Castilho Leda R., Moraes Angela Maria (Ed.) Animal Cell Technology: From Biopharmaceuticals to Gene Therapy

Подождите немного. Документ загружается.

resistance of which varies according to the temperature of its platinum

sensor (Pt100), being typically 100  at 08C and 123  at 608C.

The most common temperature control systems for cell culture are the

following:

(i) Heat blanket around the bioreactor, the electric resistance of which

can be turned on or off as required to heat the system. In this case, the

minimum culture temperature is limited by the room temperature,

which can be a problem in cultures of insect cells (with optima around

288C) in facilities without air conditioning.

(ii) Water-jacket reactor with a temperature sensor, which sends a signal

to the controller that triggers an electrical heating element within the

water of the jacket (to increase the temperature) or open a solenoid

valve that allows the inlet of colder water to the circuit (to decrease

the temperature). In this case, the minimum temperature of culture is

limited by the temperature of the water supply. If the water tempera-

ture supply is not sufﬁciently low, a thermostatically refrigerated bath

system can be used.

(iii) Water-jacket reactor with two temperature sensors, for the jacket and

for the culture medium. This is different from the control system in

(ii) in that the control is made in a cascade, where the medium

temperature controller (master) acts to send a set-point value to the

jacket water temperature controller (slave). See Section 10.5.1(iii).

Typically, the process is set to control temperature to a value that results

in high cell growth or production rate. Sometimes, however, a different

temperature may be selected, which could result in a better quality

product or a respiration rate compatible with the oxygen transfer system

(Chuppa et al., 1996). For some processes, a temperature shift batch

culture may be desirable in which there are two controlled temperature

values, one for the growth phase and another for the production phase, in

which the cell division is interrupted due to the low temperature (Fox et

al., 2004).

10.4.2 pH

pH control systems in cell culture can be classiﬁed into four basic types, as

follows.

(i) Culture medium buffered naturally by the dissolved solutes. This

may allow some pH variation during the culture and in which usually

pH electrodes are not used.

(ii) Culture medium buffered with sodium bicarbonate and carbon diox-

ide gas. In this case, the ﬂask is maintained in an incubator with

carbon dioxide gas control, and the concentration of H

. The pH will

remain relatively constant in an acceptable range, due to the equili-

brium:

þ H

O $ H

$ H

þ HCO



262 Animal Cell Technology

(iii) Controlled addition of CO

in the gas sparged into the bioreactor

(see Section 10.4.5), based on the signal of a pH electrode inserted in

the culture medium.

(iv) Controlled addition of an alkali solution (such as sodium hydroxide)

and/or acid (such as hydrochloric acid) to the bioreactor vessel, based

on the signal of a pH electrode.

The most widely used pH electrode in cell culture is the silver/silver

chloride electrochemical type with internal gel (that can be sterilized when

the whole bioreactor is autoclaved), or the combined type (which can be

pressurized during sterilization, to compensate the medium pressurization

and to avoid contamination of the sensor with culture medium) (Schu

gerl,

1991).

In long duration cultures (several weeks), a retractable housing for pH

and dissolved oxygen electrodes can be used. With the use of an internal

chamber, the electrode can be removed from contact with the culture

medium and then recalibrated, repaired, or replaced. After re-installation

in the chamber, it is sterilized with direct steam and afterwards re-enters

in contact with the culture medium, without risk of contaminations

(Sonnleitner, 2000).

There has also been an increase in the ability to measure the intracellular

pH, by ﬂow cytometry with pH-dependent ﬂuorescence indicators (Oz-

kan and Mutharasan, 2002) and nuclear magnetic resonance. There are

several reports that relate the intracellular pH value to cell behavior,

including metabolism, apoptosis, and speciﬁc growth rate (Cherlet and

Marc, 1998). Such ofﬂine measurements help to interpret the physiological

state of the cells, but still cannot be used as control systems.

