Clarkson Pharmaceuticals

Upstream Process Design for a Monoclonal Antibody Production Facility

1

1

Marie Rogers

2

2

Executive Summary

This report will discuss the design and material requirements for the upstream process of a large scale monoclonal antibody production plant. • • • Medium preparation Seed train Production reactor

3

3

4

4

Table of Contents

Executive Summary Background
Medium Preparation Cell Growth Seed Train
20 L WaveTM Perfusion Bioreactor 200 L Xcellerex Disposable Stirred Tank Reactor
TM

3 6
9 13 16
17 19 21 22

2,000 L Xcellerex Disposable Stirred Tank Reactor
TM

12,000 L Fermenter

Medium Storage Material Transportation pumps References Appendix A: Project Guidelines

26 27 28 31

Appendix B:Calculations
Exponential Cell Growth Production of MAb per Batch Carbon Balance for Glucose Consumption _
Carbon from MAb Produced Carbon from Lactic Acid Produced Carbon from Carbon Dioxide Produced

33
33 33 34
34 35 35

Total Glucose Consumption

36

Number of Reactor Lines Appendix C: MATLAB Code

36 37

5

5

Background Biopharmaceutical companies producing monoclonal antibodies perpetually investigate novel approaches for manufacturing their products. Process development investigates innovative methods to attain the goals of optimizing efficacy and preserving the product’s desirable qualities. Optimizing parameters in manufacturing has the potential to enhance cost effectiveness, improves production time, and to create manufacturing flexibility. Due to the high demand for large doses of monoclonal antibodies over an extended period of time, mass production of monoclonal antibodies is essential for fulfilling monoclonal antibody therapy regimens at an affordable cost. Marketable products in the clinic are qualified by production time and cost efficiency. In lieu of the increasing demand for large scale production, manufacturing plants containing multiple 10,000 liter or more cell culture bioreactors have become a generalized standard [1]. Monoclonal antibodies (mAbs) are a class of proteins produced by B cells, a white blood cell, in response to exposure to a specific foreign antigen, binding to their target with high affinity. The critical feature of antibodies is their biological role. The structure of the antibody determines its specific biologically relevant effector function. Monoclonal Abs are identical quaternary proteins characterized by polypeptide domains, including two heavy and two light chains and are structurally stable due to their hinge region, which creates flexibility. There are two relevant components of the antibody that are important: the variable antigen-binding cleft, comprised of two Fractionated Antigen-Binding (Fab) domains and the crystallized fraction (Fc) region, which promotes an evolutionarily conserved biological effector response. Genetic modification is a hallmark of the diversity of antibodies, typically facilitated through genetic

6

6

recombination, somatic mutation, and isotype switching, producing a monoclonal antibody possessing a high degree of specificity for binding its target molecule. The highly-affinity, specific binding of mAbs is a desirable quality required for novel therapies for a multitude of diseases; however, there are major limitations in the manufacturing of monoclonal antibodies. Several of these limitations include infusional toxicity with respect to purity, immunogenicity of chimeric proteins, limited biological features, and even mass production [B]. Traditionally, mammalian cells are noted for how difficult they are to use in manufacturing and scaling. This is because they tend to have a low yield, require complex medium and serum, and are sensitive to shear. Cell line development has been an important area of research over the past few decades. The end result of this research has resulted in cell specific productivities of over 20 pg/cell/day when media and bioreactor conditions are optimized. Through the optimization of medium composition, bioreactor operation, and conditions as well as by the selection of highly productive clones, mAb manufacturers have achieved high titers of up to about 10 g/L and cell densities of over 20 million cells/mL in fed-batch processes [1]. These achievements have caused the focus of cell culture process development to gradually shift from achieving higher titers to refining process regularity and quality at all production scales and development stages. http://pharmacologycorner.com/overview-on-monoclonal-antibodytherapy-ppt-images-and-videos/ 7

7

The production of monoclonal antibodies (mAbs) has great growth potential and is therefore of importance to chemical engineers. A recent publication by Landes Bioscience states that, “the robust implementation of this technology requires optimization of a number of variables, including (1) cell lines capable of synthesizing the required molecules at high productivities that ensure low operating cost; (2) culture media and bioreactor culture conditions that achieve both the requisite productivity and meet product quality specifications; (3) appropriate on-line and off-line sensors capable of providing information that enhances process control; and (4) good understanding of culture performance at different scales to ensure smooth scale-up.” [1] Clearly, chemical engineers face a significant and important challenge in developing and implementing processes requiring efficacy with respect to the economical production mAbs. Clarkson Pharmaceuticals is excited to approach this project. For full details of the design requirements see Appendix A. This report will deal specifically with the upstream processing part of the project, which starts, with the initial inoculum of CHO cells for harnessing their capability to produce the required titer of mAbs. This encompasses cell growth, medium preparation, design of the seed train and fermenter. The ultimate goal of this project is to be able to produce 1,000 kg of mAbs per year while achieving a final titer of 1 to 2 g/L [14]. Upstream Process Description The upstream process utilized during the large-scale manufacturing of monoclonal antibodies involves all steps of cell growth in a bioreactor, such as medium preparation, the cell culture seed train, media storage tanks, holding tanks, as well as pumps and tubing utilized in material transport of monoclonal antibodies. The major vessels employed are parameterized using

