Cytotherapy, 2013; 15: 9e19

Potency assay development for cellular therapy products: an ISCT*review of the requirements and experiences in the industry

CHRISTOPHER A. BRAVERY1,9, JESSICA CARMEN2,9, TIMOTHY FONG3,9,WANDA OPREA4, KARIN H. HOOGENDOORN5, JULIANA WODA6,SCOTT R. BURGER7,9, JON A. ROWLEY2,9, MARK L. BONYHADI8,9 &WOUTER VAN’T HOF6,9

1Consulting on Advanced Biologicals, Ltd, London, UK, 2Lonza, Walkersville, MD, USA, 3Progenitor Cell TherapyServices, Mountain View, CA, USA, 4TiGenix, Leuven, Belgium, 5Janssen Biologics, Leiden, the Netherlands,6Athersys, Cleveland, OH, USA, 7Advanced Cell & Gene Therapy, LLC, Chapel Hill, NC, USA, 8Life Technologies,Carlsbad, CA, USA, and 9Process and Product Development Subcommittee for the International Society for Cell Therapy

AbstractThe evaluation of potency plays a key role in defining the quality of cellular therapy products (CTPs). Potency can be definedas a quantitative measure of relevant biologic function based on the attributes that are linked to relevant biologic properties.To achieve an adequate assessment of CTP potency, appropriate in vitro or in vivo laboratory assays and properly controlledclinical data need to be created. The primary objective of a potency assay is to provide a mechanism by which themanufacturing process and the final product for batch release are scrutinized for quality, consistency and stability. A potencyassay also provides the basis for comparability assessment after process changes, such as scale-up, site transfer and newstarting materials (e.g., a new donor). Potency assays should be in place for early clinical development, and validated assaysare required for pivotal clinical trials. Potency is based on the individual characteristics of each individual CTP, and theadequacy of potency assays will be evaluated on a case-by-case basis by regulatory agencies. We provide an overview of theexpectations and challenges in development of potency assays specific for CTPs; several real-life experiences from thecellular therapy industry are presented as illustrations. The key observation and message is that aggressive early investment ina solid potency evaluation strategy can greatly enhance eventual CTP deployment because it can mitigate the risk of costlyproduct failure in late-stage development.

Key Words: assay development, cell characterization, cell therapy, potency, reference materials

Scope

In recent years, several cellular therapy products(CTPs) for regenerative or immune therapy applica-tions have advanced to pivotal clinical evaluation andmarket authorization (1). The successful deploymentof CTPs is hindered to a great extent by theircomplexity, which makes identification of relevantbiologic activities—and thus definition of consistentCTP quality—difficult. Although the regulatoryrequirement for potency assays is well understood byCTP developers, existing guidelines and articlesgenerally either discuss their development in isolationor mention them briefly in the context of character-ization. Feedback from International Society for CellTherapy (ISCT) members suggested the need to

*This review is the product of the ISCT Process and Product Development Subcthe Netherlands, May 18-21, 2011.Correspondence: Wouter Van’t Hof, Athersys, Inc., 3201 Carnegie Ave, Clevela

(Received 7 May 2012; accepted 9 August 2012)

ISSN 1465-3249 Copyright � 2013 published by Elsevier Inc. on behalf of Internhttp://dx.doi.org/10.1016/j.jcyt.2012.10.008

explain the available regulatory guidance and provideexamples of how others in the field have tackled theproblem. To this end, the ISCT Process and ProductDevelopment Subcommittee formed a potencyworking group at the 17th annual ISCT meeting inRotterdam, the Netherlands, in 2011 to address thiscommercial need. A resulting objective of this reviewis to provide a broader explanation of why potencyassessment is such a critical part of characterizationand the control of product quality and consistencynot only during development but also over the entireproduct life cycle.

If the intended action of the product and itsaccompanying regulatory status are overlooked,existing guidance and literature regarding potency

ommittee. It was conceived at the 17th ISCT Annual Meeting, Rotterdam,

nd, OH 44115-2634, USA. E-mail: wouter@athersys.com

ational Society for Cellular Therapy.

10 C. A. Bravery et al.

assays might be subject to certain misconceptions.For the purpose of this review, CTPs are autologous,allogeneic or xenogeneic cells for therapeutic, diag-nostic or preventive purposes in humans. Forreasons of brevity, this review excludes productsconsisting of cells combined with non-cellularcomponents (e.g., scaffolds, devices) and geneticallymodified cell products, for which additional regula-tions and characterization needs exist. Cell trans-plantation (i.e., homologous use of minimallymanipulated cells) and CTPs regulated primarilyas devices are also deliberately excluded becausetheir mechanism of action (MOA) is not intended tobe primarily medicinal. In cases where the intendedmechanism is primarily physical, determination of itsquality is primarily through physical characteristics.For example, for the living cell-based combinationproduct Apligraf (Organogenesis, Canton, MA,USA), which received pre-market approval from theU.S. Food and Drug Administration (FDA) in 1998(2), potency is defined through physical character-istics evaluated by histologic techniques (3). Like-wise, the quality of a cell transplant can moreplausibly be established by simple measurementssuch as viability because the cells are minimallymanipulated.

Because published potency strategies are rare, thefinal objective of this review is to provide examples(see later case studies and Supplementary material)that have been developed for a range of patient-specific and off-the-shelf CTPs. Although these may

Table I. Definition of key terminology for cell product characterization.

Characterization parameter

Physicochemicalcharacterization

Refers to the use of methods that measuPhysical: size, morphology, light-scatteChemical: identification of phenotypic m

Biologic characterization Refers to the use of methods that measuinfluence biologic systems). ExamplesBiologic: in vitro or in vivo measuremenmigration, tissue remodeling

Potencya Quantitative measure of relevant biologicbiologic propertiesb

Comparability testing Exercise to evaluate the impact of changeor clinical data relating to a CTP or its

Comparablec Conclusion that the product has highly schanges and that no adverse impact onoccurred

Biocompatibilityc Ability of a material to perform with an aStability testingc Determination of the shelf life under storStability Duration over which the quality of the pRelease assay Validated test method with pre-defined ac

to be released for clinical use

aAdapted from (6).bAs a measure of relevant biologic function, potency should be based ochemical measure may be used as a surrogate for potency at release orbiologic function.cAdapted from (21).

not all prove successful in the long run, theyemphasize the need for case-by-case development todeal with the differing issues and limitations.

Characterization

The first objective of CTP development should be todefine the active substance and the critical qualityattributes of the product so that these can becontrolled. To meet this objective, it is necessaryfrom the outset to characterize the product and itsmanufacturing process (including raw materials).Key terminology for cell characterization is providedin Table I. A more recent “publicly available speci-fication” (4) provides further information on char-acterization of cells including description of potentialmethodologies and their strengths and limitations.Pritchett and Little (5) also recently reviewedpotency requirements in the field.

The central principle in defining the quality ofa CTP (as with all biologic products) is that althoughphysicochemical parameters can identify the cellularactive substance, including surface markers that areinvolved in function, they cannot confirm that theproduct will actually be biologically active and potent(i.e., elicit the desired effect). Consequently, physi-cochemical measurements are primarily used foridentification and quantification of the activesubstance, intermediates, impurities and contami-nants. It is important to acknowledge this limitationof physicochemical testing to understand why broad

Definition

re physical and chemical characteristics. Examples for CTP:ring properties, tensile strength, cell number, confluencearkers and secreted substances, genotype, gene expression profile

re biologic function (i.e., how the physicochemical characteristicsfor CTP:ts of cytotoxicity, cell growth, de-differentiation, proliferation,

function of a CTP based on the attributes that are linked to relevant

s to a manufacturing process on the validity of quality, non-clinicalcomponents

imilar quality attributes before and after manufacturing processthe safety or efficacy, including immunogenicity, of the product

ppropriate host response in a specific applicationage and in use for the product and its intermediatesroduct is maintained within pre-defined parametersceptance criteria to which manufactured product needs to conform

n biologic characterization. Where this is not feasible, a physico-stability, as long as it can be correlated to a measure of relevant

Potency assay development for CTPs 11

biologic characterization should be employed toinvestigate biologic function.

In contrast, biologic characterization takes intoaccount the effect of the product on biologic systems,either modeled in vitro or in vivo in animals andultimately in the clinic. Although regulatory docu-mentation divides quality from non-clinical andclinical evaluations, they are part of the same processand should not be considered in isolation. Wherepossible, biologic characterization should also aim toinvestigate the biologic activity of intermediates andimpurities, and although such tests are not defined aspotency assays, their purpose is broadly similar. Inparticular, these biologic assays are essential toconfirm stability of intermediates such as cell banksas well as manufacturing hold steps and shipping.For impurities, biologic characterization normallyfocuses on in vivo toxicity testing, although it mayalso involve in vitro assays. Potency assays are teststhat have been demonstrated, during development,to confirm that the relevant biologic functions of theCTP (ideally correlated to efficacy) are present.

Why is potency so important?

The ultimate aim of potency assessment is to iden-tify the parameters that are critical to the efficacyof the product and to control them such thata product of consistent quality can be manufactured.

Figure 1. Central role of potency assessment in the determination of Cunderwritten by the hypothesis for MOA together with a description ofspecification and analysis of product comparability, stability and compa

Without product consistency, it is unrealistic toexpect consistent clinical effects.

The central role of potency in the various activi-ties that are required to establish product quality isdepicted in Figure 1. Confirmation of biologicactivity is needed to define the shelf life of theproduct (6) (viability may not be the most sensitiveor relevant stability-indicating parameter) and toconfirm compatibility of the product with othersubstances with which it comes in contact (e.g.delivery device, matrix material, primary container).More importantly, without measures of biologicactivity (including potency), product comparabilitycannot be established after process changes withoutnon-clinical or clinical testing (or both).

