rp1017en00

8/4/2019 rp1017en00

http://slidepdf.com/reader/full/rp1017en00 1/81www.bioprocessingjournal.com • September/October 2005

INVITED REVIEW

In today’s biopharmaceuticalpipeline, monoclonal antibod-ies are a predominant modal-

ity for a broad range of clinicalindications, including inflamma-

tory disorders, oncology, and infectiousdiseases.1 More than two dozen anti-body-based products are commercially available. In 2004, six of the 12 newbiopharmaceutics that gained approvalin the United States and Europe wereantibody-based products.2

Most antibody therapies require highdoses over a long period of time, which

requires large amounts of purified prod-

uct per patient. Therefore, manufactur-ing capacity to meet the demands of antibody production is a real challenge.It is desirable to have highly productiveand consistent manufacturing processes.In addition, speed to market is criticalto deliver health benefits to patientsquickly and to achieve business success.

For early-stage clinical studies (Phase Iand II), an initial standardized platformprocess is usually applied to satisfy thematerial demand and quality require-ments within a short period of time,

despite the relatively low productivity and robustness.3,4 Once the productcandidate is proven safe for patients in

an early-phase clinical study, the processis further optimized in order to maxi-

mize product yield and process robust-

ness, and reduce the cost of goods forcommercial manufacturing.Two major areas of commercial anti-

body process development are describedin this review article: upstream cell cul-ture and downstream purification pro-cesses. Upstream process developmentincludes cell line development, media

optimization, and cell culture processoptimization. Downstream processdevelopment comprises cell harvest,antibody capture, viral inactivation, andpolishing steps. This article provides an

updated review and discussion of thetechnologies used in recent therapeuticantibody production processes.

Cell Culture

Recombinant mammalian cell cul-

ture is the most popular expression sys-tem to produce monoclonal antibodies,and Chinese hamster ovary (CHO) andmurine myeloma (NS0) are the mostwidely used cell lines.5 A typical cell cul-ture process starts with thawing a frozenvial of a working cell bank (WCB), fol-lowed by expanding the cell population

through a series of seed trains in differ-ent culture vessels. The culture is thentransferred to a production bioreactorwhere the cells continue to grow andthe expressed product accumulates in

the culture broth. In order to achieve ahigh product titer, high cell mass and cellviability need to be maintained in the

Current Therapeutic Antibody Production

and Process Optimization

Feng Li, Ph.D.* ([email protected]) is a scientist, and Brian Lee, Ph.D., is a principal scientist in the process engineering depart-

ment; Joe X. Zhou, Ph.D.* ([email protected]), and Tim Tressel, Ph.D., are principal scientists in the purification process devel-opment department; and Xiaoming Yang, Ph.D., is a principal scientist in the cell science and technology department, Amgen,Thousand Oaks, CA.

*These two authors contributed equally to this article.

By FENG LI, JOE X. ZHOU,

XIAOMING YANG, TIM

TRESSEL, and BRIAN LEE

Figure 1. A typical cell culture process comprises vial thaw, seed expansion, and production

stages.

8/4/2019 rp1017en00

http://slidepdf.com/reader/full/rp1017en00 2/82 BioProcessing Journal • September/October 2005

production bioreactor. Two of the mostpopular process modes are fed-batch(feeding concentrated nutrient solutionsto a batch production bioreactor) and

perfusion (maintaining and recyclingthe cells to a bioreactor while continu-ously replacing spent media with a freshsupply). The fed-batch process is oftenused due to its scalability, ease of opera-tion, and high volumetric productivity.An illustration of the unit operations of a typical upstream process is shown inFigure 1.

Although stainless steel bioreactorsare still the major choice for large-scaleproduction, disposable bioreactor sys-tems have become available. The WaveBioreactor® system, which uses a plasticdisposable bag, is commonly used dur-ing seed culture expansion. Disposable

bioreactor systems can benefit the man-ufacturing process by eliminating the

clean-in-place (CIP) and steam-in-place(SIP) operations and by reducing theexpensive capital investment for stainlesssteel bioreactors. The Wave Bioreactorsystem at the 500-liter scale has achievedhigh cell densities with CHO cells, whichshows its potential application as a finalproduction vessel, especially for rapidmaterial supply during early stage clini-

cal studies.6

Cell culture process optimization isan integrated activity involving cell lineselection, medium development, andoptimization of bioreactor conditions.