8

8

calculations explained for each component involved in the upstream process. The bioprocess flow diagram is depicted at the end of this report. Medium Preparation CHO cells have specific parameters that are required for their unique ability to produce monoclonal antibodies with high productivity and appropriate quality. As with all mammalian cells, medium conditions influence the efficacy of process development during the upscale process. In order for cellular growth, nutrient availability within the liquid culture medium must be sufficient [27]. When nutrients are rapidly depleted within the medium, cells undergo early apoptosis, metabolic changes, senescence, and cell cycle arrest [1]. Some commonly used methods that can be used systematically to ensure the integrity of the process include single-component titration, spent medium analysis, and medium blending. Amino acids, vitamins, trace elements, inorganic salts, lipids and insulin or insulin-like growth factors are all important ingredients used for defining the media needed for mAb production. Hydrolysates provide nutrient supplements to the media; they are protein digests made up of amino acids, small peptides, vitamins, minerals, and carbohydrates. Inadequate nutrition can lead to early apoptosis through rapid depletion of important nutritional growth components. Animalcomponent-free hydrolysates are often used to increase cell density, culture viability, and productivity when using a chemically defined media. Soy, wheat and yeast provide non-animal derived hydrolysates which are used commonly in cell culture media and feeds; however, hydrolysates can be a significant source of medium variability because of its composition complexity and lot-to-lot variations. Animal derived raw materials and bovine serum should not be used because of safety concerns related to transmissible spongiform encephalopathy (TSE)

9

9

and other contaminants [1]. For large-scale production, a dry, scalable, serum-free medium requiring minimal supplementation is preferred [25]. Developed by Invitrogen, Gibco’s CD CHO Advanced Granulation Technology (AGTTM) medium was developed specifically for culturing CHO cells and will be the medium utilized in the largescale production of mAbs, meeting the facility’s animalfree policy, giving rise to the http://tools.invitrogen.com/content/sfs/brochures/332_021646_AGT_Bro.pdf chemically-defined nature of the purchased product. The composition of this medium is chemically-defined, possessing no traces of proteins, unknown molecules, or hydrolysates, in order to give rise to a reagent that consistently provides optimal growth signals that are reproducible, sensitive and nutritious for CHO cell lines, saving time and money. The CD CHO AGTTM is shipped in dry ice as a powder to provide consistency and productivity during the manufacturing process and does not require supplementation with thymidine and hypoxanthine, due to the CHO cell line preference for a proline rich environment [XXX]. The designed as a serum-free media, Gibco’s CD CHO AGTTM is also specifically formulated without phenol red to minimize estrogen-like effects [4]. Gibco’s CD CHO AGTTM will be stored in a dark environment, protected from light, for up to 24 months. Multiple steps must be taken to completely reconstitute the media. First, 90% 10 10

of the final volume of water for injection will be accurately measured, followed by the addition of the CD CHO AGTTM medium to the water to yield a concentration of 20 g/L, which is then mixed for 30 minutes or until completely dissolved. CD CHO AGTTM dry media was designed to dissolve almost instantly to reduce contamination with dust particles. The medium mixture will then be diluted using water its final volume and concentration to prepare for injection. In order to sterilize the medium, membrane filtration with 2 µm pores is used. Once sterile, the medium will be stored in the absence of light between 2 to 8 °C [4]. Prior to use, the medium is supplemented with 40mL/L of 8mM L-glutamine, using Gibco’s GlutaMAXTM-I that is designed for this purpose [5]. GlutaMAXTM-I does not spontaneously break down to form ammonia, like L-glutamine. Contrarily, cells produce Lglutamine, ad hoc, by cleaving dipeptide bonds, to prevent accumulation of harmful waste and providing a fresh supply of L-glutamine which is an essential requirement for extended cell culture [6]. L-glutamine is required for cell culture, as it is consumed when cells enter the division phase. Consequently, in order to maintain the process of division, L-glutamine must be replenished [3]. When the media is needed for use, it will be aseptically supplemented with GlutaMAXTM-I, adding 1 mL/L of Gibco Anti-Clumping Agent to the medium only when cell clumping is apparent [4]. Before the cells are added to the medium we will add a support surface for the cells to grow on in the larger tanks as CHO cells, like most mammalian cells, are anchorage-dependent for proper growth [7]. One surface proven to contribute to the growth of CHO cells is macroporous microcarrier beads [8]. The unique characteristic of these beads is their network of pores that provides cells with a protected area to grow by reducing cell death from shear forces brought about by agitation [9]. Clarkson Pharmaceuticals will use CytoporeTM 2 microcarriers 11 11

from GE Life Sciences because they are designed to function specifically for the culture of CHO cells [10]. Macroporous microcarrier cultures are normally comprised of 1–2 g of CytoporeTM per liter [11]. In the upstream process of monoclonal antibody production, it is required to estimate the quantity of CytoporeTM needed to for cell growth within each seed reactor and bioreactor, using the aforementioned standard for calculation. The dry CytoporeTM 2 macroporous microcarriers will be prepared for use through hydration with 10X phosphate buffered saline (PBS), purchased from GE Life Sciences, in a siliconized bottle using a volume of 50-100 https://www.gelifesciences.com/gehcls_images/GELS/Rel ated%20Content/Files/1314787424814/litdoc18113268_2 0111214164017.pdf

mL of PBS per gram of CytoporeTM[11, 12]. The solution

will be gently agitated for approximately 10 minutes at room temperature, and will be autoclaved at 121 °C for 20 minutes [11]. These prepared microcarriers have the potential to suspended in the medium for use in the seed reactors and the production bioreactor when needed. The feed composition and feeding strategy can be greater refined by interfering with the nutrient composition, by-product accumulation, and the balance between promoting growth versus volumetric productivity. Minimizing by-products such as lactate and ammonia has been achieved by sustaining low glutamine and glucose concentrations through continuous or frequent feeding strategies. However, these strategies are less desirable because of operational difficulty and cost required to validating continuous feeding strategies for large-scale manufacturing. The most widely used method to feed the production reactor is a step-wise bolus addition of the feed solution. This method is used because of its simplicity and scalability [1].

