but has been used as a positive control for in vitro biocompatibility tests using L929 mouse fibroblasts [4, 13]. Additionally, a DMSO concentration range of 0.1–1.5% (v/v) was previously stated as a common concentration range when used as a vehicle for cellular treatments [62]. Thus, DMSO concentrations between 0.1 and 5% (v/v) were investigated here.

Figure 5 depicts the course of the OTR of CHO cell line two over time for different DMSO concentrations. The control culture reached a maximum OTR of about 0.6 mmol L−1 h−1 after around 100 h of cultivation (Fig. 5, black squares).

Fig. 5
figure 5

Influence of DMSO on the oxygen transfer rate (OTR) of CHO cell line two. The OTR over time is depicted in the presence of 0% DMSO (= control) (black squares), 1% DMSO (open blue diamonds), 1.75% DMSO (closed gray circles), 2.5% DMSO (open green stars), 3.25% DMSO (closed orange triangles), or 5% DMSO (open purple circles). DMSO was added at the beginning of the cultivation (t = 0 h). Data from experiment 6 (Additional file 1: Table S1)

Full size image

Except for 1% DMSO (Fig. 5, open blue diamonds), the cultures treated with different concentrations of DMSO showed a slower increase of the OTR and reached lower maximum OTR values than the control. In contrast to CoCl2, where initial cell growth was comparable (Fig. 4), culture behavior already deviated from the beginning. With increasing DMSO concentrations, the slope of the OTR over time was lower, and consequently, the OTR ascent was delayed. The observation of an inhibited cell proliferation caused by DMSO was also observed in human cancer cells [45]. Here, the proliferation curve, measured in real time by evaluating the change in impedance, also showed different slopes after DMSO addition [45]. The results underline that attention needs to be paid if cultures are exposed to higher DMSO concentrations under culture conditions. This is consistent with the typical recommendation to quickly remove DMSO after cell thawing.

The point where the maximum OTR of the control was reached (after 100 h) was used to calculate the cytotoxicity of DMSO using Eq. 2 (Fig. 6A). Cytotoxicity was calculated by the ratio of the OTR value of the treated culture and the control culture. The resulting dose–response curve (Fig. 6A) was fitted by 4-parameter logistic fitting (Eq. 1), which resulted in an IC50 of 2.3% (Table 3).

Fig. 6
figure 6

Dose–response curves of DMSO for CHO cell line two. A Dose–response curve for DMSO obtained by curve fitting using Eq. 1 with cytotoxicity values calculated with Eq. 2 by using measured OTR values of treated and control culture after 100 h. B Dose–response curve for DMSO obtained by curve fitting using Eq. 1 with cytotoxicity values calculated with Eq. 2 using ratio of determined slopes of treated and control culture. Each slope was calculated by linear fitting of the measured OTR values between 0 and 100 h (Additional file 1: Fig. S6). Data from experiment 6 (Additional file 1: Table S1). OTR data used for calculation are depicted in Fig. 5

Full size image
Table 3 Parameters obtained by logistic fitting for evaluation of cytotoxicity of DMSO on CHO cell line two (Fig. 6)
Full size table

As shown in Fig. 5, the increase of the OTR was linear. This linear increase indicated that the cytotoxic effect was not strongly affected by the exposure time. This is also reflected when plotting the time-resolved cytotoxicity calculated from the ratio of the OTR between treated and control culture at each point (Additional file 1: Fig. S5). Consequently, the advantage of the high data density and time-resolved measurement was exploited by calculating the cytotoxicity from the slope of the OTR over time instead of using single OTR values. Accordingly, the slope of the OTR until 100 h was determined by linear fitting (Additional file 1: Fig. S6). For cultures with 5% DMSO addition (Fig. 5, open purple circles), the slope was negative (Additional file 1: Fig. S6, open purple circles). For 1% DMSO, the slope was higher compared to the control culture (Additional file 1: Fig. S5, open blue diamonds).

In the next step, the ratio of the slopes obtained from linear fitting of the OTR was used to calculate the cytotoxicity. This was performed to investigate, if the OTR data could also be exploited in another way. For this purpose, the respective slope values (Additional file 1: Fig. S6) were inserted into Eq. 2. The resulting dose–response curve is depicted in Fig. 6B. For 5% DMSO, a cytotoxicity above 100% was calculated, which could be attributed to the evaluation method applied. A negative slope is calculated, as viable cells are initially present, but the viable cell concentration is decreasing. This negative slope results in a calculated cytotoxicity of more than 100%. For 1% DMSO (Fig. 5, open blue diamonds), a negative cytotoxicity was calculated, because the slope for the culture treated with 1% DMSO was higher than the slope of the control culture. A negative response value might point toward a hormetic dose–response relation characterized by a stimulating effect of the test substance at low concentrations and toxicity at higher concentrations [9]. This stimulating effect is well reflected by the higher slope.

