Beyond Measuring Affinity:

Biacore to take Center Stage in Orthogonal Analytics for Comparability Exercises and Quality

Tasks of Biologics.

Biacore QC Workshop

Christian Maasch, Xendo

Managing Consultant/Team Lead Drug Development and RA Biologics

Bringing scientists together...

Vienna, Nov 23nd , 2017

Confucius (Chinese philosopher) said …

2

The route/way becomes the destination

… this is also true for the development of using Biacore …

Failure Is Not an Option -- It's a Requirement

2013 first Biacore paper about PK and Biomarker analysis published

1998

Retrospective Look of “Working w Biacore”

3

Christian´s personal time line

2003 first Biacore paper published

2009 Director of analytical department

2010 presented using Biacore as QC/ potency assay

2011 presented using biophysical assayas stand-alone batch release assay

2011 DiPIA Boston presented concept of bioanalytics by Biacore

• Finally we are here to meet and talk about regulatory compliant QC/ potency assay and applications beyond

• Thank you for the chance to be part of this Biacore-family for more than nearly two decades and to spread these (once) “weird ideas” for using Biacore“beyond measuring affinity”.

2000 purchased first “own” Biacore

1998 started working w Biacore at MPI Berlin

2017

4

Lessons learns by using Biacore for

more than 15 years …

5

The Protagonist and What It Measures: Biacore

Measures:

Binding rates constants ka and kd of an analyteto a sensor chip immobilized target

Data output:

Binding association rate constant ka

Binding dissociation rate constant kd

Binding affinity (KD)

Stoichiometry or number of binding sites (n)

Enthalpy ΔH

Entropy ΔS

Concentration (c) of unknown analyte

Biacore/ SPR

-60

-40

-20

0

20

40

60

80

100

-100 0 100 200 300 400 500 600

Tim e s

Re

sp

. D

iff.

RU

Spiegelmer

binding to target

ka [1/M*s] 2.72E+06 ± 2.10E+04

kd [1/s] 2.86E-03 ± 1.35E-05

KD [M] 1.05e-9

kakd

BC 3000

BC T200

Biacore sensorgram

Information-rich data, but…

Know and understand your assay; do the right controls !

from Drug Discovery to Clinical Development

Ph.II: Patients

Drug

Discovery

Preclinical

Discovery

Preclinical

Development Manufacturing

and QC

Ph.I: Healthy

volunteers

Clinical

Development

Use of Biacore Methods in Drug Discovery at

NOXXON

7

D40D32

D30

D09

D17D16

association rate log ka (M-1

s-1

)

dis

so

cia

tio

n r

ate

lo

g k

d (

s-1

)

5.8 5.9 6.0 6.1 6.2 6.3

-3.2

-3.1

-3.0

-2.9

-2.8

-2.7

-2.6

-2.5

faster association

slo

wer

dis

socia

tion

NOX-D19001-6xDNA

kc

al/m

ole

-60

-40

-20

0

20

40

G

H

-TS

Hit identification/ binding kinetics and affinity

Lead optimizationor screening

Thermodynamics/ transition-state

Selectivity/ specificity Mode-of-Action (MoA) and structure-function relationship

Biacore substantially supported the discovery of the “Spiegelmer” therapeutics

Proven Value of the Biacore Technology for

Drug Discovery

8…. but we think there is much more potential….

Accepted technology (see below recent publication of 2 articles back-to-back on April 22, 2015)

Positive Experience with Biacore …

9

Well developed biophysical/ Biacore assays …

Can we transfer our positive experiences to preclinical and clinical R&D ?

