QbD for Biologics: Learning’s from the Product Development and Realization Case Study (A- Mab) and the FDA OBP Pilot Program July 19-20 th 2010.

QbD for Biologics: Learning’s from the Product Development and Realization Case Study (A-

Mab) and the FDA OBP Pilot Program

July 19-20th 2010

Morning Panel Questions Day 1: Critical Quality Attributes

In assessing which attributes are critical and which are not, to what extent is it appropriate to apply prior knowledge from similar class molecules to a new product? When is it appropriate to leverage company-specific and literature information?

• Leveraging prior knowledge is valuable at the earliest stages of development prior to the gaining of molecule specific data as a product moves on in development

• Prior knowledge of molecular structure at early stages is useful to highlight specific product variants to look for and target the types of analytical methodology required to assess them

• Data is of value regardless of where it comes from as long as its strengths and weaknesses are understood

In assessing which attributes are critical and which are not, to what extent is it appropriate to apply prior knowledge from similar class molecules to a new product? When is it appropriate to leverage company-specific and literature information?

• Although general assumptions can be made about class specific attributes (e.g. MAb terminal heterogeneity) there is inevitably a case where a molecule will not follow that dogma

• The value of general assumptions depends on the depth that knowledge can reach (how specific to molecular structure/function one can get i.e. what is it about glycoform structure that does or does not impact Fc receptor binding)

How much additional molecule-specific information would be required to support the assessment?

• It is unlikely that the criticality of quality attributes for a molecule are going to be identical to another in either type or levels, thus it is worthwhile while investigating the unique aspects of a molecule to check that assumptions about ‘class specific’ knowledge are in fact, correct

– If looking at impact on PK or PD, might as well check all relevant attributes, but depends on resources

– Different dosing regimens, (e.g. IV versus SQ) chronic versus single dose, patient disease state, immunosuppressed etc. might all need to be considered when using ‘prior knowledge’

– When changing indication for a molecule one should revisit the CQA RA

– At what point do we accept an attribute is non critical for all class specific molecules? Regulars appear reluctant to allow this across the board, driving justification in each case

In setting and justifying acceptable ranges for CQAs, what information is required - When is pre-clinical data sufficient and when is clinical data required?

• The value of preclinical data depends on the animal model and its relevance to humans

– Does the ligand/target have the same properties as in humans – impact on PK/PD

– Impact of disease state in humans– Immunogenicity responses in animals impacting evaluation– However, one can expose animals to purified variants

• Clinical data is still the ‘gold standard’ as long as patient variability is accounted for

– Extracting product from serum samples is extremely valuable and informative for PK

– However, utility of clinical data for PD depends on available markers (e.g. increasing levels of blood cells is easier to measure than impact on tumor size/s or overall survival)

In setting and justifying acceptable ranges for CQAs, what information is required - To what extent is prior knowledge with similar class molecules applicable to setting ranges for CQAs?

• Although general assumptions can be made (e.g. MAb terminal heterogeneity) there is inevitably a case where there molecule specific differences exist and ranges on those need to be justified

• Ranges for CQAs depend on each molecule’s manufacturing process capability, patient population, dose strategy etc. It seems hard to justify a single range for a particular CQA across a whole class of molecules – only DNA and endotoxin seems to have achieved that from a safety perspective

• However, it appears the CQA risk assessment tool is now being used across the industry and is seen as a useful mechanism to incorporate prior knowledge, with the caveat that ‘non critical QAs’ or ‘less critical QAs’ are still considered in relation to CPPs and the control strategy with justification as to how they were considered (i.e. not forgotten about!)

In setting and justifying acceptable ranges for CQAs, what information is required - How does stability play into establishing acceptable CQA ranges?

• Stability must be considered for:– Understanding levels of attributes present at time

zero versus those that may change over time until expiry, thus one needs to reflect patient exposure to end of expiry material when establishing ranges (if used in clinical studies – or not!)

– Accounting for the appearance of new attributes as the product degrades over time. This would necessitate adding a Quality Attribute to the PQRA even if it is not present at time zero, and setting an appropriate range

In setting and justifying acceptable ranges for CQAs, what information is required - How do we reconcile the value of establishing broader clinical exposure to product variants with the goals of product development, which continually drives towards comparability, consistency, and higher purity?

