AU687059B2 - High performance cell culture bioreactor and method - Google Patents

Description

WO 95/21911 PCT/US94/02140 -1- HIGH PERFORMANCE CELL CULTURE BIOREACTOR AND METHOD FIELD OF THE INVENTION This invention relates to hollow fiber bioreactors. More particularly, this invention relates to a novel high performance bioreactor and to a method for its utilization.

BACKGROUND OF THE INVENTION Prior art bioreactors conventionally perfuse nutrient media through a single type of hollow fiber. The various disadvantages of such bioreactors include heterogeneous cell mass, difficult procurement of representative cell growth samples, and poor performance due to inefficient oxygenation and an inability to control oxygen levels. Moreover, microenvironmental factors such as pH cannot be effectively controlled. Mixed or co-culture of cells is not possible.

DEFINITIONS

In the specification and claims, the following terms have the meaning set forth below.

HPBr--High Performance Bioreactor.

Extracapillary Space (ECS)--The space within the bioreactor cartridge which is not occupied by the hollow fibers.

Intracapillary Space (ICS)--The totality of the space defined by the lumens of the hollow fibers in a bioreactor.

OXY-1--A hollow fiber oxygenator available from UniSyn Technologies, Inc., 14272 Franklin Avenue, Suite 106, Tustin, California 92680. The OXY-1 bioreactor is provided with 0.2 pm pore size polyethylene hollow fibers woven to form a mat. The active surface area for gas exchange is 1ft 2

I,

WO 95/21911 PC'I1US94/102140 -2- BR110--A small hollow fiber bioreactor commercially available from UniSyn Technologies, Inc. The BR110 hollow fibers are cellulosic, with molecular weight cutoff (MWc). The active surface area for cell growth is 1.5ft 2 Micro Mouse BR110--A sterile disposable bioreactor device equipped with a BR110 bioreactor with a length of silicone tubing in the ICS loop for gas exchange available from UniSyn Technologies, Inc.

Cell-Pharm 1000--A hollow fiber bioreactor instrument commercially available from UniSyn Technologies, Inc. The Cell-Pharm 1000 system includes a controller, pH sensor, gas mixture control capability, media reservoir, heater and peristaltic pump and pinch valves.

MAb--Monoclonal antibody.

Mitsubishi Oxygenator Fibers--0.2 pm pore size polypropylene single strands of hollow fibers from Mitsubishi, Japan.

Glucose Utilization Rate (GUR)--The rate at which glucose is consumed from the nutrient media in a batch feed system (in units of mass per unit time, e.g. gm/day) on the day prior to the current sample.

This is derived as follows: GUR F(Gf G) V(G 1 G (G 1 G2)/2 where Gf is the glucose concentration in IC media feeding G 1 the glucose concentration in IC recirculating media on the day prior to the current sample; G 2 the glucose concentration in IC recirculating media in the current sample, F, the IC media feeding rate (L/day), and V, the volume of IC media in reservoir Lactate Production Rate (LPR)--The rate at which lactate (or rather lactic acid) is produced anaerobically while glucose is consumed. Therefore, one molecule of glucose produces two molecules of WO 95/21911 PCT/US94/02140 -3lactic acid without oxygen involved in the reaction.

Lactate production rate (LPR), g/day is defined as: LPR [F(L-Lf) V(L-L 1 x 0.09, L (L 1

+L

2 )/2 where Lf is the lactate concentration in intracapillary (IC) media 0.1 mM); L 1 the lactate concentration (mM) in IC recirculating media on the date prior to the current sample; L 2 the lactate concentration (mM) in IC recirculating media.

SUMMARY OF THE INVENTION It has been discovered that inefficient oxygenation is a focal constraint upon the efficiency of hollow fiber cell culture devices. The invention provides a novel bioreactor which achieves greatly enhanced nutrient media oxygenation as compared with prior art devices.

Pursuant to the invention, nutrient media is supplied through an inner passage concentric to and hence surrounded by an outer bundle of oxygenator hollow fibers. The inner media passage is provided by a single, media porous tube or by an inner bundle of media supply hollow fibers. The exterior surface of the outer or oxygenator fiber bundle is spaced from the inner wall of the bioreactor housing to facilitate the mixing of cellular material and the sampling or harvesting of product cells.

