March 24 - 25, 2014 Sheraton Silver Spring Hotel Silver ......David Lansky, Precision Bioassays, Inc., USA Tsai-Lien Lin, CBER, FDA, USA Thomas Anders Millward, Novartis Pharma AG,

March 24 - 25, 2014

Sheraton Silver Spring Hotel

Silver Spring, Maryland USA

Organized by:

Welcome to Bioassays 2014: Scientific Approaches & Regulatory Strategies

On behalf of the Scientific Organizing Committee and CASSS, we are excited to welcome you to

Bioassays 2014: Scientific Approaches & Regulatory Strategies and look forward to your participation

and input March 24 - 25, 2014 at the Sheraton Hotel in Silver Spring, Maryland.

"Bioassays" has established itself as a premier conference and unique opportunity for participants and

opinion leaders to discuss and debate current regulatory and industry topics regarding bioassays.

Bioassays are a critical component of the analytical control strategies for biologics and other complex

molecules. The ability of an assay to characterize and demonstrate biological activity is essential and

developing such bioassays is becoming more difficult as biologic drugs are engineered to be more

complex and/or have multiple modes of action. Companies are continuously challenged with developing

assays that are biologically relevant for the analysis of these different mechanisms. Bioassays are also

used for lot release, stability, comparability and characterization studies, which requires that the assays

be robust and, in most cases, suitable for a QC lab.

Bioassays 2014 is structured to encourage attendee interaction. Each session includes case study

presentations followed by a panel discussion allowing for lively dialogue between attendees from

academia, industry and regulatory agencies. As in previous years, we expect this format to result in

additional focus on the technical and regulatory details of the topic. Regulatory participation from the

US FDA, Health Canada and various European agencies has grown each year. In addition, an exhibitor

showcase and poster reception at the end of Day One will give attendees the opportunity to present

additional topics and continue the day's discussion in an informal setting.

We are sure you will find Bioassays 2014 to be informative and productive, and that it will provide you

with current perspectives on bioassays.

Program Co-Chairs: Chana Fuchs, CDER, FDA, USA

Denise Gavin, CBER, FDA, USA

Helena Madden, Biogen Idec, USA

Bruce Meiklejohn, Eli Lilly and Company, USA

Scientific Organizing Committee: Evangelos Bakopanos, Health Canada, Canada

Katrin Buss, BfArM, Federal Institute for Drugs and Medical Devices, Germany

Hélène Gazzano-Santoro, Genentech, a Member of the Roche Group, USA

Xu-Rong Jiang, MedImmune, USA

Thomas Anders Millward, Novartis Pharma AG, Switzerland

Noel Rieder, Amgen Inc., USA

Sally Seaver, Seaver Associates LLC, USA

The Scientific Organizing Committee gratefully acknowledges the program partners for

their generous support of Bioassays 2014.

SUSTAINING PLATINUM PROGRAM PARTNER

Biogen Idec

PROGRAM PARTNERS

Eli Lilly and Company

Genentech, a Member of the Roche

Group

MedImmune

EXHIBITOR PROGRAM PARTNERS

Eurofins Lancaster Laboratories

Promega Corporation

Stegmann Systems

MEDIA PROGRAM PARTNERS

BioPharm International

BioProcess International

BioProcessing Journal

BioTech International

Genetic Engineering & Biotechnology

News

IPQ Publications

LCGC North America

Technology Networks Limited

The Analytical Scientist

Wiley/Journal of Separation Science

Bioassays 2014: Scientific Approaches & Regulatory Strategies

Scientific Program Summary

Monday, March 24, 2014

07:30 – 17:00 Registration in the Cypress Foyer

07:30 – 08:30 Continental Breakfast in the Magnolia Ballroom

08:30 – 08:45 CASSS Welcome and Introductory Comments in the Cypress Ballroom

Bruce Meiklejohn, Eli Lilly and Company, Indianapolis, IN USA

Bioassays 2014 Welcome and Introductory Comments in the Cypress

Ballroom


Bioassays Lessons Learned: Part One

Workshop Session One in the Cypress Ballroom

Session Chairs: Bruce Meiklejohn, Eli Lilly and Company and Noel Rieder, Amgen Inc.

08:45 – 08:50 Introduction

08:50 – 09:15 Issues to Consider When Developing Potency Assays for Biologic Products

Baolin Zhang, CDER, FDA, Bethesda, MD USA

09:15 – 09:40 Managing Acceptance Criteria Throughout the Development Lifecycle

Shea Watrin, Amgen Inc., Wellsville, UT USA

09:40 – 10:05 A Holistic Systems Approach to Controlling Bioassay: Lessons Learned

Bhavin Parekh, Eli Lilly and Company, Indianapolis, IN USA

10:05 – 10:30 Lessons Learned: Choice of Potency Assay and Differential Sensitivity to

Degradation Pathways Kirby Steger, Bristol-Myers Squibb Company, Princeton, NJ USA

10:30 – 11:00 AM Break – Visit the Exhibits and Posters in the Magnolia Ballroom

11:00 – 12:15 PANEL DISCUSSION – Questions and Answers

Katrin Buss, BfArM, Germany


Bhavin Parekh, Eli Lilly and Company, USA

Kirby Steger, Bristol-Myers Squibb Company, USA

Shea Watrin, Amgen Inc., USA

Baolin Zhang, CDER, FDA, USA

Monday, March 24 continued…

12:15 – 13:30 Hosted Lunch in the Magnolia Ballroom

Bioassays Lessons Learned: Part Two Workshop Session Two in the Cypress Ballroom

Session Chairs: Hélène Gazzano-Santoro, Genentech, a Member of the Roche Group and Thomas

Anders Millward, Novartis Pharma AG


13:35 – 14:00 Two-in-One: A Novel Approach of Bioassay Selection for Dual Specificity

Antibodies

Guoying Jiang, Genentech, a Member of the Roche Group, South San Francisco,

CA USA

14:00 – 14:25 Challenges and Strategies in Selecting MOAs-reflective Bioassays for

Bispecific Antibody Xianzhi Zhou, MedImmune, Gaithersburg, MD USA

14:25 - 14:50 Challenges in the Development of Potency Assays for ADCs and their Utility

to Detect Conjugate Variants Sonia Connaughton, ImmunoGen, Inc., Waltham, MA USA

14:50 - 15:15 Development of an Alternative, in-vitro Potency Assay for Rabies Virus

Vaccines Robin Levis, CBER, FDA, Rockville, MD USA

15:15 - 15:45 PM Break – Visit the Exhibits and Posters in the Magnolia Ballroom

15:45 - 17:00 PANEL DISCUSSION – Questions and Answers

Evangelos Bakopanos, Health Canada, Canada

Sonia Connaughton, ImmunoGen, Inc., USA

Chana Fuchs, CDER, FDA, USA

Guoying Jiang, Genentech, a Member of the Roche Group, USA

Robin Levis, CBER, FDA, USA

Xianzhi Zhou, MedImmune, USA

Exhibitor Partner Showcase in the Cypress Ballroom

Session Chairs: Xu-Rong Jiang, MedImmune and Sally Seaver, Seaver Associates LLC


17:30 – 17:45 Effective cGMP Bioassay Outsourcing

Weihong Wang, Eurofins Lancaster Laboratories, Lancaster, PA USA

Alexander Knorre, BSL BIOSERVICE Scientific Laboratories GmbH,

Planegg/Munich, Germany

Monday, March 24 continued…

17:45 – 18:00 New Capabilities of PLA

Ralf Stegmann, Stegmann Systems, Rodgau, Hesse, Germany

18:00 – 18:15 Bioluminescent Technologies for Biological Functional Analysis and Protein-

Protein Interactions

Mei Cong, Promega Corporation, Madison, WI USA

18:15 – 19:45 Exhibitor and Poster Reception in the Magnolia Ballroom

19:45 Adjourn Day One

Tuesday, March 25, 2014

07:30 – 17:00 Registration in the Cypress Foyer

07:45 – 08:45 Continental Breakfast in the Magnolia Ballroom

Bioassays to Support Biopharmaceutical Development

Workshop Session Three in the Cypress Ballroom

Session Chairs: Katrin Buss, BfArM, Federal Institute for Drugs and Medical Devices and Helena

Madden, Biogen Idec


08:50 – 09:15 Implementation of the Next Generation Effector Function Assays for

Comparability Assessments Poonam Aggarwal, Pfizer, Inc., Chesterfield, MO USA

09:15 – 09:40 The Dual Benefit of Structure Function Studies: Better Understanding of

Molecules and Help with MOA-relevant Bioassays

Carl Co, Biogen Idec, Cambridge, MA USA

09:40 – 10:05 Standards and Beyond: Challenges of Application of Old Methods to Next

Generation Products

Elena Semenova, Protein Sciences Corporation, Meriden, CT USA

10:05 – 10:30 Global Implementation of Bioassays – Things to Consider


10:30 - 11:00 AM Break – Visit the Exhibits and Posters in the Magnolia Ballroom


Poonam Aggarwaal, Pfizer, Inc., USA

Carl Co, Biogen Idec, USA



Elena Semenova, Protein Sciences Corporation, USA

12:15 – 13:30 Hosted Lunch in the Magnolia Ballroom

Tuesday, March 25 continued…

Bioassay Controls & Control Strategies

Workshop Session Four in the Cypress Ballroom

Session Chairs: Evangelos Bakopanos, Health Canada and Sally Seaver, Seaver Associates LLC


13:35 – 14:00 Assay Acceptance Criteria for Multiwell-Plate-Based Biological Potency

Assays

C. Jane Robinson, National Institute for Biological Standards and Control

(NIBSC), Hertfordshire, United Kingdom

14:00 – 14:25 “Edging Out” Edge Effects in a Cell-based Assay

Shelley Elvington, Genentech, a Member of the Roche Group, South San

Francisco, CA USA

14:25 – 14:50 Near-universal Similarity Bounds for Bioassays

David Lansky, Precision Bioassay, Inc., Burlington, VT USA

14:50 – 15:15 Health Canada Experiences with Bioassay Controls & Control Strategies

Omar Tounekti, BGTD, Health Canada, Ottawa, ON Canada

15:15 – 15:45 PM Break – Visit the Exhibits and Posters in the Magnolia Ballroom


Shelley Elvington, Genentech, a Member of the Roche Group, USA

David Lansky, Precision Bioassays, Inc., USA

Tsai-Lien Lin, CBER, FDA, USA


C. Jane Robinson, NIBSC, United Kingdom

Omar Tounekti, BGTD, Health Canada, Canada

17:00 – 17:15 Bioassays Workshop Recap

Closing Remarks and Invitation to Bioassays 2015 Helena Madden, Biogen Idec, Cambridge, MA USA

17:15 Adjournment

Bioassays Lessons Learned: Part One Session Abstract

Session Chairs:

Bruce Meiklejohn, Eli Lilly and Company and Noel Rieder, Amgen Inc.

Bioassays represent an essential part of the control strategy for assessing safety and potency of

biopharmaceuticals. There are many potential challenges in bioassay development, implementation, and

maintenance. Avoiding or dealing with these is often an integral part of bioassay development as well as

implementation into routine product testing. This session, which will be divided into two parts, will

present a series of case studies on challenges and successes experienced with bioassays. Each talk will

discuss the specifics of the case study and the key learning’s.

NOTES:

Presenter’s Abstracts

Issues to Consider When Developing Potency Assays for Biologic Products

Baolin Zhang

CDER, FDA, Bethesda, MD USA

A suitable measure of potency is critical to ensure the quality of biologics and other complex products.

To achieve this goal, sponsors are required to develop potency assays that can be used for product

characterization, lot release, in-process and stability testing. Because potency is a product-specific

measurement, potency assays are evaluated on a case-by-case basis. The assay adequacy is assessed by

taking account of multiple factors including, but not limited to, product type, history, mechanism(s) of

action (MoA), associated risk, phases of development, and quality data from physicochemical and

biochemical testing. This presentation provides an overview of regulatory expectations

regarding potency assays and discusses several case studies that highlight some of the relevant issues

commonly seen in the regulatory submissions. Emphasis will be placed on MoA and comparability

studies which represent the most challenging aspects of potency assay development.

NOTES:

3/13/2014

1

Baolin Zhang, Ph.D. Senior Investigator/Product Quality Reviewer

Division of Therapeutic Proteins

Office of Biotechnology Products

Center for Drug Evaluation and Research

CASSS Bioassays 2014: Scientific Approaches & Regulatory Strategies Washington D.C., March 24-25, 2014

Issues to Consider When Developing Potency Assays for Biologic Products

Disclaimers

The views and opinions expressed in this presentation should not be used in place of regulations, published FDA guidances or discussions with the Agency.

Outline

• Regulatory expectations

• Applications

• Issues to consider • Relevance to the mechanism(s) of action (MoA)

• Acceptance limits

• Comparability when making changes

• Case study

3/13/2014

2

Potency (21 CFR 600.3(s))

“the specific ability or capacity of the product, as indicated by appropriate laboratory tests or by adequately controlled clinical data obtained through the administration of the product in the manner intended, to effect a given result.”

Potency Tests (21 CFR 610.10)

“tests for potency shall consist of either in vitro or in vivo tests, or both, which have been specifically designed for each product so as to indicate its potency in a manner adequate to satisfy the interpretation of potency given by the definition in § 600.3(s) of this chapter.”

ICH Q6B

“for complex molecules, the physicochemical information may be extensive but unable to confirm the higher-order structure, which, however, can be inferred from the biological activity”

Examples of complex molecules: •Biological products (e.g. therapeutic proteins, mAbs)

•Mixture products wherein the proportion of “active” ingredients could not be determined by typical physicochemical/biochemical testing

Why a Potency Test?

3/13/2014

3

Complexity of Therapeutic Proteins

• Heterogeneity • post-translational modifications (e.g. glycosylation) • aggregates/degraded products • charged variants • misfolded species • oxidized species • host cell residuals (Host cell proteins, DNA) • leachables (heavy metals, resin)

• Large molecular sizes: 6 – 300 kDa • Higher-order structures (1, 2, 3) • Less defined structure-function relationships

(Compared to small molecule drugs) • Complex manufacturing processes

Regulatory Requirements for Potency of Biologics License

• 21 CFR 601.2 & FDC Act

All biological products regulated under section 351 of the PHS Act must meet prescribed requirements of safety, purity and potency for Biologic License Application (BLA) approval.

• 21 CFR 610.1

“No lot of any licensed product shall be released by the manufacturer prior to the completion of tests for conformity with standards applicable to such product,” which include tests for potency, sterility, purity, and identity (21 CFR Part 610, Subpart B).

•For all phases of IND clinical studies, data are required to assure product

•Identity, quality, purity and strength (21 CFR 312.23(a)(7)

•Stability (21 CFR 312.23(a)(7)(ii)

• ICH Q6B Specifications: Test procedures and acceptance criteria for biotechnological/biological products

Regulatory Expectations for Potency Testing - Investigational Protein Products

3/13/2014

4

Potency Tests: Applications

• Required for characterization, lot release, in-process and stability testing.

• Demonstrate product activity, quality and consistency throughout product development

• Provide a basis for assessing product comparability before and after manufacturing changes

• Evaluate product stability (expiry dating)

• Control clinical dosing consistency

Relationship Between Potency and Clinical Efficacy

21 CFR 314.126(d)

•Potency is a measure of the bioactivity of a drug product that produces a defined clinical effect.

•Potency tests are used to establish that a consistently manufactured product is administered during all phases of a clinical investigation.

•Clinical efficacy is demonstrated by “substantial evidence” from adequate and well-controlled investigations with a consistently manufactured product. Other determining factors include: – PK/PD profile

– Patient population

– Clinical end-point (e.g. overall survival in cancer treatment)

• Cell-based assays • cellular responses - proliferation, growth arrest, cell death, cytokine release • signal transduction - phosphorylation of signaling components • gene transcription (reporter genes) • ligand-receptor binding

• Animal-based assays • eg Lethal Dose 50 (LD50)

• Biochemical assays • eg enzyme activity

• Multiple assays (array matrix) • For products that have complex and/or not fully characterized mechanism of action

Typical Potency Assays

3/13/2014

5

Potency Assays for Products Targeting Cell Death or Cell Growth Pathways

A large number of protein therapeutics function through modulation of cell death or cell growth pathways in the target cells.

• Induce cell death or growth arrest – Cancer – Pain syndromes – Viral infection

• Promote cell growth or cell survival – Wound healing – Organ transplantation – Chronic heart failure – Neurodegenerative diseases

• Readouts: • Cell viability/cell proliferation • Apoptosis • Signal molecules, such as:

- Caspase activation - MAPK phosphorylation - Receptor binding

• Correlation between the readouts and the intended biological activity –Cancer drugs: Cell death vs. growth inhibition

Joslyn Brunelle and Baolin Zhang (2010) Drug Resistance Updates, 13:172-179.

Potency Assays for Products Targeting Cell Death or Cell Growth Pathways (cont’d)

Issues to Consider

• Relevance to MoA • Acceptance limits • Validation • Changes to bioassays

– Comparability

3/13/2014

6

Mechanism of Action (MoA)

21 CFR 201.57©(13)(i) • Clinical Pharmacology section of the labeling, which

states the following: • “This section must contain information relating to

the human clinical pharmacology and actions of the drug in humans.”

• This section must include the following subsections: • Mechanism of action… • Pharmacodynamics… • Pharmacokinetics…

• The MoA should be discussed at various levels, including the cellular, receptor (selectivity), target organ, and the whole body level, depending on what is known.

• Only reasonably well-characterized mechanisms should be described (21 CFR 201.56(a)(2)).

• Speculation on the mechanism of drug action must be avoided (21 CFR 201.56(a)(2)).

