Cytotherapy, 2014; 16: 1187e1196

Reference materials for cellular therapeutics

CHRISTOPHER A. BRAVERY1 & ANNA FRENCH2 1

Consulting on Advanced Biologicals Ltd. Advanced Biologicals Ltd, London, United Kingdom, and 2The OxfordeUCL Centre for the Advancement of Sustainable Medical Innovation (CASMI), The University of Oxford, Oxford, United Kingdom Abstract The development of cellular therapeutics (CTP) takes place over many years, and, where successful, the developer will anticipate the product to be in clinical use for decades. Successful demonstration of manufacturing and quality consistency is dependent on the use of complex analytical methods; thus, the risk of process and method drift over time is high. The use of reference materials (RM) is an established scientific principle and as such also a regulatory requirement. The various uses of RM in the context of CTP manufacturing and quality are discussed, along with why they are needed for living cell products and the analytical methods applied to them. Relatively few consensus RM exist that are suitable for even common methods used by CTP developers, such as flow cytometry. Others have also identified this need and made proposals; however, great care will be needed to ensure any consensus RM that result are fit for purpose. Such consensus RM probably will need to be applied to specific standardized methods, and the idea that a single RM can have wide applicability is challenged. Written standards, including standardized methods, together with appropriate measurement RM are probably the most appropriate way to define specific starting cell types. The characteristics of a specific CTP will to some degree deviate from those of the starting cells; consequently, a product RM remains the best solution where feasible. Each CTP developer must consider how and what types of RM should be used to ensure the reliability of their own analytical measurements. Key Words: reference materials, cellular therapeutics, metrology, standards

Introduction The case for appropriate standardization is clear (1), and the craft-to-mass paradigm shift described by Griffiths (2) bears many similarities to the situation we have today with cellular therapeutic (CTP) manufacturing. The term “standard” can encompass a variety of meanings and may be interpreted differently by different people, ranging from social norms to formalized agreed documents to physical materials; this imprecision of language can lead to confusion and misunderstanding as to the nature of the standard being discussed. The focus of this discussion is manufacturing and quality in product development, much of which relates to in-house standardization. However, to bring conformity to manufacturing and comply with regulatory expectations, there is a need to apply external standards. Examples of the types of standards that are covered are given in Table I, but these examples are by no means exhaustive. CTP are defined here as those cell therapy (including tissue engineered) products that are regulated as medicinal products (drugs) and thus subject to standardized quality systems, namely

Good Manufacturing Practice, and similar systems in place for other aspects of development, such as Good Clinical Practice. These quality systems standards ensure minimum standards such as hygiene, traceability of documentation and calibration of equipment and ultimately contribute to product conformity, traceability and safety. Other written standards exist such as pharmacopeia, which provide quality criteria for certain raw materials, allowing suppliers to market material of the same generally acceptable quality to many different manufacturers, reducing in-house raw material testing and facilitating material changes. Pharmacopoeias also provide some standardized test methods (eg, sterility) with reduced or no need for validation, which also facilitates regulatory review because the results are recorded in standard units and thus their meaning is immediately understood by the regulator. As will be discussed later, pharmacopoeias also supply a range of reference materials. With emerging fields such as CTP, there is often a need for standards beyond those imposed by regulation, for example, the need to agree to

Correspondence: Christopher A. Bravery, PhD, Consulting on Advanced Biologicals, Ltd, London E1 4EW, UK. E-mail: [email protected] (Received 24 March 2014; accepted 30 May 2014) ISSN 1465-3249 Copyright Ó 2014, International Society for Cellular Therapy. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcyt.2014.05.024

1188

C. A. Bravery & A. French

Table I. Examples of types of standards. Example standard Base quantity

Physical (material) standard Reference material Certified reference material Gold standard Product reference material Written standard Reference measurement procedure (reference method) Pharmacopoeia method National/international standard

