Laboratory Animals the INTERNATIONAL JOURNAL of LABORATORY ANIMAL SCIENCE, MEDICINE, TECHNOLOGY and WELFARE

Volume 54 Number 2 April 2020 ISSN 0023-6772

Laboratory Animals THE INTERNATIONAL JOURNAL OF LABORATORY ANIMAL SCIENCE, MEDICINE, TECHNOLOGY AND WELFARE

Official Journal of AFSTAL, DALAS, ECLAM, ESLAV, FELASA, GV-SOLAS, ILAF, LASA, NVP, SECAL, SGV, SPCAL

Published on behalf of Laboratory Animals Ltd. journals.sagepub.com/home/lan by SAGE Publications Ltd.

Volume 54 Number 2 April 2020 Contents

Working Party Report Genetic quality assurance and genetic monitoring of laboratory mice and rats: FELASA Working Group Report 135 F Benavides, T Ru¨licke, J-B Prins, J Bussell, F Scavizzi, P Cinelli, Y Herault and D Wedekind

Editorial Collection on score sheets, severity assessment and humane end points: Invitation to submit 149 P Jirkof, G Jarvis and B Riederer

Severity Assessment in Animal-based Research Impulse for animal welfare outside the experiment 150 L Lewejohann, K Schwabe, C Ha¨ger and P Jirkof

Original Articles Diet-regulated behavior: FVB/N mice fed a lean diet exhibit increased nocturnal bouts of aggression between littermates 159 MM Murph, S Liu, W Jia, H Nguyen, MA MacFarlane, SS Smyth, SS Kuppa and KK Dobbin

Combination of ketamine and xylazine with opioids and acepromazine in rats: Physiological changes and their analgesic effect analysed by ultrasonic vocalization 171 J Alema´n-Laporte, LA Bandini, MSA Garcia-Gomes, DA Zanatto, DT Fantoni, MAA Pereira, PE Navas-Sua´rez, TB Kirsten, RR Jimenez, G Alvarado and CC Mori

Examining compliance with ethical standards for animal research: is there a need for refinement? A qualitative study from northern Europe 183 A Brønstad and P Sandøe

News 3Rs – Reduce Reuse Recycle 194 J-P Mocho

El XV Congreso de la SECAL (Sociedad Espan˜ola para las Ciencias del Animal de Laboratorio) tuvo lugar en Sevilla del 6 al 8 de noviembre 196 E Hevia and CO Pintado

Thanks to Reviewers Thanks to Reviewers 198

Each of our customers is different

For each customer’s requirement we have a solution

Your experienced Swiss partner for all laboratory animal diets

Quality and service − Successful together

KLIBA NAFAG T. +41 61 816 16 16 Rinaustrasse 380 www.kliba-nafag.ch CH-4303 Kaiseraugst [email protected] Flawless precision

The extraordinary quality of our surgical and microsurgical instruments is the result of our relentless attention to detail. Every instrument we sell is designed to exacting specifications, forged from the strongest, lightest materials available, and tested to ensure precision performance.

FINE SURGICAL INSTRUMENTS FOR RESEARCH™ VISIT US AT FINESCIENCE.DE OR CALL +49 (0) 6221 90 50 50 ...... 125 Laboratory Animals Subscription information Annual subscription (2020) including postage: AALAS Member, (print and electronic) [£254/US$470]. Combined Institutional Rate (print and electronic) [£339/US$662]. Electronic only and print Editorial Board only subscriptions are available for institutions at a discounted Editor-in-Chief B Riederer rate. Note VAT is applicable at the appropriate local rate. Deputy Editors G Jarvis, P Jirkof Visit sagepublishing.com for more subscription details. To activate your subscription (institutions only) visit Section Section Editors online.sagepub.com. Abstracts, tables of contents and contents alerts are available on this site free of charge for all. Anaesthesia, Analgesia, M Leach, P Foley, Student discounts and single issue rates are available from SAGE Pain & Stress P Hedenqvist Publications Ltd, 1 Oliver’s Yard, 55 City Road, London EC1Y 1SP, UK, tel. þ44 (0)20 7324 8500, email [email protected] Anatomy and Neuroscience B Riederer, S Wells (neuro) and in North America, SAGE Publications Inc, PO Box 5096, Aquatic Organisms K Finger-Baier (fish), Thousand Oaks, CA 91320, USA. JP Mocho, M Crim Behaviour D Preissmann, M Gyger, Advertisement Managers L Lewejohann Biostatisics & Experimental R-D Gosselin, H Wu¨ rbel PRC Associates Ltd, 1st Floor Offices, 115 Roebuck Road, Design Chessington, Surrey KT9 1JZ, UK; Tel: +44 (0) 20 8337 3749; Fax: Education P Vergara, C Tho¨ne-Reineke +44 (0) 20 8337 7346; Email: [email protected] Imaging Techniques L van der Weerd, J Tremoleda Laboratory Animals Ltd Large Animal Models M Jensen-Waern, D Anderson, Laboratory Animals Ltd is a company limited by guarantee and has T Morris no share capital. The Memorandum of Association obliges the Management of Animal Facilities J-B Prins, M Dennis company to apply all its resources to the advancement of public Molecular & Genetic Engineering T Ruelicke, P Cinelli Nutrition and Diets G Tobin, T Nortey education in laboratory animal science, technology and welfare. It is a registered charity (Registered Charity Number 261047) and Pathology & Microbiology P Clements, D Salvatori, none of its directors may receive any fee or remuneration. A Bleich Physiology & Clinical Chemistry M Sommers Registered Office: Primates G Rainer, P Honess, C Witham Laboratory Animals Ltd, 44 Springfield Road, Horsham, West 3Rs & Ethics G Griffin, A Olsson Sussex, RH12 2PD, UK Reproductive Biology H Hedrich, B Pintado, C Gilbert Council of Management Small Animal Models M Berard, J-B Prins (temporary), S Wells Chairman J-B Prins Surgical Procedures D Bouard, R Tolba Secretary E Weir Systematic Review M Ritskes-Hoitinga, Treasurer J Gregory BS Kousholt L Antunes B Riederer Toxicology F Rutten K Applebee A Ritchie Veterinary Medicine E Rivera, J Sanchez-Morgado, M Berard M Ritkes-Hoitinga L Whitfield, N Ezov TP Rooymans N Kostomitsopoulos C Gilbert A Shortland Special Issue Microbiota guest editor: J Guillen M Wilkinson Axel Kornerup-Hansen and J Helppi S Wells Craig Franklin P Nowlan B Zevnik J Orellana Comment and correspondence relating to editorial matters may be sent to the Chairman of the Editorial Board by email: laeditorial@- ß 2020 Laboratory Animals Ltd. Apart from any fair dealing for the sagepub.co.uk; or post: LAL, PO Box 373, Eye, Suffolk, IP22 9BS, purposes of research or private study, or criticism or review, as permitted UK. under the UK Copyright, Designs and Patents Act, 1988, no part of this See also http://www.lal.org.uk publication may be reproduced, stored, or transmitted, in any form or by Laboratory Animals, (ISSN 0023-6772) is published and distributed any means, without the prior permission of the publishers, or in the case bimonthly (February, April, June, August, October, December) in of reprographic reproduction in accordance with the terms of licences both print and electronic form by SAGE Publications Ltd, issued by the Copyright Licensing Agency in the UK, or in accordance 1 Oliver’s Yard, 55 City Road, London EC1Y 1SP, UK. with the terms of licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside All manuscripts submitted for publication should the terms stated here should be sent to SAGE at the address. be prepared in accordance with the Guidelines Whilst every effort is made to ensure that no inaccurate or misleading for Authors which can be found online at data, opinion or statement appears in the journal, Laboratory Animals Ltd journals.sagepub.com/home/lan. Please submit your paper wish to make it clear that the data and opinions appearing in the articles online at mc.manuscriptcentral.com/la, and advertisements herein are the responsibility of the contributor or contributions for news items can also be made via advertiser concerned. Accordingly, Laboratory Animals Ltd and their offi- manuscript central. cers and agents accept no liability whatsoever for the consequences of any such inaccurate or misleading data, opinion or statement. Printed in Great Britain ...... 126 AFSTAL Hon. Secretary The Netherlands Jean-Philippe Mocho (www.proefdierkunde.nl) Association Franc¸aise des Sciences et Techniques de I’Animal de Laboratoire Secretariat: PO Box 372, Eye, IP22 9BR, UK SECAL President (www.felasa.eu) Sebastian Paturance Sociedad Espanˇ ola para las Ciencias del Animal de Laboratorio Vice President (Spanish Society for Laboratory Animal Elodie Bouchoux GV-SOLAS Science) Secretariat: 28, rue Saint Dominique, 75007, Gesellschaft fu¨r Versuchstierkunde (Society for Laboratory Animal Science) President Paris, France Isabel Blanco Gutierrez (www.afstal.com) President Bettina Kraenzlin Vice President Juan Rodriguez Cuesta DALAS Secretary Nicole Linklater Dutch Association for Laboratory Animal Secretary Faculty of Biology Science Julia M. Samos Jua´rez Philipps University President Karl-von-Frisch Str. 8 Treasurer Catriene Thuring 35043 Marburg Viviana bisbal Velasco Germany Secretary (www.gv-solas.de) Secretariat: c/Maestro Ripoll, 8, Linda Derks 28006 Madrid, Energieweg 19 Spain 6541 CW Nijmegen (www.secal.es) ([email protected]) ILAF Israeli Laboratory Animal Forum ECLAM President SGV Amir Rosner Schweizerische Gesellschaft fu¨r European College of Laboratory Animal Versuchstierkunde Medicine Secretary David Castel Socie´te´ Suisse pour la Science des Animaux de President Neufeld Cardiac Research Laboratoire (Swiss Laboratory Animal Science Patricia Hedenqvist Institute Association) Sheba Medical Center Secretariat: Janet Rodgers, 266 President Tel Hashomer 52621 Banbury Road, No. 314 Oxford Dr. Birgit Ledermann Israel OX2 7DL, UK (www.ilaf.org.il) Secretary Dr. med. vet. Andrina Zbinden Faculty of Science and Medicine ESLAV University of Fribourg European Society of Laboratory Animal LASA Ch. du Muse´e8 Veterinarians Laboratory Animal Science Association CH-1700 Fribourg, Switzerland (www.naturalsciences.ch/organisations/sgv) President President Peter Glerup Anne-Marie Farmer

