A ceramic microsparging aeration system for cell culture reactors

Transcription

1 Publication Series of IBPT University of Applied Sciences Giessen-Friedberg No. 1, 2 A ceramic microsparging aeration system for cell culture reactors Peter Czermak* 1, Christian Weber, Dirk Nehring Institute of Biopharmaceutical Technology University of Applied Sciences Giessen-Friedberg Wiesenstrasse Giessen - Germany Phone: Fax: peter.czermak@tg.fh-giessen.de *Department of Chemical Engineering Kansas State University Durland Hall 1 Manhattan KS , USA pczermak@ksu.edu Abstract The supply of oxygen in the cultivation of animal cells in bioreactors still poses a problem. Firstly animal cells must be supplied with the sufficient amount of oxygen, which in turn necessitates the maximum possible oxygen transfer into the liquid phase. Secondly animal cells are very shear stress sensitive. Three different procedures are used as standard in the aeration of cell culture reactors, which are basically bubble aeration, bubble-free aeration and indirect aeration. If the cell damaging impact of gas bubbles is reduced, direct aeration becomes a practical solution with scale up potential and comparatively high oxygen transfer rates. In addition all other aeration methods are technical more complicated and lead to less reliability because of the greater number of connection tubes. Thus in the production of biopharmaceuticals, gas sparging is an efficient and preferred process strategy for oxygen supply to cell cultures. In this paper, a microsparging aeration system made of porous ceramics is presented. The sparging system was used for the cultivation of mammalian cells like suspended CHO cells and adherend MDBK cells immobilized on micro carrier in 2 and L standard stirred tank reactors. The system produced bubbles of 1 μm. Already at a relatively low agitation rate of to 6 rpm in the sparged bioreactor a uniform and constant dissolved oxygen concentration was maintained in the medium. In fact, oxygen transfer measurements revealed that it greatly exceeded the cell requirements. The small bubble diameters and decreased gas flow resulted in a different pattern of foam formation as compared to standard aeration systems. Keywords: bubble aeration, oxygen supply, microsparging, ceramic sparger, CHO cells, MDBK cells, cell viability 1 Corrosponding author This paper is online available:

2 Introduction Successful culture of animal cells in bioreactors requires sufficient aeration of the culture medium. Aeration of cell culture bioreactors is usually performed by bubble aeration, bubble-free aeration or indirect aeration [1-7]. Bubble aeration delivers large amounts of oxygen into the medium. However, ascending or bursting bubbles in bubble aeration may damage shear stress sensitive cells. Cells adhere to the gas-liquid interface and are lifted to the surface where they may be damaged due to shear stresses when bubbles penetrate the surface and burst [8-11]. This leads to reduced cell viability and increased foam formation. The cells can be damaged by bursting foam bubbles due to the high surface tension. The bubble size appears have an effect on cell tendency to rise to the surface. Another disadvantage of bubble aeration is that foam is generated on the surface of the medium. This can lead to the medium being removed from the reactor. A variety of methods are available to combat foam generation. A silicon based anti-foaming agent is widely used [6]. A significant advantage of bubble aeration is the high oxygen mass transfer in water, this is due to the high volume specific phase surface area and a lower transport resistance in comparison with membrane-based techniques [1]. Due to relatively high transport resistance, bubble free aeration systems such as membranes, require large amounts of the membrane surface to achieve a sufficient oxygen supply to the medium. Compared to bubble aeration, greater effort in installation and maintenance is required and the demands of sterile operation make its use in production more problematic. Consequently, membrane aeration systems are mostly used in the laboratory and on the pilot scale [6, 7]. Other indirect aeration systems used are spin filters and vibro