…are robust and reproducible

...have a high precision and accuracy

22 repeated analysis

time [s]

resp

on

se u

nit

s [

RU

]

0 200 400 600

0

100

200

300% theoretical content of active Spiegelmer

Biacore

100% 99.69±0.21

98.04% 98.18±0.15

96.15% 96.82±0.26

94.34% 93.88±0.12

92.60% 92.26±0.20

90.90% 90.82±0.18

80.00% 79.28±0.21

10

…, but what I also learned …

Paul Belcher on Linkedin/Twitter:

“… I see Senorgrams everywhere”

11

Paul Belcher Functional Leader BiacoreGE Healthcare, US

Paul Belcher on Linkedin/Twitter

12

“… following on from my blog posts about data fitting andwhen to believe the data or not - I saw this data in apublication this week. Discussion...3...2...1 and GO”

Paul Belcher on Linkedin/Twitter

13

“… following on from my blog posts about data fitting andwhen to believe the data or not - I saw this data in apublication this week. Discussion...3...2...1 and GO”

“… we might have to define a new unit.....10e-24 is Yoctomolar” Paul B.

“It is always worth to use bothkinetic and steady state fittingsas a self-QC. And 900 binging RUwithout saturation.... scary. “Shuo W.

“I suspect those aren't rate constants,but rather the author is referring toequilibrium constants KA and KD. Thatwould agree with the numbers anyway.Still does not appear to be 0.5 pMinteraction.” Michael M.

“Association constant is10e12 M... what a fastbinding! Is it super glue?”Secil F.

“I'm going to guess at some of the other numbersgenerated by the fit: Rmax <2 RU. RI up to 1000RU. Even if this was a steady-state fit, they stillhaven't approached saturation closely enough togenerate plausible KD values.” Eric R.

“The numbers given are obviously not the rateconstants ka and kd but the equilibriumassociation and dissociation constants, asKa=1/Kd. Still the dimensionality of Ka should beM^-1, not M. Otherwise with on and off ratesbeing too fast to be measured telling trueresponse from bulk RI change is tricky if, as in thiscase, the system is far from saturation. Showingthe response vs. concentration plot is the leastthe referee should have asked for.” Andrei Z.

BBP – “Best Biacore Practice”

14

Know and understand your assay; do the right controls

Describe your assay setup and evaluation/kinetic model with respect to the MoA of your analyte (show the fit!)

Critical evaluate your data; potentially look for a “Biacoresparring partner” to discuss results

Learn “reading your sensorgram data” without software

Reading, “Dissection” and Interpretation of

Senorgram Data

15

… start “reading sensorgrams” instead of “blind trusting” the software …

Sensorgram taken from: Humanization and Characterization of an Anti-Ricin Neutralization Monoclonal Antibody, Sep 2012 · PLoS ONE

KD = 1.63 nM

Baseline: • Slope? Drift?• Memory effect?• Chip integrity?• Delay in injection? IFC or flow to slow?

Association Phase :• Saturation?• Rmax used for fitting;

calculation of ka?• Curvature sufficient?• Biphasic association

(fast/slow) ->Non-specific binding?

• Low signal for mAb@500 nM, must be saturated at lower concentrations when KD=1.63 nM

• Surface activity and control?• Initial on-rate: is it concave?

MTL?

Regeneration:• Efficient?• Monitor chip decay! Or memory effect

Evaluation:• What model?, fit and goodness of

fit not shown!• Why KD ~1.63 nM, but there is no

saturation at higher concentrations

• No standard dev (but, 3 digits for rate constants ;))

Dissociation Phase• Off-rates of differ!• NSB or bivalent or heterogeneous

ligand?

Assay setup: • Concentration range

tested feasible for estimated affinity?

16

Towards Validation of Biacore Assays

Lessons Learned by ELISA Assays, Cell-based Potency

Assays and HPLC Analysis

• Standard Curve/ “Internal” Reference Material• Calibrators/QC sample analysed in duplicates/

triplicates to monitor assay performance• LOD/ LLOQ/ ULOQ• Evaluation of sample of unknown

concentration: standard curve was calculatedby fitting a 4-parameter logistic function (4-PL-function) to the raw data of the calibrator andwas used to calculate the concentration of theunknown samples thereof.

• Evaluation of relative “binding potency” of known sample by dose-response curve: e.g. by parallel-line analysis/ IC50

ELISA and cell-based assays are steady-state analysis: How to transfer to a Biacore analysis?