Honorary Secretary Secretary General SPCAL Massimiliano Bardotti Miles Maxwell Sociedade Portuguesa de Cieˆ ncias em Animais de Laborato´ rio Honorary Secretary PO Box 524, Hull, (Portuguese Society for Laboratory Animal Frederic Decrock HU9 9HE, UK Science) (www.lasa.co.uk) Secretariat President c/o Decrock, 78,bd Gallieni Isabel Vito´ria Figueiredo 92130 Issy les Moulineaux, France (http://eslav.org) NVP Vice-President Ricardo Afonso Nederlandse Vereniging voor Proefdierkunde Secretary FELASA (Dutch Association for Laboratory Animal Catarina Pinto Reis Federation of European Science) Secretariat: Laborato´rio de Laboratory Animal Science Associations President Farmacologia President Martje Fentener van Vlissingen Faculdade de Farma´cia Hanna-Marja Voipio Largo de D. Dinis Secretary 3000 Coimbra President-elect Jan Langemans Portugal Ana Santos BPRC (www.spcal.pt) Past President Lange Kleiweg 139 Heinz Brandstetter 2288 GJ Rijswik conventional systems • air flow systems • large animal systems • enrichment and bedding • transport systems • cleaning systems • handling

Best Quality for BSL3/BSL4

ZOONLAB is one of the leading manufacturers and suppliers of complete solutions for scientific animal husbandry.

Did you know? – ZOONLAB develops best customized solutions: for example high quality isolators for tiny mice and clever monkeys for high safety and comfortable sluice solutions.

ZOONLAB cares for animal welfare, human safety and the environment.

Learn more about us – we are listening to all your need. [email protected] or +49 23 05 97 30 40

www.zoonlab.de

ZOL_Anz_Laboratory Animal_210x280_RZ_20200210.indd 1 10.02.20 12:01

Charlost tte 71 AALAS NATIONAL MEETING OCTOBER 25 - 29, 2020

Join us for the 71st AALAS National Meeting in Charlotte, North Carolina. Each fall since 1950, the American Association for Laboratory Animal Science has held its annual National Meeting. During the five days of the meeting, members and nonmembers come together to enjoy the workshops, lectures, poster sessions, and exhibits. The program is designed to have topics relevant to the entire membership. Exhibitors have an opportunity to interact with AALAS members from the academic community, research institutions, government organizations, and commercial companies. The AALAS National Meeting is the largest gathering in the world of professionals concerned with the production, care, and use of laboratory animals.

American Association for Laboratory Animal Science Phone: (901) 754 - 8620 Fax: (901) 753 - 0046 [email protected] www.aalas.org

JOIN US IN CHARLOTTE, NC OCTOBER 25 - 29, 2020

Untitled-1 1 1/17/2020 9:10:54 AM A Clear Path to Wellness with Innovative Gelled Diets

A range of highly palatable, nutritionally complete food and water gelled diets specially formulated to give optimum health to your animals. Supplied exclusively in the UK by LBS.

These special dietary products ensure animals’ survival, encourage reliable research outcomes and are easily fed to animals by placing the cup at cage level. They allow your animal to remain healthy and hydrated even through impaired conditions. Available in 1oz (40/case) or 2oz (60/case).

• Nutra-GelTM - a tried and true bacon flavoured gelled diet, based on the AIN-93G formulation. It serves as a complete food and water source for rodents.

• Pure-Water GelTM - a clear gelled water source that can be used as a complete water replacement or supplement for mobility impaired research animals.

• Electro-GelTM - an oral electrolyte hydration gel available in orange or strawberry flavours, packed with vitamins and minerals to promote recovery.

