Preface

An Expert Meeting on The Current Scientific andLegal Status of Alternative Methods to the LD50Test for Botulinum Neurotoxin (BoNT) PotencyTesting was organised by the Centre forDocumentation and Evaluation of Alternatives toAnimal Experiments (ZEBET) at the FederalInstitute for Risk Assessment (BfR), and tookplace on 27–28 April 2009, in Berlin, Germany.The goal of ZEBET is to promote primarily thereplacement of legally required animal experi-ments with valid alternative test methods, toreduce the numbers of test animals used to theabsolutely necessary level, and to alleviate theirpain and suffering. ZEBET’s tasks are the docu-mentation, assessment and promotion of alterna-tive methods to animal experiments on a nationaland international level, and the initiation ofresearch and development, including the fundingof smaller research projects.

The Expert Meeting on alternative methods forBotulinum neurotoxin (BoNT) potency testing washeld at the BfR as the result of a decree from theGerman Federal Ministry of Food, Agriculture andConsumer Protection (BMELV), after it wasinformed by German state authorities about anincrease in animal numbers being used in theLD50 potency test in mice for the purposes ofBoNT potency testing. Ethical concerns have beenraised in both Europe and the United States aboutthe animal suffering involved in this type of test-ing, especially in the context of BoNT’s aestheticapplications. There had been increased publicpressure on the German Government, due to suc-cessful campaigning by animal welfare/anti-vivi-sectionist organisations on the issue of BoNTanimal testing. As a result, the BMELV askedZEBET to assess the status of the different alter-native methods to the BoNT LD50 potency testthat are cited in the European Pharmacopoeia(EP) 6.0, Monograph 2113, Botulinum Toxin Type

The Current Scientific and Legal Status of AlternativeMethods to the LD50 Test for Botulinum NeurotoxinPotency Testing

The Report and Recommendations of a ZEBET Expert Meeting

Sarah Adler,1 Gerd Bicker,2 Hans Bigalke,3 Christopher Bishop,4 Jörg Blümel,5 Dirk Dressler,6 JoanFitzgerald,7 Frank Gessler,8 Heide Heuschen,9 Birgit Kegel,10 Andreas Luch,1,11 Catherine Milne,12

Andrew Pickett,13 Heidemarie Ratsch,14 Irmela Ruhdel,15 Dorothea Sesardic,16 Martin Stephens,17

Gerhard Stiens,18 Peter D. Thornton,19 René Thürmer,9 Martin Vey,5 Horst Spielmann,1,20 BarbaraGrune1 and Manfred Liebsch1

1Centre for Documentation and Evaluation of Alternatives to Animal Experiments (ZEBET), Federal Institutefor Risk Assessment (BfR), Berlin, Germany; 2University of Veterinary Medicine Hannover, Division of CellBiology, Hannover, Germany; 3Institute of Toxicology, Medical School of Hannover, Hannover, Germany;4Wickham Laboratories, Wickham, Hampshire, UK; 5Merz Pharmaceuticals GmbH, Frankfurt (Main),Germany; 6Department of Neurology, Rostock University, Germany; 7Allergan, Inc., Irvine, CA, USA;8Institute for Applied Biotechnology in the Tropics at the University of Göttingen, Göttingen, Germany;9Federal Institute for Drugs and Medical Devices (BfArM), Bonn, Germany; 10Paul-Ehrlich-Institut, FederalInstitute for Vaccines and Biomedicines, Langen, Germany; 11Department of Product Safety, Federal Institutefor Risk Assessment, Berlin, Germany; 12European Directorate of the Quality of Medicines and Health Care(EDQM), Strasbourg, France; 13Ipsen Biopharm Limited, Wrexham, UK; 14Landesamt für Gesundheit undSoziales Berlin (LAGeSo), Berlin, Germany; 15Akademie für Tierschutz, Neubiberg, Germany; 16NationalInstitute for Biological Standards and Control (NIBSC), Health Protection Agency (HPA), Potters Bar,Hertfordshire, UK; 17The Humane Society of the United States (HSUS), Washington, USA; 18Federal Ministryof Food, Agriculture and Consumer Protection (BMELV), Bonn, Germany; 19Home Office, ASPI, London, UK;20Free University Berlin, Berlin, Germany

ATLA 38, 315–330, 2010 315

Address for correspondence: Sarah Adler, Centre for Documentation and Evaluation of Alternatives to AnimalExperiments (ZEBET), Federal Institute for Risk Assessment (BfR), Berlin, Germany.E-mail: Sarah.Adler@bfr.bund.de

A for Injection (1), and the most promisingapproaches for their validation. Experts fromindustry, regulatory authorities, German min-istries, academia, research, national and interna-tional validation centres, and animal welfareorganisations, were invited to actively participatein the meeting.

The objective of the Expert Meeting was toreview available alternative methods for BoNTpotency testing, and to formulate recommenda-tions for making progress toward implementingthe Three Rs, i.e. Refinement, Reduction, andReplacement, in BoNT potency testing. In addition,ways in which communication on BoNT issuesbetween manufacturers, researchers and regula-tors could be encouraged, and how improvementsin regulatory harmonisation between differentcountries and continents could be achieved, werediscussed. The meeting started with presentationsby the individual participants, giving an overviewon the regulatory and scientific status of alterna-tive methods to the LD50 test for BoNT potencytesting. Afterwards, the participants were dividedinto two separate break-out groups. Break-outGroup 1 discussed the regulatory requirements forBoNT potency testing and the validation andimplementation of alternative methods. Break-outGroup 2 discussed the developed and availablealternative methods and their suitability for reduc-ing, refining or replacing the LD50 potency test.

Introduction

Botulinum neurotoxin is a biological product pro-duced by the anaerobic, Gram-positive bacteriumClostridium botulinum. There are seven major andserologically distinct serotypes (A–G) of BoNT,characterised by a complex mode of action, whichresults in blockage of acetylcholine release atneuro-muscular nerve endings, leading to flaccidparalysis. This mechanism of toxicity comprisesthe four stages of binding, internalisation, translo-cation and proteolytic cleavage of a substrate,which leads to inhibition of synaptic exocytotoxictransmitter release. The different serotypes ofBoNT act on different substrates.

BoNT-A and BoNT-B are available as licensedpharmaceuticals for the treatment of a variety ofmedical disorders, such as cervical dystonia, ble-pharospasm, spastic conditions and hyperhidrosis.However, BoNT is also used in so-called ‘aestheticmedicine’ to temporarily treat facial asymmetriesor reduce facial lines. BoNT is of biological origin,and has an extremely high toxic pharmacologicalactivity. Therefore, the safe and effective dosing ofBoNT products requires accurate and reliablemeasurement of potency before batch release ontothe market. The reference test for potency testingin the batch release of therapeutic BoNT, as well

as for the detection of the toxin in clinical cases ofbotulism, is the mouse bioassay. In diagnostic test-ing, the serotype needs to be identified and theamount of toxin quantified. The detection ofClostridium botulinum in clinical samples, withsubsequent identification and typing of the toxicculture supernatant by mouse bioassay, is oftenincluded in routine testing. This permits the iden-tification of toxico-infections and delivers isolatesfor further testing and toxin-typing, either in fur-ther animal experiments, or by molecular methods.