Biacore Evaluation:

Steady-State vs. Kinetics – do they fit?

18

KD steady state : 0.87µMKD kinetics: 0.91µM

• In an ideal world and under optimalassay conditions steady stateanalysis and kinetics fit well

-50

50

150

250

350

450

-200 0 200 400 600 800 1000

Tim e s

Re

sp.

Dif

f.

RU

Awful Fit, Chi2 = 740KD ~ 46.6 nM (kinetics) vs 481 nM (steady-state)

Steady state?

• In a real world parameters like e.g.non-specific binding, heterogeneousbinding, assay limitations, bad datafitting, … may influence that steady-state analysis does NOT fit to kineticdata

From Report Points to Standard Curve

19

time [s]

res

po

ns

e u

nit

s [

RU

]

0 200 400 600-200

0

200

400

600

Report points• Standardized and “easy-to-evaluate/transfer”

data to analysis software, e.g. Prism, PLA 3.0 or SoftMax

• Fit data to a non-linear 4/5-parameter fit

Assay Set up and Development : Report Points and Standard Curve

concentration analyte [nM]

res

po

ns

e u

nit

s [

RU

]

0.01 0.1 1 10 100 1000

0

500

1000

1500

2000

2500

Hill slope = 1.12

IC50 =318.1

Bottom

Top

DETERMINE STANDARD CURVE• Define fitting parameters and confidence

intervals for assays to “pass”

Introducing QC Samples/Calibrators to Your

Biacore assay

20

Conc. ± SD % accuracy %CV

QC 0.5nM 0.582±0.0212 105.5±4.15 3.64

QC 2.0nM 2.221±0.1329 111.1±6.6 5.98

QC 10.0nM 12.40±1.066 124.0±10.66 8.60

QC 50.0nM 55.66±3.031 110.8±7.6 5.44

Accuracy 100% Conc

MeanAcc results

Precision 100% results

results

Mean

SDCV

1

• Role of QC samples is that they represent the matrix of the samples with known amounts of the analyte; QC samples are processed in the same manner as study samples

• Used during method validation to demonstrate e.g. accuracy, precision and stability

• Subsequently used during the conduct of the study to provide batch-level quality control

Biacore Assay Acceptance Criteria:1) A minimum of 75 % of all calibration values in the working range (between the LLOQ and ULOQ) have to show an accuracy

of 80-120 %, except at the LLOQ and ULOQ where limits of 75-125 % are acceptable.2) At least 4 out of 6 QC samples at three concentrations have to show an accuracy of 80-120 % and a precision of ≤20 %. At

least one QC sample per concentration needs to meet this criterion. If this criterion is not fulfilled, the assay is invalid andthe samples must be re-analysed.

Example for Competitive Biacore Assay:

Monitor Assay Performance by Bracketing

21

injection cycle

resp

on

se u

nit

s [

RU

]

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69

0

100

200

300

400standard_start unknown samples and QC samples

no comp

200 nM

standard_end

no competition control

Sequential analysis of samples

initial 3 injections

res

po

ns

e u

nit

s [

RU

]

0

1

2

3

4

1

23

2 x standard curve

concentration analyte [nM]

res

po

ns

e u

nit

s [

RU

]

0.1 1 10 100 1000 100000

100

200

300

400standard curve start

standard curve end

Additional Information by Real-time Measurement:

Kinetics to bring Value and Confidence to your Assay

22

time [s]

res

po

ns

e u

nit

s [

RU

]

0 200 400 600-200

0

200

400

600 kakd

Qualitative and quantitative data in a sensorgram• ka [1/ M*s]: a function of the concentration or/and activity/ integrity of the analyte• kd [1/s]: reflects the binding mode of action (MoA)/identity of the analyte

KNOW YOUR ANALYTE AND ASSAY !• Measure ka and kd and define acceptance criteria

for your analyte

ka

ka [

1/M

*s]

1.0106

1.2106

1.4106

1.6106

1.8106

2.0106

kd

kd [

1/s

]

