A lifecycle approach to bioassay validation Timothy Schofield Senior Advisor, GSK Global Vaccines Technical R&D 3rd Statistical and Data Management Approaches for Biotechnology Drug Development October 12, 2016, Rockville, MD

Outline – – – – – –

Quality by design for analytical methods The analytical target profile Lifecycle stages Design and development of a bioassay Bioassay qualification Continued performance verification

IABS Lifecycle of Bioassay 10-12-16

Validation

2

Quality by Design for analytical methods AQbD – Industry and regulators have begun to recognize that analytical methods generate a product – measurements – Like pharmaceutical products, measurements should have adequate quality to meet their intended use – making a decision – The fundamental goals of product development are: – Safety and efficacy (hitting the clinical target) – Variance reduction

– The fundamental goals of analytical development are: – Accuracy (hitting the analytical target) – Variance reduction IABS Lifecycle of Bioassay 10-12-16

3

AQbD (cont.) – Many of the concepts associated with QbD for pharmaceutical products translate to concepts related to analytical methods Process Concept

Analytical Counterpart

Quality Target Product Profile (QTPP) • Target manufacturing, and commercial requirements Critical Quality Attributes(CQAs) • Potency • Purity Specifications (acceptance criteria) • 80% to 125% potency • Purity > 95%

Analytical Target profile (ATP) • Target analytical performance, testing laboratory, and customer requirements Performance attributes (validation parameters) • Precision • Accuracy Acceptance criteria • %GCV < 10% • LLOQ > 1 ng/mL

Critical process parameters • pH, time, temperature

Critical method parameters • pH, time, temperature

Process design space

Method operating design region (MODR)

Process control strategy • Comparability protocols • Tech transfer Continuous process verification • Continuous review and updating of process knowledge

Assay control strategy • Comparability protocols • Method transfer Continuous performance verification • Continuous review and updating of analytical knowledge

IABS Lifecycle of Bioassay 10-12-16

4

AQbD (cont.) – USP stimulus to the revision process - Lifecycle Management of Analytical Procedures: Method Development, Procedure Performance Qualification, and Procedure Performance Verification, PF 39(5) – Modeled after FDA Validation Guideline – Moving analytical methods into the world of QbD – Follow-on workshops – 8-9 Dec 2014, Rockville, MD – 7-9 Nov 2016, Prague, CR

IABS Lifecycle of Bioassay 10-12-16

5

AQbD (cont.) – Fundamentals – “Validation” is a demonstration of fitness-for-use – “Validation” is a continuous process – “Demonstration” should include consideration of risks – Risk of making decisions from an “invalid” procedure

– Risk of invalidating a procedure which is in control

– Risk is directly related to “uncertainty” – Movement towards metrology and ISO (International Organization for Standardization) standards

IABS Lifecycle of Bioassay 10-12-16

6

Uses of potency assays during the product lifecycle Platform

Product Lifecycle

Use(s)

in vivo (animal potency in vaccines)

Research

Developability

Development

Product characterization in vivo/in vitro concordance Release/stability

Commercial

Comparability

in vitro (infectivity/ IVRP in vaccines; cell based bioassay in biopharms)

Research

Characterization

Development

in vivo/in vitro concordance

Commercial

Release/stability

Physical/chemical

Research

Developability

Development

Product characterization

Commercial

Release/stability

IABS Lifecycle of Bioassay 10-12-16

7 11(3)

The analytical target profile (ATP) – The list of requirements throughout the bioassay lifecycle (from method selection to method retirement) – Business requirements – throughput, cost, ease of implementation – Performance requirements – specificity, precision, accuracy, LOQ

– Regulatory requirements (ICH Q2)

IABS Lifecycle of Bioassay 10-12-16

8

The analytical target profile (ATP) ala USP

– “The procedure must be able to quantify [Analyte] in [presence of X, Y, Z] over a range of A% to D% of the nominal concentration with an accuracy and uncertainty such that the reportable result falls within ±B% of the true value with at least a P probability determined with C confidence.”  – Note: -content and -confidence tolerance interval: 𝑃 𝑈−𝐿 ≥𝛽 ≥𝛾 W. Shewhart: Economic Control of Quality of Manufactured Product. Van Nostrand Company, Inc., New York, 1931, republished in 1980 as the 50th Anniversary Commemorative Reissue, 1981, by ASQC Quality Press, Milwaukee IABS Lifecycle of Bioassay 10-12-16

9

Lifecycle stages (ala USP) – Stage 1 – Method design, development and understanding – Method selection and optimization

– Stage 2 – Procedure performance qualification – “. . . confirms the analytical procedure is capable of delivering reproducible data that consistently meet the performance criteria defined in the ATP while operated subject to the noise variables that may be experienced.”

– Stage 3 – Procedure performance verification – “. . . routine monitoring of the analytical procedure's performance and evaluation to determine if the analytical procedure, as a result of any change, is still fit for purpose.”