In the manufacture of pharmaceutical BoNTproducts, the serotype and subtype of the toxin arewell defined. The focus is on the biological activityof BoNT, which needs to be determined as accu-rately as possible, for the purposes of both safetyand efficacy, to ensure that the correct potency ofthe product is supplied to clinicians for patienttreatment. The current requirements of the EP forpotency determination are based on standard prac-tice, as approved in the marketing authorisationsof products for the EU market, which stipulatesthat every production lot of BoNT must be testedin an LD50 potency test in mice. For a valid assay,the dilution range used should encompass theLD50 value. In practice, this can mean that at thehighest dose, 90% of the animals are killed, and atthe lowest dosage, 90% of the animals survive.Importantly, the precision of the test depends ondilution intervals and the number of animals usedper concentration. The product dosage for patientsis determined by the outcome of the potency assay.

In the case of BoNT LD50 potency testing, thedosing of animals is associated with severe suffer-ing. Death is generally secondary to respiratoryfailure due to paralysis of the respiratory muscles.Consequently, the introduction of alternativemethods is urgently required. A number of alter-native tests have been developed in the past fewdecades, to replace the mouse bioassay. Add -itionally, promising methods are under develop-ment, which may, either alone or in combinationwith other assays, meet the rigid requirements ofpotency testing. To encourage movement awayfrom the LD50 test, the possibility of using a vali-dated in-house alternative to the LD50 assay hasbeen included in the EP monograph (1).

The Toxic Mechanism of BoNT

BoNT induces muscle paralysis by blocking therelease of acetylcholine at neuro-muscular nerveendings (2). BoNTs are composed of a 100kDaheavy chain and a 50kDa light chain (3, 4). Theheavy chain is responsible for binding to a receptorlocated on the membrane of the neuron andtranslocation of the light chain, whereas the lightchain acts as a highly specific proteolytic enzymeonce internalised into the cell. BoNTs act via a

316 S. Adler et al.

complex mechanism (5) that comprises binding tocell surface, high affinity receptors of cholinergicnerve endings (6, 7), penetration into the plasmamembrane by endocytosis (8), and pH-dependenttranslocation of the light chain from the lumen ofthe vesicles into the cytoplasm. The light chainacts as a zinc-dependent protease that cleaves oneor more components of a membrane fusion complexcomposed of syntaxin, SNAP-25 (a synaptosomal-associated protein) and synaptobrevin, therebyblocking exocytosis (9, 10). As a result of this cleav-age event, the release of acetylcholine at the neu-romuscular junction is inhibited, leading to asubsequent flaccid muscular paralysis.

Previous Meetings on BoNT PotencyTesting

Two previous meetings were held in 2006, with thegoal of promoting alternative methods to the LD50potency test. These were: a meeting hosted by theEuropean Directorate for the Quality of Medicinesand Health Care (EDQM) of the Council of Europe;and a workshop in the USA organised by the USInteragency Coordinating Committee on theValidation of Alternative Methods (ICCVAM), theUS National Toxicology Program (NTP) Inter -agency Center for the Evaluation of AlternativeToxicological Methods (NICEATM) and theEuropean Centre for the Validation of AlternativeMethods (ECVAM), in response to a nomination bythe Humane Society of the United States (HSUS).The outcomes of these two meetings were reportedand discussed at the ZEBET meeting in Berlin.

The EDQM Meeting

Catherine Milne, from the EDQM, reported on theBotulinum Toxin 3R Strategy Meeting, hosted inApril 2006 by the EDQM, in Strasbourg, France.The participants included representatives of thethree manufacturers of BoNT-A marketed inEurope, representatives from EU official nationaltesting laboratories and licensing authorities,ECVAM and the EDQM. In addition, a conferencecall with authorities from the US Food and DrugAdministration (FDA) Center for Drug Evaluationand Research (CDER), Health Canada, and ICC-VAM was included, in order to permit the exchangeof views in a more global context. The goal of themeeting was to look to the future for potency testingof BoNT in the context of the application of theThree Rs principles, supported by the relevant con-ventions of the Council of Europe and EU Directives.

The consensus was that the LD50 assay wouldremain the ‘gold standard’ for validation of alter-native methods, but that this can only be used in aproduct-specific and method-specific setting due to

method variability between laboratories, lack ofcommon reference standards for the LD50 assay,and the need to maintain product-specific unitage.This should be taken into account as the alterna-tive methods are being developed and validated.

It was agreed that in any replacement strategy,all the functions of the toxin (cell binding, inter-nalisation, translocation and endopeptidase activ-ity) would have to be assessed. This could be via asingle assay or by a multi-assay approach.

A reduction strategy, involving a full assayapproach for the bulk toxin (covering all threefunctions) and a simplified approach for the finalproduct lot release, measuring a reduced set of cri-teria to monitor consistency, was considered to beworth investigating, as all lots would be derivedfrom the same bulk toxin. Examples of this type ofapproach are already allowed in the EP, and are inuse for human vaccines. The usefulness of thisapproach would be tightly linked to the definitionof the robustness and reproducibility of the actualproduction process and the GMP controls in place.

The potential problems that might arise in theclinic if the unitage was changed, were a signifi-cant concern. The expression of units for alterna-tive assays could be expressed in LD50 equivalentsto avoid problems in the clinic. Possibilities for thiswould need to be further explored.

Based on the data presented, the product-spe-cific aspects of the LD50 assay and the complica-tions related to unitage, no clear candidate for acommon method or reference was identified at thattime. Nevertheless, the participants were open tothe idea of exchanging reagents and data with theauthorities, in order to facilitate developmentwork, and the EDQM expressed a willingness toorganise collaborative studies on the alternativemethods in the context of the biological standardi-sation programme, once they had been furtherexplored and established.

The ICCVAM/NICEATM/ECVAMWorkshop

Martin Stephens, from the HSUS, reported on theICCVAM/NICEATM/ECVAM Scientific Work -shop on Alternative Methods to Refine, Reduce andReplace the Mouse LD50 Assay for BotulinumToxin Testing, that was held in Silver Spring, MD,USA, in November 2006. About 115 participants,mostly from Europe and North America, partici-pated in this workshop, including scientists fromgovernmental and academic institutions, nationaland international regulatory authorities, industry,and the animal welfare community.

The consensus of the panel discussions at theICCVAM/NICEATM/ECVAM Workshop (http://iccvam.niehs.nih.gov/methods/biologics/bot_workshop.htm) was that some of the alternative

ZEBET Expert Meeting 317

methods could be used in specific circumstances, orin a tiered-testing strategy, to reduce and refinethe current LD50 assay. However, none of themethods reviewed could be used individually as acomplete replacement for the mouse potency assayat that time. The panel discussions noted that,with additional development and validationefforts, one or more of the reviewed methods mightbe useful as a replacement for some of the currentBoNT testing applications in the future. Finally,some best practices were discussed for decreasingthe numbers of animals used, including the estab-lishment of reference standards, the employmentof standardised methodology, and reduction in thenumber of doses tested for assays when confirma-tion of potency is being evaluated.