Plateau values (A1, A2) varied, if single OTR values or slopes were used for calculation, because of the method used for evaluation (Table 3). Nevertheless, an IC50 of ~ 2.3% was determined for DMSO by logistic fitting independent of the evaluation method (Table 3). However, calculation using the ratio of the OTR gave a more reasonable value for the maximum cytotoxicity, expressed as A2, reached (Table 3). In addition, utilization of single OTR values enables time-resolved calculation of the cytotoxicity. The time-resolved calculation is not possible if the slopes are used for evaluation. Nevertheless, a comparison of the slopes might also be of interest, if for example, different lag phases between treated and control cultures are observed. Further, using growth rates instead of cell counts in an endpoint determination has been proposed some years ago [24]. Again, the progression of the OTR over time for different DMSO concentrations was comparable between consecutive experiments (Additional file 1: Fig. S7) and when two different devices (TOM system for experiment 6, in-house built RAMOS system for experiment 7) were used (see Additional file 1: Table S1).

The determined IC50 of ~ 2.3% is in good agreement with the literature. The IC50 was previously determined for adherent CHO cells in an application note using fluorescence evaluation [11]. A value of 1.8% was reported. Lower IC50 values for the adherent cell line might appear reasonable. Another reason for the good agreement for different CHO cell lines might be that DMSO addition largely increases the osmolality of the culture medium [63]. Consequently, the cytotoxic effect might not result from the compound itself but the effect it has on the culture medium. Thus, it seems reasonable that different CHO cell lines react comparably to an increase in media osmolality and, thus, DMSO concentrations. In addition, our experiments indicate that DMSO addition did not have a large effect on the calculated cytotoxicity between 24 and 96 h of treatment. Consequently, different treatment times may affect the IC50 less and therefore result in better comparability and similar IC50 values between different assay protocols. For a statistical sound analysis, more experiments need to be conducted.

Comparison of cytotoxicity assessment by OTR measurement with other assay methods

In this study, the first assessment of using OTR values to determine cytotoxicity for CHO suspension cells has been conducted in shake flasks. Compared to conventional MTP-based assays, which may only require a volume in the µL range, culture volumes in shake flasks are several orders of magnitude higher. Even though the number of shake flasks to be monitored in parallel may easily be numbered up, the media consumption and footprint are much larger compared to microtiter plate-based assays. Thus, it is well acknowledged, that a transfer to microtiter plate scale is very desirable to increase experimental throughput and to reduce the overall costs of the experiments. For measurement in MTPs, the µRAMOS, which enables monitoring of the OTR in 48- or 96-well microtiter plates [14, 19], may be used. However, it has to the authors knowledge neither been applied for monitoring of mammalian cells in general nor CHO cells specifically, yet.

Assuming that OTR measurement of CHO cells in microtiter plates may be achieved in the future, using measurement of the OTR for cytotoxicity assessment provides an important addition to already established measurement systems. As discussed above, measurement based on impedance (see Additional file 1: Additional Data Sect. 1.2 for a more detailed description of the technology) has been demonstrated (see, for example, [7, 10, 15, 26, 42, 45, 59]). Real-time cell analysis (RTCA) based on impedance measurement is preferably performed for adherent cells [59], while OTR-based evaluation is demonstrated for suspension cells in this study. Additionally, it was hypothesized that additives of drugs, including lactate, could interfere with impedance measurements [7]. As lactate is a common byproduct of CHO cell metabolism, impedance measurement might in general be less suited to be used with CHO cells. Here, measurement of the OTR could provide an alternative measurement option.

Another example, also mentioned above, is the Seahorse technology. It is based on evaluating mitochondrial respiration (see Additional file 1: Additional Data Sect. 1.3 for details) and is intended to be used with adherent cells [6]. This technology may be used for cytotoxicity assessment, but care needs to be taken when evaluating the data [54]. A reason for this is that data normalization to cell numbers may mask a cytotoxic effect [54]. Additionally, the technology is designed to evaluate the effect of a test substance within min to h [54]. Hence, it is not intended for monitoring culture behavior during several days of cultivation. However, as discussed above, the exposure time to a test substance is a critical factor in cytotoxicity assessment and will affect the results. Here, monitoring the OTR, applicable for long cultivation periods lasting several days could again provide a complementary measurement option.