Consensus standard

Guideline

Description and purpose Foundation of the International System of Units (SI) Quantity, measuring unit (abbreviation)
Length, meter (m)
Mass, kilograms (kg)
Time, seconds (s)
Electric current, ampere (A)
Thermodynamic temperature, Kelvin (K)
Amount of substance, mole per liter (mol/L)
Luminous intensity, candela (cd) Of the seven base units, only the kg is not (yet) derived from a physical constant of nature. A material of known or arbitrary quantity used for a specific comparison purpose, for example, international prototype kilogram. A consensus agreed reference material produced by a recognized body such as the World Health Organization, United States Pharmacopoeia (USP), European Pharmacopoeia (Phur.Eur) Historic reference of “commercial value” by which a currency unit (eg, $, £) had a value relative to a quantity of gold. Example product material from a qualified process to be used for comparative quality purposes (usually in-house only). A consensus method that if conducted exactly as described will provide a suitably reliable measurement for the intended purpose. A regional/national standard method applied to pharmaceuticals that may be a reference measurement method or may require the use of a CRM. Its use may be mandatory. Such as International Standards Organisation (ISO), European Committee for Standardization (CEN), British Standards Institution (BSI). Written standard that may or may not be mandatory for a particular industry. Such standards can, for example, describe processes (eg, quality system) or testing requirements (eg, medical device compliance). Industry or organizational agreed written standard, can cover codes of practice, standard operating procedures, agreed definitions (eg, International Society for Cell Therapy consensus definition of MSC). Usually non-binding recommendations and advice from a government agency (eg, European Medicines Agency (EMA), US Food and Drug Administration), industry group, and so on, covering a specific category or type of product.

These examples are descriptive and do not necessarily use standard terms and are used to convey a sense of the range and nature of standards available.

common terminology (3). The scientific community has, for instance, agreed on some minimum cell characteristics that define human multipotent mesenchymal stromal cells (MSC) (4). As with all standards, the purpose must be clearly stated and the standard not misapplied. Dominici et al. (5) specifically state that the criteria should only apply to research, and These identifying criteria should not be confused with release specifications for clinical studies. In a recent paper reflecting on 66 investigational new drug development submissions for MSC-based CTP, the US Food and Drug Administration acknowledges this last point but comments that many researchers seem to believe otherwise. The objective of the following discussion is to identify the need for physical reference materials for quality control of CTP along with some of the issues faced in their preparation. The diversity of methods and products means that the individual needs of developers will differ, and many reference materials

(RM) will need to be prepared in-house; consequent generalized solutions cannot be presented. Reference materials The concept of reference materials remains a fundamental principle in measurement; of the seven base units (Table I) of the international system of units (SI), only mass (the kilogram) is still dependent on a physical standard. Although a number of attempts have been made and are still under investigation to supersede this international prototype kilogram, all measurements of mass can be traced back through various copies to this one standard (6). Without standards, science would be very much less certain and observations from different labs could neither be compared nor replicated. Two measurement results are only comparable if they can be traced to the same reference (7), which should ideally be internationally recognized (eg, SI units, certified reference material). This issue affects

Reference materials for cellular therapeutics medicine regulation because many analytical methods used to control the quality of medicines rely on inhouse RM and consequently, even where available in the public domain, data from one manufacturer cannot necessarily be directly compared with data from another manufacturer. In part, for this reason, generic biological medicinal products are not possible; instead, the approach of biosimilarity must be followed and the biosimilar developer must obtain product from the innovator and directly compare it with their biosimilar product through the use of their own analytical methods. A key objective of manufacturing is product consistency; to confirm that a process results in a consistent product, it is necessary to measure the product’s critical quality attributes (CQA), those that define its quality. For simple physical objects such as paperclips or screws, this is typically straightforward; standards for length and mass are well established. For biological materials and in particular living cells, it gets much more complicated because for many attributes, no certified reference material (CRM) exists. For a developer of CTP, this means they will not only have to develop a process but also develop many of the analytical methods. This means not only will the process be unique, but also many of the methods that control it and determine the quality of the resulting product. This situation is not exclusive to CTP, the complexity and inherent heterogeneity of living cells poses novel issues including the paucity of suitable reference materials and the need to make measurements that cannot be quantified in SI units (e.g., biological activity). For developers of any product, including CTP, there are two broad uses of RM that must be considered (Table II). First, a product RM is needed, such that the overall quality of future batches can be compared with it to ensure product consistency both of the current batch and for batches over time to identify any changes (process drift).