Contact the experts: Tel: +44 (0)1293 827940 Email: [email protected] www.lbs-biotech.com

ssniff

Spezialdiäten

“The Best or Nothing”

Standard feeds – cereal based (chow) for all animal species Special / Purified diets Medicated feed

Perfect Animal – Diet models Purified diets • for every research purpose (e.g. diabetes) Obesity, metabolic syndrome, NIDDM, atherosclerosis, (high fat and/or cholesterol)

Hypertension and chronic kidney disease (± adenine) Chronic liver diseases (steatohepatosis, ASH, NASH) Diets with deficits or excesses of certain nutrients

Test compound diets with

Doxycycline Tamoxifen (flavoured diets)

Customers’ statins

Customers’ compounds for toxicity or PK studies ssniff has been the test manufacturer for several CROs Diets with hormone supplementation (17-βββ-estradiol, DHT)

Spezialdiäten GmbH Ferdinand-Gabriel-Weg 16 batch sizes as small as 1 kg DE-59494 Soest - Germany +49-(0)2921-9658-0 [email protected] / www.ssniff.com ssniff – dedicated to r esearch for decades

Profit by GVG’s high resolution STR genotyping platform

Fully characterise laboratory rodent genetics according to „Genetic quality assurance and genetic monitoring of laboratory mice and rats: FELASA Working Group Report“

 Verify authenticity and uniformity of inbred strains and substrains  Detect potential genetic contamination  Describe precise genetic composition of genetically modified lines including Y chromosomal origin  Monitor genetic drift  Obtain qualified breeding advice to speed up congenic backcrossing projects  Quantify and control genetic diversity of outbred strains

www.gvg-gm.de [email protected] Working Party Report Laboratory Animals 2020, Vol. 54(2) 135–148 Genetic quality assurance and genetic ! The Author(s) 2019 Article reuse guidelines: monitoring of laboratory mice and rats: sagepub.com/journals-permissions DOI: 10.1177/0023677219867719 FELASA Working Group Report journals.sagepub.com/home/lan

Fernando Benavides1 , Thomas Ru¨licke2 , Jan-Bas Prins3,4, James Bussell5, Ferdinando Scavizzi6, Paolo Cinelli7, Yann Herault8,9 and Dirk Wedekind10

Abstract Genetic quality assurance (QA), including genetic monitoring (GeMo) of inbred strains and background characterization (BC) of genetically altered (GA) animal models, should be an essential component of any QA programme in laboratory animal facilities. Genetic quality control is as important for ensuring the validity of the animal model as health and microbiology monitoring are. It should be required that studies using laboratory rodents, mainly mice and rats, utilize genetically defined animals. This paper, presented by the FELASA Working Group on Genetic Quality Assurance and Genetic Monitoring of Laboratory Murines, describes the objectives of and available methods for genetic QA programmes in rodent facilities. The main goals of any genetic QA programme are: (a) to verify the authenticity and uniformity of inbred stains and substrains, thus ensuring a genetically reliable colony maintenance; (b) to detect possible genetic contamination; and (c) to precisely describe the genetic composition of GA lines. While this publication focuses mainly on mouse and rat genetic QA, the principles will apply to other rodent species some of which are briefly mentioned within the context of inbred and outbred stocks.

Keywords animal facilities, genetics, quality assurance/control, refinement, rodents

Date received: 19 March 2019; accepted: 13 July 2019

1Department of Epigenetics and Molecular Carcinogenesis, The University of Texas, MD Anderson Cancer Center, USA 2Institute of Laboratory Animal Science, University of Veterinary Standardized laboratory rodents Medicine, Vienna, Austria 3The Francis Crick Institute, London, UK Inbred strains 4Leiden University Medical Centre, Leiden, The Netherlands 5 The International Committee on Standardized Genetic Biomedical and Veterinary Services Department, University of Oxford, Oxford, UK Nomenclature for Mice and The Rat Genome 6National Research Council (IBCN), Rome, Italy Nomenclature Committee considers a strain inbred 7Department of Trauma Surgery, University of Zurich, Zurich, Switzerland if it has been propagated by systematically mating 8Universite´ de Strasbourg, CNRS, INSERM, Institut de Geńe´tique brothers to sisters (or younger parent to offspring) for Biologie Molećulaire et Cellulaire, IGBMC, Illkirch, France 9Universite´ de Strasbourg, CNRS, INSERM, Institut Clinique de la 20 or more consecutive generations, and individuals of Souris, CELPHEDIA-PHENOMIN-ICS, Illkirch, France the strain can be traced to a single ancestral pair at the 10Institute of Laboratory Animal Science, Hannover Medical twentieth or subsequent generation. School, Hannover, Germany

At this point, animals within the population will aver- Corresponding author: Fernando J. Benavides, Department of Epigenetics and Molecular age 2% residual heterozygosity, and the individuals 1 Carcinogenesis, The University of Texas MD Anderson Cancer may be regarded as genetically identical (isogenic). Center, 1808 Park Road 1C, Smithville, TX, 78957, USA. However, it has been estimated that 24 generations of Email: [email protected] 136 Laboratory Animals 54(2) sib-mating are needed to reach a heterozygosity rate- the large colonies of commercial breeders. Because < 1% and 36 generations to reach (almost) complete outbred colonies, like human populations, are hetero- isogeneity.2 geneous, they are frequently used in toxicology and Isogeneity implies histocompatibility, meaning the pharmacology research.6 However, several geneticists strains are syngeneic. Syngeneic animals will perman- have disputed this use and have criticized studies in ently accept tissue transplants from any individual of which outbred mice were used inappropriately, wasting the same strain and sex. Unlike cloned animals and both animal lives and precious resources in suboptimal monozygotic twins (which are 100% identical for all experiments.7 genomic loci), inbred rodents, besides being isogenic, are also homozygous at almost all genomic loci. Other standardized strains of mice and rats Overall, each inbred strain represents a unique, although fortuitous, assortment of alleles.3 If a strain F1 hybrids result from the outcross of two separate were to be remade from scratch, using the same foun- inbred strains and are heterozygous at all loci for ders, after the same 20 generations of inbreeding it which the parental strains harbour different alleles. F1 would create a genetically distinct strain due to the littermates are genetically identical and are histocom- random assortment and fixation of alleles. Baseline patible. Congenic strains are produced by crossing two phenotypic data for the most common inbred mouse strains: the donor strain that carries the allele or strains are available through a coordinated inter- chromosomal region of interest, and the recipient or national effort initiated by The Jackson Laboratory background strain that will receive the locus of interest. and implemented through The Mouse Phenome F1 offspring generated by crossing donors and recipi- Database (http://phenome.jax.org/).4 An example of ents are then backcrossed to the recipient strain. baseline phenotypic data is presented in Offspring that carry the allele of interest are identified Supplementary Tables 1A and 1B. The Mouse and again backcrossed to the background strain. This Genome Informatics (MGI) website5 provides a list, process is typically repeated for 10 or more successive compiled by Dr Michael Festing (http://www.informat- generations (Figure 1), unless marker-assisted back- ics.jax.org/external/festing/search_form.cgi), of 420 crosses (speed congenics) are used. Repeated backcross- inbred mouse and 230 inbred rat strains (some of ing results in the chromosomes of the background which have been lost or terminated), along with brief strain progressively replacing those of the donor descriptions. The list includes widely used inbred mouse strain, except for a chromosomal region that carries strains: A/J, BALB/c, C3H/He, C57BL/6, DBA/2, the allele of interest. FVB/N and others; and rat strains: ACI, BN, F344, LE and WKY. Genetically altered (GA) rodents

