optimized feed strategies for increased titer and … world leader in serving science steve gorfien,...
TRANSCRIPT
The world leader in serving science
Steve Gorfien, PhD
Optimized Feed Strategies for Increased Titer and Product Quality
Senior Director, R&D
2
Agenda
Pain points in fed-batch culture
Addressing pain points
• Nutritional and dilutional gaps
• Process considerations
• Importance of product quality
• Factors that can influence product quality
• Latest advance in high-throughput glycan testing
Conclusions
3
Pain Points in Fed-Batch Culture
Nutritional gap
• Engineered cells have high nutritional demands
• Providing the right nutrients at the right time
• Matching basal medium and feed
Dilutional gap
• Dilution impact of feed addition
• How concentrated can you make the feed to be able to feed more?
Process
• High/low pH feeds add to preparation complexity and time
• Pick clones in a system that reflects ultimate process
Desired end point
• Cell density, product titer, product quality
• High cell density may not correlate with high titer or with desired critical quality attributes
4
Engineered Cells Have High Nutritional Demands
0
50
100
150
200
250
300
350
400
0
5
10
15
20
25
30
35
40
0 1 2 3 4 5 6 7 8 9 10 11
EP
O t
ite
r (m
g/L
)
Via
ble
ce
ll d
en
sit
y (
10
6 c
ell
s/m
L)
Run time (days)
0 2 4 6 8 10 12
Co
nc
en
tra
tio
n
Run time (days)
Arginine Asparagine Cysteine Glutamine
5
Balanced Formulations Drive Performance
Sample target process
Perform spent media analysis Analyze growth/titer
Calculate consumption rates
Analyze nutritional pathways Optimized balanced formulations
Via
ble
cell
density (
VC
/mL)
IgG
titer
(mg/L
)
Time (days)
6
Benefit of Rationally Integrated Optimization
Cell Line/control process
• CHO-K1SV–high IgG expression
• Modified catalog base medium
• Proprietary feed/process
Desired outcome
• Improved titer
• Proprietary base medium
• Formulation access
• Stay with current feed/process
Modification of basal medium alone was insufficient to improve titer
Jiang et al, BioProcess International 10(3) March, 2012
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1 2 3 4 5
Re
lati
ve
Ig
G t
ite
r
Steps
7
Process Intensification Addresses Nutritional and Dilutional Gaps
Higher total productivity by maintaining qp
Pro
tein
tit
er
Integral viable cell density
Theoretical trend
Dilutional gap
Nutritional gap
Nutrient feeds
Glucose feed only
8
IVC (1012 viable cell days)
IgG
tit
er
(g)
0 0.002 0.004 0.006 0.008 0.01
0.05
0.1
0.15
0.2
0.25
0.3
0
qp= 29 pg/cell/day
CD OptiCHO
CD FortiCHO
CD OptiCHO w/ EFC
CD FortiCHO w/ EFC
Sustaining qP Increased QP in Matched Basal/Feed Media
Elimination of nutritional gap = ~2-fold titer increase
Via
ble
ce
ll d
en
sit
y (
10
6 c
ell
s/m
L)
Culture time (days)
0
2
4
6
8
10
12
0 2 4 6 8 10 12 14 16
IgG
tit
er
(mg
/L)
0
500
1000
1500
2000
2500
3000
CD OptiCHO
CD FortiCHO
CD OptiCHO w/ EFC
CD FortiCHO w/ EFC
9
Granulated Format Improves Process and Productivity
Processing time cut in half vs. DPM
Product
Solution
delivery
Spray nozzle
Air preparation unit
Neutral pH pH range: 6.5–7.1
Super-concentrated feeds Target: 150–200 g/L
Simple to prepare Just add water
Faster hydrating ≤90 min
Flexible in its use Add less water to concentrate
Fluid granulation
Solidifying
Solid bridge
Moistening
Liquid bridge AGT granule
Finished granule Spraying
Trace liquids Powder
10
Increased Titer with Highly Concentrated Feeds
30–40% titer increase by eliminating the dilutional gap
0
20
40
60
80
100
120
Traditional Dilution Maximumworking volume
Available space
Feeding supplement