The high performance bioreactors of the invention is particularly suited for the growth of highly oxygen sensitive cells, such as insect cells used in I WO 95/21911 PCT/US94/02140 -4the production of biomolecules for commercial purposes.

CONSTRUCTION OF THE HIGH PERFORMANCE BIOREACTOR This portion of this specification describes, by reference to Figures I to 10, the best mode presently known to the inventors for the construction of a bioreactor which embodies the invention.

Description of Figures 1 to Figure 1 is a schematic illustration of a bioreactor of the invention.

Figure 2 is a cross-section of one of two substantially identical ends of a bioreactor as shown by Figure 1.

Figure 3 depicts the hollow fiber housing and the critical components which enable efficient and scaleable fabrication of high performance bioreactor.

Figure 3A is an end view of the hollow fiber housing of Figure 3.

Figure 4 shows the Figure 3 assembly in the mold with potting compound.

Figure 5 illustrates the potted ends of the cartridge after removal from the mold, with the two cut marks for removal of excess adhesive and fiber indicated. Note that the mask component is sacrificial and is removed with excess adhesive. The hollow fiber guide remains imbedded in the adhesive as a permanent part of the end seal in the hollow fiber housing.

Figure 6 depicts the cut end of the bioreactor cartridge.

Figure 7 shows the final assembly including the dual port header, which provides a means to supply media to the inner bundle of fibers and oxygen containing gas to the outer concentric ring of fibers.

WO 95/21911 I' 1C1US9/02140 Figure 8 is a detailed scale drawing of the fiber guide. This component provides the means for both organizing the two types of fibers, and enables sealing of the central fiber bundle to an independent port from that of the outer bundle of fibers.

Figure 9 is a detailed scale drawing of the sacrificial mask which fits over the portion of the fiber guide to provide a clean surface for sealing the central fiber bundle port.

Figure 10 is a detailed scale drawing of the dual port header for supplying media to the central bundle of fibers and oxygenated gas to the outer fibers independently.

Referring to Figure 1, the bioreactor 10 includes a housing 11 having substantially identical headers 12 and 13 at its input and output ends.

Headers 12 and 13 have ports 14, 15, 16 and 17 for the introduction and withdrawal of oxygen containing gas and media.

Figure 2 illustrates an outer oxygenator fiber bundle 22 concentric with an inner media supply fiber bundle 23. The exterior periphery of the oxygenator bundle 22 is separated by space 24 from the inner wall of the housing 11. Header ports 14 and 16 accommodate the introduction of oxygen containing gas into and withdrawal of oxygen containing gas from the oxygenator bundle 22. Header ports 15 and 17 accommodate the passage of nutrient media into and the removal of nutrient media from the nutrient media supply bundle 22.

Product cells or proteins are sampled and harvested from space 24 through ports 18, 19, or 21. Inoculation cells may be introduced through ports 18, 19, 20 or 21.

WO 95/21911 'CrI/US'94/021,10 -6- A preferred method of fabricating a bioreactor embodying the invention includes six major steps now described by reference to Figures 3 to i. Prepare 4 3/4 inch length fiber bundles from fiber tows with the requisite number of strands.

Wrap the end of each bundle with tape so that fibers are held securely.

2. Assemble jacket (10) with the fiber guide (10a) and mask (O10b) as shown in Figure 3.

3. Preheat and prespin cellulose fiber bundles for 20 minutes at 1800 RPM and 52°C to remove excess glycerine processing aid. Carefully insert cellulose fibers through the central hole in the fiber guides at each end of the jacket assembly shown in Figure 3A. Similarly, put three bundles of oxygenator fibers through the three lateral holes in the fiber guide and ensure that they are loosely packed.

Further drying of the bundles is done at for 14-18 hours.

4. Place the warm jacket/bundle assembly in a potting mold to encase the free ends of the bundles as shown in Figure 4. This assembly is placed in a centrifuge with premixed polyurethane resin in an applicator. Spin the entire assembly at 1500 RPM for 50-60 minutes at 38-45°C.

The resin enters the jacket assembly (10) through either ports 18 and 20, or 19 and 21.