Conclusions

This study investigated the suitability of online monitoring of the oxygen transfer rate (OTR) to assess the cytotoxic effect of different test substances on Chinese hamster ovary (CHO) cells in shake flasks. Two different CHO suspension cell lines cultured in different media were used. Qualitative evaluation of the cytotoxicity was directly possible from the OTR signal. The addition of 0.1 mM ZnCl2 did not affect growth of CHO cell line one. On the other hand, addition of 0.2 mM CoCl2 led to a decrease in the OTR. If test substances were removed from the media after 24 h of incubation, cell cultures subsequently recovered, and growth resumed with a course comparable to the control culture.

Quantitative evaluation of the cytotoxicity was possible by dividing the OTR of the treated and the control culture at each point. The cytotoxicity of CoCl2 increased with exposure time and maximum cytotoxicity was reached earlier with increasing CoCl2 concentrations. This time- and dose-dependent effect is in accordance with literature data.

Dose–response curves obtained from OTR resulted in a sigmoidal correlation between test substance concentration and cytotoxicity that is also observed in conventional assays. For CHO cell line one, 4-parameter logistic fitting of the dose–response curve resulted in an IC50 of 30 µM toward the end of treatment (140–160 h of cultivation). Evaluation of the dose–response curve after 58 h of treatment resulted in an IC50 of 111 µM. For cell line two, no IC50 could be determined, because the calculated cytotoxicity was only at 40% at a CoCl2 concentration of 200 µM after 100 h. The differences observed may be attributed to cell line one producing an IgG1, and/or different culture media used, but further experiments are required to confirm this.

For CHO cell line two, DMSO was investigated in addition to CoCl2. For DMSO, the slope of the OTR decreased with increasing DMSO concentrations at concentrations above 1%. Consequently, evaluation of the cytotoxicity was carried out based on (i) the ratio of the OTR values at each point and (ii) the ratio of the slopes obtained from linear fitting of the OTR. In both cases, an IC50 of 2.3% DMSO was determined. For a DMSO concentration of 1% the growth rate of the culture was higher than the control, pointing at a stimulation at low dose.

In conclusion, a generalized workflow to use measured OTR data to determine the IC50 of a test substance was demonstrated. Future experiments should focus on using the presented method for a statistically sound analysis of individual test substances. Additionally, transfer to the microtiter plate scale is highly desirable. For the demonstrated method, no sampling, no subsequent offline analysis, and no interruption of the cultivation process were required to calculate the IC50 and provide insights on culture behavior. In addition, assessment was possible time resolved with a resolution of one measurement point per hour.

By the presented method, a better comparison of assays performed at different conditions, especially at different treatment times or cell concentrations, is possible. In addition, the time required to reach maximum cytotoxicity can be determined and used to standardize different test protocols. In the future, the approach presented might be implemented into standardized guidelines to assess chemical cytotoxicity in vitro.

Availability of data and materials

The dataset used and/or analyzed during the current study is available from the corresponding author upon reasonable request.

References

  1. Anderlei T, Büchs J (2001) Device for sterile online measurement of the oxygen transfer rate in shaking flasks. Biochem Eng J 7(2):157–162. https://doi.org/10.1016/s1369-703x(00)00116-9

    Article  CAS  Google Scholar 

  2. Anderlei T, Zang W, Papaspyrou M, Büchs J (2004) Online respiration activity measurement (OTR, CTR, RQ) in shake flasks. Biochem Eng J 17:187–194. https://doi.org/10.1016/S1369-703X(03)00181-5

    Article  CAS  Google Scholar 

  3. Aslantürk ÖS (2017) In vitro cytotoxicity and cell viability assays: principles, advantages, and disadvantages. In: Larramendy ML, Soloneski S (eds) Genotoxicity—a predictable risk to our actual world. IntechOpen, London

    Google Scholar 

  4. Bae SH, Che JH, Seo JM, Jeong J, Kim ET, Lee SW et al (2012) In vitro biocompatibility of various polymer-based microelectrode arrays for retinal prosthesis. Invest Ophthalmol Vis Sci 53(6):2653–2657. https://doi.org/10.1167/iovs.11-9341

    Article  CAS  Google Scholar 

  5. Bohets HH, Nouwen EJ, De Broe ME, Dierickx PJ (1994) Effects of foetal calf serum on cell viability, cytotoxicity and detoxification in the two kidney-derived cell lines LLC-PK1 and MDCK. Toxicol In Vitro 8(4):559–561. https://doi.org/10.1016/0887-2333(94)90016-7

