Transcript Presentation of the Animal Cell Technology group

Fed-batch or perfusion for the production of
biopharmaceuticals by animal cell
cultivation?
Veronique Chotteau, PhD.,
Cell Technology group, KTH, School of
Biotechnology.
[email protected]
Workshop on Industrial scale cultivation of cells in pharmaceutical or antibody
production systems
April 23 2009, Royal Institute of Technology, Stockholm, Sweden
Mode of operations
Batch
bioreactor
medium & additives addition during the cultivation
O2 tension
Fed-batch
pH
From 10 to 21
days
temperature
Typically 5 days
Batch-refeed = easy option
Perfusion
continuous medium
renewal during
the cultivation
cell broth
cell-free
supernatant
cells
separation
device
harvest
collection
tank/bag
From 3 weeks
to 6 months
System with internal separation device
Perfusion devices in processes using stirred tank reactors - Filtration
fresh medium
Spin filter
cell-free
supernatant
LEVEL
CONTROL
QUI CK CONNECT
FLUI D
I NLET
ADDI TI ON
PUMP
Alternating
tangential
flow (ATF)
VALVE
FI LTRATE PUMP
FI LTRATE
HF MODULE OR SCREEN MODULE
HOUSI NG
LI QUI D
LEVEL
CONTROLLER
EXHAUS
PROCESS VESSEL
DI APHRAGM
I
ON
OFF
ATF
PUMP
FI LTER
STAND
Hollow fibre
etc,…
Sources: Refine, Microgon
Perfusion devices in processes using stirred tank reactors - Centrifugation
Centritech centrifuge
Cell-free supernatant
(to harvest tank)
Feed = cell broth
(from bioreactor)
Concentrated cells (to bioreactor)
Cell-free
supernatant (to
harvest tank)
Feed = cell broth
(from bioreactor)
Concentrated cells
(to bioreactor)
Westfalia centrifuge
Sources: Kendro, Westfalia
Perfusion devices in processes using stirred tank reactors - Sedimentation
fresh medium
cell-free
supernatant
cell
sedimentation
Inclined settler
Sedimentation by gravity
Sources: Applisens
Acoustic settler
Forced sedimentation by acoustic wave
Comparison of several perfusion systems
Properties
Spin
filter
ATF
Hollow
fibre
Centritech
Acoustic
settler
Inclined
settler
Fouling (or poor operation)
--
+
--
+
+
-
Simplicity of operation
+
++
++
-
+
+
++
++
-
-
--
++
++
+
++
+
+
-
Easiness to optimise the operation
parameters
Easiness to obtain and maintain
sterility
++
Scalability
++
Possibility for re-sterilisation
yes if
external
yes
yes
yes
yes if
external
yes if
external
Residence time of cells in
separator and connection tubing
0 if
internal
1–2
min
 10 sec.
2 – 9 min
3 – 14 min
10 – 20
min (?)
Running cost
+
-
-
-
+
+
Purchase cost
+
+
+
-
+
+
Characteristics of fed-batch and perfusion principles
Fed-batch
Perfusion
•
Continuous medium addition
•
•
Addition of selected components
•
Continuous medium addition and
removal of used medium
Addition of complete medium
energy source (e.g. glucose, glutamine),
amino acids, vitamins, salts, metal trace,
growth factors
• Removal of the by-products, etc.
– e.g. toxic lactate, ammonia, CO2
– protease
•
Dilution of the by-products, etc.
e.g. toxic lactate, ammonia
•
Changing environment for the cells
•
Constant environment for the cells
•
Environment of lower stability of the
product of interest
•
Environment of higher stability of
the product of interest
•
Alkali addition
•
Less alkali addition for pH control
Implications of the characteristics of fed-batch and perfusion at early
clinical phase development
Fed-batch
Perfusion
•
Continuous medium addition
•
•
Addition of selected components
•
Need to develop the feed strategy,
e.g. feed medium,
feed rate, …
Continuous medium addition and
removal of used medium
Addition of complete medium
Require only the selection of a
medium ‘good’ enough
•
Dilution of the by-products, etc.
presence of by-products
•
Removal of the by-products, etc.