12

12

Cell Growth Growth curves model the phases of cell growth over a period of time. The typical phases of growth are characterized by population size and rate of expansion and include: lag phase, log (exponential) phase, stationary phase, and a death phase. For my design, it is assumed that cells are indefinitely in the log (exponential) phase of growth, and was the rationale behind comfortably estimating the number of cells in the absence of CHO cell growth kinetic information. Realistically, the culture should first experience a lag phase, where growth is slow as cells adjust to the medium [1]. The lag phase is likely to occur with each transport of cells into a new vessel. Many factors influence the duration of the lag phase, such as cell species and age, inoculant size, and the conditions within the culture, i.e. nutrient concentration. Some of the hallmarks of the lag phase are a lag in division and an increase in mass, supporting growth and adjustment to media conditions. The physiological state of cell growth is characterized by an abundance of cellular machinery constituents, initiating the exponential growth phase, where the initial the division rate would accelerate to a maximum value. Growing cells in vitro has limitations regarding nutrient availability and accumulation of toxic metabolic by-products. As a result, cell growth rate declines and the rate of cell division becomes equal to the rate of cell death, eventually reaching a stationary phase. The stationary phase enables the cells to adapt to their limited environment, and is followed by a death phase where cells lacking survival advantages die due to lack of cellular energy reserves [13]. The project will begin by using a high cell density banking process to achieve an inoculum concentration of 4.5 x 108 cells per 4.5 mL, as thoroughly described in detail in Development and Implementation of a Perfusion-Based High Cell Density Cell Banking Process [2]. The number of viable cells produced by the inoculum over a time period in hours was found 13 13

using Equation 1, which is the generalized equation used to model exponential cell growth. Calculations derived from Equation 1 are found in Appendix B: Calculations. Equation 1 With respect to the notation of Equation 1, N0 is the initial number of cells, t is the time, and k is the time constant. The value for k was determined using a doubling time of 36 hours [14], typical of mammalian cells. For the purposes of this design, the value of k was found to be .019254 hr-1. The minimum and maximum cell concentrations (4.5 x 104 cells/mL, 2.5 x 106 cells/mL in the seed train, and 2.0 x 107 cells/mL in the fermenter, respectively) and Equation 1 were used to estimate the amount of time allowed for cell growth in each vessel. The logarithm of the number of cells and the logarithm of the cell concentration as a function of time are depicted in Figures 1 and 2, respectively. The inlet and outlet number of cells, as well as cell concentrations within each vessel is found in Tables 1,2,3,4.
Figure 1 : The logarithm of the number of CHO cells as a function of time

14

14

The CHO cell line produces monoclonal antibody at a rate of P=25 (pg/day*cell) [14]. For the quantification of mAb per batch, the number of cells is multiplied by the production rate and integrated with respect to time over the total number of days per batch, using the calculations shown in Appendix B: Calculations. Figure 3 describes the extent of mAb production as a function of time.
Figure 2 : The logarithm of the CHO cell concentration as a function of time.

15

15

The manufactured mAb acquired from the fermenter must be capable of yielding a titer greater than 1.0 g/L; otherwise, the production goal of acquisitioning efficacy at a reduced cost becomes an issue. The final scaling up from the seed train to the
Figure 4 : The mAb titer as a function of time Figure 3 : The amount of mAb produced as a function of time per batch

production bioreactor (fermenter) gives rise to the concentration of the titer produced, depending heavily on the length of time the cell culture is held within the vessel A target titer of 1.1 g/L is utilized in this design, and is the value corresponding to the total number of hours for the time spent within the seed train added to the hours spent within production reactor, which is approximately 686 hours. Figure 4 summarizes the mAb titer produced as a function of time, where the mAb titers at the inlet and exit of each vessel are presented in Tables 1,2,3,4. The longest portion of time spent in any one vessel during product development is the time spent in the production reactor. The number of batches per year can be approximated, dividing the number of operating days by the 8.4 days spent in the final production reactor. Assuming there are an estimated 328 operating days, the number batches per year for one operating line is around 39. Furthermore, the total mass in grams of mAbs produced per year for one operating line was determined to be 514,800 grams, as shown in Appendix B: Calculations.