• “How Therapeutic and Adverse Effects Occur” – Guidance for Industry (2009): Clinical Pharmacology Section of Labeling for Human

Prescription Drug and Biological Products— Content and Format

MoA: What is it?

MoA in Human Body: A Learning Process in the Product Lifecycle

Relevant disease models

Cell-based assays

Enzyme assays

• A bioassay may not capture all the functional attributes of a product:

– e.g. glycosylation, pegylation, ADCC, CDC, etc.

• Bioassays are performed in combination with physicochemical/biochemical tests to support product quality.

e.g. LD50

1) Phenotypic changes

2) Signal transduction

In vitro enzymatic reactions

http://media.salon.com/2012/11/shutterstock_57341071.jpg

3/13/2014

7

What Should Be Assessed for Potency?

• Relevance to the MoA(s).

– Desired MoA vs. well-characterized MoA

– Assay Matrix - Complex or unknown MoA

• Correlation between the surrogate assay(s) and the biological activity related to potency

- Signal transduction assays

• Ability to discriminate between an active and inactive product or degraded form of the product.

Issues to Consider

• Relevance • Acceptance limits • Validation • Changes to bioassays

• Comparability

Acceptance Criteria for Potency – Biologics License

• A validated potency assay or assay matrix with defined acceptance criteria must be described and justified in the BLA (21 CFR 601.2(a) and 211.165(e))

• “… should reflect the capacity of manufacturing process, and the potency limits established for product lots used in the pivotal clinical studies demonstrating clinical effectiveness” (FDC Act, Section 505(d), 21 U.S.C. 351).

3/13/2014

8

Acceptance Criteria for Potency – Early Phases

• In early development phases, potency acceptance criteria can be difficult to set:

- Limited manufacturing experience - Limited lots of drug substance (DS) and drug product (DP) - Assays not fully validated

• Broad acceptance criteria

– Evaluated with physicochemical and biochemical testing data

• As development proceeds, the acceptance criteria should be tightened to reflect the actual manufacturing capacity, clinical experience and assay performance.

Issues to Consider


• Comparability

Assay Parameters that are Usually Validated

• Robustness to assess sources of variability – reference standards (ICH Q6B) – instruments – reagents (e.g. stable cell lines)

• System suitability • Accuracy • Linearity & Range • Precision (Repeatability, Reproducibility) • Intermediate Precision (analysts, days, laboratories if more

than one will be used)

• Specificity

3/13/2014

9

Stable Cell Lines Used in Potency Assays

• Selection of cell lines – a lineage close to the cell/tissue type targeted by the drug

– a surrogate cell line if an appropriate cell surface receptor is expressed (either endogenously or via stable transfection).

– growth characteristics (e.g. homogeneity, robust growth, functional stability)

• Two-tiered cell banking system (MCB & WCB) – plasmid copy number

– cell passage number

• Standard operating procedures (SOPs) for handling the cell line, including the cell culture conditions, passage number, and procedures for detecting microbial contaminants (e.g. mycoplasma).

Dose Titration Curve

• The dose titration curve should be optimized so that the dilutions are appropriately distributed throughout the entire dose response curve with sufficient coverage in the linear portion of the curve.

• Appropriate statistical analysis (e.g. parallel line analysis) needs to be applied when generating final relative potency results.

• Once a bioassay is validated, it is important to monitor its performance over time. • trending chart for suitable parameters of the ref standard

(RS) response curve and potency of analyzed QC samples

Issues to Consider


• Comparability

3/13/2014

10

Comparability Exercise

• Demonstrate that changes to the potency assay do not interfere with the suitability of the analytical procedures for their intended purposes in terms of

– Accuracy, precision, specificity, detection limit, quantitation limit, linearity, range, robustness

– Pre-defined acceptance criteria

• Same test samples should be assayed using both the original bioassay and the proposed bioassay

– DS, DP, RS, stress and accelerated stability samples

• Appropriate statistical analysis

• In an IND Phase 1 study, a cell-based signal transduction assay (MAPK phosphorylation) was used as a potency assay for a protein product intended to treat Type 2 diabetics.

• Reviewer comment: A potency assay should reflect as much as possible the intended mechanism of action of the drug product. In this case, this would be a measure of improved glucose tolerance or increased glucose uptake in adipocytes. You should provide data demonstrating the correlation between MAP kinase activation and glucose uptake in response to the drug. Alternatively, you may develop an assay that directly measures the uptake of glucose in adipocytes.

Case study # 1

• A recombinant growth factor is tested for treating chronic heart failure. The sponsor developed a potency assay measuring phosphorylation of erbB2 receptor in a cancer cell line.

• Reviewer comment: The potency assay should be optimized to provide a more reliable quantitation of the product’s bioactivity in order to assure consistent dosing. Data should be provided to demonstrate the correlation between phosphorylation of erbB2 and the intended bioactivity of the product i.e. inhibition of cell death of cardiomyocytes.

Case study # 2

3/13/2014

11

• In a Phase 1 IND, a cell viability assay was used as a potency test with a proposed acceptance criteria: IC50 (0.1- 60 nM)

• Reviewer comment: The release specification for potency is not acceptable. The acceptance criterion is too broad to ensure lot-to-lot dosing consistency. From the data provided, it is difficult to assess whether the proposed potency range is wide due to inherent assay variability or whether there are considerable differences in lot-to-lot activity. The potency assay should either be controlled by using a suitable internal reference standard with an allowable range or the manufacturing process must be better controlled.

Case study # 3

• A potency assay was changed from a cell proliferation format to a caspase activation format

– Specification (relative to RS) was not changed

– Better assay precision

– Less plates to meet system suitability

• Validated new assay

– Accuracy (spike recovery), precision (repeatability), intermediate precision, linearity, range and robustness

• Comparability demonstrated by assessing

– Lot release, drug substance, drug product, stability and stressed samples (e.g. heat, light)

Case study # 4

• One applicant developed a cell-based assay to replace animal-based assays for lot release of a licensed protein toxin.

• Issues with the mLD50 assay - highly variable - high rate of assay failure - limited detection range - nonspecific to the product - suffering and death of animals

Case study # 5

3/13/2014

12

mLD50 Bioassay Cell Based Assay Specificity for toxin No Yes, depend on cell lines Sensitivity LD50 1 U ≦ 0.5 U (LD50 equivalent) Range 0.5 – 2.0 U ≦ 0.5 - >100U Precision +/-20-30% <10% Validity 70 – 80 % > 90%

• Comparability demonstrated using a variety of samples (drug substance, drug product, reference standards, stressed samples).

• Improved performance

Case study # 5 (cont’d)

• The potency assays for a licensed recombinant growth factor product have evolved from animal-based assays to cell-based proliferation assay to cell-based gene expression assay.

Case study # 6

Pros Cons

Animal-based assays

• Direct manifestation of MoA • Manifestation of the active

glycosylated forms

• Requires many animals • Highly variable

Culture cell-based assays

• Less laborious • More sensitive • More robust • Fast

• Respond to non-glycosylated product

• Not suitable for measuring drug levels in plasma

Case study #7: Bioassays for complex small molecule drugs

• A bioassay was included as part of the release specifications for a mixture cancer drug, because typical physicochemical/biochemical assays could not determine the proportion of “active” ingredients in the product.

• The approved potency assay uses xenograft tumor mouse models implanted with a murine tumor cell line.

• The assay was used for > 15 years as a release test until recently when tumors failed to grow appropriately upon implantation.

• A cell viability assay using a human cancer cell line is under development:

• Is the selection of such a cell line acceptable?

• Can this new assay be used to replace the animal-based assay for lot release?

3/13/2014

13

Progressive implementation of potency assays

•Limited data on relevant biological attributes

•Broader acceptance criteria

•Release and stability testing

Early phase: pre-clinical

phase 1 phase 2

Later phase: phase 3 pivotal

Biologics License

• A validated potency assay or assay matrix

• Defined acceptance criteria

• Multiple bioassays • Validation • Narrower limits to

ensure lot-to-lot consistency

• Stability testing of validation lots to establish expiry dating prior to licensure

Final thoughts on potency tests

• Potency tests are product-specific, and the adequacy of these assays is evaluated on a case-by-case basis.

• Potency tests may evolve and change significantly in the course of product development and in the product lifecycle.

• It is recommended that sponsors seek timely advice from the FDA on designing, evaluating and validating potency assays.

Acknowledgements

• Serge Beaucage

• Susan Kirshner

• Amy Rosenberg

• Ennan Guan

• Dov Pluznik

Managing Acceptance Criteria Throughout the Development Lifecycle

Shea Watrin

Amgen Inc., Wellsville, UT USA

One of the most significant challenges, when developing a bioassay and shepherding it through to

commercialization, is setting appropriate assay acceptance criteria. Early in development assay

acceptance criteria may be set based on limited experience with the assay, these criteria may be found to

be not ideal as the assay begins to be used by a broader variety of labs, reagent lots, and equipment.

Cases will be presented where criteria need to be adjusted as the assay matures and discuss appropriate

practices.

NOTES:

3/13/2014

1

Managing Acceptance Criteria Throughout the Development Lifecycle

Shea Watrin, Cecilia Chin, Julie TerWee

Amgen Quality

Outline

• Acceptance criteria primer

• Criteria on day 1

• What necessitates change?

• Is it ok to change?

• Case Study

4 Parameter Logistic (4PL) Curve

50

1

left asymptote

slope

right asymptote

b

a dy d

dose

c

a

b

c ED

d

4-Parameter

Logistic

3/13/2014

2

Fundamental Assumption Checked Before Relative Potency Calculation

• Similarity of the test and standard material

• Desire is to show that the only difference between the samples is concentration

• This assumption is checked by evaluating the parallelism or similarity of the dose response curves

Acceptance Criteria

• ‘Parallelism’ of curves. • Do the curves have the same a, b, & d.

• F Test of full vs. reduced models

• Equivalence test for comparisons of a, b, & d.

• Sufficient Assay Response

• Max-to-Min

• Fold Stimulation

• Etc.

• Goodness-of-Fit

• R2

• Residual Sum of Squares

Initial Acceptance Criteria

• Difficult to set without experience

• Need something generic • F-test for parallelism

• Wide limits for Pair Wise Comparisons

3/13/2014

3

Update Criteria

• After a reasonable set of data has been collected • Reasonable: Broad and sufficient experience (Labs, reagents,

equipment, etc.,)

• Limits are setting off inappropriate signals • Good assays fail

• Bad assays pass

Information Missing Early in Development

• Assay development timelines may be so short as to not allow for sufficient experience with all important factors:

• Numerous critical reagent lots

• Variety in analysts

• Variety in labs

• Variety in equipment (plate readers, pipettes, etc.,)

Appropriate to Set New Limits?

• If assays start failing criteria, determine which situation you are in:

• 1. Bad assays are appropriately failing

• Investigate causes of failures and improve

• 2. Good assays are inappropriately failing

• Adjust criteria to allow good assays in the future to pass

• Definition of good is important

3/13/2014

4

What is a good assay?

• Accurate

• Precise

• Ensure that curves are parallel

Set New Limits

• Data in hand • Look at historic data and set limits at edges of experience

• Quantify variability

• Setup studies to capture appropriate variation in factors of concern (Analysts, labs, reagents, etc…)

• Design Space

• Attempt to push criteria to extremes and determine if the quality of results is maintained

Set New Limits

• Calculate limits based on any one of the 3 procedures on previous slide

• Make sure assays are still accurate and precise at or near the proposed criteria

3/13/2014

5

Case Study 1 (AlphaScreen® Binding Assay)

• Max-to-Min limit was set in development lab (Max to Min ≥ 40)

• Assay transferred to new lab

• Shortly after transfer, Max-to-Min failures began to occur

Example Curve

• Curves and Potencies Looked

Good Despite Max-to-Min

Failures

• Max-to-Min = 17.8

• Relative Potency = 0.995

Investigation Findings

• The receiving lab for the method transfer was using a different lot of AlphaScreen® beads

• The beads come from a single source and only one lot was used during the method development

• Determined that results with new lot of beads were still accurate and precise

• A study was set up to determine how low the Max-to-Min could go and still have accurate and precise results

3/13/2014

6

Accuracy and Precision vs. Max-to-Min

• Determined that assay was

accurate and precise with

much lower Max-to-Min values

• Lowered the acceptance

criteria and monitored the

assay

Case Study 2 (Gene Expression Bioassay)

• Assay began to have a few failures and several values right at the limit for Max-to-Min

• Investigation was initiated and hypotheses were tested

Max-to-Min Pre-Investigation

3/13/2014

7

Investigation Findings

• Cell density identified as the most likely cause of low Max-to-Min

• Cell Density was then increased and the data were monitored

Max-to-Min

Relative Potency

Increased Cell Density

Implemented

Higher Ppk = lower probability of Out-of-Specification result

3/13/2014

8

Conclusion

• Acceptance criteria that are created and met during development may not be appropriate through the method lifecycle.

• Key factors influencing acceptance criteria variability including laboratory, analyst, instrument, reagent lot, reagent concentration, etc. should be monitored and evaluated for accuracy and precision.

• Fitness for use of the method must be ensured as changes are made to criteria

A Holistic Systems Approach to Controlling Bioassay: Lessons Learned

Bhavin Parekh

Eli Lilly and Company, Indianapolis, IN USA

Bioassays are one of the most complicated assays implemented as part of the analytical control strategy

for the development of bioproducts. Unfortunately, they can unexpectedly falter when you least expect. I

will present case examples highlighting several areas (eg., assay platform, curve characteristics, assay

design, criticial reagents) that need to be taken into account to control the assay from a holistic

perspective.

NOTES:

1

A Holistic systems approach to controlling Bioassays: Lessons Learned CaSSS-Bioassay 2014 Bhavin Parekh, Ph.D. Group Leader-Bioassay Development Eli Lilly and Company Indianapolis, IN 46221

Outline

Background

Strategic considerations in developing bioassays to support a biotech portfolio

Lesson Learned over the last decade

3 Case studies

Summary

Parekh, CaSSS conf., 2009 Copyright © 2009 Eli Lilly and Company

Why do need bioassays?

Biological therapeutics (proteins, vaccines, Ab, etc.) are complex and heterogeneous in composition. They can exist in multiple physical and chemical conformations.

Even though physiochemical analysis (HPLC, MS, AAA, etc.) can be extensive, it cannot measure activity (or impact of changes in activity) of bioproducts.

Bioassays need to be sensitive to structural changes and/or to product degradation that may impact biological activity, efficacy, or safety

By measuring potency using a bioassay, we can infer the structural integrity of a complex biological, thus bioassays are a measure of the quality of the therapeutic

3/13/2014

2

Case Study 1

Product X is a protein product

Assay ulitizes UMR-106 (osteosarcoma) cell line which expresses the receptor endogenously

Method was validated to support clinical development and post-approval


13-Mar-14 File name/location

Method Description: Cell Induction

Intracellular cAMP released

Product X

binding

UMR-106 cell covered

with Product X receptors

= Product X

= intracellular cAMP produced in

response to Product X

Cell Lysis

cAMP detection


Method Description: cAMP Detection

Assay Plate coated with

goat anti-rabbit antibody

Rabbit anti-cAMP antibody

binds cAMP and cAMP-AP

AP AP

AP

= cAMP

AP AP AP

Cellular cAMP and Alkaline

Phosphatase-conjugated cAMP

added to Assay Plate

= Rabbit anti-cAMP = Goat anti-rabbit

AP AP AP

CSPD/Subs

= Light

Wash

Figure 1. rhPTH(1-34) Dose Response Curve

Product X (nM)

0.001 0.01 0.1 1 10

RLU

's

0

5

10

15

20

25

rhPTH(1-34) Reference Standard

rhPTH(1-34) Unknown Sample

3


Historical Performance: Assay Response

Valid Run Ratio: 8.3 Slope: 1.6 L-term 0.0506

Invalid Run Ratio: 3.2 Slope: 1.4 L-term: 0.3257

0

25000

50000

75000

100000

Y

.01 .006 .1 .07 .04 .02 1 .7 .5 .3 .2 2 3 4 6

Concentration

0

5000

10000

15000

20000

25000

Y

.01 .006 .1 .07 .04 .02 1 .7 .5 .3 .2 2 3 4 5 7

Concentration

Y Median Response S

Median Response UPASS FAIL


Historical Performance: L-term Analysis


Historical Performance: Outcomes

Decline in assay run critical parameters 5-fold decrease in assay response (100K to 20K)

2-fold reduction in S/N (asymptote ratio) from 6.0 to 3.0

Poor quality dose-response curves

High L-term values

Assay valid rate dropped from >90% to <25%

Created significant assay backlog and meet market supply chain demands

4


Root Cause Investigation

Focused on both the analytical method and the cAMP detection (Vendor) kit

Comprehensive technical assessment 8 months in duration

Cross-functional effort (QCL, Bioassay Development, Vendor, TPO Lab)

Every aspect of the assay was assessed

Cells losing response?

Cell health/viability

Pipetting technique, sample preparation, wash steps, blocking, substrate incubation, etc, etc

Root Cause was Determined Conjugate dilution buffer identified as primary cause


Conjugate Buffer Diluent Improvement

0

50000

100000

150000

200000

Y

.01 .006 .1 .07 .04 .02 1 .8 .6 .4 .3 .2 2 3 4 5 7

Concentration

L-term: 0.1035 Slope: 1.7 Asymptote Ratio: 6.1 Potency: 69.80%

.

