Quality by Design for Biopharmaceuticals E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Serie: Wiley Series on Biotechnology
• Sprache: Englisch

Seitenzahl: 487

Veröffentlichungsjahr: 2011

Beschreibung

The concepts, applications, and practical issues of Quality by Design Quality by Design (QbD) is a new framework currently being implemented by the FDA, as well as EU and Japanese regulatory agencies, to ensure better understanding of the process so as to yield a consistent and high-quality pharmaceutical product. QbD breaks from past approaches in assuming that drug quality cannot be tested into products; rather, it must be built into every step of the product creation process. Quality by Design: Perspectives and Case Studies presents the first systematic approach to QbD in the biotech industry. A comprehensive resource, it combines an in-depth explanation of basic concepts with real-life case studies that illustrate the practical aspects of QbD implementation. In this single source, leading authorities from the biotechnology industry and the FDA discuss such topics as: * The understanding and development of the product's critical quality attributes (CQA) * Development of the design space for a manufacturing process * How to employ QbD to design a formulation process * Raw material analysis and control strategy for QbD * Process Analytical Technology (PAT) and how it relates to QbD * Relevant PAT tools and applications for the pharmaceutical industry * The uses of risk assessment and management in QbD * Filing QbD information in regulatory documents * The application of multivariate data analysis (MVDA) to QbD Filled with vivid case studies that illustrate QbD at work in companies today, Quality by Design is a core reference for scientists in the biopharmaceutical industry, regulatory agencies, and students.