•
Changing environment for the cells
•
Constant environment for the cells
•
Environment of lower stability of the
product of interest
•
Environment of higher stability of
the product of interest
•
Alkali addition
•
Less alkali addition for pH control
Consequences of fed-batch and perfusion techniques
Fed-batch
Perfusion
•
Accumulation of the product of
interest
•
Continuous dilution of the product of
interest  lower concentration
•
•
Single harvest
Smaller harvest volume
•
•
Multiple harvests  variation?
In total larger volume of harvest 
increase work load of down-stream
• Larger bioreactor (up to 20000 L)
 less available, less ‘convenient’
• Smaller bioreactor (up to 500 or 1000 L)
 more available, even disposable
•
•
Technically less complex
Technically more complex 
– higher risk for failure
– higher risk for contamination
– requires perfusion device knowhow
Perfusion in 2 L bioreactor with ATF
(Karin Tördahl, Véronique Chotteau) - The first two runs
Medium
sponsorin
g
IrvineScie
ntific, US,
CA
Collaborati
on &
cartridge
sponsorin
g, GE
Healthcare
Selection of perfusion or fed-batch in biopharmaceutical industry
Obvious selection of perfusion if
• Instable protein
– production of proteases by the cells, e.g. serine protease, metalloprotease
– physical instability at cultivation pH and temperature
•
Product of interest toxic for the cells
Selection of perfusion or fed-batch in biopharmaceutical industry
•
Preference for fed-batch
– High titer and high cell density
– Simplicity (technical) of fed-batch process
– Lower contamination & failure risk
– Dominance of antibodies  stable proteins
– Usage of technical platforms with fed-batch process
– Large companies  availability of large bioreactors
3 perfusion, 6 batch/fed-batch, 4 unknown process in ‘Table 2: Numerical
summary of BLA product generated in mammalian cell-culture systems
1996-2000’, Chu & D.K. Robinson, Curr. Opinion in Biotechn. 2001,
12:180.
•
Selection of perfusion when
– Know-how present in company
– Previous experience of product
– Equipment adapted for perfusion
– Desire to use smaller bioreactors
Simulation of a comparison of fed-batch and perfusion – case of cell
specific productivity = 5 pg/cell/day
Simulation of a comparison of fed-batch and perfusion – case of cell
specific productivity = 5 pg/cell/day
Simulation of a comparison of fed-batch and perfusion – case of cell
specific productivity = 5 pg/cell/day
Simulation of the product mass in a 1 L bioreactor in fed-batch or
perfusion as a function of the cell specific productivity
Notice: Konstantinov (2006) cell specific productivity  20 to 60 pg/cell /day (Bayer),
Wurm (2004) cell specific productivity  20 pg/cell /day
High viability & constant environment  favorable for protein
Example of fed-batch (Wurm 2004
cited ‘courtesy of Lonza’)
Example of perfusion (Konstantinov 2006
perfsuion rate from 6 RV/day  2 RV/day)
Basic development for simple fed-batch or perfusion processes
•
Assumptions: small company, cell producing 5 pg/cell/day, ‘simple’ protein
•
Purpose: process ‘good enough’ for production of phase I material (up-scale
at 10 L bioreactor)
•
Rough estimate of USP & DSP development cost
•
Conclusions
•
Perfusion
– best knowledge within big/middle biotech/big pharma
– very attractive due to smaller bioreactor volumes and disposable
bioreactors (up to 1000 L)
•
Perfusion can be an interesting alternative for small companies
– not higher cost
– favorable for proteins
•
Requires perfusion device
– simple
– robust
– scalable
– affordable cost
 ATF
Questions?