It was reported at the 2009 Expert Meeting inBerlin that the US workshop in 2006 succeeded inbringing together international experts who workon disparate applications of the LD50 assay andalternatives for BoNT testing, and provided asnapshot of the status quo. It was less successful infulfilling its stated goal of developing a set ofstrategic recommendations for future actions.

The Regulatory Framework

Regulation on the protection of animalsused for scientific purposes, according toDirective 86/609/EEC

The protection of animals used for experimentaland other scientific purposes is regulated inEurope by Directive 86/609/EEC, which is cur-rently under revision (11). Among the aims of thedirective is the reduction of the numbers of ani-mals used for experiments, by requiring that ananimal experiment should not be performed whenan alternative method exists, and by encouragingthe development and validation of alternativemethods to replace animal methods. Furthermore,it requires the minimisation of any pain, sufferingand distress caused to test animals. The MemberStates of the European Union have implementedthe Directive 86/609/EEC into their legislations.However, some of the member states, includingGermany and the UK, have implemented nationalanimal welfare acts that go beyond the require-ments stated in Directive 86/609/EEC.

Regulations for the marketing of pharmaceutical products

European regulations

In Europe, the LD50 testing of BoNT, i.e. the bulkdrug substance and the final drug product lot, is a

requirement of the EP Monograph 2113, Botul -inum Toxin Type A for Injection (1). EP mono-graphs and texts are the legally binding qualitystandards for all medicinal products in the EU.The individual marketing authorisations for prod-ucts marketed in the EU must apply these stan-dards according to the EU Directive for HumanMedicines 2001/83/EC (12), as amended byDirective 2004/27/EC (13).

The method currently used for the potency test-ing of all BoNT-A pharmaceutical products on theEU market, as accepted in the marketing authori-sations by the European regulatory authorities, isthe LD50 assay. Nevertheless, the specific mono-graph and the general notice open the possibility ofdeveloping alternative methods, in particular inthe interest of animal welfare and the Three Rs.Since 2005, three different alternative approacheshave been noted in the EP Monograph 2113,Botulinum Toxin Type A for Injection (1), but usersare not limited to these methods, which are: “anendopeptidase assay in vitro, an ex vivo assayusing the mouse phrenic nerve diaphragm, and amouse bioassay using paralysis as the endpoint.”Furthermore, it is stated in the EP that alternativemethods must be suitably validated in comparisonto the reference method, which is the LD50 assay:“After validation with respect to the LD50 assay(reference method), the product may also be assayedby other methods that are preferable in terms ofanimal welfare, including one of the above men-tioned.”

EP Monograph 2113, Botulinum Toxin Type Afor Injection (1), is currently under considerationfor revision. The revision provides additionaldetails for a refined in vivo assay as an alternativeto the LD50 potency assay that uses local paralysisas an endpoint. Moreover, a new monograph forBoNT type B for injection is under development,which presents a similar strategy for potency test-ing as for BoNT type A. The new and revised mono-graphs were published in July 2009 inPharmeuropa 21.3, for public consultation (14, 15).

FDA regulations

In the USA, FDA regulates BoNT products asdrugs. The FDA requires the submission of data onBoNT potency, as well as on safety, purity, steril-ity and other parameters for these products. Theregulations are specified in the Code of FederalRegulations 21 Chapter I (16). Although the FDArequires the appropriate potency testing of BoNTproducts, and specifies the performance attributesof such a test, it does not state which particulartest should be used. While recognising that theinternational standard is the mouse LD50 potencyassay, the FDA will also review data submitted in

318 S. Adler et al.

support of alternative potency assays and, whenwarranted, issue formal approval for their use.

Quality Guidelines (ICH and EMA)

The LD50 potency test is not only performed forrelease testing of BoNT, but also for other manufac-turing purposes (Figure 1). For the marketingauthorisation application, stability testing of drugsubstance and drug product is required. In addition,according to the principles of Good Manu facturingPractice (GMP), annual stability studies have to beperformed after the medicinal products have beenauthorised. International (Europe, USA and Japan)Conference on Harmonisation of Technical Require -ments for Registration of Pharmaceuticals forHuman Use (ICH) Guideline Q5C, Stability Testingof Biotechnological/ Biological Products, requires,for products that are linked to a definable and meas-urable biological activity, that testing for potencyshould be part of the stability studies (17).

Furthermore, for the marketing authorisationapplication process in Europe, validation studiesare required by the regulatory authorities of theMember States that demonstrate the suitabilityand reproducibility of the manufacturing process.The development of new formulations or productsalso requires additional process validation andstability studies, where LD50 tests are performed(ICH Q5E; 18).

Validation Requirements forAlternative BoNT Potency Assays

The basis for successful validation studies foralternative bioassays is ICH guideline Q2Validation of Analytical Procedures (19). Since theLD50 tests of the different BoNT-producing com-panies vary and are product-specific (20, 21), thespecific alternative methods have to be validatedfor each individual medicinal product. In this con-text, no standard validation programme can beapplied, but individual validation plans have to bedeveloped.

For the successful validation of an alternativeassay, the specificity, reproducibility, precision,range and robustness have to be demonstrated.During the development and validation of thealternative method, sources of variability shouldbe identified and variability should be minimised.The number of batches required for a successfulvalidation and cross-validation with the LD50 test,and acceptable statistical limits, are key discus-sion points for validation requirements.

The Federal Institute for Drugs and MedicalDevices (BfArM), and other European RegulatoryAgencies, each offer the opportunity for companiesto discuss their development plans and strategiesat an early stage. These early discussions betweenmanufacturers and regulatory authorities can bebeneficial in facilitating the process of regulatoryacceptance of alternative methods, and can help

Figure 1: A schematic overview of the BoNT production process, demonstrating whenpotency tests are required to meet regulatory requirements

Potency tests are required in the EP for the bulk purified toxin and final reconstituted product. In addition, potencytests have to be performed according to the ICH, to demonstrate the product stability over time (ICH Q5C) and forprocess validation (ICH Q5E).

Bulk Final purified Final reconstituted

Seed lot toxin bulk product

toxinpurification

Potency test (EP)

Proof ofactive toxin

Potency test (EP)Potency test (ICH Q5C)Potency test (ICH Q5E)

> 100,000 folddilution

drying

ZEBET Expert Meeting 319

manufacturers to plan appropriate validationstudies on methods, such that they ultimatelymeet the regulatory requirements.

In this context, the question arose at the ZEBETExpert Meeting, of whether a general guidancedocument on the validation of alternative methodsto the BoNT LD50 potency test would be helpful.However, the experts agreed that validation ofsuch methods could not be generalised, since guid-ance for product-specific validation can only begiven by the regulatory authorities in close com-munication with the manufacturers before andduring the validation process, and will be specificfor the particular product, the proposed alternativepotency testing method used, and the data set gen-erated.

Workshop participants from the regulatoryauthorities stated at the meeting that after thesuccessful product-specific validation of a suitablealternative method and regulatory acceptance of aspecific product, it would be easier for later appli-cants to achieve regulatory acceptance of the samemethod for use with other products. However, com-prehensive data will need to be provided by eachapplicant. The experts also agreed that, from apharmaceutical point of view, there would be noobvious reason why a method that could replacethe LD50 potency test for one specific productshould not work for the other products, as theBoNT-A medicinal products currently on the mar-ket do not differ in their modes of action.