Outbred stocks Before presenting the different types of GA rodents, it is worth mentioning that there are basically two differ- Outbred stocks are populations of laboratory animals ent approaches to characterizing gene function. that differ from inbred strains in that they are genetic- Forward genetics (from phenotype to genotype) aims ally heterogeneous. Compared with inbred strains or to characterize the gene alteration that is responsible F1 hybrids, the genetic constitution of a given animal, for a specific mutant phenotype (typically from spon- taken randomly from an outbred stock, is not known a taneous or chemically-induced mutations). Reverse gen- priori. However, all of the animals in the group share etics is the opposite approach and aims to characterize group characteristics (identity), such as being albino the function of a gene by analysing the consequences (at (although not all outbred mice or rats are albino), the phenotypic level) of alterations normally engineered good breeders and relatively tame compared to other by researchers at the DNA level. This section intro- strains; features that make these animals very popular duces the four basic types of GA rodents, those created as foster mothers for assisted reproductive techniques. by: (a) pronuclear microinjection, (b) vector- mediated Examples of outbred stocks of mice are ICR (CD-1), transgenesis (c) homologous recombination in embry- CFW and NMRI (all derived from the original ‘Swiss’ onic stem (ES) cells, (d) gene editing nucleases, and (e) mice imported to the USA by Clara J. Lynch in 1926) either chemically induced or spontaneous mutations. and (non-Swiss) CF-1. Examples of outbred rat stocks Detailed descriptions of the technologies used to are Sprague Dawley (SD), Wistar (WI) and Long- create GAs have been published.8 Before selecting a Evans (LE). Since outbred stocks are not genetically gene-editing technique to create a genetically modified defined, quality control is commonly based on assessing animal, it is important to check an appropriate data- expected phenotypic traits, such as coat colour, growth base such as those hosted by The Jackson Laboratories and reproductive characteristics, based on data from and the International Mouse Phenotyping Consortium Benavides et al. 137 as to whether a suitable animal model already exists and off. Examples of this strategy are the Tet-on and (see Supplementary Table 2 for the complete list of Tet-off expression systems. In these systems, transcrip- online resources for laboratory mouse and rat strains). tion of a given transgene is placed under the control of a tetracycline-controlled trans-activator protein, which Transgenesis by pronuclear microinjection can be regulated, both reversibly and quantitatively, by exposing the transgenic mice to either Tetracycline (Tc) Transgenic mice were introduced in the early 1980s9 or one of its derivatives, such as Doxycycline (Dox). and were the first transgenic animals. It is advisable to Both Tet-on and Tet-off are binary systems that require use the term ‘transgenic’ only for animals whose gen- the generation of double transgenic (bigenic) mice.11 omes have been altered by the random insertion of DNA. (There are numerous terms used to describe gen- Vector-mediated transgenesis etic changes in animals: genetically engineered mice (GEM) or genetically modified mice (GMM) are typic- Alternative methods for transgenesis by random inte- ally used to describe any type of genetic modification in gration are based on vectors of different origin. Most the mouse. We use the term GA rodent here to also important and very efficient are retroviral/lentiviral vec- include those carrying spontaneous or chemically tors12 and transposons.13 Also pre-treated spermatozoa induced mutations, and ‘line’ instead of ‘strain’ for have been successfully used as vectors in combination GA rodents.) Transgenic rodents are almost exclusively with ICSI (intracytoplasmic sperm injection).14 Each created by the pronuclear microinjection of foreign technique has advantages and disadvantages and the DNA fragments directly into one of the two pronuclei corresponding principle of transgene integration may of one-cell embryos (zygote), a technique that is still affect the quality of the resulting GA models. Viral vec- widely used. In this process of additive transgenesis, tors and transposons for instance integrate as a single the microinjected transgene randomly integrates into copy, however multiple integrations, randomly distrib- the genome as a single copy or more often as a conca- uted in the genome, are not uncommon. Major con- temer with variable copy number. The mouse and rat cerns exist regarding the impact of sperm-mediated models created with this system typically express or, in gene transfer on the sperm genetic material, possibly the resultant concatemer, overexpress a transgene induced by the pre-treatment of spermatozoa.15 placed under the control of a tissue-specific, developmental-stage-specific, or ubiquitous promoter (along Targeted mutagenesis by homologous with other regulatory elements), all contained in the recombination using ES cells transgene DNA construct. The recommended generic symbol for a transgenic Another important technology utilizes murine ES cell insertion is Tg. The founder transgenic animals are lines. ES cells are undifferentiated, pluripotent, embry- hemizygous for the DNA segment and are designated onic cells derived from the inner cell mass of pre- Tg/0. Transgenes are extra segments of DNA that have implantation blastocysts that can participate in forming no corresponding ‘wild-type’ sequence in the unmodi- the germ-cell lineage of chimeric mice, an indispensable fied homologous chromosome in hemizygous animals, step in generating founder mice carrying the targeted that is why the use of ‘0’ instead of ‘þ’ (typically used to mutation. Historically, the first ES cell lines were denote wild-type alleles) is recommended. Each trans- derived from embryos of the 129 family (129S2, genic line generated via random integration creates a 129P3, etc.), that is inbred strains originally bred for unique animal model and each putative founder must the isolation of embryonic carcinoma (EC) cells. Today be developed independently. Traditionally, to distin- ES cell lines are available from many mouse strains and guish between homozygous (Tg/Tg) and hemizygous those of the C57BL/6N origin have become widespread (Tg/0) mice, the mouse of interest was crossed to a and are often selected for trans-national projects (e.g. non-transgenic partner and the progeny were statistic- EUCOMM). ally analysed for Mendelian segregation of the In cases where constitutive null alleles lead to com- transgene. A more modern technique uses quantitative plex phenotypes, reduced viability, or have other draw- real-time polymerase chain reaction (qPCR) to distin- backs, conditional alleles may be used, allowing one to guish hemizygous from homozygous transgenic mice.10 control the time and tissue where a gene is turned off, In order to achieve a pure genetic background (recom- typically using the Cre/loxP system.16 Production of mended), the transgene must be introduced into conditional KOs requires two independent lines: one embryos derived from an inbred strain. providing a source of Cre recombinase, an enzyme A later improvement on the constructs used in the derived from bacteriophage P1, in the tissue under transgenesis approach was the introduction of inducible study, and another containing loxP (locus of X-ing systems in which transgene expression can be turned on over P1) sites flanking the DNA segment of interest 138 Laboratory Animals 54(2)

Figure 1. This scheme represents the successive steps in the establishment of a congenic strain. The initial step is a cross between the donor strain (albino in the example) carrying the gene of interest (e.g. a targeted gene or a transgene) and a recipient or background strain (black in the example). At each generation, a breeder carrying the gene of interest (*) is backcrossed to a partner of the recipient strain (genetically linked genes are transferred with it and the size of the introgressed fragment can be many thousands or millions of bases, and include many genes). The degree of grey colour indicates that, after each backcross generation, the offspring have an increased amount of the background genome (average percentage is indicated in each N generation). When the modified gene is not resulting in an easily recognizable phenotype (e.g. skin or behavioural changes), molecular genotyping is necessary to select the carrier (heterozygous) mice. that needs to be crossed to generate double mutant briefly described here, provide ES cell-independent mice. The Cre enzyme cuts and recombines the methods to create targeted mutations in laboratory ‘floxed’ DNA at loxP sites. The Cre transgene can be mice, rats and other species. made inducible, adding more sophistication to the To make mutations using zinc-finger nucleases system. The tamoxifen-inducible CreERT2 which can (ZFN), two complementary and sequence-specific be activated in a spatio-temporal manner by adminis- multi-finger peptides containing the FokI nuclease tration of tamoxifen, is widely used.17 The Cre-loxP domain must be designed. Each peptide is designed to strategy can also be used to regulate the expression of recognize a specific DNA sequence spanning 9–18 base reporter genes. For example, the lacZ gene can be pairs (bp) on either side of a 5–6 bp sequence, which driven by a ubiquitous promoter (e.g. Rosa 26) with a defines the targeted region. When injected into a pro- floxed stop sequence, containing several terminator nucleus or cytoplasm of zygotes, the ZFN assemblies codons inserted between the promoter and the lacZ bind tightly, one on each strand, on both sides of the coding sequence. target site. The dimerized FokI endonuclease then creates double strand DNA breaks (DSBs) triggering cel- Gene editing using nucleases lular mechanisms to repair the damage. Damage is normally repaired by either homology-directed repair Over the last 10 years, a number of new techniques have (HDR) or non- homologous end joining (NHEJ). been developed for the production of targeted muta- HDR requires a homologous template to guide the tions using engineered nucleases. These techniques, repair and thus re-establishes the original sequence. Benavides et al. 139