High titers up to approximately 5 g/Land cell densities of more than 20 mil-lion cells/ml have recently been reportedin fed-batch cultures.7,8 The specific

productivity of over 20 pg/cell/day canbe routinely achieved for productioncell lines.9 Enhancement of specificproductivity per cell is accomplishednot only by selecting highly produc-tive cell lines, but also by optimizingmedium compositions and bioreactoroperating conditions.

Clone Selection

Cell line development and mastercell bank (MCB) generation compriseone of the most critical steps of processdevelopment. After transfection withplasmids bearing the antibody light- and

heavy-chain genes, cells are screenedfor highly productive cell lines through

growth recovery, single cell cloning,serum-free and suspension adaptation,amplification, and final clone selection.Screening and selecting a highly produc-tive and stable clone from the transfec-tant population in a limited time frameis a major challenge. Most commonly,the transfected cells are diluted and cul-tivated in 96-well plates with a basal

growth medium and screened to identify

those with fast growth. The productiv-ity among the fastest growing cell linesis then analyzed for product titer. Thecandidates with the best productivity

and growth characteristics are adaptedto suspension culture using serum-freemedium in shake flasks.

At this stage, in order to predict the

performance of each clone candidate ina large-scale production bioreactor, anenriched medium similar to the finalproduction medium formulation with asimilar feeding regime can be employedin shake flasks. Once promising candi-dates are identified, the cell lines can beamplified to increase the copy numberof the antibody gene by using higher

concentrations of inhibitors of selectivemarkers. For example, increased con-centrations of methotrexate (MTX), adihydrofolate reductase (DHFR) inhibi-tor, are added to culture medium whereCHO cells are growing. Under theseconditions, only the cells containing a

high copy number of the DHFR gene cancontinue to grow. During amplification

of the selective marker gene, the genes of interest expressing the antibody may beco-amplified, which consequently resultsin a higher expression level and enhancedproductivity of the antibody. A similarapproach can be taken with NS0 cellscontaining glutamine synthetase (GS)by using its inhibitor, methioninesulf-oximine (MSX). The amplified cell lines

need to be evaluated for genetic stabil-

ity in the absence of the selective pres-sure because the production stage usually does not employ selective pressure.

High-throughput screening andselection technologies have been devel-oped to shorten the time required forclone selection. Screening for cells withhigh expression and secretion levels can

be accomplished using a fluorescently tagged antibody against the productexpressed either on the cell surface orsecreted in a micro gel bead, followedby fluorescence-activated cell sorting

(FACS).10,11 Another approach involvesminiaturized bioreactors or shake flasksthat can simulate standard productionbioreactor conditions, including nutri-

ent feeding. This technology usual-ly employs high-speed automation tostreamline liquid handling and analyti-cal capability for early evaluation of growth and productivity profiles forcandidate clones. Once the productioncell line is identified, MCBs and WCBsare generated under GMP conditionsFigure 2. Typical pH and lactate profiles in a fed-batch process.

8/4/2019 rp1017en00


before initiating process developmentfor commercial production.

Medium and Feeding StrategyDevelopment

In general, medium development fora fed-batch process involves batch medi-um and feed concentrate development,as well as feeding strategy optimization.Optimization of cell culture processesis often regarded as cell line-depen-dent and is, therefore, based on the

metabolism and nutrient consumptionof a specific cell line. Several approachescan be used systematically for mediumoptimization, such as single-componenttitration, spent medium analysis, andmedium blending.12 Due to regulatory concern regarding serum and other ani-

mal-derived components, the addition of animal-component-free hydrolysates to

chemically defined media is a commonapproach to increase cell density, cultureviability, and productivity. Hydrolysatesare protein digests composed of aminoacids, small peptides, carbohydrates,vitamins, and minerals, which providenutrient supplements to the media. Inaddition, hydrolysate peptides can alsoact as growth factors and stimulate pro-

duction. Non-animal-derived hydroly-

sates from soy, wheat, and yeast are com-monly used in cell culture media andfeeds. However, due to their complexcomposition and lot-to-lot variations,hydrolysates can be a significant sourceof medium variability.