• Producing ‘more variable’ product lots early in development can provide patient exposure and the ability to better understand the impact of different levels of attributes on PK/PD (and maybe safety if the study is powered enough)

• However, this level of variability may not reflect commercial process capability, especially at the time of licensure – although it may be important for future changes and provides an expanded CQA ‘design space’

• There is a cost and time associated with producing greater numbers of smaller lots early in development

• Can we justify patient exposure to potentially negatively impacting levels of attributes beyond what is normally designed into dose escalation studies?

• The use of different molecular candidates early in development is another route to understand criticality of QAs

In setting and justifying acceptable ranges for CQAs, what information is required - How does a company broaden CQA ranges based on safety and efficacy considerations?

• Since the definition of CQA is Critical Quality Attribute the assumption is that the attribute might have some impact on safety and efficacy – thus one has to understand at what point the impact has relevance to patients (e.g. aggregates)

• Leverage preclinical and clinical serum samples– Detecting variant clearance over time– Maximize assessment of dose ranging studies– Linking levels of quality attributes to

immunogenicity/safety/efficacy (very hard to do! Most clinical studies are not powered enough, nor able to track giving a patient a single set of levels of variants)

– Epitope mapping if an immune response

In setting and justifying acceptable ranges for CQAs, what information is required - How does a company broaden CQA ranges based on safety and efficacy considerations?

• Introduce increased levels of attributes into an appropriately powered preclinical study to see if impacting levels do get reached

• Use relevant in vitro studies to show limits that do/do not impact PK/PD (e.g. Fc receptor binding, potency assays etc.)

• It must also be considered that data derived from the clinic may lead to attempts to reduce the levels (or strengthen control) of an attribute, if the link of safety/efficacy to a QA can be made post the original CQA risk assessment

What aspects should be considered when assessing interactions between quality attributes? Can the interaction of non-critical attributes render them critical? What information would be required to establish the absence of interactions?

• One should use the design space of say, fermentation, to get an idea of the true ‘QA design space’ in relation to relative levels of QAs being produced before needing to do extensive interaction studies of QAs that are not realistically mamufactured at different levels by the process

• It may be the case that attributes on their own may not appear critical but may interact and become critical - although no specific examples were given

• One can use forced degradation to create high levels of one particular QA (e.g. oxidation) and look at the impact on another QA (e.g. highly oxidized protein aggregating or not)

• It will require the creation of various purified molecules with each of the QAs at specified levels and test in animals or in vitro (if feasible) to show lack of impact to PK/PD and maybe safety – but this is extremely costly and time consuming

Afternoon Panel Questions Day 1: Design Space

Design Space Definition What types of information/data can be used to define a design space (e.g., manufacturing data, DoE, platform/prior knowledge).

• Manufacturing data – Pilot scale runs– Engineering runs– Full scale clinical/commercial runs

• DoE and process characterization• Platform/prior knowledge

– Internal data, published data etc.

• Formulation development• Stability studies• Comparability studies• Product related data used to assign criticality of

quality attributes

Design Space Definition Should a Design Space only consist of CPPs or should non-CPPs be included in the Design Space? When might this be appropriate?

• Design space is for assurance of quality – not just CPPs – however if design space is just CPPs, then how you define CPPs will require a much greater level of data

• If you include some form of control of non CPPs or inclusion in some way into design space, data requirements are less, as opposed to having to have thorough data package to convince everyone that one can avoid controls or inclusion in design space for non CPPs

• However, non CPPs are still controlled in the process within the manufacturing procedure, it is how they are monitored, what their ranges are, how deviations are dealt with etc, that will be different

Design Space Definition Should a Design Space only consist of CPPs or should non-CPPs be included in the Design Space? When might this be appropriate?

• Each company may have different risk acceptance profiles to define what is critical or not – hard for regulators to accept each risk assessment without in depth review

• It is still not clear how to deal between statistical significant impact to a CQA versus practically significant – it appears to be in the eye of the beholder and not something that may be able to be defined universally

• There is concern that the definition of CPP is heavily dependent on the operating range studied that defined its criticality and needs to be reviewed and changes beyond that need to be managed appropriately

Design Space Definition How should parameters that are not included in the design space be handled? Is an infinite range applied to them?

• Such parameters should be controlled within the overall Quality System

– MPs, in process monitoring, change control assessments, risk assessments etc.