1189

Because process changes ranging from different sources of critical raw materials through to process improvements are expected to occur, both during development and over the product life cycle, a product RM is also critical for confirming comparability (Figure 1). This becomes particularly important for complex medicinal products, such as those made from cells, because their structure and function, and thus quality, cannot be completely determined. A typical approach is provided in Figure 2; an early master RM is prepared once the process is sufficiently developed to initiate preclinical safety studies, such that a bridge between those initial safety data is maintained as the process evolves toward the final commercial process. During pivotal studies, a new master RM is typically prepared from material made by the same process used for those studies, such that after authorization a link is retained to the product that was used to demonstrate efficacy. To reduce the amount of master RM needed, usually a working RM is prepared and qualified against the master, and this is used for routine applications. Ideally, the master RM would be sufficient to last for the product life cycle, perhaps 20 or 30 years in the case of medicines, though in many cases the RM will not be sufficiently stable or may not be available in sufficient quantities, necessitating the preparation of a new master RM from time to time. The second use of RM applies to analytical methods. A discussion on metrology is beyond the scope of this review, but a simplified overview of some of the sources of measurement uncertainty and error is provided in Figure 3. The choice of RM is dependent both on the nature of the measurand (see Table III) and the analyte (substance to be measured, eg, interleukin [IL]-2, endotoxin) to be measured; these must be suitably similar. Importantly, a method RM is not intended to determine assay accuracy but to facilitate the analysis of assay

Table II. Reference materials categorization. Product RM Representative product batch material Uses include:
Comparability following process changes, including technology transfer
Identification of process drift in quality over time
Establish a link from early-stage clinical trials through to commercial phase
Ensure that quality attributes are tracked through product lifecycle
Product batch is representative of intended product and is of acceptable identity, purity, strength (content) and potency
Stability testing

Measurement RM Representative of analyte (eg, purified analyte or product RM) Uses include:
Assay qualification and validation
As a calibrator in quantitative analytical methods
Determination of relative biological activity (including potency)
Define assay acceptance criteria (positive control)
System suitability testing
Identification of method drift over time

Reference materials can be categorized on the basis of use as a product or measurement RM; however, in practice, the product RM can be the most suitable RM for certain analytical tests, in particular, potency methods.

1190

C. A. Bravery & A. French

Figure 1. Simplified concept of the use of a product RM confirms product quality after process changes over the product life cycle. Process changes are an essential part of product life cycle management and are required for a variety of reasons such as on-going regulatory compliance, change of supplier for critical raw materials, process improvements to reduce cost of goods and changes of manufacturing scale or strategy to meet market demands. Such changes require regulatory agency approval (European Union, variation; United States, supplement) with supporting evidence for comparability. Comparability demonstrates that A ¼ B, D ¼ E or G ¼ H but cannot confirm that H ¼ B without access to a suitable RM prepared from process iteration B.

reliability. Certain characteristics will be unique to the product, such as potency, and so the product itself may in these circumstances be the only material from which a suitable measurement RM can be prepared. For example, the concentration (mass/volume) of IL-2 in a sample can be measured by means of enzyme-linked immunoassay (ELISA); the analytical principle used being permanent binding of a specific antibody that is in some way tagged (eg, fluorochrome, enzyme) and the complex immobilized (eg, bound to plastic) and when necessary incubated with a substrate (eg, for enzyme). The relationship between concentration of IL-2 and signal (eg, optical density) typically exhibits a sigmoid curve, the working range of the assay being that part where the relationship between concentration and signal is (sufficiently) linear. The signal will also vary, depending on the experimental conditions (eg, substrate incubation time, temperature, pH), many of which cannot be completely controlled. Consequently, the method would have little utility without an RM that could be

compared with the test sample because the measurement would otherwise not be linked to a known IL-2 concentration. In the case of IL-2, a range of CRM are available of documented purity, allowing a standard curve to be prepared from which an estimate of the sample IL-2 concentration (mass per unit volume) can be determined along with some estimate of the uncertainty. The added benefit of the use of a CRM, or a RM qualified against a CRM (with documented traceability), is that results can also be compared between laboratories with the use of the same CRM. Beneath this simplicity there are other considerations; it is important to understand the nature and purpose of the RM used. For instance, the presence of glycosylation on a native protein may alter its binding to the capture/detection antibodies of an ELISA compared with an aglycosylated recombinant protein RM, such that resulting quantification may be inaccurate. In contrast, the use of an RM to assign units of activity in a bioassay, whether the test protein is glycosylated or not, does not alter the validity of the measurement because it is “relative” (to the

Reference materials for cellular therapeutics

1191

Figure 2. Typical approach taken for product RM. During early development, a product master RM is prepared from an early batch/es of product used for preclinical and/or early clinical safety studies. During pivotal studies, a new product master RM is prepared from batch/es used to demonstrate efficacy, such that later batches (throughout product life cycle) can be compared with material known to be effective.