Alternative Methods in ProductStability and Process Validation

There was discussion of whether alternative meth-ods, unacceptable for product release alone (sincethey do not cover all functions of the toxin — e.g.the endopeptidase assay), could nonetheless beemployed for other purposes, such as the monitor-ing of product stability or process validation. Theparticipants had different opinions on this, butagreed that, in principle, it would be desirable touse the same assay for all testing requirements,i.e. stability, process validation and productrelease, in order to provide comparable data.Nevertheless, this issue was also discussed withregulatory experts. They stated that a reduced sta-bility programme could be performed that couldemploy alternative assays, e.g. the endopeptidaseassay, as a replacement for the LD50 potencyassay for at least every second control test peryear. Intermediate stability control tests wouldstill have to be performed with the LD50 assay, oran alternative test reflecting fully functional activ-ity of the toxin.

Such an approach might be acceptable inEurope, if sufficient data were available to provideinformation on the stability of the products.

However, it was mentioned that the lack of globalharmonisation among regulatory authorities couldbe a problem.

An International Reference Standard

Due to the current lack of standardisation, theintroduction of an international reference stan-dard was discussed, to facilitate better comparisonof different products and standardised unitage.The EP requires the use of in-house reference stan-dards, as a common standard is currently unavail-able. This is partly because of the differencesobserved with the LD50 assays used for the differ-ent products, which create difficulties in establish-ing such a standard. There have also been concernsover the problems that could arise in the clinic, ifthe existing unitage for dosing were changed forthe individual products due to the use of a new ref-erence unit. Nevertheless, it was argued that sucha standard could help to promote the acceptance ofalternative methods to the LD50 assay.

However, the majority of the participants sug-gested that the development of an internationalreference standard would be time-consuming andcurrently not necessary, since, for the acceptanceof alternative potency assays, product-specific val-idations are required that would use the in-housestandard. It was argued that the work on an inter-national reference standard would currently delayvalidation efforts.

In contrast to the development of an interna-tional standard for the harmonisation of BoNTunits, the experts recommended the developmentof BoNT reference materials that would be essen-tial for comparing alternative potency assays andtheir performance in different laboratories.Reference materials already available at theNational Institute for Biological Standards andControl (NIBSC) could be considered for interimuse in planned method comparison studies (21).

Harmonisation

A main concern expressed at the meeting was thelack of global consensus between regulatoryauthorities on the implementation of alternativemethods for BoNT potency testing. In Europe, theimplementation of an alternative test methodrequires a variation application either at thenational or at the European level. Moreover, theworld-wide implementation of alternative methodsfor medicinal products is much more complex, andis considered to be an obstacle for globally-actingmanufacturers. Some of the experts expressedtheir wish to promote international harmonisationand mutual acceptance criteria for the validationof alternative methods.

320 S. Adler et al.

Establishment of a ‘BoNT ExpertWorking Group’

In order to overcome current difficulties in thedevelopment, optimisation, validation and regula-tory acceptance of alternative methods to theBoNT potency test, the experts discussed theestablishment of a ‘BoNT Expert Working Group’.This group would deal with the essentials foracceptability, comparability and implementationof the Three Rs in BoNT potency testing. Therewas a common agreement to initially focus on thesituation in Europe, as this could speed up theprocess of the implementation of alternative BoNTpotency assays in this region.

A core group should be established, comprisingexperts from European regulatory authorities,Three Rs and validation institutions, manufactur-ers and scientists. Experts from overseas should beinvited as observers. In addition, a network ofadditional experts and stakeholders should beestablished, who would join the ‘BoNT ExpertWorking Group’ on a case-by-case basis. The Groupshould meet regularly up to three times per yearduring a proposed time-frame of four years, andshould publish its recommendations on a regularbasis.

The objectives of the ‘BoNT Expert WorkingGroup’ would include, among others, providingadvice and guidance on validation requirementsfor the proposed alternative methods, and definingminimum standards for a validation framework inorder to implement the Three Rs in BoNT potencytesting. The Group would assess replacement,refinement and reduction strategies, and reviewdata, reference materials, methods and scientificapproaches, and would promote the standardisa-tion of protocols. The Group would also providemore transparency between the stakeholders andregulatory authorities.

All the attending participants expressed thewish that an independent institution should beresponsible for the establishment and manage-ment of the ‘BoNT Expert Working Group’, and theBfR/ZEBET and BfArM agreed to accept this task.

Alternative Methods to the LD50Potency Test

Refinement assays

There are a number of opportunities for refine-ment of the LD50 potency assay in the potencytesting of BoNT. These could include, for example,the use of lower species, shortening of the durationof the procedure, and endpoints earlier than death,which would imply an increased frequency of ani-mal inspections and intervention when death is

impending, as has often been the practice withother lethal dose tests. In addition, refinement canbe achieved by alternative experimental designsand analyses. To date, a number of these opportu-nities have been identified and exploited.

Humane endpoints

In the interests of the Three Rs, the conventionalLD50 test was deleted from the OECD TestGuidelines Programme for the testing of chemicalsin 2001. It was replaced by three alternative meth-ods, one of which includes the use of non-lethalhumane endpoints (22). This deletion was only pos-sible after weighing the risks and benefits, thusensuring that all three alternative methods couldserve the regulatory needs.

For BoNT, while complete deletion of the test isnot yet possible, careful evaluation of the clinicalsigns that indicate imminent death has resulted insome success in implementing earlier (morehumane) endpoints. Such signs include cyanosis,indrawn scaphoid abdomen, gasping, and behav-ioural responses such as immobility and social iso-lation. The application of such humane endpointscould be undertaken without affecting the resultsof the LD50 test, and data were presented thatshowed that up to 21% of the animals could be suc-cessfully culled by a humane method, rather thanbeing allowed to die (Table 1). However, due to thefact that it is not yet feasible to establish such end-points at a sufficiently early time-point (Figure 2),the current employment of humane endpoints doesnot permit a significant decrease in the suffering ofthe animals used.

In addition, the above-mentioned clinical signsthat should indicate imminent death have not yetbeen sufficiently investigated. A change from thelethal endpoint to the refined humane endpointswould also require a product-specific validation

Table 1: The results of experimental studiesthat applied humane endpoints

Study number % Humane endpoint

1 16%2 11%3 11%4 15%5 15%6 21%

Average 15%

% Humane endpoint specifies the percentage of micethat could have been culled by a humane method ratherthan going on to die, without affecting the results of theLD50 test.

ZEBET Expert Meeting 321

study, and most experts recommend that furthereffort should not be invested in the development ofrefined assays based on humane endpoints at pres-ent, but that the focus should be on other refine-ment or replacement assays.

Local paralysis assays

Changing the LD50 potency assay to an alterna-tive, less severe, test by using non-lethal end-points, is complicated. Such alternative testsinclude the mouse flaccid paralysis assay (23), therat muscle force (RMF) assay (24), the digitalabduction score (DAS) assay (25, 26), and the ratgastrocnemius assay. From a Three Rs point ofview, these assays would be preferred to the LD50potency test, as only sub-lethal doses of the toxinare injected and animals generally fully recoverfrom the treatment, even at the highest dose used.