Cut the warm cured resin at the potted ends of the jacket/bundle assembly along the two "cut" lines shown in Figure 5. Remove the cut pieces along with the mask (10b shown in Figure to expose the clean surface of the fiber guide (10a in Figure 3).

This surface participates in sealing the header.

L-

WO 95/21911 PCI/US94./02140 -7- 6. The header is installed after applying polyurethane resin to the clean surface exposed by removing the mask, and to the outer rim of the jacket Curing occurs at room temperature over a 24 hour period.

Media Supply Fiber Types Hollow fibers represent a specific group of tubular articles that are within the scope of this invention. Therefore, any porous tube or fiber that is capable of being integrated into a recirculating (ICS) loop, while enabling perfusion of nutrient media to cells growing on the outside, is included in this invention. The following are examples of tubes or fibers preferred for use in this invention: Hollow fibers and tubes with dialysis, ultrafiltration and microfiltration properties, molecular weight cutoff (MWc) ranging from 1 kD to 1,000 kD; and pore size ranging from 0.01 pm to 5.0 pm, The most preferred range is from 10 kD MWc and about up to 1 pm pore size.

(ii) Materials of construction include polymers, graphite, ceramics (including porous glass fiber) and metals stainless steel).

Typically, polymers are required to have good physical properties tensile strength, melt temperature and glass transition temperature).

These include cellulose, polyethylene, polypropylene, polysulfone (and other engineering thermoplastics), polymethyl methacrylate, polyacrylonitrile, various polymer blends and the like.

Specifically preferred media supply fibers are from cellulose polymers, have a MWc of 10-1,000 kD and a pore size of 0.01-1.0 pm.

W()O 95/21911 I 102 1,10 -8- Oxygenator Fiber Types Hollow fibers are the only tubular configuration selected for oxygenation in this invention. These hollow fibers can be porous or non-porous. Porous hollow fibers for oxygenation are typically hydrophobic (like those made from polyethylene, polypropylene, polytetrafluoroethylene (Teflon) and the like). Examples of non-porous hollow fibers include (but are not restricted to) silicone, and silicone copolymer capillary tubing. They typically have high gas permeability through the solid wall structure.

Porous hydrophobic hollow fibers are most effective for gas exchange when the pore size is in the range of 0.01 to 0.5 p m Optimally, pore size shculd be in the range of 0.1 to 0.2 pm.

Specifically preferred oxygenator fibers are formed from polypropylene and a pore size of 0.15 pm.

Discussion of Operation The HPBr is equipped with the means for recirculating nutrient media via the lumen (or ICS) of a hollow fiber bundle or other tubular structures centrally located in the bioreactor jacket. These hollow fibers or tubes are selected for their permeability properties, and in some instances for their surface properties hydrophilicity or hydrophobicity). Media profused through the porous wall of these structures bath cells that have been inoculated in the ECS.

0 2 /C0 2 gas mixtures are supplied to the HPBr through the lumen tube or a bundle of hollow fibers that are situated concentric to the centrally placed hollow fibers or tubes. The gas mix flows counter current to the flow of media. One important operating parameter is the gas pressure in the

-I

WO 9/21911 P(IT/IIlS94/0l21,10( -9oxygenator fibers. Tho optimal pressure Is a function of the number of strands of fibers used, and the pressure of the perfusing media in the ECS. For example, a bioreactor embodiment described in the examples that had 180 Mitsubishi polypropylene fibers showed a bubble point of >2.5-3 psi. For a HPBr with 540 oxygenator fibers, the bubble point was closer to about 2 psi. As cell mass increases and cells adhere directly to the fibers, the bubble point will tend to increase.

Table 1 illustrates the relationship between the number of oxygenator and media fibers for two types of HPBr's described in Example II Type I and II). Two additional HPBr configurations (Type III and IV), and an OXY-1/BR110 combination are included for comparison. The figure demonstrates that as the ratio of oxygenator fiber surface area to media fiber surface area increases, oxygenation efficiency also increases at the expense of the extracapillary space available for cell culture.