Vendor conjugate dilution buffer

Lilly conjugate dilution buffer (Mg2+ and Zn2+)

0

5000

10000

15000

20000

25000

Y

.01 .006 .1 .07 .04 .02 1 .7 .5 .3 .2 2 3 4 5 7

Concentration

Invalid Run Ratio: 3.2 Slope: 1.4 L-term: 0.3257


Lesson Learns

Risk 1: Reliance on Vendor-supplied kit Single sourced

Complex biological reagents

Non-GMP supplier

Look at the business model. Are kits designed for discovery/screening efforts

Does “QC tested” meet your standards

Reliability concerns

Lot to lot differences

Timely delivery of kits

Risk 2: Assay complexity Multiple liquid transfer and pipetting steps

Labor intensive

Prone to errors

Sensitive to analyst technique

5


Lessons Learned: Short and long term

improvements

Short-term

Implement new conjugate dilution buffer

Bioassay Development group maintain on-going communication with vendor kit vendor

Long-term

Proactively monitor assay performance and react to trends

Develop new assay based on a different technology


Long term strategy: Outcomes

Developed reporter gene assay in same parental cell line

Demonstrated comparability between cell-based ELISA vs. reporter gene assay

Validated and transferred method to testing lab

Reporter gene assay Significant improvement in assay run parameters

4-fold increase in response level

Enhanced S/N ratio

L-term values consistently below limit (NMT 0.2000)

Acceptable dose-response curves

Assay valid run rate >90%

Completed transfer of assay to CRL

ATP

cAMP G-Protein

X-Receptor Adenylate Cyclase

PKA

Activated PKA

CREB Phospho-CREB

(activated)

CRE-Sequence

(promoter)

CRE-Luciferase Construct

Luciferase mRNA

Luciferase (Protein)

Product X

Case Study 2: Drug Y Bioassay

Reporter gene assay with promoter of validated gene linked to luciferase

Gene is stimulated by a constant concentration of agonist.

Drug Y inhibits the stimulation by the gene and this can be measured in a dose-dependent manner.

Method qualified in QC lab to support early phase CT lot release and stability Accuracy <20%, Precision <10%. Linearity (R2) = >0.95

Stimulating

molecule

Drug Y

Luciferase

6

Testing of DS and DP lots revealed a bias in potency values

%P

ote

ncy

100

120

140

160

Bioassay Results

Avg Potency= 116%± 13%

High Bias and variability not observed with sys. Suit sample that is also testing no each plate

High Bias and variability observed with DS and DP testing (Lot and stability)

Avg Potency= 102%± 8%

DS/DP lot testing Sys. Suit. testing

Bioassay Curve Characteristics for Product Y

0

100000

200000

300000

400000

500000

600000

0.001 0.01 0.1 1 10 100 1000 10000 100000

Re

lati

ve

Lig

ht

Un

its

Conc.

Ref Std

Drug X

5% RLU change = 14%

Potency difference

Potency = 152%

Shallow slope (<1)

Two Linear points

Summary of Assay Improvements and learnings

Difficult to change the dynamics of the biological response , i.e. slope since it may be an inherent biological property

Instead a series of improvements to provide robustness, and reduce variability and bias.

Improvements

• Implement gravimetric preparation of samples to reduce pipetting error associated

with viscosity

• Add more robustness to the curve by adding dilution points in the linear portion

• Implement twice the cell number per well to increase overall signal strength

• Remove Phenol Red from medium to reduce quenching of luminescent signals

7

Experiment Conducted to Determine Effects of

Volumetric vs Gravimetric in Drug Y Bioassay

40

60

80

100

120

140

160

180

200

20 uL Volumetric 100 uL Volumetric Gravimetric

%R

ela

tiv

e P

ote

ncy

Individual Run Bioassay Results by Dilution Method

Reduced Bias and Variability

Avg = 116%±13% Avg = 105%±8%

Company Confidential Copyright © 2013 Eli Lilly and Company

Kamikura 2013

Case 3: Reporter gene assay- Stability of recombinant cell lines

Instability likely mediated by gene silencing

8


Kamikura 2013

Developing stable reporter genes assays

•vector design (eg. bicistronic vectors to link selection to promoter activity, lentivirus) – increasing

transgene stability

• appropriate gene promoter (preferred to transcription factor binding sites) – reduced variability

• Requires baseline expression for selection –open chromatin and maintenance of response

Luciferase Neo/Kan

1.85kb reporter of choice

Ori

promoter of choice

Luciferase Neo/Kan

Vector

1 mRNA

Neo/Kan Luciferase 2 proteins

Signal remains stable over several passages


Passaged 2x week for 16 weeks. Acceptable even in the absence of selection = stable genome integration


Kamikura 2013

Example: Bi-cistronic Vector Stability

x axis

1 10 100

0

50000

100000

150000

System Suitability graph

4-P Fit: y = (A - D)/( 1 + (x/C)^B ) + D: A B C D Rel. Pot.

Plot#1 (Reference Standard: Concentration vs Medi... 1.72e+05 6.29 14.2 2.22e+04 1

Plot#2 (System Suitability: Concentration vs Median... 1.72e+05 6.29 13.8 2.22e+04 1.03__________

Weighting: Fixed

PLA (Std. Curve: Plot#1) Degrees of Freedom: parallel = 11 free = 8 non-parallel = 3

R^2 = 0.994 F-stat = 0.296 F-prob = 0.827

Low Passage (p4)

x axis

1 10 100

10000

20000

30000

40000

50000

60000

70000

80000

System Suitability graph

4-P Fit: y = (A - D)/( 1 + (x/C)^B ) + D: A B C D Rel. Pot.

Plot#1 (Reference Standard: Concentration vs Medi... 8.46e+04 6.13 13.4 1.17e+04 1

Plot#2 (System Suitability: Concentration vs Median... 8.46e+04 6.13 13.3 1.17e+04 1__________

Weighting: Fixed

PLA (Std. Curve: Plot#1) Degrees of Freedom: parallel = 11 free = 8 non-parallel = 3

R^2 = 0.993 F-stat = 0.495 F-prob = 0.696

High Passage (p48)

Passaged 1-2x week for ~11 months Acceptable even in the absence of selection = stable genome integration

* no selection

9

Future technology exploration….


Several genomic editing technologies allow generation of reporter gene at gene-specific sites

TALENS Zn Fingers Crispr-Cas9

Learnings


Platform assays Helps build expertise Leverage cross project learnings Shared instrumentation: likely to catch problems

How much control do you want of critical reagents (cells, antigens/ligands, consumable, substrates) Qualify vendors for critical reagents? Make i Additional contractual agreements regarding quality of reagents and timely

delivery?

Take a long term view to assess control of bioassay method Qualifications and validations are ‘snap-shots’ in time Tracking and trending of parameters such as curve properties, passage

numbers, analysts, split schedules, FBS lots, critical reagents, etc are an important tool to assess control


Acknowledgement Bioassay development group

Darren Kamikura

Sharon Sibley

Jeanne Helmreich

Piyush Vyas

Denise Lyons

Liming Shi

Jane Sterner

Global Quality Labs

Robert Beckmann

Liying Lin

Katie Singer

Linda Wolters

Lessons Learned: Choice of Potency Assay and Differential Sensitivity to Degradation Pathways

Kirby Steger

Bristol-Myers Squibb Company, Princeton, NJ USA

Potency assays represent key components in the design of a well-defined control strategy for protein

therapeutics. During drug development, various methods iterations may be in use to determine the

potency of the therapeutic. Care must be taken to fully understand the mechanistic relevance of the

chosen systems for the definition of Critical Quality Attributes (CQAs). A case study will be presented

to illustrate how the choice of potency assays for release and characterization can influence the

designation of CQAs.

Slides were not available at the time of printing.

NOTES:

Bioassays Lessons Learned: Part One Workshop

PANEL DISCUSSION – Questions and Answers

Katrin Buss, BfArM, Germany


Bhavin Parekh, Eli Lilly and Company, USA

Kirby Steger, Bristol-Myers Squibb Company, USA

Shea Watrin, Amgen Inc., USA

Baolin Zhang, CDER, FDA, USA

Questions to be discussed:

In your experience, which have been the most important lessons learned with respect to bioassay

development?

What are the most important points to consider during bioassay development and/or

implementation for product testing?

What are proven tips and tricks for getting a bioassay to run reproducibly?

How do you mitigate bioassay performance issues if they can’t be avoided?

Did you ever encounter bioassay problems that could not be solved or mitigated?

From a health authority perspective: what are the most frequent bioassay-related challenges

faced by sponsors?

From a sponsor perspective: what are the most frequent bioassay-related questions received from

health authorities?

NOTES:

NOTES:

Bioassays Lessons Learned: Part Two Session Abstract

Session Chairs:

Hélène Gazzano-Santoro, Genentech, a Member of the Roche Group and Thomas Anders Millward,

Novartis Pharma AG

Bioassays represent an essential part of the control strategy for assessing safety and potency of

biopharmaceuticals. There are many potential challenges in bioassay development, implementation, and

maintenance. Avoiding or dealing with these is often an integral part of bioassay development as well as

implementation into routine product testing. This session, which will be divided into two parts, will

present a series of case studies on challenges and successes experienced with bioassays. Each talk will

discuss the specifics of the case study and the key learning’s.

NOTES:


Two-in-One: A Novel Approach of Bioassay Selection for Dual Specificity Antibodies

Guoying Jiang

Genentech, a Member of the Roche Group, South San Francisco, CA USA

Dual-specificity antibodies that simultaneously recognize two different antigens are being considered as

potential therapeutics for human diseases in recent years. How to design and develop a bioassay that is

reflective of the mechanism of action (MoA) and can measure the dual activities poses unique and

exciting challenges. For example, how many bioassays will be needed, one or two? Here we presented a

novel approach of having only one bioassay for a dual-specificity antibody by using a cell line that

expresses both receptors. It was found that the assay was able to measure the antibody effects on both

targets and was stability-indicating. Furthermore, the assay was demonstrated to be able to detect

potency changes in either target with assessment of antibody variants. This “single bioassay” approach

is reflective of the MoA of the intended therapeutic indications and has the potential to be used for other

dual specificity antibodies.


NOTES:

Challenges and Strategies in Selecting MOAs-reflective Bioassays for Bispecific Antibody

Xianzhi Zhou

MedImmune, Gaithersburg, MD USA

Abstract and slides were not available at the time of printing.

NOTES:

Challenges in the Development of Potency Assays for ADCs and their Utility to Detect Conjugate

Variants

Sonia Connaughton

ImmunoGen, Inc., Waltham, MA USA

Antibody-drug conjugates (ADCs) are anticancer agents that comprise a tumor-targeting antibody with a

cytotoxic agent attached. ADCs are designed to deliver the cytotoxic agent selectively to the tumor

tissue by targeting an antigen expressed on the surface of a cancer cell. The talk will discuss the

challenges of developing bioassays for ADCs containing a maytansinoid cytotoxic agent. This talk will

also focus on the complementary role of bioassays and analytical assays to study the in vitro activity of

conjugate variants


NOTES:

Development of an Alternative, in-vitro Potency Assay for Rabies Virus Vaccines

Robin Levis

CBER, FDA, Rockville, MD USA

The potency assay for currently licensed rabies virus vaccines, The NIH Test, is an animal challenge

assay in mice. Following two immunizations with the test vaccine, mice are challenged via IC

inoculation with a virulent strain of rabies virus. A potency value is assigned based on protection against

rabies virus disease and is quantitated relative to a reference vaccine of known potency. This potency

assay requires the use of many animals, takes a long time, and is highly variable with results ranging

from 25 – 400 %. In addition, the validity criteria for the assay are very stringent leading to many

invalid tests requiring multiple re-tests. The development of an alternative potency assay for rabies virus

vaccines has been ongoing for over three decades. Recently, a new collaborative study has been

initiated to define and validate an alternate potency assay for rabies virus vaccines. Critical scientific and

regulatory considerations for the adoption of an alternative assay will be presented.

NOTES:

1

Development of an alternative,

in vitro potency assay for rabies

virus vaccines.

Robin Levis

Division of Viral Products

Office of Vaccines Research and Review

Food and Drug Administration

March 24, 2014

Disclaimer

“My comments are an informal

communication and represent my own best

judgment. These comments do not bind or

obligate FDA.”

Regulations Related to Potency

for Human Vaccines

No specific tests are defined in the CFR for potency of vaccines. CFR 610.10 states: “Tests for potency shall consist

of either in vitro or in vivo tests, or both, which have been specifically designed for each product so as to indicate its potency in a manner adequate to satisfy the interpretation of potency given by the definition in 600.3(s) of this chapter.

Assigned potency must be shown to correlate with clinical efficacy.

2

Introduction

Rabies vaccines

Potency assignment

NIH potency test

History of alternate test development

WHO collaborative study using SRID – 1980s

FDA/NIBSC working group – 2000s

EDQM/EMA collaborative working group - 2012

Current work on alternate test development

ELISA assay development

Regulatory pathway for licensure

Licensed Rabies Vaccines – US

• sanofi pasteur - Human Diploid Cell Vaccine

• IMOVA X rabies - licensed in 1999

• Pitman-Moore Strain

• grown in MRC-5 cells

• Novartis Vaccines and Diagnostics - Purified

Chick Embryo Cell Vaccine

• RabAvert – licensed in 1989

• Fixed Rabies Virus Strain - Flury LEP

• grown in primary chicken fibroblasts

Rabies virus vaccine potency

Rabies virus vaccine efficacy was originally defined as

protection from death by rabies disease

based on a field study

Iranian wolf study (published 1976)

Potency assignment based on survival – 1 IU/dose

WHO recommends vaccines have a potency of > 2.5 IU/mL

Current rabies vaccines are licensed with a potency

specification of > 2.5 IU/mL (as determined using the

NIH potency test)

Efficacy of currently licensed vaccines has been

demonstrated in controlled clinical trials

No vaccine failures* using current post-exposure

treatment regimen

3

Successful protection of humans exposed to

rabies infection. Postexposure treatment with the

new human diploid cell rabies vaccine and

antirabies serum.

JAMA. 1976 Dec 13;236(24):2751-4.

Bahmanyar M, Fayaz A, Nour-Salehi S, Mohammadi M, Koprowski H.

Abstract

Forty-five persons severely bitten by rabid dogs and wolves in Iran

were treated after exposure with a new rabies vaccine produced in

cultures of human diploid cells. All except one also received one

injection of rabies immune serum. This treatment, in contrast to past

experience with other vaccines, resulted in protection of all individuals

against rabies. Thus, almost a century after the postexposure

treatment of humans was initiated, an effective tool for protecting

man against rabies has finally been developed.

NIH Potency Test (Monogr Ser World Health Organ. 1966;23:145-51.)

Current potency test:

Originally developed for neural tissue based vaccines

Animal-based immune challenge assay: Immunize mice at day 0 and 7 (5 groups/5-fold

dilutions)

I.C. Challenge on day 14

Observe for rabies disease*

Potency is calculated based on survival relative to a reference vaccine.

Assay is currently used as release test and as a stability indicating test

Why do we need a

replacement test? NIH potency test has never been considered a great

assay.

Test uses ~600 mice.

High degree of variability: 25 – 400%

Takes up to 6 weeks to complete.

Pass on potency is the geometric mean of two valid

tests in the US and Canada.

Single test used everywhere else.

Ongoing discussions on test replacement for several

decades

Collaborative studies to establish an alternative test started in

the early 1980s

4

A collaborative study on the use of single radial

immunodiffusion for the assay of rabies virus

glycoprotein.

J Biol Stand. 1984 Jul;12(3):283-94.

Ferguson M, Seagroatt V, Schild GC.

Abstract

The single radial immunodiffusion (SRD) technique has been applied to the

assay of the glycoprotein content of rabies vaccines produced in cell cultures.

Fourteen laboratories in seven countries participated in a collaborative study to

evaluate the reproducibility of the SRD technique; some laboratories also

examined vaccines in the mouse protection (NIH) test and by enzyme

immunoassay. Good agreement was found between potency estimates using

the SRD technique: the geometric coefficients of variation for combined

potency estimates of all laboratories were about 10%. SRD assays appear to

have a role for the in vitro assay of antigen content of vaccine and could

complement results obtained in in vivo assays which are subject to wide

variability.

Use of the single radial immunodiffusion test as a

replacement for the NIH mouse potency test for

rabies vaccine.

Dev Biol Stand. 1986;64:73-9.

Fitzgerald EA, Needy CF.

Abstract

The method currently recommended by the World Health Organization (WHO) for the

potency assay of rabies vaccine is the NIH mouse potency test, a highly variable test

requiring large numbers of animals. The Single Radial Immunodiffusion (SRID) test, an in

vitro test, has been used successfully for the quantitation of hemagglutinin in inactivated

influenza vaccine and is being evaluated for its utility as an assay for the rabies virus

glycoprotein, considered to be the major protective antigen, of rabies vaccine. Potency

values calculated using the SRID test were compared with those calculated using the

NIH test for rabies vaccines produced in cell culture. The within-test variability was

significantly lower with the SRID test but the potency values were generally higher than

those from the NIH test. Vaccines which assay below the minimum acceptable

potency value (2.5 International Units/ml) in the NIH test generally gave values

above that level in the SRID test. The implications of these results on rabies

vaccine control testing are discussed.

Alternatives to the NIH Rabies

Vaccine Potency Test

Working group re-convened

FDA sponsored workshop - September 2000

Representatives from industry (Chiron

Behring and Aventis Pasteur), CDC, Thomas

Jefferson University, Kansas State University,

NIBSC, AFSSAPS, PEI, EDQM

Collaborative study between CBER, NIBSC,

and two industry sponsors

Goal of study was to develop an ELISA assay that

would potentially serve as an alternative potency

assay

5

Overview of Study (2000)

Development of an in vitro ELISA to test for

antigen content in rabies vaccines

ELISA to replace NIH test as a release test

for potency

No requirement for correlation with NIH test

Consistency of manufacturing

ELISA to replace NIH test to test shelf life

and stability

Show that ELISA will identify sub-potent lots

Initial ideas for replacement

potency test (2000)

Development of uniform protocol

Development of uniform reagents

Establishment of standard values as

compared to a reference standard

Mathematical determination of potency

from ELISA results

Mass measurement vs. protection in animals

Lots to test:

normal production lots

sub-potent lots

Lessons learned from

collaboration (2000)

Developed ELISA assay with available reagents

Common reagents were difficult to obtain

Even published ones

Vaccine strain differences matter

Potency relative to a reference standard were

different based on vaccine strain and reagents

used

Able to identify sub-potent lots

Data correlated with NIH test results

Industry sponsors continued with ELISA testing

for information purposes.