1.010- 4

1.210- 4

1.410- 4

1.610- 4

1.810- 4

2.010- 4

EXAMPLE:

Biacore as Surrogate Potency Assay

23

12.5nM – 6.25 – 3.13 – 1.56 – 0.78 – 0.39 – 0 nM

-10

0

10

20

30

40

50

60

-100 0 100 200 300 400 500 600 700 800

respo

nse u

nits [R

U]

time [s]

double injection of 12.5nM

Read-out report pointsP1, P2 – slope, baseline stabilityP3 – RU after injection time vs concentration (Req?!)P4, P5 – regeneration efficiency

Biacore fingerprint: -> sensorgram comparison tool;Sum of evidence for quality of data

P1 P2 P3

Read-out binding kineticska – association rate constantkd (1) – early dissociation rate constant kd (2) – late dissociation rate constant P4

regeneration

P5

ka kd (1) kd (2) Read-out data fittingApply global fitting model !!!RI – bulk or unspecific binding contributionChi2 – parameter for “goodness of fit”, ruol ≤ 1(-5)% of Rmax

Adjust kt value if appropriate

Intra assay controls/assay performance• 6 QC samples (4 in linear range)• accuracy of 80-120 % and a precision of ≤20 %.

Biacore Assay KPIsEstablish ranges (e.g. CI) - pass or fail?

24

KPIs = Key Performance Indicators (kinetic analysis)

• Know AND understand how your assay setup/parameters (and their correlation) affect the performance of the assay (DoE/QbD approaches)

• Establish reliable ranges and statistics for the definition of acceptance criteria

QC1 QC2 QC3 QC4 QC5 QC6 QC1 QC2 QC3 QC4 QC5 QC6

* Relative “binding potency”: KDrel = mean KD samples/ KD internal standard

Further Biacore KPIs

25

• Acceptance criteria for baseline, blanks/matrices, non-functional analytecontrols…

• Acceptance criteria regeneration efficiency• Is curvature for appreciate number of concentrations established?• Rmax consistency vs injected concentrations -> monitor NBS contribution• Double injections -> chip integrity• monitored Rmax, Req (local) -> validity of kinetic data• …

Giordana Bruni is going to talk about that topic and share her experience

Example: Clinical IMP Batch Release

26

Cell-based Assay

Relative potency = 76.25 %

Comparable results between biophysical and cell-based assay format

Spiegelmer (nM)

flu

ore

scen

ce

(rf

u)

0.01 0.1 1 100

10000

20000

30000

40000

50000

reference IC50 = 0.4159 nM

IMP_7 batch IC50 = 0.5454 nM

inter- and intra-assay variations

% a

cti

vit

y o

f re

fere

nc

en=1 n=2

0

50

100

150

% IMP activity ofreference batch

% a

cti

vit

y o

f re

fere

nc

e

IMP_7 batch

0

50

100

150

Biophysical: Biacore

p= 0.3415, ns

Relative binding „potency“ = 78.49 % ± 4.867 (SD)CI95% of mean = 75.40 – 81.58%

Why a Biophysical Assay ?

27

Technically less challenging

No batch-to-batch variation as in animal bioassays

No limitations of handling cells in cell-based bioassays

Higher precision and sensitivity

Not subject to inherent variability of bioassays (e.g. passage number)

Assay-to-assay consistency of critical reagents

Reagents may be frozen and used over a long period of time

Combines robustness with high sensitivity, precision, accuracy and low sample consumption

Regulatory Perspective:

Biacore Beyond Lead Generation?

28

new guidance document for analytical method validation (02/2014)

What´s essentially new?

Section III on 'Analytical Methods Development' Emphasis on getting the method right at the beginning with respect to specificity, linearity, limits of detection (LOD) and quantitation limits (LOQ), range, accuracy, and precision and method robustness

Encourages to find (new) analytical methods with best analytical performance characteristics

Development Research IND Enabling Clinical Post-market

IND e.g. BLAIND = investigational new drug; BLA = biologics license application

Regulatory requirements

RA Compliance:

Biophysical Assay as Surrogate Potency Assay?