Leseprobe

CONTENTS

FOREWORD

PREFACE

PREFACE TO THE WILEY SERIES ON BIOTECHNOLOGY AND RELATED TOPICS

CONTRIBUTORS

1: QUALITY BY DESIGN: AN OVERVIEW OF THE BASIC CONCEPTS

1.1 INTRODUCTION

1.2 CRITICAL QUALITY ATTRIBUTES

1.3 AN OVERVIEW OF DESIGN SPACE

1.4 RAW MATERIALS AND THEIR IMPACT ON QbD

1.5 PROCESS ANALYTICAL TECHNOLOGY

1.6 THE UTILITY OF DESIGN SPACE AND QbD

1.7 CONCLUSIONS

REFERENCES

2: CONSIDERATIONS FOR BIOTECHNOLOGY PRODUCT QUALITY BY DESIGN

2.1 INTRODUCTION

2.2 QUALITY BY DESIGN

2.3 RELEVANT PRODUCT ATTRIBUTES

2.4 MANUFACTURING PROCESS

2.5 DEVELOPING A DESIGN SPACE

2.6 UNCERTAINTY AND COMPLEXITY

2.7 FUTURE HORIZONS

2.8 QbD SUBMISSION THOUGHTS

2.9 IMPLEMENTATION PLANS

2.10 SUMMARY

ACKNOWLEDGMENTS

REFERENCES

3: MOLECULAR DESIGN OF RECOMBINANT MALARIA VACCINES EXPRESSED BY Pichia pastoris

3.1 INTRODUCTION

3.2 THE MALARIA GENOME AND PROTEOME

3.3 EXPRESSION OF TWO MALARIA ANTIGENS IN P. pastoris

3.4 SUMMARY

ACKNOWLEDGMENTS

REFERENCES

4: USING A RISK ASSESSMENT PROCESS TO DETERMINE CRITICALITY OF PRODUCT QUALITY ATTRIBUTES

4.1 INTRODUCTION

4.2 EXAMPLES OF CRITICALITY DETERMINATION

4.3 CONCLUSION

ACKNOWLEDGMENTS

REFERENCES

5: CASE STUDY ON DEFINITION OF PROCESS DESIGN SPACE FOR A MICROBIAL FERMENTATION STEP

5.1 INTRODUCTION

5.2 APPROACH TOWARD PROCESS CHARACTERIZATION

5.3 RISK ANALYSIS

5.4 SMALL-SCALE MODEL DEVELOPMENT AND QUALIFICATION

5.5 DESIGN OF EXPERIMENT STUDIES

5.6 WORST CASE STUDIES

5.7 DEFINITION OF DESIGN SPACE

5.8 DEFINITION OF VALIDATION ACCEPTANCE LIMITS

5.9 REGULATORY FILING, PROCESS MONITORING, AND POSTAPPROVAL CHANGES

ACKNOWLEDGMENT

REFERENCES

6: APPLICATION OF QbD PRINCIPLES TO TANGENTIAL FLOW FILTRATION OPERATIONS

6.1 INTRODUCTION

6.2 APPLICATIONS OF TFF IN BIOTECHNOLOGY

6.3 TANGENTIAL FLOW FILTRATION OPERATING PRINCIPLES

6.4 TFF DESIGN OBJECTIVES

6.5 MEMBRANE SELECTION

6.6 TFF OPERATING PARAMETER DESIGN

6.7 TFF DIAFILTRATION OPERATING MODE DESIGN

6.8 SUMMARY

REFERENCES

7: APPLICATIONS OF DESIGN SPACE FOR BIOPHARMACEUTICAL PURIFICATION PROCESSES

7.1 INTRODUCTION

7.2 ESTABLISHING DESIGN SPACE FOR PURIFICATION PROCESSES DURING PROCESS DEVELOPMENT

7.3 APPLICATIONS OF DESIGN SPACE

7.4 CELL HARVEST AND PRODUCT CAPTURE STEPS

7.5 PROTEIN A CAPTURE COLUMN

7.6 HYDROPHOBIC INTERACTION CHROMATOGRAPHY

7.7 ANION EXCHANGE CHROMATOGRAPHY

7.8 SUMMARY

ACKNOWLEDGMENTS

REFERENCES

8: VIRAL CLEARANCE: A STRATEGY FOR QUALITY BY DESIGN AND THE DESIGN SPACE

8.1 INTRODUCTION

8.2 CURRENT AND FUTURE APPROACHES TO VIRUS CLEARANCE CHARACTERIZATION

8.3 BENEFITS OF APPLYING DESIGN SPACE PRINCIPLES TO VIRUS CLEARANCE

8.4 TECHNICAL LIMITATIONS RELATED TO ADOPTION OF QbD/DESIGN SPACE CONCEPTS IN VIRUS CLEARANCE

8.5 DEVELOPING A VIRUS CLEARANCE DESIGN SPACE

8.6 STAYING IN THE DESIGN SPACE

8.7 CONCLUSION

ACKNOWLEDGMENTS

REFERENCES

9: APPLICATION OF QUALITY BY DESIGN AND RISK ASSESSMENT PRINCIPLES FOR THE DEVELOPMENT OF FORMULATION DESIGN SPACE

9.1 INTRODUCTION

9.2 QUALITY BY DESIGN (QbD) APPROACH

9.3 TARGET PRODUCT PROFILE (TPP)

9.4 MOLECULAR DEGRADATION CHARACTERIZATION

9.5 ACTIVE PHARMACEUTICAL INGREDIENT (API) CRITICAL PROPERTIES

9.6 PREFORMULATION CHARACTERIZATION

9.7 INITIAL FORMULATION RISK ASSESSMENTS

9.8 FORMULATION OPTIMIZATION AND DESIGN SPACE

9.9 SELECTION OF SOLUTION FORMULATION COMPOSITION

9.10 SUMMARY

ACKNOWLEDGMENTS

REFERENCES

10: APPLICATION OF QbD PRINCIPLES TO BIOLOGICS PRODUCT: FORMULATION AND PROCESS DEVELOPMENT

10.1 INTRODUCTION: QbD IN BIOLOGICS PRODUCT DEVELOPMENT

10.2 RISK ASSESSMENT PROCESS

10.3 EXAMPLES

10.4 CONCLUSIONS

REFERENCES

11: QbD FOR RAW MATERIALS

11.1 INTRODUCTION

11.2 BACKGROUND

11.3 CURRENT PRACTICE FOR RAW MATERIALS

11.4 QbD IN DEVELOPMENT

11.5 QbD IN MANUFACTURING

11.6 QbD FOR ORGANIZATIONS

11.7 TESTS AVAILABLE

11.8 CONCLUSIONS AND FUTURE PROSPECTS

ACKNOWLEDGMENTS

REFERENCES

12: PAT TOOLS FOR BIOLOGICS: CONSIDERATIONS AND CHALLENGES

12.1 INTRODUCTION

12.2 CELL CULTURE AND FERMENTATION PAT TOOLS

12.3 PURIFICATION PAT TOOLS

12.4 FORMULATION PAT TOOLS

12.5 PAT TOOLS FOR BIOPROCESS STARTING MATERIALS, DEFINED MEDIA, AND COMPLEX RAW MATERIALS

12.6 CHEMOMETRICS AND ADVANCED PROCESS CONTROL TOOLS

12.7 THE POWER OF PLS AND PCA

12.8 “RELEVANT TIME” COLUMN INTEGRITY MONITORING (MOMENTS ANALYSIS VERSUS HETP)

12.9 CHALLENGES FOR IMPLEMENTATION OF PAT TOOLS

12.10 FUTURE PAT TOOLS

ACKNOWLEDGMENTS

REFERENCES

13: EVOLUTION AND INTEGRATION OF QUALITY BY DESIGN AND PROCESS ANALYTICAL TECHNOLOGY

13.1 INTRODUCTION

13.2 EVOLUTION OF PAT AND QUALITY BY DESIGN (QbD): EMERGING GUIDELINES AND STANDARDS

13.3 PROCESS ANALYTICAL TECHNOLOGY (PAT)

13.4 QUALITY BY DESIGN

13.5 IMPLEMENTING QbD AND PAT

13.6 CONCLUSIONS

ACKNOWLEDGMENTS

REFERENCES

INDEX

Copyright © 2009 by John Wiley & Sons, Inc. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New JerseyPublished simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone], or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at [phone], outside the United States at [phone] or fax [phone].

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Rathore, Anurag S. (Anurag Singh); 1973-Quality by design for biopharmaceuticals : principles and case studies / by Anurag S.Rathore and Rohin Mhatre.p. cm.Includes index.ISBN 978-0-470-28233-5 (cloth)1. Pharmaceutical biotechnology–Quality control. I. Mhatre, Rohin. II. Title.

RS380.R38 2009615’.19–dc22

2008045484

To our family:Bhawana, Payal, Parul, and Jyoti

FOREWORD

These are truly exciting times to be involved in the development of biopharmaceutical products. As the research community expands our understanding of the biological basis of health and disease, those who turn this knowledge into medical treatments are providing safer and more effective health care options. Over the relatively short history of biologically derived drugs, this trend is clearly apparent. The first biological products to be developed were natural products such as antisera and hormones purified directly from animal tissues. The development of hybridoma technology in the 1980s allowed the preparation of monoclonal antibody products and significantly reduced the structural variability characteristic of polyclonal antibody products. The molecular purity of these products allowed them to be extremely well characterized and also led to a much better understanding of the biological activities of their structural features. Subsequently, the application of recombinant DNA technology to biopharmaceutical development has allowed manufactures to design proteins with specific structural and functional characteristics that give them desired beneficial therapeutic properties and reduce their potential adverse reactions.

These changes in product development and expression system technology have driven, and relied upon, parallel advances in the manufacturing sciences. Biopharmaceutical manufacturers have always been faced with the challenges of finding ways to make living systems produce proteins with desired characteristics, purifying them from complex mixtures with economically feasible yields, and formulating them in to stable, medically useful products. These challenges are compounded by the variabilities in raw material quality, equipment components, environment within the manufacturing facility, and capabilities of operators. As those who have struggled with these issues know so well, the quality of biological products depends to a large extent on the design and control of the manufacturing process.

It is crucial to public health that the drugs upon which we depend are safe, efficacious, and of consistent high quality. Safety and efficacy determinations are based on toxicological data, clinical study results, and postmarketing evaluation-based performance. Because the quality of a drug product can have a major impact on its clinical performance, successful drug development and manufacturing must focus on quality. In this regard, the concept of quality is twofold. One aspect of product quality is the design of the drug itself as defined by specification of the characteristics it needs to have to treat a disease. This includes the structure of the pharmaceutically active molecule itself, as well as the formulation and delivery system that allow the therapeutic to reach its target. The other aspect of quality is the consistency with which the units of a batch or a lot of product meet the desired specifications. As was alluded to earlier, within-batch variability and batch-to-batch variability depend, to a large extent, on the quality of the raw materials and the design of the manufacturing process and its control systems. Incorporation of these two aspects of quality into product and process development is the essence of quality by design.

To realize the full benefits of quality by design, one must develop a thorough understanding of the interrelationship between the attributes of the input materials, the process parameters, and the characteristics of the attribute of the input materials, the process parameters, and the characteristics of the resultant products. With this information in hand, it is possible to manufacture with a very high degree of assurance that each unit of product will have the desired quality. Of particular note in this regard is the quality control system known as process analytical technology (PAT) that has been applied with great success to manufacturing operations outside of the pharmaceutical industry. A cornerstone of PAT is the use of rapid analytical techniques and process control systems to monitor and control product quality during manufacturing. In 2004, the FDA published guidance for industry on PAT1 to encourage the development and implementation of the agency’s “Product Quality for the 21st Century Initiative,” as PAT can provide the assurance of quality in a flexible manufacturing environment conducive to streamlined implementation of innovative technologies. The use of correlated metrics of quality, such as bioreactor conditions, within the process control system is quite familiar to biopharmaceutical manufacturers. However, future strides in rapid, real-time analytical technologies promise to make direct control of product quality during manufacturing a reality and open the door to efficiencies such as continuous processing and real-time release.