The most common approach to devel-oping a feed medium uses concentrated

basal media without salts (to avoid highosmolality). Certain key feeding com-ponents (e.g., phosphate) have also beenidentified.13 During medium prepara-tion, pH and/or temperature may need to

be adjusted to completely dissolve somelow-solubility components.14 To opti-mize a feeding strategy, considerationshould be given to nutrient consump-

tion, by-product accumulation, and thebalance between growth and production.Previous studies indicated that byprod-ucts, such as lactate and ammonia, couldbe minimized by maintaining low glucoseand glutamine concentrations throughfrequent feeding.15 However, frequentfeeding is less desirable for large-scale

production in manufacturing due to itscomplexity. Stepwise bolus addition of the feed solution to the production bio-reactor is widely used in industry due toits simplicity and scalability. In general,medium development is labor-intensiveand time consuming. Approaches that

combine high-throughput screeningplatforms with statistical design of exper-iment (DOE) are commonly applied toshorten the development time.

Bioreactor Monitoring and Control

Process improvement can be achievedthrough optimization of bioreactor physi-

cochemical environments. Optimizationof culture conditions needs to balancecell growth with antibody production. Apreferred cell growth condition may notbe the optimal condition for productiv-ity and vice versa.

For mammalian cell culture, pH isnormally controlled within the range

of 6.5–7.8. High pH (≥7.0) is pre-ferred for initial cell growth, whereaslow pH can facilitate antibody produc-tion.16 However, high pH is usually associated with increased anaerobic cellmetabolism that converts more glucoseinto lactate. When lactate accumula-

tion exceeds the medium’s buffer capac-ity, the pH of the medium decreases.Because most bioprocesses are pH con-trolled at a certain level, lower pH of the culture medium induces base addi-

tion to the medium, which increasesosmolality. This can cause delayed cellgrowth and accelerated cell death. Thus,the initial pH control condition needs

to be optimized at an early stage of thecell culture process in order to maxi-mize initial cell growth. Once fast cellgrowth is achieved, the pH set point canbe shifted to a lower level to facilitateantibody production. Figure 2 showstypical pH and lactate profiles in a fed-batch process. As lactate was consumed

Figure 3. A typical antibody downstream process.

8/4/2019 rp1017en00


in the later stage of the process, pH wasallowed to increase.

Manipulating the cell cycle in CHOcell culture through a temperature shifthas been reported to extend culture lon-

gevity. A temperature shift from 37° Cto 30° C after 48 hours post-inocula-

tion can retain cells in the G1 phaselonger and therefore delay the onset of apoptosis.17 Dissolved oxygen (DO)is usually considered as a less criticalparameter, which can be controlled ina wide range (between 20–100%) of air saturation. Cell growth may appearunaffected under a certain DO level;however, DO can have significant impact

on product quality. For example, it has

been reported that a reduction in DOcaused decreased glycosylation of anti-body N-glycan chains.18

Although a bioreactor’s physicalenvironmental parameters such as pH,DO, and temperature can be monitoredon-line and controlled, monitoring of chemical and biological parameters is

limited due to the lack of reliable on-line sensors. The measurement of celldensity, metabolites, and protein con-centrations usually occurs through daily bioreactor sampling and subsequent

analysis by off-line instrumentationassays such as the Cedex cell counter(Innovatis, Bielefeld, Germany; NovaBioProfile, Nova Biomedical, Waltham,

MA), YSI analyzer (YSI Incorporated,Yellow Springs, OH), and high-pressureliquid chromatography (HPLC). Withinthe last decade, many new on-line moni-toring and control techniques for biore-actors have been developed. Some novelbiosensors are commercially available(e.g., biomass sensors and YSI’s dissolved

CO2 probe). These can measure on-lineoxygen uptake rates to optimize nutri-ent feeding strategy and at-line HPLC isused to analyze cell culture broth aminoacids and glucose.19,20 Near-infrared

(NIR) spectroscopy is a nondestruc-tive technique that can monitor mul-

tiple components simultaneously froma single probe. With recent advances inNIR instrumentation and data analysisalgorithms, its applications have beendemonstrated in the measurement of glucose, glutamine, lactate, ammonia,and even recombinant protein concen-trations in cell culture broth.21,22 WithFDA’s regulatory framework on Process

Analytical Technology (PAT), these new

technologies for in-process, on-line, andat-line monitoring are being heavily investigated as tools to identify criticalprocess variables and initiate the appro-priate control actions to maintain pro-cess and product consistency.