    Article  CAS  Google Scholar 

  6. Braissant O, Astasov-Frauenhoffer M, Waltimo T, Bonkat G (2020) A review of methods to determine viability, vitality, and metabolic rates in microbiology. Front Microbiol. https://doi.org/10.3389/fmicb.2020.547458

    Article  Google Scholar 

  7. Cai L, Qin XJ, Xu ZH, Song YY, Jiang HJ, Wu Y, Ruan HJ, Chen J (2019) Comparison of cytotoxicity evaluation of anticancer drugs between real-time cell analysis and CCK-8 method. ACS Omega 4(7):12036–12042. https://doi.org/10.1021/acsomega.9b01142

    Article  CAS  Google Scholar 

  8. Calabrese EJ, Baldwin LA (2001) U-shaped dose-responses in biology, toxicology, and public health. Annu Rev Public Health 22:15–33. https://doi.org/10.1146/annurev.publhealth.22.1.15

    Article  CAS  Google Scholar 

  9. Calabrese EJ, Baldwin LA (2003) The hormetic dose-response model is more common than the threshold model in toxicology. Toxicol Sci 71(2):246–250. https://doi.org/10.1093/toxsci/71.2.246

    Article  CAS  Google Scholar 

  10. Ceriotti L, Ponti J, Broggi F, Kob A, Drechsler S, Thedinga E et al (2007) Real-time assessment of cytotoxicity by impedance measurement on a 96-well plate. Sensors and Actuators B-Chemical 123(2):769–778. https://doi.org/10.1016/j.snb.2006.10.024

    Article  CAS  Google Scholar 

  11. Christmann T, Lüke J, Hummel S, Pirsch M (2017) Versatile cytotoxicity assay for animal cells with the NYONE(R). https://synentec.com/media/livedead_an-f340-xx-08_.pdf. Accessed 26 Apr 2021

  12. Conolly RB, Lutz WK (2004) Nonmonotonic dose-response relationships: mechanistic basis, kinetic modeling, and implications for risk assessment. Toxicol Sci 77(1):151–157. https://doi.org/10.1093/toxsci/kfh007

    Article  CAS  Google Scholar 

  13. Da Violante G, Zerrouk N, Richard I, Provot G, Chaumeil JC, Arnaud P (2002) Evaluation of the cytotoxicity effect of dimethyl sulfoxide (DMSO) on Caco(2)/TC7 colon tumor cell cultures. Biol Pharm Bull 25(12):1600–1603. https://doi.org/10.1248/bpb.25.1600

    Article  Google Scholar 

  14. Dinger R, Lattermann C, Flitsch D, Fischer JP, Kosfeld U, Büchs J (2022) Device for respiration activity measurement enables the determination of oxygen transfer rates of microbial cultures in shaken 96-deepwell microtiter plates. Biotechnol Bioeng 119(3):881–894. https://doi.org/10.1002/bit.28022

    Article  CAS  Google Scholar 

  15. Dokumaci AH, Fayemi PO, Yerer MB (2019) Real time monitoring of cytotoxicity of Callistemon citrinus against Colo-205 cell Line. Istanbul J Pharm 49(1):25–32. https://doi.org/10.26650/IstanbulJPharm.2019.418892

    Article  Google Scholar 

  16. Dorfman NA, Civin CI, Wunderlich JR (1980) Susceptibility of adherent versus suspension target cells derived from adherent tissue culture lines to cell-mediated cytotoxicity in rapid 51Cr-release assays. J Immunol Methods 32(2):127–139. https://doi.org/10.1016/0022-1759(80)90066-6

    Article  CAS  Google Scholar 

  17. Duellman SJ, Zhou WH, Meisenheimer P, Vidugiris G, Cali JJ, Gautam P, Wennerberg K, Vidugiriene J (2015) Bioluminescent, nonlytic, real-time cell viability assay and use in inhibitor screening. Assay Drug Dev Technol 13(8):456–465. https://doi.org/10.1089/adt.2015.669

    Article  CAS  Google Scholar 

  18. Fellows MD, O’Donovan MR (2007) Cytotoxicity in cultured mammalian cells is a function of the method used to estimate it. Mutagenesis 22(4):275–280. https://doi.org/10.1093/mutage/gem013

    Article  CAS  Google Scholar 

  19. Flitsch D, Krabbe S, Ladner T, Beckers M, Schilling J, Mahr S, Conrath U, Schomburg WK, Büchs J (2016) Respiration activity monitoring system for any individual well of a 48-well microtiter plate. J Biol Eng. https://doi.org/10.1186/s13036-016-0034-3