Media working volume
~20% improvement in
bioreactor use
3x
3x
71 L
29 L 19 L
10 L
71 L 88 L
12 L
mAb A, EFC (45%)
mAb B, EFC (45%)
mAb A, 1X EFC+ (45%)
mAb B, 1X EFC+ (45%)
mAb A, 2X EFC+ (22.5%)
mAb B, 2X EFC+ (22.5%)
Culture time (days)
6 8 10 12 14 16 0
500
1000
1500
2000
2500
3000
An
tib
od
y t
ite
r (m
g/L
)
Vo
lum
e (
L)
11
Clone Ranking Varies with Process
0
200
400
600
800
1000
1200
1400
1600
1800
34F05 54A05 02B07 05B09 05D09 25F10 05C07 60D06 55F02 05A03
Simple Fed-batch
Fed-batch 1
Fed-batch 2
Top 10 MabB Clones in SFB & FB
Clone ID
Tit
er
(mg
/L)
12
Product Quality is Important
Critical quality attribute:
“A physical, chemical, biological or microbiological property or characteristic that should be within an appropriate
limit, range or distribution to ensure the desired product quality”*
* ICH, ICH Harmonized Tripartite Guideline Q8: Pharmaceutical Development, Step 4 version (August 2009)
Examples of critical quality attributes of biotherapeutics:
• Glycosylation
• Charge variants
• N- and C- terminal structural features
• Fragmentation
• Aggregation
13
Glycosylation is a Key Critical Quality Attribute
Glycosylation can affect:
• Protein folding
• Solubility
• Immunogenicity
• Enzymatic activity
• Transport
• Turnover rate
Activity and availability of the protein are impacted by glycans
NA2G1F
NA2FA2F
A1F
NGA2F
A1
NA2G1
Man-9
Man-8
Man-7
Man-5
Man-6
NA2G1F
NA2FA2F
A1F
NGA2F
A1
NA2G1
Man-9
Man-8
Man-7
Man-5
Man-6
14
Productivity Increase While Maintaining Quality Profile
No change in key glycan structures
0%
10%
20%
30%
40%
50%
60%
70%
80%
CD OptiCHO medium + 30% EFA CD OptiCHO medium + 30% EFA
+ 10% FMax
Percent glycosylation at harvest
G0F G1F G2F G0 M5
CD OptiCHO medium
+ 30% EFA
0 50 100 150 200 250
0
50%
100%
150%
200%
250%
Pro
tein
tit
er
Integral viable cell density
0
50%
100%
150%
200%
250%
Pro
tein
tit
er
Culture duration (days)
EFA = Gibco™ CHO CD EfficientFeed™ A Supplement and Fmax = Gibco™ FunctionMAX™ Titer Enhancer
CD OptiCHO medium
+30% EFA
+10% FMax
15
Addressing Media and Process Impact on Glycosylation*
• pH
• DO, pCO2
• Temperature
• Low glucose/glutamine
• Mn, NH3, glycerol, nucleotide sugar content
• Cell viability at harvest
• Shear stress
• Perfusion vs. fed batch
* From review by Hossler P, Khattak SF and Li KJ, Glycobiology, 19(9):936-949, 2009
16
Preventing Ammonia Accumulation Galactosylation
0
10
20
30
40
50
60
70
80
90
100
Condition 1 Condition 2 Condition 3
G0F
G1F
G2F
G0
M5
Day 14: NH4+: 9 mM Day 14: NH4+: 6 mM Day 14: NH4+: 4.5 mM
Pe
rce
nt
(%)
17
Controlled Conditions in Bioreactors Galactosylation
0
10
20
30
40
50
60
70
80
90
Bioreactor Shake flask
G0F
G1F
G2F
G0
G1
M5
Pe
rce
nt
(%)
18
Lower Harvest Viability G0F, G1F
0
10
20
30
40
50
60
70
80
90
Day 14 Day 19
G0F
G1F
G2F
G0
G1
M5
Pe
rce
nt
(%)
95% Viability 45% Viability
Harvest day
19
CHO Transcriptome Comparison
Glycosylation differences between cell lines linked with transcriptome profiles
Increased transcripts in DG44: ALG2, ALG3, ALG9 and ALG12, ALG8 and ALG10, ALG14, and B4GALT5
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Day 5 Day 7 Day 11 Day 14
G0F
G1F
G2F
G0
G1
M5
CHO-S
0%
10%
20%
30%
40%
50%
60%
70%
80%
Day 5 Day 7 Day 11 Day 14
G0F
G1F
G2F
G0
G1
M5
DG44 CHO
Higher terminal galactose in DG44
20
Media/Feed Composition Influences Glycosylation
0
10
20
30
40
50
60
70
80
90
G0F
G1F
G2F
G0
G1
M5
Pe
rce
nt
(%)
21
Rational Glycan Modulation
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Hu
nd
red
s
G0F
G1F
G2F
G0
M5
% G
lyc
osyla
tio
n
22
Additives Only Provided Partial Modulation of Glycans