16

16

Thus, the number of reactor lines needed to surpass the design requirements of producing 1,000 kg of product per year is 3 as shown in Appendix B: Calculations. Seed Train The seed train is used to build up the cell density and volume before going into the fermenter, where a large portion of the mAb production will take place. Of the seven vessels used in the upstream processes, four are described below. The cell culture seed train is where proliferation occurs, altering the density of cells within the vessel. The capacity for CHO cells to grow and divide is optimal when the cellular density is between 3 to 5 x105. When the cell density approaches the upper limit, the cell culture is transferred into a larger vessel for maintaining growth and division. 20L WaveTM Perfusion Bioreactor

https://www.gelifesciences.com/gehcls_images/GELS/Related%20Content/Files/1314823637792/litdoc28979083_20111214164418.pdf

The first vessel used for culture growth is a WaveTM perfusion bioreactor, manufactured by GE HealthCare. This disposable bioreactor is purchased with the 20/50 system along with a 20 L cell bag to allow for growth with perfusion, temperature and pH control. Due to the nature of the working volume in which half of the total Cellbag volume may be occupied, the 4.5 mL of culture along with 9.996 L of medium will be transferred to this bioreactor and allowed to grow for 182

17

17

hours. Temperature and pH will be maintained at 37 ± 0.5˚C and 7.0 ± 0.1 using integrated sensors to project the measured in vivo conditions and acquire higher viable cell densities [15]. Using a disposable WaveTM Bioreactor is advantageous. First, the WaveTM Bioreactor has a simple user interface, requiring minimal training, and is considerably less costly to operate than similar technologies. The cell culture bags are designed for single use and are pre-sterilized, eliminating the time and resources required for cleaning and sterilization by reducing the risk associated with culture contamination through conventional approaches. In addition, utilities such as steam, cooling water or process gases are not required [16]. The specifications and calculations for the design of this bioreactor are listed in Table 1.

Table 1. Specifications and Calculations for 20L WaveTM Perfusion Bioreactor 20L WaveTM Perfusion Bioreactor Working Volume (L) 10 Amount of Inoculum Added (L) 4.5 X 10-3 Amount of Medium Added (L) 9.996 Number of Doubling Periods 5.06 Residence Time (hours) 182 Initial Final 8 Number of Cells 4.5 X 10 1.5 X 1010 Cell Concentration (cells/mL) 4.5 X 104 1.5 X 106 MAb Titer (g/L) 0 .0787 MAb Product (g/batch) 0 .787

One component of the WaveTM perfusion bioreactors is the container for the cell culture, which is a flexible, disposable pre-sterilized plastic bag called the Cellbag®. After partially filling the Cellbag with media and culture, it is stabilized on a rocking platform for gentle agitation. The volume remaining in the Cellbag is filled with process gas, a mixture consisting of CO2 and O2. The process behind the addition these gases is facilitated by a pre-attached sterile 18 18

inlet filter, which elicits oxygenation and gas exchange while also maintains pH control and CO2 extraction. A backpressure control valve is attached to the exhaust outlet sterile filter as a way to secure appropriate inflation of the Cellbag. The back and forth rocking motion of the Cellbag is responsible for liquid mixing and mass transfer of gasses, generating waves at the liquid-air interface. The presence of waves at this interface expands the surface area needed to augment gas transfer. In order to parameterize the degree of rocking motion, the angle and number of rocks per minute must optimized and controlled. Initially, the parameters will be characterized by 7.5° rocking angle and a 20-30 rpm rocking speed [2]; these settings will be adjusted for optimizing liquid mixing and gas exchange if necessary. The internal temperature of the Cellbag is controlled and maintained using a heater and a non-invasive sensor that are integrated into the unit base [17]. 200 L and 2000 L XcellerexTM Disposable Stirred Tank Reactors

http://www.xcellerex.com/pdf/XDR-profile.pdf

After the cells have been cultured in the 20 L WaveTM perfusion bioreactor for 182 hours, the cell culture will be transferred to an XcellerexTM 200L Disposable Stirred Tank Reactor. The seed reactor is designed for single-use and has a working volume of 200 L [18]. The stirred tank reactor is a multi-component reactor with a stainless steel shell, the appropriate, fitted single-use 19 19

bag, and an impeller that is shielded with a sterile sleeve. Together, each component ensures the appropriate mixing of the cell culture and reduces undesirable cell death as a result of stagnancy for too long [19]. The transfer of 10 L of dculture into this vessel is accompanied by the addition of 190 L of medium. The contents within the reactor will be grown for 157 hours. This reactor’s design specifications and calculations are displayed in Table 2. The temperature maintenance within this vessel, specified at 37.0 ˚C, is accomplished by the XDRTM’s external temperature control unit, which will be programmed to automatically heat or cool the culture if the temperature is altered. Table 2. Specifications and Calculations for the 200L XDRTM. 200L XDRTM Working Volume (L) Amount of Inoculum Added (L) Amount of Medium Added (L) Number of Doubling Periods Residence Time (hours)

200 10 190 4.32 157

Initial Final Number of Cells 1.5 X 1010 3.0 X 1011 Cell Concentration (cells/mL) 7.5 X 104 1.5 X 106 MAb Titer (g/L) 0.0039 .0810 MAb Product (g/batch) .787 16.2 During this portion of the seed train, there may be an unacceptable deviation in the cell culture’s pH range as a consequence of accumulating carbon dioxide produced by the exponential increase and perpetually increasing number of proliferating cells. When the concentration of carbon dioxide is high enough, the pH of the culture is reduced, and has the advantage of being adjusted with ease using a base. CD CHO Medium AGT™ is made up of many ingredients, including sodium bicarbonate – a common buffer used to maintain pH [4]. Thus, if the pH were to falter outside the acceptable range and exhibit a skewing toward an