As biotechnology moves ahead, the concepts of William Edwards Deming and others that quality must be built into products will continue to be applied to the design of novel products and dose delivery systems as well as to the design and engineering of more effective and reliable manufacturing methods. Technological advances in this field will undoubtedly occur in an evolutionary manner, with successful systems serving as the foundation of even more valuable systems. However, this steady progression will, nearly as surely, be punctuated by revolutionary discoveries of magnitudes equal to hybridoma technologies that introduced monoclonal antibody production or the polymerase chain reaction that has made genetic engineering a relatively facile process. To ensure that we have the safest and most efficacious medications to treat today’s disease, and those of tomorrow, we must not only continue developing innovative products and technologies but also take them to the manufacturing plant and the marketplace as quickly as possible. The sharing of ideas, information, and experience through books such as this is essential to the success of this endeavor.

Keith O. WebberDeputy Director, Office of Pharmaceutical Safety,Food and Drug Administration

1FDA Guidance for Industry: PAT—A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance (September 2004).

PREFACE

Quality by Design (QbD) is receiving a lot of attention in both the traditional pharmaceutical and biopharmaceutical industries subsequent to the FDA published “Guidance for Industry: Q8 Pharmaceutical Development” in May 2006. Key challenges in successfully implementing QbD are requirements of a thorough understanding of the product and the process. This knowledge base must include understanding the variability in raw materials, the relationship between the process and the critical quality attributes (CQAs) of the product, and finally relationship between the CQA and the clinical properties of the product. This book presents chapters from leading authorities on a variety of topics that are pertinent to understanding and successfully implementing QbD.

Chapter 2 by Kozlowski and Swann provides a summary of QbD and related regulatory initiatives. Approaches to relevant product quality attributes and biotechnology manufacturing have been discussed along with some thoughts on future directions for biotechnology products.

Chapter 3 by Narum presents a case study where QbD principles have been applied to make significant improvements in the capacity of recombinant expression systems to produce malarial proteins by introducing synthetic genes for Pichia pastoris as well as Escherichia coli. It is shown that the use of synthetic genes not only makes possible the expression of a particular protein but also allows the gene designer to make appropriate modifications to increase product quantity and quality.

In Chapter 4, Schenerman et al. present a risk assessment approach for the determination of the likelihood and extent of an impact of a CQA on either safety or efficacy. Examples are used to illustrate how nonclinical data and clinical experience can be used to define the appropriate risk category for each product quality attribute. The attribute classifications then serve as a rationale for product testing proposals, associated specifications, and process controls that ensure minimal risk to product quality.

Chapter 5 contributed by Hoek et al. presents a case study involving a cell culture step. All operational parameters were examined using a risk analysis tool, failure mode and effects analysis (FMEA). The prioritized parameters were examined through studies planned using design of experiments (DOEs) approach. Qualified scale-down models were used for these studies. The results were analyzed to create a multivariate model that can predict variability in performance parameters within the “design space” examined in the studies. The final outcome of the effort was identification of critical and key operational parameters that impact the product quality attributes and/or process consistency, respectively, along with their acceptable ranges that together define the design space. Chapters 6 and 7 define approaches to establishing design space for a filtration and chromatography unit operation, respectively.

Sofer and Carter present a strategy in Chapter 8 for applying QbD principles for virus clearance. It is concluded that implementation of the proposed strategy will require an extended and coordinated effort, primarily by manufacturers and regulators. The mutual investment in moving to a QbD approach holds promise of better understood, and therefore better controlled, unit operations. The QbD design space describes a full range of manufacturing conditions within which changes may be made with relative ease and modest regulatory oversight, freeing both manufacturers and regulators’ limited resources. Intrinsically, enhanced process control and process understanding represents a benefit to the patient population.

Chapter 9 by Ng and Rajagopalan presents the different considerations to remember while designing a formulation process. Some of the key steps include identification of target commercial drug product profile; preformulation and forced degradation studies to characterize molecular stability properties, impact of formulation variables, and other factors; preliminary stability risk assessment with emphasis on direct impact on the activity based on preformulation and forced degradation studies results; initial formulation risk assessment to establish the cause–effect relationship of different factors and solution formulation stability via Ishikawa (Fishbone) diagram; multivariate DOE studies to optimize the formulation composition and define a robust design space to meet the expected shelf life of 24 months at 5°C; establishing formulation design space based on DOE results and stability properties projections; and finally selection of commercial solution formulation based on design space, molecule knowledge, and risk assessment.

In Chapter 10, Singh et al. present case studies illustrating a systematic work process for application of risk-based approaches to formulation development for biologics.

Lannan addresses the application of multivariate data analysis (MVDA) to analysis of raw materials in Chapter 11.

Chapter 12 by Molony and Undey provides a review of various PAT tools and applications for the biopharmaceutical industry. Finally, Chapter 13 by Low and Phillips provides the background for PAT and also how it relates to QbD.

Anurag S. RathoreRohin Mhatre

Thousand Oaks, CaliforniaCambridge, MassachusettsMarch 2009

PREFACE TO THE WILEY SERIES ON BIOTECHNOLOGY AND RELATED TOPICS

Significant advancements in the fields of biology, chemistry, and related disciplines have led to a barrage of major accomplishments in the field of biotechnology. The Wiley Series on Biotechnology and Bioengineering focuses on showcasing these advances in the form of timely, cutting-edge textbooks and reference books that provide a thorough treatment of each respective topic.

Topics of interest to this series include, but are not limited to, protein expression and processing; nanotechnology; molecular engineering and computational biology; environmental sciences; food biotechnology, genomics, proteomics, and metabolomics; large-scale manufacturing and commercialization of human therapeutics; biomaterials and biosensors; and regenerative medicine. We expect these publications to be of significant interest to the practitioners both in academia and industry. Authors and editors are carefully selected for their recognized expertise and their contributions to the various and far-reaching fields of biotechnology.

The upcoming volumes will attest to the importance and quality of books in this series. I thank the fellow coeditors and authors of these books for agreeing to participate in this endeavor. Finally, I thank Ms Anita Lekhwani, Senior Acquisitions Editor at John Wiley & Sons, Inc., for approaching me to develop such a series. Together, we are confident that these books will be useful additions to the literature that will not only serve the biotechnology community with sound scientific knowledge but will also inspire them as they further chart the course of this exciting field.

Anurag S. RathoreAmgen, Inc.