    Article  Google Scholar 

  20. Freshney RI (2005) Cryopreservation. In: Freshney RI (ed) Culture of animal cells. Wiley-Liss Inc., New York. https://doi.org/10.1002/0471747599.cac020

    Chapter  Google Scholar 

  21. Galvao J, Davis B, Tilley M, Normando E, Duchen MR, Cordeiro MF (2014) Unexpected low-dose toxicity of the universal solvent DMSO. Faseb j 28(3):1317–1330. https://doi.org/10.1096/fj.13-235440

    Article  CAS  Google Scholar 

  22. Gigout A, Buschmann MD, Jolicoeur M (2008) The fate of pluronic F-68 in chondrocytes and CHO cells. Biotechnol Bioeng 100:975–978. https://doi.org/10.1002/bit.21840

    Article  CAS  Google Scholar 

  23. Goldrick S, Lee K, Spencer C, Holmes W, Kuiper M, Turner R, Farid SS (2018) On-line control of glucose concentration in high-yielding mammalian cell cultures enabled through oxygen transfer rate measurements. Biotechnol J. https://doi.org/10.1002/biot.201700607

    Article  Google Scholar 

  24. Hafner M, Niepel M, Chung M, Sorger PK (2016) Growth rate inhibition metrics correct for confounders in measuring sensitivity to cancer drugs. Nat Methods 13:521–527. https://doi.org/10.1038/nmeth.3853

    Article  CAS  Google Scholar 

  25. Hayashi M, Honma M, Takahashi M, Horibe A, Tanaka J, Tsuchiya M, Morita T (2013) Identification and evaluation of potentially genotoxic agricultural and food-related chemicals. Food Safety 1(1):2013003–2013003. https://doi.org/10.14252/foodsafetyfscj.2013003

    Article  Google Scholar 

  26. Hazekawa M, Nishinakagawa T, Kawakubo-Yasukochi T, Nakashima M (2019) Evaluation of IC50 levels immediately after treatment with anticancer reagents using a real-time cell monitoring device. Exp Ther Med 18(4):3197–3205. https://doi.org/10.3892/etm.2019.7876

    Article  CAS  Google Scholar 

  27. Hsiao IL, Huang Y-J (2013) Effects of serum on cytotoxicity of nano- and micro-sized ZnO particles. J Nanopart Res 15(9):1829. https://doi.org/10.1007/s11051-013-1829-5

    Article  Google Scholar 

  28. Ihling N, Munkler LP, Berg C, Reichenbaecher B, Wirth J, Lang DT, Wagner R, Büchs J (2021) Time-resolved monitoring of the oxygen transfer rate of Chinese hamster ovary cells provides insights into culture behavior in shake flasks. Front Bioeng Biotechnol. https://doi.org/10.3389/fbioe.2021.725498

    Article  Google Scholar 

  29. Ihling N, Munkler LP, Paul R, Berg C, Reichenbächer B, Kadisch M, Lang D, Büchs J (2022) Non-invasive and time-resolved measurement of the respiration activity of Chinese hamster ovary cells enables prediction of key culture parameters in shake flasks. Biotechnol J. https://doi.org/10.1002/biot.202100677

    Article  Google Scholar 

  30. Kauffmann K, Gremm L, Brendt J, Schiwy A, Bluhm K, Hollert H, Büchs J (2020) Alternative type of Ames test allows for dynamic mutagenicity detection by online monitoring of respiration activity. Sci Total Environ. https://doi.org/10.1016/j.scitotenv.2020.137862

    Article  Google Scholar 

  31. Kauffmann K, Werner F, Deitert A, Finklenburg J, Brendt J, Schiwy A, Hollert H, Büchs J (2020) Optimization of the Ames RAMOS test allows for a reproducible high-throughput mutagenicity test. Sci Total Environ. https://doi.org/10.1016/j.scitotenv.2020.137168

    Article  Google Scholar 

  32. Kirkland D, Pfuhler S, Tweats D, Aardema M, Corvi R, Darroudi F et al (2007) How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: Report of an ECVAM Workshop. Mutat Res Genet Toxicol Environ Mutagen 628(1):31–55. https://doi.org/10.1016/j.mrgentox.2006.11.008

    Article  CAS  Google Scholar 

  33. Kirkland DJ (1994) Statistical evaluation of mutagenicity test data: recommendations of the U.K. Environmental Mutagen Society. Environ Health Perspect 102(1):43–47. https://doi.org/10.1289/ehp.94102s143

    Article  Google Scholar 

  34. Kochan J, Scheidle M, van Erkel J, Bikel M, Büchs J, Wong JE, Melin T, Wessling M (2012) Characterization of antibacterial polyethersulfone membranes using the respiration activity monitoring system (RAMOS). Water Res 46(16):5401–5409. https://doi.org/10.1016/j.watres.2012.07.019

    Article  CAS  Google Scholar 

  35. Krampe B, Al-Rubeai M (2010) Cell death in mammalian cell culture: molecular mechanisms and cell line engineering strategies. Cytotechnology 62(3):175–188. https://doi.org/10.1007/s10616-010-9274-0

    Article  Google Scholar 

  36. Larsson P, Engqvist H, Biermann J, Ronnerman EW, Forssell-Aronsson E, Kovacs A, Karlsson P, Helou K, Parris TZ (2020) Optimization of cell viability assays to improve replicability and reproducibility of cancer drug sensitivity screens. Sci Rep. https://doi.org/10.1038/s41598-020-62848-5

    Article  Google Scholar 

  37. Lewis NE, Liu X, Li Y, Nagarajan H, Yerganian G, O’Brien E et al (2013) Genomic landscapes of Chinese hamster ovary cell lines as revealed by the Cricetulus griseus draft genome. Nat Biotechnol 31(8):759–765. https://doi.org/10.1038/nbt.2624

    Article  CAS  Google Scholar 

  38. Li AP (1981) Simplification of the CHO/HGPRT mutation assay through the growth of Chinese hamster ovary cells as unattached cultures. Mutat Res 85(3):165–175. https://doi.org/10.1016/0165-1161(81)90032-7

    Article  CAS  Google Scholar 

  39. Lopez-Meza J, Araiz-Hernandez D, Carrillo-Cocom LM, Lopez-Pacheco F, Rocha-Pizana MD, Alvarez MM (2016) Using simple models to describe the kinetics of growth, glucose consumption, and monoclonal antibody formation in naive and infliximab producer CHO cells. Cytotechnology 68(4):1287–1300. https://doi.org/10.1007/s10616-015-9889-2

    Article  CAS  Google Scholar 

  40. Lutz W, Lutz R, Gaylor D, Conolly R (2020) Dose–response relationship and extrapolation in toxicology. Mechanistic and statistical considerations. In: Reichl F-X, Schwenk M (eds) Regulatory toxicology. Springer, Berlin, pp 1–23

    Google Scholar 

  41. Martinez-Monge I, Roman R, Comas P, Fontova A, Lecina M, Casablancas A, Cairo JJ (2019) New developments in online OUR monitoring and its application to animal cell cultures. Appl Microbiol Biotechnol 103(17):6903–6917. https://doi.org/10.1007/s00253-019-09989-4

    Article  CAS  Google Scholar 

  42. Martinez-Serra J, Gutierrez A, Munoz-Capo S, Navarro-Palou M, Ros T, Amat JC et al (2014) xCELLigence system for real-time label-free monitoring of growth and viability of cell lines from hematological malignancies. Onco Targets Ther 7:985–994. https://doi.org/10.2147/Ott.S62887

    Article  Google Scholar 

  43. Meier K, Herweg E, Schmidt B, Klement T, Regestein L, Büchs J (2013) Quantifying the release of polymer additives from single-use materials by respiration activity monitoring. Polym Testing 32(6):1064–1071. https://doi.org/10.1016/j.polymertesting.2013.06.004

    Article  CAS  Google Scholar 

  44. Mühlmann MJ, Forsten E, Noack S, Büchs J (2018) Prediction of recombinant protein production by Escherichia coli derived online from indicators of metabolic burden. Biotechnol Prog 34(6):1543–1552. https://doi.org/10.1002/btpr.2704

    Article  CAS  Google Scholar 

  45. Nguyen ST, Nguyen HTL, Truong KD (2020) Comparative cytotoxic effects of methanol, ethanol and DMSO on human cancer cell lines. Biomed Res Ther 7(7):3855–3859. https://doi.org/10.15419/bmrat.v7i7.614

    Article  Google Scholar 

  46. Oberly TJ, Bewsey BJ, Probst GS (1987) A procedure for the CHO Hgprt mutation assay involving treatment of cells in suspension culture and selection of mutants in soft-agar. Mutat Res 182(2):99–111. https://doi.org/10.1016/0165-1161(87)90058-6