NHEJ is much less precise and cause nucleotide dele- through the observation of an abnormal phenotype, tions that lead to frame shifts that create potential loss- present several advantages. First and foremost, they of-function or truncation mutations. Mice and rats are produced at virtually no cost and are generally carrying null alleles or sequence-specific modifications freely available. Second, they usually have an obvious have already been produced using ZFN technology.18 phenotype, as they are identified based on observation. Like ZFNs, transcription activator-like effector nucle- Third, spontaneous mutations represent a great variety ase (TALEN) technology involves the combination of a of molecular events, such as deletions, insertions and nonspecific DNA endonuclease fused to a DNA-bind- point mutations, generating not only loss-of-function ing domain, but can be more easily engineered (com- alleles but also hypomorphic and hypermorphic alleles. pared to ZFN) to target a particular DNA sequence. Finally, mutations arise in a variety of backgrounds The CRISPR (clusters of regularly interspaced short including inbred strains and outbred stocks. Several palindromic repeats)/Cas system, commonly imple- spontaneous mutations have provided rodent models mented as CRISPR/Cas9, is based on a primitive for human conditions. These include classical muta- defence mechanism that allows bacteria and archaea tions such as, nude (Foxn1nu), scid (Prkdcscid), hairless to fight against infection from viruses, plasmids and (Hrhr), diabetes (Leprdb), obese (Lepob) and X-linked phages.19 CRISPR-based guide RNAs (gRNAs) are muscular dystrophy (Dmdmdx) in the mouse; and the designed to target a Cas endonuclease to cut DNA at mutations behind the Rowett nude (Foxn1rnu) and the desired site through RNA-guided DNA cleavage. Zucker diabetic fatty (Leprfa) models in the rat. The RNA-guided endonucleases can be engineered to The discovery of the extraordinary virtues of the cleave virtually any DNA sequence by appropriately alkylating agent N-ethyl-N-nitroso urea (ENU) as a designing the gRNA, for example to generate KO mutagen was a milestone in the history of mouse gen- mice.20 CRISPR/Cas technology has several advan- etics. Researchers using ENU have generated and pro- tages over ZFNs and TALENs. The main advantage pagated numerous mutant alleles for protein-coding is the ease of design and the flexibility of using a genes, thus establishing a precious tool for genome sequence-specific RNA interacting with the Cas annotation. Because ENU typically creates point muta- enzyme instead of a complex sequence-specific protein tions, it has been widely used in forward genetic (DNA-binding domain) fused to a nuclease. Also, screens. The major drawback of ENU-induced muta- mutations in multiple genes can be generated in a genesis is that it creates random mutations rather than single step by injecting mice with multiple gRNAs targeted mutations. Several projects have been under- that simultaneously target different genes.21 Such multi- taken to systematically and extensively phenotype the plex gene editing has been successful in cells, as well as offspring of ENU-mutagenized males. Large ENU mouse and rat embryos.20 CRISPR/Cas9 has been used mutagenesis programmes have been conducted in to create insertions, deletions and point mutations. The Germany, England and the USA.25 system is highly flexible, fast and efficient, and is revo- 22 lutionizing genomic engineering in mammals. It Quality assurance and exchange allows making KO and KI lines in any genetic back- of GA-rodents ground. DNA can be electroporated (with size restrictions) or injected into either the cytoplasm or pronuclei What to ensure after (in-house) generation or upon of 1-cell or 2-cell stage embryos, thus avoiding the use arrival? The possibility of crossing different GA lines of ES cells and chimeras. However, as each engineered combined with the increasing complexity of targeting animal is unique, this technology requires extensive approaches has greatly increased the number of avail- sequence analysis to characterize multiple putative able GA models. The need to cross different GA lines founders to ensure the presence of the desired mutation together for a particular study generates additional and the absence of undesired on- and off-target muta- complexity, especially at the genetic background level. tions or unpredictable larger genome alterations,23,24 Many mutants have been and are still generated on a while also identifying mosaic founders (G0). Once iden- hybrid genetic background. Therefore, it is essential to tified, the selected founder should be bred with wild- keep adequate records of detailed information for all type animals to evaluate transmission of the mutation. genetically modified strains. This information must be transferred with the strain to all collaborators and Spontaneous and chemically-induced users. The most important information includes the mutations correct strain name, a complete description of the mutation, the genetic background of the animals, a A list of GA rodent types is not complete without genotyping protocol and observed phenotypic changes. including both spontaneous and chemically-induced Together, these provide the minimum information for mutations. Spontaneous mutations, generally identified the recommended ‘rodent-passport’, and several forms 140 Laboratory Animals 54(2) have been designed for cataloguing this information. alternative DNA sequences (alleles) at a locus among We recommend the data sheet developed by the individuals, groups, or populations, at a frequency FELASA Working group on the refinement of methods >1%. Two types of genetic markers are commonly for genotyping genetically modified rodents.26 used in association studies and genetic quality control: Every mutant strain name must provide precise microsatellites and single nucleotide polymorphisms information on the affected gene, the type of mutation (SNPs) (see ‘Marker systems’ below). and the genetic background. For in-house generated strains, one must provide a specific Institute for Genetic drift and the generation Laboratory Animal Research Laboratory (ILAR) of substrains Code Registration for the laboratory where the mutant originated. An overview on the importance of While permanent inbreeding effectively eliminates a nomenclature can be found in the ‘FELASA guidelines proportion of new mutant alleles, another undetected for the production and nomenclature of transgenic fraction may become progressively fixed in the homo- rodents’.27 A name designed according to the inter- zygous state, replacing the original allele, a process national nomenclature rules is the only means to unam- known as genetic drift. Genetic drift contributes inex- biguously distinguish strains from each other. This is orably to strain divergence and the generation of sub- important when the same strain is held in different facil- strains when the same strain is propagated ities around the world and/or they are listed in archives independently in different places.28 Examples of and databases. Further, it is imperative that strains be mouse substrains are abundant, for example there are properly described in publications using a universal c. 10 documented BALB/c substrains and c. 15 C57BL/ nomenclature. Without a common nomenclature, it 6 substrains including the J and N substrains from The becomes impossible to accurately communicate scien- Jackson Laboratory (Jax) and the National Institutes tific results. Vague or incomplete names create errors of Health (NIH), respectively.29 In the same way, many rendering experiments irreproducible. rat inbred strains present at least two substrains, for example SHR has at least four substrains (including Origin and consequences of genetic SHR/Ola and SHR/NCrl), and WKY and F344 have variation at least three substrains each. Substrain variability has been confirmed by sequencing analysis for these rat A serious challenge facing rodent animal facilities is substrains,30 with WKY showing the highest degree keeping inbred strains genetically pure and GA lines of substrain variation (this is in part due to the on a defined background. Changes in the genetic con- supply of the model prior to the prescribed 20 gener- stitution of inbred strains can be produced by (a) con- ation inbreeding requirement). tamination by accidental outcrosses and (b) genetic drift due to residual heterozygosity or fixation of de Undesirable passenger mutations novo spontaneous mutations. Mutations that are hidden in the genomes of substrains Genetic contamination or GA lines and can affect the outcome of an experiment are sometimes referred to as passenger muta- The accidental mating of individuals from one inbred tions.31 There are many examples in the literature strain with animals of another origin is by far the most where substrains originating from the same inbred important source of genetic profile alteration in inbred strain have acquired new phenotypes as a consequence strains. Genetic contamination of this type, which of genetic drift.32 For example, mice of the C57BL/ always results in a sudden and massive exchange of 6JOlaHsd substrain are homozygous for a deletion of alleles, is more likely between strains that have similar the a-synuclein (Snca) and multimerin (Mnrn1) coat colour (i.e. albino (Tyrc/Tyrc), agouti (A/A), or genes.33,34 Likewise, some spontaneous mutations dif- non-agouti (a/a)). Where lines have the same coat ferentially segregate in C57BL/6J and C57BL/6N, the colour alleles, extra care must be taken when housing most common substrains of C57BL/6, separated in them in close proximity of each other. 