Development of CTPs is an expensive and time-consuming endeavor. Funding is a precious com-modity, and the needs and expectations of investorsmay frequently conflict with the CTP developers.Investors are not likely to view critical early devel-opment milestones such as the identification ofa suitable potency assay as important as the regula-tory approval of clinical trials or positive outcomes ofthe trials. From the investor perspective, this attitudemakes sense; however, driving early clinical devel-opment at the expense of manufacturing and qualitydevelopment brings significant risks. Identifying anddeveloping potency assays yield critical tools thatenable process changes in the future. Without these

TP quality. Potency is central to biologic characterization, which,the physicochemical properties, provides the platform for producttibility.

Table II. U.S. Food and Drug Administration core requirementsfor potency testing.

Potency testing of CTP must:� Indicate product-specific biologic activity� Measure identity and activity of active component� Provide test results for product release� Provide data to establish stability specifications� Meet labeling requirements� Comply with biologics regulations and good manufacturing

practice

Potency testing methods must:� Have pre-defined acceptance or rejection criteria

12 C. A. Bravery et al.

tools, even minor process changes could requireclinical qualification, which would be costly andpotentially commercially disastrous. Consequently,fast-to-phase III development programs that bypassdetailed biologic characterization in the race tothe clinic have a higher risk of failure or at best maybe saddled with manufacturing limitations thatcompromise commercial success. Early and sus-tained investment in a bona fide potency program isessential for maximizing a product’s commercialsuccess.

� Include appropriate reference materials, standards, controls� Be amenable to validation� Have established and documented accuracy, sensitivity, speci-

ficity and reproducibility� Provide quantitative data

Regulatory expectations for CTPs

The essential principle that active substances whosemolecular structure cannot be fully defined requirean evaluation of their potency before release onto themarket is enshrined in the pharmaceutical legislationof both the United States (7) and European Union(8,9). It is a legal requirement to evaluate the potencyof each batch of a licensed CTP. Although there arevarious references to the importance of potencyassays throughout the International Conference onHarmonisation, European Medicines Agency(EMA) and FDA guidelines, to date only the FDAhas compiled a comprehensive guidance documenton potency testing for CTPs and gene therapyproducts (10). The EMA, in contrast, has notyet written a general guideline for potency testingof CTPs, although the core multidisciplinary guide-line (11) has a short section that also refers toa guideline on potency testing for immunotherapyproducts (12).

The FDA and EMA recognize the challengesinherent in developing potency testing for CTPs andhave adopted a flexible, although still rigorous,regulatory approach. Specifically, regulatory agenciesacknowledge that potency evaluation is determinedby the individual product characteristics and that theadequacy of potency assays consequently need to beevaluated on a case-by-case basis. All CTP devel-opers have a responsibility to define and test the mostsuitable criteria and measures for potency assessmentof their individual products; in certain cases, this mayrequire establishing specific novel standards orprocedures not yet covered by regulatory guidance.These existing guidelines are reviewed next with thekey messages outlined. Core requirements forpotency testing as defined by the FDA are presentedin Table II.

Progressive implementation

CTP potency testing is expected to be developedover the course of pre-clinical and clinical develop-ment, becoming more sophisticated and defined

with increasing understanding of the product and itsfunctions. In the course of phase I and phase II trials,multiple characteristics and candidate assays shouldbe evaluated, guided by proposed MOA and datafrom proof-of-concept studies. Early in clinicaldevelopment, defined acceptance criteria, althoughdesirable, are not required for potency testing. Boththe FDA and the EMA expect potency testing withdefined acceptance criteria to be in place before thestart of pivotal clinical trials and potency testingvalidation to have been completed before submissionof a market authorization.

This flexible regulatory approach to potencytesting development must be matched by the CTPproduct developer. It may be necessary to adjustpotency test methods and specifications in the courseof clinical development to reflect manufacturing andclinical experience.

Regulatory core requirements

Although there is flexibility during development, theFDA specifies certain core requirements for CTPpotency testing (Table II). Potency testing mustdemonstrate the relevant biologic activity or activitiesof the product. It is not a requirement for potencytesting to reflect all of the product’s biologic func-tions, but it should indicate one or more relevantbiologic functions. Testing all biologic functions ofa product is rarely feasible, given the multimodalMOA of CTP. Testing a subset of the most relevantbiologic function or functions is a practical, scien-tifically meaningful strategy that meets regulatoryexpectations.

Potency testing results are a required element ofproduct release testing (8,9,13). Reflecting progres-sive implementation of potency testing, in earlyclinical development, results may be presented “forinformation only” without a specification. Various

Potency assay development for CTPs 13

analytic methods may be used, including in vitro or invivo functional (biologic) assays, non-functional(physicochemical) assays or a matrix of functionaland non-functional assays. Because the turnaroundtime of many functional assays may be problematicfor use in product release, both the FDA and theEMA accept non-functional (physicochemical)assays for release and stability testing, if appropriatelyqualified against functional assay results. Analyticmethods are expected to be quantitative, incorpo-rating suitable controls, reference materials andstandards (6). Although fully validated potencytesting is not required until the end of pivotal trials,eventual validation must be foreseen even at earlystages of development, and analytic methods that areamenable to validation should be chosen. Similarly,it is expected that accuracy, sensitivity, specificityand reproducibility will be established for the analyticmethods used in potency testing and that they aresuitably robust. Although quantitative analyticmethods are a core requirement for potency testing,this is not always possible for living biologic prod-ucts. Regulatory agencies recognize this and incertain cases may accept semi-quantitative assays forpotency testing, although demonstration of assayaccuracy and precision could be challenging with thisapproach. The following sections discuss thecommercial need for potency assays and how theexpectations of the regulators are aligned with theneeds of the developer.

Potency assay strategy for CTPs

To facilitate biologic characterization of the product,it is first necessary to develop a theory for the MOAand to outline how and why the product should

Table III. Pathway to developing a potency strategy for cell therapy pro

Preclinical research and development1. Understand the biologic basis of the disease indication your CTP a2. Define the realm of relevant biologic functions that your CTP poss3. Develop a hypothesis-driven, scientific rationale for the manner

indication of choice4. Create a matrix of assays that quantitatively measure the relevant bio

of the disease by your CTPProduct development and clinical assessment

5. Use the matrix of assays throughout product development to monia. The manufacturing process (e.g., culture vessels, cell propagatib. Donor-to-donor or batch-to-batch variability

6. Where feasible and appropriately powered, attempt to correlate cliniis a clear correlation, it will be straightforward to select the final “potscenario. A direct correlation between measured function (potencymid-stage development, and it is currently not a requirement

7. Set specifications based on all relevant manufacturing and clinicala. During early-stage clinical testing wide specifications can be

robustness and relevance of any one assayb. Specifications can be tightened after more process knowledge is

or batch-to-batch variability

work. As indicated in Figure 1 and Table III, thesetheories must be based on scientific data and prin-ciples. Source data may be derived from the scientificliterature, non-clinical studies in animals or otherbiological systems and possibly clinical data in caseswhere the data are already available (e.g., later stagesof development). Given the inherent complexity, itmay be tempting for CTP developers to argue thatbecause they cannot fully define the relevant MOA,they cannot develop a bona-fide potency assay.However, the key issue is not if any, but rather, whatlevel of knowledge is required to start defining thepotency of a CTP. If a CTP is being considered fortherapeutic use, the science should be matureenough to identify a therapeutic rationale, which bydefinition encompasses an assumed MOA. Withouta plausible therapeutic rationale, not only would it beunethical to administer cell-based products topatients, but also it would be highly unlikely that anyregulatory agency would approve such a clinical trialor that any investor would fund it.

The (hypothesized) mode of action is the foundation forpotency development

In the early stages of product development, there arelikely to be several theoretical MOA, and the MOA islikely to be multifactorial for most CTPs. Thesetheories all should be considered when embarking onbiologic characterization, and quantitative methodsshould be explored to measure them. Certain assayswould be unsuitable for defining specificationsbecause biologic assays are often complex in them-selves, and read-outs require days or weeks.However, assays that correlate with the intendedclinical effect or at least measure a characteristic for

ducts.

ims to treatesses (literature searches, screening experiments)in which your CTP will correct or modify the pathology in the

logic functions that align with the scientific rationale for treatment

tor if and how the biologic functions vary depending onon length, cell densities at seed and harvest, media composition)

cal outcome with one or more assays for biologic functions. If thereency assay” for pivotal studies and beyond. This is a low-probability) and clinical outcome is in reality rarely observed in early- or

dataused in light of the uncertainties associated with the (lack of)

acquired throughout clinical testing with respect to donor-to-donor

14 C. A. Bravery et al.

which there is clear evidence that it is part of theMOA are valid starting points. Successful biologiccharacterization not only would guide the identifi-cation of appropriate potency assays but also wouldprovide an arsenal of methods that can be employedin comparability studies (discussed later) to supportfuture process changes during development andpost-marketing. To ensure that the relevant needscontinue to be addressed, it would be necessary toevaluate the employed methods continually againstnew data relating to the MOA and adapt as neces-sary. It might even become prudent to change thepotency assay at a later stage if clear evidencesuggests the original theory was flawed or ifsubstantially better analytic methods become avail-able. In certain cases (e.g., when the MOA is hard todefine), discovery-driven methods such as micro-array, proteomics or other types of analysis may beable to identify potential potency biomarkers.However, in most cases, the best starting point is thedefinition of one or more theoretical MOA.