RM), not an absolute quantification (concentration). For this reason, all RM should have a certificate of analysis that includes metrological traceability and defines the purpose (uses for which it was prepared) of the RM, including those prepared in-house. Thus, the suitability of an RM for a method necessitates that quantities of the same kind are compared, for example, mass, volume. When a complete understanding of the measurand is not available (eg, CTP function), the use of an RM for that measurand is only valid if undertaken through the use of an identical or proven equivalent measurement procedure. To illustrate with an example (7,8), the crude fiber content of animal feed (the measurand) cannot be defined by structure but is instead defined empirically as “amount of fat-free organic substances which are insoluble in acid and alkaline media.” An arbitrary RM can be prepared such that relative measurements can be made to it to allow common units of measurement either in-house or after qualification of the RM though collaborative studies, between labs. However, the measurand is defined by the method (briefly, grind, weigh, acid digest, alkaline digest, dry, weigh residue then ash residue and re-weigh). Undertaking the method differently (eg, different approach, reagents of different specification) will not result in comparable results, even if the same RM is used. The implications for this are that changing an analytical method used for a CTP must be undertaken with great care, where the measurand can be defined, for example, mass of protein secreted, then applying a different or modified method can be shown to be valid. For biological activity in particular, the measurand

probably will not be definable, so changes in the method must be handled with great care, and extensive data probably will be necessary to demonstrate equivalence. In some cases, it may be necessary to accept that the method measures something different, which in no way invalidates its usefulness but merely that measurements from the two methods are not comparable. Another approach, perhaps less amenable to the types of analytical methods used with cell products, is that of a reference method (reference measurement procedure). A reference method provides standardized detail of how to undertake the test and prepare calibrators such that the results are comparable over time and between laboratories with acceptable uncertainty (for the intended use). Such methods have been developed, for example, for total cortisol in blood (9) and blood hemoglobin content (10). In general, reference methods tend to be based on simple chemistry; however, for cellular products there is optimism on the basis of efforts to devise a reference method for determination of cell concentrations (11) and leukocyte differential counts (12). Issues faced by developers of CTP Developing RM for living cell products poses some unique challenges, primarily because of the inherent instability and heterogeneity of even uniform populations of cells and their ability to respond dynamically to their environment, together with the need for a RM to be suitably representative of the analyte to be tested. Because living cells do not have ideal characteristics for an RM, we will first consider how

1192

C. A. Bravery & A. French

Figure 3. Simplified concept for the use of a measurement RM with analytical methods. All analytical methods are subject to source variance and error leading to measurement uncertainty or error. The use of a measurement RM of a known (or arbitrary) value ensures that results on different occasions can be compared (for a validated method). Examples (non-exhaustive) of sources of error and variation are shown.

non-cellular RM might be used in analytical methods and consider their advantages and limitations. For example, flow cytometry is a workhorse for all developers of CTP, yet the apparent simplicity of the user-interface belies the complexities that lie within and are outside this discussion. When fluorescence intensity is used to assign a cell as positive or negative (ie, percentage of positive cells), accurate calibration may be of lesser importance (although still important), but when the measurement is mean or median fluorescent intensity, reliable calibration is essential. In principle, standardization of fluorescent measurements would allow comparison of results between devices and laboratories; however, a recent collaborative study gives insights into the difficulties in doing this (13). A range of calibration beads are available, but none of these can be traced to a CRM, and bead surfaces stained with the same fluorochromes used to tag antibodies are generally not very stable. Beads with imbedded fluorochromes are more stable, but the excitation and emission spectra of these typically do not match the fluorochromes used to stain cells. Nevertheless, acceptable reproducibility can be achieved with a single cytometer through the use of a combination of beads and a suitable biological reference. This brings us back to the issue of a suitable biological reference for flow cytometry; one possibility is to use a receptor known to be consistently and stably expressed. For example, in one study, the relative fluorescent intensity of lymphocyte CD4

staining was measured on 57 healthy individuals and found to have a biological and technical coefficient of variance of only 4.9% (14). More recently, a lyophilized CD4 lymphocyte standard has been prepared and its use as a quantification standard explored (15). A similar approach might prove useful for cell therapy developers, either by the use of normal donor material or potential cell lines. More detailed discussion on the development and validation of flow cytometry (16e20) and interpretation of data (21) are provided elsewhere. Molecular biology techniques likewise often lack suitable CRM, although some actives are underway (22,23). However, the use of relatively little cellular material, a significant amount of in-house RM can usually be prepared fairly easily and can be stored in a form that should be stable for many years. In most cases, for quality control purposes, it is likely that the product itself would be the most suitable source for such material, and a scheme similar to that described for the product RM probably would be followed (Figure 2). The real challenge comes from assays of biological function (24), including those used to confirm potency, because these rely on the activity of living cells. Typically, for biological products, such methods use the product reference material because CRM for function rarely exist, so function/potency must be determined relative to the product RM. Nevertheless, for some recombinant proteins such as erythropoietin, granulocyte colony-stimulating factor and interferon-a, there are World Health

