“Successful transfer and scale up of a mammalian cell Abstract

culture process producing a monoclonal antibody from

Large scale mammalian cell culture is typically performed in suspension
culture. Two configurations that are used in industry are stirred tank and
air-lift fermenters. The debate about the advantages and disadvantages of
these configurations is one that has received attention in recent years.

a stirred tank reactor to an air-lift fermenter” This poster examines how a fed-batch transfectoma process operated in a
stirred tank reactor was successfully transferred to an air-lift fermenter.
Cell growth and metabolism, productivity and product quality are
compared. Optimisation of both modes of operation to ensure a similar
by S.Abraham(1), H.Bland(1), A.Westlake(2), L.Smith(2), M.Brown(2), D.Velez(3), J.Goldstein(3), B.Hornberger(3), A.Daus(3), Q.Zhou(3) physicochemical environment for the cells maximises the ability to
maintain equivalent product quality and yield. The process of transferring
technology from customer to contract manufacturer and between
Address : (1) : Lonza Biologics Inc, 101 International Drive, Portsmouth NH 03801. development and manufacturing is described with particular reference to
areas that require careful consideration to ensure seamless process
(2) : Lonza Biologics plc, 228 Bath Road, Slough SL1 4DY, United Kingdom. (3) : ImClone Systems, 22 Chubb Way, Somerville NJ 08876. transfer to manufacturing scale.

Results
Introduction & Aim Program
• Objectives : Cell Growth 1.E+07
IEF Analysis
• Process review
1200L STR (n=4)
10L ALF (n=3) 1 2 3 4 Lane Sample
– To transfer and scale-up a fed-batch mammalian cell culture process from 1200 L 1 pI Markers

(cell.mL )
130L ALF (n=1)

-1
Stirred Tank Reactor (STR, ImClone) to 5000 L Air-Lift Fermenter (ALF, Lonza) • Raw materials qualification 5000L ALF (n=4)
2 Product ImClone 1200 L
STR
– Maintain process and product equivalence • Laboratory evaluation in flask culture and 10 L ALF 1.E+06
3 Product Lonza 5000 L ALF

Viable Cell Concentration

• Scale process to 130 L ALF 4 pI Markers
• Key determinants of equivalence :
• Demonstrate equivalence of product derived from laboratory and
– Product quality
pilot scale ALF to 1200 L STR
1.E+05

– Productivity (volumetric, specific)

• Process scale-up to 5000 L ALF pilot lot
– Cell growth
– Cell metabolism
• Demonstrate equivalence and consistency of 5000 L ALF
manufacturing scale to 1200 L STR
1.E+04
0 50 100 150 200 250 300 Non-Reduced SDS-PAGE Analysis Reduced
Elapsed Time (hours) 1 2 3 4 5 6
• Strategy : 1 2 3 4 5 6

Product Accumulation
Lane S ample
– Minimise changes to the process 1 Molecular Weight Markers
500
2 Blank
– Evaluate impact of any process modifications
– Control critical raw materials during scale-up
Materials & Methods 400
1200L STR (n=4)
10L ALF (n=3)
130L ALF (n=1)
5000L ALF (n=4)
3 Product ImClone 1200 L
S TR
4 Product Lonza 5000 L ALF
• Cell line : Mouse SP2/0 AG14 Transfectoma 5 Blank

Product Concentration
6 Molecular Weight Markers
• Product : Murine/Human Chimeric IgG1 300

Comparison of STR vs. ALF • Serum-free medium with several concentrated nutrient feeds 200

Stirred Tank Reactor
• Low aspect ratio (typically 1:1)
• Control parameters : Temp = 36.7°C, pH 7.1 to 6.8, DO 25% saturation with air
Air-Lift Fermenter
• High aspect ratio (typically 5:1)
• Product concentration determined by Protein A HPLC
100

Conclusions
• Greater flexibility in terms of • Restricted flexibility in operating volume • Glucose, glutamine and lactate concentrations determined by autoanalyser 0

• Successfully transferred fed-batch mammalian cell culture

0 50 100 150 200 250 300

operating volume • Ballast gas ensures CO2 does not accumulate • Ammonia determined by colorimetric assay Elapsed Time (hours)

• CO2 accumulation may be an issue to toxic levels • pCO2 determined by Blood Gas Analyser process from STR to ALF
• Mixing time : 3 minutes (Ref 1) • Mixing time : 5 min • Product quality : Isoelectric Focusing (IEF) and Sodium Dodecyl Sulfate
Carbon Dioxide Partial Pressure 140

