presentation overview
TRANSCRIPT
An evaluation of Protein A and Non-Protein A methods for the recovery of monoclonal antibodies and considerations for process scale-up
Martin Smith
LONZA, 224 Bath Road, Slough, SL14DX
Presented at “Scaling-up of Biopharmaceutical Products”, 26/27th January 2004, The Grand, Amsterdam
13 Apr 2004 / 2
LONZA CUSTOM MANUFACTURING
At present, LCM offers production and purification of monoclonal antibodies at two sites
Portsmouth, New Hampshire, USA3*20,000L Reactors, associated purification train2*5000L, 1*2000L, 2*1500L (perfusion)
– associated purification trains
Typically Phase 3 launch & In-market supply
Slough, UKmix of 200 and 2000L reactorsMultiple batch dedicated purification facilities
Typically Phase 1 & 2, rapid speed to clinic/market
13 Apr 2004 / 3
Background
At early phase development, speed to clinic is critical
At all scales of operation, low cost of goods with reliable, scaleable manufacturing operations is essential
Protein A capture widely regarded as fastest route for purification process development
LONZA experience with >100 antibody processes has resulted in generation of truly “generic” approach to antibody purification
13 Apr 2004 / 4
Potential Scale constraints
US, 3*20,000L reactorsFacility designed to handle wide range of current and future monoclonal antibody processes
As fermentation titres increase, need to address impact on downstream requirements
Product and buffer hold tanks
UK, mix of 200, 2000L reactorsFacility designed to be very flexible across scales of operationSeek to maximise facility throughput to increase customer base and speed products to clinic
13 Apr 2004 / 5
Presentation overview
Introduce case study for generic purification process upgrade
Highlight key operational challenges with traditional recovery technology
Demonstrate improvements resulting from process development to a high throughput process upgrade
Examine future potential to improve purification costs.
13 Apr 2004 / 6
Introduction to case study
Case Study, IgG1 from GS-NSO cell line
2,000L reactor at ~0.5g/L, 1kg into purification
“generic” purification process, average capacities, average number column cycles, typical fluxes…etc
PnA – VI - IEX – UF – IEX – VRF – UF – SF
LONZA generic approach traditionally utilises compressible matrices during chromatography operations
13 Apr 2004 / 7
What’s the problem with compressible matrices
0 100 200 300 400 500 600 7000
5
10
15
20
25
30
35
40
Pressure Drop versus Flowrate
Packed bed Height = 15cmMatrix = Sepharose 4 Fast FlowBuffer =PBS @ 22°C
Pres
sure
Dro
p (p
si)
Linear velocity (cm/h)
1.6cm 2.6cm 5.0cm 10cm 20cm 28cm 40cm
ColumnDiameter
d1400
Chromatography scale-up usually achieved by increasing diameter only
Bed height & flow rate kept constantThis reduces supportive wall effectsResults in increased pressure drop at same flow velocity
Imposes severe limits to usable bed heights and flow rates
Stickel JJ, Fotopoulos A.
Biotechnol Prog. 2001 Jul-Aug; 17(4): 744-51.
13 Apr 2004 / 8
Implementing High Productivity Process upgrade
Numerous strategies are available
Simplest route is to upgrade compressible matrices to high flow rate rigid matrices
Should improve schedule but not buffer consumption
Targeted initial two column steps for matrix upgrade evaluation.
Imposed constraint…..chromatography buffer chemistry should remain the same despite change in matrices.
Reflects a typical rapid development project
13 Apr 2004 / 9
Purification schedule constraints
Duration of any purification batch is dependent on many factors.
Binding capacity vs flow rate of chromatography matricesColumn size, aspect ratio
Operational philosophyShift patterns, labour availabilityEquipment availability/turn aroundHold pointsTarget throughput per annum
13 Apr 2004 / 10
Batch productivity
For purposes of case study, duration of typical batch is represented as 100%
Improvements due to upgrades provided as % decrease in overall batch length
Normalised Batch time = 100%
VI UF IEX UF VRF IEX UFPrA
13 Apr 2004 / 11
PnA Sepharose vs Rigid Upgrade
0 10 20 30 40 50 600
20
40
60
80
100 Rigid Upgrade, 0.5g/L
Bre
akth
roug
h (C
/Co)
%
IgG loaded (mg/mL.matrix)
PnA Seph, 0.5g/L
PnA load concentration ~ 0.5g/L representing for example very low fermentation titres
PnA Sepharose DBC=18g/L
Rigid Upgrade DBC=18g/L
However Rigid Upgrade results obtained at 3x higher flow rate
Greater differences exist at higher breakthroughs
But at expense of yield
13 Apr 2004 / 12
Effect of PnA load concentration on PnAcapacity
PnA load concentrations >5g/L result in dramatic increases in PnAdynamic binding capacity
PnA Sepharose DBC=33g/L(83%higher than at 0.5g/L)
Rigid Upgrade DBC=23g/L(23% higher than at 0.5g/L)
Rigid Upgrade results obtained at 3x higher flow rate
Direct impact on throughput… 0 10 20 30 40 50 600
20
40
60
80
100
Rigid Upgrade, 5g/L
IgG loaded (mg/mL.matrix)
PnA Seph, 5g/L
Bre
akth
roug
h (C
/Co)
%
13 Apr 2004 / 13
Effect of PnA upgrade on process throughput
First target for high throughput upgradeImplement rigid matrix for PnA step
Maintained column geometry (& cost)Increased number of cycles (4 to 5)Decreased PnA step time (by 60%)Increase in buffer consumption (by 25%)
Overall productivity improvement of 15%
PrA VI UF IEX UF VRF IEX UF
VI UF IEX UF VRF IEX UFPrA
-15%
13 Apr 2004 / 14
Upgrade IEX productivity
Evaluated impurity and contaminant clearance ofSepharose based IEX step vs.Rigid IEX matrix
Experimentally verified…Equal clearance of contaminants, impuritiesEqual protein purificationIdentical product quality and step yield
Major improvement…3x increase in step productivity due to 3x increase in linear velocity for same column volume.