20

20

acidic pH, the weak base, sodium bicarbonate, has the potential to return the cell culture’s pH within the acceptable range. Conversely, should the culture’s pH increase to a more basic degree, CO2 has been shown to reduce the pH. Designed for convenience, XDRTM incorporates a control to monitor pH levels using sterile probes that activate a sterile peristaltic pump to feed base or CO2 from storage units to the system when the pH balance is offset. Essentially, the need for automated control is high because if the pH falls outside of the acceptable range for an extended period of time, the cells may die and are unviable giving rise to difficult troubleshooting [19]. The next vessel is a 2000 L XcellerexTM Disposable Stirred Tank Reactor, where the cell culture will be transferred, and possess many of the same features and operating procedures as the 200 L vessel. During the scale-up process, 1800 L of medium and the 200 L of culture will be added to the bioreactor. As previously mentioned, CytoporeTM macroporous microcarriers will be used to optimized cell growth and reduce cell death from shear forces at this large-scale level. To the seed reactor, 2.5 kg of CytoporeTM will be added in order to maintain exponential proliferation and growth of the CHO cells within the reactor for 146 hours. The specifications and calculations involved in the design of this bioreactor are found in Table 3. Table 3. Specifications and Calculations for 2000L XDRTM. 2000 L XDRTM Working Volume (L) Amount of Inoculum Added (L) Amount of Medium Added (L) Number of Doubling Periods Residence Time (hours) Number of Cells Cell Concentration (cells/mL) MAb Titer (g/L) MAb Product (g/batch)

2000 200 1800 4.06 146

Initial Final 11 3.0 X 10 5.0 X 1012 1.5 X 105 2.5 X 106 0.0081 .1352 16.2 270.4

21

21

12,000 L Fermenter Once the time requirements within the final growth vessel have been met (146 hours), the cell culture must be transported to a tank reactor with stirring for the production of mAbs by the CHO cells. The bioreactor, also referred to as a fermenter, is constructed with stainless steel and is a reusable reactor with a known working volume of 12,000 L, where CytoporeTM macroporous microcarriers will be added. The purpose of these microcarriers is to promote cell growth and reduce cell death from the shear forces exhibited by the production bioreactor. Suggestively, the calculations affirm that 10.0 kg of CytoporeTM should be added to this production bioreactor for efficacy in optimizing cell growth and productivity. The cell culture is grown for 201 hours in the production bioreactor as a mechanism for achieving an antibody titer of 1.08 g/L. The summary of the specifications and calculations applied to this bioreactor are located in Table 4. This bioreactor will be stirred using a 45° pitched blade turbine for continuous mixing, resulting in the reduction of eddy formation while avoiding the implications accompanied by excessive shear stress [20, 21]. Table 4. Specifications and Calculations for the 12,000 L Fermenter. 12,000 L Reusable Reactor Working Volume (L) 12,000 Amount of Inoculum Added (L) 2,000 Amount of Medium Added (L) 10,000 Number of Doubling Periods 5.59 Residence Time (hours) 201 Initial Final 12 Number of Cells 5.0 X 10 2.4 X 1014 Cell Concentration (cells/mL) 4.17 X 105 2.0 X 107 MAb Titer (g/L) 0.0225 1.08 MAb Product (g/batch) 270.4 12,960 22

22

http://www.bioprocessintl.com/journal/2012/June/Vertical-Integration-of-Disposables-in-Biopharmaceutical-Drug-Substance-Manufacturing331152

Glucose Consumption and Cell By-products Formation There is a high degree of glucose consumption that accompanies the monumental number of proliferating CHO cells that use glucose to carry out necessary metabolic functions. Glucose levels act as a threshold that must be met for the support cellular growth and production of mAbs within the production bioreactor. In order to counteract the diminishing availability of glucose within the fermenter, it is possible to supplement these levels with glucose during their time of residence within the fermenter. The quantitative carbon balance of each of the following metabolites is used to estimate the relative rate of glucose consumption using quantitative measurements relating to the release of by-products during the production of monoclonal antibodies within the fermenter: consumed glucose, monoclonal antibodies produced, and the 23

23

production of lactic acid and carbon dioxide. These calculations are summarized explicitly in Appendix B: Calculations. The rationale behind the design of production assumes each monoclonal antibody produced has the same physical properties as another mAb, AvistinTM, manufactured by another pharmaceutical company, Genentech [22]. The physical properties assumed for determining the relative need of glucose within production bioreactor is determined by utilizing multiple constituents within the reactor, including their physical properties. Monoclonal antibodies share physical properties; the molecular weight of a typical mAb is 149 kD and its chemical composition yields 6.64 x 103 moles of carbon for every mole of mAb [23]. Fermentation is the conversion of glucose into carbon dioxide and lactic acid. The initial nutrient conditions are modified as cells produce mAbs under proper fermentation conditions. Carbon can be analyzed over the course of time spent in the production bioreactor to optimize the upstream process by using monoclonal antibodies as readouts. Furthermore, costs associated with maintaining the cellular environment during fermentation can be estimated using glucose consumption rates and mAb production rates. The molecular weight of glucose is approximately 180 g/mol, with 6 carbon atoms for each glucose molecule. Ideally, the approximation of glucose consumption assumes the carbon atoms from the glucose are used in the production of the mAbs, lactic acid, and carbon dioxide; however, the energetic potential of glucose is allocated toward a variety of cellular by-products, including mAbs. The rate of glucose consumption per day was estimated for each cell to be 4.92 x 10-11 (grams of glucose/cell*day). The details of this calculation can be found in Appendix B: Calculations. According to the literature, experimental findings suggest the rate of consumption is approximately 4.65 x 10-11 (grams of glucose/cell*day) [24]. The experimental data supports our theoretical calculation and depicts the conservational balance of carbon from nutrient to metabolites. Optimizing the administration of