mixers. Furthermore, systems with external aeration are also used [6]. These procedures all involve high capital costs which make them unsuitable for large scale production. A scale up can also be prone to errors. Oxygen transfer then becomes progressively more difficult as one scale up the internal volume, since the mass transfer efficiency of stirred tank bioreactors will generally degrade as scale is increased [1]. While all other aeration methods are technically more complicated and lead to less reliability because of the increased number of connection tubes, direct sparging is still the method of choice for all scales [2, 6]. Kunas and Papoutsakis [11] reported that in bioreactors completely filled with microbubbles from to 3 µm in diameter damage to the cultured hybridomas was very low. A new system for the generation of micro bubbles was developed [12-14], in order to avoid the problems of convential bubble aeration. A porous ceramic with homogenous micro porous distribution was designed and put into production. Hydrophilic materials, such as porous ceramics for example, are very well suited to the production of higher phase surface areas on the basis of their small wetting angle of contact in water based medium. The aim was to produce the smallest possible micro bubbles equally distributed across the surface area and then apply this system to various scales of cell cultivation. Finally the system should be tested for cultivation of cultivation of mammalian cells like suspended CHO cells and adherend MDBK cells immobilized on micro carrier in standard stirred tank reactors. Special attention was given to cell viability and foam formation at the medium surface. Further, the k la-values of ceramic sparger and stainless steel sparger were compared under different conditions such as medium, aeration rate and agitation speed. Figure 1 a-c: Bubble aeration using the ceramic sparger in phosphate buffer. Material and Methods Cell Culture: The cells used were a CHO-K1 derived line (CHO-easyC, Cell Culture Technologies GmbH, Zürich, Switzerland). The original adherent cells were brought to suspension by growing in the Chomaster HP-1 Medium (Servichen GmbH, Gaggenau-Hörden, Germany). Throughout the experiments Chomaster HP-1 medium supplemented with.1 % Pluronic-F-68 (Applichem GmbH, Darmstadt, Germany) was used. 2

3 In addition Madin-Darby-Bovine-Kidney (MDBK) cells were used. The cells were grown in Dulbecco s Modified Eagle s medium () (Gibco/BRL, Eggenstein, Germany), supplemented with % fetal calf serum (FCS) (PAA, Germany) and kept at a temperature of 37 C and % CO 2 surface aeration when cultured in flasks. All cells were pre-cultivated in T-flasks. MDBK cells are strictly adherend cells and are considered to be shear sensitive cells. They were immobilised on micro carrier (Cytodex 3, 3gr/L, Amersham Bioscience). Initially, cells were grown in spinner flasks, incubated in a % CO 2 atmosphere (Medcenter Einrichtungen GmbH, Gräfelfing, Germany) at 37 C, to a cell density of about 1.4 x 1 6 cells/ml. Ceramic Sparger: The ceramic porous tube with an aeration length of 2 (fig. 1a) or 4 mm (fig. 1b, c) and an outer diameter of 1 mm consisting of titan dioxide or zirconium dioxide was fitted to a stainless steel construction (Applikon Biotechnology, Schiedam, Netherlands), as can be seen in figure 1. The open porous structure of the ceramic shows a homogenous pore size distribution, the average pore size is between 2 and μm. The SEM image (figure 2) was taken after fracturing the ceramic tube perpendicular to the long axis and sputtering the fracture surface with gold. Stainless Steel Sparger: Two standard stainless steel spargers were used, one from B.Braun International, Germany, and the other one from Applikon Biotechnology, Schiedam, Netherlands. Fermentation Equipment for Cell Cultivation: A 2L bioreactor system (Cell Bundle Europe, Bio Console ADI12, Bio Controller ADI11, Applikon Biotechnology, Schiedam, Netherlands) was used for the serum-free CHO cultivation. The bioreactor system was equipped with a marine impeller for mixing, and a ceramic microsparger (Figure 1) or a stainless steel sparger (Applikon Biotechnology, Schiedam, Netherlands) for bubble aeration. A 1L bioreactor