1. No category

Biosimilars: Case studies
on how to hit the target
Joerg Windisch, Ph.D.
Chief Science Officer
Sandoz Biopharmaceuticals
DIA EuroMeeting
Amsterdam, March 6, 2013
25th Annual
EuroMeeting
4-6 March 2013
RAI, Amsterdam
Netherlands
The views and opinions expressed in the following PowerPoint slides are
those of the individual presenter and should not be attributed to Drug
Information Association, Inc. (“DIA”), its directors, officers, employees,
volunteers, members, chapters, councils, Special Interest Area
Communities or affiliates, or any organization with which the presenter is
employed or affiliated.
These PowerPoint slides are the intellectual property of the individual
presenter and are protected under the copyright laws of the United States of
America and other countries. Used by permission. All rights reserved. Drug
Information Association, DIA and DIA logo are registered trademarks or
trademarks of Drug Information Association Inc. All other trademarks are
the property of their respective owners.
Drug Information Association
www.diahome.org
2
Biosimilars must be systematically engineered to match
the reference product
2. Target
directed
development
Biological variability
Clinical
Recombinant cell line development
PK/PD
3.
Confirmation
of
biosimilarity
Process
development
Bioprocess development
Preclinical
Purification process
development
Biological
characterization
Analytics
Drug product
development
Physicochemical
characterization
Target range
Reference
product variability
1. Target definition
Drug Information Association
www.diahome.org
3
Structure: High resolution, orthogonality and redundancy in
analytical characterization provide full understanding
Attributes:













Primary structure
Mass
Disulfide bridging
Free cysteines
Thioether bridging
Higher order
structure
N- and C-terminal
heterogeneity
Glycosylation
(isoforms, sialic
acids, NGNA,
fucosylation, alpha
gal, site specific)
Glycation
Fragmentation
Oxidation
Deamidation
Aggregation
Drug Information Association
Proteins can be well characterized at least
up to the complexity of monoclonal
antibodies
 Primary structure determined from recombinant
DNA sequence and fully accessible to analytical
verification
 Set of orthogonal analytical methods available to
characterize the identity and amount of related
variants with high sensitivity
 Glycosylation profile can be comprehensively
determined with regard to identity and content of
individual glycans with high sensitivity
 Accurate and relevant bioassays for pivotal
biological functions available
www.diahome.org
Methods e.g.:
 MS (ESI, MALDITOF/TOF, MS/MS)
 Peptide mapping
 Ellman‘s
 CGE
 SDS-PAGE
 CD
 H-D exchange
 FT-IR
 HPLC
 HPAEC
 IEF
 2AB NP-HPLC
 SE-HPLC
 FFF
 AUC
 DLS
 MALLS
4
Function: Broad range of sensitive biological, biophysical
and immunological assays provides full understanding
Attribute
Functiona Assay
l element
Binding to
target
Fab
Formats
Binding assay Cellular binding assay
Surface plasmon resonance
(Biacore®)
Programme Fab
d cell death
Apoptosis
assay
Cell based apoptosis assay
CDC
Fab and
Fc
CDC assay
Cell based CDC assay
Fc
C1q binding
Surface plasmon resonance
(Biacore®)
Fab and
Fc
ADCC assay
Cell based ADCC assay
Fc
FcgR binding
Surface plasmon resonance
Fc
FcRn binding Surface plasmon resonance
ADCC
PK
Drug Information Association
www.diahome.org
5
Understanding the target: Variability is significant in
reference biologics – antibody example
2,0
 Monitoring batches of
an approved mAb
revealed a shift in
quality
Post-Shift
Unfucosylated G0
[% of glycans]
1,6
1,2
Pre-Shift
 Shift in glycosylation
(structure) pattern
results in different
potency in cell-based
assays (function)
0,8
0,4
0,0
08.2007
12.2008
05.2010
09.2011
Expiry Date
140
ADCC Potency
[% of reference]
 Indication of a change
in the manufacturing
process
Post-Shift
120
Schiestl, M. et al., Nature
Biotechnology 29, 310-312, 2011)
100
80
 Products found to be
equally safe and
effective post-shift by
regulators (EMA, FDA)
Pre-Shift
60
08.2007
12.2008
05.2010
09.2011
Expiry Date
Drug Information Association
 Such shifts observed
in several original
products
www.diahome.org
6
Defining the target: Variability in reference biologic
defines very narrow goal posts for biosimilarity
ADCC (%of Reference)
700
600
500
400
300
200
100
0
0
2
4
6
8
bG0-F [rel. %]
Variability of
reference product
Variability observed during
cell line development
Very narrow
goalposts for
biosimilar
Biologically possible variability
Understanding the criticality of the quality attributes is
essential
 Risk assessment for ranking and prioritizing quality attributes
 General concept described in A-MAb case study (Tool #1)
Criticality Score = f(Impact,Uncertainty)
e.g.: Criticality Score = Impact x Uncertainty (A-MAb)
Range
Impact
Uncertainty
Criticality Score
Known or potential consequences
on safety and efficacy, considering:
• Biological activity
• PK/PD
• Immunogenicity
• Safety (Toxicity)
Relevance of information
e.g.