13 Apr 2004 / 15
Effect on production schedule of PnA and IEX rigid matrix upgrades
PrA VI UFIEX
UF VRF IEX UF
PrA VI UF IEX UF VRF IEX UF
VI UF IEX UF VRF IEX UFPrA 100%
-15%
-23%
23% Productivity increase was achieved using the upgraded high throughput processNo detrimental effect on purification, yield or product quality observed!
13 Apr 2004 / 16
Potential for increased productivity through upgrading 2nd IEX step
Future plans to evaluate high throughput matrices for 2nd IEX step in generic process
Expected to realise similar productivity increases…
Normalised Batch time = 100%
VI UF IEX UF VRF IEX UFPrA
PrA VI UFIEX
UF VRF IEX UF - 44%
13 Apr 2004 / 17
Potential issues arising from use of rigid matrices
Rigid matrices…Generally smaller particles sizes
Slightly higher pressure dropsPacking complicationsShear sensitivityCleaning regimes£/LGeneric applicability to wide range of Mab’s and cell culture feed stocks.
Extensive re-development programs might be required in some cases
13 Apr 2004 / 18
Scale-up Pilot Plant-PurificationSlough UK
HYDRAULIC SKID
PACKING SKID
40cm RESOLUTE COLUMN
55L SLURRY VESSEL
10mm CHROMATOGRAPHY SKID
13 Apr 2004 / 19
LSBO Purification suite P6B
Purification suite P6-B:
1.4m diameter stainless steel columnChromatography skid to left
2m diameter stainless steel chromatography column
Filtration rig to left
13 Apr 2004 / 20
Operational issues implementing high productivity processes
Supply of buffers in a reduced time can shift operational bottlenecks to:
Buffer make-upBuffer holdProduct hold tank sizesPurification suite floor space
(for increase buffer volume per unit time)Subsequent steps
13 Apr 2004 / 21
Disposable buffer makeup and supply capabilities-pilot scale
Buffers prepared 1-2 days ahead of use, in dedicated buffer prep facility using disposable componentsBuffer prep/hold located adjacent to process area.
500L makeup tank 200 and 500L prep tanks with storage vessels
13 Apr 2004 / 22
LSBO Media Preparation tanks
Media Prep
Visible tanks are:1*17,500L
(far left)
2*3,500L(centre/right)
Buffer PrepSimilar in design, vastly increased number of tanks!!
13 Apr 2004 / 23
Reducing buffer make-up by inline dilution
LONZA has successfully implemented in-line dilution for PnA equilibration buffers
Reduced buffer prep requirements by 10x
CIR101 CIR102 AIR121pH FIR141 TIR101 SetMark
0
50
100
150
mS/cm
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
pH
0.0 2.0 4.0 6.0 8.0 10.0 min
Wat
er ri
nse
1x b
asel
ine
10x_
conc
entr
ate
Feed
back
ena
bled
21.0
22.0
23.0
24.0
25.0
26.0
27.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
6.0 7.0 8.0 9.0 10.0 min
Feed
back
ena
bled
13 Apr 2004 / 24
The Story, so far…
Where schedule bottlenecks occur, rigid matrices can dramatically improve individual batch productivities
Scale-up no longer restricted to diameter increases alone
Watch for where the bottlenecks move next…Usually buffer makeup, hold, then product tank sizesAs fermentation titres continue to rise…
Column diameters must inevitably increaseAssociated matrix costs therefore rise
13 Apr 2004 / 25
Non-Protein A recovery processes
Major cost centre for Protein A processes is the cost of the Protein A itself (and CIP reagents)
Currently evaluating numerous non-Protein A alternatives that compete favourably with Protein A wrt
CostDevelopment timelinesPurification (protein, impurities, contaminants)Product qualityOverall process productivity
13 Apr 2004 / 26
Prometic Biosciences-MabSorbent A2P
Provided material to Prometic Biosciences under MTA
Developed two step purification strategy
Fractogel + Mabsorbent A2P
Key benefit is reduced matrix cost per litre(approximately 3-5x reduction in raw material costs)
13 Apr 2004 / 27
Prometic 2 step process for case study IgG
Two step process outline…
1:3 dilution of clarified harvest materialLoad onto S Fractogel (Bind/Elute)Direct load onto MabSorbent A2P
No intermediate UF!
Initial process evaluation stopped after A2P
13 Apr 2004 / 28
Productivity comparison between Protein A and Non-Protein A process
15% Faster
35% Slower
PrA VI UFIEX
S -Fract
VI UF IEXPrA
VI UF VRF IEX UFA2P
Productivity of Prometic two step process is competitive with high throughput PnA upgrade
13 Apr 2004 / 29
Non-Protein A purification performance
Ion exchange antibody capture(Fractogel SO3-)
Mabsorbent A2P polishing step High capacities and flow rates, 150-200 cm/hr)
High recovery (95% 2 step process recovery)High purity (>95%)Contaminant removal (DNA <1pg/mg protein, 2 HCP)No leached affinity ligand detected!
13 Apr 2004 / 30
Conclusions
Capacity constraints in the purification of monoclonal antibody products may be alleviated by employing high throughput rigid matrices.
Increased throughput can be achieved with little increased cost of goods
Capacity bottlenecks shift to other operational areas
Non-Protein A media are becoming competitive with accepted traditional recovery strategies for monoclonal products