system (Biostat B, B.Braun International, Melsungen, Germany) was used for the MDBK cultivation and a L vessel for k La measurement. The bioreactor system was equipped with a marine impeller for mixing, and a ceramic microsparger (Figure 1) or a stainless steel sparger (B.Braun International, Melsungen, Germany) for bubble aeration. For bubble free membrane aeration a silicon tube membrane (Raumedic, Rehau, Germany) with an outer diameter of d a= 3.2 mm and an inner diameter of d i=2.4 mm was used. The membrane ratio in the bioreactors was kept at 2. m/l. The cells were cultured in batch mode at ph of 7.2, temperature of 37 C, agitation speed of or 6 rpm, aeration rate of.4 vvm or, vvm (volume gas per reaction volume per minute at STP standard temperature and pressure) by using the ceramic sparger. Dissolved oxygen concentration in the medium was controlled by pulse modulation of the gas flow through the sparger. ph was controlled by adding gaseous CO 2 to the gas flowing in the sparger and NaOH to the medium. Medium samples were taken once or twice daily and analyzed for cell number and viability, and glucose concentration. Cell density and viability were determined by means of a hematocytometer (Neubauer chamber) and tryphan blue exclusion. Glucose concentration was determined using an enzymatic test (Glucose Assay Kit, PN: GAGO-2, SIGMA, Taufkirchen, Germany). Sample absorbance was assessed in a microplate reader (Sunrise, Tecan, Grödig, Austria) at a wavelength of 74 nm. Figure 2a + b: SEM images of the porous structure of a zirconium dioxide microsparger cross section (The images were taken after breaking the ceramic tube across the diameter). Evaluation of Oxygen Transport Efficiency: Oxygen transport from the gas phase into the medium was assessed in terms of phase surface specific transport resistance k La. The k La value was measured according to the dynamic method [3, 6]. In short, the medium was saturated with nitrogen until no more oxygen was detected in it. Then, a constant gas volume flow was fed to the sparger at a constant rotation speed of the impeller (6 rpm) and the dissolved oxygen concentration measured with time. Dissolved oxygen concentration was measured with an electrode sensor (Applikon 3

4 Biotechnology, Schiedam, Netherlands). The transient oxygen balance on the medium yields: k la (t 2 t 1) = ln [(c* - c t1)/(c* - c t2)] [1] k la: oxygen transfer coefficient [h -1 ] t: time [h] c*: oxygen saturation concentration [mmol m -3 ] c t: oxygen concentration at time t [mmol m -3 ] The k la-value was estimated from the slope of the logarithm of the dissolved oxygen concentration ratio vs. time difference curve. above all by the larger phase surface area and by the fact that the gas bubbles remained longer in the bioreactor. Results of Cell Culture Experiments: The CHO cells were cultured in batch mode for about seven days at ph of 7.2, temperature of 37 C, O 2- saturation of 4%, agitation speed of 6 rpm, aeration rate of.4 vvm by using the ceramic sparger and. vvm by using the stainless steel sparger. Figure 3 show the results of fermentations in respect to the cell density and the viability over the cultivation time and Figure 4 the cultivation parameter. Results and Discussion Oxygen Transport: Ceramic microspargers, such as that shown in figure 1 offer the possibility to deliver air in small bubbles with diameters from 1 µm to µm [6, 12]. The bubble size when aerated with stainless steel spargers produced diameters of 8-3 µm [6]. The developed aeration strategy is based on the development of porous ceramic materials. Through the homogenous pore diameters (Figure 2) of the ceramic micro bubbles are distributed equally in a small range of sizes with a large phase surface area. The aeration system consists of porous ceramic materials (Figure 1) like Al 2O 3, TiO 2 or ZrO 2. These are integrated, in combination with other biocompatible materials, into any reactor system. A significant advantage, in comparison to the bubble free, indirect aeration systems such as silicon tubes or Accurel membranes, lies in the smaller required aeration surface area. The aeration elements can be easily sterilised with standard sterilisation procedures and without the need