29

According to ICH Guideline Q6B, a biological assay to measure thebiological activity of the product may be replaced by biophysicaltests only in those instances where:

sufficient mode-of-action (MoA) information about the drug…

and relevant correlation to biologic activity can be demonstrated…

and there exists a well-established manufacturing history…

Animal-based Cell culture-based Biophysical ???

Accepted as potency assay

Correlation of Binding Affinity to Biologic Activity?

30Kd (nM;biophysical)

IC5

0 (

nM

; c

ell

-ba

se

d a

ss

ay

)

0.1 1 10

0.1

1

10slope = 0.9052 0.1675

NOX-A12

NOX-D20

NOX-L41 NOX-E36

NOX-H94 NOX-G16

NOX-S93

STHKRTG a )ln(

R

S

TR

HKa

1)ln(

Short reminder / tutorial – binding and thermodynamics belong togethervan’t Hoff analysis:

Excellent correlation of binding affinity and biological activity/

potency for oligonucleotide-based scaffolds

Temperature matters ! – measure at 37°C !

Long-term Experince of Using a Biophysical

(Surrogate) Potency Assay

31 Competitive binding assay, low intra- and inter-assay variations, higher precision and accuracy than cell-based assay and linear correlation with biological potency assay

… used since 2010 as stand-alone surrogate potency assay for batch release

1 10 100 10002.0

2.5

3.0

3.5NOX-H94

80/20

50/50

IC50

NOX-H94

22.57

80/20

~ 33.21

50/50

~ 2.768e-012

SPM (nM)+20nM HEP

Ferr

itin

(n

g/µ

g P

rote

in)

1 10 100 10002

3

4

5

6NOX-H94

80/20

50/50

IC50

NOX-H94

13.12

80/20

13.78

50/50

29.98

SPM (nM)+20nM HEP

Ferr

itin

(n

g/µ

g P

rote

in)

1 10 100 10002.0

2.5

3.0

3.5

4.0NOX-H94

80/20

50/50

IC50

NOX-H94

26.27

80/20

23.55

50/50

56.52

SPM (nM)+20nM HEP

Ferr

itin

(n

g/µ

g P

rote

in)

week 1 week 2 week 3

Cel

l-b

ased

ass

ay

0.1 1 10 100 1000 100000

20

40

60

80

100

120

nM SPM

Bin

din

g o

f D

IG-l

ab

ele

d N

OX

-H9

4

to

bio

-HE

P(%

)

0.1 1 10 100 1000 100000

20

40

60

80

100

120

nM SPM

Bin

din

g o

f D

IG-l

ab

ele

d N

OX

-H9

4

to

bio

-HE

P(%

)

0.1 1 10 100 1000 100000

20

40

60

80

100

120

nM SPM

Bin

din

g o

f D

IG-l

ab

ele

d N

OX

-H9

4

to

bio

-HE

P(%

)

week 1 week 2 week 3

50/50

80/20

NOX-H94

50/50

80/20

NOX-H94

50/50

80/20

NOX-H94

Bio

ph

ysic

al a

ssay

32

Further tasks using Biacore in CMC

Biacore Assays used to Support CMC for …

QC/ batch release

IPC = In process control/ QbD support (e.g. AMBR studies= rapidly screen and optimize cell culture processparameters) -> identification of CQAs/ comparability studyfrom process performance to product quality attributes

Formulation development/Stability tests (e.g. total vs active DS)

Comparability studies

AP

I/D

S

Surrogate potency assay (clinical batch release)

Formulation development/Stability tests

QC/ re-analysis/ in-process controls

IMP/

DP

New process Establish process and analytics Establish ranges/ QbD Identify CQAs

Biosimilars … coming back to Confucius

34

The route/way becomes the destination… this is especially true for the development of Biosimilars

Biosimilar

QTPP Biosimilar-> GOAL: Innovator QTPP in mind !!!

Biological/ Innovator drug

QTPP = Quality Target Product Profile

Innovator drug challenge:QTPP not known/available

Are we opening Pandora´s Box by applying new processes and technologies?