Reference materials for cellular therapeutics

1193

Table III. Terms and definitions. The following terms are broadly synonymous:
Reference standards
Reference materials
Laboratory standards
Analytical standards
Standard reference materials The preferred term used here is reference materials (RM), but, when it is helpful to distinguish RM made by recognized organizations (such as European Directorate for the Quality of Medicines (EDQM), United States Pharmacopoeia (USP), World Health Organization), these will be referred to as certified reference materials (CRM). In addition, the following terms are defined: Master RM: highest order of in-house RM, for example, master product RM (also called primary RM). Working RM: batch of suitable material qualified against a master RM (also called secondary RM), for routine use to conserve master RM. Measuranda: quantity to be measured. Quantitya: property or phenomenon, body or substance, when the property has a magnitude that can be expressed as a number and a reference (eg, g, mL, IU). Analytea: specific substance to be measured (eg, IL-2, hemoglobin). Process drift: an unintended, unexplained or unexpected trend of measured process parameter(s) and/or resulting product attribute(s) away from its intended target value in a time-ordered analysis over the lifetime of a process or product (27). Method drift: An unintended, unexplained or unexpected trend of the metrological properties of a measuring system to change over time. a

Adapted from Barwick and Pritchard (7).

Organization CRM, and so in some cases in which an assay read-out involves measuring a secreted cytokine or growth factor, it may be possible to use these (eg, IL-2 example discussed previously). However, such RM will only serve as a calibrator for the analyte (secreted protein) and so are more akin to the enzyme/substrate system of an ELISA and cannot control for the cell response that leads to secretion of the factor (measurand). Clearly, the ideal RM here probably would be a product RM because it is unlikely that such assay systems could be sufficiently standardized to be reliable over many years. An alternative, in the absence of a product RM, might be the use of a cell line exhibiting the same activity by the same mechanism (although not necessarily the same magnitude), which could perhaps be clonally expanded to improve consistency and banked. These and other approaches will have to be determined for each specific assay associated with a product. Reliable measurement RM and fastidious assay development and standardization can bring significant improvement to analytical measurement reliability and remove at least some of the uncertainty as to whether a test result today can be compared with one in a decade’s time. However, the manufacturing process itself is also subject to most if not all the same sources of potential variance identified for analytical methods (Figure 3). Combined with the probability that not all (or perhaps none) of the true CQA of the product are understood, the need for a product RM is perhaps greater than for other types of medicine in which their structure and CQA be more easily resolved. It is also highly likely that the understanding of the mechanism of action will evolve, even after approval, leading to new or improved analytical methods. Without a product RM, it will not be

possible to understand these characteristics in the context of the primary clinical evidence. More importantly, deliberate process changes aside, there is a need to monitor the process for process drift (ie, trending) and discern this from any assay drift, again over the entire product lifecycle. The first obvious problem for a living cell RM is stability; currently, the only realistic option would be for the RM to be cryopreserved. When the CTP is cryopreserved, this should not pose an obvious problem; when the CTP is released fresh, a cryopreserved RM will bring a host of additional issues. The second obvious problem is the quantity of material needed; even allogeneic CTP tend to be manufactured at a relatively small scale, autologous CTP typically at a scale of a single dose, and commonly with limited additional material even to undertake release testing. Because the product RM should be representative of the process used for clinical manufacture (including process impurities), increasing the manufacturing scale is not a realistic option because of the uncertainty associated with establishing comparability. Additionally, this approach would then remove the direct link to product used in preclinical and/or clinical studies. As shown in Figure 2, pooling of batch material is possible, and for autologous probably advantageous to create an “average” product. However, for autologous in particular, donors are typically patients and collection of material other than to treat the patient may be unethical. Normal or even cadaveric donor material might be acceptable as an alternative but will require careful testing to ensure it is representative of the product. Commercially, in particular for autologous products in which a large number of batches are tested, the cost of preparing sufficient product RM may be prohibitive.