10.4.3 Dissolved oxygen

Probably the most important state variable for the process control is the

dissolved oxygen concentration (DO) in contact with the cells. Because of

the very low solubility of oxygen in water and in culture medium, around

6–8 mg/L at cell culture temperature, it is clear that DO is a nutrient that

needs to be supplied continuously to the culture medium, and that will be

consumed simultaneously with the carbon source in aerobic cultures

(Henzler and Kauling, 1993; Moore et al., 1995).

There are several systems of oxygen transfer, but some are speciﬁcally

used in cell culture in agitated tanks: diffusion through liquid surface,

direct sparging, microsparging, and diffusion through membranes.

Although the speciﬁc respiration rate of animal cells is considerably lower

than with microbial cells, it is important to supply oxygen with minimal

potential shear stress to the cells caused by agitation and by the rupture of

the gas bubbles.

Nowadays, the most commonly used DO electrode is the polarographic

that employs an electrochemical method, through the reduction of mole-

cular oxygen at the platinum electrode (cathode) generating electrons for

the silver oxidation at the silver/silver chloride electrode (anode). In this

type of sensor, it is necessary to impose an external voltage and the

resulting current is measured and converted into dissolved oxygen concen-

Monitoring and control of cell cultures 263

tration (Figure 10.2). Another common type is the galvanic sensor, in

which the oxidation reaction at the anode is capable of generating a voltage

spontaneously (Pb! Pb

þ2

+2e

–

) (Johnson et al., 2000).

More recently, other electrodes have been developed, the most notable

among them are those that use the ﬂuorescence quenching by oxygen on

ruthenium compounds (O’Neal et al., 2004).

Dissolved oxygen control strategies are associated with the oxygen

transfer term in the oxygen mass balance (Equation 1):

¼ k

a:(C

 C)  Q

:X (1)

where C is dissolved oxygen concentration in culture medium, t is time,

a is volumetric coefﬁcient of oxygen transfer, C

is oxygen concentra-

tion in balance with the gas (saturation), Q

is speciﬁc respiration rate

and X is cell concentration (a good explanation of that equation can be

obtained from Bailey and Ollis, 1986).

Typically, oxygen transfer can be modiﬁed acting through the rate of

impeller agitation or through the aeration or micro-aeration ﬂow rate,

which change the k

a term of Equation 1. The dissolved oxygen controller

VP connector

straight

Thread sleeve

Pg 13.5

529-100-64

Washer

O-ring

Æ 12 mm

120

Anode

Cathode

Membrane body

Cop sleeve

(Type N)

Protection cap

Figure 10.2

Dissolved oxygen polarographic electrode (InPro

6050, Mettler-Toledo, 200 6).

The cathode, where the half-reaction with O

+2H

O+4e

–

! 4OH

–

)takes

place, is in contact with a membrane, that allows the transport of the dissolved

oxygen from the external medium. At the anode, the silver oxidation

(Ag + Cl

–

! AgCl + e

–

) takes place.

264 Animal Cell Technology

can calculate new values for a set-point based on the agitation motor speed

or the ﬂow rate controller. A disadvantage of this control method is that it

interferes with the shear stress imposed by higher values of agitation or

aeration, which can be particularly critical in animal cell cultures.

To avoid the above problem a system involving the control of a mixture

of two gases can be used, for instance, oxygen and nitrogen (Figure 10.3).

The dissolved oxygen concentration value is controlled by acting on the

two gas ﬂow rates ratio, which varies the C

value and the oxygen transfer.

In this way, it is possible to keep the agitation and the total gas ﬂow rate

admitted into the reactor constant and, consequently, the k

a and the

stress induced to the cell are also constant.

Alternatively, a membrane of material allowing oxygen diffusion can be

used (usually silicone tubing, a microporous material, or polypropylene),

avoiding bubble formation. If aeration is accomplished by a bubble-free

system (see Chapter 9), using membranes or tubings made of materials

that allow oxygen diffusion (e.g. silicone tubing or microporous mem-

branes made of polypropylene), bubble formation is avoided. In this

situation, the transfer is not regulated by gas ﬂow or agitation, but instead

by the partial pressure of oxygen, which is related to the internal pressure

in the membrane and its gas composition and will determine the C

value

(Qi et al., 2003).