24

24

glucose to CHO cells in culture to maintain glucose requirements are is done once the parameters have been experimentally determined by sampling the CHO cell culture at arbitrary time points of feeds over the production period and measuring the availability of glucose in the culture, adjusting the concentration of glucose until an optimal glucose feed is found to correspond to optimized mAb production. The doubling period for each cell increases the total number of cells in culture at an exponential rate, and each cell consumes glucose within the bioreactor creating an extraordinary need for glucose. In order to approximate the total glucose consumed in the bioreactor, the total number of cells is multiplied by the glucose consumed by each cell and are integrated with respect to time over the total time the cells spend inside the bioreactor, as depicted below in Equation 2. Equation 2

Calculations for this equation are found in Appendix B: Calculations. The estimation of the total glucose consumed was found to be 2.56 x 104 g for each batch inside the production bioreactor, which requires a minimum glucose level of 2.0 g/L. Maintaining glucose levels at or above 2.0 g/L involves the addition of glucose to the bioreactor. Assuming initial presence of glucose in culture meets the 2.0 g/L requirement, the single addition of glucose to homogeneously maintain the minimum glucose level in the culture required for cell growth within the fermenter is equivalent to the quantity of glucose consumed, with the mass of glucose being 2.56 x 104 g. The delivery of glucose to the reactor can be facilitated either during the transfer of culture or step-wise as a continuous feed to maintain the minimum required glucose concentration. The molecular weight of glucose and the hours spent in the bioreactor can 25 25

be used to calculate each feed over the period of monoclonal antibody production resulting in a 1.08 g/L mAb titer, and was calculated to be 7.07 x 10-1 moles of glucose/hour for the residence time. Calculations regarding glucose consumption are depicted in Appendix B: Calculations. Since the vessel is not disposable, contamination between batches is possible; however, contaminants are eliminated with standard protocol involving sanitary connections in the event that a component is removed for maintenance, replacement, and sanitation – just as the 2,000 L reusable seed reactor does [21]. Reducing the risk of contamination by collecting in the vessel of its process piping, intensive cleaning and sterilization will be overseen after each batch. Medium Storage The remarkably small quantity of medium permitted for use in the first four reactors elicits manual preparation and delivery of the medium to these reactors prior to each batch formation. The quantity of medium required for the larger seed train vessels, including the 1,000 L XDRTM, the 4,000 L reusable seed reactor, and the 10,000 L fermenter, will be transported to the culture using aseptic connectors to medium storage tanks. Medium tanks will be made of stainless steel capable of storing medium. Harvesting ports can be attached to aseptic connections, which will be located on bottom of the reactors. Powdered media and water for injection will be added each tank, where its contents will be mixed with a pitched blade turbine accommodating both low and high flow axial mixing [33]. The agitator blades will be inclined at 45˚to ensure adequate mixing [36]. The agitator blades are important by eliminating any need for baffles by reducing the formation of eddies [33]. Mounting the covered impeller on the top of the tank’s interior along with the positioning of the harvesting port at the bottom of the tank is responsible for the ease of transferring medium out of 26 26

the tank to reduce waste. The specifications and calculations for the medium storage tanks are stated below in Table 7. Figure 17 shows a typical process of size exclusion permitted by the dead end filter for the passage of fluid through a semi-permeable membrane, trapping particles larger than the size of the pore. When the medium exits each storage tank, dead end filtration will be used to remove any accumulated undissolved media [40]. SteamThru disposable aseptic lines will then be used to allocate and direct media into each reactor [32]. Each storage tank will require one dead end filter, in which the media is stored for multiple reactors, the number of dead end filters is smaller than the number required, given each vessel had its own medium storage tank. Figure 18 depicts a scanning electron micrograph of the GibcoTM CD CHO AGTTM medium. The size of the medium granules can range from 20 µm to 300 µm, and are sufficiently prevented from transport by the dead end filter, exhibiting a 2 µm porous membrane. Material Transportation Pumps Large-scale manufacturing requires a significant volume of medium during the transfer of cell cultures through the seed train and into the production bioreactor. In order to accomplish efficacy in the transportation process, material transportation pumps can be of great utility. The contamination associated with liquid transport through pumps limits the quality of the pharmaceutical product; however, an ideal transportation pump should be characterized by minimal contact with fluids – eliminating any potential contact-dependent contamination. Not only may medium be transported from a medium storage tank, but it may also be transported from the production bioreactor for storage or purification purposes through sterilized tubing designed for this material transport process.