for dismantling of the elements. The aeration elements were sterilised with saturated steam. The cleaning was successful with.1n sodium hydroxide solution. Table 1 shows the k la-values in dependence of the aeration rate, kind of medium and agitation speed. The k la-values which results by using the ceramic sparger are up to five times higher than the k la-values using the stainless steel sparger (see also [6, 12-14]). Table 1 kla values measured in stirred tank reactors (* from [6] and [12]) Aeration [vvm] Impeller speed [rpm] Reaction volume [L] 2 2 Medium ChoMaster HP1 ChoMaster HP1 k l a ceramic sparger [h -1 ] 31* 33* 33* * 2 - k l a stainless steel sparger [h -1 ] 9* 1* 1* 1* Table 1: k la values measured in stirred tank reactors (* from [6], [12-14]) The k la -values of the ceramic microsparging aeration system clearly show that a significantly better oxygen transfer into the medium was achieved. This is explained Time [h] Figure 3: The Graph shows the viability and cell density of two fermentations using a stainless steel sparger and two fermentations using the ceramic sparger. The trend lines are for visual interpretation only Viability Trend Line, Ceramic Sparger, ---- Viability Trend Line, Stainless Steel Sparger; Cell Density Trend Line, Ceramic Sparger; Cell Density Trend Line, Stainless Steel Sparger; Viability Ceramic Sparger, Cultivation 1; Δ Viability Ceramic Sparger, Cultivation 2; Cell Density Ceramic Sparger, Cultivation 1; Cultivation 1, Cell Density Ceramic Sparger, Cultivation 2; Viability Stainless Steel Sparger, Cultivation 1; Viability Stainless Steel Sparger, Cultivation 2; Cell Density Stainless Steel Sparger, Cultivation 1; Cell Density Stainless Steel Sparger, Cultivation 2 Temperature [ C] ph ph1 Time [h] Figure 4: CHO cell cultivation parameter Dissolved Oxygen Viability [%] [%] T emperature Dissolved Oxygen The results show that the viability is greater by using the ceramic sparger as well as the final cell density. By the cultivation of adherend MDBK-cells in a 1L reactor the ceramic microsparging aeration system produced the same growth curves as those produced by membrane aeration (Figure ). 4

5 Therefore no cell damage was observed as a result of the aeration. Conclusion Some publications indicate that direct bubble aeration in cell culture causes cell injury and consequently cell death and lower productivity [11, 9, 1]. Assuming that bubble size is an influential parameter for cell damage with the used ceramic micro sparging system it is possible to minimize the damaging effect of bubble breakup events and coalescence in the bulk in different bioreactors. The holistic approach to determine the effect of bubble aeration on cell death published by D. Martens and J. Tramper [16] suggests bubble break up at the surface is the main cause of cell depletion. However Michaelis et al. [17] and Kunas et al. [11] reported that smaller bubbles were not harmful to cells and were in fact beneficial in reducing cell damage. Cell density / ml 1E+7 1E+6 by the cultivation of MDBK cells adhered on Cytodex 3 microcarriers in 1 L scale [12] and by the cultivation of MDCK cells in 2L, L, 3L and 1L scale in comparison with membrane aeration [6]. Ceramic microspargers, such as that shown in Figure 1, offer the possibility to deliver oxygen as small bubbles. The open porous structure of the ceramic shows a homogenous pore size distribution. The average pore size is between 2 and μm with resulting bubble diameters from 1 to µm (measured photometrically in phosphate buffer [6, 12]). The microbubble size results in an increased gas-liquid contact area, increased gas mean residence times in the bioreactor, and increased oxygen transfer rates into the liquid medium (i.e., high k la). Under these conditions, a lower gas flow suffices to meet the cell oxygen requirement. Additionally, smaller bubble volumes result in lower relative speeds between cells and bubbles and reduce the probability of cellbubble collisions. As stated previously, culture methods preventing cell viability loss and foam formation are available. Due to the different foam forming tendency caused by the lower gas flow of the ceramic sparger compared to standard aeration systems, we were able to develop an appropriate process control strategy. 