An additive approach alone did not provide the ability to target a desired glycan profile
0
10
20
30
40
50
60
70
80
90
100
A1F/MAN5 G0F G1F_1 G1F_2 G2F
control
0.03X
0.17X
0.36X
0.68X
0.87X
1.00X
% R
ela
tive
qu
an
tity
23
Another Approach Was Tried Using Gibco™ GlycanTune™ C+ Total Feed
0
2
4
6
8
10
12
14
16
18
VC
D (
x1
06 c
ell
s/m
l)
0 2 4 6 8 10 12 14 16
Culture time (days)
A
Glucose control
20% 2X EFC+ Control
15% 3X GTC+ Transition day 15
15% 3X GTC+ Transition day 13
15% 3X GTC+ Transition day 11
15% 3X GTC+ Transition day 9
15% 3X GTC+ Transition day 7
15% 3X GTC+ Transition day 5
15% 3X GTC+ Transition day 4
B
4 6 8 10 12 14 16 0
20%
40%
60%
80%
100%
120%
Culture time (days)
IgG
tit
er
(% o
f c
on
tro
l)
24
Another Approach Was Tried Using GlycanTune C+ Total Feed
The timing of transition from Gibco™ EfficientFeed™ C+ (EFC+) to GlycanTune C+ (GTC+)
makes it possible to target specific glycosylation profiles.
0
10
20
30
40
50
60
70
80
EFC+Control
Day 15 Day 13 Day 11 Day 9 Day 7 Day 5 Day 4
G0F
G1F_1
G1F_2
G2F
Mann5/A1F
Day of transition from EFC+ to GTC+
% R
ela
tive
qu
an
tity
25
Results Solutions
Case Study: Modulating Glycan Profile
Problem
Starting point:
Customer was satisfied with yield
but required product quality
modulation to match the innovator
molecule (4/12 glycan attributes
within an acceptable range) In-house
program
Analyze
product
quality data
DoE using
in-house
IP/literature
Evaluate in
ambr 15 HT
system
Client
analyzes PQ
Optimization
SA2
SA2F
SA1
SA1F
M5
G0
G0F
G1
G1Fa
G1Fb
G2
G2f
PQ improvement
Outcome: 9 of 12 attributes within range,
remaining 3 within 80% of target.
26
Current Glycan Analysis Solutions
• Single channel (CE & LC) low-throughput (days/96 samples) separation
• Poor data quality of high-throughput methods
• Labor intensive workflow (pipetting, centrifugation, etc.)
• Use of toxic sodium cyanoborohydride chemistry
• Use of vacuum centrifugation steps
• Generic software taking long analysis time
• Non integrated solution from multiple vendors
• Commercial kits with high cost per sample
• No validation support for result/solution
Kit supplier Instrument
supplier
Generic
software
Non integrated
solution
27
Key Features Desired for a High-Throughput Glycan Analysis Platform
Easy sample prep
• Magnetic bead-based sample prep
• Hands-on time <3 hrs for 96 samples
• Fewer pipetting steps
• No sodium cyanoborohydride use
• No vacuum centrifugation steps
High throughput
• Sample prep and data of 96 samples in 7–9 hrs
Resolution
• Sialylated glycans
• Structural isomers
• Fucose species
• High-mannose species
Dye labeling
• Multiple dyes with superior sensitivity
High sensitivity
• Analyze as little as 1 µg of IgG
Low cost
• Robust instrument and capillaries with low running cost
Fully integrated
• Full sample prep, hardware, and software solution
28
Data analysis
Fit for use software
Robust alignment
algorithms
Data collection
Multi-capillary
Run-to-run reproducibility
Cap-to-cap concordance
Dye labeling
Multiple dyes
Reactivity
Sensitivity
Resolution
No dry down
Glycan purification
Reproducible & quantitative
(Magnetic beads)
Novel CE-based HT Glycan Analysis System
Glycan release
Fast <1 hr IgG inputs
(1.6–50 µg)
Complete sample prep and data of 96 samples in 7–9 hours
29
Different Dyes for Better Specificity
Separation of major glycans was possible
APTS
Turquoise™
Teal™
30
Conclusions
• Improved volumetric productivity (Qp) can be achieved by use of rationally
designed, matched basal and feed media to sustain specific productivity (qp)
• A granulated feed format enables highly concentrated feeds which can