Developing a strategy

The process of potency assay development for CTPshould start with a “potency strategy,” based ona scientific rationale and basic research and takinginto account regulatory principles. This strategy,summarized in Table III, is typically built in stagesfrom preclinical research and development to clinicaltesting. For instance, if the product is thought toelicit a particular response in the recipient (e.g.,inhibit an inflammatory response), one or more testsshould be developed that specifically measure thatresponse. If the cells might suppress an inflammatoryresponse that might be mediated through anti-inflammatory cytokines, cell-cell contact or both,investigation into the exact mechanism may identifya potential potency assay. When part of the MOA isknown, biologic characterization should focus onthat, with the intention to develop a potency assayassociated with that characteristic. When the MOA isnot yet understood, a matrix of more generalmeasures of biologic activity likely needs to beemployed until the MOA is sufficiently understood.Regardless of the initial situation, potency determi-nation needs to be constantly reviewed and refined inthe light of new data throughout development.

Putative measures of potency should be identifiedduring pre-clinical evaluations when efficacy studiesare being evaluated in animal models. Typically,a general understanding of MOA is elucidatedduring this phase of development when data areemerging to suggest that the cell product can elicita curative effect in an animal. Identifying potentialmeasures of potency early in clinical development

should evolve from pre-clinical studies (14). Forproduction of material for early clinical trials,potential assays should ideally be tracked with a tar-geted acceptance range. However, as indicatedpreviously, in cases where it is not yet possible todefine a preliminary specification, it may be justifiedto collect data for information only. For later phase(pivotal) clinical trials in which efficacy is beingevaluated, it is important to demonstrate manu-facturing consistency, which can be ensured onlythrough validated assays with defined acceptancelimits. Without this assurance in place, doubt wouldremain as to whether all trial subjects were exposedto the product with a consistent potency that is stableover the shelf life, leading to difficulties in assessingthe overall risk versus benefit.

Practical considerations

Potency—worth the investment?

First, the objective is to manufacture a consistentproduct that elicits a consistent clinical response.Without product consistency, there is little chance ofobserving consistent clinical effects or successfullydemonstrating efficacy. As discussed, showing consis-tencyofphysicochemical characteristics is importantbutshould be secondary to consistency of biologic function.

After meaningful measures of biologic functionhave been established, process development canfocus on maximizing the desired biologic character-istics of the product and minimizing the undesiredeffects, allowing processes that result in a higherpotency per unit of content (e.g., number of cells) tobe identified. Additionally, because the relationshipbetween cell number and potency is often obscure orunknown, potency measures may allow the rela-tionship between cellular content and potency to bemore clearly understood and the dose (typicallydefined in units of potency for biologic products) tobe more clearly defined such that the minimumeffective dose can be established. Identifying theminimum effective dose of the product makescommercial sense because the cost per dose is relatedto the number of cells required. This approach is alsoaligned with current agency expectations (11,15). Inconclusion, the characterization required to explorethe relevant biologic functions has greater utility thanjust the identification of potency assays in that itprovides the basis for development and evolution ofthe product that enables commercial exploitation ofthe CTP.

Comparability—addressing inevitable change

The second, and perhaps equally important, role ofbiologic characterization is to provide relevant tools

Potency assay development for CTPs 15

that allow comparability to be established followingprocess changes, which inevitably occur during bothdevelopment and post-market authorization. When-ever changes are made to the manufacturing process,potency determination forms an integral part of theconfirmation that the relevant biologic characteristicsare retained. This confirmation needs to be supportedby additional biologic assays (in vitro and wherenecessary in vivo). This support is particularlyimportantwhere constraints of time andmaterial (e.g.,for autologous products or other products witha short shelf life) limit the tools available for potencydetermination to “simple” or surrogate physico-chemical parameters. Even in the absence of thesekinds of constraints, potency evaluations are neededto address potentially routine changes, such assuppliers or sources of critical raw materials, withoutnecessitating clinical validation before they can beimplemented. Temporary cessation of supply, whiledata are collected and regulatory approval is sought,could seriously damage the commercial reputation ofthe developer and result in significant commerciallosses.

During development, significant changes maybecome clinically qualified by the next clinical studyphase. However, following market authorization,implementing process changes with an insufficientcharacterization tool set would necessitate confirma-tory non-clinical or clinical data (or both), addingsignificant cost and time to any process changes or sitemoves. In contrast to small molecules and some otherbiologics, conventional pharmacokinetics and phar-macodynamics cannot be measured, and so in theabsence of a suitable biomarker, clinical qualificationmay require a lengthy clinical follow-up, depending onthe indication and nature of the product.

Potency is central to confirming product stability

Although viability may appear to be the obviouscharacteristic indicating stability, viable cells maylose biologic function during storage. Whether lossof biologic activity precedes cell death by minutes,hours or days cannot be known unless a reliablemeasure of biologic function is included in stabilitystudies. Even in cases where other characteristicsmay be more sensitive stability indicators, it is stillrelevant to include potency in all stability programsbecause ultimately no product should be releasedthat does not have the stated potency for the fullduration of the shelf life; this is also a fundamentalregulatory principle (13). Until a potency assay isdefined and qualified, product shelf life cannotbe confirmed, and there is a risk that a productused during development might not be potent forthe whole assumed shelf life or may be more

sensitive to storage conditions than other parame-ters suggest.

The stability of a CTP is usually a key driverwhen considering product formulation (e.g., freshversus frozen). This is a key decision that needs to beaddressed early in development because it hasa significant impact on the approach to commer-cialization. If viability is used early on to determinethat the product has a sufficient shelf life to becommercially viable, it might be disastrous to findlater that the shelf life is significantly shorter becausepotency is lost over time in viable cells. Such a reve-lation might postpone the development programmonths or years (if trials need repeating) while theproduct is re-formulated and work is undertaken todemonstrate the altered product is comparable.

Reference materials—confirming relative potency

The need for reference materials is a fundamentalconcept in thequality control of allmedicinal products,yet this poses significant and specific challenges forCTPs. The normal approach is to allocate a represen-tative batch of product, use this as themaster referencematerial (RM) and qualify working RM that can beused for batch release, comparability and other needs.In situationswhere no international reference standardexists (as with CTPs), RM is usually used to assignrelative potency to the potency assay, ensuring themeasurement of potency is normalized over manybatches. Likewise, RM can be important for otherspecifications and comparability studies.

With CTPs, preparing RM can bring significantchallenges because manufacture of large quantities ofproductmay be impossible even ifmultiple batches arecombined. However, any measurement of potency isseriously compromised if no RM is available toconfirm performance of the assay and to allow relativepotency to a known batch to be confirmed. In theabsence of the ability to prepare a product RM, otherapproaches to develop suitable RM need to beconsidered; this is a very relevant consideration butgoes beyond the immediate scope of this review.

The difficulties of developing RM should not bedismissed lightly; potency assays are often complexwith wide variations, and there is a risk of assay driftand, perhaps more importantly, process drift.Trending alone is unlikely to be successful becausemany analytic methods for CTPs have broad rangesmeaning it could be months or years before evensignificant trends can be identified, if at all. Consid-ering that CTPmanufacturing processes often involvemultiple manual steps and that bioassays can also besensitive to operator technique, it is essential tocontrol these potential sources of variability. If a batchof product fails release testing, without RM it may be

16 C. A. Bravery et al.

difficult to identify whether the root cause of theproblem is within the process or the assay.

No potency assay is complete until suitable RMhas been developed to ensure the results from batchto batch can be reliably compared. RM is centralto confirming product consistency in general andis vital to confirm the consistency of potencydetermination.

Considerations for the assay itself

It is important to calculate an overall cost impactfor a particular assay or matrix of assays to releasea batch of product with a final metric of “cost perproduct unit.” The cost per product unit is affectedmost dramatically by batch size and how many unitsare required for testing to release the batch for use.Choosing the number of product units to be testedis not trivial, and many factors are likely tocontribute to the decision. Testing most often isperformed on the final product after packaging. Insome instances, in-process testing may also berequired before fill or finish. Release testing ofa product is meant to ensure that the batch asa whole meets specifications as well to confirm thatthe first and last product units manufactured are ofequivalent quality.

These calculations should not deter developersfrom employing a matrix of potency testing duringdevelopment because, as discussed at lengththroughout this review, identifying assays that definepotency may be difficult. By using a matrix of assays,the assays that provide the most useful measure ofpotency can be selected and, where necessary,developed further into a potency test appropriate forcommercial needs and satisfactory to regulators. Thedata collected from the matrix of assays employedover development provide a basis as well as historicdata for comparability. The value of these data can behard to judge in the short-term, but if not collected atthe time, the data are lost forever.

Another important commercial consideration forpotency assays are the reagents used. Some materialsused in analytic methods, especially for potencyassessment, may be as critical to product quality asthe raw materials used in the manufacture process.Consequently, their supply continuity and qualityshould be ensured in the same way. Key reagents suchas indicator cell lines andmonoclonal antibodies canbehard to replace and perhaps impossible without RM.

Reported potency assays in the field and casestudies

The previous sections focused on practical anddevelopmental issues concerning potency assays,their development and use. This section illustrates

the realities, pitfalls and solutions for the variousobstacles in this iterative process through five real-lifecase studies. These examples span various classes ofcell products, including autologous, allogeneic,adherent and non-adherent products from differenttissue sources, and highlight a diversity of assays andsurrogates that were identified to support potencyevaluation for each individual CTP.

Various CTP potency assays have been reported inthe literature and public meetings, and some exam-ples from the field are summarized in Table IV. Fivecase studies are described; detailed descriptions,illustrated with relevant data sets, are presented asSupplemental material to this review, available online.

Case study 1

Case study 1 concerns development of a potency assayfor ChondroCelect (Tigenix, Leuven, Belgium), acommercial CTP for autologous chondrocyteimplantation in cartilage repair. This autologouschondrocyte product is indicated for the treatment ofcartilage defects. In vivo cartilage formation wascorrelated with marker gene expression and in vivochondrogenesis, supporting the validation ofa potency assay based on specific molecular markerexpression in a quantitative reverse transcriptasepolymerase chain reaction-based assay.