Thousand Oaks, CaliforniaJanuary 2009

CONTRIBUTORS

Milton J. Axley, MedImmune, Gaithersburg, Maryland

Amit Banerjee, Pfizer Corporation, Chesterfield, Missouri

Jeffrey Carter, GE Healthcare, Westborough, Massachusetts

Douglas J. Cecchini, Biogen Idec, Cambridge, Massachusetts

Jean Harms, Amgen Inc., Thousand Oaks, California

Carol F. Kirchhoff, Pfizer Corporation, Chesterfield, Missouri

Steven Kozlowski, Food and Drug Administration, Silver Spring, Maryland

Maureen Lanan, Biogen Idec, Cambridge, Massachusetts

Duncan Low, Amgen Inc., Thousand Oaks, California

Rohin Mhatre, Biogen Idec, Cambridge, Massachusetts

Michael Molony, Allergan Corporation, Irvine, California

David L. Narum, National Institutes of Health, Rockville, Maryland

Kingman Ng, Eli Lilly and Company, Indianapolis, Indiana

Cynthia N. Oliver, MedImmune, Gaithersburg, Maryland

Joseph Phillips, Amgen Inc., Thousand Oaks, California

Natarajan Rajagopalan, Eli Lilly and Company, Indianapolis, Indiana

Kripa Ram, MedImmune, Gaithersburg, Maryland

Anurag S. Rathore, Amgen Inc., Thousand Oaks, California

John Rozembersky, Rozembersky Group, Inc, Boxborough, Massachusetts

Mark A. Schenerman, MedImmune, Gaithersburg, Maryland

Satish K. Singh, Pfizer Inc., Chesterfield, Missouri

Gail Sofer, Consultant, SofeWare Associates, Austin, Texas

Patrick G. Swann, Food and Drug Administration, Silver Spring, Maryland

Cenk Undey, Amgen Inc., West Greenwich, Rhode Island

Pim van Hoek, Amgen Inc., Thousand Oaks, California

Xiangyang Wang, Amgen Inc., Thousand Oaks, California

Gail F. Wasserman, MedImmune, Gaithersburg, Maryland

Peter K. Watler, JM Hyde Consulting, Inc., San Francisco, California

Keith Webber, Food and Drug Administration, Silver Spring, Maryland

1

QUALITY BY DESIGN: AN OVERVIEW OF THE BASIC CONCEPTS

Rohin Mhatre and Anurag S. Rathore

1.1 INTRODUCTION

The premise of Quality by Design (QbD) is that the quality of the pharmaceutical product should be based upon the understanding of the biology or the mechanism of action (MOA) and the safety of the molecule [1]. The manufacturing process should then be developed to meet the desired quality attributes of the molecule, hence the concept of “design” of the product quality versus “testing” the product quality. Although testing the product quality after manufacturing is an essential element of quality control, testing should be conducted to confirm the predesired product attributes and not to simply reveal the outcome of a manufacturing process. The ICH Q8 guideline provides an overview of some of the aspects of QbD [2]. The guideline clearly states that quality cannot be tested into products; that is, quality should be built in by design.

Although the task of designing a complex biological molecule such as a monoclonal antibody may seem daunting, the experience gained in the past roughly 30 years of the biotechnology industry history has laid the foundation for the QbD initiative [3, 4]. The industry has come a long way in identifying and selecting viable drug candidates, in developing high-productivity cell culture processes, in designing purification processes that yield a high-purity product, and in analyzing the heterogeneity of complex biomolecules. As all these activities are the building blocks of QbD, the concept of QbD has in fact been practiced for the last few years and has in turn led to the development of highly efficacious biopharmaceuticals and robust manufacturing processes. The issuance of the ICH Q8 guideline was an attempt to formalize the QbD initiative and to allow manufacturing flexibility based on the manufacturer’s intricate knowledge of the molecule and the manufacturing process. The concept of obtaining intricate knowledge of the molecule along with the manufacturing process and the resulting flexibility in manufacturing, the eventual goal of the QbD initiative, requires an understanding of the various elements of QbD.

The two key components of QbD are [4]

1. The understanding of the critical quality attributes (CQAs) of a molecule. These are the attributes of the molecule that could potentially affect its safety and efficacy profile.

2. The design space of the process defined as the range of process inputs that help ensure the output of desired product quality.

An overview of these components is discussed further in this chapter and elsewhere in this book.

1.2 CRITICAL QUALITY ATTRIBUTES

The starting point of QbD is developing a good understanding of the molecule itself. Biomolecules are quite heterogeneous due to the various post-translational modifications that can occur and have been commonly observed. These modifications arise from the glycosylation, oxidation, deamidation, cleavage of labile sites, aggregation, and phosphorylation, to name a few. As many of these modifications could impact the safety and efficacy of the molecule, defining the appropriate CQAs of the molecule is an important starting point in the development cycle of a biopharmaceutical. Although the understanding of the CQAs evolves during the life cycle of the product, understanding the CQAs at an early stage of the development of the molecule is clearly desirable. Studies conducted during the early research stages of development of a potential biopharmaceutical may entail evaluating various forms of a particular biomolecule in animal studies. The outcomes of such studies help “design” a biomolecule with the desired quality attributes so as to be safe and highly efficacious.

Since the CQAs can impact the safety and clinical efficacy of a molecule, data gathered in animal studies, toxicological studies, and early human clinical trails become the starting point for defining the CQAs. On the basis of the safety and efficacy readout of a clinical trial, one can start to define the product profile of a molecule. The assumption is that if the CQAs of the molecule are similar to those used in preclinical and clinical trials, the safety and efficacy will be comparable as well. Furthermore, historical data from clinical trials of similar molecules can also provide valuable insight into the CQAs. Evaluation of the in vitro biological activity via bioassays, reflecting the mechanism of action, can provide a good assessment of how the various product attributes could potentially impact the in vivo activity of a molecule. The molecule can be altered by conducting stress studies to induce higher level of aggregation; oxidation, deamidation, and the glycosylation pattern can be varied as well. The impact of changes in the molecular structure on the biological activity can then be evaluated via various bioassays. This study is referred to as structure–activity relationship. The evaluation of in vitro activity is often the relatively easiest means of determining the CQAs. However, in vitro assessments can only provide an understanding of the potential changes in the activity of the molecule, and the correlation between this change in activity and the impact of efficacy in patients is often unclear. Further assessment of the molecule in animal studies to evaluate clearance, efficacy, and safety is often a good indicator of the behavior of the molecule in human trials and is a better tool for understanding the CQAs. Additional details of the determination of the CQA can be found in Chapter 4.

1.3 AN OVERVIEW OF DESIGN SPACE

After defining the CQAs, the next and more critical step is the development of a manufacturing process that will yield a product with the desired CQAs [4, 5]. During the process development, several process parameters are routinely evaluated to assess how they could impact product quality [6]. The design space for the process eventually evolves from such a study. For example, during the cell culture development, ranges for process inputs such as temperature, pH, and the feed timing can be evaluated to determine if operating within a certain range of temperature and pH has an impact on product quality. The design of experiments (DOE) is conducted in a manner so as to evaluate the impact of the multiple variables (multivariate) and also to understand if and how changes in one or more of the process inputs have an effect on the product quality and/or if a process input is independent of changes in other inputs.