27

27

References [1] Feng Li, Natarajan Vijayasankaran, Amy (Yijuan) Shen, Robert Kiss, and Ashraf Amanullah. Cell culture processes for monoclonal antibody production. Landes Bioscience. September/October, 2010, 466-477. [2]Yiwen Tao, Jennifer Shih, Marty Sinacore, Thomas Ryll, and Helena Yusuf-Makagiansar. Development and Implementation of a Perfusion-Based High Cell Density Cell Banking Process. Cell Culture Development, Biogen Idec Inc. Published Online April 2011. [3] Fike, R. Nutrient Supplementation Strategies for Biopharmaceutical Production, Part 2: Feeding for Optimal Recombinant Protein Production. BioProcess Technical. December, 2009, 46-52. [4] CD CHO Medium. Form 3868 June, 2011. Gibco: Invitrogen Cell Culture. Life Technologies Corporation. [5] GlutaMAXTM vs. Glutamine. (2012) Life Technologies Corporation. Retrieved 4/12/2013 from: http://www.invitrogen.com/site/us/en/home/Products-and-Services/ Applications/Cell-Culture/Mammalian-Cell-Culture/media-supplements/GlutaMAXMedia/glutamax-vs-glutamine.html [6] GlutaMAX™ Media: Healthier Cells Live Longer. (2011) Gibco: Invitrogen Cell Culture. Life Technologies Corporation. [7] Ruaan RC, Tsai GJ, Tsao GT. Surface area and anchorage-dependent growth of Chinese hamster ovary cells. Biotechnol Prog. 1993. Jul-Aug; 9 (4): 362-365. [8] Cytopore from GE Healthcare, Life Sciences. 2012. Biocompare: A Buyers Guide for Life Scientists. Retrieved 4/8/13 from: http://www.biocompare.com/20095Microcarriers/96734-Cytopore-1/

28

28

[9] Senger RS, Karim MN. Effect of shear stress on intrinsic CHO culture state and glycosylation of recombinant tissue-type plasminogen activator protein. Biotechnol Prog. 2003. Jul-Aug; 19 (4):1199-1209. [10] Cytopore 2: Dry Powder. GE Healthcare, Life Sciences. Retrieved 4/7/13 from http://www.gelifesciences.com/webapp/wcs/stores/servlet/catalog/en/GELifeSciencesus/products/AlternativeProductStructure_16976/17127101 [11] Cytopore™: Microcarriers for Cell Culture Instructions. (2002) Amersham Biosciences. [12] Material Safety Data Sheet: PBS Buffer, 10X. (2011) GE Healthcare. [13] Lee, J. M. (2009) Biochemical Engineering. Pullman, WA: Washington State University. [14] Wilcox, W. R. (2011) Design of a Plant to Manufacture Monoclonal Antibodies. Potsdam, NY: Clarkson University [15] Guidelines for Maintaining Cultured Cells.2012. Life Technologies Corporation. Retrieved 4/12/13 from http://www.invitrogen.com/site/us/en/home/References/gibco-cell-culturebasics/cell-culture-protocols/maintaining-cultured-cells.html [16] AB ReadyToProcess™ WaveTM Bioreactor™ 2/10 and 20/50 systems. (March 2011) GE Healthcare, Life Sciences. [17] Pierce, L. N., Shabram, P. W. Scalability of a Disposable Bioreactor from 25L – 500L Run in Perfusion Mode with a CHO-Based Cell Line: A Tech Review. BioProcessing Journal. 2004. July/August: 51-56. [18] Application Brief: Scale-Up to 1000 L Perfusion Cell Culture in a Disposable Stirred-Tank Bioreactor. (April 7, 2008). Xcellerex.

29

29

[19] Supplement: Guide to Disposables, A Next Step in Implementing Disposables: Transfer Lines. November 2007. Printed in the U.S.A. BioPharm: The Science & Business of Biopharmaceuticals. [20] Stainless Steel - Grade 316 - Properties, Fabrication and Applications. Azom.com: The A to Z of Materials. Retrieved 4/14/13 from http://www.azom.com/article.aspx?ArticleID =863 [21] Arcuri, L. Clean Steam Sterilization for Bioprocess Equipment.2012. Bioresearch Online. Retrieved 4/12/13 from http://www.bioresearchonline.com/doc.mvc/Clean-SteamSterilization-For-Bioprocess-0002 [22] Avastin. (2012). Genetech, Inc. Retrieved 4/8/13 from: http://www.gene.com/gene/ products/information/oncology/avastin/ [23] Bevacizumab. (2012)> Wikipedia. Retrieved 4/8/13 from: http://en.wikipedia.org /wiki/Bevacizumab [24] Ahn, W. S; Antoniewicz, M. R. Metabolic flux analysis of CHO cells at growth and nongrowth phases using isotopic tracers and mass spectrometry. Metabolic Engineering. 13 (2011): 598–609. 47 [25] Advanced Granulation TechnologyTM (AGTTM) Media for Bioproduction. 2002 Invitrogen Corportation. Retrieved 4/8/13 from: http://tools.invitrogen.com/content/sfs/brochures/ 332_021646_AGT_Bro.pdf [26] Alkan, S.S. Monoclonal antibodies: the Story of a Discovery that Revolutionized Science and Medicine. Nature Reviews Immunology. 4 (2004): 153-156. [27] Milstein, C.; Kohler, G. Continuous cultures of fused cells secreting antibodies of predefined specificity. Letters to Nature. 256 (1975): 496-497.