Cell Culture’s Impact on ProductQuality

Monoclonal antibodies are compli-cated glycoproteins subject to glycosyl-ation heterogeneity and other modifica-

tions such as N-terminal pyroglutamateand C-terminal lysine variants, methio-nine oxidation, asparagine deamidation,disulfide bond scrambling, aggregation,

and fragmentation. Multiple processfactors have the potential to change theseproduct quality attributes, potentially resulting in clinical implications includ-ing efficacy and safety. While optimizingprocesses to achieve high yield, it is criti-cal to monitor product quality changesat every stage of development.

Glycosylation variation, which canimpact in vivo IgG functions and prod-uct stability, is one of the most sen-sitive quality-related attributes that is

dependent on cell lines and cell cultureconditions. Most mouse-derived celllines are known to add Gal-α1,3-Galonto antibody heavy chain N-glycan.High Gal-α1,3-Gal content, which isnot found in human antibodies, raisesimmunogenicity concerns. It has beenreported that this glycoform is varied indifferent NS0 clones. A high-through-

put screening approach was used toselect low Gal-α1,3-Gal clones in early process development stages.23

The effects of cell culture mediaand conditions on antibody glycosyl-ation have been extensively studied.24,25 Factors such as medium serum, glucose,

ammonia, DO, dissolved CO2, and cul-ture osmolality have been reported to

cause glycosylation changes in differentcell lines. These factors can likely affectthe activity of monosaccharide transfer-ases and/or sugar transport to the Golgiapparatus, which is the major glycosyl-ation site in mammalian cells.

Although the integrity of a protein’sbackbone is usually unchanged acrossdifferent cell lines and culture condi-

tions, some modifications can occur

during cell culture processes. The IgGheavy-chain C-terminal is conservedas lysine, which is normally cut off through post-transcriptional modifica-tion. The presence or absence of C-lysine can result in product charge vari-ants. Carboxypeptidase B specifically cleaves C-terminal lysine residues. The

activity of such metal enzymes can beaffected by some medium trace elementconcentrations.

Downstream Processes and

Optimization

As enhancement of the final producttiter and cell density of a cell culture

process continues, the subsequent puri-fication stage can become a bottleneckthat impedes achieving a cost-effectiveand robust manufacturing process. Acell culture process with high producttiter generates a heavy burden on thedownstream process not only due toincreased non-product impurities, but

Figure 4. A typical harvest diagram for an antibody process.

8/4/2019 rp1017en00


also because of a high amount of unde-sirable product-related isomers. Majorchallenges to developing an efficientdownstream process for monoclonalantibody manufacturing with highly productive cell culture materials includethe following three areas:

1. Process capacity: The equip-ment and manufacturing facility need to have an adequate capac-ity to handle the high-titer cellculture broth. This includes

resin-binding capacity, processtime, number of process cycles,overall facility size, schedule andfacility flexibility, buffer and

water consumption, and compli-ance with cGMP regulations.

2. Removal of product-related

impurities: Heterogeneity iscommon in antibodies expressedby recombinant mammalian cellculture processes. These impuri-

ties include dimers, aggregates,and various isoforms resultedfrom amidation, deamidation,oxidation, and shuffling of disul-fide bonds. Removal of thesespecies cannot be achieved by protein A affinity chromatog-raphy alone, and requires addi-

tional polishing steps.

3. Removal of process impurities: Process impurities include hostcell protein (HCP), nucleic acids

(DNA, RNA), leached protein A,and potential viral contamina-tion, all of which need to becleared according to FDA

guidelines. Currently acceptedlevels of impurities for therapeu-tic antibodies reported in theliterature are: <5 ppm for HCP,<10 ng/dose for rDNA, <0.5%for dimers or aggregates, and <5ppm for leached protein A.26

A typical antibody downstream pro-cess flow is shown in Figure 3. The mostcommon industrial purification processfor an antibody uses protein A affinity chromatography as a capture step fol-lowed by two or three additional chro-matography polishing steps. The bulkproduct is then filtered through a 20-nm

filter for viral removal, concentrated to acertain target concentration in pre-for-mulation buffers using an ultrafiltration/diafiltration device, and finally sterilefiltered through a 0.2-µm membrane.

The entire purification process for GMPmanufacturing must be validated by demonstrating a 12 to 20 log reductionof viral load using at least four different

model viruses following FDA guidelines.

