Quality by Design A Perspective From the Office

Quality by Design: A Perspective From the Office of Biotechnology Products ADVISORY COMMITTEE FOR PHARMACEUTICAL SCIENCE October 26, 2005 Barry Cherney, Ph. D. Deputy Director DTP/OBP/CDER

Overview • Introduction of Biotech Products – defining the • • issues OBP Practice Designing a Quality Product Designing a Quality Process Implementation

Office of Biotechnology Products • Therapeutic Proteins • • m. Abs (related products) These proteins are produced from recombinant or non recombinant cell culture expression systems and from transgenic and non transgenic systems Products transferred from CBER to CDER in October 2003 Excludes ONDCQA regulated protein products • • – Growth Factors –Enzymes –Cytokines –Chemokines –Angiogenic factors –Toxins –Soluble Receptors/Receptor antagonists

Biotechnology Products Biotechnology products tend to be: • Large, complex molecules • Mixtures of many active ingredients - Subject to extensive heterogeneity in quality attributes of the API • Dependent on higher ordered structures and many • times, flexibility (e. g. changes in conformation) Are sensitive to small changes in manufacturing and impurity profiles, conformation stability limited

Product Variability • Carbamylation • Amino Acid Substitution • Truncation • Carboxylation • Mismatched S-S bonds • Formylation • N- and C-terminal difference • Gamma Carboxyglutamic acid • Aggregation • O-linked Glycosylation • Multimer Dissociation • N-linked Glycosylation • Denaturation • Methylation • Acetylation • Oxidation • Acylation • Phosphorylation • Addition of lipid • Sulphation • Amidation/Deamidation

Biotechnology Products • Generally, have poorly understood structure/function relationships • These properties of the API are hard to fully characterize resulting in uncertainty • Formulations: majority liquid presentations, less complexity then other formulations (stability a main issue, sampling size needs improvement ) Control of the API is a major source of concern for Biotech products.

Current OBP Practice

Paradigms • Quality is ensured by testing and rejecting lots that fail to meet its stated quality (insufficient) • A guiding principle for the Biotech industry has been that the process is the product (can be too restrictive) Quality by design concept: • Quality cannot be tested into a product; it has to be built by design. This design incorporates knowledge of the product and the process to ensure all critical quality parameters are adequately controlled

Quality Control Strategy Product Testing - Method Validation - Release Testing - Characterization - Stability Testing

How Much of the Iceberg (desired product) Can We See? • Release tests • Characterization ? • Process

Comprehensive Quality Control Strategy Process - Facilities and Equipment - Control of Raw Materials - In-Process Testing (PAT) - In-Process Controls - Process Validation (FED) - c. GMPs (QC/QA) Product - Method Validation - Release Testing - Characterization - Stability Testing

Designing a Quality Product • Design a high quality product that maximizes efficacy while minimizing adverse affects • Design a robust quality process to efficiently deliver a consistant product with the expected Q, S, and E profile

Q by D General Requirements for Biotech Products • Full Characterization of the product’s attributes (establish • product variability – the earlier the better) Understanding the relationship between the product’s quality attributes and safety and efficacy – Understanding the mechanism of action both in terms of efficacy and safety (Biological characterization) • Understand how process affects critical quality attributes This knowledge is limited for many Biotech products

The Desired Product • Dosage form is usually a given, liquid (some • • vialed as lyophilized power) Excipients vary from product to product but mostly affect product stability Desired attributes of the API (Focus for Biotech) - Opportunity for protein engineering - understanding protein structure/function relationship - Limit variability for attributes that negatively impact on product quality (via process or product)

Protein Engineering (rational design) • Increase manufacturability • Improving function/new properties - Increase specificity/affinity • Increasing Bioavailability - Pegylation - Glycoslation - Adding protein domains with increased half life - (Fc) Adding domains that bind to endogenous long lived proteins

Protein Engineering • Reduce tendency for aggregation • Increase conformational stability • Reducing immunogenicity - Eliminate sequences that promote aggregation Humanizing foreign proteins (m. Ab) Pegylation Incorporate structures that are less immunogenic (disulfide bond scaffolds) - T cell epitope engineering

Protein Engineering • OBP has encouraged development of innovative products (not a regulatory requirement) • Less enthusiastic concerning the use of products whose design increases uncertainly and has no expected value clinically (premise: limit product variability) - Histidine tag proteins (Quality versus Manufacturability) - Protein domains that potentially adversely impact safety

Designing A Quality Process

Examples of Problematic Process Designs • • Manufacturing capacity to clear viruses is limited Following elimination of aggregates by SEC, the manufacturer performs a heat treatment step for viral inactivation thus reintroduces aggregates back into the process • Process performed at room temperature with negative impact on quality • Roller bottle processes (open, multiple fermentations difficult to control) • Recloning is used to establish new cell banks introducing variability Manufacturer recognized the limitations but regulatory hurdles are difficult to overcome particularly after approval