The design space (process range) is then established for each of the above process inputs. This can be further explained using an example of a design space for a purification column. If a column used to purify a protein is expected to reduce the level of protein aggregate to 2%, the various column operating parameters such as flow rate, pH and strength of the buffer, load volume, and so on are evaluated such that operating within a certain range of these parameters yields an aggregate level of less than or equal to 2%. If it turns out that pH above 6 or below 4 results in aggregate levels above 2%, then the design space for the pH of the buffer is defined as between 4 and 6. One can similarly envision a design space for the flow rate and other inputs for the particular purification step. Eventually, the entire production process for a molecule will have a defined design space, and operating within that design space should lead to a product of acceptable quality. Operating beyond the design space of a particular process input may result in an unacceptable product quality.

Since the production process for a biomolecule entails multiple steps starting from the cell culture process to the final purification and eventually to the formulation and fill in the desired container, the development of a design space for a particular step is not usually independent of other steps in the production process. Since the output of one step becomes the input for the next step, the development of the design space for a process should be evaluated in a holistic manner. One such approach may be to determine the desired product quality from the final process step and to work backward in the process to ensure that each step of the process delivers the required product quality needed for the next step to meet the quality target of the final step. To provide an example of this approach, we can revisit the above example of a desired level of 2% aggregate in the final drug product. In this particular case, design space should be developed for all parameters of the production process that can potentially impact the level of aggregate in the final drug product. The maximum level of aggregate resulting from each of the steps in the process does not need to be less than 2%, particularly steps that are upstream in the process such as the protein A purification, the first columns used in the purification of an antibody. The development of the design space including the design of experiments is discussed in detail in Chapters 5–7.

1.4 RAW MATERIALS AND THEIR IMPACT ON QbD

In addition to the design space and CQAs, other factors also play an important role in implementing QbD, and raw material is one such factor. Cell culture processes used to make recombinant proteins use complex growth media such as hydrosylates and also feeds such as vitamins for the cell. The understanding of how the various components of these complex raw materials affect the productivity of the cells and the quality of the product is not a trivial task. It requires a thorough analysis and quantitation of the various components of the raw material. Raw material analysis and correlation between raw material components and the productivity of cells and product quality is an area that has not been sufficiently explored by the biotechnology industry. However, the evolution of instruments such as high-resolution nuclear magnetic resonance spectroscopy, near infrared spectroscopy, and mass spectrometry has provided an opportunity to analyze complex mixtures of raw materials. In addition, the availability of sophisticated statistical tools for deconvolution and pattern matching of complex data sets has further refined the approach to analyzing raw materials. Once the correlation between the critical components of the growth or feed media and the performance of cells is understood, the ultimate goal of raw material analysis in the context of QbD would be to fortify the media as needed with the relevant component so as to ensure the desired productivity and product quality. Further details of analysis of complex raw materials are provided in Chapter 11.

1.5 PROCESS ANALYTICAL TECHNOLOGY

Since one of the goals of QbD is to maintain control of the process to achieve the desired product attributes, process analytical technology (PAT) is an important tool for QbD. PAT entails analysis of product quality attributes during the various stages of the manufacturing process of a biomolecule. The analysis is often conducted online using either probes inserted into the bioreactor to monitor critical components such as the cell density or sterile sampling devices to divert the stream from a purification column to assess the product purity [7]. In either case, the online analysis enables operators on the manufacturing floor to make real-time adjustments to the process parameters so as to obtain the desired product profile at every stage of the manufacturing process. For example, a PAT tool to monitor a purification column would entail periodically sampling the elution stream from the column via a sampling device and diverting the sample to an online HPLC system [8]. The results of the online HPLC analysis, indicative of the product purity, would be used to determine the eluate volume that should be collected. In this particular example, the fraction of the eluate of purity below a predetermined criterion would provide a trigger to stop collection of the eluate and to divert the elution stream to waste. The advantage of such a PAT tool would be that the collection of the column eluate would be based on the required product purity and would help to ensure a consistent product quality for every production batch of the biomolecule [8]. Further applications of PAT can be found in Chapters 12 and 13.

1.6 THE UTILITY OF DESIGN SPACE AND QbD

Prior to development of the design space, the questions to ask are the following: How would the design space be used? What is the advantage for a company of developing a design space for any of its products? What would be the driver for regulatory agencies to promote the concept of design space and QbD?

As seen in Fig. 1.1, limits that establish the acceptable variability in product quality and process performance attributes would also serve as the process validation acceptance criteria [4, 5]. After the design space has been established, the regulatory filing would include the acceptable ranges for all key and critical operating parameters (i.e., design space) in addition to a more restricted operating space typically described for pharmaceutical products. After approval, CQAs would be monitored to ensure that the process is performing within the defined acceptable variability that served as the basis for the filed design space. The primary benefit of an expanded design space would be a more flexible approach by regulatory agencies. Process changes are often driven by changes in the manufacturing equipment and raw materials, to name a few. At present, changes in the process require formal filings and approvals from regulatory agencies and often require a significant commitment of both time and resources for the industry and the regulatory agencies. The outcome of the design space development (as stated in the ICH Q8 guideline) would be that upon the approval of the design space for a particular product by a regulatory agency, process changes within the design space would not require additional regulatory filing and approval. This shift in the paradigm of using enhanced process knowledge to enable process changes with a limited burden of regulatory approval is clearly beneficial to both the manufacturer and the regulatory agencies. Chapter 2 further reviews the regulatory relief and implications of the QbD initiative.

Process improvements during the product life cycle with regard to process consistency and throughput could take place with reduced postapproval submissions. As manufacturing experience grows and opportunities for process improvement are identified, the operating space could be revised within the design space without the need for postapproval submission. This is illustrated in Fig. 1.2, which shows that if the process creeps outside the design space, process changes may be required to be made and may require process characterization, validation, and filing of the changes to the approved process design space.

Figure 1.1 Illustration of the creation of process design space from process characterization studies and its relationship with the characterized and operating spaces. The operating range denotes the range in the manufacturing procedures and the characterization range is the range examined during process characterization. The acceptable range is the output of the characterization studies and defines the process design space. Adapted from Ref. [5], by permission of Advanstar Communications.

Figure 1.2 Application of the design space concept in process characterization, validation, monitoring, and regulatory filing. Adapted from Ref. [5], by permission of Advanstar Communications.

Figure 1.3 The various elements of QbD. The boxes in the bottom row show all the relevant information that is used to develop the critical quality attributes. The CQA and DOE data are then used to develop the design space. The design space and PAT tools help establish QbD.

1.7 CONCLUSIONS

Figure 1.3 depicts the various components of QbD discussed above and the correlation between the various components. As shown in the figure, the outcome of the QbD exercise is the establishment of the design space for the process and the operating ranges (ORs) that help achieve the desired product quality. As mentioned earlier, the reader is referred to the various sections of the book to gain further understanding of the various aspects of QbD. The editors hope that this book will help establish a good framework for any researcher to build Quality by Design into a manufacturing process for a biomolecule.