1194

C. A. Bravery & A. French

Discussion The objective of this discussion is to identify the need for RM for CTP, both to provide measurement reliability and to confirm product quality consistency over the product life cycle. The intent is not to provide solutions but to highlight why RM are needed, with examples when useful. Each developer needs to evaluate the best solutions for their product and quality control approach; however, there is scope for industry-wide efforts to develop RM for common methods such as flow cytometry. RM are a regulatory requirement because reliable measurement is key to demonstrating product consistency and hence reliability, something that must be maintained over many years or decades. Product RM are also important to provide a link from early safety studies through to pivotal safety and efficacy studies and from those throughout marketing. The absence of a product RM makes later process changes more difficult and could even mean that additional clinical data are required when comparable quality cannot be reliably confirmed. Others have identified the need for RM for CTP, most notably the MSC Reference Materials Working Group (25), who propose preparing a consensus MSC RM, although they specifically state this would not be a “gold standard” MSC. Their proposal appears to envisage an RM that can be used for a wide range of measurements, including the sorts of complex immune assays proposed by Krampera et al. (26). As discussed here, care must be taken when developing complex biological assays for complex biological substances, such as CTP, to define carefully both the analyte and the measurand [Table III (27)], and the measurand itself must be carefully defined. For example; “protein (analyte) in urine” (measurand) will not provide results that can be meaningfully compared because it does not define the conditions sufficiently. In contrast, “protein in 24-hour urine” will result in measurements that can be meaningfully compared as long as the same RM or reference method is used for the analyte. Defining the measurand will be the first challenge for any method that might be used by a range of different laboratories. In many cases, the analyte measured will be a secondary outcome of the assay, for example, secreted IL-2 in a mixed lymphocyte reaction, requiring RM both for the assay and to calibrate the measuring system. The “crude fiber” example given previously seems a good analogy in that the measurand may need to be defined by the method, meaning the method itself may need to be standardized. The second challenge will be demonstrating that the method is sufficiently reproducible, even with an RM. Assuming that it can be agreed

even one measure of biological activity is worth the work to develop it as a standardized method, considerable collaborative work will be involved to qualify and understand whether it is valid to use with a diverse range of MSC-derived CTP. These complexities tend to limit the utility of an individual biological RM to one or a few specific uses, and, for potency assays in particular, these are likely to be unique to the product and indication (24). As discussed, a product RM has a very specific purpose that is distinct from that of a measurement RM, although the product RM is often the most appropriate measurement RM for potency methods. Whether a consensus living cell RM such as one for MSC would be of value to the academic and CTP community beyond one or a few analytical methods is worth considering. First, an agreed phenotypic definition for a particular type of cell could more easily be defined through a written standard, which could, when appropriate, define the method and even specific antibodies along with available measurement RM to use for the method. Wherever possible, nonliving RM probably would be more reliable or, when these could be devised, reference method approaches such as those currently being explored (11,12,15). Any consensus on the definition of a particular type of cell will probably need to come from cells isolated by a variety of methods from a variety of sources and tested with standard methods and appropriate RM to ensure that results can be compared. Second, as already discussed, cellular functions are probably best defined through standardized methods because otherwise, comparisons probably are not possible even with an RM. These methods will still require measurement RM, but whether a primary viable cell would be the best solution must be determined for each proposed test method. The more challenging question is whether, and to what degree, any of this can enable comparability between CTP, even based on the same cellular starting material, such as MSC. It is clear that appropriate use (that intended and qualified) of an RM will result in measurements that can be meaningfully compared within and between labs, subject to the difficulties identified above being addressable for each measurement. For fresh starting material, comparison to a written standard and the use of appropriate RM for each test method should allow reliable comparisons. However, CTP by their nature are the product of a manufacturing process that may or may not have the objective of retaining the characteristics of the starting cell population. Aside from changes caused by their manufacture, the mechanisms by which the CTP acts on the body will be dependent on other factors, such as cell density,