The CMC Challenge of a Biological

35

The very nature of a biologic means It is practically impossible for two different manufacturers to produce two

identical biopharmaceuticals if identical host expression systems, processesand equivalent technologies are not used

This has to be demonstrated in an extensive comparability programfor biosimilars

Biosimilars do have a few Developments in their Favor

Technological advances to characterize biologic molecules and theanalytics used to evaluate and prove biosimilarity.

Improvements are being made in the manufacturing techniques used toproduce biosimilars.

Structure Function Relationship – Orthogonal

Approaches

36

• Understanding the structure-function relationship of a biotech product isvital for its successful development.

• Awareness of this is important, and by selecting the right assays at theright time, it is possible to reduce significantly the costs and risks of theR&D process.

• By applying orthogonal methods and approaches for characterization,information regarding the structural, binding interactions and functionalproperties of the biological product are linked – providing a greaterunderstanding and interpretation of the determined characteristics.

PotencyPurity

Structure

Effector function MoA In vivo, eg. PK/PD/Immunogenicity

Additional Tasks for Biacore in Analytics

37

Affinity andBinding kinetics

Concentration/Purity(total vs active)

Surrogate Potency Assay

IPCs/ process development/CQA

Identity/Integrity

Stability

Structure-Functionrelationship

Receptor binding assays

Formulation

(Accelerated) stability studies

Off-Target binding

Immunogenicity

Example: Structure-Function

Relationship

38

Orthogonal Use of Binding Analysis (SPR) for Structure-Function Relationship

Target binding: Affinity, rate constants and selectivity, specificity by SPRAdvantage: High accuracy, precision and sensitivity, low inter- and intra

assay variations

Purity: HPLC/ SEC data linked to total and active DS/DPAdvantage: Linking the information of total to active DS/DP and refer to

potency

Identity/integrity: LC-MS, Disulphide bridge, DSC analysis linked to SPR rate constants as “fingerprint” and prediction of MoA

Advantage: Confidence in structure-function relationship

Glycosylation: Abundance of a glycosylation species was linked to effector receptor binding as determined by SPR

Advantage: Bridging glycosylation pattern and effector function by descriptive (binding) MoA

39

Potency: Cell-based assay linked to SPR assayAdvantage: Only One Cell-Based Assay acceptable to demonstrate

bioactivity

Effector function: Binding of FcgRs linked to cellular cytotoxity (ADCC)Advantage: Linking effector function as determined by cell-based assays to

descriptive (binding) MoA

FcRn Binding: Used as total sum of evidence for identify/ integrity and linked to FcRn binding assays as prediction of pharmacokinetic properties

Stability: see SPR analytics above

Example: Structure-Function

Relationship

Stuart Knowling (BioOutsource) and Mike Willcox (Eurofins) have already shared their experience and insights

Additional Information from Sensorgrams linked to

other Analytics

40

CONFIDENCE IN DATA AND ”PIVOTAL” EVIDENCE OF ANALYTICAL SIMILARITYby best practise analytics and orthogonal approaches

AGENCIES are happy

REDUCTION OF ANALYTICAL EFFORTSdriving the reduction of costs and risks

SPONSORs are happy

Where Do We Go? –

The Challenge of Crude Sample Analysis

41

PharmacokineticsWhat is the half-life t1/2 of a drug in the blood ?

Pharmacodynamics/ BiomarkersHow long is the drug effective ?

Bio-distribution/ADMEWhere is the drug distributed in the body and where is it metabolized or cleared ?

ImmunogenicityDoes the drug induced anti-drug antibodies (ADAs) ?

The Problem of Plasma Binding and Cut-

Point Analysis

42

Binding does not equal functional/ biological significance

Unspecific binding of biological matrices reduces the assay’s detection limit

Unspecific binding to various surfaces differs

From: Biacore Immunogenicity Package Manual

-50

0

50

100

150

200

-50 0 50 100 150 200 250 300 350 400

Tim e s

Re

sp

. D

iff.