6

Regulatory Catch 22

What is the approval pathway for an alternate

test?

Sponsors want to know NRAs will approve

alternate test before expending resources to

develop and validate test.

NRAs (in this case – FDA) would like to see

data prior to confirming adequacy of test as a

replacement.

Is it possible to institute a replacement

test for the rabies potency assay?

We have successfully approved the replacement of

several animal-based immunogenicity assays with

ELISA-based assays for the measurement of viral

vaccine potency

Neutralizing epitopes were well defined – or –

Antibody used in the assay bound to critical conformational

epitopes

Clear correlations could be shown between amount of

antigen required to induce immune response in animals vs

amount of antigen measured using alternative in vitro

assays vs immune response in human vaccinees

Can we do this with rabies vaccines?

Replacement of NIH Test

Currently potency is defined by protection

against challenge in animals –

By virtue of survival after challenge, the NIH potency

assay measures a protective response (animals are

doing the work for us)

Potency/dose should correlate with clinical efficacy

Correlation between protection against disease in animals

and potency in humans was established in clinical trials -

If neutralizing epitopes are well defined then it should

be possible to correlate the amount of antigen with

immune response measured in animal. Then,

It should be possible to define potency based on the

amount of antigen in a dose.

7

Replacement of NIH Test

Attributes of alternate assay: Neutralizing epitopes are well defined for rabies

Reagents are defined that recognize these epitopes and distinguish appropriate conformation of virus

Protective immune response has been defined – well accepted for human vaccines

Clear correlations HAVE NOT BEEN shown between amount of antigen required to induce protective immune response in animals vs amount of antigen measured using alternative in vitro assays (how does this translate to vaccine efficacy)

How do we show correlation between potency defined by protection in animals vs potency defined by alternate methods?

Is there a necessity for clinical studies?

EPAA meeting

(The European Partnership for Alternative

Approaches to Animal Testing)

Archachon, France - October 2012

Re-initiate collaborative discussion on

alternate test development

Establish timeline for reagent and assay

development

Two phase development approach

Phase 1- reagent selection - labs are testing

reagents with individual, in house assays.

Phase 2 – collaborative study to define assay

Development of an

alternative potency test

Development of study protocol

Development/availability of reagents

Difficult step*

Development of reagents at CBER

Standard values based on International reference(s)

Definition of potency from test results

Establishment of test specifications

Requirement for clinical data

Lots to test:

Normal production lots

Lots on stability

8

Considerations for Approval

Should there be a requirement to show comparability/equivalence to current mouse potency test?

If tests are not comparable, then the new test must be well qualified. Clear definition of potency: antigen units vs.

international units per dose.

Antibodies used for detection must correlate with protection.

Antibody binding affinity and vaccine strain differences must be well defined (common reference) Due to strain differences, it may be necessary to utilize

different reagents for each vaccine

Depending on the reagents used, the level of free G protein vs. virus associated G protein must be determined.

Test must be able to distinguish potent vs. non-potent lots.

Considerations for Approval

Is there a necessity for clinical data to support the new potency assay? For human vaccines – immunogenicity trial to show

antibody response to vaccines with potency measured using alternative assay

Should this be required for the currently licensed vaccines?

History of manufacturing consistency

History of clinical efficacy

Summary

Renewed global effort by both human and veterinary vaccine manufacturers and control authorities to establish alternative rabies vaccine potency tests.

FDA has licensed non-animal based replacement tests for several vaccine products

FDA is working with sponsors and other regulatory authorities in this endeavor regarding the replacement of the NIH potency test for rabies virus vaccines.

Bioassays Lessons Learned: Part Two Workshop


Evangelos Bakopanos, Health Canada, Canada

Sonia Connaughton, ImmunoGen, Inc., USA


Guoying Jiang, Genentech, a Member of the Roche Group, USA

Robin Levis, CBER, FDA, USA

Xianzhi Zhou, MedImmune, USA


In your experience, which have been the most important lessons learned with respect to bioassay

development?

What are the most important points to consider during bioassay development and/or

implementation for product testing?

What are proven tips and tricks for getting a bioassay to run reproducibly?

How do you mitigate bioassay performance issues if they can’t be avoided?

Did you ever encounter bioassay problems that could not be solved or mitigated?

From a health authority perspective: what are the most frequent bioassay-related challenges

faced by sponsors?

From a sponsor perspective: what are the most frequent bioassay-related questions received from

health authorities?

NOTES:

NOTES:

Exhibitor Partner Showcase

Eurofins Lancaster Laboratories, Inc.

2425 New Holland Pike

Lancaster, PA 17601 USA

Phone: 717-656-2300

Website: www.EurofinsLancasterLabs.com

Company Description

Eurofins Lancaster Laboratories, a global leader in comprehensive cGMP-compliant laboratory services,

enables bio/pharmaceutical companies to advance candidates from development through

commercialization, ensuring regulatory compliance, cost effectiveness and achievement of timelines.

See why 800+ leading pharmaceutical and biotech customers continue to trust us with their product

tesing needs at www.EurofinsLancasterLabs.com.

NOTES:

Effective cGMP Bioassay Outsourcing

Alexander Knorre2; Weihong Wang

1

1Eurofins Lancaster Laboratories, Lancaster, PA USA;

2BSL BIOSERVICE Scientific Laboratories

GmbH, Planegg/Munich, Germany

The development and qualification/validation of cell based potency assays is critical for measuring

biological activity and ensuring consistent product quality. Due to the complex nature of biological

assays, managing outsourced cell based potency assay projects can be challenging.

This presentation will offer solutions for outsourcing cell based potency assays with Eurofins Lancaster

Laboratories, including a highly efficient assay group with more experienced PhDs than any other lab

and assay specific training with clearly expressed performance expectations and a greater than 95%

success rate on assay transfers.

With the largest breadth of services and facility capacity for cGMP biologics stability and release in the

industry, Eurofins Lancaster Laboratories provides world wide cGMP Bioassay testing support with

tailored services to meet any client’s specific needs.

NOTES:

1

www.LancasterLabsPharm.com

Effective GMP

Bioassay Outsourcing

Eurofins Lancaster Laboratories, a global leader in comprehensive laboratory services,

enables pharmaceutical and biopharmaceutical companies to advance candidates from

development through commercialization while ensuring regulatory compliance, cost

effectiveness, and achievement of timelines.

cGMP Bioassay Services

Potency Assays

Assay Capability & project experience

Facilities and Equipment

Data analysis Software

Supporting Assays

Staff and Training

Keys to Success

EU Laboratories

Potency Assays – Project Experience

Assay Types

Proliferation/cytotoxicity assays – multiple substrate type

Apoptosis assay (via caspase3/7 activity)

Signaling molecule (cAMP, phospho-protein, AP)

Cell surface receptor binding

Reporter gene assay

ADCC, CDC

Binding ELISA (direct and competitive)

TR-FRET binding assays

Detection Methods: colorimetric, fluorescence, time-resolved

fluorescence, luminescence

2

Potency Assays – Capability

Tailored Services to Meet Client Needs

Complete assay development

Method optimization

Method transfer

Method qualification and validation

Routine testing to support GMP manufacturing and release

Support for stability testing

Capability to bank cells for client assays

Dedicated cell banking team

Preparation of Master and Working cell banks

Testing/Characterization of cell banks – Sterility, Mycoplasma

(rapid or compendial), IVAA

Potency Assays –Equipment & Data Analysis

Software

Plate Readers – 5 units

Molecular Devices M2 – 2 units

Molecular Devices M5e- 1 unit (additional unit on order)

Molecular Devices L – 2 units

Automated Cell Counter

Beckman Coulter Vi-Cell – 2 units

Automated Plate Washers – 7 units

Biotek Precision Robot

RT-qPCR

Applied Biosystems (2-7500s, 1-7900HT)

Data analysis software (with PLA analysis):

Softmax, StatLIA, PLA 2.1

All instruments and software are fully validated and 21 CFR part 11

compliant

Supporting Services

qPCR

Validated in house qPCR assays

Residual DNA (CHO, E. coli, Human, BHK)

Selected viruses (fPERT, MMV, xMuLV, PCV 1/2)

Development and validation of client specific assays

Segregated PCR facilities

– Reagent prep lab, samples prep lab, instrument lab

– HEPA filtered and pressure controlled

ELISA for host cell and process residuals

HCP assays using generic kits (Cygnus)

Process residual ELISAs (Protein A, Benzonase etc)

Method development and validation available for process/product specific HCP ELISA (later phase projects)

Rapid Mycoplasma - Milliprobe

rRNA targeted Transcription Mediated Amplification system with a 5 day TAT for GMP mycoplasma results – Fully Validated

In Vitro Cytotoxicity – USP <87>

3

Facilities

~4,000 square feet of Laboratory space for analysis

Three dedicated cell based assay laboratories

Two labs HEPA filtered, Positively pressured

One lab BSL-2 classified, HEPA filtered, Negatively Pressured

Dedicated ELISA laboratory

Dedicated Molecular laboratory

Dedicated positives and negatives prep rooms

Four dedicated cell banking suites

2 - Class A/B, 2 – ISO 5/7

>2,000 square feet total space

HEPA Filtered

Positive Pressure to surrounding areas

Validated disinfection, cleaning and environmental monitoring

Staff – Cell & Molecular Biology Services

Highly efficient single-site assay group

27 full time employee

20+ FTE at bench

5 group leaders

7 Ph.D. scientists

2 cell culture technician

All assay scientists with Bachelor or Master degree

Cross training in ELISA, cell based assay and qPCR

Training

Core Training modules completed by all analysts include

Aseptic Technique

Tissue Culture

General related lab equipment (pipettes, balances etc)

ELISA, PCR and cell based assay training modules

Assay Specific Training

One-on-One training at ELLI with experienced analyst

Training at client’s facility

Training by client at ELLI

Data Analysis Software Training

4

Keys to Success

More than a typical customer-vendor relationship

Start early and allow sufficient time for transfer and validation

• Perform Gap analysis of current client method if requested to

prepare for qualification/validation

Performance Expectations - clearly articulated from the onset

• Use of project timeline spreadsheets (i.e. Gantt charts)

Conduct training as a first step for method transfer - ELLI

analysts trained by client’s technical expert at the client site or

at LLI whenever necessary

Direct communication between assay expert at your facility and

the principal scientist managing your assay at ELLI

Weekly/biweekly conference calls to discuss status and resolve

issues

BSL BIOSERVICE

Bioassay Services (EU) Munich, Germany

Instrument Readout Capabilities

Absorbance

Luminescence

Fluorescence (incl. TRF and HTRF)

AlphaScreen ®

AlphaLisa®

Analysis by flow cytometry (HTS module equipped, eight colour)

Radioactivity (HTS module equipped scintillation counter)

BSL – Capabilities

Established Bioassays

Cell Proliferation

Cell Survival/ Cell Apoptosis

Reporter Gene

ADCC and CDC

Cell Migration Assay

Binding Assays (ELISA)

Viral CPE Assay

In vivo Bioassay Biological Assay

Tailor-made bioassays

Pharmacopeial Methods (EP/USP):

G-CSF (Filgrastim and PEG- Filgrastim, EP and USP)

Interferon Beta (EP)

Interferon Alpha 2b (EP)

Erythropoietin (EP)

FSH (Urofollitropin) (EP)

Insulin (USP)

5

Bioassay Personnel - BSL

Each project team is set up depending on specific project requirements

Bioassay core team

Head of department (1 FTE)

5 Scientists (4.5 FTE)

6 Analysts (5.5 FTE)

2 Team assistants (1.6 FTE)

Total: 14 (12.6 FTE)

Experience with Therapeutic Antibodies (including Bispecific), Antibody

Drug Conjugates, Hormones/Peptides and Viral (Like) Particles

GMP compliance trained

2006 and 2011: US FDA two day inspection regarding bioassay validation

and bioassay routine testing, no 483 observations

Bioassay: Facts - BSL

Longstanding experience in development/optimization (DOE), validation and performance of bioassays (1984 - present) with global client base

Performance of bioassays using primary cells, permanent cell lines and animals (mice, rats, rabbits)

Project planning and tracking using Gantt charts

Assay monitoring: Statistical process control of bioassay and cell culture parameters using Minitab

Experience in statistical analysis (in-house and external biostatistician)

Brand-new facility with dedicated laboratory space for cell based bioassays and for ELISA (duplication of previous capacity)

New Space expansion - BSL

March 13, 2014 BSL BIOSERVICE Facts 15

3,500 m² animal facility for

rodent & non-rodent species,

in vivo bioassay

700 m² for in vitro testing

1,000 m² for in vitro bioassay,

viral clearance & biosafety

3 operational facilities

south of Munich,

Germany

6

Conclusion

ELLI has the capacity, facilities, project experience and

scientific intelligence to successfully support client assay needs

throughout the lifespan of the project

Working together with Eurofins BSL, we provide world wide

cGMP Bioassay testing support

Constant scientific exchange and project communications

High efficiency method transfer between sites

Cost effective co-validation

Thank you!

Stegmann Systems

Raiffeisenstrasse 2, C1/C2

D-63110 Rodgau/Hesse, Germany

Phone: 49 61 0677 0100

Website: www.bioassay.de

Company Description

Stegmann Systems is a leading software vendor in the field of bio-statistical software for biological

assay since 1996. The main product of Stegmann Systems is the well-known PLA Software for Analysis

of Biological Assays. PLA is currently used by over 400 companies worldwide in GxP and non-GxP

environments. About 20 developers (software specialists, statisticians, quality specialists) are currently

involved with the development of PLA. Their primary focus is the development of easy-to-use software

that supports the bio-statistical tasks and compliance needs of our clients from the beginning of product

development to production in GxP and non-GxP environments.

NOTES:

New Capabilities of PLA

Ralf Stegmann

Stegmann Systems, Rodgau/Hesse, Germany

PLA is Stegmann Systems leading software solution for the analysis of biological assay. In this

presentation a larger number of updates for the analytical features and technical capabilities of PLA will

be presented. These updates cover several classes of biological assay and advanced calculations as well

as the platform: Parallel-Line Assay, Parallel-Logistic Assay (3-,4- and 5 parameter non-linear

regression), Dichotomous Assay (Quantal Response Assay), Slope-Ratio Assay, Control Charting,

Equivalence Margin Development etc..

NOTES:

Promega Corporation

2800 Woods Hollow Road

Madison, WI 53711 USA

Phone: 608-274-4330

Website: www.promega.com

Company Description

Promega Corporation is a world leader in providing innovative solutions to the life sciences industry.

We develop bioluminescent technologies that deliver more biologically relevant data for biologics and

small molecule drug discovery, and we bring best-in-class technology platforms that deliver data with

least amount of variability, in off-the-shelf and custom formats. Speak with a Promega representative to

learn how we can develop a solution for your discovery needs. For more information about Promega,

visit www.promega.com.

NOTES:

Bioluminescent Technologies for Biological Functional Analysis and Protein-Protein Interactions

Mei Cong

Promega Corporation, Madison, WI USA

Bioluminescence reporter gene assays have been widely adapted in quantifying biological activities. The

application of the new NanoBiT technology in monitoring protein-protein interaction will be discussed

as a new method of measuring biological functions.

NOTES:


Session Abstract

Session Chairs:

Katrin Buss, BfArM, Federal Institute for Drugs and Medical Devices and Helena Madden, Biogen Idec

Measurement of biological activity is required throughout all stages of biopharmaceutical and cellular

and gene therapy product development programs. Typically development and implementation of

bioassays takes place in a phase-based manner. Simple methods adequately support early stage

development and are followed by a progression to an MOA-reflective cell-based functional method or

methods. Bioassays are used for release and stability testing, to assess comparability during process

changes, and for understanding the impact of structural changes. This session will discuss strategies for

selection, development, maintenance, and execution of bioassays to support successful development of

biological drugs. Regulatory perspectives on measurement of bioactivity at all stages of development

will be shared and discussed.

NOTES:


Implementation of the Next Generation Effector Function Assays for Comparability Assessments

Poonam Aggarwal

Pfizer, Inc., Chesterfield, MO USA

The development of a biotherapeutic program necessitates an in-depth understanding of the mechanism

of action(s) of the product. In vitro cell-based and binding studies are performed in order to characterize

the potential biological activities of mAb products that can perform distinct functions through their Fab

and Fc regions. The assay panels include binding to the target antigen, target-mediated biological

activity, binding to the Fcγ receptors, and complement, as well as functional assays for Fc-associated

effector functions (i.e., ADCC and CDC assays).

Examples will be presented that includes strategy and development of the bioassays to assess antibody

effector function.