REFERENCES

[1] PAT Guidance for Industry: A Framework for Innovative Pharmaceutical Development, Manufacturing and Quality Assurance. U.S. Department of Health and Human Services, Food and Drug Administration (FDA), Center for Drug Evaluation and Research (CDER), Center for Veterinary Medicine (CVM), Office of Regulatory Affairs (ORA), September 2004.

[2] Guidance for Industry: Q8 Pharmaceutical Development. U.S. Department of Health and Human Service, Food and Drug Administration (FDA), May 2006.

[3] Kozlowski S, Swann P. Current and future issues in the manufacturing and development of monoclonal antibodies. Adv Drug Deliv Rev 2006;58:707–722.

[4] Rathore AS, Winkle H. Quality by Design for Pharmaceuticals: Regulatary Perspective and Approach. Nature Biotechnology 2009;27:26–34.

[5] Rathore AS, Branning R, Cecchini D. Design space for biotech products. BioPharm Int 2007;36–40.

[6] Harms J, Wang X, Kim T, Yang J, Rathore AS. Defining design space for biotech products: case study of Pichia pastoris fermentation. Biotechnol Prog 2008;24(3):655–662.

[7] Munson J, Stanfiled CF, Gujral B. A review of process analytical technology (PAT) in the U.S. pharmaceutical industry. Curr Pharm Anal 2006;2:405–414.

[8] Rathore AS, Yu M, Yeboah S, Sharma A. Case study and application of process analytical technology (PAT) towards bioprocessing: use of on-line high performance liquid chromatography (HPLC) for making real time pooling decisions for process chromatography. Biotechnol Bioeng 2008;100:306–316.

2

CONSIDERATIONS FOR BIOTECHNOLOGY PRODUCT QUALITY BY DESIGN

Steven Kozlowski and Patrick Swann

2.1 INTRODUCTION

In August 2002, the Food and Drug Administration (U.S. FDA) announced a significant new initiative, pharmaceutical Current Good Manufacturing Practices (CGMPs) for the twenty-first century [1]. This initiative is intended to enhance and modernize pharmaceutical manufacturing and product quality. Specific areas of focus include facilitating industry adoption of risk-based approaches, technological advances, and modern quality management techniques. As part of this initiative, the FDA will use state-of-the-art pharmaceutical science in developing review, compliance, and inspection policies and will coordinate these activities under a quality systems approach.

Concurrently with the CGMPs for the twenty-first century initiative, process analytical technology (PAT), a system to improve pharmaceutical manufacturing was being discussed at the advisory committee of the Office of Pharmaceutical Science at CDER and at the FDA Science Board [2].

Process Analytical Technology is a system for designing, analyzing, and controlling manufacturing through timely measurements (i.e., during processing) of critical quality and performance attributes of raw and in-process materials and processes with the goal of ensuring final product quality [3]. In 2004, the FDA published guidance on PAT [4] that described multivariate tools for design, data acquisition and analysis, process analyzers and controllers, continuous improvement, and knowledge management tools.

Systematic approaches to pharmaceutical manufacturing may be of benefit even if they do not use each of these specific tools. Although PAT may allow for greater flexibility, manufacturing may still be improved in the absence of real-time analysis of material attributes and without real-time linkage to process control. The term Quality by Design (QbD) [5, 6] is used to describe a more general approach to systematic pharmaceutical manufacturing. As described by Dr. Janet Woodcock, the desired state that drives all these manufacturing initiatives is a maximally efficient, agile, flexible pharmaceutical manufacturing sector that reliably produces high-quality drug products without extensive regulatory oversight [7].

Over the last few years, there has been significant progress in moving forward with these initiatives for small molecules, including a pilot program for QbD submissions [8]. However, biotechnology products are a growing part of the drug development pipeline [9]. It is important to consider how to approach the “desired state” for more complex products, such as biotechnology products. The principles of Quality by Design should be applicable to all pharmaceuticals including biotechnology products [10].

2.2 QUALITY BY DESIGN

Quality by Design is defined as a systematic approach to development that begins with predefined objectives and emphasizes product and process understanding and process control based on sound science and quality risk management [11]. Dr. Moheb Nasr has summarized QbD in a diagram [12] (Fig. 2.1). A systematic approach to pharmaceutical development should start with the desired clinical performance and then move to product design. The desired product attributes should then drive the process design, and the process design should drive the strategies to ensure process performance. This systematic approach may be iterative and thus the circular design as shown in Fig. 2.1. The inner circle interacts with many other specific measures of pharmaceutical manufacturing, such as specifications, critical process parameters, and so on. This QbD circle can be divided into two major semicircles, product knowledge and process understanding. A critical tool for enabling QbD manufacturing is a defined way of linking these two semicircles.

The International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) has bridged this gap using the concept of a design space. A design space is the multidimensional combination and interaction of input variables (e.g., material attributes) and process parameters that have been demonstrated to provide assurance of quality [3]. This is the scientific definition of a design space. Design space also has a regulatory definition. Movement within a design space is not considered as a change that requires regulatory approval. However, change within a design space does need oversight by the sponsor’s quality system. Design space is proposed by the applicant and is subject to regulatory assessment and approval. A design space could potentially link process performance to variables such as scale and equipment. The design space is thus a very flexible tool that links process characteristics and in-process material attributes to product quality. A recent definition of product quality was given by Dr. Janet Woodcock [13], “Good pharmaceutical quality represents an acceptably low risk of failing to achieve the desired clinical attributes.” As indicated at the top of Fig. 2.1, QbD starts with the desired clinical performance.

Figure 2.1 Quality by Design is a systematic approach to development that begins with predefined objectives and emphasizes product and process understanding and process control based on sound science and quality risk management (ICH Q8R1). Quality by Design begins by defining the desired product performance and also by designing a product that meets those performance requirements. The characteristics of the designed product are the basis for designing the manufacturing process, and the performance of the manufacturing process also needs to be monitored. Each of these steps may impact each other. For example, process performance may provide knowledge regarding manufacturability that could impact product design in an iterative manner. These steps also relate to specific quality measurements and tools. Product specifications should ideally be based on desired clinical performance. The product design should be defined in terms of product quality attributes. The criticality of these attributes and the relationship of these attributes to specifications may evolve over the product life cycle. Process parameters are important in defining the process, and process controls are important in ensuring process performance. This circle of QbD can be split into two general areas, product knowledge and process understanding. These two areas meet in the design space and the interaction of product knowledge and process understanding allows for continuous improvement. The QbD circle was developed by Dr. Moheb Nasr. (See the insert for color representation of this figure.)

2.3 RELEVANT PRODUCT ATTRIBUTES

In the draft annex to ICH Q8, ICH Q8(R1) [11], the target product profile is described as a starting point for Quality by Design. The target product profile is based on the desired clinical performance. A number of considerations for the target product profile are described including dosage form, strength, release characteristics, and drug product quality characteristics (e.g., sterility, purity). All of these are also considerations for biotechnology products; however, there may be a difference in focus for these complex products.