RU

Report point 1

Report point 2

Rat plasma binding todextran Biacore surface

res

po

ns

e u

nit

s [

RU

]

report point 1 report point 20

50

100

150

200

0%

0.0039%

0.0078%

0.0156%

0.0313%

0.0625%

0.125%

0.25%

0.5%

1%

The Real Challenge: Individual Cut-points!

43

Unspecific binding of sera from healthy volunteers to an immobilized chemokine

Reducing unspecific binding is mandatory for further assay development

res

po

ns

e u

nit

s [

RU

]

healthy volunteers0

1000

2000

3000

4000

5000

Unspecific binding of 250 individual sera from healthy volunteers differs dramatically

Definition of a cut-point is challenging in terms of the assay sensitivity (more likely it is impossible)

The only way out: pre-dose vs treated normalization of data

What if there is no pre-dose sample available?

• Unspecific binding wasreduced to a negligiblelevel at NOXXON Pharmaby optimized bufferconditions and the use ofunspecific competitors(e.g. soluble dextran)

• Methods established forthe analysis of plasma,serum, urine and issuehomogenates

Example: Compound (Scaffold)

Pharmocokinetic/ Pharmacodynamic In Rats

44

0´ 10´ 1h 3h 8h 24h 48h 96h

Blood sampling

Organ Dissection

Biacore analysisPlasma pharmacokineticTissue distributionTarget/ Biomarker plasma levels

Compound (Spiegelmer): L-nucleic acid linked to PEG

Biomarker

Overview of Assay Performances for

Bioanalytics by Biacore

45

Assay L-Nucleic Acid Qualification PEG Qualification Biomarker Quantification

Assay type Direct binding Biacore assay Competitive Biacore Assay Competitive Biacore Assay

Matrices Plasma, liver, urine Plasma, liver, urine Plasma

Sample preparation None; simple dilution None; simple dilution

On-target effects No influence; up to 1µM tested No influence; up to 1µM tested No influence; up to 1µM tested

Biological matrices No influence; (up to 50%) No influence; (up to 30%) No influence; (up to 50%)

In-use stability of chip > 500 analysis cycles > 350 analysis cycles > 310 analysis cycles

LLOQ ≤ 0.2 nM ≤ 1.95 nM 0.2 nM

Precision %CV < 3 %CV < 8 % CV < 7

% Accuracy 98 -103% 93 -110% 97 -104%

Dilution linearity % Recovery

Yes> 98%

Yes > 96%

Yes>93%

Time per sample 9 min 7 min 7 min

concentration Spiegelmer [nM]

res

po

ns

e u

nit

s [

RU

]

1 10 100 1000 10000 100000

0

500

1000

1500

2000

2500

Hill slope = 1.12

Spiegelmer

Spiegelmer + 1µM target

Spiegelmer concentration [nM]

% b

ind

ing

of c

on

tro

l

0.1 1 10 100 1000 100000

20

40

60

80

100

IC50 = 121.1 nM

Hill slope = -0.78

4-PL fitted standard curve

concentration of biomarker/ target [nM]

re

sp

on

se

un

its

[R

U]

0.01 0.1 1 10 100 10000

50

100

150

200

250 target

target + 1µM Spiegelmer

no comp

Hill slope = -1.040/ -1.025

Questions Answered by Biacore

Bio-Analytics

46

sampling time [h]

Sp

ieg

elm

er

pla

sm

a c

on

c. [

nM

]

0 24 48 72 96

100

1000

10000

100000

quantified by oligonucleotide

quantified by PEG moiety

Plasma pharmacokinetics

How long is the compoundin the blood?

Target/ Biomarker pharmacodynamics

What is the efficacy of the compound?

sampling time [h]

Bio

ma

rke

r in

pla

sm

a [

nM

]

SD

0 24 48 72 96

0

200

400

600

800

1000

1200

1400

release rate = 40.81 2.916 nM/ hour

Cmax = 890.1 273.8 nM at 48 hours

Tissue distribution

How is the compound cleared from the body?