NOTES:

The Dual Benefit of Structure Function Studies: Better Understanding of Molecules and Help with

MOA-relevant Bioassays

Carl Co

Biogen Idec, Cambridge, MA USA

Structure activity relationship (SAR) and stability studies must be performed to understand the critical

product quality attributes that are responsible for the efficacy and safety of a molecule. In addition, these

studies generate sample variants that are extremely useful in the development of new MOA-relevant

bioassays. Two case studies are presented which describe two approaches to structure function studies:

in one study, long term stability samples were used to assess the effect of product quality changes on

bioactivity; in a second study, Fc variants and forced degraded samples were generated to better

understand the Fc function of the molecule and to aid with the development of more MOA-relevant and

stability-indicating bioassays


NOTES:

Standards and Beyond: Challenges of Application of Old Methods to Next Generation Products

Elena Semenova; Penny Post; Tim Fields; Manon Cox

Protein Sciences Corporation, Meriden, CT USA

Propagation of the virus in developing chicken embryos is the predominant way to manufacture

influenza vaccines. Despite the fact that recombinant technology has been used in the manufacturing of

biopharmaceuticals for more than two decades, the first recombinant influenza vaccine was only

approved early 2013. Difficulties in adaptation of classical and widely accepted from 1978 release

bioassay - single-radial-immunodiffusion (SRID), optimized for traditional products to the new

recombinant vaccine were among many reasons for a lag in market appearance of recombinant vaccine,

which make influenza immunization a reality for people with severe egg-allergies and have a greater

potential to provide necessary number of doses fast enough during pandemics. SRID signal depends on

ability of the influenza antigens to diffuse into the agarose gels, form insoluble complexes with specific

antibodies in pores of the gel and remain there after gel washing. The vast majority of reference antigens

for SRID are produced by WHO Essential Regulatory Laboratories are made from egg-derived HA

proteins without purification step. Among factors which may present challenges in applicability of SRID

assay to new technologies are differences in expression systems, level of purification, and chemical

agents used for inactivation of reference antigens. Regulations and guidance, developed and optimized

for traditional products, may need adjustment when applied to novel innovative and often fundamentally

different products.

NOTES:

3/13/2014

1

Standards and beyond: challenges of

application of old methods to next

generation products

Elena Semenova, Penny Post,

Tim Fields, Manon Cox

March 25, 2014

Recombinant Flu vaccines

Propagation of the influenza virus in chicken embryos is the predominant way to manufacture influenza vaccines.

Various problems, associated with the use of embryonated chicken eggs for inactivated vaccines, such as getting high yield egg-adapted reassortant and possible antigenic change during egg adaptation (one of the reason of low vaccine effectiveness in 2012-2013 season, VRBPAC 2013 transcript), and the necessity to attenuate pathogenic strains of the influenza virus (for live vaccines) could be solved through the use of highly purified recombinant subunit vaccines.

A recombinant influenza vaccine has a greater potential to provide necessary number of doses faster during pandemics and also makes influenza immunization a reality for people with severe egg-allergies.

Flu vaccine and recombinant products timeline

1942-1943 1945 1978 2012 2013 1982

First recombinant Flu

vaccine is approved (Protein Sciences

Corporation)

First recombinant product

(insulin) is approved (Eli Lilly)

Difficulties in adaptation of classical and widely accepted release bioassay (SRID),

optimized for traditional products to the new recombinant vaccine were among

many reasons for a more than 30 years lag in market appearance of recombinant

Flu vaccine

3/13/2014

2

SRID as a potency bioassay for Flu vaccines

The single-radial-immunodiffusion (SRID) assay has been

adopted world-wide in 1978.

SRID measures amount of hemagglutinin (HA): viral surface

protein. Antibodies against HA protects from the Influenza

SRID was optimized to provide consistent results when tested

against all of the licensed influenza vaccine preparations at

time of implementation: whole virus (currently discontinued)

and split vaccines. Both vaccines were propagated in eggs

and inactivated by chemical agents (MS Williams 1993).

Advantages of SRID as bioassay

Simple in performance

Robust

Reproducible

An international collaborative study (Wood et al. 1981) demonstrated test reproducibility

(geometric coefficient of variation between laboratories less than 10% ).

Two EU collaborative studies in 1989 and 1990 have reaffirmed test reproducibility (J. Wood Textbook of Influenza 1998).

Correlation with clinical efficacy Numerous clinical trials have validated the test (Ennis et al. 1977; Pandemic Working Group

of the MRC 1977; Nicholson et al.1979)

Numerous follow up clinical studies have confirmed that the test measures immunologically

active HA (J. Wood Textbook of Influenza 1998).

Measure biologically relevant potency,

Ability to detect individual strains in multivalent vaccine

Proven stability-indicating capacity

5

Disadvantages of SRID

SRID requires the production of antibodies to a strain-specific HA (up to 4 months)

SRID is relative bioassay and requires reference standard with assigned potency,

which production and standardization must be performed after production and

characterization of strain-specific antibodies.

For seasonal influenza virus vaccines, which typically contain three constantly

changing sub-types, new antisera and reference standard must be made and

standardized each time a new strain is incorporated into the vaccine formula

during annual reformulation.

Each of these steps are time consuming and costly. More importantly reagent

preparation may be a major factor in delaying of supplies of a vaccine during

pandemic.

Suboptimal for vaccine which are produced in other than egg-based systems:

in cell culture (J. Wood, Textbook of Influenza, 1998 Chapter 25) and using

recombinant DNA technology.

6

3/13/2014

3

This presentation summarizes points to consider for

adaptation of established standards, which are produced for

traditional Flu vaccine, for measuring of potency of novel

innovative and often fundamentally different products.

Description of SRID assay

Calculation of Flu vaccine potency

Standard Sample

1:1

1:1.5

1:2

1:4

dilution

Potency of sample is assigned by

comparing ring diameter with the

standard with US FDA assigned

potency

3/13/2014

4

Standard and Vaccine in SRID assay

Different vaccines – the same standard

Flu vaccines, which are required to use SRID as a potency bioassay

Inactivated vaccine •Inactivated either by alkylating or crosslinking agents

•Predominately egg-produced Recombinant

vaccine Produced in insect cells

Contains>90 % of rHA Split vaccine •Virion split by detergent •Virion contains 38-44 % of HA (Ruigrok, Textbook of Influenza 1998)

Subunit vaccine •After splitting vaccine is enriched for surface antigens (HA and NA)

Standard for SRID is a freeze-dried inactivated whole virus,

initially optimized for split and whole vaccines.

Whole vaccine •Discontinued at the end of 1970’s due to side effects

Calibration of seasonal/pandemic influenza

antigen working standards

Reagents are prepared by the four WHO Essential Regulatory

Laboratories (ERLs):

• Australia – Therapeutic Goods Administration (TGA)

• Japan – National Institute for Infectious Disease (NIID)

• United Kingdom – National Institute for Biological Standards

and Control (NIBSC)

• USA – Center for Biologics Evaluation and Research (CBER)

One of ERLs (lead ERL) prepares Primary Liquid Standard (PLS) and

a large batch of freeze-dried antigen (inactivated whole virus), which

are distributed to all other ERLs for independent calibration.

Data generated by the ERLs are collected by the lead ERL and

compiled for the final potency value agreement and confirmation.

Manufacturers’ data may be considered, if available. The lead ERL has

final authority to assign a potency value. EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION (Geneva, 17-21 October 2011), Generic protocol WHO 2012

3/13/2014

5

Primary Liquid Standard (PLS)

PLS is inactivated whole virus

HA is 38 - 44 % of

total virion mass (Textbook of Influenza, R.W.H.

Ruigrok 1998)

Scheme is based on “Generic protocol for the calibration of

seasonal/pandemic influenza antigen working reagents” WHO 2012

PLS is calibrated by physicochemical means

Points to consider for recombinant Flu

vaccine at PLS calibration step

Co-migration effect for other proteins (different proteins at the same band on SDS-PAGE) can be as high as 25 % (Getie-Kebtie et al, An. Biochemistry 2011).

General guidance for confirmation of accuracy of PAGE band analysis for ERLs: HA content should be between 20% and 50% of total protein. Recombinant protein is > 90 % of HA in the sample and free from influenza viral proteins.

Ratio between HA and total protein for PLS is assigned based on band densitometry analysis of sample which have less than 50 % of HA, thus co-migration effect from other proteins is possible. Even moderate co-migration effect could put potency number “off” for highly pure HA recombinant vaccine.

Freeze-dried working reference standard

PLS with assigned

HA content

freeze-dried antigen

(working reference standard)

SRID assay

Assigning potency to working reference standard

Distribution to manufacturers of Flu vaccine

3/13/2014

6

Points to consider for recombinant

Flu vaccine

As reported previously, the SRID gave unreliable potency results when different chemical agents used for inactivation of the virus were used for preparation of reference standard and vaccine samples (R. Gupta, and W. McCormick, US FDA, WHO/FDA/HC Workshop, June

2010, Ottawa, Canada). These agents modify proteins in a different ways, and therefore inactivated viral preparations may have different gel mobility. The recombinant HA vaccine does not require an inactivation step.

Traditional influenza vaccines contain significant amount of other viral and host proteins, which may be involved in formation of stable complexes. This further increases complexity of the sample post treatment with inactivating agents.

All these facts may contribute to a higher or lower gel mobility of the highly purified recombinant HA proteins compared with

egg-based reference antigens.

Points to consider for recombinant

Flu vaccine (cont)

The protein ring intensity and appearance, which is also likely dependent on the way of HA presentation in the sample, may be different between standard and measured samples, produced in heterologous systems.

Such differences could create problems in automated reading of the gels, which negatively affects accuracy and precision of

the bioassay.

CONCLUSIONS

Regulations and guidance, developed and optimized for traditional products, may need adjustment when applied to novel innovative and often fundamentally different products.

Points to consider outlined above may require either an adaptation of traditional method for the new products, for example through the preparation of compatible reference reagents, or the development of standard-independent physicochemical assays measuring the physiologically active form of the proteins present in the vaccine.

Global Implementation of Bioassays – Things to Consider

Bruce Meiklejohn

Eli Lilly and Company, Indianapolis, IN USA

To be competitive in today’s pharmaceutical industry companies need to be able to develop and

distribute their products on a global basis and can no longer focus on a specific region or country.

Multiple regions add complexity to the already difficult development process. The development of a

biotechnology pharmaceutical product requires multiple areas of scientific expertise and regulatory

understanding to be successful. The biotechnology industry continues to mature and analytical

methodologies advance and are more discriminating for the characterization of complex biological

products. Different countries and regions are in different stages of acceptance of analytical control

strategies where historically the process defined the product and clinical experience established the

safety and efficacy. As the development and commercialization of these products becomes a global

activity the need to meet regional requirements is becoming more diverse. In the past decades ICH

helped to normalize expectations in certain regions of the globe. As the industry continues to grow

regional or specific country expectations emerge making the global development process gain in

complexity. These differences include specifications, testing requirements, shelf life and sample

requirements for the health authorities to name a few. Examples of differences in expectations and

strategies will be discussed to help with clinical trials and post approval life cycle management.

NOTES:

3/13/2014

1

Global Implementation of Bioassays, Things to Consider

Bruce Meiklejohn, Eli Lilly and Company, Indianapolis IN USA

25 March 2014

The opinions expressed in this presentation are solely those of the individual presenter, and do not necessarily reflect the views of Eli Lilly & Company.

Confidential 2014 Eli Lilly and Company

• To be competitive in today’s pharmaceutical industry, companies need to be able to develop

and distribute their products on a global basis and can no longer focus on a specific region or

country. Multiple regions add complexity to the already difficult development process. The

development of a biotechnology pharmaceutical product requires multiple areas of scientific

expertise and regulatory understanding to be successful. The biotechnology industry

continues to mature and analytical methodologies advance and are more discriminating for

the characterization of complex biological products. Different countries and regions are in

different stages of acceptance of analytical control strategies where historically the process

defined the product and clinical experience established the safety and efficacy. As the

development and commercialization of these products becomes a global activity the need to

meet regional requirements is becoming more diverse. In the past decades ICH helped to

normalize expectations in certain regions of the globe. As the industry continues to grow

regional or specific country expectations emerge making the global development process gain

in complexity. These differences include specifications, testing requirements, shelf life,

sample requirements for the health authorities to name a few. Examples of differences in

expectations and strategies will be discussed to help with clinical trials and post approval life

cycle management.


3/13/2014

2

Background

• To be competitive in today’s pharmaceutical industry companies need to be able to develop and distribute their products on a global basis.

• The development of a biotechnology derived pharmaceutical product requires multiple areas of scientific expertise and regulatory understanding to be successful.

• Companies can no longer focus on a specific region or country.

• Multiple regions add complexity to the already difficult development process.


• The biotechnology industry continues to mature and analytical methodologies advance and are more discriminating for the characterization of complex biological products.

• Countries and regions have different expectations on analytical control strategies

• Historically the process defined the product and clinical experience established the safety and efficacy.


Background

Background

• As the development and commercialization of biotechnology products become a global activity, the need to meet regional requirements is becoming more diverse.

• ICH has helped to harmonize expectations in certain regions of the globe.

• Regional and specific country expectations emerge making the global development process increase in complexity.


3/13/2014

3

Development Complexity

Needed Studies

Release Tests In-Process Tests Stability

Clinical Efficacy/Safety

Bioassay

Additional Characterization

Animal PK/PD

Human PK/PD

CQA Impact Potential

Significant

Minimal

Moderate

Phase of Development

Post Approval Commercial

Pivotal

Tox

FHD

FED

Manufacturing Process

Cell Line

Purification

Manufacturing Site

Fermentation

Scale-Up

Formulation

Chemically Syn Peptide

Structure Complexity

Antibody drug Conjugate Complex Protein

Simple Protein

Recombinant Peptide

Fusion Protein


CMC Inputs for Global Registrations

Site

Reg

istr

atio

ns

CM

C “

Plu

s” D

ocu

men

ts

Pro

cess

Val

idat

ion

Met

ho

d T

ran

sfer

s

Dru

g M

aste

r Fi

les

Pro

du

ct S

amp

les

Cer

tifi

cate

of

Ph

arm

Pro

du

ct C

PP

CTD

Mo

du

le 2

an

d 3


Manufacturing Site Registrations


• A manufacturing site may require additional information for a CT or registration submission for certain countries

For example:

―Site registration and the need for GMP certification

• Prior to submission

3/13/2014

4


• CTD content for some countries outside the United States require additional information

• CMC “Plus” documents include:

COAs

Reagents (Company generated)

Representative Standard Curves

Batch Records

CMC “Plus” Documentation


• In certain countries the local Quality Control Labs need to release commercial Drug Product: e.g., Japan, Korea, Mexico, Brazil, Argentina

• Transfer of methods to the Quality Control Labs is needed prior to launch [not submission]

• Some health authorities will perform testing in MOH labs, so •

method , reagents ,& instrument details will be needed ―e.g., China, Russia, UAE

Companies may be asked to provide samples and reference

standards

Analytical Method Transfers


Country Performed By Comments (including whether full retesting or different tests/standards used)

Albania MOH lab Test parameters listed on COA

Algeria MOH lab For the registration, the national laboratory of control tests all the parameters but once the product is registered, the test will be on standard parameters and on each batch imported

Armenia MOH lab only selected finished products; selected tests including ID

Azerbaijan MOH lab only selected finished products; selected tests including ID

Belarus external no other information available

Bosnia MOH lab Quality control of the first imported batch includes full testing of all parameters in accordance with approved Specification. Every subsequent analysis of every subsequent imported batch in

most cases is partial-Control laboratory makes decision.

China State lab National lab determines testing for import to be tested by State lab.

Costa Rica MOH lab Testing on the first imported lot only

Croatia external Only perform partial testing.

Egypt MOH lab Full testing required.

Georgia MOH lab only selected finished products; selected tests including ID

India MOH lab All insulin and insulin analog lots tested.

Iran MOH lab Tests are not performed on each shipment, rather they are done randomly on one shipment

each while MOH informs the agent in advance

Iraq MOH lab Tenders are usually divided in several shipments, tests are done on the first shipment of each

tender

Israel MOH lab Full testing. The MoH will test only first batch of every new product, marketed products

following any variation and all biological products. All other batches will be released by the company's Qualified Person

Jordan MOH lab testing is performed on the first seven shipments to the country

Kazakhstan MOH lab Full retesting of each lot

MOH Testing Upon Import (not registration)

3/13/2014

5

Responsible for Routine Testing

Country Regulatory Requirement for Testing upon

Importation Comments

Argentina Company Test all parameters for all imported batches.

Brazil Company After Nov 2012, biotech products received as fully finished products

are not tested

Canada Company Test all parameters for all imported batches.

Chile Biotech - external lab and MOH Test some parameters for all imported batches. Can reduce testing if

you have 2 years/20 lots of consistency.

EU Company Test all parameters for all imported batches.

Japan Any lab Use JP standards

Korea Company Test all parameters for all imported batches.

Mexico biotech performed in lab authorized by MOH Must transfer methods to company in Mexico who transfers them to MOH. Reduced testing possible after 3 years of importation or 20

lots.

Peru Company

No biotech testing. Micro testing is required on all imported batches. Legislation was for MOH to perform testing only in the first lot

imported after initial approvals and 5 year renewals.

Uruguay Company

Not yet implemented but proposed to test all imported batches in

future.



• Registration Samples – Samples typically need to come from commercial supply chain and will need a COA from final packaging and distribution site that reflect commercial specifications

• Other considerations:

Samples need sufficient dating for shipping and testing

Need an import license before you can send samples

Registration Samples

MOH Testing (registration samples only)

Country Performed By Comments (including whether full retesting or different tests/standards used)

Colombia MOH lab Draft regulation would require testing for registration samples

Dominican Republic MOH lab Currently test registration samples only.

Ecuador MOH lab Currently test registration samples only.

El Salvador MOH lab Currently test registration samples only.

Guatemala MOH lab Currently test registration samples only.

Honduras MOH lab Currently test registration samples only.

Nicaragua MOH lab Currently test registration samples only.

Panama MOH lab Currently test registration samples only.


3/13/2014

6

Certificate of Pharmaceutical Product (CPP)

• The Certificate is needed by the importing country when the product is intended for registration

• Certification is issued by the source country in the WHO format

• Used by countries to assess the quality of pharmaceutical products for registration or importation

• WHO recommends to national authorities to ensure that analytical methods can be confirmed by the national laboratory

• Single product

• Establishes the status of the pharmaceutical product and of the applicant for this certificate in the exporting country.