– Not a regulatory commitment but filed in the development section

• There must be some consideration of ranges for parameters not included in design space – there is always a point at which a process/parameter can be pushed to have an impact, even if it is very wide

• Such ranges may be based on limits that have been tested beyond normal operations – ‘knowledge space’, although justification of wider ranges may be based on prior knowledge

Design Space Definition What actions should be taken if a unit op response is not as expected, either at pilot or mfg. scale?

• This may mean that prior knowledge of the function/operation of that unit and/or its impact on product may not be correct

• It also depends at what stage of development this occurs – the earlier on it is, the more likely the impact can easily be rectified – late stage failures/unexpected results may require a more comprehensive evaluation of assumptions and data on which design space (or process understanding) is based

• All data relating to that unit operation should be reassessed in light of the failure. Depending on the results, other unit ops, RAs, PQA assessments might need revisiting

Design Space Definition How might a design space change across the lifecycle of a product? What types of new information might identify a new design space limit?

• Knowledge is gained over time that can influence assumptions or add to existing data that can modify the design space (both expand or contract)

• Processes nearly always undergo change and new/altered processes may provide new data that can influence the design space

– NCs, comparability data, stability data, testing at different limits/conditions

• Additional manufacturing, preclinical or clinical data may enhance product knowledge

– CQAs may become non CQAS or visa versa– Process/product impact may become evident with more

manufacturing experience at scale

Design Space Definition How can modifications to a design space be filed throughout the lifecycle of a product?

• It depends on when design space is initially ‘fixed’ – if any changes occur between then and license application, the changes would be described in the MA.

• Should changes occur after the MA and the design space are ‘approved’, filing changes should be related to the extent and type of change (AR, CBE-30, PAS etc.)

– This filing strategy can be preapproved in the ECP as part of the MA and built into the Quality System

• There is a need to have a common understanding of what is required to be submitted in regards to description of the QMS and how that will influence the need to file modifications to design space

Regulatory/ or Submission Impact How should the Design Space be described in a submission?

• Provide justification of parameter scoring from RAs used to design process characterization experiments, including data on which decisions were made

• Justify small scale model qualification against large scale• It is still not clear exactly what parameters to include (the

CPP and non-CPP argument) and how much detail (non-CPP limits tested?)

• Design space description only applies to where a CQA is impacted

• Describe linking of individual steps across process to ensure CQA control

• Process steps with design space are part of license claims with parameter ranges and mathematical models

• Include graphical representations? Data summaries?

Regulatory/ or Submission Impact How should the Design Space be described in a submission?

• There needs to be a balance between more data required versus flexibility for change without reporting

• Data for filing vs data to be available on inspection

• Description of manufacturing and process controls S.2.2

– CPPs vs non CPP, CQA vs non CQA– Description of design space– Input variables, process parameters and QAs covered by

design space– Input material controls and process controls– Model representation, equations, combination of ranges?

• Control of Materials S.2.3– Detailed input material controls– CQA for starting material

• S.2.4 Control of critical steps and intermediates– Input controls etc

Regulatory/ or Submission Impact How should the Design Space be described in a submission?

• Development S.2.6 – Development strategy, CQA, CPP selection, QRM, Prior

knowledge, DOE, MVA/univariate, lot/process history, comparability

• S.2.5 Process Validation and or Evaluation– Evaluation of operating units– Storage/hold times column lifetime, compatability, viral safety

etc. – Evaluation of design space– Validation – confirmation of consistency (in process and end

product) - movement to continuous process verification– Continuous process verification protocol?– Change management protocol for changes?

• Also show design space model not impacted by change

– Stability protocols?

Regulatory/ or Submission Impact How can movement at the edges of the Design Space be justified/implemented (e.g., 'adaptive' control strategy, statistically justified, etc)?

• Statistical limits can be bound into the design space (statistical boundaries, CPKs etc.) to provide a level of confidence when approaching edges of design space

• When at the limits of design space, the qualification of small scale models (where edges are usually defined) is even more essential

• One could increase testing if at the edges of design space to assure product quality

Regulatory/ or Submission Impact How can movement at the edges of the Design Space be justified/implemented (e.g., 'adaptive' control strategy, statistically justified, etc)?