HPBr Type IV represents a more extreme case covered by this invention, where the ann:lar cell culture space is within the range for a BR110 (conventional) bioreactor. The ECS of 11.8 mL is within the range expected from about 70% (maximum) fiber packing density for a BR110 or OXY-1. (The maximum space in the potted jacket with absolutely no fiber present is about 22 mL.) For the embodiment described in Example II, the oxygenator/media fiber surface area ratio range is between 0.03 to 5.5. The preferred range is between 0.7 and 2.6. As this data shows, the invention may be regarded as an oxygenator device within which cells are cultured, and media supply is provided by a relatively few porous hollow fibers or a large porous tube, a graphite tube.

1 L I1I~ y III WO }5/2 1 K"17119,S'1/ l ,IZ O I TABLE 1 Relationship Between Oxygenator and Media Supply Fibers in HPBr OXY-1/ High BR110 Com- Performance Bioreactors bination Type Type Type Type I II III IV* Oxygenator 1.0 0.1 0.3 0.5 1.1 Fiber S.A. (ft 2 Media Fiber S.A. 1.5 0.2 0.2 0 2 1.2 (ft 2 Oxy/Media Fiber 0.7 0.5 1.5 2.0 Ratio Extracapillary n/a 19.5 18.0 16.4 11.8 Space (mL) *Note: Based on 2000 media fibers at a maximum packing packing density The gas composition is important in defining the efficiency of oxygenation as well as the pH in the micro-environment for cell growth created in the ECS. Illustrative data showing the effect on pH is presented in Example II. This invention permits more accurate monitoring and control of pH in the optimal range required than had been possible previously.

Provision is made in the bioreactor for an annular space between the outer most oxygenator fibers and the inner wall of the jacket. This space is a specially designed part of the ECS for containing cell mass. This culture space facilitates mixing when the bioreactor is (optimally) periodically rotated to-und-fro (by 120 degrees).

Further, it facilitates sampling and harvesting of products (particularly intact viable cells, for cellular therapy applications). This space also enables microcarriers for anchorage dependent cells to be contained in the HPBr. A specific instance SUBSTITUTE SHEET (RULE 26) Wo ,/21911 V/ti1.i94/02 I.(0 -11where this is critical is in the culture of hematopoietic stem cells for cellular therapy applications. Stem cells are grown better in the presence of anchorage dependent stromal cells (including endothelial cells). However, stem cells are anchorage independent. Therefore, stromal cells immobilized in or on microcarriers can be co-cultured with stem cells. The micro-environment thus created will facilitate the hitherto unrealized goal of long term ex-vivo expansion of stem cells. High yields and viability for cells from this cell culture space are achievable.

Finally, expensive growth factors and cytokines that are vital to the growth and differentiation of various cell types can be applied in relatively small quantities to the ECS to further tailor the micro-environment.

Important and unique features of the invention, both separately and in combination, include: a relatively high ratio of gas to media supply fibers in a single jacket; control of oxygenation and pH by gas composition, pressure and flow rate; relatively low fiber packing density to provide an extracapillary space for growing large cell mass and containing microcarriers; porous tubes or hollow fibers are employed to supply media and porous or non-porous hollow fibers are concentrically arranged to supply oxygenated gases.

Two different types of hollow fibers are used, that are made from different materials. This is represented by the examples in this invention.

One type of fiber microporous polypropylene), is utilized where media perfusion and gas exchange in the respective bundles are achieved and controlled by regulating pressures. In WO 95/21911 PCT/US94/02140 -12applications where anchorage independent cells are being cultured at very high densities, hydrophobic fiber may be preferred so that cells grow primarily in suspension and not on the fibers.

Media supply may be provided by a single porous tube or by a bundle of hollow fibers.

Coatings are applied to hollow fiber or tube surfaces to influence cell attachment (either positively or negatively). For example, cellular matrix molecules found in connective tissue between cells in-vivo coated onto bioreactor fibers facilitates adhesion of anchorage dependent cells like endothelial cells and Chinese hampster (CHO) cells. Other molecules containing the active amino acid sequence arginine-glycine-aspartic acid, called "RGD" found in matrix molecules also provide this benefit. Cellular matrix molecules coated onto HPBr surfaces facilitate a microenvironment conducive to efficient gene transfer into hematopoietic stem cells.

Additional surface area for anchorage dependent cells is achieved by introducing microcarriers to the culture space in the ECS. Co-culturing of different cell types to gain synergistic benefit is permitted stem cells and stroma).