Centrifugation and Depth Filtration

The harvest process separates the anti-body product released in the culture broth(conditioned medium) from the cells. Atypical harvest diagram is illustrated in

Figure 5. Comparison of antibody separation in CEX with pH-gradient and salt-gradient. A: Salt-gradient; B, C, D: pH-gradient with different

buffer B concentrations. Binding Capacity: 40g/L, Linear Velocity: 300cm/hr. Condition C was chosen due to low pool conductivity and small pool

volume. Arrows indicate pool conductivities by conditions B, C, and D, respectively.

8/4/2019 rp1017en00


Figure 4. Continuous centrifugation iscommonly used to harvest large-scalecell culture broth ranging from 500 L to20,000 L, resulting in a cell-free superna-

tant. Generation of a supernatant withlow cell turbidity depends on the choicebetween a full and partial shot, as wellas the total number of shots and buffervolume per shot during the continuouscentrifugation process. The supernatantis then further clarified through depthfiltration followed by a 0.2-µm filtration.For optimal depth filtration, selection

of filter type, flux, and total membranesurface area need to be considered. Thedepth filtration step with a well-selectedfilter can not only reduce the cell turbid-ity, but can also remove up to 50% of DNA impurities and 15% of host cellprotein at neutral pH. The filtrate is then

loaded on a purification column.

Protein A Affinity Chromatography

Protein A chromatography is themost efficient purification step for anti-bodies — it can purify the product tomore than 98% purity and remove mostprocess impurities, including proteases.

Two major types of protein A res-ins are commercially available: Agarose-

based resin from GE Healthcare and

glass bead-based resin from Millipore.Both resins are robust enough to handlea high flux and have acceptable chemi-cal resistance, including resistance tohigh concentration of urea, GuHCl,and reducing agents. Dynamic bind-ing capacity is in the range of 20 to 50g per liter of resin, and both resins can

be reused up to 200 times. Because of different physical properties, each typeof resin has its advantages and disad-vantages. For example, agarose-based

resin has very low non-specific binding.However, high operational backpressureis observed when the resin is used in acolumn with an inner diameter (i.d.) of

greater than 45 cm and a bed height of greater than 20 cm. In order to maintainan operating backpressure below 30 psi,a low linear flow rate of less than 200cm/hr and a bed height lower than 20cm need to be employed. In contrast,glass-bead resin can be used with a highflow rate in a high bed height columnwith an i.d. of 45 to 160 cm without

significant backpressure. However, theglass-bead type resin is sensitive to caus-tic solutions and has higher non-spe-cific binding with process contaminantscompared to the agarose-based resin.Additional washes with a solvent-con-taining buffer are required in order to

reduce the HCP level in the protein Apool. A low loading temperature around

15±3° C is recommended to minimizepotential proteolytic degradation of theproduct and the protein A resin.27

Operating conditions of the proteinA affinity chromatography step need tobe optimized in three areas: resin typeand sample loading residence time foroptimal binding capacity; composition,pH, and volume of wash solution to

minimize the HCP and rDNA levels in

the pool; and composition and pH of elution solution to minimize turbidity,conductivity, dimer, and aggregate levels,as well as volume of the final pool, whichis important for the subsequent step.

Low pH Viral Inactivation

A viral inactivation step at low pHusually follows the protein A affinity chromatography step; two model virus-es, minute virus of mice (MVM) and

murine leukemia virus (MuLV), arecommonly used for process optimiza-tion. The protein A column pool istitrated with 10% acetic acid or 1M

citric acid to a pH between 3.3 and 3.8and incubated for 45 to 60 minutes atroom temperature.28 The choice of pH level largely depends on the stabil-ity profile of the antibody product andbuffer components. Use of low pH forviral inactivation is effective for MuLV,whereas, MVM viral particles are notefficiently killed at low pH.

After viral inactivation, the pH of theprotein A pool is titrated up with 3M Trisbuffer prior to the next step. Turbidity of the solution increases as the pH of the pool rises. The degree of turbidity of the solution after pH adjustment var-ies depending on process conditions for

each product. In order to remove theturbidity, viral inactivation is followed by

depth filtration or 0.2-µm filtration. Anoptimized depth filtration step can great-ly reduce turbidity, as well as HCP andrDNA levels in the post-viral inactivationpool solution. In addition, depth filtra-tion can efficiently remove MVM andMuLV viruses with log reduction values(LRV) of four and two, respectively.28,29 Optimization of the depth filtration step

must consider the filter type, surface area,

and the operation flux rate.