further improve productivity and ease of use
• Product quality can be impacted by choice of media, feed, and
process conditions
• High-throughput glycan analysis by capillary electrophoresis simplifies and
improves the workflow
Future challenges:
• Methods for custom optimization of other critical quality attributes, including
charge variant and aggregation profiles
• Rapid, high-throughput, and sensitive methods for analysis of other
critical quality attributes
31
Acknowledgments
• Ryan Boniface
• Michael Gillmeister
• Jaime Goldfuss
• Shawn Barrett
• Mark Stramaglia
• Serena Smith
• Doan Tran
• Baburaj Kunnummal
• Shaheer Khan
• JenKuei Liu
The world leader in serving science
Caution: For use as a raw material in further manufacturing applications
© 2016 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
For more information on glycosylation, visit thermofisher.com/GlycanTune
The world leader in serving science
Appendix
34
Another Approach Was Tried Using Gibco™ GlycanTune™ A+ Total Feed
B
18
A
0
2
4
6
8
10
12
14
16
VC
D (
x1
06 c
ells
/mL
)
0 2 4 6 8 10 12 14 16
Culture time (days)
18
Culture time (days)
0
20%
40%
60%
80%
100%
120%
IgG
tit
er
(% o
f c
on
tro
l)
0 2 4 6 8 10 12 14 16
Glucose control
20% 2X EFC+ control
15% 3X GTC+ transition day 15
15% 3X GTC+ transition day 13
15% 3X GTC+ transition day 11
15% 3X GTC+ transition day 9
15% 3X GTC+ transition day 7
15% 3X GTC+ transition day 5
15% 3X GTC+ transition day 4
A. The growth profiles for the DG44 cell in Gibco™ CD OptiCHO™ medium were similar
with both EFA+ and with the use of or when transitioning to GTA+.
B. DG44 titer comparison between feeding conditions. 100% titer is the titer of 15% 3X
EFA+ on day 16. Titer results indicate that the use of and transition to GTA+ does not
negatively affect protein production.
0
10
20
30
40
50
60
70
80
EFA+Control
Day 15 Day 13 Day 11 Day 9 Day 7 Day 5 Day 4
G0F
G1F_1
G1F_2
G2F
Mann5/A1F
Day of transition from EFA+ to GTA+
% R
ela
tiv
e q
uan
tity
The timing of transition from Gibco™ EfficientFeed™ A+ (EFA+) to GlycanTune A+ (GTA+) makes
it possible to target specific glycosylation profiles. Modulating G0F from 72% down to 30%,
while increasing G1F (1 and 2) and increasing G2F.
35
Another Approach Tried with Gibco™ GlycanTune™ B+ Total Feed
Also works with other matched basal/feed media
Culture time (days)
Culture time (days)
B
A
A. The growth profiles for the DG44 cell in CD CHO medium were similar with both EFB+ and when transitioning
to GTB+. The large error bar at day 10 was due to clumped cells.
B. DG44 titer comparison between feeding conditions. 100% titer is the titer of 15% 3X EFB+ on day 16. Titer
results indicate that the use of and transition to GTB+ does not negatively affect protein production.
Glucose control
20% 2X EFC+ control
15% 3X GTC+ transition day 15
15% 3X GTC+ transition day 13
15% 3X GTC+ transition day 11
15% 3X GTC+ transition day 9
15% 3X GTC+ transition day 7
15% 3X GTC+ transition day 5
15% 3X GTC+ transition day 4
0 2 4 6 8 10 12 14 16
4 6 8 10 12 14 16
IgG
tit
er
(% o
f c
on
tro
l)
VC
D (
x1
06 c
ells
/mL
)
0
2
4
6
8
10
12
14
16
18
0
20%
40%
60%
80%
100%
120%
0
10
20
30
40
50
60
70
80
90
EFA+Control
Day 15 Day 13 Day 11 Day 9 Day 7 Day 5 Day 4
G0F
G1F_1
G1F_2
G2F
Mann5/A1F
Day of transition from EFA+ to GTA+
% R
ela
tiv
e q
uan
tity
The timing of transition from Gibco™ EfficientFeed™ B+ (EFB+) to GlycanTune B+ (GTB+) makes it
possible to target specific glycosylation profiles. Modulating G0F from 80% down to 41%, while
increasing G1F (1 and 2) and increasing G2F.