Case study 2

Case study 2 concerns development of a potencyassay for Xcellerated T Cells (Xcyte Therapies,Seattle, WA, USA), an autologous activated T celltherapy product for treating chronic lymphocyticleukemia (CLL). The proposedMOA for XcelleratedT Cells on leukemic B cells in CLL was that T celldefects were corrected via stimulation using magneticbeads coated with anti-CD3 and anti-CD28 anti-body, leading to restored T cell functionality andsubsequent TNF receptor superfamily, member 6(FAS)-mediated killing of leukemic B cells aftercontact with infused CD154-expressing T cells.Response of Xcellerated T Cells to re-stimulation invitro with plate-bound anti-CD3 and anti-CD28monoclonal antibodies (mimicking in vivo reac-tivation) was shown to be tightly correlated to restoredfunctionality. A cell-based potency assay was devel-oped using flow cytometry for CD154 expression.This plate-bound re-stimulation assay was validatedand implemented for quality control batch releasepotency testing of Xcellerated T Cells.

Case study 3

Case study 3 concerns development of a potency assayfor a regulatory T cell (Treg) product Athelos

Tab

leIV

.Examples

ofrepo

rted

potenc

yassays

inthecelltherap

yindu

stry.

Platform

Indication

orstatus

Poten

cyassay

Referen

ces

Autolog

ous

Cho

ndroCelect(T

iGen

ix).Autolog

ous

chon

droc

ytes

EMA

approv

ed,kn

eecartila

gerepa

irExp

ressionof

molecular

markers

(PCR)

Sup

plem

entalinform

ation

PROVENGE,sipu

leuc

el-T

(Den

dreo

n).Primed

,expa

nded

Tcells

FDA

approv

ed,treatm

entof

horm

onerefractory

prostate

canc

erCD54

expression

(FACS)

(16)

Xcelle

ratedT

Cells(X

cyte

The

rapies).Activated

Tcells

Pha

seI/II,treatm

entof

chroniclymph

ocytic

leuk

emia

CD15

4expression

(FACS)

Sup

plem

entalinform

ation

Amorcyte,AMR-001

(Neo

Stem).Bon

emarrow-derived

cells

Pha

seI,cardiacrepa

iraftermyo

cardialinfarct

Invitromigration

ofCD34

þ/C

XCR4þ

cells

inSDF-

1grad

ient

Sup

plem

entalinform

ation

Athelos,na

turalregu

latory

Tcells

(Neo

Stem).

Exvivo

expa

nded

naturalTregs

Pha

seI,treatm

entof

graft-versus-hostdisease

Sup

pression

ofCD69

andCD15

4in

Tcell

activation

assay(F

ACS)

Sup

plem

entalinform

ation

Allo

gene

icProch

ymal

(Osiris).Adu

ltstromal

stem

cells

Pha

seIII,treatm

entof

graft-versus-hostdisease

Secretion

ofTNFR1(E

LISA)

(17,18

)M

ultistem

(Athersys).B

onemarrowstromal

stem

cells

Pha

seI,treatm

entof

isch

emic

cardiacdisease

Secretion

ofVEGF,IL

-8an

dCXCL5(E

LISA)

(19)

GRN-001

(Geron

).HESde

rivedoligod

endroc

yte

prog

enitor

cells

Pha

seI,spinal

cord

repa

irSecretion

ofne

urotroph

icfactor

pane

l(E

LISA)

(20)

ELISA,en

zyme-lin

kedim

mun

osorbe

ntassay;

HES,hu

man

embryo

nicstem

cell;

PCR,po

lymerasech

ainreaction

;SDF-1,stromal

cell-de

rivedfactor

1;TNFR1,

tumor

necrosisfactor

type

1receptor;VEGF,vascular

endo

thelialgrow

thfactor.

Potency assay development for CTPs 17

(NeoStem, NewYork, NY, USA) for the treatment ofgraft-versus-host disease (GVHD) and otherimmune-based diseases. Athelos is in the pre-clinicalstage for development of a Treg product for thetreatment of GVHD, autoimmune diseases and solidorgan transplant tolerance. The development ofaTreg potency assay for final product characterizationand release will be based on a mechanism involvingcell-cell contact to use the ability of Tregs to inhibitanti-CD3/anti-CD8 activated conventional T cell(Tcon) proliferation in vitro. Suppression of T cellactivation markers CD69 and CD154 will be deter-mined by flow cytometry as a measure for Tregpotency.

Case study 4

Case study 4 concerns development of a potencyassay for Amorcyte, AMR-001 (NeoStem), anautologous adult bone marrow-derived CTP forcardiac repair after myocardial infarction. AMR-001is being evaluated for the treatment of damagedheart muscle after acute myocardial infarction(AMI). The potency assay for AMR-001 wasdeveloped based on the putative MOA of infusedCD34þCXCR4þ cells that home to damaged tissuevia a gradient of stromal-derived factor 1 (SDF-1)to facilitate tissue repair and vascular regeneration.During phase I studies, several cell characteristics ofAMR-01 were evaluated to correlate cell productrelease criteria to clinical activity. The relativemobility in an SDF-1 gradient of CD34þ cells inthe AMR-001 product was the only parameter thatcorrelated to clinical benefit, and an in vitro migra-tion assay of CD34þCXCR4þ cells in an SDF-1gradient was developed as the potency assay forAMR-001.

Case study 5

Case study 5 concerns development of a potencyassay for MultiStem (Athersys, Cleveland, OH,USA), an adult allogeneic bone marrow-derivedstromal CTP, in treatment of myocardial infarction.MultiStem is an adult adherent stem cell product.Pre-clinical models of ischemia demonstrated thatMultiStem induces angiogenesis in vivo and thatincreased vessel formation correlates with efficacyand treatment. An in vitro angiogenesis assayestablished vascular endothelial growth factor(VEGF), chemokine ligand 5 (CXCL5) and inter-leukin (IL)-8 secreted by MultiStem product to berequired for the induction of angiogenesis. Anecessary threshold of angiogenic factor expressionwas established using the in vitro angiogenesis assay,and detection of these factors by enzyme-linked

18 C. A. Bravery et al.

immunosorbent assay (ELISA) was defined as thesurrogate potency assay (19).

Conclusions

The first hurdle in potency determination is tounderstand the MOA of the CTP, and because thisrequires biologic characterization, such studiesshould by their nature identify putative assays.Biologic characterization encompasses not only invitro assays developed as part of manufacturingquality but also non-clinical and clinical databecause the purpose of these assays is to under-stand the relevant (and undesired) biologic effectsof the product. In a well-conceived developmentprogram, potential potency assays should bea natural by-product, albeit requiring some adap-tation into test methods suitable for productrelease. Although the preferred potency assay forproduct release would provide a quantitative bio-logic measure of relevant function, such assays arenot always suitable for release purposes because ofproduct-specific issues such as batch size and timelimitations, a point that is acknowledged by regu-lators. It is reasonable where such arguments arejustified to use simpler physicochemical methods assurrogates for potency but only where they can becorrelated to a true measure of relevant biologicfunction.

More broadly, this review puts the measurementof relevant biologic function and potency intocontext within the broader topic of product devel-opment. The primary purpose of potency determi-nation is to confirm that the relevant biologicactivity of the product is present to the expectedlevel at release. Likewise, it is essential to confirmthat the relevant biologic activity is retainedthroughout the claimed shelf life because viabilityalone does not achieve this. The second andperhaps equally important role for potency deter-mination is to provide the tools for comparabilityassessment after process changes (e.g., processimprovements, new sites, changes of scale).Although potency alone is unlikely to be sufficientin many cases, the additional biologic assaysexplored during characterization provide an arsenalof methods to confirm that the essential character-istics of the product have been retained. Finally,potency assays are critical to confirming otheraspects of stability (e.g., in-use stability) andcompatibility of the product with devices (whereapplicable). The return on early, systematic andprolonged investment in a bona-fide potencyprogram for CTPs would mitigate the risk of costlyproduct failure in late-stage development andenhance the likelihood of commercial success.

Disclosure of interest: WO is a paid employeeand stock option holder of TiGenix (Leuven,Belgium). JW and WvH are paid employees andstock option holders of Athersys (Cleveland, OH,USA). MB is a paid employee and stock optionholder of Life Technologies (Carlsbad, CA, USA).JC and JR are employees of Lonza (Walkersville,MD, USA) and TF is employed by PCT (Moun-tain View, CA, USA), both of which companiesprovide contract services for certain cell therapydevelopers mentioned in this review. The otherauthors have no commercial, proprietary or finan-cial interest in the products or companies describedin this article.

References

1. Maciulaitis R, D’Apote L, Buchanan A, Pioppo L,Schneider CK. Clinical development of advanced therapymedicinal products in Europe: evidence that regulators mustbe proactive. Mol Ther. 2012;20:479e82.

2. Premarket Approval (PMA). Apligraf (Graftskin). Available at:http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMA/pma.cfm?id¼13855. Accessed August 8, 2012.

3. USP. Graftskin Monograph, http://www.pharmacopeia.cn/v29240/usp29nf24s0_m35846.html. Accessed August 8, 2012.

4. PAS 93: 2011. BSI. Characterization of human cells for clinicalapplications—guide. Available at: http://feedback.bsigroup.com/Home/Details/850. Accessed August 8, 2012.

5. Pritchett T, Little L. “Hard cell” potency testing for cellulartherapy products. BioProcess Int. 2012;10:36e48.

6. International Conference on Harmonisation. Guidance onspecifications: test procedures and acceptance criteria forbiotechnological/biological products (ICH Q6B). 64 FR 44928.Available at: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm073488.pdf. Accessed August 18, 1999.

7. 21CFR610.10. Available at: http://www.gpo.gov/fdsys/pkg/CFR-2011-title21-vol7/pdf/CFR-2011-title21-vol7-sec610-10.pdf. Accessed August 8, 2012.