30

30

Appendix A: Project Guidelines MEDIUM PREPARATION “You will need to determine whether you will use a proprietary medium formulation that is specific to your company or whether you will use an off-the-shelf medium provided by a company such as Gibco or Lonza. Your facility is considered to be an animal-free facility, so you will be using a Chemically-Defined Medium in your process. The large quantities of medium that will be needed for your process will be made onsite from powdered components. Please design the Medium Preparation Area and select whether to use steam-in-place vessels, disposable vessels, or both. SEED TRAIN You will obtain one vial of CHO (Chinese Hamster Ovary) cells for each batch of product that will be produced in your facility. You will be using CHO cells because the current platform for making biologicals is often CHO in the mammalian realm. The first MAbs were made using hybridomas, but the yields tended to be low and for this and other unnamed reasons your company decided many years ago to proceed with a CHO platform. The vial size can range from 1 mL and up based on the company’s method of banking their master cell banks. (Genentech recently presented work at the American Chemical Society where they used bags instead of vials that contained 125 mL of culture in them.) In order to account for both old and new cell line banking procedures, set the vial volume size to be 1 mL. Each vial should contain 1x106 viable cells/mL. The number of cells will need to be increased by passing into larger and larger volumes. (How will you do this?) After the final scale-up in the seed train, the culture will be placed in the production reactor(s) (fermenter). Note that the seed train will be a batch process, i.e., there will be no addition of feeds, including glucose. The doubling time for the CHO cells that you will be using is 36 hours. Note that doubling time varies based on the cell line, so, for example, one may obtain another cell line and it may double in 24 hours or perhaps 48 hours, while ours is 36 hours for this design problem. PRODUCTION REACTORS (FERMENTERS) Once your cells/cultures leave the seed train they will enter the production reactors. Currently there are a number of types of reactors used to culture cells. A very common reactor used in industry is a stirred-tank reactor. You may decide to use a stirred-tank reactor or another type of reactor such as a perfusion reactor, a packed-bed reactor, or an air-lift reactor. Please note that you will need to design your production reactors to yield a titer of 1 to 2 g/L and, in the future, 5 to 10 g/L. You will need to base your production bioreactors on a worst-case scenario of 1 to 2 g/L of product, but if titers increase as expected, your facility should be able to account for this. (Hint: Use multiple smaller production bioreactors instead of one very large production bioreactor.) For design purposes, you may assume that your cell line produces your product at a rate of 25 pg/(cell*day). For design purposes, you must be able to produce a minimum of 1,000

31

31

kg of product a year, which is approximately the estimated quantity produced for both Enbrel™ and Remicade™ in 2007 (Jagschies, 2009). Your new production facility will need to have the capability to run both batch and fed batch processes, but you should design a process in which the cultures are fed. You will need to feed at a minimum glucose rate to your production bioreactors. The glucose level in your reactors should not go below 2 g/L.” [14]

32

32

Appendix B: Calculations Exponential Cell Growth

The equation for exponential cell growth is stated above. The initial number of cells is N0 = 4.5 X 108 (viable cells). The number of cells as a function of time, N(t), can be determined after the growth constant is calculated to be k = 1.9254 X 10-2 (hours-1) by using the doubling time of 36 (hours) as shown below:

Now, we can write our equation for the number of cells as a function of time which is N(t) = 4.5 X 108 e.019254 t with the time variable in hours. Production of mAb per Batch
In order to find the amount of product per batch, the production rate must be multiplied by the number of cells and integrated with respect to time over the amount of time per batch, t. The production rate for our cell line is P = 25 (pg/cell * day) [14].

33

33

Carbon Balance for Glucose Consumption There are three carbon containing products that are considered in the carbon balance. These are carbon from the MAbs, the lactic acid, and the carbon dioxide produced with glucose consumption. The molecular weight of glucose is 180.16 g/mol and there are 6 carbon atoms for each glucose molecule. The molecular weight of the MAb product is 149 kD and contains 6.64 X 103 moles of carbon for each mole of MAb [23]. This is based off of the assumption that the MAb product has the same properties as AvistinTM. Carbon from MAb

Carbon from Lactic Acid
The two methods of lactic acid formation from glucose are homolactic fermentation, , and heterolactic fermentation, . By averaging the ratio of the stoichiometric coefficients, the amount of lactate produced per glucose consumed is approximately 1.5 mol/mol [24]. The amount of carbon from the production of lactic acid is calculated as follows:

34

34

Carbon from Carbon Dioxide Carbon dioxide can be produced as a product of heterolactic fermentation or through the oxidation of glucose, . By averaging the ratio of the stoichiometric coefficients of both reactions, the approximate amount of carbon dioxide produced per glucose consumed is 3.5 mol/mol.

Glucose consumption can be estimated as follows

Total Glucose Consumption The total amount of glucose consumed in the bioreactor is estimated as the number of cells multiplied by the glucose consumption rate and integrating with respect to time over the days the cells spend in the bioreactor. For this calculation tin is 485 hr and tout is 686 hr.

35

35

The glucose will be added in a step wise function to the fermenter in order to maintain the desired glucose concentration of 2 g/L . Since the culture will remain in the bioreactor for 201 hours, the hourly addition of feed to the fermenter is calculated below.

Number of Reactor Lines In order to meet the production requirement of 1000 kg of product a year, the number of reactor lines needs to be determined. Assuming the final titer will remain at 1.1 g/L, that the number of operating days is about 328 days which takes into account down time for cleaning, that the longest amount of time spent in a single bioreactor is 8.4 days, and assuming an efficiency of 80 percent, the number of reactor lines needed is calculated below.

36

36

Appendix C: MATLAB Code

37

37