IABS Lifecycle of Bioassay 10-12-16

10

Stage 1: Design and development of a bioassay – Selection of method(s) – Bases – release, stability, process/formulation development, characterization, lifecycle management – Potency assay methods – In vivo bioassay (mouse potency or challenge assays) – In vitro bioassay (antigencity in vitro – IVRP; cell based bioassay) – Physical/chemical – properties related directly to potency (e.g., oacetyl content in vaccines; fucose content in monoclonals)

– Design – relative potency bioassay

Relative Potency Determination 220

– Parallel curve bioassay (4PL)

– Complex assessment of similarity

Standard

140 120

Dilutional Linearity

RP

100 80 60

200

40 20

Response

– Robust to changes in sensitivity

Test

180 160

Response

– Range should support its uses

200

150

0 1

10

100

Concentration

1000

10000

100 50

0

IABS Lifecycle of Bioassay 10-12-16

1

10

100

Concentration

1000

11

Stage 1: Design and development of a bioassay (cont.) – 26-2 fractional factorial (Resolution IV)

– Optimization using DOE – Screening design and results – Factors and levels

– Precision profiles

Exp. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Capt. 250 250 250 750 250 750 750 250 750 750 250 750 750 250 250 750

Biot. 250 600 600 600 600 600 250 250 600 250 250 250 600 600 250 250

EnzCult 750 300 750 300 750 300 300 750 750 300 300 750 750 300 300 750

Vol. 100 100 50 50 50 50 100 100 100 100 50 50 100 100 50 50

Incub. 1 1 3 1 3 1 3 1 3 3 3 1 3 1 3 1

NBCl 1 4 1 1 1 1 1 1 4 1 4 4 4 4 4 4

– Factor effects

12

Stage 1: Design and development of a bioassay (cont.) – The method operates reproducibly at the optimum, leaving room to determine a range which ensures quality measurements (adequate range) – Using Bayesian analysis to determine the design space:

There is room left for the factors to vary while ensuring adequate performance

𝑓(𝑋1 , 𝑋2 , … ) = 𝑏0 + 𝑏1 𝑋1 +∙∙∙

𝑓(𝑋1 , 𝑋2,… )

±e

DSp=0.65

𝑒(𝑏0 , 𝑏1 , … ) X 10,000 IABS Lifecycle of Bioassay 10-12-16

13

Stage 1: Design and development of a bioassay (cont.) Arrhenius Plot of Recombivax IVRP

– in vivo/in vitro concordance

1.5

1 0.5 0 ln(b)

-0.5

-1 -1.5 -2 -2.5 2.9

3

3.1

– Strategic studies can be performed to establish the concordance between methods

– Accelerated stability, manufacturing & clinical data show direct proportionality

3.4

3.5

Stability, manufacturing & clinical concordance Manufactured Lots

3.20

90

Stability Experiment Unit Slope Line

ED50

1.60

188

Clinical Experience

0.80

375

0.40

750

0.20

1500

0.10

IVRP: 0.10

– Relationship used to “calibrate” in vitro specification Schofield, T. (2002) in vitro versus in vivo concordance: A case study of the replacement of an animal potency test with an immunochemical method, Dev. Biol., 111, 295-304 IABS Lifecycle of Bioassay 10-12-16

3.3

ln(ED50) = -1 ln(IVRP)

– Concordance: direct proportionality between methods – y = bx – Arrhenius analysis of accelerated stability data used to target potency levels

3.2

1000/Absolute Temperature

GMT

2.8

3000 0.20

0.40

0.80

1.60

IVRP limit (0.50) corresponding to in vivo specification (1.5) 2.00 Old Seed (n=6) Redemo (n=12)

1.50

New Seed (n=158)

ED50

– in vivo methods are inherently more variable than in vitro assays, and thus less effective controls of product quality

`

0.50 0.25

0.25

0.50

IVRP

1.00

2.00

14

Stage 2: Bioassay qualification – Historically known as bioassay validation – “Validation” is a demonstration of fitness-for-use – One use is to control product through specifications – Scientifically/clinically justified minimum and maximum quality requirements

Maximum Requirement

Release Specifications

Control Lim its

Minimum Requirement

-6

0

6

IABS Lifecycle of Bioassay 10-12-16

Shelf-Life

12

18

24

30

36

Requirements

– Release limits determined to ensure (with specified confidence) that the CQA meets its quality requirements at release and throughout shelflife – Control limits associated with process 15 control

Stage 2: Bioassay qualification (cont.) – “Classical” (parameters) approach

Cpm 

– Requirements on accuracy and precision

Upper SpecificationLimit Lower SpecificationLimit 2 2 6  ˆ Pr  RB 2  Re oduct lease

2 where ˆ Pr is an estimate of product variability, RB is relativebias, oduct

– Approach in USP Biological Assay Validation

2 and Re is release assay variability. lease Cpk and probability of OOS for various restrictions on RB and IP

– Combinations assessed to address manufacturing risk – Prob(OOS)

LSL-USL

IP (%)

RB (%)

Cpk

Prob(OOS)

0.7-1.43

0.20

0.20

0.56

(9.30%)

0.7-1.43

0.08

0.12

0.98

(0.33%)

0.7-1.43

0.10

0.05

1.62

(0.0001%)

1

– Requirement to meet specifications can be represented as isobars of constant probability (acceptable region)

– Allows lab to balance accuracy and precision (e.g., can tolerate more variability for a very accurate procedure) – Designed to manage study risks and effectively assess ruggedness factors n

IABS Lifecycle of Bioassay 10-12-16

(t ,n1  t / 2,n1 )2  s 2 ac 2



(1.89  2.36)2  log(1.08)2 log(1.12)2

 8.