1951. These include a retinal degeneration mutation in the Crb1 gene (Crb1rd8), present only in the N sub- Spontaneous mutations and polymorphisms strain, and a deletion in the Nnt gene, present only in the J substrain.35,36 Berghe and colleagues recently Spontaneous mutations are a source of uncontrolled reported that passenger mutations are also common genetic variation that is often impossible to detect by in most GA lines derived from 129 ES cells, and that simple phenotypic observation or routine genetic moni- these mutations persist even after the creation of fully toring (GeMo). Genetic polymorphism is the presence of congenic strains.37 This is not trivial; Berghe et al. Benavides et al. 141 estimated that close to 1000 protein-coding genes could dozens of times creating allelic diversity among stains. be aberrantly expressed in the 129-derived chromo- Genomic DNA primers are designed to unique somal segments that are still segregating in these consequences flanking the repeats. The PCR products, typ- genic lines. This finding emphasizes the need for ically around 100–300 bp in size, are analysed using properly chosen control animals to identify phenotypes agarose or polyacrylamide gel electrophoresis. The due to background mutations or the combination of MGI webpage has comprehensive SSLP information, background mutations and the genetic modification of including primer sequences and size variations in bp interest, rather than the modification itself. for several inbred mouse strains (http://www.informatics.jax.org/marker). A collection of mapped, highly Importance of using standard polymorphic, SSLP markers for inbred laboratory rat nomenclature strains is available in The National BioResource Project – Rat database and is linked to the Map Rules guiding nomenclature were established by the Report of the Rat Genome Database (RGD) (http:// International Committee on Standardized Genetic rgd.mcw.edu). See Supplementary Table 2 for the com- Nomenclature for Mice and Rats and are continuously plete list of online resources for laboratory mouse and updated. These rules, last revised in January 2016, are rat strains. described on the MGI webpage under ‘Guidelines for SNP genotyping is an alternative to microsatellites Nomenclature of Mouse and Rat Strains’ (http://www. that is now widely used for GeMo. SNP genotyping is informatics.jax.org/mgihome/nomen/strains.shtml). A inexpensive and can be performed in most research helpful and visual Mouse Nomenclature Quick Guide institutions or outsourced. SNPs are the most is available at https://www.jax.org/jax-mice-and-ser- common genetic variation and exist in both coding vices/customer-support/technical-support/genetics- and non-coding regions. Almost all SNPs are bi-allelic, and-nomenclature#. For more details on nomenclature presenting one of only two possible nucleotides (e.g. refer to the Supplementary material. homozygous G/G or T/T), or both (e.g. heterozygous G/T) in an individual. Petkov and co-workers from The Genetic quality control programmes Jackson Laboratory have described the allelic distribution of 235 SNPs in 48 mouse strains and selected a The current gold standard for genetic quality control of panel of 28 SNPs sufficient to characterize the majority laboratory rodents depends on polymorphic genetic of the c. 300 inbred, wild-derived, congenic, consomic markers to distinguish between different genetic back- and recombinant inbred strains maintained at The grounds. Genetic markers are specific DNA sequences Jackson Laboratory.40 Several publications have with a known location on a chromosome and are essen- reported useful SNPs for the rat. For example, tial tools for genetic quality control. Genetic quality Zimdahl and colleagues described a map with control is essential to determine the genetic compos- >12,000 gene-based SNPs from transcribed regions.41 ition of an animal and to screen for uniformity and authenticity of a strain. Please note that outbred colo- GeMo of inbred strains and outbred stocks nies cannot be tested for authenticity. Instead, the colony is screened for its level of genetic heterogeneity Most GeMo techniques used currently are based on to detect genetic contamination and to monitor the pro- microsatellites or SNPs. However, GeMo should not gress of breeding programmes and to select future rely solely on molecular techniques, but should take a breeders. broader view that includes phenotypic parameters such as coat colour, behaviour and breeding performance. Marker systems Commercial breeders are extremely sensitized to the risk of genetic contamination and regularly monitor Many polymorphisms have been described in the their strains for genetic contamination, but not neces- mouse and rat; however, only microsatellites and sarily genetic drift, by using different sets of SNPs to SNPs are used as genetic markers in current QA pro- monitor their nucleus colonies. The Jackson grammes. Microsatellite markers, also known as Simple Laboratory incorporated a unique, patented, Genetic Sequence Length Polymorphisms (SSLPs) or Short Stability Program42 designed to effectively limit cumu- Tandem Repeats (STRs), are still used in modern lative genetic drift by rebuilding their foundation stocks GeMo programmes because they are inexpensive and from pedigreed, cryopreserved embryos every five gen- easy to type.38,39 Animals are genotyped by analysing erations. For example, starting in 2005, they began sell- PCR-products amplified from short, tandemly ing only C57BL/6J mice derived from two chosen mice arranged, repeating DNA sequences. These repeats through hundreds of frozen embryos of the duo’s are typically 2–6 bp long and are repeated a few to grandchildren (enough to last for 25–30 years). It 142 Laboratory Animals 54(2) should be noted that when recovering strains from of mouse SSLPs that could be used to authenticate frozen stocks good GeMo should be carried out to some classical inbred strains. assure oneself that genetic contamination or wrong Interpreting SSLP data is straightforward. Because genotypes were not present prior to freezing. inbred animals are isogenic and homozygous, they will For outbred stocks, GeMo helps preserve the genetic present only one band in the electrophoresis gel, repre- heterogeneity and allele pool of a colony. This complex senting a single allele, when genotyped for a particular process requires analysing a large number of animals SSLP. The presence of any heterozygosity, indicated by and comparing this data with historical data document- two bands, or bands that do not coincide with those of ing the alleles present, their frequency and the level of the strain control DNA, should be considered as indi- heterozygosity in that particular colony. In some cases, cating potential strain contamination (Figure 2). How the results can reveal a loss of genetic variability result- frequently colony strain identity should be evaluated ing in a colony with very low heterogeneity. The degree depends on the size of the colony, the generation inter- of genetic heterogeneity in outbred colonies depends on val, etc. Generally, testing once every two years is rea- their history. Low heterogeneity can result from poor sonable for a facility maintaining a small number of selection of future breeding stock, deviation from colonies well-separated in terms of coat colour, and approved (rotational) breeding systems or the bottle- with low numbers of importations. neck effect caused by a small breeding pool, as is common when a small group of breeders is imported Using a small panel of SNPs. For GeMo purposes or being used to rederive a colony. In contrast, high only, 40 polymorphic SNPs, evenly distributed across heterogeneity can result from a recent outcross. In gen- the chromosomes is a reasonable number for detecting eral, outbred stocks are characterized by measuring recent genetic contamination (this suggestion should be phenotypic traits and calculating the corresponding modified dependent on the conditions or risks in each mean and standard deviations. Essentially, genetic con- facility). SNP genotyping is currently available on dif- trol of outbred stocks is directed at avoiding inbreeding ferent platforms, that vary in cost and automation cap- and stabilizing genetic diversity over many generations. abilities. Kompetitive Allele Specific PCR (KASP), a variation on allele-specific PCR, uses allele-specific GeMo of inbred mice and rats bred in-house. The oligo extension and fluorescence resonance energy best recommendation here is to purchase animals transfer,43 has the advantage that it can be automated from reliable vendors and replace the breeding stock using 96- or 384-well plates and pipetting robots for the with animals from the same vendor after 10 gener- PCR reactions (Supplementary Figure 3). Another ations, rather than to maintain independent colonies option, real-time PCR (TaqManÕ) technology, uses of classical inbred strains. As an additional benefit, specific primers coupled with a sequence-specific, fluor- using animals from the same vendor prevents the for- escent TaqMan probe, is effective and easy to automation of substrains harbouring potential mutations mate; however, the cost per individual assay is and maintains a similar microbiome. Nevertheless, inexpensive compared with KASP assays, and requires house colonies should always be tested with a small set a more costly real-time thermocycler. Finally, microar- of informative microsatellite markers or SNPs to con- ray-based SNP genotyping is not typically used for firm integrity. small scale, in-house GeMo, but may be an option for vendors of inbred mice. When using or requesting Using a small panel of microsatellites microarray genotyping services, be aware that only a (SSLPs). Microsatellites can be used to verify that the percentage of the SNPs will be polymorphic between animals in an inbred colony are essentially pure, with the strains under analysis (e.g., c. 40% for some clas- no traces of genetic contamination. This is especially sical inbred strain combinations). Information regard- important in facilities that maintain strains with the ing which alleles (C, G, A or T) to expect for a same coat colour in the same room, a particularly dan- particular SNP/strain combination, and their genomic gerous practice especially when not using individually location are available for hundreds of thousands of ventilated cage (IVC) systems. Microsatellite testing SNPs and for the common mouse and rat inbred strains can normally be performed in-house. The number of in easily accessed databases and genome browsers markers to use for testing has not been standardized: (Supplementary Table 2). each situation and facility differs in how many and which strains are kept. Nonetheless, a panel of 40 poly- GeMo of outbred colonies. GeMo of outbred stocks is morphic SSLPs, evenly distributed across the auto- much more complex, because these animals are not gen- somes, will rule out recent genetic contamination, if etically uniform. Outbred colonies are essentially a the markers can distinguish among the strains being group of closely related animals, with shared ancestors analysed. Supplementary Table 3 presents a small set and group identity, but that exhibit some level of Benavides et al. 143