Polishing

Two to three polishing steps are usual-ly required to remove impurities. Typicalpolishing objectives are to remove:

• high molecular weight aggregates,

• trace amounts of host cell proteins,

• isomers of the product,

• residual rDNA,

• leached protein A, and

• viral contaminants.

Cation-exchange chromatography (CEX), hydrophobic-interaction chro-matography (HIC), and anion-exchangechromatography (AEX), are commonly used for polishing (Table 1).

Table 1. Applications of chromatography as polishing steps.

8/4/2019 rp1017en00


CEX Chromatography CEX chromatography used for bind-

and-elution has proven to be a powerfultool for removing product-related impu-

rities left behind by protein A affinity chromatography. CEX resin screeningcriteria include less HCP binding, highproduct dynamic binding capacity at arelatively high conductivity, and highresolution to remove target protein vari-ants. An ideal resin will leave about70–80% of HCP and most of the DNA/RNA and endotoxin in the flow-through

fraction. Deamidated or acidic mate-rial can be separated in the front peakwhile the amidated or basic material anddimers/aggregates can be isolated in thetailing part.

Elution can be performed stepwiseor as a linear gradient. Linear gradient

elution provides better purity, processcontrol, process monitoring, reproduc-

ibility, and PAT conformance.30 In con-trast, stepwise elution has proven tobe mechanically simpler and results inhigher product pool concentration.31

Recently, a linear pH gradient elutionmode in a narrow pH range has beendeveloped. Compared to salt gradientelution, pH gradient elution can providesimilar product purity, higher yield, a

smaller pool volume (up to 50% of the

volume produced by salt gradient), andthe lower pool conductivity required forthe next step. pH gradient CEX chroma-tography has been demonstrated at largescale with a recovery rate of over 95%.32 Figure 5 illustrates the separation powerof Fractogel COO– using a pH gradientfrom 5.0 to 6.0. Fractogel COO– is a

weak CEX resin with high antibody-binding capacity (>50 g/L).

HIC and CHT Chromatography Hydrophobic-interaction chroma-

tography (HIC) is an efficient modeto remove dimers and aggregates inbind-and-elution fashion. However, thismode has relatively low yield and separa-

tion resolution for other product-relatedisomers, and has to contend with highsalt concentration in the elution pools.Therefore, bind-and-elution throughHIC is becoming less popular in anti-body production. Instead, HIC chro-matography in a flow-through modeis gaining interest as a way to remove a

large percentage of aggregates. A slowflow rate is necessary because, like otherprotein binding during HIC chromatog-

raphy, aggregate binding on HIC resin isresidence time-dependent.

Ceramic hydroxyapatite (CHT) chro-matography also can be a robust step toremove dimers and aggregates.33 Severaltechnical issues regarding CHT need tobe addressed, however, including lot-to-lot viability, extractables, and the half-life of the resin.

AEX Chromatography Several chromatography modes have

proven very useful to remove traceamounts of impurities (e.g., DNA andendotoxin) and viruses. Among these,anion-exchange chromatography (AEX),is perhaps the most powerful. In mostcases, AEX chromatography is carriedout using flow-through (FT) fash-ion, in which impurities bind to the

resin and the product of interest flowsthrough. However, the use of conven-tional packed-bed chromatography withFT-AEX requires columns with a very large diameter to permit high volumetric

flow rates which are required to avoida process bottleneck at the polishingstep.26 In order to minimize the effecton the column operation by inadequate

header design, a specific bed height isrequired for proper flow distribution.34 This leads to a large column volume,which is needed for fast flow but is notoptimized for binding capacity. Thisdisadvantage with AEX columns has ledto the development of membrane chro-matography or membrane adsorbers.

Current membrane chromatography offers a convenient alternative to resinchromatography in the purification of

antibodies.35

Q membrane chromatography devic-

es have been in development for almost15 years and are a very useful approachin viral vaccine production and DNApurification for gene therapeutics. Thedevices have also been used for endotox-in removal in both laboratory and large-scale processes. Current Q-membranetechnology offers major advantagesover packed-bed resin chromatography

in antibody purification, including fast

operation, no requirement for clean-ing or storage validation, significantly low buffer consumption, easy and quickqualification, good scalability, and anti-body recovery higher than 98%.35,36

It was demonstrated that, with anewly developed representative scale-down model, a process capacity of

>3,000 g/m2 (total membrane surface)or >10.7 kg/L (membrane volume), aMVM viral log reduction of >6.0, and aMuLV viral log reduction of ~5.0 couldbe achieved.37,38 Q membranes can sig-

nificantly save operating time and bufferconsumption. Table 2 compares theperformance between Q Sepharose FastFlow column chromatography and the

Sartobind Q membrane device.