Renal clearance

% o

f to

tal d

os

e

SD

24 h 96h0

2

4

6

liver

% o

f to

tal d

os

e

SD

24 h 96h0

2

4

6

kidney

% o

f to

tal d

os

e

SD

0-24 h 24-48h0

1

2

3

4

Bound-free Analysis

When do I have to dose again for the best effect?

sampling time [h]

pla

sm

a c

once

ntr

atio

n[n

M]

0 24 48 72 96

0.1

1

10

100

1000

10000

100000

target

Spiegelmer

sampling time [h]

% a

cti

ve

of

co

ntr

ol

0 24 48 72 96

0

50

100

Summary/ Take-Home Messages

Biacore/ biophysical assays have the potential to support the leadidentification and pharmaceutical development of biologics and otherscaffolds from hit to clinical trials

Biophysical analyses with their high accuracy and precision - compared tobiological assays - increasingly gain importance as surrogate potency, QC andCMC assays

The most critical issue to generate highly sensitive Biacore data withbiological samples was overcome by intensive buffer optimization

For therapeutic scaffolds Biacore could be used for a variety of analyticalcomparability and QC tasks

Biacore 8k

Acknowlegdement

Axel VaterSandra MahrSven KlussmannLukas BethgeSimone SellWerner PurschkeJohannes HoosKlaus BuchnerDirk ZboralskiLisa BauerKathrin SchindeleGabi AnlaufDörte VossmeyerStefan ZöllnerStefan VonhoffFrank SchwöbelDirk EulbergAnna KruschinskiMatthias Baumann

Xenia Freifrau von MaltzanXiaoxi ZhuHarm Hermsen

GermanyAnja Drescher

Uwe Bierfeund

Uwe Roder

Thomas Von Kaenel

U.S.Brian Lang

Michael Murphy

Paul Belcher

JapanMami Okudaira

Discussions, help and inspiration: Biacore colleagues & friendsMichael Schraeml, Roche, Penzberg

Christian Stegmann, Bayer Healthcare, Berlin

Amaury Fernandez, Bayer Healthcare, Berlin

Jörg Bomke, Merck KGaA, Frankfurt

Stefan Geschwindner, Astra-Zeneca, Sweden

GlobalFredrik Sundberg

SwedenRobert Karlsson

Markku Hämäläinen

Olof Karlsson

Liselotte Molander

I still love to see great new ideas, data and also weird results

…and thank you for your attention

IndiaVinutha Devendran

U.K.Tim Fagge

ChinaHelen He

Abstract and Contact

49

Beyond Measuring Affinity: Biacore to take center stage in orthogonal analytics for comparability exercisesand quality tasks of biologics.

Christian Maasch, Xendo, Berlin, GermanyUnderstanding the structure-function relationship of a biotech product is vital for its successful development. Awareness of this is important, and byselecting the right assays at the right time, it is possible to reduce significantly the costs and risks of the R&D process. The diversity of therapeutic biologicsand their complexity not only poses a challenge for robust manufacturing, but also comprehensive characterization of these molecules.

SurfacePlasmonResonance (SPR)-based methods (Biacore) have shown to combine robustness with high sensitivity, precision, and accuracy andsubstantially assisted the lead identification and characterization of biological and chemical compounds over the last decade. By recent SPR methoddevelopments, these assays gained increasing importance and acceptance to support pharmaceutical QC, CMC, bioanalytics and comparability studies.

By applying orthogonal methods and approaches for characterization, information regarding the structural, binding interactions and functional propertiesof the biological product are linked – providing a greater understanding and interpretation of the determined characteristics.

Herein we show examples and insights on how SPR with information-rich data could support the analysis of biologics in terms of Critical Quality Attributes(CQAs) and drug release testing, as well as for comparability exercises for the development of biosimilars.

Christian Maasch, XendoManaging Consultant Drug Development & RASiemensdamm 6213627 Berlintel +49 (0) 30 856 068 78 – 24mob +49 (0) 1511 0348793 email christian.maasch@gmail.comemail christian.maasch@xendo.com