Figure 2. Example of genetic contamination detected by SSLP PCR. The picture shows a 4% agarose gel with the characteristic bands obtained after PCR amplification using genomic DNA from four mice supposedly belonging to the BALB/c strain (first four lanes), plus a standard DNA control for BALB/c (last lane). In this example, only five SSLP loci are shown, located in chromosomes 1 to 5. Note the presence of heterozygosity (two bands) and homozygosity for bands that do not match the standard for BALB/c. This is a clear case of loss of authenticity due to genetic contamination. The PCR products are compared with a 100 bp DNA ladder.

Figure 3. This chart explains the typical timeline for a marker assisted (speed congenic) backcross process. The pre- diction of >98% recipient genome at N5 is based on the use of 20 best breeders (carriers) at each generation,55 however, this number is not always available and fewer breeders can be used, with disparate results, depending also on chance. PI: Principal Investigator (laboratory). Service: the laboratory providing the genome scan with SNP markers. genetic heterozygosity. Since outbred colonies form a animals in the breeding colony, causing a reduction population rather than a strain, it is difficult to establish of alleles represented in the population. This may a standard GeMo programme with only a few genetic impact genetic drift and increase the inbreeding markers. However, with an adequate number of SNPs coefficient. Such colonies are neither truly outbred or SSLPs, allele frequencies within the population nor inbred. Although SSLPs or SNPs can be used to could indicate the identity of the stock.44 One of the estimate the level of heterozygosity within the colony, main problems of in-house outbred stocks is that they if it is not possible to keep an appropriate number are often maintained with a very small number of of breeders, it is better to purchase outbred 144 Laboratory Animals 54(2) rodents from vendors that maintain a very large colony strain of choice; (b) modifying the gene of interest in and use recommended breeding schemes to reduce ES cells from the preferred background strain (e.g. inbreeding. using C57BL/6 ES cells); and (c) crossing chimeras and KO/KI founders with mice of the same strain as Background characterization (BC) for GA the ES cells used for the targeting. Finally, if the GA and mutant lines line has already been developed or acquired from a collaborator or repository, a BC should be per- The explosion in the number of GA lines is exacerbat- formed, and if needed, a fully congenic strain should ing the problem of undefined ‘mixed backgrounds’ in be developed, either by classical or marker-assisted experimental rodents. This is particularly worrisome backcrossing. Periodic backcrossing of a congenic for inducible and conditional models that require the strain to the background strain (of reputable source) crossing of two independent lines (e.g. Cre-expressing also minimises divergence and keeps the congenic lines crossed with floxed lines). Given that genetic strain genetically close to the strain background of con- background influences phenotype, especially through trol animals. the influence of modifier genes; mutations, both spontaneous and induced, transgenes, and targeted Marker-assisted backcrossing for quality alleles that are introgressed into a new background assurance and refinement may not exhibit the expected phenotype.45,46 One of the first cases reporting this phenomenon involved the The use of DNA markers has allowed for a much more classical diabetes (Leprdb) mutation that presented rapid and rigorous process of congenic strain develop- transient diabetes in a C57BL/6 background but ment called marker-assisted backcrossing or speed con- overt diabetes in C57BLKS.47 Other examples genics.53 This process relies on using polymorphic include background effects on survival rate in Egfr genetic markers covering the whole genome to deter- (epidermal growth factor receptor) KO mice48 and mine the percentage of donor genome present in the tumour incidence in Pten KO mice.49 For this reason, animals, then selecting the animals with the lowest every GA line should be characterized in terms of percentage of donor DNA for the next backcross to their genetic background. Moreover, the knowledge the recipient strain. This relies on the regions between of the genetic background of a mutation is also import- the polymorphic genetic markers being those of the ant for the selection of the corresponding control donor genome: the denser the number of markers the animals.50 higher the donor genome can be inferred. Common Genetic markers evenly distributed and covering the practice is the use of 100–300 markers. This process entire genome can be used in a genome scan to estimate reduces the number of generations to reach full con- the percentages of genome coming from different genicity (e.g. from N10 to N5), and therefore strain inbred origins. This process of BC is provided by development time, by approximately half. Using some commercial enterprises and institutional core marker-assisted backcrosses and the right number of facilities. A typical BC employs polymorphic markers animals we can obtain c. 80% recipient background at to distinguish between the suspected inbred strains. In N2, c.94%atN3,andc. 99% at N4 (compared to the most mouse cases, these strains are C57BL/6 and 129 classical mean values of 75.0, 87.5 and 93.7%). A substrains because, historically, the ES cells used for the flowchart depicting a standard speed congenic proto- development of KO and KI mice through homologous col is shown in Figure 3. Ideally, the backcross pro- recombination (section ‘‘Targeted mutagenesis by cedure is started with a donor female and a recipient homologous recombination using ES cells’’ above) male. Then, F1 mutant males will carry the correct were derived exclusively from 129 substrains51 whereas Y-chromosome and after mating to a recipient WT C57BL/6 females were typically used to prove female, males of the N2 generation will carry the cor- germline transmission from the chimeras. Without sub- rect X- and Y-chromosome of the recipient strain sequent backcrosses, this scheme resulted in a B6;129 (avoiding the use of genetic markers on these chromo- mixed background. However, the availability of ES cell somes).54 It was predicted by Markel et al. that if 20 lines derived from other strains (particularly from best breeders (carriers) are used at each generation of C57BL/6) and the arrival of genome editing techniques the speed congenics protocol >98% recipient gen- (section ‘‘Gene editing using nucleases’’) that allow ome can be attained at N5.55,56 However, the chromo- direct production of targeted alterations in any mouse somal segments flanking the selected locus tend to or rat strain52 is slowly changing this scenario. In any remain associated with it and this is a limitation of case, the problem of mixed background can be circum- the congenic lines due to the potential presence vented altogether by (a) injecting transgenes or nucle- of modifier genes in this segments, the so-called ases (Cas9-sgRNA) into inbred embryos from the ‘flanking gene problem’.57 Benavides et al. 145