Process Control • Current OBP expectations are that critical sources of variation should be identified and controlled (raw materials/ unit operations) • Controlled through in-process testing (PAT or other tests), monitoring operating parameters and process validation Based on long standing paradigm that process consistency = product consistency

Biotechnology Process Control Some steps controlled by volume or time few measure product attributes directly Turbidity Conductivity Harvest Fermentor D 02 p. H Chromatography Columns Proteolytic Steps 280 nm ABS Conductivity Renaturation Diafilt. /Conc. Formulation Lyophilization

The Essence of PAT • Process decisions (in real time) are based on assessments of critical material attributes - Forward-feed of incoming material - Feedback by in-process monitoring • Product quality is monitored and controlled during the manufacturing process • End points = achievement of the desired material attribute Currently, limited use of PAT in Biotech products but applicability is promising

Process Control of Unit Operations • Identify intended functions of unit operations and the • • • critical product attributes potentially affected Establish desired limits of attribute (typically established by estimates of process capability) Identify critical variables for the process step Establish the range of the variables that provides assurance that you can meet your quality expectations - First principles ? ? - Empirical approach using multi variant analysis FED, but can you extrapolate to larger scales?

Design Space (Fermentation) Critical process parameters Time Media composition Agitation

Expanding the Design Space • Characterize a quality attribute with regard to relevant, clinically important parameters, i. e. it’s affect on: - Potency - Bioavailability - Biodistribution - Immunogenicity • This information can be used to set specifications to ensure product quality as it relates to S and E and expand the design space

Examples from Biotech • • • For a highly glycoslyated protein various isoforms were isolated and monitored for relevant bioactivity in a animal model suitable for Pk measurements. Outcome: widen specs for isoform profile Monitored product isoforms from human serum samples over time, showed rates of decay were similar concluded isoforms did not impact bioavailability Outcome: broaden acceptance criteria Use of multiple lots of drug product in clinical trials to establish a link between variability of product attributes and clinical performance

Multiple binding/cellular assays Small Animal/Complex Bioassay Clinical/Clin Pharm Validated bioassay Developmental lots Stressed lots Purified/induced variants Many to all lots Clinical lot extremes One to some lots Clinical lots Biological Activity Matrix

Implementation

Regulatory Relief (based on process understanding) • • Validate the process is capable of impurity removal to appropriate levels (non toxic impurities) Relief: Impurity is not routinely measured when operating under the validated state (removed from specifications) Different approaches depending on the nature of the impurity - Validate capacity to remove those impurities that are added at fixed concentrations (fixed input) - Validate excess capacity for removal of impurities that variable (alternatively control of input levels of impurities) Examples: Host Cell Proteins/DNA

Regulatory Relief (based on product understanding) • Understanding of the relationship between the quality attribute and its impact on safety and efficacy can reduce regulatory requirements Relief: If no likely impact on S and E don’t include as a specification (no rejection limit) - use as a process consistency measure, where exceeding a limit initiates an investigation - if not a consistency measure, drop the test entirely • Transitioning to this new paradigm of action versus rejection limits • Need to discuss more extensively in-house and provide reviewer training

Implementation of Q by D • Q by D “a major fear by industry is that reviewers will not understand or be receptive to the submission” paraphrased from Dr. Ken Morris, Q by D presentation October 17, 2005 • OBP review is based on scientific merits of the proposal and not simply reliance on existing practice. Guidance helps frame the issue but science and knowledge dictates the outcome. • For example, we try to stay away from proscriptive rules i. e. “rejection limits can be established +/- 3 SD”. Instead, we evaluate the proposal using our best scientific judgment and are open to other statistical analysis but links between the attribute and what is known regarding its impact on S and E are important. Lack of knowledge increases uncertainty and may result in tighten controlled.

Implementation of Q by D Structure of OBP • Product reviewers a mixture of research/reviewers and full time reviewers • Research conducted in molecular and cellular biology and pharmaceutical science • Expertise in biological characterization of protein products is critical for meaningful risk assessment • Provides hands on experience with latest techniques familiarity with fermentation/purification processes • Expertise in biological characterization relevant to other CDER products - Consultations across CDER

PAT Future Directions Many steps controlled by measuring product attributes (or by monitoring all DP samples) Data Analysis In-line Fermentor Ion exchange chip SPR chip MS

Continued and Future Directions • Training of OBP product reviewers in PAT (4 OBP • • product reviewers will undergo extensive training for Biotech products), Q by D, and new analytical techniques (Biosensors SPR) for biotech products Q by D discussions within and outside Agency Encourage biological characterization of products • Encourage industry to incorporate new or under utilized analytical methods for control of in-process materials and purified proteins