Reference materials for cellular therapeutics excipients, process and product impurities, route and schedule of administration and the underlying disease of the patient. During development, a range of measures of potency will need to be developed that are specific to the CTP and indication (24). Thus, CTP, whether identified as derived from a particular cell phenotype or not, cannot be adequately defined to allow meaningful comparisons; even direct sideby-side analysis will have limitations. It would therefore be very dangerous to assume that the clinical effects of two CTP derived from a similar cell phenotype could be compared except through controlled comparative clinical studies, such as those done for marketing reasons (typically Phase IV). A consensus cell RM for a particular cell type would therefore have little utility for CTP and could not be a substitute for a product RM. Conclusions Most methods used to characterize and control the quality of CTP need suitable RM to ensure measurement reliability over time; currently, in many cases, suitable RM are not available and will need to be prepared in-house. A product RM is also an essential tool to ensure that the quality of each batch of product is maintained and to support comparability after necessary process changes over the product life cycle. However, in many circumstances, the classic approach used to date poses significant challenges because CTP are composed of living cells that do not have ideal characteristics for an RM; consequently, alternative solutions are required. What happens to the phenotype or function of a cell population when they are manufactured into a CTP is a matter for the developer to reconcile for each product and intended use. Whether the cell population of a CTP still meets the phenotypic definition of the starting cell population is largely irrelevant, though perhaps of academic interest. It is important that the CTP is demonstrated to have a consistent quality and have an acceptable risk/benefit in the chosen indication. Each CTP developer must consider how and what types of RM should be used to ensure the reliability of their own analytical measurements. The case for a consensus ‘cell-type’ RM appears to be weak, and resources could more usefully be used in developing RM for commonly used methods such as flow cytometry. An RM for a particular biological activity would only make sense if a particular method was applied to a number of sufficiently similar CTP; but the diversity of mechanisms of action and indications under exploration suggests that this probably would not be the case. Whereas CRM are available for some common protein

1195

therapeutics (both follow-on and biosimilar) such that their potency can be compared, these products can be shown to have the same structure (though any glycosylation pattern will be less certain). Manipulated cells are subjected to a variety of differences in their manufacture, and, because manufacturing details are usually proprietary and undisclosed, two CTP based on the same starting cells should not be assumed to be similar. Without such information, even relative measurements to a cell RM would at best be unreliable. Acknowledgments We sincerely thank Karin Hoogendoorn, Julian Braybrook (LGC), Jurjen Velthuis (Kiadis Pharma), David Brindley (University of Oxford/CASMI) and Kim Bure (Sartorius Stedim) for their useful comments while preparing this manuscript. Disclosure of interests: The authors have no commercial, proprietary, or financial interest in the products or companies described in this article.

References 1. Allen RH, Sriram RD. The role of standards in innovation. Technol Forecast Social Change. 2000;64:171e81. 2. Griffiths B. Manufacturing paradigms: the role of standards in the past, the present and the future paradigm of sustainable manufacturing. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture October 2012 vol. 226 no. 10 1628e1634. 3. BSI. Cell therapy and regenerative medicine glossary. Regen Med. 2012;7:S1e124. 4. Dominici M, Le Blanc K, Mueller I, Slaper-Cortenbach I, Marini F, Krause D, et al. Minimal criteria for defining multipotent mesenchymal stromal cells: the International Society for Cellular Therapy position statement. Cytotherapy. 2006;8:315e7. 5. Mendicino M, Bailey AM, Wonnacott K, Puri RK, Bauer SR. MSC-based product characterization for clinical trials: an FDA perspective. Cell Stem Cell. 2014;14:141e5. 6. Judson LVH. Weights and measures standards of the United States: a brief history. Washington, DC: Department of Commerce, National Bureau of Standards: for sale by the Supt of Docs, US Government Printing Office; 1976. iii:36. 7. Barwick VJ, Pritchard E, editors. Eurachem Guide: Terminology in Analytical Measurement: Introduction to VIM 3. Available at: www.eurachem.org; 2011. 8. Ellison SLR, William A, editors. Eurachem/CITAC Guide: Quantifying Uncertainty in Analytical Measurement. 3rd edition. Available at: www.eurachem.org; 2012. 9. Tai SS, Welch MJ. Development and evaluation of a candidate reference method for the determination of total cortisol in human serum using isotope dilution liquid chromatography/ mass spectrometry and liquid chromatography/tandem mass spectrometry. Anal Chem. 2004;76:1008e14. 10. Zwart A, van Assendelft OW, Bull BS, England JM, Lewis SM, Zijlstra WG. Recommendations for reference method for

1196

11.

12.

13.

14.

15.

16.

17.

18.