Conclusion

Recent developments and improve-ments on cell culture processes andseparation technologies make it pos-sible to lower the manufacturing cost of

Table 2. Comparison of performance and operation between column.

8/4/2019 rp1017en00


goods for antibody manufacturing, andsignificantly reduce the overall devel-opment time from clone selection toclinical manufacturing. High product

titers greater than 2 g/L by the end of cellculture and purification recovery yieldsgreater than 75% are readily achiev-able. While productivity and processefficiency continues to be improved, itis important to monitor product quality attributes to ensure the comparability with any process changes during com-mercial manufacturing. Recent publica-

tion of the Q5E comparability guide-line for biotechnological and biologicalproducts by FDA provides framework toconsider for development and manufac-turing processes.39

ACKNOWLEDGEMENT

The authors sincerely acknowledge Tony Hong for

proofreading.

REFERENCES

1. Chadd HE, Chamow SM. Therapeutic antibody

expression technology. Curr Opin Biotechnol 2001;12:

188–194.

2. Walsh G. Biopharmaceuticals: Approvals and

approval trends in 2004. BioPharm International 2005;

18(5):58–65.

3. Hubbard B, Shukla A. Platform approaches to

monoclonal antibody purification. Paper presented

at the IBC Antibody Production and Downstream

Processing meeting : 2005 March; San Diego, CA.

4. Tressel T. Development of a generic platform and

use of statistically designed experiments to enable

rapid development of several antibodies and increase

throughput for first in human antibodies. Paper present-

ed at the IBC Antibody Production and Downstream

Processing meeting: 2004 February; San Diego, CA.

5. Chu L, Robinson DK. Industrial choices for pro-

tein production by large-scale cell culture. Curr Opin

Biotechnol 2001;12:180–187.6. Pierce LN, Shabram PW. Scalabi lity of a disposable

bioreactor from 25L – 500L run in perfusion mode with

a CHO-based cell line: A tech review. BioProcessing J

2004;4:51–56.

7. Gay R, Birch J. Approaches to improving the perfor-

mance of mammalian cell cultures for protein produc-

tion. Paper presented at the BioLOGIC conference :

2004 October; Boston, MA.

8. Li F, Pendleton R, Lee B. Scale-down model devel-

opment and process characterization for a fed-batch

cell culture process. Paper presented at the ACS BIOT

conference : 2005; San Diego, CA.

9. Burky J. Improving NS0 fed-batch culture produc-

tivity. Paper presented at the IBC Cell Culture and

Upstream Processing conference : 2003; San Diego,

CA.

10. Akselband Y, Trnovsky J, Trnovsky PM. Transfected

cell line enrichment using the gel microdrop (GMD)

secretion assay. BioProcessing J 2003;2:83–88.

11. Holmes P, Al-Rubeai M. Improved cell line develop-

ment by a high throughput affinity capture surface dis- play technique to select for high secretors. J Immunol

Meth 1999;230(1-2):141–147.

12. Fletcher T. Designing culture media for recombinant

protein production: A rational approach. BioProcess

International 2005;3:30–36.

13. deZengotita VM et al. Phosphate feeding improves

high-cell-concentration for monoclonal antibody pro-

duction. Biotechnol Bioeng 2000;69:566–576.

14. Robinson DK et al. Optimization of a fed-batch pro-

cess for production of a recombinant antibody. Annals

NY Acad Sci 1994;745:285–296.

15. Xie L, Wang DIC. High cell density and high mono-

clonal antibody production through medium design

and rational control in a bioreactor. Biotechnol Bioeng

1996;51:725–729.

16. Sauer PW et al. A high-yielding, generic fed-batch

cell culture process for production of recombinant anti-

bodies. Biotechnol Bioeng 2000;67:585–597.

17. Moore A et al. Effects of temperature shift on cell

cycle, apoptosis, and nucleotide pools in CHO cell

batch culture. Cytotechnol 1997;23:47–54.