8. Commission Directive 2003/63/EC. Amending Directive2001/83/EC of the European Parliament and of the Councilon the Community code relating to medicinal products forhuman use. Annex I, Part I, section 3.2.2.1. Available at:http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri¼OJ:L:2003:159:0046:0094:en:PDF. Accessed June 25, 2003.

9. Commission Directive 2009/120/EC. Amending Directive2001/83/EC of the European Parliament and of the Councilon the Community code relating to medicinal products forhuman use as regards advanced therapy medicinal products.Annex I, Part IV, section 3.3.2.3. Available at: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri¼OJ:L:2009:242:0003:0012:EN:PDF. Accessed September 14, 2009.

10. Guidance for Industry. Potency tests for cellular andgene therapyproducts. Available at: http://www.fda.gov/downloads/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/CellularandGeneTherapy/UCM243392.pdf. Acc-essed January 2011.

11. Guideline on human cell based medicinal products. Availableat: http://www.emea.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003894.pdf. AccessedSeptember 2008.

12. Guideline on potency testing of cell based immuno-therapy medicinal products for the treatment of cancer.

Potency assay development for CTPs 19

Available at: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003814.pdf.Accessed May 2008.

13. International Conference on Harmonisation. Final guidelineson stability testing of biotechnological/biological products(ICH Q5C). 61 FR 36466. Available at: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm073466.pdf. Accessed July 10, 1996.

14. Carmen J, Burger SR, McCaman M, Rowley JA. Developingassays to address identity, potency, purity and safety: cellcharacterization in cell therapy process development. RegenMed. 2012;7:85e100.

15. Guidance for FDAReviewers and Sponsors.Content and reviewof chemistry,manufacturing, and control (CMC) information forhuman somatic cell therapy investigational new drug applications(INDs). Available at: http://www.fda.gov/downloads/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/Xenotransplantation/ucm092705.pdf. AccessedApril 2008.

16. Higano CS, Small EJ, Schellhammer P, Yasothan U,Gubernick S, Kirkpatrick P, et al. Sipuleucel-T. Nat RevDrug Discov. 2010;9:513e4.

17. Danilkovitch A. Potency assay development for a noveltherapy product: prochymal adult mesenchymal stem cells.

Presented at Cellular, Tissue and Gene Therapies Advi-sory Committee Meeting; Gaithersburg, MD; February 9,2006.

18. Kebriaei P, IsolaL,Bahceci E,HollandK,Rowley S,McGuirk J,et al. Adult human mesenchymal stem cells added to cortico-steroid therapy for the treatment of acute graft-versus-hostdisease. Biol Blood Marrow Transplant. 2009;15:804e11.

19. Lehman N, Cutrone R, Raber A, Perry R, Van’t Hof W,Deans R, et al. Development of a surrogate angiogenicpotency assay for clinical-grade stem cell production. Cyto-therapy. 2012;14:994e1004.

20. Zhang YW, Denham J, Thies RS. Oligodendrocyteprogenitor cells derived from human embryonic stem cellsexpress neurotrophic factors. Stem Cells Dev. 2006;15:943e52.

21. PAS 84:2012. Cell therapy and regenerative medicine—glos-sary. Available at: http://shop.bsigroup.com/Browse-by-Sector/Healthcare/PAS-84/. Accessed August 8, 2012.

Supplementary Data

Supplementary data related to this article can befound at http://dx.doi.org/10.1016/j.jcyt.2012.10.008

19.e1 C. A. Bravery et al.

Case studies for potency assay development forcellular therapy products

In this supplemental section, we present the five casestudieswithmoredetail, including certain relevant datasets. They highlight different hypotheses and ap-proaches used for the development of potency assaysfor different classes of CTPs at different stages ofproduct development. Products include autologouschondrocytes for cartilage repair, T cell products fortreatment of leukemia or immune disorders, andstromal or non-adherent bone marrow-derived cellproducts for treatment of AMI. Each case studydescribes adifferent path fromMOAhypothesis towardthe identificationof potencymarkers that enable simpleand accurate quantitation in vitro. The resultingpotency assays include quantitative polymerase chainreaction, fluorescent activated cell sorter (FACS),ELISA and cell migration assays. These case studiesdirectlyunderscore the concept that potency evaluationshould be based on the specific characteristics ofeach individual CTP and that the assay adequacy needsto be developed and evaluated on a case-by-case basis.The five specific case examples are as follows:

� Case study 1: development of a potency assayfor ChondroCelect (Tigenix, Leuven, Belgium),a commercial CTP for autologous chondrocyteimplantation in cartilage repair

� Case study 2: development of a potency assayfor Xcellerated T Cells (Xcyte Therapies, Seat-tle, WA, USA), an autologous activated T-celltherapy product for treating CLL

� Case study 3: development of a potency assay fora Treg cell product for the treatment of GVHDand other immune-based diseases (Athelos,a NeoStem Company, New York, NY, USA)

� Case study 4: development of a potency assay forAMR-001 (Amorcyte, a NeoStem Company,New York, NY, USA), an autologous adult bonemarrow-derived CTP for cardiac repair aftermyocardial infarction

� Case study 5: development of a potency assay forMultiStem (Athersys, Cleveland, OH, USA), anadult allogeneic bone marrow-derived stromalCTP, in treatment of myocardial infarction

Case study 1: development of a potency assayfor ChondroCelect, a commercial CTP forautologous chondrocyte implantation incartilage repair

Product specifics

ChondroCelect received final marketing authoriza-tion as an Advanced Therapy Medicinal Product bythe EMA in October 2009. It is indicated for the

treatment of symptomatic, isolated, full-thicknesscartilage defects of the femoral condyle and deliveredas a sterile suspension of ex vivo expanded, autologouschondrocytes in excipient medium. ChondroCelect issurgically applied in a two-step procedure known asautologous chondrocyte implantation. In the firststep, a cartilage biopsy sample is obtained arthro-scopically from a minor weight-bearing region of thejoint to provide the starting material for the cellexpansion process. In the second step, the isolated andexpanded chondrocytes are implanted in the defectivecartilage in a mini-arthrotomy.

Strategy

The use of isolated and expanded chondrocytes inautologous chondrocyte implantation has beendocumented for many years, with original clinicalinvestigations performed by Brittberg et al. (1). Amajor challenge in the field is obtaining a sufficientnumber of high-quality cells at the end of the ex vivoculture process that retain the capacity to formstable, structured cartilage in vivo. It has been widelyreported that culture of isolated chondrocytes resultsin progressive loss of the chondrogenic phenotypeand consequently loss of good cartilage repair abili-ties (2,3). It was known at the outset of developmentthat limiting cell expansion and assessing the potencyof the final, expanded cellular product would becrucial. In developing a potency test for Chon-droCelect, a series of biologic tests relevant forassessing cartilage quality were investigated. Subse-quently, the correlations between in vivo cartilage-forming capacity and surrogate in vitro and molecularassays were determined, where the latter could serveas the routine potency measure.

Potency assay for ChondroCelect

The first step in developing a potency test forChondroCelect was to establish an assay to measurecartilage-forming capacity in an in vivo environmentbut using a simpler model than surgical orthotopicautologous chondrocyte implantation in a largeanimal (Supplementary Figure 1). To this end, theectopic cartilage formation assay (ECFA) wasdeveloped. In this model, cells are injected intra-muscularly, and cartilage formation is evaluatedbased on a histologic score. Using the ECFA,different populations of cultured chondrocytes couldbe functionally distinguished, ranging from chon-drocytes that have retained the ability to form stable,hyaline-like cartilage tissue to chondrocytes that havelost all cartilage-forming capacity through de-differ-entiation during culture. These populations werescreened for molecular marker expression usinga micro-array approach to identify genes that

Supplementary Figure 1. Multi-assay analysis of ChondroCelect. Multiple cell populations, ranging from phenotypically stable chon-drocytes to increasingly de-differentiated cells with reduced cartilage-forming capacity were evaluated in vivo, in vitro and at a molecularlevel. (A) In a goat model of autologous chondrocyte implantation, stable cartilage formation was observed using phenotypically stablechondrocytes, whereas the quality of the repair tissue was reduced using de-differentiated chondrocytes. No significant repair could beobserved in the negative control dermal fibroblast group. (B) The same cell populations were evaluated in an ECFA in the nude mouse. Inthis model, phenotypically stable chondrocytes formed tissue displaying all the hallmarks of hyaline cartilage, whereas de-differentiatedchondrocytes formed a less mature and more fibrous tissue. Only fibrous tissue was evident in the negative control of dermal fibroblasts. (C)De-differentiation status was shown to correlate with gene expression profile based on specific molecular markers. Based on correlations withthe in vivo models, this assay was validated as a surrogate potency assay for routine product testing. PCR, polymerase chain reaction.

Potency assay development for CTPs 19.e2

correlated positively or negatively with cells able toproduce stable hyaline cartilage in the ECFA model.Using this approach, a ChondroCelect potency assaywas established based on molecular marker expres-sion. The relevance of this surrogate assay wasconfirmed through correlations with the ECFA,which was correlated with the large animal efficacymodel. Additional work showed that the surrogatemolecular assay also correlated well with other rele-vant in vitro models of chondrocyte functionality,such as the well-described three-dimensional chon-drogenic pellet assay (4). The surrogate potencyassay was developed and validated for use in routineproduction, whereas the additional in vivo and invitro assays validated during development providea more complete picture of product biologic activityand can be used as important tools to supportprocess and assay changes.

In conclusion, based on the combination of theaforementioned studies, an overall picture of cellularpotency could be constructed, where in vivo cartilage

formation was correlated with marker gene expres-sion and in vitro chondrogenesis. Taken together,these studies supported the validation of a potencyassay based on specific molecular marker expression.This quantitative reverse transcriptase polymerasechain reaction-based assay is both practically feasiblein a routine manufacturing setting and relevantbecause it correlates with an extended set of assaysdemonstrating chondrogenic functionality.