literature
prior knowledge
in vitro
preclinical
clinical
or combination of information
Quantitative measure for an
attribute‘s impact on safety and
efficacy.
2 (very low) – 20 (very high)
1 (very low) – 7 (very high)
Using best possible surrogates
for clinical safety and efficacy
2 - 140
Accumulated experience, relevant information, data
e.g. literature, prior & platform knowledge, preclinical and clinical batches,in vitro studies, structure-function
relationships
Drug Information Association
www.diahome.org
8
Quality attributes can be influenced at all stages of
cell line and process development...
1
Cell line
■ Host cell line.
■ Transfection/amplification pool
■ Genetic “set up” of production cell line (clone).
Quality
2
Process
■ Growth medium composition.
■ Culture conditions (pH, T, aeration,...)
■ USP type (batch, fed batch, perfusion,...)
■ USP regime (duration, fed type,perfusion rate..)
■ Culture history (genetic stability, process stability..)
■ Individual DSP steps
■ Hold times
■ Storage (buffer, container, conditions..)
Drug Information Association
www.diahome.org
9
...but it is the biology that decides if the molecule will
make the grade
Variability of product quality attributes at project start
Screening of host cell lines
Genetics
Screening of transfection pools
Screening of clones
Physiology
Media
development
Bioprocess
development
Chemistry
DSP
dev.
Target spec
Drug Information Association
www.diahome.org
10
Cell line development: An elaborate, target-directed
multi-stage process
Clones
Clones
Clones
Selected
Cell lines
Clone
Hundreds
Thousands
96/24/6
Shake flask
well plates
Drug Information Association
Hundreds
www.diahome.org
Tens
Process Dev
.
Pools
Pools
One
Lab-scale bioreactor
11
Cell line development: Multifactorial selection of best
clone aided by software tool focuses on quality
% glycan structure
39
38
39
39
41
43
63
52
37
Pool
44
60
K33/p17
68
3.9 2.2
K32/p17
66
4.2
K40/p17
64
K48/p17
65
K42/p17
64
4.0 4.6 4.4
bG1(-N)
Man 8
K36/p17
36
1.7
K53/p17
62
40
K30/p17
3.8
Man5
bG2S1
K29/p17
0.9 1.2 1.3
K28/p17
63
3.8
K24/p17
63
4.2
1.4
bG0
unk8
K23/p17
Gmap GP2017 clones (SF500 FB screening)
40
3.0
K7/p17
K11/p17
K15/p17
80%
39
K3/p17
46
K126/p26
36
bG0 (-F)
Man 7
K122/p26
49
K106/p26
36
K84/p26
K104/p26
45
K59/p26
38
bG0 (-N)
bG2
K53/p26
46
K7/p26
K52/p26
44
Man 4
unk6
K1/p26
K22/p26
47
K75/p17
44
K72/p17
62
K69/p17
100%
47
K64/p17
35
bG0 (-N-F)
unk5
K59/p17
35
K38/p17
60%
46
2.3
K37/p17
2.0 2.3 1.9 1.6 1.8 1.8 3.3 2.2 2.4 2.9
0.9 1.7
2.2 2.5 3.1 2.5 2.8 2.8 3.1 0.9
Unk group
Man 6
K25/p17
K31/p17
63
1.1 1.5
K22/p17
66
5.2
K13/p17
K17/p17
71
0.6 3.1
Man3*
bG1 (1-3)
K9/p17
40%
20%
0%
GP2017
K56/p17
58
5.4
bG1(-F)
K60/p17
50
49
Similarity
Safety
Productivity
Unique
Optimization
P17
SSF3
2.40
-0.01
0.17
0.02
0.49
0.56
0.53
2
P29
K1-PD
2.27
0.26
0.78
-0.14
-0.09
0.44
0.06
3
P28
K1-PD
1.47
-0.04
0.79
0.02
-0.12
0.12
0.05
4
5
P30
P56
Cell_line
K1-PD
HPT2
MTX
MTX
1.43
1.29
0.00
0.38
0.78
-0.23
Efficacy
0.01
0.59
-0.10
-0.06
0.01
0.03
0.05
0.50
6
7
P56
P57
HPT2
HPT2
MTX
1.11
1.10
0.18
0.27
-0.07
-0.16
0.02
0.56
-0.10
0.17
0.90
-0.40
8
9
P26
P6
K1-PD
K1
MTX
1.04
0.91
-0.09
0.32
0.78
-0.24
0.02
0.14
-0.05
0.11
-0.34
-0.02
0.07
0.40
10
11
12
P34
P33
P43
K1
K1
SSF3
MTX
MTX
0.78
0.69
0.68
0.08
0.14
0.47
0.74
-0.02
-0.03
-0.12
1.00
-0.35
-0.13
0.20
-0.13
-0.22
-0.97
0.10
-0.27
0.01
0.31
13
14
15
16
17
18
19
20
P55
P17
P20
P28
P43
P27
P29
P61
HPT2
SSF3
SSF3
K1-PD
SSF3
K1-PD
K1-PD
HPT2