Currently, many protein products are delivered parenterally, although orally administered enzymes [14, 15] and novel dosage forms such as inhaled insulin [16] have been developed. Even for parenteral dosage forms, there are important considerations regarding formulation (e.g., liquid versus lyophilized) and route (e.g., intramuscular, subcutaneous, or intravenous). The choice of delivery system, such as prefilled syringes, is also important. For complex as well as for simple drugs, desired pharmacology, targeted patient population(s), and disease state(s) should be carefully considered in decisions on drug product dosage form, strength, and route of administration.

For biotechnology products, the complex processes, raw materials, and biological substrates used in manufacturing can lead to a broad range of process-related impurities. These impurities may impact product performance beyond direct toxicities. Impurities may impact the attributes of the active pharmaceutical ingredient, such as protein aggregation by tungsten moieties [17]. Contaminating proteases may also impact stability of a protein product. Some impurities can act as adjuvants and thus may have the potential to alter protein immunogenicity. Such impurities have been suggested as playing a role in erythropoietin immunogenicity [18, 19] although other possible causes have also been suggested [20, 21].

Product quality characteristics for complex products encompass a wide variety of product variants that include product-related substances and product-related impurities [22]. The three-dimensional structure of proteins is important for receptor interactions. Changes in folding may alter receptor binding and/or signaling. Protein multimerization may change receptor blockade to receptor activation. Abnormally folded proteins may impact immunogenicity through aggregation, generation of novel epitopes, and/or altered uptake by antigen presenting cells. Protein folding depends on low-energy interactions such as hydrogen bonds. Thus, minor environmental changes could impact protein structure and generate structural variants. Environmental excursions during processing or shelf life may interact with other impurities such as trace proteases and impact degradation.

In addition to variants in higher order structure, proteins can have many post-translational modifications. Biotechnology products are often heterogeneous mixtures with many variants that have different sets of modifications. Although many post-translational modifications may not impact product performance, others may alter pharmacokinetics, activity, or safety (e.g., immunogenicity). In Fig. 2.2a, a schematic of a monoclonal antibody is shown with a subset of potential post-translational modifications, such as N-terminal pyroglutamines, oxidations, deamidations, glycations, C-terminal lysines, and glycosylations. One of these modifications, an oxidation site, is considered in a decision tree regarding a potential impact on performance (Fig. 2.2b). Such a decision tree can be assigned for each of the modifications. Decision trees can also be constructed regarding specific safety concerns, such as immunogenicity. Ideally, a probability or risk ranking can be applied to the various possibilities in each of these decision trees. For example, immunogenicity is difficult to predict on the basis of attributes, but the impact of immunogenicity can be evaluated in terms of clinical risk [23–25]. Risk assessments, whether for activity or safety, are challenging and require the use of many sources of information such as prior knowledge, related product or platform data, in vitro and in vivo biological characterization of product variants, and clinical data. In many cases, no one source of information would allow a meaningful risk assessment. However, the integration of multiple sources of information may facilitate a useful assessment of risk. A matrix approach can be informative [26], evaluating product lots generated at varying points throughout the development for biological impact across a variety of studies. This information can give confidence to proposed mechanisms of action and structure–function relationships. These data may be integrated around mechanism of action models and potentially use Bayesian statistical approaches.

Figure 2.2 Biotechnology products have a large number of structural attributes that may impact product performance. (a) Some structural attributes of monoclonal antibodies are indicated, such as pyroglutamine, oxidation, glycation, deamidation, glycosylation, and clipping of C-terminal lysines. An oxidation at one site is circled in blue. (b) A decision tree for the impact of that oxidation on product activity is shown. (See the insert for color representation of this figure.)

For products that share significant sequence homology to related products, a platform approach to attribute impact may be very useful. Monoclonal antibodies may present opportunities [26] for this, but the nature of the targets, patient population, disease state(s), and the role of effector functions need to be considered in any extrapolations. Studies demonstrating that a product attribute is rapidly modified in vivo (e.g., deamidation or oxidation of a specific residue) may allow for increasing levels of that attribute over product shelf life. Comparisons to related endogenous molecules may also be informative.

Clinical studies on product variants would provide the strongest linkage between product attributes and clinical performance. In this book [27], an example of using structural characterization of variants in timed samples for variant pharmacokinetics is given. Such studies do not require purification of variants and additional clinical studies. These studies may allow the evaluation of variant effects on pharmacokinetics and, in some cases, variant effects on pharmacodynamics. In addition, during standard clinical development many different lots may be used at different times after manufacture. The use of multivariate statistical analysis for evaluating product attributes and clinical findings may be informative [28].

There are many potential strategies for understanding the input of complex product attributes on clinical performance. Attributes may interact and a multivariate attribute space may be more useful than univariate ranges for such attributes. However, complex products may have many thousands of possible attribute states. Even using all these strategies, this large number of attributes and interactions cannot be fully evaluated. Thus, a risk assessment, utilizing prior knowledge in conjunction with other approaches, will be needed. An important consideration for complex products is that the more the attributes dismissed based on assumptions, the greater the uncertainty that no critical attribute or interaction was missed. This greater uncertainty needs to be considered in defining attributes as unimportant. For complex products, mechanism of action and biological characterization may contribute to understanding the importance of product attributes.

We have discussed assigning risk and importance to many attributes of complex products. For regulatory purposes, clear ways to distinguish attributes will be needed. Draft guidance [11] has defined a critical quality attribute (CQA) as a physical, chemical, biological, or microbiological property or characteristic that should be within an appropriate limit, range, or distribution to ensure the desired product quality. This CQA definition is process independent and does not consider the control system; a CQA is only defined as a property or characteristic of the product. Other suggestions have included an intermediate level of attribute importance, analogous to intermediate levels of parameter importance or key parameters [29]. For complex products with greater uncertainties regarding attribute importance, intermediate attribute categories, between noncritical and critical, may be of value. Linkage of potential attribute risk to the appropriate control strategy may be best achieved with a variety of attribute risk categories. Attribute importance can be assigned through a risk assessment, as described above. CQAs do not necessarily need to be controlled by end-product testing (classical specifications), but the sensitivity of biotechnology product attributes to environmental conditions should be considered in any upstream approaches to CQA control.

Assessment of product attributes using the above methods may require significant efforts. However, the knowledge gained from linking product attributes to clinical performance can be leveraged into other products and may facilitate discovery and design of new products. This information does not need to be complete at the time of marketing application submission. As product knowledge grows, CQAs may evolve. Not all sponsors may invest in an extensive risk assessment of all their products before approval. Historically, sponsors have used the ranges of attributes in clinical study lots to assure clinical performance. This approach may still be appropriate in some cases but could limit flexibility.