Addition of oxygen carriers (particularly hemoglobin and perflourochemical emulsions) to the ECS (or ICS), greatly enhances the ability to accurately control oxygen delivery to the cells and the rapid growth of cell mass.

EXAMPLE I Two 15 day cell culture experiments were conducted with 3C11 hybridoma cell line, that secrets IgG 1 monoclonal antibody in a UniSyn Cell Pharm 1000 pilot scale bioreactor system. In run 1, three UniSyn BR110 bioreactor cartridges (1.5 ft 2 cellulose -L I WO 95/21911 CT/)1US94/0214(0 -13hollow fibers) were equipped with a UniSyn OXY-1 oxygenator (1 ft 2 0.2 pm pore size polyethylene ,xygenator fibers) for oxygen mass transfer. In run 2, three high performance bioreactor (HPBr) equivalently sized cartridges and were fitted to evaluate the capacities of HPBr in 3C11 cell culture. These HPBr's all had 450 cellulose hollow fibers (with 10 kD M.W. cut-off) but #1 and #2 had different amounts of oxygenation fibers (in the form of a mat, as used in OXY-1); and, 504 polyethylene oxygenation fibers.

Figure 11 is a schematic of the flow path for three BR110 cartridges with one OXY-1 cartridge used for run 1. Media recirculation rate was established by setting the RPM of the media recirculation peristaltic pump. The mixed gas containing CO 2 and air was pumped through the oxygenator counter current to media recirculating in the ICS loop of the bioreactors by using the gas peristaltic pump.

Calibrating pH and temperature probes were performed according to the operator's manual. To control the pH in the range of 7.0 to 7.4 in ICS media, the control unit of the Cell Pharm 1000 adjusted the CO 2 percentage in the mixed gas automatically. Also, this control unit maintained the constant temperature at 37"C. Figure 12 is a schematic of the flow path for the three HPBr cartridges in run 2. No change has been made for ICS media flow path relative to run 2. However, a modified gas flow path was designed. This involved connecting the gas mixture supply to both the three HPBr's and the media reservoir. It should be noted that this crude modification of the existing Cell Pharm 1000 did not provide optimal control of gas flow and back pressure. Connecting the gas supply with the ICS ~P WO 95/21911 PCT/US94/02140 -14recirculation media reservoir facilitated pH control by the Cell Pharm 1000 system controller.

The entire flow path including all tubing, filters, probes, and reservoirs, renew, and waste bottles were autoclaved at 120"C for thirty minutes on the liquid cycle. Final assembly included mounting the bioreactor cartridge into the Cell Pharm 1000 System. Contamination check and system flush were set to run for 24 hours to ensure sterility, removal of glycerine (processing aid for cellulose fibers), and allow for system stability.

Media conditions: ICS--5% fetal bovine serum (FBS) 1% penicillin/streptomycin 2% glutamine in 1000 mL

DMEM.

(ii) ECS--20% FBS 1% penicillin/streptomycin 2% glutamine in 1000 mL

DMEM.

Each bioreactor cartridge, including the BR110 and HPBr, were inoculated with the same 3 x 108 cells (over 90% viability) into the ECS by using two sterile, 10 mL syringes with attached 18 gauge needles. One syringe contained 5 mL of ECS media and the c l.s. The second syringe was empty and used to collect the media displaced during the inoculation.

Product harvest was accomplished in a manner similar to the inoculation procedure. Harvesting was typically done three times a week and 10 mL/each, and the standard operating procedure for this Cell Pharm 1000 System was followed for maintaining cell culture daily.

Process parameters in the cell culture systems were set up to monitor glucose uptake, lactate production, NH 3 production, MAb production and perfusion rate of ICS media. Since one cannot e 1 e I WO 95/21911 IPCT/US94/02140 directly estimate the cell mass present in the hollow fiber bioreactor at any given time, the MAb production may be regarded as a means for measuring the effect of oxygen mass transfer within the two types of bioreactors in run 1 and 2. MAb concentration in the harvest was analyzed by using radial immunodiffusion assay, glucose, lactate, and NH3 sampled from the ICS of the bioreactor were analyzed by using a Kodak Ektachem Analyzer. The perfusion rate was determined by measuring the media level of renew bottles. Initial perfusion rate was started at 240 mL/day and was increased step by step and maximized up to 1000 ml/day until the end of the runs according to either the glucose concentration (when it fell as low as 1.5 g/L) or lactate concentration (when it reached as high as 20 mM).