MTX
MTX
MTX
MTX
MTX
MTX
MTX
0.65
0.57
0.52
0.49
0.47
0.40
0.32
0.31
0.13
-0.17
0.33
0.10
0.15
0.04
-0.10
0.48
-0.06
0.02
-0.36
-0.10
0.17
-0.09
-0.02
-0.82
-0.11
-0.17
-0.42
0.26
-0.25
0.32
0.01
-0.12
-0.06
0.56
0.16
0.27
-0.08
0.27
0.26
0.18
0.62
0.06
0.58
-0.29
0.39
-0.34
-0.02
0.36
0.14
-0.14
0.08
-0.02
-0.15
0.00
0.07
0.36
50
1.1 1.5 1.4
K79/p26
60
K80/p17
K93/p17
68
3.0
bG1 (1-6)
4.0
K77/p17
K92/p17
66
K62/p17
2.0
All unk peaks
K55/p17
Total Score
1
QUALITY, growth,
productivity, stability
CQA risk assessment
Drug Information Association
www.diahome.org
12
0.14
0.39
Cell line development: Multiple selection rounds required
to hit the target („evolution in the test tube“)
Individual Value Plot of Total Score
5.0
2.5
Pool 7
Total Score
0.0
Pool 13
-2.5
Selected clone and backup
clone for further process
development
-5.0
-7.5
-10.0
-12.5
1. Pools 50 mL
Drug Information Association
2. Clones 120 mL
3. Clones 1 L
www.diahome.org
4. Clones 5 L
5. Originator
13
Cell line development case study: Minor glycan structures
and ADCC bioactivity – attention to detail is essential...
Characterization of mAB glycosylation heterogeneity
High resolution identification and
quantification of major
(G0,G1,G2) and minor glycan
structures
(down to a level of 0.1 rel.%)
2x
Targeting ADCC activity and fucosylation by clone selection
10
600
8
500
bG0(-F) [%]
ADCC (%of Reference)
700
400
300
200
6
4
2
100
0
0
0
2
4
6
8
Originators
bG0-F [rel. %]
Drug Information Association
www.diahome.org
Parental
Cells
Pool 18
Pool 16
Clone 19
14
... and products that do not fulfill these standards are not
biosimilars and will not be approved by EMA
 Higher host cell protein content
 Content of aggregates not
comparable
 Charge distribution not comparable
 Glycosylation not comparable
 ADCC effector function not
comparable
 Clinical data: Only PK/PD study in 17
patients
Source: Mike Doherty, Global Head Regulatory Affairs, Roche Pharmaceuticals, at
Roche Investor Day 2010, March 18, 2010, see
http://www.roche.com/investors/ir_agenda/rid_2010.htm?track=8 and
www.roche.com/irp100318_md.pdf
Drug Information Association
www.diahome.org
15
Media development: Strategy is product specific, builds
on what is known, and evolves as more is learned
 Literature research
 Use know-how from related projects
 Study of metabolic pathways – e.g. evaluation
of enzyme cofactors, substrates, by-products
 Spent medium analysis
 DoE with product quality profiling over
cultivation time
 Application of metabolomics
Drug Information Association
www.diahome.org
16
Media development case study:
Getting rid of potentially immunogenic glycans
 Alpha-galactosylation (Gal-α1,3-Gal) of glycan structures bG2 and
bG2(-F) might potentially evoke some immunogenic reactions
(Chung et al, 2008; Macher, 2008)
 Aim of process development: increase bG2 (reduce unfucosylated
component), keep bG2_Gal at minimum
 Only component A was capable of increasing bG2 without
also increasing alpha-Gal
Drug Information Association
www.diahome.org
17
Bioprocess development case study:
Getting rid of potentially immunogenic glycans
Screening experiment (fractional factorial design:
Agitation, temperature, pH, pO2)
Optimization experiment (full factorial design:
pH, temperature)
 Lower levels of alpha-galactosylation
Drug Information Association
www.diahome.org
18
Raw materials are critically important
Inoculum cultivation in three different
lots of a commercial medium
1,4E+06
100