• Not all edges are designed equal – some may be a cliff, others may just be a gradual difference, so statistical limits can be applied as appropriate

• The QMS may treat excursions to the design space differently depending on the excursion and potential impact to the product – one could file a strategy of how such excursions would be handled (more studies, based on existing knowledge, risk assessment etc.) – or how the QMS will deal with movement near the edges and the associated uncertainty

Quality System and Lifecycle Implications What is the role of the Quality System in approaching CPP vs. non-CPPs - Impact to deviations or excursions

• A Quality System change management program is essential to assure regulators and the manufacturer that changes within the design space will be dealt with appropriately and may not have to be reported, or reported with a reduced reporting category

• Deviations/excursions should be dealt with through the normal NC/deviation QMS, with enhancements required to ensure adherence to design space and/or eCPs – the impact of a deviation to a non CPP may not require the level of investigation to that of a CPP depending on the nature of the deviation (e.g. within the knowledge space) – actually not much different than now going within or beyond validation limits

Quality System and Lifecycle Implications What is the role of the Quality System in approaching CPP vs. non-CPPs - Impact to deviations or excursions

• Deviations that require revision of the design space (shrink or expand) may require some sort of a filing (level TBD depending on type of change) to ‘reapprove’ it, or at a minimum to keep the agency informed

• If a deviation reveals that a non CPP is in fact a CPP, then the design space and other related systems (e.g. risk, small scale model qualification etc,) would need revising through the QMS CC procedure

Quality System and Lifecycle Implications What is the role of the Quality System in approaching CPP vs. non-CPPs - Planned movement within the design space or approved protocol

• The QMS should be able to handle movement within the approved design space through preapproved enhancements to such systems as Change Control or Process Monitoring that ensure appropriate documentation, process control and product monitoring etc. to ensure that there are no shifts in process capability or product quality

• Obviously the level of change management will be different for non-CPPs versus CPPs as far as how the change control QS handles movement (level of assessment, testing data required, post change monitoring, reportability etc.)

• A non CPP movement beyond the range that defined its criticality would need enhanced scrutiny – perhaps a defined limit to movement within the range (e.g. 50%) would be a compromise to allowing totally free movement

• Aspects of the QMS enhancements required can be filed whilst others should be available on inspection

Quality System and Lifecycle Implications What is the role of the Quality System in approaching CPP vs. non-CPPs - Post-approval, how should the quality system manage and document oversight of their continuous monitoring process and how should process improvements/optimization be implemented and communicated to the Agency?

• As described, the QMS can be enhanced to include improved process monitoring (e.g. holistic monthly product review), statistical trending and appropriate actions should trends be found. These enhancements can be filed.

• The management of process improvement filings can be predefined as part of the change control process depending on the level of change and can also be filed (e.g. as part of an eCP)

• Differentiating between what applies to CPPs versus non CPPs is as discussed previously

• Could S.2.2 include a commitment to update say, the design space equations through the APR

Morning Panel Questions Day 2: Control Strategy / Lifecycle Management

How would a control strategy look different in a traditional submission vs. a QbD submission?

• A QbD control strategy is based on a holistic, comprehensive assessment of the criticality of quality attributes, linking that to impact of process and defining process controls and product testing to assure quality, safety and efficacy

– Includes risk assessments, prior knowledge, enhanced molecule and process understanding, leveraging preclinical and clinical data and testing capability

– Therefore IPCs, specifications (product and raw materials) and the stability program will be based on criticality of quality attributes and probably be more streamlines and have less items (or less with high stringency) than a traditional approach

How would a control strategy look different in a traditional submission vs. a QbD submission?

• A QbD control strategy should consider the impact of unit operations to product across the manufacturing process and thus interactions between unit operations

• Moving control to the process to deliver high quality product rather than testing quality into the product

• A QbD control strategy will also include the concepts of continuous verification (increased multivariate analysis etc.) and continuous improvement – a lifecycle approach

• The QbD strategy would inevitably include more data and justification in the process characterization, process control and justification of specifications sections in a filing

• A QbD control strategy also needs to deal with the various levels of uncertainty for the design space

How would parameters that are not specified in the license be handled and what will be the agency involvement?

• Non specified parameters should be handled by the Quality Management System

– Process monitoring, change control, NCs etc.

• How the QMS will deal with these parameters (non CPPs, non critical inputs and outputs, non critical quality attributes etc.) can be described in a filing or is available on inspection

Are there additional considerations beyond criticality of a given attribute that factor into the control strategy development?

• An attribute that indicates process consistency that can not be easily measured through another parameter (e.g. glycosylation) may need to be considered as part of process monitoring or on comparability, but not necessarily routine lot release or stability – the old ‘If variation is occurring, what you don’t know may be happening’…….