Note: ChondroCelect is a commercial medicinalproduct; some aspects of the potency assay areproprietary and not disclosed.

Case study 2: development of a potency assayfor Xcellerated T Cells, an autologous activatedT cell therapy product for treating CLL

Product specifics

Xcellerated T Cells were developed by Xcyte Ther-apies (now defunct) for the treatment of multipleindications, including CLL. A hallmark of CLL is

19.e3 C. A. Bravery et al.

the accumulation of leukemic B cells concomitantwith immunologic abnormalities in the T-cell com-partment (5,6). The Xcellerate treatment hypothesiswas that potent activation of T cells in the patientwould reverse T-cell functional defects, allowingnormal T-cell orchestration of the immune response,which would reverse abnormal accumulation ofleukemic B cells through restoration of normalhomeostatic mechanisms. T cells from the patientswere activated using Dynabeads CD3/CD28 CTSimmunomagnetic beads (Life Technologies, GrandIsland, NY, USA) bearing covalently attached anti-bodies against the T-cell surface receptors CD3 andCD28. Activated cells were expanded in culture andshowed restored expression of key markers andcytokines along with restoration of response to allo-geneic stimulation (mixed lymphocyte reactionresponse). Co-culture of Xcellerated T Cells withleukemic B cells resulted in the up-regulation of keyreceptors on leukemic B cells, including CD54,CD86 and CD95 (FAS), rendering the leukemic Bcells sensitive to FAS-ligand induced cell death (5).FAS induction was shown to depend on engagementwith CD154 (CD40L) expressed on Xcellerated TCells (7), similar to earlier observations by Wierdaet al. (8) following co-culture of CD40L-expressinggene-modified B cells with leukemic CLL B cells.In a phase I/II clinical trial, expanded T cells demon-strated normalized T-cell receptor repertoire, andCLL patients infused with Xcellerated T Cellsdemonstrated sustained decreases in lymphadenop-athy (12 of 17 patients) and splenomegaly (11 of 13patients) in conjunction with increased levels ofcirculating T cells.

Strategy

The proposed MOA for Xcellerated T Cells onleukemic B cells in the setting of CLL was that T-celldefects were corrected via stimulation using Dyna-beads CD3/CD28 CTS, leading to restored T-cellfunctionality and subsequent FAS-mediated killingof leukemic B cells after contact with infusedCD154-expressing T cells. Response of XcelleratedT Cells to re-stimulation in vitro with plate-boundanti-CD3 and anti-CD28 monoclonal antibodies(mimicking in vivo reactivation) was shown to betightly correlated with restored functionality. A cell-based potency assay was developed using thisapproach. Both flow cytometry-based and ELISA-based assays were used to assess a panel of markersassociated with normal functionality, including rapidup-regulation of surface markers (CD25, CD54,CD137, CD154), rapid secretion of cytokines(granulocyte-macrophage colony-stimulating factor,interferon-a, tumor necrosis factor-a) and rapid

down-modulation of CD62L. The assay was set upto compare the response of Xcellerated T Cells frompatients with CLL with their non-manipulated T-cellcounterparts, with the end goal of creating a simple,reproducible, low-variance assay that would measurea single marker as the correlate for restored T-cellfunctionality.

Potency assay for Xcellerated T Cells

To enable routine use of the potency assay in qualitycontrol lot release and stability testing, developmentfocused on establishment of a standardized methodwith a robust, consistent read-out. Assay parametersthat were investigated and optimized included theconcentration of the plate-bound anti-CD3 and anti-CD28 antibodies for re-stimulation, the time ofexposure to re-stimulation and the incorporation ofpositive and negative controls to assess system suit-ability. The induction of cytokine secretion andexpression of various key surface markers all corre-lated well with normalization of T-cell function(Supplementary Table I). CD154 expression wasselected as the ultimate potency assay markerbecause CD154 expression was believed to bedirectly correlated with the proposed therapeuticMOA. The assay was validated for analysis of CD154(CD40L) induction because the sensitivity, rapidity,reproducibility and low variance were optimal forCD154.

In conclusion, when analyzed in the potencyassay, final product Xcellerated T Cells exhibit fasterkinetics and greater levels of up-regulation of CD154than the autologous peripheral blood T cells used asthe starting cells for the Xcellerate Process (Supple-mentary Figure 2). Xcellerated T Cells producedfrom patients with B-cell CLL demonstrate a robustand reliable response in the potency assay, witha mean fold increase in expression of CD154 of21.0 � 4.3 for CD4þ T cells and 11.9 � 3.7 forCD8þ T cells (n ¼ 26). The plate-bound re-stimu-lation assay was validated and implemented forquality control lot release potency testing of Xcell-erated T Cells.

Case study 3: development of a potency assayfor a regulatory T cell product for the treatmentof GVHD and other immune-based diseases

Product specifics

Athelos, which is 80% owned by NeoStem, is devel-oping a Treg product for the treatment of GVHD,autoimmune diseases, and solid organ transplanttolerance (9e11). The Athelos Treg program is inthe pre-clinical and phase I development stages.Natural Tregs (nTregs) are thymus-derived T cells

Supplementary Table I. Evaluation of potency marker for Xcellerated T Cells.

Potency correlate marker

Assay kinetics(time requiredfor maximumresponse)

Parameters

Assay read-out(fold increase/decreasein marker expression)

Inter-samplevariance

Bestmarker

Secreted growthfactors (ELISA orLuminex assay[Life Technologies])IL-2 [ >16 h >200 Medium þIFN-g [ 18e21 h >24 High

TNF-a [ >18 h >15 HighGM-CSF [ 21e25 h >200 Medium þ

Cell surface markers (FACS)CD25 [ >16 h High Low þþCD54 [ >16 h High Medium þCD62L Y 4e6 h High Medium þþCD154 [ 4e6 h High Low DDD

Allo-response restorationCD25 [ 48e72 h Low High

Comparing normalized fold increase or decrease in marker expression between Xcellerated T Cells and non-Xcellerated T Cells using ananti-CD3/anti-CD28 coated plate re-stimulation format or allo-response (mixed lymphocyte reaction) assay. Rapid assay kinetics, maximumfold change in marker expression and low variance were considered desirable read-out attributes (reflected by bold in table).GM-CSF, granulocyte-macrophage colony-stimulating factor; IFN, interferon; TNF, tumor necrosis factor.

Potency assay development for CTPs 19.e4

that have been characterized as primarily CD4þ

CD25hiCD127loFoxP3þ cells (12e14). InducedTregs (iTregs) are CD4þ or CD8þ cells that can beinduced in vitro with transforming growth factor-b,IL-10 (Tr1) or other factors to suppress activatedconventional T cells (Tcons) (15,16). One of theAthelos Treg cell therapies will use an in vitroexpanded nTreg product. nTregs have been shown tosuppress activated CD4þ and CD8þ effector cells andprevent or treat GVHD or autoimmune disease inseveral animal models (9). Several phase I trials areunderway at the present time or have been completedat academic institutions where the safety of Tregs has

Supplementary Figure 2. CD154 expression on CD4þ cells before anconcentrations were evaluated for optimized marker induction, as were kthis example, induction of CD154 (CD40L) on CD4þ T cells was mea

been shown, and efficacy in preventing GVHD hasbeen suggested (17).

Strategy

Studies suggest that nTreg suppression of Tconsmay be mediated through several different MOA, butcell-cell contact appears to be an important mecha-nism (18). In vitro, this activity can be demonstratedby the inhibition of in vitro polyclonally activatedTcons when co-cultured with Tregs. The develop-ment of a Treg potency assay for final productcharacterization and release was based on this

d after expansion and re-stimulation. Various antibody coatinginetics of induction to design final re-stimulation assay format. Insured by FACS.

19.e5 C. A. Bravery et al.

mechanism to use the ability of Tregs to inhibitanti-CD3/anti-CD8 activated Tcon proliferation invitro. Earlier versions of the in vitro assay to deter-mine Treg suppression of Tcon proliferation usedeither 3H-thymidine or carboxy-20,70-dichloro-fluorescein diacetate, succinimidyl ester (CSFE)intracellular dye to monitor cell expansion. Em-ploying the recently developed commercial FACS-based assay kit developed by Becton DickinsonBiosciences (Franklin Lakes, NJ, USA), suppressionof T cell activation markers CD69 and CD154 willbe used to assess Treg potency (19). The FACS kithas been performed in parallel with assays usingCSFE and has shown consistent comparability indetermining Treg activity (20). The advantages ofusing this kit are same-day or next-day results (vs. upto 5 days using 3H-thymidine or CSFE) and reducedtechnical hurdles in preparing and performing theassay. Both of these enhancements result in anoverall reduction in assay variability.

Potency assay

Several components of the FACS-based system arecritical to the reproducibility and robustness of thisassay. Two particularly important factors are thesource and quality of the Tcons and the type ofstimulation used to activate them. For the treatmentof GVHD or solid organ transplant tolerance inwhich the nTreg product is donor-specific andpatient-specific, freshly isolated Tcons may be usedto determine potency of the final product. However,in the scenario of developing an allogeneic “off-the-shelf” product, the donor-to-donor variability ofusing fresh Tcons may make lot-to-lot comparisonsdifficult. Tregs can suppress third-party Tcons, andthe generation of a bank of frozen peripheral bloodmononuclear cells from a single or defined numberof donors is an approach that would reduce inter-assay variability for both patient-specific and “off-the-shelf” products. For the Athelos potency assay,the strategy is to create a qualified bank of theperipheral blood mononuclear cells with the bestsignal-to-noise ratio for use in subsequent potencyassays. Assay development during the pre-clinicalresearch phase evaluated several methods to activateTcons, including soluble anti-CD3 antibody, plateor bead bound anti-CD3 with soluble anti-CD28and beads bound with both antibodies. Titrationstudies evaluating several bead-to-Tcon ratiosdemonstrated that commercial anti-CD3/anti-CD28beads (Dynabeads CD3/CD28 CTS; Life Tech-nologies) gave the most consistent activation ofTcons. Development of the Athelos assay will seekthe optimal level of Tcon activation that is stillsuppressible by Tregs.