18. Kunkel JP et al. Dissolved oxygen concentra-

tion in serum-free continuous culture affects N-linked

glycosylation of a monoclonal antibody. J Biotechnol

1998;62:55–71.

19. Zhou W et al. Fed-batch culture of recombinant

NS0 myeloma cells with high monoclonal antibody

production. Biotechnol Bioeng 1997;55:783–792.

20. Larson TM. Chemometric evaluation of on-line

HPLC in mammalian cell cultures: analysis of amino

acids and glucose. Biotechnol Bioeng 2002;77:553–

563.

21. Yeung KS et al. Near-infrared spectroscopy for

bioprocess monitoring and control. Biotechnol Bioeng

1999;63:684–693.

22. Harthun S, Matischak K, Friedl P. Determination

of recombinant protein in animal cell culture super-

natant by near-infrared spectroscopy. Anal Biochem

1997;251:73–78.

23. Zhu GJ. A high-throughput system to analyze N-glycans of NSO cell produced antibodies. Paper pre-

sented at the IBC Cell culture and upstream processing

conference : 2003; San Diego, CA.

24. Jenkins N, Parekh RB, James DC. Getting the

glycosylation right: Implications for the biotechnology

industry. Nature Biotechnol 1996;14:975–981.

25. Jefferis R. Glycosylation of recombinant antibody

therapeutics. Biotechnol Prog 2005;21:11–16.

26. Fahrner RL. Industrial Purification of Pharmaceutical

Antibodies: Development, Operation, and Validation of

Chromatography Process. Biotechnol Gen Eng Rev

2001;18:301–327.

27. Laverdiere, A. Reduction of ligand leaching during

protein A affinity chromatography by controlling proteo-

lytic degradation of protein A. Paper presented at the

229th ACS national meeting : 2005 March 13–17; San

Diego, CA.

28. Zhou J et al. Development of A Cost-Effective,

Robust Process for Recombinant Monoclonal Antibody

Production. Paper presented at the Williamsburg

BioProcessing 2nd Annual Europe Antibodies &Recombinant Proteins meeting : 2005 May 22–24;

Amsterdam, The Netherlands.

29. Tipton B et al. Retrovirus and parvovirus clearance

from an affinity column product using adsorptive depth

filtration. BioPharm 2002;Sept:43–50.

30. Blank GS. The future of process development

for MAb manufacturing. Paper presented at IBC’s

14th International Antibody Production & Downstream

Processing : 2005 March 9–11; San Diego, CA.

31. Gagnon P. The quest for an industrial purifica-

tion method for generic IgG. Paper presented at

the Williamsburg BioProcessing 2nd Annual Europe

Antibodies & Recombinant Proteins meeting : 2005 May

22–24; Amsterdam, The Netherlands.

32. Zhou J et al. pH-gradient cation exchange chroma-

tography for process-scale antibody purification. Paper

presented at the 229th ACS national meeting : 2005

13–17; San Diego, CA.

33. Franklin S. Removal of aggregate from an IgG 4

product using CHT ceramic hydroxyapatite. BioProcess

International 2003;January:50–51.

34. Yuan QS et al. Flow distribution in chromato-

graphic columns. J Chromatog A 1999;831:149–165.

35. Gottschalk U, Fishcher-Fruehholz S. A cutting

edge process technology at the threshold. BioProcess

International 2004;May:56–65.

36. Knudsen H et al. Membrane ion-exchange chro-

matography for process-scale antibody purification. J

Chromatog A 2001;907:145–154.

37. Tressel T, Zhou J. Process development of an

anion exchange membrane. Paper presented at the

North American Membrane Society Conference : 2005

June 4–6.

38. Dermawan S et al. Process development of a flow-

through anion exchange membrane chromatography in

protein purification. Paper presented at the 229th ACS

national meeting : 2005 March 13–17; San Diego, CA.

39. Food and Drug Administrat ion [FDA]. FDA guid-

ance for industry: Q5E comparability of biotechno-

logical/biological products subject to changes in their manufacturing process. <www.fda.gov/cder/guidance/

index.htm> June, 2005.

This article is reprinted courtesy of TheWilliamsburg BioProcessing Foundation(WilBio), BioProcessing Journal , Vol. 4No 5.

RP1017EN00

Documents

rp1017en00