In the flaccid paralysis assay, different sub-lethal concentrations of BoNT are injected subcu-taneously into the top of the left hind leg, theinguinocrural region of the lower abdomen of amouse. The animals are then scored by two or moreobservers according to the size of the local abdom-

inal bulge (Figure 3) at 24 and 48 hours, by usinga five-point scale (Table 2). This scoring systemwas introduced because the method had been pre-viously criticised for its subjective and non-quanti-tative read-out.

The dose–response, suitable for calculation ofpotency, is obtained by plotting the scores againstthe log toxin dose for the test and reference sam-ples. Then the relative potency is calculated by aparallel line method. This approach has been usedat the NIBSC, primarily to confirm activity inproduct-specific reference preparations. At theNIBSC, the LD50 test has not been performed formore than 10 years (27).

The advantages of the flaccid paralysis assay arenot only a refined endpoint, but also ease of use,fast one-off read-out (48 hours), and increased sen-sitivity with comparable precision, coupled withthe use of only 20% of the number of animals usedin a comparable LD50 test. Furthermore, the dos-ing is considered more relevant to clinical use, asmuscle paralysis rather than systemic toxicity isevaluated. Recent potency data presented by theNIBSC confirmed an excellent agreement withLD50-determined potency for seven different ther-apeutic products.

Figure 2: The numbers of dead mice over time, following BoNT injection

The graph shows the mean results of six independent runs.

deat

h (

%)

hours

Sample 6

Sample 5

Sample 4

Sample 3

Sample 2

Sample 1

240

10

8

6

4

2

0

32 40 48 56 64 72 8088 96

322 S. Adler et al.

Figure 3: The mouse flaccid paralysis assay

The mouse on the left has been injected with BoNT to assay the potency of a test sample, and shows an abdominalbulge at the site of injection (stage 3, Table 2), compared with the uninjected control mouse on the right (stage 0,Table 2) (illustrations provided by D. Sesardic et al. [29]).

However, criticism was expressed by some expertsduring the meeting. One of the concerns was that,although local paralysis was a relevant endpoint, thescoring could still be too subjective. Furthermore, itwas mentioned that the low precision of 20% —which was the rate achieved with an 80% reductionin animal number — may necessitate an increase inthe number of test animals used, so that the numberof animals necessary may eventually be close to thatfor the LD50 test. The performance of a validationstudy involving several laboratories, to confirm theprecision of the test that is referred to in the EP, wasconsidered to be a useful proposal.

The experts also discussed the rat muscle force(RMF) assay, the digital abduction score (DAS)

assay and the rat gastrocnemius assay. All threemethods were evaluated as being not yet suitablefor BoNT potency testing, and the RMF assay andthe rat gastrocnemius assay were recommendedonly for research purposes.

Replacement assays

Ex vivo assays

1. The mouse hemidiaphragm assay (HDA): TheHDA (28) operates with an isolated nerve musclepreparation (Figure 4). When BoNT is added tothe organ bath in which the muscle has been

Table 2: An example of a recommended scoring scheme that is used to define the responseof individual mice in the flaccid paralysis test

Stage Description

0 No signs, normal

1 Just detectable bulge, e.g. covering an area of approximately 0.5cm in diameter, or less.

2 More pronounced bulge, e.g. covering an area greater than 0.5cm in diameter, but less than the maximum radius of the hind leg heel.

3 More extensive bulge extending over a larger area. Extending below hips and top of thigh when viewed from the side and beyond the maximum radius.

4 Maximal local effect. More extensive bulge extending over a larger area. Will often extend as far as the bottom of the rib cage, or over a large area with extensive distension or bulging.

ZEBET Expert Meeting 323

placed, the contraction amplitude of the nerve-stimulated muscle gradually declines until it dis-appears. The contraction of the diaphragm isrecorded isometrically, by using a force trans-ducer, and the contraction amplitude is assessedover time. The time at which the muscle is paral-

ysed by 50% of the original contraction amplitudeis determined, following the application of differ-ent concentrations of BoNT. The recorded time ofthe onset of 50% paralysis (t1/2) decreases withincreasing concentrations of BoNT.

The advantages of the HDA are fast read-out (a

Figure 4: The hemidiaphragm assay

The hemidiaphragm assay operates with an isolated nerve muscle preparation that is placed in an organ bath. Thecontraction of the diaphragm is recorded isometrically, by using a force transducer, and the contraction amplitude isassessed over time (illustrations provided by H. Bigalke).

324 S. Adler et al.

few hours) and full functionality. This is an organ-otypic assay — which means that the animals arehumanely killed, and the determination of BoNTpotency is then carried out on the isolated nervemuscle preparation. The method is technicallychallenging and requires sophisticated and expen-sive equipment. The robustness of the assay wastested at the Medical School of Hannover, in thepresence of human serum with various protein andlipid concentrations. The test was validated formeasurement of the presence of antibodies to neu-rotoxin in sera from BoNT-treated patients, andwas shown to fulfil the criteria of specificity, selec-tivity and precision for this purpose. A validationstudy for BoNT potency testing by using the HDAbegan in 2009, in collaboration with MerzPharmaceuticals.

The discussions at the Expert Meeting revealeda number of questions that could not be answeredat present, e.g. the product specificity of theassay, the strain specificity of the ex vivo mate-rial, and the comparability of the steepness of theconcentration–response curves of the HDA andthe LD50 assay. Recent published data from theNIBSC identified improved precision with aninbred mouse strain, but also observed supplier-specific differences in A toxin activity in the HDA(29).

2. The intercostal neuromuscular junction (NMJ)assay: Another ex vivo assay discussed at theExpert Meeting was the intercostal NMJ assay,which uses ex vivo rat intercostal muscle prepar-ations (30). As described above for the HDA, this exvivo assay still requires the killing of animals(rats), but successfully addresses the question oftheir suffering. Up to six individual rat intercostalrib sections can be obtained from a single rib cage.Upon treatment with BoNT injected directly intothe tissue sections, the contractile response ismonitored upon electrostimulation of the inter-costal nerve over time. The NMJ assay was evalu-ated by the experts as an assay that employs arelevant endpoint and reduces the suffering of theanimals used. However, a pre-validation studywith this assay, performed by one manufacturer,was unsuccessful. The experts at the meeting con-sidered that this might have been due to the rela-tively small size of the muscle preparations. Thetissue size employed in the HDA, for example, issubstantially bigger, and the phrenic nerve is a rel-atively robust nerve.

In vitro assays

1. Cellular assays: Potency assays based on cellsoffer an advantage over other in vitro models, asthey provide information on all the essential stepsin the BoNT intoxication process, i.e. binding,

internalisation and intracellular activity. In the-ory, only the cell-based assays have the potentialto fully replace animal-based potency tests for allapplications. Furthermore, they could be appliedto all stages of the production process, includingthe testing of the purified bulk active toxin com-ponent.