In conclusion, an nTreg in vitro potency assaybased on the cell contact-dependent mechanism ofsuppression of Tcon activation and proliferation hasbeen developed. For both patient-specific and alloge-neic “off-the-shelf” Treg products, an nTreg potencyassay using a commercial flow cytometry assay will beused. Third-party peripheral blood mononuclear cellswill be screened and banked as a source of Tconresponders to reduce inter-assay variability. The assaydescribedhere is currently indevelopment and is beingimplemented for research use only.

Case study 4: development of a potency assayfor AMR-001, an autologous adult bonemarrow-derived CTP for cardiac repair aftermyocardial infarction

Product specifics

AMR-001 is an adult bone marrow-derived cellproduct manufactured by PCT (Allendale, NJ, USA)for Amorcyte, a NeoStem Company. AMR-001 isbeing evaluated for the treatment of damaged heartmuscle after AMI. Over the last few years, severalclinical trials evaluated the infusion of stem cells forAMI and showed promising results (21,22). AMR-001 has completed a phase I clinical trial that showeda dose-related significant improvement in perfusion(23). Patients received doses of 5, 10 or 15 millionCD34þ cells. Patients receiving�10million cells (n¼9) showed significant improvement in resting perfu-sion rates at 6 months compared with patientsreceiving 5 million cells (n ¼ 6) and control (n ¼ 15),as measured by the single photon emission computedtomography (SPECT) total severity score (�256vs.þ13; P¼ 0.01). The data also showed that patientsreceiving �10 million cells experienced a trendtoward improvement in ejection fraction (þ4.5%vs. þ0.69%), end-systolic volume (�5.7 mL vs. þ3.5mL) and infarct size and tissue death secondary to lossof adequate blood supply (�7.4% vs. �5.3%) at 6-month follow-up. No study-related significantadverse events were reported.

Strategy

The potency assay for AMR-001 was developed basedon the putative MOA of AMR-001 that the infusedCD34þCXCR4þ cells home to damaged tissue viaa gradient of stromal cell-derived factor-1 (SDF-1)where they facilitate tissue repair and vascular regen-eration (24e29). After AMI, damaged and dyingmyocardiocytes under ischemic stress and hypoxia inthe peri-infarct zone produce hypoxia-induciblefactor 1 (HIF-1) and vascular endothelial growthfactor (VEGF), two pro-angiogenic factors. HIF-1also induces the production and release of SDF-1,

Potency assay development for CTPs 19.e6

a ligand for CXCR4. SDF-1 produced in thedamaged region attracts mobilized CD34þCXCR4þ

stem cells to the infarct site to initiate repair ofdamaged tissue. AMR-001 is infused via the infarct-related artery 5e11 days after ST segment elevationmyocardial infarction—the optimal time frame forcellular intervention, after the pro-inflammatory “hotphase” and before permanent scar formation. Theinfused CD34þCXCR4þ cells home to the at-risktissue via the SDF-1 gradient, inducing neoangio-genesis and resulting in functional benefit to thepatient.

Development of the potency assay is based on theputative MOA of AMR-001, in which greatercardiac repair is correlated with the ability of stemcells to migrate to the peri-infarct zone. Duringphase I studies, several cell characteristics ofAMR-01 were evaluated to correlate criteria forcell product release to clinical activity. Parametersevaluated included number of CD34þ cells,CD34þCXCR4þ cells and CD34þVEGFR2þ cells,colony-forming unit-granulocyte macrophage form-ing ability in vitro and relative activity in an SDF-1migration assay. As shown in SupplementaryFigure 3, the relative mobility in an SDF-1 gradientof CD34þ cells in the AMR-001 product was theonly parameter that correlated to clinical benefit asmeasured by the resting hypoperfusion score(Resting Total Severity Score) and reduction in theinfarct size as a percentage of the left ventriculararea.

Supplementary Figure 3. Correlation between CD34þ mobilityand clinical benefit of AMR-001. The correlation is shownbetween potency assay assessing CD34þ cell mobility in vitro andclinical observations after AMR-001 therapy in Resting TotalSeverity Score (RTSS) (A) and reduction of left ventricular infarctarea (B). (From Quyyumi AA, et al. CD34(þ) cell infusion afterST elevation myocardial infarction is associated with improvedperfusion and is dose dependent. Am Heart J. 2011;161:98e105,with permission from A. Pecora.)

Potency assay

The in vitro migration assay of CD34þCXCR4þ cellsin an SDF-1 gradient was developed based on theassay as originally described by Jo et al. (30).Conversion of the research assay into an assay suit-able for product release and quality assurancerequired the evaluation and standardization ofreagents including qualification of the culture mediaand recombinant SDF-1. Additional factors exam-ined included optimization of the kinetics and timeneeded for sufficient CD34þ cell migration andstandardization of enumerating migrating cells. Toreduce variability in enumerating migrating cells inthe assay by microscopic visualization and colonycounting, the AMR-001 assay incorporates flowcytometry analysis of viable CD34þ cells to deter-mine the number of migrating cells.

In conclusion, development of the potencyassay for AMR-001, a bone marrow-derivedCD34þCXCR4þ cell product for the treatment ofAMI, is described. The use of the SDF-1 migrationassay is based on pre-clinical data supporting thehypothesis that repair of ischemic myocardium

requires migration of infused stem cells to the site ofinjury. The ability of the cell product to migrate inthe SDF-1 mobility assay was correlated to clinicalbenefit. Although migratory ability of the stem cellsmay not be directly involved with the actual mecha-nism of repair of cardiac tissue, the data suggest thatthe number of stem cells homing to the site of injuryis an important factor in the overall efficacy of thetherapy. During a phase I study, cell products wereassessed for numerous characteristics, includingphenotype (CD34þ and CXCR4þ or VEGFR2þ),granulocyte macrophage colony formation and SDF-1 mobility. Potency of the AMR-001 productcorrelated to the quality of the cell product based onthe mobility assay.

19.e7 C. A. Bravery et al.

Case study 5: development of a potency assayfor MultiStem, an adult allogeneic bonemarrow-derived stromal CTP, in treatment ofmyocardial infarction

Product specifics

MultiStem is an adult adherent stem cell productmanufactured by Athersys. MultiStem is currentlyunder clinical evaluation in phase I and II studies asadjunct therapy for treatment of ischemic andinflammatory processes that occur after AMI,hematopoietic stem cell transplantation and strokeand during ulcerative colitis. MultiStem has shownbeneficial effects in a broad array of pre-clinicaldisease models, which lends support to a model ofmultimodal therapeutic activity of the cell productthat, depending on the disease pathology, mayconsist of modulation of inflammation and cellularimmune activity, cytoprotection and support ofangiogenesis (31e34). Clinical results in patientswith AMI receiving MultiStem demonstrate a strongsafety and benefit profile (35). The development ofa potency assay is presented in which production ofa panel of cytokines required for the specific angio-genec potential of MultiStem is measured. Thisinformation is presented to highlight the rationaleand potential strategy for stromal cell therapypotency assessment. This assay is not as yet a regu-latory approved procedure.

Supplementary Figure 4. Mechanismof the angiogenic potential ofMultiStem. VEGF, CXCL5 and IL-8 were individually immune-depleted from MultiStem conditioned media. These factors weresubsequently added back in a dose-response curve to identify theminimum level needed to restore the angiogenic response. Endo-thelial growth factor medium (EGM) is a positive control, whereasendothelial basal medium alone (EBM) is a negative control.

Strategy

Preclinical models of ischemia demonstrated thatMultiStem induces angiogenesis in vivo, and theincrease in vessel formation induced by MultiStemtreatment correlates with efficacy and treatmentbenefit (31,33,36). A well-established angiogenesisassay, human umbilical vein endothelial cell tubeformation assay (37), was used to demonstrate thatMultiStem can induce angiogenesis in vitro and invivo. Multiple pro-angiogenic factors were shown tobe secreted by MultiStem, including VEGF, CXCL5and IL-8. Using the human umbilical vein endothelialcell tube formation assay, immunodepletion studiesdemonstrated that these three angiogenic factors,VEGF, CXCL5 and IL-8, are necessary for Multi-Stem-induced angiogenesis, with significant reduc-tion in tube formation in the absence of any of thesefactors. Adding back increasing amounts of thesecytokines into depleted conditioned media estab-lished the lower limits of each of these cytokinesrequired to induce angiogenesis (SupplementaryFigure 4). Mesenchymal stromal cells, producedaccording to classic culture procedures, lackedsignificant production of CXCL5 and IL-8, andconditioned media from mesenchymal stromal cell

cultures did not provide stable tube formation activityin vitro. Detection of VEGF, CXCL5 and IL-8 byELISA was deemed to provide an adequate set ofsurrogate potency markers with specificity for Multi-Stem activity in ischemic injury indications.

Potency assay

The lower limit of each of these factors required toinduce angiogenesis was set as pass/fail criteria forthe cells. ELISAs for VEGF, CXCL5 and IL-8 arecommercially available and standardized for accu-racy, precision and specificity by the manufacturer.To validate these assays further for potency testingand establish a positive reference for use in future lottesting, qualification of the assay using two operatorsand on two separate days was performed. These testsvalidated the assay as accurate and reproducible bycorrelating the amount of these factors in theconditioned media to levels secreted during clinicalmanufacturing runs. The levels of these factorsdetected across manufacturing runs are consistentand demonstrate that the cells produced significantlymore of these factors than the minimum required toinduce angiogenesis. Clinical manufacturing lotshave shown consistent production of this cytokine

Potency assay development for CTPs 19.e8

panel at levels that range between 10-fold and 50-fold above required in vitro threshold levels.