Cumulative MAb production in each HPBr cartridge (run 2) for 3C11 cells, are compared to that in each BR110 (run 1) in Figure 13. The ratio of lactate production rate to glucose consumption rate was developed as an indicator to evaluate the effect of oxygen mass transfer. This ratio was significantly lower in Run 2 (shown in Figure 4) compared to Run 1. These results suggest that the oxygenator fibers in HPBr cartridges help to reduce the resistance of oxygen mass transfer to the grow cells. This was especially true for the HPBr cartridge with the Mitsubishi oxygenator fibers that gave higher MAb production levels than those in the other two HPBr's which have polyethylene oxygenator fibers. It was assumed that the relatively high dissolved oxygen concentration in the ECS lets cells grow more efficiently.

I

WO 95/21911 PCT/IUS94/02140 -16- EXAMPLE II A 25 day cell culture experiment (run 1) was conducted with 3Cl1 hybridoma cell line, that secrets IgG 1 monoclonal antibody. A test system was designed to evaluate the performance of three HPBr cartridges under controlled conditions. As a reference, a UniSyn Micro Mouse BR110 small scale bioreactor with a UniSyn OXY-1 oxygenator was employed in run 2.

Run 3 consisted of a UniSyn Micro Mouse BR110 small scale bioreactor with a silicone tubing for oxygen mass transfer by diffusion.

Run 1 used the test system that was designed to operate HPBr's with the media flow path and the gas flow rate and pressure manually controlled (see Figure 14). A peristaltic pump assembly was used with three pump heads to provide the same media recirculation flow through each HPBr at a flow rate of 100 mL/min. Temperature in the ECS of each HPBr was controlled at 37°C by immersing the three media reservoirs in a thermostatically controlled recirculating water bath. The ICS media pH was controlled in the range from 7.00 7.30 by aerating the space above the media in the reservoir with CO 2 in 75% to 85% air. This gas mix was also used to supply gas to the three HPBr cartridges with a constant flow rate of 38 mL (std(/min and back pressure of 1.5 psi.

The three HPBr cartridges each had 450 cellulose hollow fibers for supplying media to the cells by perfusion. Both HPBr #1 and #2 had 180 Mitsubishi oxygenator fibers. The number of Mitsubishi oxygenator fibers in HPBr #3 was three times higher than that in #1 and For run 1, the contamination check and system flush were preformed according to L 1 L WO 95/21911 I'CT/US94/021.I0 -17the instructions found in the Micro Mouse operating manual.

Run 2 was assembled in an oven for maintaining a constant temperature Premixed air and CO 2 was passed through the OXY-1 oxygenator at a rate of about 70 mL(std)/min counter current to the direction of the loop recirculating media in the ICS of the BR110 bioreactor. To control pH within the range from 7.0 to 7.4, the ratio of CO 2 to air was adjusted by using valves on two gas flow meters.

Run 3 was mounted in a CO 2 incubator for maintaining a constant temperature (37*C) and constant pH (from 7.0 to 7.2) in the IC media was established by setting the ratio of 7% CO 2 to 93% air in incubator.

Media conditions: ICS--5% FBS 1% penicillin/ streptomycin 2% glutamine in 1000 mL DMEM.

(ii) ECS--20% FBS 1% penicillin/ streptomycin 2% glutamine in 1000 mL DMEM.

3 x 108 cells (91% viability) were inoculated in the ECS by using two sterile, 10 mL syringes with 18 gauge needles. One syringe contained 5 mL of ES media and cells. The second syringe was empty and used to collect the media displaced during the inoculation. Both run 2 and 3 were inoculated with 4 x 108 cells (95% viability) in the ECS using the method illustrated above. The standard operating procedure for run 2 and 3 was followed for maintaining cell culture daily. For run 1, daily adjustment of back pressure and gas flow rate, sampling from each ICS media and harvesting from ECS of each cartridge were carried out.