Despite active research in this field, no suitableassays with the desired sensitivity and character-istics for the potency assessment of toxin-basedproducts are yet available. Primary neuronal cells,derived from rat spinal cords, have been exten-sively studied and confirmed to offer the desiredsensitivity. However, robustness and reproducibil-ity have been limiting factors so far in the intro-duction of such methods within a quality controlsetting. Human cell lines of neuronal lineageswould provide the best option in the testing ofBoNT potency, but, up to now, none of the avail-able cell lines have offered the required sensitivity,unless they were fully differentiated to neuronalcell types.

A number of cell-based assays for potency test-ing are currently under development by expertsfrom both the NIBSC and Allergan, Inc. Strategiesat the NIBSC have focused on the differentiatedhuman neuroblastoma cell line, SH-SY5Y. Withthese cells, the inhibition of depolarisation-evokedneurotransmitter release is measured after incu-bation with botulinum toxin. In addition, collabo-rative studies with the NIBSC Scientific Divisionof Cell Biology and Imaging (CBI), have providedthe opportunity to study differentiated cell linesand human stem cells that were grown in multi-electrode dishes. With such dishes, the sponta-neous electrical activity of the cells can be recordedbefore and after BoNT exposure. Although theseapproaches showed increased sensitivity to pico-molar levels of BoNT, it was anticipated at themeeting that they are not sufficiently robust forquality control testing, but are likely to remainuseful research tools.

Another strategy focused on by the NIBSC, wasthe development of a new generation cell-basedendopeptidase assay that employs antibody reagentsalready used in the SNAP-25 ELISA format. In addi-tion, new antibodies were generated for the captureand detection of SNAP-25, and an optimum protocolfor the sensitive capture of recombinant SNAP-25and cleaved SNAP-25 from cell lysates was devel-oped. The sensitivity of the method was confirmed inthe fully-differentiated human neuronal cell line,SH-SY5Y, although primary rat spinal cord cellsoffered greater sensitivity in Western blotting (31)and immunodetection of cleared SNAP-25 (unpub-lished data). However, robustness and precision incomparison to the conventional LD50 assay have notyet been assessed, so further studies are necessary.

Allergan Inc. have developed a novel cell-basedscreening Western blot assay for measuring BoNT

ZEBET Expert Meeting 325

activity, based on a SNAP25197-specific antibody.They have reported substantial progress in theoptimisation of the assay, resulting in increasedsensitivity and improvements in assay read-out.

Another ongoing project, involving collaborationbetween the University of Veterinary MedicineHannover and the Georg-August University ofGöttingen, focuses on the development of a BoNTtesting assay based on the human teratocarcinomacell line, Ntera2 (NT-2). These cells can be differ-entiated into fully functional, post-mitotic neurons,which display a variety of neurotransmitter phe-notypes (Figure 5; 32). In this project, the processto generate post-mitotic neurons was significantlyshortened to approximately four weeks’ duration(33). During differentiation, the NT-2 acquiredtypical neuronal markers, such as beta-tubulintype III, MAP-2 and phosphorylated tau, and grewneural processes which expressed punctateimmunoreactivity for synapsin and synaptotag-min, suggesting the formation of synaptic struc-tures. A major subset of the model neurons showedimmunoreactivity to the cholinergic markers,choline acetyltransferase and vesicular acetyl-choline transporter. Moreover, spontaneous post-synaptic currents in glia-free cell cultures weredetected, which result from the firing of excitatory

and inhibitory NT-2 neurons (34). These sponta-neously-active neural networks may be a usefultool in the detection of BoNT potency in blockingneurotransmission, by using cell physiologicaltechniques.

The experts have extensively reviewed the use ofcell-based assays, in order to determine acceptablecell types for the BoNT potency assays. Theyagreed that primary neuronal cells are notfavoured from a regulatory point of view, as theyare too variable. No primary cells are currentlybeing used by industry to develop new BoNTassays. Cell lines, on the other hand, are easier tostandardise than primary neuronal cells, but oftenare not as sensitive when used in assays. If used inBoNT potency assays, cell lines must first be verywell characterised.

In addition, the experts discussed the cell typesthat would be best suited for testing the potency ofBoNT. They expressed different opinions about thesuitability of neuronal cells, either alone or in com-bination with muscle cells. Some experts thoughtthat cells applied in a potency assay should resem-ble, as closely as possible, motor neurons.However, other experts believed that any cells thatcould reflect the action of BoNT, from receptorbinding, to SNAP-25 cleavage or blocking of trans-

Figure 5: The in vitro differentiation of NT-2 cells into post-mitotic neurons

a) Expansion of NT-2 cells in free floating cell spheres. This technique resembles the sphere culture method used inembryonic stem cell differentiation. b) Post-mitotic NT-2 neurons, showing the immunofluorescent staining of aneuronal marker. Spindle shaped cell bodies send out processes that form a neural network (illustrations provided byG. Bicker, H. Böhnel and F. Gessler).

a) b)

100μm 50μm

326 S. Adler et al.

mitter release, would be useful. In addition, theimportance of the species of origin of the cells wasdiscussed. The question arose as to whether neuro-muscular junctions should be cholinergic in aBoNT potency test, or could be also glutaminergic,as, for example, found in some lower species. Someexperts thought it would be fundamental for a testto be based on a system with cholinergic neuro-muscular junctions. However, others thought thatthe transmitters would not be of importance, butsolely the substrate, SNAP-25, as BoNT would alsoact on different types of glands.

There are a number of challenges associatedwith the use of cellular BoNT potency assays. Theexperts agreed that a key problem with the cell-based assays currently under investigation, wasthe relatively flat dose–response curves seen, incomparison with the LD50 assay. This observationwas explained by the minimum signal resultingfrom a cell due to the low amount of transmittersreleased. This effect is amplified in vivo or inorganotypic assays, due to the muscle activity,especially in the voltage-dependent channels.

A steep concentration–response curve was con-sidered necessary, as a good separation of thetested concentrations is of great importance. TheEP monograph requires a potency specificationbetween 80% and 125% (1). Therefore, three to fiveconcentrations have to be tested in this range.

The experts discussed how the cell signals couldbe amplified in vitro. They agreed that transgeniccell lines, e.g. Green Fluorescent Protein (GFP)-expressing cell lines or cells over-expressing recep-tors or substrate, could result in steeperconcentration–response curves and therefore pro-vide suitable in vitro models for BoNT potencytesting. In addition, the incubation time of BoNTin the potential assay will have an influence on thesteepness of the concentration–response curve.

Finally, it was emphasised by the experts thatcells are the minimum accepted unit that can coverall the steps of the BoNT mechanism of toxicity. Togo beyond this level to molecular assays wouldrequire a combination of different test methods.However, no cell-based assays are as yet at a stageappropriate for a validation study.

Molecular assays

Endopeptidase assays

The information on the intracellular mode ofaction of BoNTs provided the opportunity todevelop enzyme cleavage assays. A series of paperswere published, in which endopeptidase assayswere applied to BoNT drug products (35–37). Theapproach has been validated in-house at theNIBSC for a particular product, and has been in

use for the last five years, to confirm the manufac-turer’s claim of potency against a product-specificreference preparation. The NIBSC reported that,by using the endopeptidase assay, it was possibleto detect nanogramme quantities of toxin in thepresence of high concentrations of bulking agentsand stabilisers in the final product, with superiorsensitivity and excellent agreement with themouse LD50. Furthermore, the assay afforded ahigh precision that was superior to that of the invivo models and, in addition, ease of use, robust-ness and ready transferability. The NIBSC hasrecently improved the assay protocol (38), and nowconsiders this to be an excellent method for theevaluation of trends and consistency.