C. A. Bravery & A. French

haemoglobinometry in human blood (ICSH standard 1995) and specifications for international haemiglobinocyanide standard (4th edition). J Clin Pathol. 1996;49:271e4. Kammel M, Kummrow A, Neukammer J. Reference measurement procedures for the accurate determination of cell concentrations: present status and future developments/ Referenzmessverfahren für die genaue Bestimmung von Zellkonzentrationen: Status und zukünftige Entwicklungen. Laboratoriums Medizin. 2012:25. Roussel M, Davis BH, Fest T, Wood BL. (ICSH) ICfSiH. Toward a reference method for leukocyte differential counts in blood: comparison of three flow cytometric candidate methods. Cytometry A. 2012;81:973e82. Hoffman RA, Wang L, Bigos M, Nolan JP. NIST/ISAC standardization study: variability in assignment of intensity values to fluorescence standard beads and in cross calibration of standard beads to hard dyed beads. Cytometry A. 2012;81:785e96. Hultin LE, Matud JL, Giorgi JV. Quantitation of CD38 activation antigen expression on CD8þ T cells in HIV-1 infection using CD4 expression on CD4þ T lymphocytes as a biological calibrator. Cytometry. 1998;33:123e32. Wang L, Abbasi F, Ornatsky O, Cole KD, Misakian M, Gaigalas AK, et al. Human CD4þ lymphocytes for antigen quantification: characterization using conventional flow cytometry and mass cytometry. Cytometry A. 2012;81:567e75. Davis BH, Wood B, Oldaker T, Barnett D. Validation of cellbased fluorescence assays: practice guidelines from the ICSH and ICCS, part I: rationale and aims. Cytometry B Clin Cytom. 2013;84:282e5. Davis BH, Dasgupta A, Kussick S, Han JY, Estrellado AGroup IIW. Validation of cell-based fluorescence assays: practice guidelines from the ICSH and ICCS, part II: preanalytical issues. Cytometry B Clin Cytom. 2013;84:286e90. Tanqri S, Vall H, Kaplan D, Hoffman B, Purvis N, Porwit A, et al. Validation of cell-based fluorescence assays: practice guidelines from the ICSH and ICCS, part III: analytical issues. Cytometry B Clin Cytom. 2013;84:291e308.

19. Barnett D, Louzao R, Gambell P, De J, Oldaker T, Hanson CA, et al. Validation of cell-based fluorescence assays: practice guidelines from the ICSH and ICCS, part IV: postanalytic considerations. Cytometry B Clin Cytom. 2013;84:309e14. 20. Wood B, Jevremovic D, Bene MC, Yan M, Jacobs P, Litwin V, et al. Validation of cell-based fluorescence assays: practice guidelines from the ICSH and ICCS, part V: assay performance criteria. Cytometry B Clin Cytom. 2013;84:315e23. 21. Herzenberg LA, Tung J, Moore WA, Parks DR. Interpreting flow cytometry data: a guide for the perplexed. Nat Immunol. 2006;7:681e5. 22. Baker SC, Bauer SR, Beyer RP, Brenton JD, Bromley B, Burrill J, et al. The External RNA Controls Consortium: a progress report. Nat Methods. 2005;2:731e4. 23. Jiang L, Schlesinger F, Davis CA, Zhang Y, Li R, Salit M, et al. Synthetic spike-in standards for RNA-seq experiments. Genome Res. 2011;21:1543e51. 24. Bravery CA, Carmen J, Fong T, Oprea W, Hoogendoorn KH, Woda J, et al. Potency assay development for cellular therapy products: an ISCT review of the requirements and experiences in the industry. Cytotherapy. 2013;15:9e19. 25. Viswanathan S, Keating A, Deans R, Hematti P, Prockop D, Stroncek DF, et al. Soliciting strategies for developing cell-based reference materials to advance MSC research and clinical translation. Stem Cells and Development. June 1, 2014,23(11):1157e67. 26. Krampera M, Galipeau J, Shi Y, Tarte K, Sensebe L(ISCT) MCotISfCT. Immunological characterization of multipotent mesenchymal stromal cells: the International Society for Cellular Therapy (ISCT) working proposal. Cytotherapy. 2013;15:1054e61. 27. Szymczak MM, Friedman RL, Uppoor R, Yacobi A. FDA/ Industry Perspectives: detection, Measurement, and Control in Manufacturing. Pharmaceutical Technology [Internet]. 2011; 35:70e4. Available at: http://www.pharmtech.com/pharmtech/ RegulatoryþCorner/FDAIndustry-Perspectives-DetectionMeasurement-and/ArticleStandard/Article/detail/742137.