• Exporting and importing country

• Name, dosage form and composition of the product

• Registration and marketing status of the product in the exporting country

• Number of product license and date of issue

• Product information

• Details on the applicant

• Statement to confirm whether or not the document is issued in the format

Examples of the type of information found on a Certificate of Pharmaceutical Product

Certificate of Pharmaceutical Product


• Many non US countries require a CPP prior to submission, but some can submit without it if local clinical development occurred.

• Examples;

China will allow the company to submit without a CPP if all development [CMC & Clinical] was performed in China

Mexico will allow a company to submit without a CPP if Phase 3 studies occurred in Mexico

Some countries will allow a company to submit without the CPP but need it prior to approval [need to be negotiable]

3/13/2014

7

Regional Considerations include:

• Specifications • Stability and shelf life • Samples/shipping • Import/export • Requirements for the local health authorities testing • Method transfer/Reagents/Training/Equipment • Submission Content • Comparators • ICH acceptance • Language


Comparators

• Comparators are used to establish the safety and efficacy of a drug in clinical trials

• Regional source (US vs. OUS)

Site

Test Methods

Limits

Comparability


Case Study - China MOH Testing

• Local regulations require that 3 registration samples be provided to Center for Drug Evaluation (CDE) Any cells needed for Bioassays also need to be provided

Importation laws will apply and required testing and certificate

• A product standard document is created to describe the procedures to be used for quality testing by the MOH There may be an opportunity for a company to prepare this

standard document to include additional detail that is not included in the analytical procedure.

• A positive outcome of this testing is needed before the Import Drug License will be granted


3/13/2014

8

Case Study - Russia

• Dossier is submitted to Ministry of Health (MOH) for review.

• MOH checks for completeness

• MOH sends the document to its expert organization for evaluation.

• The expert institution completes its review and send results back to MOH. • MOH approves clinical trials applications

• MOH makes final decision on issuance of a registration certificate

Applicants are allowed to communicate with MOH only. Limited interactions MOH controls and regulates all steps of the registration process


Case Study – Russia Information included in submission

• Application

• •IP rights (patents, trademarks)

• •Manufacturing sites (names and addresses )

• •GMP certificates of manufacturing sites (API, DP, packaging,

testing)

• •CPP issued for Russia

• •CoA for DP


Case Study - Russia

• Companies must provide drug samples and reference

standards for MOH testing of Drug Substance and Drug Product. Import License

• Required at the time of submission

• MOH sends material to expert lab for testing of drug and API

samples

• Successful testing by the MOH is the first part of the marketing authorization process


3/13/2014

9

Case Study - Russia


• A company submitted a registration dossier which contained the full analytical procedures

• No ability to provide additional method details

• Rejection notice from Russia for the application based on the outcome of the quality testing

• There is no provision in the Russian law to allow for the MOH to ask questions to resolve a laboratory issue during testing

Case Study – Russia Key Learning Points


• A successful marketing application is dependent on MOH testing of DS & DP

• There is no ability for the MOH to contact a company with questions or problems encountered during the testing The only option is to reject the application

• “New” Pre-submission consultations with MOH officials is now allowed since July 28, 2011

• Question - Is there a process that additional information can be provided in future applications when methodology is complex?

Case Study - Korea

• A company did not want to repeat all of the release tests, so they were asking for reduced testing

• The company was asked for a technical rationale and data to support why certain tests were not needed upon importation

• A scientific rationale was provided, KFDA accepted the approach


3/13/2014

10

Conclusions

• Early planning for method transfers to QC/MOH labs will be critical for complex biotechnology products

• Communication with the local MOH labs need to define the methods that will be transferred

• If all tests will not be repeated at local labs, a robust dataset will need to be available as justification

• For certain countries, i.e. Russia and China, more detailed instructions for complex methods is advised


Acknowledgements

• Susan Stolz

• Allison Wolf


Thank you



Workshop


Poonam Aggarwaal, Pfizer, Inc., USA

Carl Co, Biogen Idec, USA



Elena Semenova, Protein Sciences Corporation, USA


What are the regulatory expectations for correlating early and late-stage assays?

What studies are needed to replace target binding assays with more complex functional methods?

When would a binding assay be considered sufficient throughout development?

What are some strategies for development of bioassays to support development of drug

candidates with multiple MOAs

Can Quality by Design principles be used to support development and implementation of robust

bioassays?

When should structure activity relationship studies be performed?

How can results of structure activity relationship studies support establishment of specifications?

Is it possible to replace potency assays by physicochemical tests based on structure activity

relationship studies?

NOTES:

NOTES:

Bioassay Controls & Control Strategies Session Abstract

Session Chairs:

Evangelos Bakopanos, Health Canada and Sally Seaver, Seaver Associates LLC

Most bioassays that measure the potency of a therapeutic product measure the reaction of a living

system, either cells or animals, to different doses of the therapeutic. The response of this system to a

sample is compared to the response of the reference standard; the potency of the sample is proportional

to the relative shift in the sample curve from the reference standard curve. These assays also have

several procedural steps and custom components which make them highly susceptible to variability and

can impact their ability to generate reliable relative potency estimates.

Every well-developed bioassay has a series of controls or control strategies that ensure that they are

operating as intended. Typical controls for bioassays include procedural measures and assay design

elements to minimize assay variability, such as system/assay and sample suitability or acceptance

criteria. There are also statistical criteria for assessing the similarity of the reference standard and sample

curves.

This session will focus on bioassay controls and control strategies to assess assay validity and sample

validity for monitoring assay performance and ensuring that a given assay remains “fit for use”

throughout the various stages of its lifecycle.

NOTES:


Assay Acceptance Criteria for Multiwell-Plate-Based Biological Potency Assays

C. Jane Robinson

National Institute for Biological Standards and Control (NIBSC), Hertfordshire, United Kingdom

All analytical techniques require criteria to judge objectively whether a test has been executed properly.

Guidance has been developed previously for physicochemical assays but this guidance is not necessarily

appropriate or sufficient for bioassays. A paper entitled “Assay acceptance criteria for multiwall-plate-

based biological potency assays” has been published in BioProcess International 12(1) January 2014,

p30-41, as a draft for consultation. This paper lays out proposed guidelines for Assay Acceptance

Criteria and Sample Acceptance Criteria for bioassays. To prevent the document becoming unwieldy,

this set of guidelines has been limited to biological potency assays performed in multiwell plates.

Analytic dilution assays using multiwell plates are the most widely used platform for in vitro bioassays

and immunoassays, but multiwell plate formats introduce specific artifacts to the measured responses, so

assay design and appropriate acceptance criteria are of critical importance. This presentation will

address issues raised in the draft for consultation, including commonly used criteria, typical limits and

problems which can arise from setting inappropriate criteria. Comments are invited from all interested

parties to assist in compiling the final guidelines and can be submitted to the Biopharmaceutical

Emerging Best Practices Association at www.bebpa.org.

NOTES:

13/03/2014

1

Assay Acceptance Criteria for Multiwell-

Plate–Based Biological Potency Assays

Jane Robinson

[email protected]

National Institute for Biological Standards and Control, UK, www.nibsc.org .

CASSS Bioassays 2014: Scientific Approaches & Regulatory Strategies

24 – 25 March 2014, Silver Spring, Maryland USA

Quantitative biological assays (potency assays, bioassays) to

measure biological activity are essential

……for specifications, comparability & stability studies,

product & process development, …………

Multiwell plates - most common platform for in vitro bioassays &

immunoassays

CASSS Bioassays 24 – 25 March 2014 Silver Spring MD

• Convenient for handling large

numbers of doses & replicates

• Wide range of plate types with

standardized footprint

• Supporting equipment and

measurement systems

Multiple components of overall assay system must be within defined

limits for execution of a valid assay. Tested before or during assay =

System Suitability Testing (SST)

Available guidance primarily for physicochemical techniques

Bioassay responses susceptible to wide variety of factors

need to use data produced by individual assay to judge whether that

assay executed correctly

Focus on criteria that can be applied to assay results

Assay Acceptance Criteria (AAC) (often considered subclass of

SST). Classification as SST or AAC not necessarily important

SST & AAC must be appropriate to the assay system & purpose &

design of assay


For any analytical technique, need to judge objectively

Is the assay valid?

before considering the results for the test samples

13/03/2014

2

BioProcess International 12(1) January 2014 pages 30-41


Multiwell plate formats

…… introduce specific artifacts to the measured responses

• individual plates & well positions within a plate may be subject to

different conditions. Non-random distribution of samples & doses

(eg. all dilution curves for one sample in edge rows, or reference

standard in one plate & test sample in another) can introduce bias to

the measured relative potencies. Assay design is crucial

• Selection of acceptance criteria is inextricably linked to assay design

• This discussion uses mainly examples from cell-based potency

assays in 96-well plates, but applies to other assay systems (non-cell-

based functional, binding, immunoassays, ….) & plate formats


Assay acceptance & sample acceptance

Propose: 2 separate sets of acceptance criteria:

1) Assay Acceptance Criteria (AAC) based on responses of control

samples and reference standard

2) Sample Acceptance Criteria (SAC) based on responses of each

separate sample

If the plate fails AAC, then there is no processing of test sample data

If one test sample fails its SAC, then that particular test sample potency

quantification fails. Other test samples on the plate are assessed

separately


AAC pass

fail

SAC each

sample

13/03/2014

3

For most biopharmaceuticals, potency is assessed in a bioassay by

comparison of

dose-response curves of test material & reference standard

log dose

resp

on

se

fundamental requirement for obtaining a valid relative potency: reference

standard & test sample must behave similarly in the assay system

The dose–response curves have the same

mathematical form. Any displacement between curves

along the log-concentration axis is constant. The

displacement is used to calculate the relative potency

test sample standard

Assay control sample

AAC based primarily on comparison of the dose-response curves of

control sample(s) & a reference standard

at least one control sample should behave similarly to the standard &

the (expected behavior of) the test samples in this assay system. We

propose the name “assay control sample” for this control

assay control sample & standard are tested at multiple dilutions; other

controls may be tested at multiple / a few / a single dilution

AAC applied to each plate independently dilution curve for

assay control sample & dilution curve for standard run on every plate


Assay control sample

assay control sample as independent of standard & test samples in

origin as possible … with constraint all behave similarly in the assay

standard & assay control sample commonly different lots from the

same production method

using 2 dilution series of the standard is not sufficient: this tests only

the dilution procedure & subsequent handling

……..however ….in early assay development, 2 dilution series of the

standard may be the only possibility

assay control sample & standard should be as similar as possible to

test samples wrt formulation, concentration, etc.

assay control sample, standard & test samples should all be prepared &

tested in same fashion …. but …. an initial step of reconstitution,

dilution, etc, may be necessary


13/03/2014

4

Criterion of similarity of dose-response curves

log dose

resp

on

se

of assay control sample & reference standard & = essential AAC

of test sample & reference standard = essential SAC

• measure the response of each at several doses spread over an

appropriate range

• assessing similarity depends on statistical analysis

• for some pharmacopeial assays, the method of analysis may be

specified

test sample

standard

assay control sample

Four-Parameter Logistic (4PL) non-linear regression

curve fitting

concentration (x)

resp

on

se (y

)

• 4PL curve fit widely used

• sigmoidal curve symmetrical around the inflection point (5PL includes

asymmetry factor)

• if data heteroscedastic, can apply weighting algorithms

• each parameter may provide an acceptance criterion for determining

similarity of the dose-response curves

D = maximum asymptote

A = minimum asymptote

B = Hill Slope C = inflection point

y = D + A – D

1 + (x/C)B

Absolute versus relative values

• A potency assay is comparative: potency is measured relative to that of

the reference standard.

• ……… therefore absolute values for characteristics of a response

curve should not be critical

• ……… however, most assay systems function adequately over only a

limited range of conditions.

• The assay is validated for a defined range of conditions

• This is reflected by a limited range of dose-response curve

characteristics

• An unusually high or low value can indicate that the assay system is

not behaving in the usual manner within the validated range.

Examples: ED50, slope, ratio of upper asymptote to lower asymptote


13/03/2014

5

Evolution of Acceptance Criteria during Assay

& Product Development

Acceptance criteria & assigned values change during assay & product

development process

due to improvement in assay performance, accumulation of

more data on assay performance & stricter requirements at later

stages of product development

• limits on some criteria may be tightened

• some limits may need to be widened: initial data may not reflect the full

variation in assay conditions

• additional criteria may be adopted

• some criteria may be removed: it may become evident that they do not

reflect the validity of the assay & they could cause the rejection of

assays that are fit for purpose. Their values can still be recorded, often

“for information only” & can be useful for monitoring & trending.


Acceptance Criteria in Assay Monitoring &

Trending

The values observed for AAC & SAC can be used for assay & sample

monitoring & trending, helping identify aberrant results & whether the

assay is stable

Monitoring & trending a wide range of assay response characteristics is

useful, particularly in early assay development

Monitoring or trending a characteristic should not carry an expectation

that it will be adopted as an acceptance criterion. This discourages

investigation & selection of optimum acceptance criteria

Statistical analysis of the observed values, giving a

statistical process control (SPC) chart, permits

objective analysis of variation in assay performance

Statistical process limits & AAC & SAC are not the same as

product specification limits but they must be appropriate to

accommodate the product specification limits


Setting values & limits for acceptance criteria

using an assay control chart

Example: assay control sample potency

• multiple independent executions of the assay are performed with the

assay control sample included as a test sample

• the acceptance criterion for the potency of the

assay control sample is set as the

mean ± some multiple of the SD

commonly the statistical process control (SPC) limits are set at

± 3 SD …. but …..if the SD is based on only a few measurements, it

is a very uncertain estimate

• can use tolerance intervals - wide when based on a few measurements

& narrowing with more data

table of tolerance limits:

from 5 assays (multiplier ± 10.75) to 199 assays (multiplier ± 3.03)

Orchard T, Biopharm International, 19(11) 2006: 22-29

can establish initial acceptance criteria with (for example) 25 assays &

then revise as more data are collected


13/03/2014

6

Common current practices which cause

problems

Acceptance too dependent on an individual data point.

Eg: upper asymptote = mean of duplicate responses at max dose.

Max % difference between upper asymptotes assay control sample

& standard.

1 aberrant well at max dose of assay control sample on one plate.

Asymptote is aberrant compared with that of standard.

Plate is rejected.

Reportable value = mean of potencies from 2 replicate plates.

No reportable value is obtained

Inclusion of unnecessary, or unnecessarily tight, criteria.

Eg. in some systems ED50 can vary widely between assays which

are fit for purpose. Setting tight limits on an absolute value for

ED50 can result in rejection of an assay that is fit for purpose



Commonly used acceptance criteria

incorporating information from the delegates at the 2013 BEBPA Bioassay conference

Similarity of dose-response curves

Essential AAC & SAC. Used almost

universally

Slope

Widely used. Ratio – part of similarity testing &/or absolute value. Confidence

intervals or ranges eg 0.8-1.25.

Lower asymptote

Ratio – part of similarity testing &/or absolute value

Upper asymptote

Ratio – part of similarity testing &/or absolute value

Ratio upper to lower asymptote Ratio (=ratio of ratios) part of similarity testing &/or absolute value

Inflection point Not commonly used

ED50

Absolute value, widely used for trending but not commonly as AAC or SAC

……..continued…….


Commonly used acceptance criteria (continued)

Goodness of fit

essential AAC & SAC, used almost

universally. R2 common eg. R2≥0.95 to

R2≥0.98, simple but not sensitive

Note: error “R2≤” in publication of draft

Potency of control Assay control sample, full dose response curve, should be in every assay (& every

plate). At present, not always the case.

Typical limits 70-130% to 90-110%.

Commonly based on historical data

Ratio of values for different control samples

Not used commonly

Minimum number of doses used in curve fit

Common AAC & SAC, all doses used or exclusion of a few permitted, resulting in eg.

a minimum of 6 to 10 doses

Minimum number of doses in “linear” part of dose-response curve

Common AAC & SAC, both for linear & full curve fits. Values generally from 3 to 6

Minimum number of doses in upper &/or lower asymptote

Used in some assays, more during development. Most common value is 2

……..continued…….

13/03/2014

7


Commonly used acceptance criteria (continued)

Minimum dose range used in curve

fit

Sometimes explicit AAC & SAC (eg. dose

range 50-150% should lie within the linear

range). Often implicit, by specifying doses

tested & min number of doses in curve fit

Maximum number of statistical

outliers excluded

Common AAC & SAC, defined as eg.

number of doses or individual points per

curve or number of replicates per dose.

Many assays do not allow the exclusion of

any outliers or only if anomaly attributed to

experimental error

Variability of replicates

Essential AAC & SAC. Replicate wells / dilution series / aliquots, analysed from one

plate or across several. Expressed as %CV,

eg. 10 - 30%, or relative 95% confidence

interval around the mean eg. 80-125% or

75-133%.

Assay Acceptance Criteria for Multiwell-

Plate–Based Biological Potency Assays

Draft for Consultation C. Jane Robinson, Michael Sadick, Stanley N. Deming,

Sian Estdale, Svetlana Bergelson and Laureen Little

BioProcess International 12(1) January 2014 pages 30-41

http://www.bioprocessintl.com/journal/2014/January/Assay-

Acceptance-Criteria-for-Multiwell-PlateBased-Biological-Potency-

Assays-349245

( http://www.bioprocessintl.com/ then Archive then Archive by Issue )

or

http://www.bebpa.org/wp-

content/uploads/2012/11/BPI_A_141201AR05_O_231158a.pdf

( www.bebpa.org then White Papers )


Please send your comments on the draft

guidance ……..