1E Time [d] Figure : MDBK-cells fixed on Cytodex 3 carrier in a 1L stirred tank vessel with -medium at 37 C, ph 7,2, impeller speed rpm; ( ) membrane aeration; ( ) ceramic microsparging aeration system Despite of the conclusions of other researchers which consider that smaller bubbles are more detrimental to animal cells [18], we believe that there are some physical advantages of smaller bubble sizes. These apparent inconsistencies derive from the fact that several parameters affect cell damage. Due to smaller bubble size, the relative surface of the gas phase in the liquid increases which leads to a higher mass transfer of oxygen into the liquid phase. Less gas flow is required to suit the oxygen demand for the living cells. Additionally, a smaller bubble volume causes a lower relative speed between the cells and the bubbles. The probability of cell bubble collision decreases and the mean dwell time of the bubbles in the medium increases ensuring that complete dissolving is possible. In this study we have not investigated the mechanism of cell damage and bursting bubbles, but we have tried to keep the energy dissipation as low as possible by using a high powerful gas dissipation system to offer a high oxygen transfer rate. Figure 7: Foam formation while CHO cell cultivation in a 2L stirred tank reactor The microsparging system can be used as an alternative aeration system and will fulfil easily the design considerations for large scale cell culture [, 7]. Compared to membrane aeration, the ceramic microsparger is easier to handle and more cost effective due to the fact that no expendable items (i.e. membrane tubes) are necessary. Due to the small bubble diameter a low agitation rate was applied, thus preventing cells from additional mechanical stress while increasing viability and decreasing foaming (Figure 7). In addition similar results were shown earlier

6 References 1 Fenge et al (1993): Agitation, aeration and perfusion modules for cell culture bioreactors, Cytotechnology 11: Henzler H.J, Kauling D.J. (1993): Oxygenation of cell cultures, Bioprocess Eng. 9: Krahe, M. (23): Biochemical Engineering, Reprint from Ullmann s Encyclopedia of Industrial Chemistry; VCH Publishers: Weinheim, Germany, Vol. 6 4 Vorlop, J. et al (1989): Entwicklung eines Membranrührers zur blasenfreien Begasung und Durchmischung von Zellkulturreaktoren im Pilotmaßstab, PhD Thesis, Technical University Braunschweig, Germany Nelson, K.L. et al (1988): Industrial scale mammalian cell culture, Part I: Bioreactor design considerations. BioPharm 1: Nehring D., Czermak, P. et al (24): Experimental study of a ceramic microsparging aeration system in a pilot scale animal cell culture, Biotechnol. Progress 2: Marks, D. M. (23): Equipment design considerations for large scale cell culture. Cytotechnology 42: Chisti, Y. (2): Animal-cell damage in sparged bioreactors. Trends Biotechnol. 18: Papoutsakis, E.T. et al (1991): Fluid-mechanical damage of animal cells in bioreactors, Tibtech 9: Chalmers, J.J. et al (1994): Cells and bubbles in sparged bioreactors, Cytotechnology 1: Kunas, K.T. et al (199): Damage mechanisms of suspended animal cells in agitated bioreactors with and without bubble entrainment, Biotechnology and Bioengineering 36: Nehring, D., Czermak, P. (22): Ein neues Begasungssystem für Zellkulturen im Laborund Produktionsmaßstab, in Beckmann, D. (Hrsg.): Technische Systeme für Biotechnologie und Umwelt Biosensorik und Zellkulturtechnik, S , ISBN , Erich Schmidt Verlag, Berlin 13 Patent DE A1 14 Patent DE A1 1 Zhang, S. et al (1993): A comparison of oxygen methods fot high-density perfusion cultures of animal cells. Biotechnology and Bioenginering, 41: Martens, D.; Tramper, J. (1991): Lethal events during gas sparging animal cell culture. Biotechnology and Bioengineering, 37: Michaels, J.D. et al (1996): Sparging and agitation-induced injury of cultured animal cells: do cell-to-bubble interactions in the bulk liquid injure cells? Biotechnology and Bioengineering, 1: Meier, S.J.; Hatton, T.A.; Wang, D.I.C. (1999): Cell death from bursting bubbles: role of cell attachment to rising bubbles in sparged reactors, Biotechnology and Bioengineering, 62: