Several important biopharmaceuticals such as interferons, hormones, vaccines and monoclonal antibodies are produced with cell submerged culture systems in stirred tank bioreactors under aerobic conditions. The success of gas-liquid oxygen transfer depends greatly on aeration and mixing intensity, while sensitive animal cells require gentle mixing and aeration.  Studies of oxygen transfer under these conditions are much less in number than the plethora of articles published covering microbial cultivations over the last 60 years¹.

The influence of impeller mixing and aeration on the oxygen mass transfer coefficient (k_La) was first studied more than sixty years ago and k_La was originally related to the gassed power consumption per unit volume of broth (P_g/V) and the superficial gas velocity (v_g)^2,3. In fermentation systems, an optimal ratio between the aeration rate and batch volume (vvm) is usually kept constant during the fermentation process transfer from small to large scale. Taking this into account, a correlation for k_La as a function of P_g/V and vvm as key parameters was proposed ⁴. A correlation in terms of agitator speed (N) and air flow rate (Q) was also recently proposed in relation to the conditions used for cell culture in miniature ambr^TM bioreactors, particularly in cases where surface aeration considerably contributes to the oxygen mass transfer ⁵.

Due to low oxygen solubility in the medium, the oxygen supply or oxygen transfer rate (OTR) required may become limiting at higher cell concentrations. Insufficient oxygen supply at higher cell concentrations is mainly due to the gentle mixing and aeration required, and the resulting low mass transfer coefficients. This problem can be solved by adding pure oxygen to the inlet gas stream, causing a few-fold increase in oxygen supply to be achieved⁶.

A five-liter stirred tank bioreactor system Biostat B, equipped with a Biostat B Control Unit and a Gas Mix Unit (B. Braun Biotech International GmbH, Germany), described in detail elsewhere⁶, were used in this investigation. The stirrer speed, temperature, pH and partial oxygen pressure were measured during the experiments. A stirrer speed of 60 < N < 180 rpm and a gas flow rate of 0.1 < Q < 0.9 L/min were used during the experiments. Inlet gas mixtures were prepared from gas cylinders with pure N₂, O_2, and CO₂ and measured with flowmeters. The MFCS/win supervisory control and data acquisition software package system was used for logging pO₂, stirrer speed, pH and temperature data online.

Power consumption was determined in a geometrically similar vessel and impellers of the same size in separate experiments on a Benco-ELB system (Bench Scale Equipment Co. USA). A dynamic gassing method with pure nitrogen was used to determine the mass transfer coefficient in a cultivation media without culture at various impeller speeds and aeration rates, as reported elsewhere⁶. Standard cultivation media based on sugars, vitamins, minerals, amino acids, salts and other additives were used. The inoculum of the Chinese Hamster Ovary (CHO) cell culture was prepared in a CO₂ incubator at 37°C and 5% CO₂ and the 5 L bioreactor was inoculated so that the starting cell concentration was 0.2×10⁵ cells/mL. All the experiments were performed under sterile conditions, with oxygen concentration in the air not dropping below 60%. A dynamic method was used daily for the oxygen uptake rate evaluation at various cell concentrations by following the course of the pO₂.The medium was first saturated with oxygen by gassing with air or an oxygen enriched gas mixture.

The experimental k_La data during our investigation were in the range 2–10 h^-1, while the P_g/V for example was lower than 15 W/m³ in all experiments. The data were processed with three types of equations, correlating k_La with the various operating parameters discussed above^2-5. Microsoft Excel Solver was used for this purpose. The correlation of k_La with (P_g/V; W/m³)) and (v_g; m/s) gives the following equation with the regression coefficient r² = 0.99:

k_La = 0.52 (P_g/V)^0.10 v_g^0.75                                                                 (1)

Considering that k_La is a function of Pg/V and vvm (min) as key parameters, the result is (r² = 0.99):

k_La = 29.6 (P_g/V)^0.11(vvm)^0.77                                                            (2)

In terms of agitator speed (N) and gas flow rate (Q), where surface aeration considerably contributes to the oxygen mass transfer, the equation is (r² = 0.99):

k_La = 2.51(N)^0.33Q^0.73                                                                        (3)

The value of the correlation coefficient for all three equation types is identical since their mathematical type is essentially the same. All three correlation coefficients show excellent agreement with the experimental results which may be due to the relatively low mixing and aeration intensity and the narrow range of operating conditions investigated here. All the results were obtained in one five-liter reactor, where v_g, vvm and Q are in linear correlation. Consequently, the effect of aeration through the last term in all three equations was very similar – exponents in the range of 0.73 – 0.77. Considering that the mixing power of a stirrer in a turbulent regime, and at low aeration rates, is almost proportional to P_g α N³, the exponent 0.33 on N in equation 3 correlates well with the exponents on P_g/V in the first two equations. According to the geometry of the system used, the contribution of surface aeration to the oxygen mass transfer may not be important in our case. Since vvm is usually used as a key parameter with regard to aeration in fermentation systems, correlation 2 may be the most appropriate for k_La estimation in the similar systems investigated herein.

The required oxygen transfer rate (OTR) may be considered as a product of two factors as stated below:

OTR = k_La (C* - C)                                                                          (4)

The upper value of k_La is limited due to the gentle mixing and aeration required in the case of mammalian cell cultivation. However, the second factor, which is the oxygen concentration difference between equilibrium and the actual liquid oxygen concentration (C* - C) can be increased by adding pure oxygen to the inlet gas stream. In this way a considerable increase in oxygen supply can be achieved with the same gentle mixing conditions. During our investigation, the highest estimated oxygen uptake rate was around 50 mg/Lh. Under possible mixing and gassing conditions for our culture system, only about 20 mg/Lh of oxygen could be supplied with air. By increasing inlet oxygen concentration to 50%, a sufficient oxygen supply was achieved even at the lowest stirrer speed. By further increasing the inlet oxygen concentration and stirrer speed, more than 100 mg/Lh of oxygen could be supplied to the media in our experimental system.

Oxygen consumption, together with oxygen supply in terms k_La and (C* - C), can be graphically presented and applied as a control strategy during the cultivation of mammalian cells in stirred tank bioreactors.

Selected references

1. Marks D.M., Equipment design considerations for large scale cell culture, Cytotechnology, 2003, 42, 21-33

2. Cooper C.M., G.A.Fernstrom, S.A.Miller, Performance of agitated gas-liquid contactors, Ind. Eng. Chem., 1944, 36, 504-509.

3. Van't Riet K., Review of measuring methods and results in nonviscous gas-liquid mass transfer in stirred vessels, Ind.Eng.Process.Des.Dev., 1979, 18, 357-364.

4. M.Moresi and M.Patete, Prediction of kLa in conventional stirred fermenters, J.Chem.Tech.Biotechnol, 1988, 42, 197-210

5. A.W.Nienow, C.D.Rielly, K.Bronsan, N.bargh, K.Lee, K.Coopman, C.J.Hewitt, The physical characterization of a microscale parallel bioreactor platform with industrial CHO cell line expressing an IgG4, Biochemical Engineering Journal, 2013, 76, 25-36.

6. M.Tisu and A.Pavko, Oxygen transfer in a laboratory stirred tank bioreactor during mammalian cell culture cultivation, Acta. Chim. Slov., 2010, 57, 123-128.