Designs in 8 runs

Factors

Design

2

2 x 22

3

23

4

24-1

Hoffman, D., et.al., (2007) A Total Error Approach for the Validation of Quantitative Analytical Methods, Pharm. Res., 24(6), 1157-64.

Stage 2: Bioassay qualification (cont.)



– VC analysis to acknowledge factors in the study – Confidence interval on the overall variability can be compared to the REML Estimates Component acceptance criterion Variance Estimate Var(operator) 0.00000 Var(cell) – Components associated with study Var(operator*cell) 0.00245 0.00000 Var(run(operator*cell)) 0.00217 factors yield higher estimate of IP Var(Error) 0.00076 – An “equivalence” approach can be used to assess conformance to the acceptance criterion on relative accuracy – Mixed effects model combining levels to reduce risk due to correlation among results across levels – Total error (results) approach

UCL = 11.3%

VC Intermediate Pr ecision  100   e  i  1  



 100  e

0.00245 0.00217 0.00076



 1  7.6%

12%

0%

-11%

0.50

0.71

1.00

1.41

2.00

– Tolerance interval on combined accuracy and precision IABS Lifecycle of Bioassay 10-12-16

17

Design & development of a bioassay (Revisited) – The “reportable value” is the composite of assays performed in replicate – Reportable values are associated with uncertainty – Expressed as confidence/tolerance intervals – Managed through design

𝑆𝐸 = 100 ∙ 𝑒

𝑉𝑎𝑟 𝐴𝑠𝑠𝑎𝑦 𝑛+𝑉𝑎𝑟(𝑅𝑒𝑝) 𝑛𝑘

Number of assays (n)

– Assays x reps – Other design (assay) factors: – Analysts – Instruments

−1

Reps (k)

1

2

3

6

1

7.2%

5.1%

4.1%

2.9%

2

6.4%

4.5%

3.6%

2.6%

3

6.0%

4.2%

3.4%

2.4%

6

5.7%

4.0%

3.3%

2.3%

– VC’s can be used to design studies for other uses of bioassays in the most effective/efficient manner (e.g., method transfer, standard qualification/calibration, process development, etc.) IABS Lifecycle of Bioassay 10-12-16

18

from Schofield, De Montfort QbD for Biopharmaceuticals, 13Jan2015

Stage 3: Continued performance verification – Statistical process control on system suitability parameters which are related to bioassay performance (e.g., precision) – Hill coefficient (sometimes called slope) of the reference curve – Together with the upper asymptote; or as effective range (A-D) – Related to precision and range

– Lower limit of quantitation (LLOQ) from precision profile – Related to sensitivity

– LLOQ and ULOQ from precision profile – Related to range

– Independent control sample – Holistic measure of method performance

– Response at a selected dose – Geometric mean titer of mice in an bioassay – Preferably close to the ED50 IABS Lifecycle of Bioassay 10-12-16

19

Stage 3: Continued performance verification (cont.) – Method changes (transfer, technology, new standard)

Shelf-life Limit

– A margin () can be established which is associated with low risk of an OOS if the margin is reached (manufacturer's risk) while the patient is protected by the release limit

Release Limit

0

IABS Lifecycle of Bioassay 10-12-16

6

12

18

24

30

36

Time (mos.)

– The “comparability study” design addresses the risks of making a bad decision – An equivalence test (or noninferiority) is performed to demonstrate conformance to the equivalence margin (upper confidence bound falls within the equivalence margin)

Control Limits



Equivalent )

-1

0

1

)

Not equivalent 20

Summary – Bioassay validation follows the stages developed for process validation, and represents a method of ensuring fitness for use(s) throughout its lifecycle

– Strategic design and development of the bioassay, together with appropriate system suitability parameters, combine to build and maintain quality of potency measurement – Potency may be measured in vivo, in vitro, or chemically; careful method selection during development and QC facilitates process development and patient efficacy – Statistical approaches to validation help reduce and communicate risks of bioassay performance

IABS Lifecycle of Bioassay 10-12-16

21

Acknowledgements – – – – –

Amin Khan Barbara Capecchi Cristiana Campa Alex Pysik Bill Egan

– Rick Burdick

IABS Lifecycle of Bioassay 10-12-16

22

Thank you [email protected]

IABS Lifecycle of Bioassay 10-12-16

Tim Schofield is an employee of the GSK group of companies.

23