In conclusion, an in vitro angiogenesis assay hasbeen developed based on the level of cytokines,VEGF, CXCL5 and IL-8, secreted by MultiStemand demonstrated to be required for the induction ofangiogenesis. A necessary threshold of angiogenicfactor expression was established using the in vitroangiogenesis assay. By correlating the levels of thesecytokines required to induce tube formation in vitrowith levels of these factors found in the spent mediafrom MultiStem manufacturing production runs,detection of these factors was defined as a surrogatepotency assay with defined pass/fail criteria. Furtherassay development to qualify and validate this assayfully for regulatory approval is currently underway.

References

1. Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson O,Peterson L. Treatment of deep cartilage defects in the kneewith autologous chondrocyte transplantation. N Engl J Med.1994;331:889e95.

2. Dell’Accio F, De Bari C, Luyten FP. Molecular markerspredictive of the capacity of expanded human articularchondrocytes to form stable cartilage in vivo. Arthritis Rheum.2001;44:1608e19.

3. Dell’Accio F, Vanlauwe J, Bellemans J, Neys J, De Bari C,LuytenFP.Expandedphenotypically stable chondrocytes persistin the repair tissue and contribute to cartilage matrix formationand structural integration in a goat model of autologous chon-drocyte implantation. J Orthop Res. 2003;21:123e31.

4. Johnstone B, Hering TM, Caplan AI, Goldberg VM, Yoo JU.In vitro chondrogenesis of bone-marrow derived mesen-chymal progenitor cells. Exp Cell Res. 1998;238:265e72.

5. Bartik MM, Welker D, Kay NE. Impairments in immune cellfunction in B cell chronic lymphocytic leukemia. SeminOncol. 1998;25:27.

6. Cantwell M, Hua T, Pappas J, Kipps TJ. Acquired CD40-ligand deficiency in chronic lymphocytic leukemia. Nat Med.1997;3:984.

7. Bonyhadi M, Frohlich M, Rasmussen A, Ferrand C,Grosmaire L, Robinet E, et al. In vitro engagement of CD3and CD28 corrects T cell defects in chronic lymphocyticleukemia. J Immunol. 2005:2365e75.

8. Wierda WG, Cantwell MJ, Woods SJ, Rassenti LZ,Prussak CE, Kipps TJ. CD40-ligand (CD154) gene therapyfor chronic lymphocytic leukemia. Blood. 2000;96:2917.

9. TangQ, Bluestone JA,Kang S-M.CD4þFoxp3þ regulatoryTcell therapy in transplantation. J Mol Cell Biol. 2012;4:11e21.

10. Leslie M. Regulatory T cells get their chance to shine.Science. 2011;332:1020e1.

11. Ukena SN, Velaga S, Geffers R, Grosse J, Baron U,Buchholz S, et al. Human regulatory T cells in allogeneic stemcell transplantation. Blood. 2011;118:e82e92.

12. Hartigan-O’Connor DJ, Poon C, Sinclair E, McCune JM.HumanCD4þ regulatoryT cells express lower levels of the IL-7receptor alpha chain (CD127), allowing consistent identificationand sorting of live cells. J Immunol Methods. 2007;319:41e52.

13. Seddiki N, Santner-Nanan B, Martinson J, Zaunders J,Sasson S, Landay A, et al. Expression of interleukin (IL)-2and IL-7 receptors discriminates between human regulatoryand activated T cells. J Exp Med. 2006;203:1693e700.

14. Liu W, Putnam AL, Xu-Yu Z, Szot GL, Lee MR, Zhu S, et al.CD127 expression inversely correlates with FoxP3 andsuppressive function of human CD4þ T reg cells. J Exp Med.2006;203:1701e11.

15. Huter EN, Stummvoll GH, DiPaolo RJ, Glass DD,Shevach EM. Cutting edge: antigen-specific TGFb-inducedregulatory T cells suppress Th17-mediated autoimmunedisease. J Immunol. 2008;181:8209e13.

16. Cao O, Dobrzynski E, Wang L, Nayak S, Mingle B,Terhorst C, et al. Induction and role of regulatory CD4þCD25þ T cells in tolerance to the transgene productfollowing hepatic in vivo gene transfer. Blood. 2007;110:1132e40.

17. Di Ianni M, Del Papa B, Cecchini D, Bonifacio E, Moretti L,Zei T, et al. Immunomagnetic isolation of CD4þCD25þFoxP3þ natural T regulatory lymphocytes for clinical appli-cations. Clin Exp Immunol. 2009;156:246e53.

18. Bluestone JA, Tang Q. How do CD4þCD25þ regulatory Tcells control autoimmunity? Curr Opin Immunol. 2005;17:638e42.

19. Ruitenberg JJ, Boyce C, Hingorani R, Putnam A,Ghanekar SA. Rapid assessment of in vitro expanded humanregulatory T cell function. J Immunol Methods. 2011;372:95e106.

20. Canavan JB, Afzali B, Scottà C, Fazekasova H, Edozie FC,Macdonald TT, et al. A rapid diagnostic test for humanregulatory T-cell function to enable regulatory T-cell therapy.Blood. 2012;119:e57e66.

21. BrittenMB, Abolmaali ND, Assmus B, LehmannR, Honold J,Schmitt J, et al. Infarct remodeling after intracoronaryprogenitor cell treatment in patients with acute myocardialinfarction (TOPCARE-AMI). Circulation. 2003;108:2212e8.

22. Martin-Rendon E, Brunskill SJ, Hyde CJ, Stanworth SJ,Mathur A, Watt SM. Autologous bone marrow stem cells totreat acute myocardial infarction: a systematic review. EurHeart J. 2008;29:1807e18.

23. Quyyumi AA, Waller EK, Murrow J, Esteves F, Galt J,Oshinski J, et al. CD34(þ) cell infusion after ST elevationmyocardial infarction is associated with improved perfusionand is dose dependent. Am Heart J. 2011;161:98e105.

24. Geisler T, Fekecs L, Wurster T, Chiribiri A, Schuster A,Nagel E, et al. Association of platelet-SDF-1 with hemody-namic function and infarct size using cardiac MR in patientswith AMI. Eur J Radiol. 2011.

25. Seeger FH, Rasper T, Koyanagi M, Fox H, Zeiher AM,Dimmeler S. CXCR4 expression determines functionalactivity of bone marrow-derived mononuclear cells for thera-peutic neovascularization in acute ischemia. ArteriosclerThromb Vasc Biol. 2009;29:1802e9.

26. Seeger FH, Tonn T, Krzossok N, Zeiher AM, Dimmeler S.Cell isolation procedures matter: a comparison of differentisolation protocols of bone marrow mononuclear cells used forcell therapy in patients with acute myocardial infarction. EurHeart J. 2007;28:766e72.

27. Zhang G, Nakamura Y, Wang X, Hu Q, Suggs LJ, Zhang J.Controlled release of stromal cell-derived factor-1 alpha insitu increases c-kitþ cell homing to the infarcted heart. TissueEng. 2007;13:2063e71.

28. Wojakowski W, Tendera M, Zebzda A, Michalowska A,Majka M, Kucia M, et al. Mobilization of CD34(þ),CD117(þ), CXCR4(þ), c-met(þ) stem cells is correlatedwith left ventricular ejection fraction and plasma NT-proBNPlevels in patients with acute myocardial infarction. Eur HeartJ. 2006;27:283e9.

29. Britten MB, Abolmaali ND, Assmus B, Lehmann R,Honold J, Schmitt J, et al. Infarct remodeling after intra-coronary progenitor cell treatment in patients with acute

19.e9 C. A. Bravery et al.

myocardial infarction (TOPCARE-AMI): mechanisticinsights from serial contrast-enhanced magnetic resonanceimaging. Circulation. 2003;108:2212e8.

30. Jo DY, Rafii S, Hamada T, Moore MA. Chemotaxis ofprimitive hematopoietic cells in response to stromal cell-derived factor-1. J Clin Invest. 2000;105:101e11.

31. Van’t HofW,Mal N,Huang Y, ZhangM, Popovic Z, Forudi F,et al. Direct delivery of syngeneic and allogeneic large-scale expanded multipotent adult progenitor cells improvescardiac function after myocardial infarct. Cytotherapy. 2007;9:477e87.

32. Kovacsovics-Bankowski M, Streeter PR, Mauch KA,Frey MR, Raber A, van’t Hof W, et al. Clinical scale expandedadult pluripotent stem cells prevent graft-versus-host disease.Cell Immunol. 2009;255:55e60.

33. Kovacsovics-Bankowski M, Mauch K, Raber A, Streeter PR,Deans RJ, Maziarz RT, et al. Pre-clinical safety testing

supporting clinical use of allogeneic multipotent adultprogenitor cells. Cytotherapy. 2008;10:730e42.

34. Medicetty S, Wiktor D, Lehman N, Raber A, Popovic ZB,Deans R, et al. Percutaneous adventitial delivery of allogeneicbone marrow derived stem cells via infarct related arteryimproves long-term ventricular function in acute myocardialinfarction [published online ahead of print October 14, 2011].Cell Transplant.

35. Penn MS, Ellis S, Gandhi S, Greenbaum A, Hodes Z,Mendelsohn FO, et al. Adventitial delivery of an allogeneicbone marrow-derived adherent stem cell in acute myocardialinfarction: phase I clinical study. Circ Res. 2012;110:304e11.

36. Aranguren XL, McCue JD, Hendrickx B, Zhu XH, Du F,Chen E, et al. Multipotent adult progenitor cells sustain func-tion of ischemic limbs inmice. JClin Invest. 2008;118:505e14.

37. Taraboletti G, Giavazzi R. Modelling approaches for angio-genesis. Eur J Cancer. 2004;40:881e9.