An additional advantage of this DO control system is the possibility of

having a real-time estimation of the cells’ respiration rate. As can be

observed in Equation 1, if C is constant, the term of the transfer k

a.(C

C) is equivalent to the consumption Q

.X. With the actuation value of

the controller, the transfer rate can be calculated, and also the consumption

at each moment (Kamen et al., 1996). There are several reports in the

literature dealing with the use of respiration measurements to estimate cell

concentration online, based on the assumption that, during the growth

phase, each cell consumes a constant amount of oxygen (Yoon and

Konstantinov, 1994; Rufﬁeux et al., 1998; Jorjani and Ozturk, 1999).

10.4.4 Cell concentration

In every bioprocess, monitoring biomass concentration is fundamental,

because usually the growth is required to be maximized. Substrate con-

O flow rate

controller

N flow rate

controller

C (%)

QQQ

T12

⫽⫹

Figure 10.3

Dissolved oxygen contr ol by the adjustment of the oxygen fraction in the sparged

gas. Its total ﬂow rate is kept constant and corresponds to the sum of the two

controlled gases Q1 and Q2.

Monitoring and control of cell cultures 265

sumption or product synthesis rates can be analyzed but it is important

that variations in these parameters can be related to the cell concentration.

For this reason speciﬁc rates are often used, such as the speciﬁc rate of

product formation:



(2)

where 

is the speciﬁc rate of product formation, t is time, X is cell

concentration, and P is product concentration.

For microbial cultures it is common to express cell concentration as dry

cell weight by volume, in animal cell cultures the number of cells per unit

volume is more commonly used (for instance, 1.3 3 10

cells/ml). For the

determination of cell number a microscope graduated slide can be used

such as a hemocytometer or Neubauer chamber. In these, the thickness

between the slide and the coverslip is ﬁxed and known (typically 0.1 mm)

and graduation marks on the slide allow the number of cells to be counted

in a certain area. The volume can be determined by multiplying this area

by the known thickness. With the aid of a dye (such as trypan blue), it is

possible to estimate cell viability. This is expressed as the percentage of

viable cells determined as those with an intact membrane capable of

excluding the dye from inside the cell and so remaining colorless (see

Chapter 2 and Freshney, 2005).

Although counting by hemocytometer is the most common method for

the determination of cell concentration, it does not allow any action to be

taken during the culture process, because it is typically an ofﬂine measure-

ment. When a real-time estimation is needed, it is common to use an

indirect measurement, such as the optical density, or turbidity. It is

possible to correlate the absorbance measured by a sensor introduced into

the bioreactor with the cell concentration (Konstantinov et al., 1994; Wu

et al., 1995; Olsson and Nielsen, 1997). Several sensors are commercially

available for this function, notably those that measure reﬂected light

(Aquasant AS81/44 and Metler FSC402) and transmitted light (ASR Laser

Turbidimeter). Optical sensors can also be used for monitoring other

variables, even in small bioreactors (Kostov et al., 2001).

Other forms of online estimation of the cell concentration are:

(i) A microscope directly inserted into the culture medium, using image

analysis software to evaluate cell concentration and size distribution

for each acquired image (Joeris et al., 2002; Gueza et al., 2004).

(ii) Capacitance measurement of the culture, that is correlated with the

concentration of viable cells. Each cell with an intact membrane

behaves as a small electric capacitor (Dowd et al., 2003).

(iii) Fluorescence measurement of the medium. This is based on the

assumption of a constant content of a ﬂuorophor such as NAD(P)H

per cell.

As well as the estimation of the cell concentration, other parameters

related to the biomass can be obtained. However, at present these are

limited to ofﬂine measurements. Among these are:

266 Animal Cell Technology

(i) Distribution of the size of the cells, based on the Coulter Counter

(Beckman Coulter, USA) or similar equipment like Casy TTC

(Innovatis, Germany), in which a dilute cell suspension ﬂows through

a capillary tube containing two electrodes. Each cell is detected by a

disturbance in the electric current, the intensity of which is measured.