    Article  CAS  Google Scholar 

  47. Ollig J, Kloubert V, Weßels I, Haase H, Rink L (2016) Parameters influencing zinc in experimental systems in vivo and in vitro. Metals 6(3):71. https://doi.org/10.3390/met6030071

    Article  Google Scholar 

  48. Prabhu A, Gadgil M (2019) Nickel and cobalt affect galactosylation of recombinant IgG expressed in CHO cells. Biometals 32(1):11–29. https://doi.org/10.1007/s10534-018-0152-0

    Article  CAS  Google Scholar 

  49. Ramaiah L, Hinrichs MJ, Skuba EV, Iverson WO, Ennulat D (2017) Interpreting and integrating clinical and anatomic pathology results: pulling it all together. Toxicol Pathol 45(1):223–237. https://doi.org/10.1177/0192623316677068

    Article  Google Scholar 

  50. Rana U, Kothinti R, Meeusen J, Tabatabai NM, Krezoski S, Petering DH (2008) Zinc binding ligands and cellular zinc trafficking: apo-metallothionein, glutathion, TPEN, proteomic zinc, and Zn-Sp1. J Inorgan Biochem 102(3):489–499. https://doi.org/10.1016/j.jinorgbio.2007.10.030

    Article  CAS  Google Scholar 

  51. Rehberger B, Wodarczyk C, Reichenbächer B, Köhler J, Weber R, Müller D (2013) Accelerating stable recombinant cell line development by targeted integration. BMC Proc 7(Suppl 6):P111–P111. https://doi.org/10.1186/1753-6561-7-S6-P111

    Article  Google Scholar 

  52. Reinhart D, Damjanovic L, Kaisermayer C, Sommeregger W, Gili A, Gasselhuber B et al (2019) Bioprocessing of recombinant CHO-K1, CHO-DG44, and CHO-S: CHO expression hosts favor either mAb production or biomass synthesis. Biotechnol J 14(3):1700686. https://doi.org/10.1002/biot.201700686

    Article  CAS  Google Scholar 

  53. Robison SH, Cantoni O, Costa M (1982) Strand breakage and decreased molecular-weight of DNA induced by specific metal compounds. Carcinogenesis 3(6):657–662. https://doi.org/10.1093/carcin/3.6.657

    Article  CAS  Google Scholar 

  54. Rogers WG, Winer L, Schwalfenberg M, Romero N, Kam Y (2022) Principle of mitochondrial toxicity assessment using Agilent Seahorse XF solution. https://www.agilent.com/cs/library/whitepaper/public/wp-principle-of-mitochondrial-toxicity-assessment-5994-4732en-agilent.pdf

  55. Saha S, Brannath W (2019) Comparison of different approaches for dose response analysis. Biom J 61(1):83–100. https://doi.org/10.1002/bimj.201700276

    Article  Google Scholar 

  56. Sebaugh JL (2011) Guidelines for accurate EC50/IC50 estimation. Pharm Stat 10(2):128–134. https://doi.org/10.1002/pst.426

    Article  CAS  Google Scholar 

  57. Shridhar S, Klanert G, Auer N, Hernandez-Lopez I, Kandula MM, Hackl M et al (2017) Transcriptomic changes in CHO cells after adaptation to suspension growth in protein-free medium analysed by a species-specific microarray. J Biotechnol 257:13–21. https://doi.org/10.1016/j.jbiotec.2017.03.012

    Article  CAS  Google Scholar 

  58. Sofuni T, Matsuoka A, Sawada M, Ishidate M Jr, Zeiger E, Shelby MD (1990) A comparison of chromosome aberration induction by 25 compounds tested by two Chinese hamster cell (CHL and CHO) systems in culture. Mutat Res 241(2):175–213. https://doi.org/10.1016/0165-1218(90)90122-i

    Article  CAS  Google Scholar 

  59. Stefanowicz-Hajduk J, Ochocka JR (2020) Real-time cell analysis system in cytotoxicity applications: usefulness and comparison with tetrazolium salt assays. Toxicol Rep 7:335–344. https://doi.org/10.1016/j.toxrep.2020.02.002

    Article  CAS  Google Scholar 

  60. Tan EL, Williams MW, Schenley RL, Perdue SW, Hayden TL, Turner JE, Hsie AW (1984) The toxicity of sixteen metallic compounds in Chinese hamster ovary cells. Toxicol Appl Pharmacol 74(3):330–336. https://doi.org/10.1016/0041-008x(84)90286-2