2.4 MANUFACTURING PROCESS

In Fig. 2.3, a schematic for a typical biotechnology manufacturing process is described. The upstream process begins with expansion of the cell substrate that produces the product and continues through the final production cell culture to the harvesting of unpurified bulk product. Downstream process steps include initial purification (e.g., chromatography), modifications (e.g., conjugation), and final polishing steps leading to the bulk drug substance. Filling and/or lyophilization of the drug substance are then performed. Many of the important attributes of a biotechnology product are impacted by the upstream manufacturing. The cells used to manufacture the product are miniature factories affected by many variables including subtle differences in media composition, aeration, metabolites, shear forces [30], and cell density. Differences in these factors can impact a wide variety of product structural attributes including glycosylation, oxidation, cellular proteolytic processing, and so on. Many challenging process impurities are generated during upstream manufacturing, such as host cell proteins, DNA, and media components. Systems biology approaches to clonal selection or alteration [31], cellular metabolism, and biosynthesis along with better defined media components, improved models for bioreactor fluid dynamics [32], more sophisticated monitoring (e.g., beyond pH and dissolved O2), and multivariate statistical analysis [33] may improve the understanding associated with these complex process steps. An improved process understanding linking process parameters to important product attributes could allow for a better design of an upstream process. Improved process understanding could also allow for broad design spaces and opportunities for changes in scale, equipment, and so on without prior FDA approval.

Figure 2.3 A schematic for manufacturing of a biotechnology product. The manufacture of a biotechnology product generally involves many steps. Manufacture often being with the thawing of frozen cells then continues through a series of cell culture expansions into the final production bioreactor. A harvesting step to remove cells and other culture components precedes downstream purification. For purification, the product then undergoes a series of purification steps, such as affinity, ion-exchange, hydrophobic, or size exclusion chromatography. The purified bulk product is then concentrated and formulated and may be lyophilized. Additional processing steps such as proteolysis or conjugation may occur as part of manufacturing. These steps would generally be followed by some purification to remove step residuals.

Although there are many potential opportunities for upstream processing, many challenges remain in understanding such complex process steps. Downstream processing for biotechnology products may afford more immediate opportunities for generation of large design spaces. To illustrate approaches to a downstream unit operation, a hypothetical column purification step will be considered. The use of a design space approach is compared to more traditional process limits in Fig. 2.4. Although manufacturing has always had ranges for operating parameters, they have generally been univariate. The pH and protein load of a chromatography column are used as example parameters. The linkage of these parameters to quality has been empirical, often based on the limited ranges used during manufacture of clinical trial material (Fig. 2.4a). Complex biological products have been defined by their manufacturing processes [34]; the process is the product. Changes in the manufacturing process often required a clinical trial to maintain the empirical link between process characteristics and product quality. By the mid-1990s [35], a specified subset of well-characterized biological products could utilize biochemical comparability to allow some manufacturing changes in the absence of new safety and efficacy studies. However, the parameter ranges used in manufacturing clinical or to-be-marketed material are often limited and do not explore the full extent of the ranges leading to acceptable product quality. Small-scale studies can be used to support wider parameter ranges (Fig. 2.4b). However, the validity of the scale-down models needs to be demonstrated and the column performance measurements used should link to product performance. Although small-scale studies can expand univariate ranges, they do not account for interactions between parameters. Multivariate experiments at small scale can define a more meaningful space for parameters that may interact (Fig. 2.4c). Such a space can be generated with an efficient number of experiments using design of experiment (DOE) approaches [36, 37]. As with univariate small-scale experiments, the validity of the scale-down and the linkage of measured responses to product performance should be described. In addition to these considerations, the modeling used in DOE approaches should be justified.

Figure 2.4 Spaces or ranges that can be used in product manufacturing. (a) Biologic and biotechnology products were historically defined by their manufacturing process. The process and process ranges were based on the ranges used for the product used in clinical trials showing safety and efficacy (S&E). Comparability studies allowed for some process changes. (b) These ranges were often narrow and could be expanded in small-scale models. The use of these wider ranges in manufacturing was dependent on the validity of the scale-down models and the performance criteria. (c) Since many variables interact, a multivariate space is more reflective of reality and may be generated using design of experiments. Generally, these experiments are also done at small scale and depend on the validity of the DOE models as well as the scale-down model and performance criteria.

An important advantage of generating a multidimensional design space is described in Fig. 2.5. An empirically derived manufacturing process is static with locked-in parameters (Fig. 2.5a). Any variability in process inputs is transferred to the product since there is no way to compensate. In a dynamic process, information on variability of inputs or outputs can be used to tune the process (Fig. 2.5b). This information can be real time or based on off-line testing. Variable input parameters can then be compensated for. In addition to process monitoring, a dynamic manufacturing system needs flexibility in setting process parameters or a design space. To allow for process adjustments, the design space also needs to predict how movement within the space will impact product attributes. A simple example of compensating for variability, using the design space shown in Fig. 2.4c, is as follows. The product yield from an upstream process was higher than expected, and material would need to be discarded based on protein load limits. However, adjusting the pH could support a higher protein load in the upper left corner of the design space. This would facilitate taking advantage of variability in the upstream process.

Figure 2.5 A design space allows for a dynamic approach to manufacturing that transfers variability from important product attributes to process parameters. (a) The traditional paradigm for pharmaceutical manufacturing utilizes a fixed process. Thus, any variability in inputs may result in variable product. (b) In a dynamic manufacturing process, input variability can be monitored either directly or through product impact. Information regarding this variability can then be used to adjust the process parameters to compensate and produce high-quality product. A knowledge-rich design space that has been explored during development and throughout the product life cycle will allow the process flexibility necessary to compensate for variable process inputs. This figure was adapted from a diagram by Jon Clark.

A design space can be generated for one unit operation, as done for the chromatography example above, or for an entire process [11]. In an entire process, such as shown in Fig. 2.3, final product quality attributes can drive the design of the final manufacturing step and process inputs for the final step can drive the outputs of the preceding step. This approach to unit operations can continue upstream until it defines acceptable inputs and outputs from the initial thawing of a cell bank vial.

2.5 DEVELOPING A DESIGN SPACE

In the previous section, a hypothetical design space for chromatography was presented. Figure 2.6 describes some of the initial steps for developing such a design space. It is important to first define the outputs that the manufacturing step will need to achieve. These will be the responses evaluated in studying the chromatography step. This requires establishing the CQAs that the final product must meet. After that, the chromatography output performance measures must be set so that the complete process will deliver the final product CQAs. The general requirements of the manufacturing step should also have been considered in the initial process design (e.g., choice of the methodology).

In Fig. 2.6a, the step requirements are removal of subpotent charge variants and impurities. A flow-through ion-exchange column is then chosen to achieve these requirements (Fig. 2.6b). The potential factors that could impact ion-exchange performance are listed in a cause and effect diagram, such as the fishbone or Ishikawa diagram (Fig. 2.6c