In this way, not only are the cells counted, but size distribution is

estimated, based on a precalibration relating the cell diameter to the

disturbance of electric intensity (Seewo

ster and Lehmann, 1997).

Digital image analysis based on equipment such as Cedex (Innovatis)

also provides ofﬂine cell size distribution estimation.

(ii) Flow cytometry is a method through which it is possible to count

cells presenting certain characteristics (Davey and Kell, 1996; Brezins-

ky et al., 2003). FACS (ﬂuorescence activated cell sorter) is a variant

of ﬂow cytometry, in which besides counting each cell with certain

characteristics (ﬂuorescent), it is possible to separate them into

subpopulations with similar characteristics (Figure 10.4; see color

section) (Herzenberg et al., 2000; Park et al., 2006).

10.4.5 Other variables of interest

Besides the variables previously mentioned, the following deserve to be

discussed.

(i) Partial pressure of dissolved carbon dioxide gas (pCO

). Carbon

dioxide is a compound present in the culture medium that can inhibit

cell growth. In small-scale cultures (Schott or T-ﬂasks), it is common

to keep the ﬂask in an incubator with a controlled CO

concentra-

tion. Such control is made through a sensor (infrared spectrophoto-

metry or thermal conductivity) and a valve that releases carbon

dioxide gas into the incubator. Through the presence of bicarbonate

in the culture medium, it is possible to maintain a reasonably constant

culture pH (see Section 10.4.2). In bioreactors, pCO

can be con-

trolled through the addition of CO

to the sparged gas mixture. Cell

growth inhibition will occur at very high or low pCO

. Therefore, it

is desirable to maintain its value between 40 and 200 mmHg, depend-

ing on the cell line. There are commercially available sensors designed

for pCO

, which can be sterilized and inserted into any bioreactor

port, such as the InPro

5000, with the principle of operation similar

to a pH electrode (Mettler-Toledo, 2006). In the literature there are

reports of sensors based on other technologies (Pattison et al., 2000).

(ii) Metabolites can be measured in real time, through techniques such as

FIA (ﬂow injection analysis) that allow online measurement of glu-

cose and other substrates or products. In this system, a small ﬂow of

culture medium is continuously pumped to a unit composed of a

tangential ﬁlter (for cell separation), injection pump, and detector

(Reinecke and Stephanopoulos, 2000). Another possibility is to con-

nect ofﬂine equipment to automated sampling systems (Figure 10.5),

to provide information on the metabolite concentration for a control

system, for instance, for controlling the ﬂow of fresh medium in a

fed-batch culture (Konstantinov et al., 1996; Ozturk et al., 1997). It is

Monitoring and control of cell cultures 267

also possible to estimate cell concentration, including those immobi-

lized on microcarriers based on measurements of real-time substrate

consumption (Ducommun et al., 2001).

10.5 Strategies of control

10.5.1 Traditional control

There are several strategies to control a process variable. Among them, the

most important are the following.

Bioreactor

Sanitizing agent

Online

sample

device

YSI

Glucose/lactate

analyzer

Process control

computer

Medium

Harvest

Weight

Figure 10.5

Feeding ﬂow rate control system based on glucose concentration measur ed in real time (adapted

from Ozturk et al., 1997).

268 Animal Cell Technology

(i) On-off controller, in which the action can only assume two states (on

or off). This controller is used, for instance, in incubators to keep

temperature in a small range. When the temperature sensor indicates a

value lower than set up, the system turns an electric resistance on,

heating up its interior. In this situation, it is not possible to take an

action larger or smaller, that means, it is warming up or it is not.