I I ru WO 95/21911 PC'f/USi94/02110( -18- Process parameters in the cell culture systems were set up to monitor glucose uptake, lactate and

NH

3 production and MAb production. Cumulative MAb production was monitored to determine the effect of different oxygen mass transfer methods exemplified in runs 1-3.

ICS media for the bioreactors was replaced with fresh media when either the glucose concentration fell below 2.0 g/L or lactate concentration was higher than 20 mM. For run 2 and run 3, harvesting of MAb from the ECS of the bioreactor was done three times weekly and a volume of 10 mL/each (as defined by the Micro Mouse protocol). MAb production was analyzed by using radial immunodiffusion assay, and glucose, lactate, NH 3 concentration in harvesting of the ECS of HPBr's were also analyzed. The difference between the levels for these cell metabolites found in the ICS and the ECS of HPBr were noted.

Figure 15 shows the cumulative IgG 1 produced in each HPBr cartridge and in runs 2 and 3. HPBr #3 appeared to be superior in MAb production compared with HPBr #1 and A similar conclusion was drawn when the HPBr was also compared with BR110 bioreactor in run 2 and run 3. This suggests that the increase of oxygenator fibers in HPBr can significantly enhance cell growth and MAb production due to the increased oxygen supply. Also the ratio of LPR to GUR was calculated for HPBr #3 and found to be significantly lower than other bioreactors. This signifies more efficient utilization of glucose in the presence of increased oxygen in the ECS.

Figure 16 compares the lactate/glucose ratio for various bioreactors. The lowest ratio was shown by HPBr devices, and these low values remain below that of the conventional bioreactors and oxygenation ~e I WO 95/21911 PCT/US94/02140 -19methods for prolonged cell culture runs. Ideally, the ratio should be as close to zero as possible throughout the cell culture run, signifying optimal metabolic utilization of glucose.

Table 2 is a comparison of bioreactor and oxygenator properties. It also contains the relative antibody productivity of bioreactor/oxygenator combination as a function of their physical properties. In this embodiment, as the number (or surface area) of oxygenator fibers increase, and the ratio of oxygenator to media fiber surface areas increase, bioreactor productivity also increases.

L II r II--I I I _I TABLE 2 Comparative Bioreactor and Oxygenator Properties Property BR110 OXY-1 HPBr Type 1 HPBr Type II HPBr Type III (Bioreactor) (Oxygena- Media Oxygenator Media Oxygenator Media Oxygenator tor) Poly- Poly- Poly- Poly- Material Cellulose ethylene Cellulose propylene Cellulose propylene Cellulose propylene MWC or Pore Size 10 kD 0.2 um 10 kD 0.2 um 10 kD 0.2 um 10 kD 0.2 um Surface Area (ft 2 1.5 1.0 0.2 0.1 0.2 0.3 0.2 Number of Fibers 3150 1684 450 180 450 540 450 900* Permeability of Media Fibers (mL/hr. mmHg) 0.2 n/a 2.8 x 10 2 n/a 2.8 x 10- 2 n/a 2.8 x 10-2 n/a Typical IgG 1 Productivity (mg) for 3C11 Hybridoma 23 32,22 68 Data not available after 14 days at this time.

of culture (batch run) *Note: Maximum number of oxygenator fibers possible in this embodiment of HPBr is about 2000 1.1 ft 2 This assumes a packing density of 70%, which is in the range that is typical of conventional hollow fiber devices in general.

I

WO95/21911 PCT/US94(/02140 -21- Figure 17 illustrates the level of pH contrachievable with the HPBr. The particular data depicted is for HPBr pH control regulates biomass production vs. MAb production. Lower pH, 7.0, tends to encourage cells to remain in the exponential growth phase to produce high biomass.

Higher pH, 7.4, tends to encourage cells to remain in the bseady state phase and therefore exhibit higher productivity.

r I I -r I

Claims (16)

1. A bioreactor comprising: a housing comprising an inner wall defining a substantially tubular cell culture chamber, the chamber having first and second ends, a longitudinal axis, and a port providing access to said chamber; a central bundle of porous hollow fibers surrounding the longitudinal axis of said chamber, each of the fibers of the central bundle having an inlet end and an outlet end, the inlet ends of the fibers in fluid communication with a common axially-extending liquid inlet port; an annular bundle of gas permeable hollow fibers disposed concentric with, and surrounding, said central bundle, each of the gas permeable fibers of the annular bundle having an outlet end and an inlet end, the outlet ends of the gas permeable fibers in fluid communication with a common annular space surrounding the axially-extending liquid inlet port, a gas outlet port in fluid communication with the common annular spnao.