The method has been criticised for its inability toreflect all the functional domains of the toxinknown to contribute to the in vivo activity, as onlythe light chain activity is measured. The expertsagreed that it could be promising in support ofpotency testing for process validation and stability,but might not be suited for product release. On theother hand, some experts considered the endopep-tidase assay to be suitable for assigning thepotency of the final product after successful prod-uct-specific validation, if the potency of the finalbulk was assessed by an assay able to cover thewhole toxic mechanism of BoNT.

Binding and translocation assays

Binding and translocation assays are available,but are still at the research stage. Ipsen, for exam-ple, has developed a ganglioside binding assay.The reagents for such assays are not yet commer-cially available and may be of limited use, in viewof the fact that gangliosides might not be highaffinity receptors for the toxins.

Combination of molecular assays

Due to the fact that molecular assays can only pro-vide information on single steps in the toxicologicalmechanism of BoNT, various tests would have tobe combined in order to provide all the necessaryinformation on potency. The experts discussed howdifferent assays could be combined, and agreedthat this would be a very difficult task, as differenttests would provide different potency values(units) that would have to be linked or weighted insome way.

Public Funding

The experts recommended increased funding onresearch into cellular neurotoxicity mechanisms,in order to identify relevant mechanisms for novel

ZEBET Expert Meeting 327

assays, and into feasibility and prevalidation stud-ies of existing methods. Furthermore, it wasreported that funding possibilities exist from theGerman Federal Ministry of Education andResearch, and that also the US National Instituteof Health is currently funding small enterprises forthe development of alternative assays for BoNTpotency testing.

Reduction of Animal Numbers andRefinement of the LD50 Assay

Allergan and Ipsen reported a substantial reduc-tion in the numbers of animals used in the LD50potency test. Allergan have improved their dosingrange, which has resulted in the testing of fewerdoses and the requirement for fewer animals perdose. By using this approach, they could reduce theanimal numbers by 50%. Ipsen stated that theyhave reduced the testing time from 96 to 72 hours,which resulted in a reduction of suffering (refine-ment). Merz informed the participants of the meet-ing that these approaches have also been inpractice at their facilities, since the beginning oftheir own BoNT product development.

Conclusions and Recommendations

1. The industry and scientists have made progressin implementing reduction and refinement inBoNT potency testing, and in the developmentof alternative potency assays.

2. General guidance on the product-specific vali-dation of alternative methods for BoNT potencymethods cannot be provided. Guidance on prod-uct-specific validation is given continuously byregulatory authorities to manufacturers, bothbefore and during the validation process.

3. Regulatory authorities should actively beinvolved in international harmonisation and indefining mutual acceptance criteria for alterna-tive methods to LD50 potency testing.

4. A ‘BoNT Expert Working Group’ should be estab-lished, to provide advice on validation require-ments for alternative methods, define minimumstandards for a validation framework, assessrefinement and reduction strategies, review data,test methods, reference materials and scientificapproaches, and promote the standardisation ofprotocols. BfR/ZEBET, in collaboration withBfArM, will be responsible for establishing the‘BoNT Expert Working Group’.

5. Funding should be made available and its usecoordinated, to develop and validate alternative

methods for BoNT potency testing according tothe Three Rs principles. In particular, thedevelopment of a replacement alternative forBoNT testing should have the highest priority.

6. Reference materials should be developed to pro-mote the comparability of different BoNTpotency assays.

7. Alternative methods that are not yet able toreplace the LD50 test in all steps of the produc-tion process may still be useful in testing forproduct stability, process validation or lotrelease.

References

1. Council of Europe (2006). Botulinum Toxin Type Afor Injection 01/2005:2113. In European Pharma -copoeia 6.0, pp. 1117–1119. Strasbourg, France:Council of Europe.

2. Simpson, L.L. (1981). The origin, structure, andpharmacological activity of botulinum toxin.Pharmacological Reviews 33, 155–188.

3. Lacy, D.B. & Stevens, R.C. (1997). Recombinantexpression and purification of the botulinum neuro-toxin type A translocation domain. ProteinExpression & Purification 11, 195–200.

4. Hanson, M.A. & Stevens, R.C. (2000). Cocrystalstructure of synaptobrevin-II bound to botulinumneurotoxin type B at 2.0Å resolution. NatureStructural Biology 7, 687–692.

5. Simpson, L.L. (1980). Kinetic studies on the inter-action between botulinum toxin type A and thecholinergic neuromuscular junction. Journal ofPharmacology & Experimental Therapeutics 212,16–21.

6. Dolly, J.O., Black, J., Williams, R.S. & Melling, J.(1984). Acceptors for botulinum neurotoxin resideon motor nerve terminals and mediate its internal-ization. Nature, London 307, 457–460.

7. Simpson, L.L. (1986). Molecular pharmacology ofbotulinum toxin and tetanus toxin. Annual Reviewof Pharmacology & Toxicology 26, 427–453.

8. Black, J.D. & Dolly, J.O. (1986). Interaction of 125I-labeled botulinum neurotoxins with nerve termi-nals. II: Autoradiographic evidence for its uptakeinto motor nerves by acceptor-mediated endocyto-sis. Journal of Cell Biology 103, 535–544.

9. Schiavo, G., Rossetto, O., Santucci, A., DasGupta,B.R. & Montecucco, C. (1992). Botulinum neurotox-ins are zinc proteins. Journal of BiologicalChemistry 267, 23,479–23,483.

10. Schiavo, G., Matteoli, M. & Montecucco, C. (2000).Neurotoxins affecting neuroexocytosis. Physiolog -ical Reviews 80, 717–766.

11. Anon. (1986). Council Directive 86/609/EEC of 24November 1986 on the approximation of laws, regu-lations and administrative provisions of theMember States regarding the protection of animalsused for experimental and other scientific purposes.Official Journal of the European CommunitiesL358, 18/12/1986, 1–28. Available at: http://eur-lex.europa.eu/smartapi/cgi/sga_doc?smartapi!celex-plus!prod!DocNumber&lg=en&type_doc=Directive&an_doc=86&nu_doc=609 (Accessed 22.06.10).

328 S. Adler et al.

12. Anon. (2001). Directive 2001/83/EC of the Euro -pean Parliament and of the Council of 6 November2001 on the Community code relating to medicinalproducts for human use. Official Journal of theEuropean Communities L311, 28/11/2001, 67–128.Available at: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32001L0083:EN:NOT(Accessed 24.06.10).

13. Anon. (1996). Directive 2004/27/EC of the Euro -pean Parliament and of the Council of 31 March2004 amending Directive 2001/83/EC on theCommunity code relating to medicinal products forhuman use. Official Journal of the EuropeanCommunities L136, 30/04/2004, 34–57. Availableat: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?ur i=OJ:L:2004:136:0034:0057:EN:PDF(Accessed 05.05.10).