(ii) Controller modulated by pulse width (pulse width modulation or

PWM). In this control type, the action is also on or off but the time

that the actuator stays on within a certain cycle can be adjusted

continuously, allowing a ﬁnal operation of different intensities. Fre-

quently the pH control in bioreactors uses this strategy. The greater

the difference between the pH set-point and the measured value, the

larger would be the fraction of the cycle (usually a few seconds) in

which the base pump is on, creating prolonged action within the

cycle. This strategy prevents an exaggerated control action that would

cause substantial deviation from the set-point value (called over-

shoot). Some dissolved oxygen control systems operate this way,

through the controlled mixing of different gases by the opening and

closing of solenoid valves.

(iii) Cascade controller, composed of one master and one slave loop. This

is used when a more rigid control of a process variable is required, for

instance, the temperature of the culture medium. Here the action of

the master controller (primary loop) to keep bioreactor temperature

controlled is simply to calculate the adequate value for the slave

controller set-point, which takes action in the process (Figure 10.6).

(iv) PID or proportional integral derivative controller. This is the most

common system used in industry and the one that best keeps control

of the process variable. It is based on the principle that the action is

taken not only on how large is the error (difference between desired

and measured values), but also on the sum of past errors (integral of

Controller

(slave)

Actuator

electric

heater

Controller

(master)

Process

Sensor

(slave)

jacket

temperature

Sensor

(master)

culture

temperature

Set-point

process

Slave (secondary) loop

Set-point

jacket

Figure 10.6

Cascade control. The actuator of the slave controller (electric heater) maintains the temperature of

the jacket water controlled in a value set by the master controller.

Monitoring and control of cell cultures 269

the error) and to the rate that the error is changing (derivative of the

error). Mathematically it can be expressed as in Equation 3:

actuation ¼ P:error þ I:

error:dt þ D:

d error

(3)

where actuation is the controller output (the value that goes to the

actuator, according to Figure 10.1), error is the difference between the

desired value (set-point) and the one measured by the sensor, t is time and

P, I, and D are constants that need to be adjusted for each system. The

adjustment of the constants for a process, that is, the algorithm to

determine the numeric values that will make the controller operate better,

is called PID controller tuning. The most common procedure for this is

called Ziegler-Nichols. For further details to tune a controller for a given

process, a textbook of process control is recommended (Seborg et al.,

2003).

10.5.2 Advanced control

For chemical processes (without cells) several advanced control strategies

have been developed, some of which can be applied to bioprocesses and,

especially, in animal cell cultures. Two of them are described below.

Adaptive control. When there are big changes during a process, a

certain set of values of constants like P, I, and D, may not be adequate for

the whole culture. The controller may be tuned, for instance, to the

beginning of the culture, but not to the end. In this situation, the

controlled variable may start to oscillate signiﬁcantly, which can interfere

with the process. For those cases, algorithms have been developed to tune

these constants during the process. Such a procedure is called adaptive

control (Diaz et al., 1996).

Control based on neural network. Similar to fuzzy logic modeling,

neural network analysis uses a series of previous data to execute simula-

tions of the process, with a high degree of success, without however using

formal mathematical models (Chen and Rollins, 2000). To this goal, it is

necessary to deﬁne inputs, outputs, and how many layers of ‘‘neurons’’

will be used, which depends on the number of variables and the available

data.

References

Bailey JE, Ollis DF (1986), Biochemical engineering fundamentals, 2nd Ed., McGraw-

Hill, Singapore.

Brezinsky SCG, Chiang GG, Szilvasi A, Mohan S, Shapiro RI, MacLean A, Sisk W,

Thill G (2003), A simple method for enriching populations of transfected CHO

cells for cells of higher speciﬁc productivity, J. Immunol. Methods 277:141–155.

Chen VCP, Rollins DK (2000), Issues regarding artiﬁcial neural network modeling for

reactors and fermenters, Bioprocess. Eng. 22:85–93.

Cherlet M, Marc A (1998), Intracellular pH monitoring as a tool for the study of

hybridoma cell behavior in batch and continuous bioreactor cultures, Biotechnol.

Prog. 14:626–638.