2. The bioreactor of Claim 1, wherein the annular bundle is spaced from the inner wall of the cell culture chamber to provide a cell culture annular space, the bioreactor capable of rotating around the axis thereof to facilitate mixing of fluid in the annular space.

3. The bioreactor of Claim 2, further comprising a plurality of microcarriers within said cell culture annular space.

4. The bioreactor of Claim 1, wherein the ratio of the surface area of said gas permeable hollow fibers to the surface area of said porous hollow fibers is from about 0.03 to about The bioreactor of Claim 1, wherein the central bundle of porous hollow fibers are arranged in a packing density of about AMEiNDED Sft PCnUS 9 02 1 -23- IPEUS 26 FB'96

6. The bioreactor of Claim 1, wherein the annular bundle of pas permeable hollow fibers comprises hydrophobic fibers having a pore size in the range from about 0.01 to about 0.5 microns.

7. The bioreactor of Claim 1, wherein the porous hollow fibers of the central bundle comprise a cellulose polymer having a molecular weight in the range 10 kD to 1000 kD.

8. The bioreactor of Claim 7, wherein the hollow fibers have a pore size in the range 0.01 to 1.0 microns.

9. The bioreactor of Claim 1, wherein an extracapillary space comprises from 55 to 90 volume percent of an interior volume of the bioreactor. A bioreactor comprising: a housing comprising ,i inner wall defining a substantially tubular cell culture chamber, the cl having first and second ends, a longitudinal axis, and a port providing access to said chamber; a central bundle comprising porous hollow fibers surrounding the longitudinal axis of said chamber; an annular bundle of gas permeable hollow fibers disposed concentric to, and surrounding said central bundle, outermost fibers of the annular bundle spaced from the inner wall to define an annular space for cell culture and mixing, when the bioreactor is rotated about the longitudinal axis; means disposed at said first and second ends for passing nutrient media through said porous hollow fibers; and means disposed at said first and second ends for passing an oxygen-containing gas through said gas permeable hollow fibers.

11. The bioreactor of Claim 10, further comprising a plurality of microcarriers within said annular space for cell culture. AMmN)D SHatr PCo 9 0 2 14 IPEAU S 6 FEB '96

12. The bioreactor of Claim 10, wherein the ratio of the surface area of said gas permeable hollow fibers to the surface area of said porous hollow fibers is from about 0.03 to about

13. The bioreactor of Claim 10, wherein the central bundle of porous hollow fibers are arranged in a packing density of about

14. The bioreactor of Claim 10, wherein the annular bundle of gas permeable hollow fibers comprises hydrophobic fibers having a pore size in the range from about 0.01 to about 0.5 microns. The bioreactor of Claim 10, wherein the porous hollow fibers of the central bundle comprise a cellulose polymer having a molecular weight in the range 10 kD to 1000 kD.

16. The bioreactor of Claim 15, wherein the hollow fibers have a pore size in the range 0.01 to 1.0 microns.

17. The bioreactor of Claim 10, wherein an extracapillary space comprises from 55 to 90 volume percent of aa interior volume of the bioreactor. AI*EWED SHEEI c 25 17. A method for the ex vivo expansion of stem cells which comprises providing a hollow fiber bioreactor having extracapillary space said extracapillary space containing microcarriers having stromal cells immobilized thereon, and concurrently culturing stem cells and said stromal cells in said

18. hereinbefore

19. described. bioreactor. A bioreactor according to any one of claims 1 to 16 substantially as described. A method according to claim 1' 7 substantially as hereinbefore r o o o o r r e c r or o o o a DATED: 17 November 1997 CARTER SMITH BEADLE Patent Attorneys for the Applicant: UNISYN TECHNOLOGIES, INC. MAW PP #21287 RSI 17 November 1997 L~ L