14. Anon. (2009). Botulinum toxin type A for injection.Pharmeuropa 21.3, 361–363.

15. Anon. (2009). Botulinum toxin type B for injection.Pharmeuropa 21.3, 363–365.

16. FDA (undated). The Code of Federal Regulations.Title 21 — Food and Drugs. Chapter I — Food andDrug Administration, Department of Health andHuman Services. (4-1-10 Edition). Available at:http://www.access.gpo.gov/nara/cfr/waisidx_10/21cfrv1_10.html (Accessed 10.08.10).

17. ICH (1996). ICH Harmonised Tripartite Guideline.Quality of Biotechnological Products: StabilityTesting of Biotechnological/Biological Products,Q5C, 10pp. Geneva, Switzerland: InternationalConference on Harmonisation of Technical Require -ments for Registration of Pharmaceuticals forHuman Use. Available at: http://www.ich.org/LOB/media/MEDIA427.pdf (Accessed 24.06.10).

18. ICH (2005). ICH Harmonised Tripartite Guideline.Comparability of Biotechnological/Biological Prod -ucts Subject to Changes in Their ManufacturingProcess, Q5E, 16pp. Geneva, Switzerland: Inter -national Conference on Harmonisation of TechnicalRequirements for Registration of Pharmaceuticals forHuman Use. Available at: http://www.ich.org/LOB/media/MEDIA1196.pdf (Accessed 24.06.10).

19. ICH (1994). ICH Harmonised Tripartite Guideline.Validation of Analytical Procedures: Text andMethodology, Q2(R1), 17pp. Geneva, Switzerland:International Conference on Harmonisation ofTechnical Requirements for Registration ofPharmaceuticals for Human Use. Available at:http://www.ich.org/LOB/media/MEDIA417.pdf(Accessed 24.06.10).

20. McLellan, K., Das, R.E., Ekong, T.A. & Sesardic, D.(1996). Therapeutic botulinum type A toxin: Factorsaffecting potency. Toxicon: Official Journal of theInternational Society on Toxinology 34, 975–985.

21. Sesardic, D., Leung, T. & Gaines Das, R.E. (2003).Role for standards in assays of botulinum toxins:International collaborative study of three prepar-ations of botulinum type A toxin. Biologicals 31,265–276.

22. OECD (2001). OECD Guideline for the Testing ofChemicals. No. 420. Acute Oral Toxicity: Fixed DoseProcedure, 14pp. Paris, France: Organisation forEconomic Co-operation and Development. Availableat: http://lysander.sourceoecd.org/vl=942322/cl=12/nw=1/rpsv/cw/vhosts/oecdjournals/1607310x/v1n4/contp1-1.htm (Accessed 24.06.10).

23. Sesardic, D., McLellan, K., Ekong, T.A.N. & GainesDas, R. (1996). Refinement and validation of an

alternative bioassay for potency testing of thera-peutic botulinum type A toxin. Pharmacology &Toxicology 78, 283–288.

24. Pickett, A., O’Keeffe, R., Judge, A. & Dodd, S.(2008). The in vivo rat muscle force model is a reli-able and clinically relevant test of consistencyamong botulinum toxin preparations. Toxicon:Official Journal of the International Society onToxinology 52, 455–464.

25. Aoki, K.R. (2001). A comparison of the safety mar-gins of botulinum neurotoxin serotypes A, B, and Fin mice. Toxicon: Official Journal of the Inter -national Society on Toxinology 39, 1815–1820.

26. Aoki, K.R. (2002). Botulinum neurotoxin serotypesA and B preparations have different safety marginsin preclinical models of muscle weakening efficacyand systemic safety. Toxicon: Official Journal of theInternational Society on Toxinology 40, 923–928.

27. Sesardic, D. & Gaines Das, R.E. (2009). Botulinumtoxin: Applying the 3Rs to product potency testing,8pp. London, UK: NC3Rs. Available at: http://www.nc3rs.org.uk/downloaddoc.asp?id=853&page=1023&skin=0 (Accessed 24.06.10).

28. Wohlfarth, K., Göschel, H., Frevert, J., Dengler, R.& Bigalke, H. (1997). Botulinum A toxins: Unitsversus units. Naunyn-Schmiedeberg’s Archives ofPharmacology 355, 335–340.

29. Rasetti-Escargueil, C., Jones, R.G., Liu, Y. &Sesardic, D. (2009). Measurement of botulinumtypes A, B and E neurotoxicity using the phrenicnerve-hemidiaphragm: Improved precision with in-bred mice. Toxicon: Official Journal of theInternational Society on Toxinology 53, 503–511.

30. Huber, A., France, R.M., Riccalton-Banks, L.,McLaren, J., Cox, H., Quirk, R.A., Shakesheff, K.M.,Thompson, D., Panjwani, N., Shipley, S. & Pickett, A.(2008). The Intercostal NMJ Assay — A new alterna-tive to the conventional LD50 assay for the determi-nation of the therapeutic potency of botulinum toxinpreparations. ATLA 36, 141–152.

31. Pellett, S., Tepp, W.H., Clancy, C.M., Borodic, G.E.& Johnson, E.A. (2007). A neuronal cell-based botu-linum neurotoxin assay for highly sensitive andspecific detection of neutralizing serum antibodies.FEBS Letters 581, 4803–4808.

32. Podrygajlo, G., Tegenge, M.A., Gierse, A., Paquet-Durand, F., Tan, S., Bicker, G. & Stern, M. (2009).Cellular phenotypes of human model neurons (NT2)after differentiation in aggregate culture. Cell &Tissue Research 336, 439–452.

33. Paquet-Durand, F. & Bicker, G. (2007). Humanmodel neurons in studies of brain cell damage andneural repair. Current Molecular Medicine 7,541–554.

34. Podrygajlo, G., Song, Y., Schlesinger, F., Krampfl,K. & Bicker, G. (2010). Synaptic currents and trans-mitter responses in human NT2 neurons differenti-ated in aggregate culture. Neuroscience Letters 468,207–210.

35. Ekong, T.A., Feavers, I.M. & Sesardic, D. (1997).Recombinant SNAP-25 is an effective substrate forClostridium botulinum type A toxin endopeptidaseactivity in vitro. Microbiology 143, 3337–3347.

36. Sesardic, D., Jones, R.G., Leung, T., Alsop, T. &Tierney, R. (2004). Detection of antibodies againstbotulinum toxins. Movement Disorders: OfficialJournal of the Movement Disorder Society 19,Suppl. 8, S85–S91.

37. Gaines Das, R.E., Heath, A.B., Martin, H. &

ZEBET Expert Meeting 329

Sesardic, D. (1999). Validation of in vitro assays forbotulinum toxin: a case study. Developments inBiological Standardization 101, 267–276.

38. Jones, R.G.A., Ochiai, M., Liu, Y., Ekong, T. &

Sesardic, D. (2008). Development of improvedSNAP25 endopeptidase immune-assays for botu-linum type A and E toxins. Journal of Imm -unological Methods 329, 92–101.

330 S. Adler et al.