    Article  CAS  Google Scholar 

  61. Thompson LH, Fong S, Brookman K (1980) Validation of conditions for efficient detection of Hprt and Aprt mutations in suspension-cultured Chinese-hamster ovary cells. Mutat Res 74(1):21–36. https://doi.org/10.1016/0165-1161(80)90188-0

    Article  CAS  Google Scholar 

  62. Tuncer S, Gurbanov R, Sheraj I, Solel E, Esenturk O, Banerjee S (2018) Low dose dimethyl sulfoxide driven gross molecular changes have the potential to interfere with various cellular processes. Sci Rep. https://doi.org/10.1038/s41598-018-33234-z

    Article  Google Scholar 

  63. Weng LD, Li WZ, Zuo JG, Chen C (2011) Osmolality and unfrozen water content of aqueous solution of dimethyl sulfoxide. J Chem Eng Data 56(7):3175–3182. https://doi.org/10.1021/je2002607

    Article  CAS  Google Scholar 

  64. Wurm FM (2013) CHO quasispecies—implications for manufacturing processes. Processes 1(3):296–311. https://doi.org/10.3390/pr1030296

    Article  Google Scholar 

  65. Wurm FM, Hacker D (2011) First CHO genome. Nat Biotechnol 29(8):718–720. https://doi.org/10.1038/nbt.1943

    Article  CAS  Google Scholar 

  66. Wurm FM, Wurm MJ (2017) Cloning of CHO cells, productivity and genetic stability—a discussion. Processes 5(2):20

    Article  Google Scholar 

  67. Zhang M, Aguilera D, Das C, Vasquez H, Zage P, Gopalakrishnan V, Wolff J (2007) Measuring cytotoxicity: a new perspective on LC50. Anticancer Res 27(1A):35–38

    Google Scholar 

  68. Zou WG, Yan MD, Xu WJ, Huo HR, Sun LY, Zheng ZC, Liu XY (2001) Cobalt chloride induces PC12 cells apoptosis through reactive oxygen species and accompanied by AP-1 activation. J Neurosci Res 64(6):646–653. https://doi.org/10.1002/jnr.1118

    Article  CAS  Google Scholar 

Download references

Acknowledgements

The provision of a Kuhner TOM device by Kuhner AG (Birsfelden, Switzerland) is gratefully acknowledged. NI would also like to thank Andreas Schulte for his support in relation to data acquisition with the TOM device.

Funding

Open Access funding enabled and organized by Projekt DEAL. This project has received funding from the Innovative Medicines Initiative 2 Joint Undertaking (JU) under Grant agreement No. 777397. The JU receives support from the European Union’s Horizon 2020 research and innovation program and EFPIA partners Bayer, Byondis, GSK, Pfizer, Rentschler Biopharma SE, Sanofi, and UCB. This project has also received funding by the Federal Ministry of Education and Research (BMBF) and the Ministry of Culture and Science of the German State of North Rhine-Westphalia (MKW) under the Excellence Strategy of the Federal Government and the Länder.

Author information

Authors and Affiliations

  1. AVT - Biochemical Engineering, RWTH Aachen University, Forckenbeckstr. 51, 52074, Aachen, Germany

    Nina Ihling, Lara Pauline Munkler, Richard Paul & Jochen Büchs

  2. Rentschler Biopharma SE, Erwin-Rentschler-Str. 21, 88471, Laupheim, Germany

    Dietmar Lang

Authors
  1. Nina Ihling
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Lara Pauline Munkler
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Richard Paul
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Dietmar Lang
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jochen Büchs
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

NI contributed to conceptualization (lead), funding acquisition (equal), data curation (lead), formal analysis (equal), investigation (equal), methodology (lead), visualization (lead), and writing—original draft (Lead); LPM and RP were involved in data curation (supporting), investigation (equal), visualization (supporting), and writing—review and editing (supporting); DL was involved in conceptualization (supporting), resources (supporting), supervision (supporting), and writing—review and editing (equal); JB contributed to conceptualization (supporting), funding acquisition (lead), project administration (lead), supervision (lead), and writing—review and editing (equal). All the authors read and approved the final manuscript.

Corresponding author

Correspondence to Nina Ihling.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

DL is employed by Rentschler Biopharma SE. The remaining authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1.

Additional Data, Additional Table S1, Additional Figures S1–S7.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ihling, N., Munkler, L.P., Paul, R. et al. Introducing oxygen transfer rate measurements as a novel method for time-resolved cytotoxicity assessment in shake flasks. Environ Sci Eur 34, 97 (2022). https://doi.org/10.1186/s12302-022-00673-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12302-022-00673-5

Keywords