Please note:

• Comments should be received by 31 May 2014

• All comments must be accompanied by a name, affiliation & email

address

• All comments will be given due consideration by the authors

• Any comment submitted may be posted on the BEBPA website

www.bebpa.org/white-papers/

• Comments may be posted in their entirety or extracts may be posted. In

the latter case, it will be noted that only part of the comment is posted

• If you would like your name & affiliation to be posted with your

comments, please state that BEBPA has your permission to do so


by email to the Biopharmaceutical Emerging Best Practices Association

at

[email protected] starting subject line with: AAC

http://www.bioprocessintl.com/journal/2014/January/Assay-Acceptance-Criteria-for-Multiwell-PlateBased-Biological-Potency-Assays-349245



















http://www.bioprocessintl.com/

http://www.bebpa.org/wp-content/uploads/2012/11/BPI_A_141201AR05_O_231158a.pdf



http://www.bebpa.org/

http://www.bebpa.org/white-papers/



mailto:[email protected]

“Edging Out” Edge Effects in a Cell-based Assay

Shelley Elvington


For many cell-based assays, long incubations at 37oC can lead to edge effects. One approach to mitigate

these effects is to limit the number of wells used in the assay to the inner 60. While this approach is

effective in avoiding edge effects, it also limits the amount of space available on the plate for use. Here I

present case studies in which alternative assay plates and cell-handling techniques were used to reduce

edge effects. These approaches can be used to alleviate edge effects without reducing the number of

usable wells on the plate, thereby increasing sample throughput and flexibility in plate format.


NOTES:

Near-universal Similarity Bounds for Bioassays

David Lansky

Precision Bioassay, Inc., Burlington, VT USA

Equivalence tests for similarity in bioassay are now broadly accepted; setting bounds for the tests is still

considered challenging. Sensitivity analyses illustrate factorial combinations of non-similarity

constructed via scaled parametric shifts and their impact on potency bias. Parameter-specific and

composite measures of similarity are compared for their ability to detect non-similarity that causes

potency bias. A simple combination to supplement scaled parametric measures will control potential

bias from correlated shifts in non-similarity parameters. With this modification we propose a complete

method for setting equivalence bounds informed by assay performance requirements (product

specifications) rather than assay capability.

NOTES:

Near-universalsimilarity bounds

for bioassays

D. Lansky

Abstract

Introduction

SensitivitySimulation:Experiment Plan

Results

Summary

Acknowledgements

Near-universal similarity boundsfor bioassays

David Lansky, Ph.D.

Burlington, Vermont, [email protected]

March, 2014

1 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Abstract

Equivalence tests for similarity in bioassay are now broadly

accepted; setting bounds for the tests is still considered

challenging. Sensitivity analyses illustrate factorial combinations of

non-similarity constructed via scaled parametric shifts and their

impact on potency bias. Parameter-specific and composite

measures of similarity are compared for their ability to detect

non-similarity that causes potency bias. A simple combination to

supplement scaled parametric measures wil control potential bias

from correlated shifts in non-similarity parameters. With this

modification we propose a complete method for setting equivalence

bounds informed by assay performance requirements (product

specifications) rather than assay capability.

2 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

The Problem

I Equivalence (now) non-controversialI Setting equivalence bounds?

I Set so (known) similar samples passI Set so historical pass rate okI Set to match reject rate of difference test

I None recognize or exploit the meaning ofequivalence measures

I Products w/wide therapeutic window haveI wide potency specificationI wider similarity bounds?

3 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Assumptions and Approach

I Parallel four parameter logistic model

I i.i.d. N(0, σ2) residualsI Examine nonsimilarity and measures via:

I Sensitivity analysis to nonsimilarityI Bias and variance of non-similarity estimatorsI Ease of setting equivalence boundsI Do equivalence bounds protect potency?

5 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

This work in context:

I Assay capability requires limits on potencybias (CP in USP<1033>).

I Sample acceptance criteria should ensuregood (i.e.; unbiased) potency.

I We demonstrate setting similarityequivalence bounds:

I that control bias in potency, andI can be set without data.

7 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Four Parameter Logistic

y ∗ =Ai

1 + e−Bi (log(x)−Ci )+ Di + ε

A = Response Range, B = ”Slope”, C = Log EC50, and D = No-dose Asymptote

Universal Scaled Similarity Parameters:

I %∆A = 100 × (ATest − ARef)/ A∗Ref

I %∆D = 100 × (DTest − DRef)/ A∗Ref (Not a typo)

I %∆B = 100 × (BTest − BRef)/ B∗Ref

∗ Long term averages

9 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Scaled Shifts w/consistent meaning

A: Range

B: Slope

D: No-dose Asy.

Black/Magenta pairs are reference/testA and D × (2/3, 1, 3/2) + 10%,B × (1/3, 1, 3) + 50%

11 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

No-dose Asymptote -5% shift

: B { -35 } : A { -5 }

: B { 0 } : A { -5 }

: B { 35 } : A { -5 }

: B { -35 } : A { 0 }

: B { 0 } : A { 0 }

: B { 35 } : A { 0 }

: B { -35 } : A { 5 }

: B { 0 } : A { 5 }

: B { 35 } : A { 5 }

13 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

No-dose Asymptote 0% shift

: B { -35 } : A { -5 }

: B { 0 } : A { -5 }

: B { 35 } : A { -5 }

: B { -35 } : A { 0 }

: B { 0 } : A { 0 }

: B { 35 } : A { 0 }

: B { -35 } : A { 5 }

: B { 0 } : A { 5 }

: B { 35 } : A { 5 }

15 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

No-dose Asymptote +5% shift

{ -35 }{ -5 }

{ 0 }{ -5 }

{ 35 }{ -5 }

{ -35 }{ 0 }

{ 0 }{ 0 }

{ 35 }{ 0 }

{ -35 }{ 5 }

{ 0 }{ 5 }

{ 35 }{ 5 }

17 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Experience w/”Scaled shift”bounds

I %∆A:B:D Range:Slope:no-dose AsymptoteI Excellent assays cannot pass 5:35:5

equivalence boundsI Noisy assays do not pass 10:50:10

equivalence bounds

I Visual sensitivity analysis:I 5% for no-dose asymptote and range, OK?I 35% slope seems okI Some combinations look non-similiarI Certain combinations likely to induce bias

19 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Sensitivity Experiment Plan

I CRD DesignI 6 replicates, 2 samples, 10 dilutionsI Residual SD: 1, 3, 6, and 9% of range

I Factorial Sample PropertiesI potency: 1/2, 1, 2I Shift in no-dose asy: -5%, 0, 5%I Shift in range: -5%, 0, 5%I Shift in slope: -35%, 0, 35%

I Performance Measures:I Geometric bias of potencyI % sample failure

I 999 simulated assays per condition

21 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Percent Geometric Bias of PotencyPGbias of Potency

Slope.delta

Perce

nt

−10

−5

0

5

10

15

−40 −20 0 20 40

11 1

11

1

1

33 3

33 3

33

3

66 6

66 6

66 6

99 9

99 9

99

9

: A.delta { −5 } : D.delta { −5 }

1 1 11 1 111 1

3 3 33 3 333 3

6 6 66

6 66 66

99 9

99 99 9

9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { 5 } : D.delta { −5 }

1 1 11 111

3 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { −5 } : D.delta { 0 }

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { 0 } : D.delta { 0 }

−10

−5

0

5

10

15

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { 5 } : D.delta { 0 }

−10

−5

0

5

10

15

1 1 111

1 13 3 333 3

3 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

11 1

11

11 1 13

3 33

33

3 3 366 6

66

66

6 699 9

9 99

99 9

: A.delta { 0 } : D.delta { 5 }

11 1

11

1

11 1

33 3

33

3

33 3

66 6

66

6

66 6

99 9

99

9

99 9

: A.delta { 5 } : D.delta { 5 }

Symbols (1, 3, 6, & 9) indicate the residual SD as % of response range

Color indicates sample potency (magenta=0.5, black=1, green=2).

Background color marks bias: Orange: over 20%, Yellow: 10%-20%

23 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Bias of Scaled Non-sim. w/1% SD

Median universal Range sim.

Slope.delta

Perc

ent

2

4

6

8

−40 −20 0 20 40

1 1 11 11 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { −5 } : D.delta { −5 }

1 1 11 1 11 1 13 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 9

9 99

: A.delta { 5 } : D.delta { −5 }

1 1 11 11 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 9

9 9 9

: A.delta { −5 } : D.delta { 0 }

1 1 11 1 11 1 13 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 0 } : D.delta { 0 }

2

4

6

8

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9

99 99

: A.delta { 5 } : D.delta { 0 }

2

4

6

8

1 1 11 11 13 3 33 3 33 3 36 6 66 6 66 6 69 9 9

9 9 99

9 9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1 1 11 1 11 1 13 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 0 } : D.delta { 5 }

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 9

9 99

9 9 9

: A.delta { 5 } : D.delta { 5 }

25 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Ratio Non-sim. w/1% SDMedian ratio Range sim.

Slope.delta

Perce

nt

5

10

15

20

−40 −20 0 20 40

1 1 11 11 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { −5 } : D.delta { −5 }

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 5 } : D.delta { −5 }

1 1 11 11 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 9

9 9 9

: A.delta { −5 } : D.delta { 0 }

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 0 } : D.delta { 0 }

5

10

15

20

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 5 } : D.delta { 0 }

5

10

15

20

1 1 11 11 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 9

9 9 9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 9

9 9 9

: A.delta { 0 } : D.delta { 5 }

1 1 11 1 11 1 1

3 3 33 3 33 3 3

6 6 66 6 66 6 6

9 9 99 9 99 9 9

: A.delta { 5 } : D.delta { 5 }

%RatioA = 100× ATest/ARef, %RatioD = 100× DTest/ARef,

%RatioB = 100× BTest/BRef

Median ratio non-similarity >2X true value with residual SD=1%.

Ratio estimates more sensitive to residual SD than universal.

27 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Chi-Square Equivalence Bound?Median Chi−square sim.

Slope.delta

Chi−s

quare

0.0

0.2

0.4

0.6

−40 −20 0 20 40

1

1 1

1

1

1

1

3

3 3

3

3 3

3

3 3

6

6 6

6

6 6

6

66

9

9 9

9

9 9

9

9 9

: A.delta { −5 } : D.delta { −5 }

1

11

1

11

1

11

3

33

3

33

3

33

6

66

6

66

6

66

9

99

9

99

9

99

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

11

11

111

1

133

33

333

3

366

66

666

6

699

99

999

9

9

: A.delta { 5 } : D.delta { −5 }

1

1 1

1

1

1

1

3

3 3

3

3 3

3

3 3

6

6 6

6

6 6

6

6 6

9

9 9

9

9 9

9

9 9

: A.delta { −5 } : D.delta { 0 }

1

11

1

11

1

11

3

33

3

33

3

33

6

66

6

66

6

66

9

99

9

99

9

99

: A.delta { 0 } : D.delta { 0 }

0.0

0.2

0.4

0.6

11

111

11

1

133

333

33

336

6

666

66

669

9

999

99

99

: A.delta { 5 } : D.delta { 0 }

0.0

0.2

0.4

0.6

1

1 1

1

1

1

1

3

3 3

3

3 3

3

3 3

6

6 6

6

6 6

6

6 6

9

9 9

9

9 9

9

9 9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1

11

1

11

1

11

3

33

3

33

3

33

6

66

6

66

6

66

9

99

9

99

9

99

: A.delta { 0 } : D.delta { 5 }

11

111

1

11

133

333

3

33

366

666

6

66

699

99

9

9

99

9

: A.delta { 5 } : D.delta { 5 }

With one test sample χ2nonsimilarity = (SSRFull − SSRReduced) /3

Estimates not sensitive to residual, sensitive to potency (potential for bias)

29 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Sample Pass Rate for 10:50:10

Universal pass rate at 10:50:10

Slope.delta (and target potency)

Perc

ent

0

20

40

60

80

100

−40 −20 0 20 40

1 1 11

1

11

1

13 3 33

3

3

3

3

36

6

6

6

6 6

6

6 6

9

9

99 9

9

9 9

9

: A.delta { −5 } : D.delta { −5 }

1 1 11

1

11

1

13 3 33

3

33

3

366

666 6

6

6 6

9

9

999

9

99

9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 1

1

1 1

1

1 13 3 3

3

3

3

3

3

3

66

6

6

6

66

6

6

9

9

99

9

99

9

9

: A.delta { 5 } : D.delta { −5 }

1 1 11

1

11

1

13 3 33

3

3

3

3

36

6

6

6

6 6

6

6 6

9

9

99 9

9

9 9

9

: A.delta { −5 } : D.delta { 0 }

1 1 11

1

11

1

13 3 33

3

33

3

366

666

6

6

6 6

9

9

99

9

9

9 9

9

: A.delta { 0 } : D.delta { 0 }

0

20

40

60

80

1001 1 1

1

1 1

1

1 13 3 3

3

3

3

3

3

3

66

6

6

6

66

6

6

9

9

99

9

99

9

9

: A.delta { 5 } : D.delta { 0 }

0

20

40

60

80

100 1 1 11

1

11

1

13 3 33

3

3

3

3

36

6

6

6

6 6

6

6 6

9

9

99 9

9

9 9

9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1 1 11

1

11

1

13 3 33

3

33

3

366

66

66

66 6

9

9

999

9

99

9

: A.delta { 0 } : D.delta { 5 }

1 1 1

1

1 1

1

1 13 3 3

3

3

3

3

3

3

66

6

6

6

66

6

6

9

9

99

9

99

9

9

: A.delta { 5 } : D.delta { 5 }

31 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Universal Similarity - Improved

Universal similarity (10:50:10) w/10% bound on A+D


Perc

ent

0

20

40

60

80

100

−40 −20 0 20 40

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { −5 } : D.delta { −5 }

1 1 11

1

11

1

13 3 33

3

33

3

36

6

666 6

66 6

9

9

99

999

99

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 1

1

1 1

1

1 13 3 3

3

3

3

3

3

3

66

6

6

6

66

6

6

9

9

99

9

99

9

9

: A.delta { 5 } : D.delta { −5 }

1 1 11

1

11

1

13 3 33

3

3

3

3

36

6

6

6

6 6

6

6 69

9

99 9

9

9 9

9

: A.delta { −5 } : D.delta { 0 }

1 1 11

1

11

1

13 3 33

3

33

3

366

666

6

6

6 6

9

9

99

9

9

9 9

9

: A.delta { 0 } : D.delta { 0 }

0

20

40

60

80

1001 1 1

1

1 1

1

1 13 3 3

3

3

3

3

3

3

66

6

6

6

66

6

69

9

99

9

99

9

9

: A.delta { 5 } : D.delta { 0 }

0

20

40

60

80

100 1 1 11

1

11

1

13 3 33

3

3

3

3

36

6

6

6

6 6

6

6 6

9

9

99 9

9

9 9

9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1 1 11

1

11

1

13 3 33

3

33

3

36

6

666

6

66 69

9

99 9

9

9 9

9

: A.delta { 0 } : D.delta { 5 }

1 1 11 1 11 1 13 3 33 3 33 3 36 6 66 6 66 6 69 9 99 9 99 9 9

: A.delta { 5 } : D.delta { 5 }

33 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Adjusted Geometric Total ErrorPercent Geometric Total Error adj. for failure, universal sim. 10:50:10


PGTEF

0

5

10

15

−40 −20 0 20 40

1

1 1

11 3

3 3

33 6

6 6

9

9 9

: A.delta { −5 } : D.delta { −5 }

11 1

1 11

1

33 3

66 6

9

9 9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 113 3 36 6 69

9 9

: A.delta { 5 } : D.delta { −5 }

1 1 111 33 333 66 6

99

9

: A.delta { −5 } : D.delta { 0 }

1 1 11 11 13 3 36 6 69

99

: A.delta { 0 } : D.delta { 0 }

0

5

10

15

1 13 3 36 6 69

9 9

: A.delta { 5 } : D.delta { 0 }

0

5

10

15

1 1 111 3

3 333 6

6 69

99

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

11 1

1

1

1 13

3 3

66 6

9

9 9

: A.delta { 0 } : D.delta { 5 }

1

1 1

3

3 3

6

6 6

9

9 9

: A.delta { 5 } : D.delta { 5 }

PGTEF = PGE“RMSTE

“log“R̂”””

= 100 ∗

0BBBBB@2

vuuutBias2

log2

“R̂”+ σ2

log2

“R̂”,

Psimilar

− 1

1CCCCCA ,

where:I Bias

log2

“R̂” = log2

“R̂”− log2 (R),

I R̂ and R are Estimated and True Potency,I Psimilar = Proportion Similar, and

I PGE (RMSTE) = Percent Geometric Expected Root Mean Square Total Error.

35 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Total Error After Improvement

PGTEF, universal sim. 10:50:10 w/10% bound on A+D


PGTE

F

0

10

20

30

40

−40 −20 0 20 40

: A.delta { −5 } : D.delta { −5 }

1 1 11 111

3 3 36 6 69 9 9

: A.delta { 0 } : D.delta { −5 }

−40 −20 0 20 40

1 1 113 3 36 6 69 9 9

: A.delta { 5 } : D.delta { −5 }

1 1 111 3 3 333 6 6 69 9 9

: A.delta { −5 } : D.delta { 0 }

1 1 11 11 13 3 36 6 69 9 9

: A.delta { 0 } : D.delta { 0 }

0

10

20

30

40

1 13 3 36 6 69 9 9

: A.delta { 5 } : D.delta { 0 }

0

10

20

30

40

1 1 111 3 3 333 6 6 69 9 9

: A.delta { −5 } : D.delta { 5 }

−40 −20 0 20 40

1 1 111

1 13 3 36 6 69 9 9

: A.delta { 0 } : D.delta { 5 }

9 : A.delta { 5 } : D.delta { 5 }

37 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Summary & Recommendations

I Use equivalence bounds for similarity

I Use scaled shifts for non-similarity

I Bias limits from assay purpose

I Chi-square (alone) not adequate

I 10:50:10 reasonable starting bounds

I Do limit positively correlated shifts inRange and No-dose Asymptote

39 / 42


for bioassays

D. Lansky

Abstract

Introduction


Results

Summary

Acknowledgements

Acknowldgements

I Consulting clients

I USP and USP bioassay panel members

I Carrie Wager

I Ramiro Barrantes

I Mark Blanchard

I NSF EPSCoR

I NIH SBIR 3R44RR02198-03S1

41 / 42

Health Canada Experiences with Bioassay Controls & Control Strategies

Omar Tounekti

BGTD, Health Canada, Ottawa, ON Canada

Bioassays are complicated analytical methods. They are based on living systems and require several

procedural steps and components which make them highly susceptible to variability. The degree of

variability displayed by bioassays directly impacts their ability to generate reliable relative potency

estimates. Although the development of any bioassay involves identifying sources of variability and

minimizing their impact, not all sources of variability can be defined. Moreover, certain sources of

variability, even when minimized, are unavoidable. Therefore, bioassays require a high level of control

in order to ensure that they are operating as per validation.