Main-stage process performed
with three different base powder
lots of an in-house medium
90
1,2E+06
70
60
8,0E+05
50
6,0E+05
40
--- viability [%]
Galactosylation
30
4,0E+05
20
2,0E+05
10
0,0E+00
0
0
1
2
3
4
5
6
7
8
9
10
duration [days]
Supplementation of „bad“ lot with Fe
[%]
VCD [cells/ml]
80
1,0E+06
Lot A_1
Drug Information Association
www.diahome.org
Lot A_2
Lot B_1
Lot B_2
Lot C_1
Lot C_2
19
Downstream process development case study:
Systematically removing unwanted components
Aggregate removal by Ion Exchange Chromatography (IEC)
Prior
Knowledge
Prior knowledge included in risk
assessment and set-up of the screening
design study (28-4 – design)

3 main input parameters statistically
significant for aggregate removal

Response surface model providing good
model quality and predictive power
(R2 = 0.95; Q2 = 0.89)

Verification in a challenge study with
aggregate spike under worst case
conditions
Drug Information Association
Response
surface
model
Screening
design
Condition A
www.diahome.org
Challenge
study
Condition B Condition C
20
How to evaluate biosimilarity at the quality level
mAb case study 1
Drug Information Association
www.diahome.org
21
How to evaluate biosimilarity at the quality level
mAb case study 2
Drug Information Association
www.diahome.org
22
Fingerprint: All methods can contribute to demonstrating
similarity beyond the main components
Primary structure e.g.:
LC-MS intact mass
LC-MS subunits
Peptide mapping
Higher order structure e.g.:
NMR
CD spectroscopy
FT-IR
Impurities e.g.:
CEX, cIEF acidic/basic variants
LC glycation
Peptide mapping deamidation,
oxidation, mutation, glycation
SEC/FFF/AUC aggregation
Post translat. modif. e.g.:
NP-HPLC-(MS) N-glycans
AEX N-glycans
MALDI-TOF N-glycans
HPAEC-PAD N-glycans
MALDI-TOF O-glycans
HPAEC-PAD sialic acids
RP-HPLC sialic acids
Biological activity e.g.:
Binding assay
ADCC assay
CDC assay
Combination of attributes e.g.:
MVDA, mathematical algorithms
Drug Information Association
www.diahome.org
23
Clinical trials provide reassurance, but analytics is much
more sensitive in detecting differences
Clinical trial design
Bioanalytics
 Traditional clinical trials cannot even
differentiate between completely
different anti-TNFs at week 24 (all with
similar response)
 Can determine overlapping function
justifying smaller clinical trial focused on
“similarity”
ACR20 Response Rate [%]
 Not useful to document biosimilarity
 Bioanalytics easily differentiates
between molecules
Kaymakcalen, et al: Clinical Immunology, (2009)
131, 308-316
Drug Information Association
www.diahome.org
24
Biosimilars expand patient access already today
UK example: Access to G-CSF
UK G-CSF volume growth
Percent change vs. previous year
 G-CSF prevents hospital readmission due to infection
Sept. 2008
biosimilars
approved
by EMA
2007
2008
 More physicians start G-CSF
treatment earlier (primary
prophylaxis) due to more affordable
cost
2009
2010
2011
Source: IMS, NHS
Drug Information Association
Biosimilar C-CSF is not yet approved in the US.
www.diahome.org
25
Conclusions
Today, biologics can be thoroughly characterized and
understood both structurally and functionally
The goal posts are set by the variability in the
reference product and CQA understanding
Knowledge of how process parameters influence
product attributes is used to achieve matching quality
Analytics is more sensitive in detecting differences
than clinical studies
Drug Information Association
www.diahome.org
26