• Process shows high degree of equivalence and consistency

1200L STR (n=4)

• kLa typically up to 10 hr-1 (Ref 1) • kLa typically up to 20 hr-1 Polyacrylamide Gel Electrophoresis (SDS-PAGE) Analysis (Data shown) 120
10L ALF (n=3)
130L ALF (n=1)
5000L ALF (n=1)

• Tip speed » 0.5m.sec-1 • Liquid velocity » 0.1m.sec-1 • More extensive product characterisation performed (Data not shown) 100

• Impeller requires mechanical seals, • No moving parts or mechanical seals for pCO 2 (mm Hg)
80
• Cell specific productivity and nutrient metabolism were
maintenance impeller statistically equivalent between 1200 L STR and 5000 L ALF
Process Flow Process Transfer
60

• Product quality was equivalent between STR and ALF

ImClone Lonza 40
Manufacturing Manufacturing
Scale Scale

Ampoule
20
Ampoule

Key Areas for Consideration

CONTRACT

• Process Import and Transfer performed in approximately six

CUSTOMER 0
MANUFACTURER
Static T-Flask Suspension
IMCLONE SYSTEMS
T-25 to T-175 Shake Flask
125 mL to 2 L
LONZA BIOLOGICS 0 50 100 150 200 250 300

Elapsed Time (hours)

KNOWLEDGE AND
Agree Scope of work
Agree Schedule
Define deliverables
Monitor Progress
months
• Gas Flow Rate Differences
Product and Metabolite Summary
UNDERSTANDING DEVELOPMENT QUALITY

• Raw Materials
Spinner Flask Spinner Flask
0.5 L to 8 L 3L q Expertise in cell q Raw materials
culture and testing
q Cell Line process scale up q Product Testing
q Product

– 0.02 vvm (STR) vs 0.04 vvm (ALF)

q Compliance

– Identify Critical Raw Materials

q Process q Validation
q Raw Materials
Transfer of
q Regulatory
ImClone Lonza Lab Lonza Pilot Lonza
MANUFACTURING

Acknowledgements
Manufacturing Scale (n=3) Scale (n=1) Manufacturing
– (NB : vvm = maximum total gas flow rate
knowledge,
REQUIREMENTS q cGMP expertise

– Vendor Qualification
understanding and
data
30 L Seed STR 80 L Seed ALF
q Product Quantity SALES AND Scale (n=4) Scale (n=4)
per unit liquid volume)
LOGISTICS
q Product Quality MARKETING

• Inoculum Methodology
q Timeline q Contract Management q Procurement of cGMP
q Budget
q Project Co-ordination
q
raw materials
Scheduling
qproduct 100 ± 18 90 ± 9 79 82 ± 4 • Dr.Jerry Tong • Lonza Quality Services Laboratories
– potential differences in pCO2
Proposal
q Regulatory Package Data Reporting

• Lonza Fermentation Development • Lonza Purification & Assay Development Staff

– Static vs. shake flask culture
Identify and resolve
operational normalised units
differences 9 -1
.(10 cells.hr)
• Foaming : Antifoam usage • Lonza US Manufacturing Fermentation • ImClone Systems Manufacturing, Process Development
– Method of gas exchange 300 L Seed STR 580 L Seed ALF
qglucose 10.7 ± 3.0 19.6 ± 4.1 14.2 14.8 ± 0.7
Suite 3 Staff and Analytical QS Method Development Groups
– Total sparge gas flow rate higher for ALF
9 -1
mg(10 cells.hr)

• Subculture Ratio than STR qglutamine 6.0 ± 0.7 5.9 ± 3.7 8.0 6.5 ± 0.5
– 1 in 6 (STR) vs 1 in 9 (ALF)
9 -1

References
mg(10 cells.hr)

– Effect of antifoam usage on Fermentation 5

subculture ratio
4

qlactate 8.0 ± 1.5 10.2 ± 1.0 7.8 8.0 ± 0.8

3
2
1

and Downstream Operations

Batch Records Data
Final Product
1200 L 9 -1

1 Nienow, A.W, et al (1996) Cytotechnology 22, 87

5000 L
Production STR Production ALF mg(10 cells.hr)

qammonia 0.7 ± 0.1 0.7 ± 0.0 0.7 0.6 ± 0.1 “Homogenisation and oxygen transfer rates in large agitated and sparged
animal cell bioreactors : Some applications for growth and production”
9 -1
mg(10 cells.hr)