Typical controls for bioassays include procedural measures and assay design elements, as well as system

and sample suitability criteria. The specific set of controls (control strategy) applied to each bioassay

will depend primarily on assay type, assay knowledge and development data. This presentation will

focus on case studies where the proposed bioassays controls and control strategy were insufficient to

ensure that the bioassay method performed consistently from assay to assay and in agreement with

validated parameters. More specifically, this presentation will provide examples of gaps in the assay

design, system and sample suitability criteria and will highlight how these deficiencies precluded any

reasonable assessment of data contained in the regulatory file.

NOTES:

1

Health Canada Experiences with Bioassay Controls & Control Strategies

Omar Tounekti, PhD Senior Biologist/Evaluator Biologics and Genetic Therapies Directorate Health Canada

Disclaimer

The opinions of this presentation represent the speaker’s experience.

The contents do not necessarily reflect Health Canada official policy.

Biossays 2014 March 24-25, 2014

Presentation Outline

• Introduction

• Overview of biossays

• Bioassays controls and control strategies

• Case studies

• Concluding remarks


2

• Health Canada's BGTD is the Canadian federal authority responsible for the regulation of biological drugs and radiopharmaceuticals for human use.

• Products regulated by BGTD include: • biotherapeutics (cytokines, hormones, enzymes & monoclonal

antibodies)

• radiopharmaceuticals

• cell and genetic therapies

• viral, bacterial and combination vaccines

• blood and blood products

• cells, tissues and organs for transplantation


Organization

ICH Guideline Q6B

• Potency is the quantitative measure of the biological activity using a suitably quantitative biological assay (also called potency bioassay), based on the attribute of the product which is linked to the relevant biological properties.

• A relevant, validated potency assay should be part of the specifications for a biotechnological or biological drug substance and/or drug product.

• Products”


Test method development and the product life cycle


Preclinical

Clinical

Phase I Phase II Phase III

Post-approval

Definition/Design Evaluation Validation Verification

Knowledge

3

Use of Bioassays in the development and production of biologics

Drug candidate selection

Product characterization

Process validation

In-process testing

Lot release testing

Stability testing

Comparability studies


Examples of procedures used to measure biological activity:

•Animal-based biological assays, which measure an organism's biological response to the product.

•Cell culture-based biological assays, which measure biochemical or physiological response at the cellular level.

•Biochemical assays, which measure biological activities such as enzymatic reaction rates or biological responses induced by immunological interactions.

•Other procedures such as ligand and receptor binding assays, may be acceptable.


ICH Guideline Q6B

Bioassays are indirect (i.e. comparative) quantitative procedures: Bioassays are complex test systems that are susceptible to

many variables.

The performance of bioassays (and hence their biological readouts) can vary from day to day and especially from laboratory to laboratory.

The response of a bioassay system to a test material can not be used by itself to assign an absolute potency value.

The biological response of a test material is measured relative to that of a reference preparation.


Bioassays for Biotherapeutics

4

Relative potency assessment


• Bioassays can exhibit a greater variability than do chemically-based tests due to their reliance on biological substrates (e.g. animals, living cells, or functional complexes of target receptors).

• Inherent variability from instruments, reagents, day-to day, lab-to-lab variation.

• Intrinsic variability inherent in manufacturing biologicals

• Relative potency methodology


Why potency assay controls are critical?

1. Assay design

2. System suitability

3. Sample suitability


Potency assay controls & control strategies

5

Assay design:

• While assay development should be focused primarily on the properties of potency efforts to identify and control variation in the concentration-response relationship are also appropriate.

• Example of factors that may affect bioassay response: cell thawing; plating density and confluence; culture vessels; growth, staging, and assay media; serum requirements; incubation conditions (temperature, CO2, humidity, culture times from thaw); cell harvesting reagents and techniques; cell sorting; cell counting; determination of cell health; cell passage number and passaging schedule; cell line stability; and starvation or stimulation steps.


USP <1032> Design and Development of Biological Assays

System suitability:

• System suitability in bioassay, as in other analytical methods, consists of pre-specified criteria by which the validity of an assay (or perhaps a run containing several assays) is assessed.

• Analysts can assess system suitability by determining that some of the parameters of the Standard response are in their usual ranges and that some properties (e.g., residual variation) of all the data are in their usual range.



Sample suitability:

• Sample suitability in bioassays generally consists of the assessment of similarity, which can only be done within the assay range.

• Relative potency may be reported only from samples that both show similarity to standard, exhibit requisite quality of model fit, and have been diluted to yield an EC50 (and potency) within the range of the assay system.



6

Bioassay controls and control strategy are insufficient to

ensure that the method is performing consistently from

assay to assay and in agreement with validated

parameters and precludes any reasonable assessment

of the data contained in the submission/file. More

specifically:

• Gaps in the assay design

• Inadequate system suitability

• Inadequate sample suitability


Recurrent issue related to bioassay controls & control strategies

1. Parallelism (similarity)

2. Assay control sample

3. Replicate variability

4. Parameters of the reference standard response

5. Assay design


Case studies involving bioassay controls & control strategies

• Questionable release, stability and comparability data

provided in the file.

• Request of additional information

• Influence product ‘s Lot Release categorisation

• May lead to the issuance of a Notice of Deficiency

• Sponsor is typically requested to provide an action plan

and an impact assessment of applying new criteria to the

data supplied in the original file.


Consequences of inadequate controls and control strategies

7

• Bioassay results are key at all stages in the product life

cycle, from early research work to final quality control of

finished products.

• Assay choice and design are crucial. Oversights result

often in assays that are difficult to perform, time

consuming and that generate questionable data.


Concluding Remarks

Acknowledgments

Dr Maria Baca-Estrada

Dr Evangelos Bakopanos


Thank you

Bioassay Controls & Control Strategies Workshop


Shelley Elvington, Genentech, a Member of the Roche Group, USA

David Lansky, Precision Bioassays, Inc., USA

Tsai-Lien Lin, CBER, FDA, USA


C. Jane Robinson, NIBSC, United Kingdom

Omar Tounekti, BGTD, Health Canada, Canada


What aspects of a bioassay are controlled and how is this accomplished?

How and when are system and sample suitability (acceptance) criteria set? When are they

revised?

How and when are statistical requirements set?

What methods are used to monitor trends in system or sample suitability (acceptance) criteria?

How can bioassay controls be used to identify problems?

How can bioassay controls be used to troubleshoot problems?

What happens to bioassay controls following assay improvements?

NOTES:

NOTES:

Poster Abstracts

P-01

Optimization of Assay Design and Data Analysis for a Competitive Antibody-antigen Binding

Assay with Higher Variability

Wei Zhang; Deepthi Kanuparthi; Svetlana Bergelson

Biogen Idec, Cambridge, MA USA

We have developed a competitive binding assay to measure the binding of an antibody to its antigen.

Due to the existence of different forms of the antigen and the different binding affinities of the antibody

to them, this assay showed high variability compared with similar assays. In order to develop an assay

that can accurately measure the antibody-antigen binding, we tested different assay formats, conditions,

plate layouts and data analysis methods. In the end, we overcame the higher variability and developed an

electrochemiluminescent (ECL) competitive binding assay on the Meso Scale Discovery (MSD)

platform for the antibody with satisfactory accuracy, precision, linearity and specificity by (1) using a

plate layout with four independent data blocks; (2) plotting curves with data from two replicates; (3)

transforming raw data with square root; (4) using F-test for curve parallelism test.

NOTES:

P-02

Successful Transfer of a Cell Based Potency Assay for a Biological Product

Evelyn Kilareski1; Anthony Burkholder

1; Weihong Wang

1; Robert Donatelli

1; Miriam Franchini

2; Larry

Anderson2

1Eurofins Lancaster Laboratories, Inc., Lancaster, PA USA;

2Smith & Nephew, Inc., London, United

Kingdom

A cell based potency assay measures the physiological response elicited by a given product. It is often

the preferred format for determining the activity of biological products and is commonly employed for

lot release as well as stability testing. Transfer of a cell based potency assay between laboratories can

pose significant challenges due to the complexity of the assay. For a marketed product, method transfer

is under direct scrutiny by regulatory agencies and therefore is an even more significant undertaking.

This poster presents a case study to demonstrate how a cell-based potency assay can be successfully

transferred to a third party contract testing laboratory.

NOTES:

P-03

A Reporter Gene Bioassay for Potency Assessment of a Therapeutic Monoclonal Antibody

Tianmeng Shao; Hyun Jun Kim; Xu-Rong Jiang; Michael Washabaugh


We developed and optimized a novel reporter gene bioassay for quantifying the bioactivity of a

therapeutic monoclonal antibody. The antibody blocks the binding of a ligand on the antigen-presenting

cells to a receptor on T cells. The inhibition will sustain T cell activation and T cell immune function by

blocking the negative regulatory signals generated by the binding of the ligand to its receptor. A Nuclear

Factor Kappa-light-chain-enhancer of activated B cells (NFB) cell line has been engineered that

expressed both the receptor and an NFB-luciferase reporter gene. The reporter gene assay can measure

NFB activity in T cell that is proportional to T cell activation. Test samples caused a concentration-

dependent induction of NFB activity in T cells that is measured relative to a reference standard. The

current bioassay format is highly robust, simple to perform, and amenable for use in a regulated

environment.

NOTES:

P-04

Better Cell-Based Assays for Anti-CTLA-4, Anti-PD1/PD-L1, and Bispecific Immunotherapy

Drug Studies

Mei Cong; Pete Stecha; Natasha Karassina; Jim Hartnett; Zhi-Jie Jey Cheng; Frank Fan

Promega Corporation, Madison, WI USA

Cancer immunotherapy was named the 2013 “Breakthrough of the Year” by Science. It aims to stimulate

a patient’s own immune system to treat cancer. Key inhibitor drugs in the market or under clinical

investigation such as nivolumab and lambrolizumab (anti-Programmed cell Death protein 1 [PD1]

antibodies), and ipilimumab (Yervoy, an anti-Cytotoxic T-Lymphocyte Antigen 4 (CTLA-4) antibody)

are providing long-term benefits to cancer patients. Catumaxomab, a bispecific and bifunctional

antibody fusion protein, is another unique way to engage T cells for cancer therapy by allowing targeted

delivery and activation of immunoregulatory signaling pathways. Cytokines and interleukins also belong

to this category by nonspecifically stimulating T cell proliferation. Traditional inhibitory blockade

studies commonly use animal models or freshly isolated peripheral blood mononuclear cells (PBMCs) to

quantify antibody drug biological activity. However, animal models can be cost-prohibitive and cell

acquisition and preparation is labor intensive and the bioassays incorporating such cells have high

inherent variability. Here we demonstrate multiple bioluminescent reporter-based assays that can be

used to rapidly measure potencies of multiple biological immunotherapy drugs. We also demonstrate

that these bioassays reflect mode of action of each drug, and quantify potencies of on-market

monoclonal antibody drugs for cancer.

NOTES:

P-05

Characterization of FcRn Binding Kinetics of Therapeutic Antibodies

LeeAnn Machiesky; Kenneth R. Miller; Xu-Rong Jiang; Michael Washabaugh


The neonatal Fc receptor (FcRn) recycles IgGs within endothelial cells and rescues them from a

degradative pathway thus increasing antibody half-life. A binding assay has been developed using label-

free surface plasmon resonance (SPR) technology to assess the binding affinity and kinetics for the

interaction between human FcRn and IgG molecules. FcRn - IgG binding kinetics are fit using a

heterogeneous ligand model, which assumes there are two classes of non-interacting sites on FcRn. At

lower IgG concentrations, the higher affinity site is occupied, and at higher IgG concentrations the lower

affinity sites are also occupied. This model generates both a high and low affinity binding constant (KD1

and KD2) for the interaction. We evaluated the binding affinity and kinetics for several different IgG

molecules, across multiple manufacturing lots, including ones which had been engineered to have

enhanced FcRn binding. There was no observable difference in the association rate constant for any of

the IgG molecules; however, an observable difference was measured in the dissociation rate constant

(kd), with IgG molecules that were engineered to have enhanced FcRn binding, having a 10-fold lower

kd value.

NOTES:

P-06

Establishment of an ADCC Screening/Characterization Assay and a CDC Assay for Anti-

TNFalpha Therapeutic Antibodies (e.g. Humira®)

Frances Brauer; Manuela Schmid; Alexander Knorre

BSL BIOSERVICE Scientific Laboratories GmbH, Planegg / Munich, Germany

Cytotoxicity is a mechanism of action of antibodies through which virus-infected or other diseased cells

can be killed, either by components of the cell-mediated immune system (ADCC, Antibody-dependent

cell-mediated cytotoxicity) or by the complement system (CDC, complement-dependent cytotoxicity).

An ADCC surrogate assay using Promega’s ADCC Reporter Assay and antibody-specific target cells

was established for use as screening/characterization assay. Potency determination is performed using 4

parameter logistic fit. This assay can be used for anti-TNFalpha therapeutic antibodies, e.g. Humira

(Adalimumab), Enbrel (Etanercept) and Remicade (Infliximab). The anti-TNFalpha therapeutic antibody

CIMZIA (Certolizumab pegol) which lacks the ADCC-inducing Fc part was used as negative control.

Assay dose response curves with continuously-cultured or with frozen, thaw-and-use target cells will be

shown. In addition, a CDC assay with luminescence readout and human serum as complement source

for anti-TNFalpha therapeutic antibodies was established. A representative dose response curve with 4

parameter logistic fit analysis will be shown.

NOTES:

P-07

Automation of a Bioassay in QC Laboratories for Routine Lot Release and Stability Testing

Lichun Huang; Ariel Margulis; Ai Shih; Wei-Meng Zhao; Parth Sampathkumar; Joseph Marhoul


As part of continuous improvement efforts, an automated method to determine clot lysis activity has

been developed and validated to replace the plate based manual method. The ACL TOP system is a

fully automated, stand-alone random-access multiparameter coagulation analyzer. Using this system,

the potency method fully utilizes the automated pre-dilution feature and data generated are exported to

enable parallel line analysis for potency calculation. The automated assay passes tightened accuracy

(recovery of 95-105%) and precision (inter-assay CV of <5%) criteria. A high level of inter-instrument

precision allows the instruments to be used interchangeably. The stability indicating properties are

confirmed using a panel of samples stressed under various conditions. The design of comparability is

based on two one-sided tests (TOST) with a criterion of 95% confidence interval of mean difference

between methods falling within +/- 1.7% of the mean. Comparability testing performed using 44

samples that include representative samples from lot release and stability program, stability retains that

past normal shelf-life, stress panel, and high concentration samples to simulate hyper-activities confirms

equivalence between the automated and plate based manual methods. The transfer of the assay to QC

laboratory is evaluated using data tested at recipient laboratory to data obtained at the donor laboratory

in the comparable timeframe. Since there was limited history for the Clot Lysis by ACL TOP assay at

the time of transfer, the maximum acceptable difference between the laboratories was conservatively

established as the 99% confidence limit of the validation SD. Since the implementation and transfer,

method monitoring data showed a tightened control trend with 100% success rate.

NOTES:

P-08

Bioassay Design and Analysis Strategies

David Lansky


Bioassay designs are constrained by practical considerations, statistical principles, and variation in

biological materials. Good design, lab technique, and analyses combine to yield high performance

bioassays. Simulation experiments with properties that mimic real bioassays illustrate the impact of

strategic design and analysis choices on bioassay performance over useful ranges of potency and

variation (both within and between assays). This presentation illustrates good designs and analysis

strategies for slope ratio, parallel line, and parallel four parameter logistic curve bioassays. Each of these

strategic approaches has advantages and disadvantages; these are compared.

NOTES:

P-09

Near-universal Similarity Bounds for Bioassays

David Lansky


Several versions of composite and parameter-specific measures of non-similarity will be described and

compared. Two approaches to sensitivity analyses yield new ways to examine the potential impact of

various amounts of non-similarity and offer a good rationale for nearly universal similarity equivalence

acceptance limits (scaled parametric shifts). Factorial combinations of various amounts of parametric

non-similarity are used in simulations to compare different candidate measures of non-similarity.

Candidate non-similarity measures are compared using median simulated sample failure rates and

median geometric bias of potency. Scaled parametric equivalence bounds require a modification to

control potential bias from correlated shifts in non-similarity parameters; with this modification we

propose a complete method for setting equivalence bounds informed by assay performance requirements

(product specifications) rather than assay capability.

NOTES:

NOTES:

Documents

March 24 - 25, 2014 Sheraton Silver Spring Hotel Silver ......David Lansky, Precision Bioassays, Inc., USA Tsai-Lien Lin, CBER, FDA, USA Thomas Anders Millward, Novartis Pharma AG,