• An attribute that provides data about the ability to supply patients (e.g. yield) would require some form of assessment (in-process)

Are Expanded Change Protocols (eCP) and EU Post Approval Change Management Protocols (PAMP) the same? If not, what are the key differences?

• No idea• The definition of the eCP is not rigorous

at this stage, nor fully fleshed out• PAMP is also very new

Afternoon Panel Questions Day 2:Quality by Design (QbD) for Non-Monoclonals

What challenges would there be to justifying the described immunogenicity design space for a vaccine?

• Need to understand how the fragments/3-D structure/vaccine truly impact the immune system – i.e. stimulating the bits we want and not others (want natural protective immune response)

• Need to design additional studies to further examine how the product works

• QbD can be applied to vaccines and one needs to know how to manufacture the product and how it works – continuous supply for vaccines is no different than any other product

• May need to go beyond the traditional potency assay to better characterize and predict response

• Immune response is a biomarker but may not necessarily reflect efficacy

• An understanding of the patients and their response to a vaccine is also important

What studies would be needed for a therapeutic protein to justify an ‘immunogenicity design space’ where immunogenicity is undesirable?

• First and foremost is the attempt to understand what actually causes immunogenicity for any particular product

– Epitope mapping of antibodies to try to identify where in the molecule the antibodies bind

– Monitoring which lot of material each patient gets and controlling the levels of quality attributes patients get

– Patient specific responses to products (e.g. MHC contributions)

• Preclinical or non clinical studies to understand the immunogenic potential of the product (in silico, T-cell, in vivo etc.)

• A thorough understanding of the product variants and process related impurities

• Any prior knowledge

QbD has several elements that can be applied across multiple product types and associated systems. What are the essential components of QbD that can be applied most generally?

• Plan/Design– Molecule design, equipment, facility etc.

• Execute– Training, clear SOPs, streamlined processes/methods

• Monitor– SPC, MVA etc.

• Continuous Improvement– Data review to drive actions, CAPAs, CAPA EV

• Risk assessments

• Some elements of QbD are becoming regulatory expectations – feeling we have been doing ‘QbD-lite’ for years (e.g. process/product interactions, criticality of in process controls etc.) and cGMP is an expectation

– Better justification/reassessment of specifications and IPCs– Risk assessments– Good science and common sense

What elements of QbD appear to be the most difficult/costly/time consuming?

• Design of experiments, data accumulation and reporting of design space

• Multiple risk assessments• Developing extensive eCP rather than one off

comparability protocols• Not having a global acceptance of QbD leading to

different files in different jurisdictions

• We should not allow the above to give the impression that QbD is not worthwhile – many aspects of QbD are very cost effective (molecular design, CQA understanding) and can be applied by companies regardless of size, product or process

How are companies making decisions over how much, if at all, QbD will be applied to a particular product, especially considering early phase, late phase and licensed products?

• All aspects of QbD may not be financially viable to be applied to all molecules (e.g., attrition of molecules during development, cost, time to carry out specific aspects of QbD etc.)

• There is a need to establish a strategic framework in order to guide circumstances to apply QbD

Probability of Success

• Knowledge of biological pathway (e.g., established understanding, proven pathway)

• Availability of clinical data• Market position (e.g., first vs third)

Process / Product Complexity

• Platform vs off-platform• Lyo vs liquid drug product• Established vs new technology• Regulatory risks (e.g., patient population, indication)

Material Demand• Chronic vs acute• High potency vs low potency• High vs low plant utilization

Global acceptance of QbD by regulators is one barrier to the implementation of QbD on a holistic basis. How are companies managing with global filings?

• Essentially producing two different files or

• File all data and wait for the questions………..

What are the main concerns companies have in implementing QbD?

• The impact to the expense and time to develop a product

– Will QbD really allow for more rapid development if platform knowledge is applied?

– Certainly a benefit for molecular design and process development

• Questions from regulators as increased data is provided in filings

• Will the cost/time of QbD be recognized from a COG perspective as opposed to ‘a better process and product understanding’

– Will QbD have inherent value to industry?– Does the traditional approach really lead to multiple failures

that QbD prevents?– Is a QbD product of better quality?

• If not designed in, not necessarily true

• Concern seems to be focused around design space, the efforts to create it and what regulatory/QMS relief, if any, will there be

Apart from process and product, on what other applications can QbD have an impact?

• Equipment design, implementation and execution• Facility design• Utilities• Raw Materials• Containers• Transport• Quality Management Systems