Chuppa S, Tsai YS, Yoon S, Shackleford S, Rozales C, Bhat R, Tsay G, Matanguihan

270 Animal Cell Technology

C, Konstantinov K, Naveh D (1996), Fermentor temperature as a tool for control

of high-density perfusion cultures of mammalian cells, Biotechnol. Bioeng.

55:328–338.

Davey HM, Kell DB (1996), Flow cytometry and cell sorting of heterogeneous

microbial populations: the importance of single-cell analyses, Microbiol. Rev.

60:641–696.

Diaz C, Dieu P, Feuillerat C, Lelong P, Salome M (1996), Simultaneous adaptive

predictive control of the partial pressures of dissolved oxygen (pO2) and dissolved

carbon dioxide (pCO2) in a laboratory-scale bioreactor, J. Biotechnol. 52:

135–150.

Dowd JE, Jubb A, Kwok KE, Piret JM (2003), Optimization and control of perfusion

cultures using a viable cell probe and cell speciﬁc perfusion rates, Cytotechnology

42:35–45.

Ducommun P, Bolzonella I, Rhiel M, Pugeaud P, von Stockar U, Marison IW (2001),

On-line determination of animal cell concentration, Biotechnol. Bioeng. 72:

515–522.

Fox SR, Patel UA, Yap MG, Wang DI (2004), Maximizing interferon-gamma produc-

tion by Chinese hamster ovary cells through temperature shift optimization:

experimental and modeling, Biotechnol. Bioeng. 85:177–184.

Freshney RI (2005), Culture of Animal Cells: A Manual of Basic Technique, 5th Ed.,

Wiley-Liss, New York.

Gueza JS, Cassar JP, Wartelle F, Dhulster P, Suhr H (2004), Real time in situ

microscopy for animal cell-concentration monitoring during high density culture

in bioreactor, J. Biotechnol. 111:335–343.

Henzler HJ, Kauling DJ (1993), Oxygenation of cell cultures, Bioprocess Eng. 9:

61–75.

Herzenberg LA, De Rosa SC, Herzenberg LA (2000), Monoclonal antibodies and the

FACS: complementary tools for immunobiology and medicine, Immunol. Today

21:383–390.

Joeris K, Frerichs J-G, Konstantinov K, Scheper T (2002), In-situ microscopy: online

process monitoring of mammalian cell cultures, Cytotechnology 38:129–134.

Johnson MJ, Borkowski J, Engblom C (2000), Steam sterilizable probes for dissolved

oxygen measurement, Biotechnol. Bioeng. 67:645–656.

Jorjani P, Ozturk SS (1999), Effects of cell density and temperature on oxygen

consumption rates for different mammalian cell lines, Biotechnol. Bioeng. 64:

349–356.

Kamen AA, Be´dard C, Tom R, Perret S, Jardin B (1996), On-line monitoring of

respiration in recombinant-baculovirus infected and uninfected insect cell bior-

eactor cultures, Biotechnol. Bioeng. 50:36–48.

Kilburn DG (1991), Monitoring and control of bioreactors, In: Butler M (Ed.)

Mammalian Cell Biotechnology. A Practical Approach, Oxford University Press,

New York, pp. 159–185.

Konstantinov K, Chuppa S, Sajan E, Tsai Y, Yoon S, Golini F (1994), Real-time

biomass-concentration monitoring in animal-cell cultures, Trends Biotechnol.

12:324–333.

Konstantinov KB, Tsai Y, Moles D, Matanguihan R (1996), Control of long-term

perfusion Chinese Hamster Ovary cell culture by glucose auxostat, Biotechnol.

Prog. 12:100–109.

Kostov Y, Harms P, Randers-Eichhorn L, Rao G (2001), Low-cost microbioreactor

for high-throughput bioprocessing, Biotechnol. Bioeng. 72:346–352.

Mettler-Toledo (2006), Available at: http://www.mt.com (accessed 20 December

2006).

Moore A, Donahue CJ, Hooley J, Stocks DL, Bauer KD, Mather JP (1995), Apoptosis

in CHO cell batch cultures: examination by ﬂow cytometry, Cytotechnology

Monitoring and control of cell cultures 271