Genetic stability and cryopreservation References programmes 1. Silver L. Mouse Genetics. Concepts and Applications. Oxford: Oxford University Press, 1995. For inbred, co-isogenic and congenic strains, breeding 2. Simecek P, Forejt J, Williams RW, et al. High-resolution methods and genetic stability programmes help to min- maps of mouse reference populations. G3 (Bethesda) imize substrain divergence due to genetic drift, and also to 2017; 7: 3427–3434. prevent genetic contamination by accidental crosses with 3. Guenet JL, Benavides F, Panthier J, et al. Genetics of the other strains. To reduce genetic drift, the number of gen- Mouse. Berlin: Springer, 2015. erations of in-house breeding should be minimized, and 4. Paigen K and Eppig JT. A mouse phenome project. the lines submitted to repositories such as, JAX, EMMA, Mamm Genome 2000; 11: 715–717. 5. Eppig JT. Mouse Genome Informatics (MGI) Resource: MMRRC, IMSR or RIKEN, to be archived as frozen Genetic, genomic, and biological knowledgebase for the embryos and/or sperm. This secures the line and provides laboratory mouse. ILAR J 2017; 58: 17–41. a means of replacing the breeding stock every 10 gener- 6. Chia R, Achilli F, Festing MF, et al. The origins and uses ations as recommended by The Jackson Laboratory of mouse outbred stocks. Nat Genet 2005; 37: 1181–1186. Genetic Stability Program (GSP) in order to slow down 7. Festing MF. Inbred strains should replace outbred stocks cumulative genetic drift.42 For outbred stocks, the intent in toxicology, safety testing, and drug development. is to minimize inbreeding, maintain heterozygosity and Toxicol Pathol 2010; 38: 681–690. manage genetic drift that would otherwise lead to 8. Nagy A, Gertsenstein M, Vintersten K, et al. colony divergence. Ideally, outbred colonies should be Manipulating the Mouse Embryo: A laboratory manual. maintained with 25 breeding pairs, all of which have 3rd ed. New York: Cold Spring Harbor Press, 2003. to contribute to the next generation, in order to avoid 9. Gordon JW and Ruddle FH. Integration and stable germ an increase of the inbreeding coefficient per generation line transmission of genes injected into mouse pronuclei. Science 1981; 214: 1244–1246. of more than 1%. Smaller colonies drift fast toward 10. Ballester M, Castello A, Ibanez E, et al. Real-time quan- homozygosity because breeders are closely related.58 titative PCR-based system for determining transgene Cryopreservation strategies have been adopted for copy number in transgenic animals. Biotechniques 2004; long-term storage of embryos and gametes in several 37: 610–613. large centralized repositories including the EMMA/ 11. Gossen M and Bujard H. Tight control of gene expres- INFRAFRONTIER (European Mutant Mouse sion in mammalian cells by tetracycline-responsive pro- Archive), the Knock Out Mouse Project (KOMP) moters. Proc Natl Acad Sci USA 1992; 89: 5547–5551. Repository, The Jackson Laboratory Repository, The 12. Lois C, Hong EJ, Pease S, et al. Germline transmission Center for Animal Resources and Development and tissue-specific expression of transgenes delivered by (CARD) and the Riken Bio Resource Center, which lentiviral vectors. Science 2002; 295: 868–872. can provide cryopreserved material or live mice to 13. Ivics Z, Mates L, Yau TY, et al. Germline transgenesis in laboratories. These repositories facilitate the availabil- rodents by pronuclear microinjection of Sleeping Beauty ity of these strains to the worldwide scientific commu- transposons. Nat Protoc 2014; 9: 773–793. 14. Perry AC, Wakayama T, Kishikawa H, et al. Mammalian nity and provide a backup for a potential loss of a transgenesis by intracytoplasmic sperm injection. Science strain. The International Mouse Strain Resource 1999; 284: 1180–1183. (IMSR) is a searchable online database of mouse 15. Fernandez-Gonzalez R, Moreira PN, Perez-Crespo M, strains, stocks and mutant ES cell lines available world- et al. Long-term effects of mouse intracytoplasmic wide, including inbred, mutant and genetically engin- sperm injection with DNA-fragmented sperm on health eered strains (http://www.findmice.org/). and behavior of adult offspring. Biol Reprod 2008; 78: 761–772. Declaration of Conflicting Interests 16. Sauer B and Henderson N. Site-specific DNA recombin- The author(s) declared no potential conflicts of interest with ation in mammalian cells by the Cre recombinase of bac- respect to the research, authorship, and/or publication of this teriophage P1. Proc Natl Acad Sci USA 1988; 85: article. 5166–5170. 17. Feil S, Valtcheva N and Feil R. Inducible Cre mice. Funding Methods Mol Biol 2009; 530: 343–363. 18. Carbery ID, Ji D, Harrington A, et al. Targeted genome The author(s) received no financial support for the research, modification in mice using zinc-finger nucleases. Genetics authorship, and/or publication of this article 2010; 186: 451–459. 19. Jinek M, Chylinski K, Fonfara I, et al. A programmable Supplementary Material dual-RNA-guided DNA endonuclease in adaptive bacter- The full report, including 160 references, is available as ial immunity. Science 2012; 337: 816–821. Supplemental Material to this publication online via https:// 20. Yang H, Wang H, Shivalila CS, et al. One-step gener- journals.sagepub.com/doi/full/10.1177/0023677219867719. ation of mice carrying reporter and conditional alleles 146 Laboratory Animals 54(2)

by CRISPR/Cas- mediated genome engineering. Cell 37. Vanden Berghe T, Hulpiau P, Martens L, et al. 2013; 154: 1370–1379. Passenger mutations confound interpretation of all gen- 21. Wang H, Yang H, Shivalila CS, et al. One-step gener- etically modified congenic mice. Immunity 2015; 43: ation of mice carrying mutations in multiple genes by 200–209. CRISPR/Cas-mediated genome engineering. Cell 2013; 38. Benavides F, Glasscock E, Coghlan LG, et al. PCR-based 153: 910–918. microsatellite analysis for differentiation and genetic 22. Yoshimi K, Kaneko T, Voigt B, et al. Allele-specific monitoring of nine inbred SENCAR mouse strains. Lab genome editing and correction of disease-associated Anim 2001; 35: 157–162. phenotypes in rats using the CRISPR-Cas platform. 39. Mashimo T, Voigt B, Tsurumi T, et al. A set of highly Nat Commun 2014; 5: 4240. informative rat simple sequence length polymorphism 23. Kosicki M, Tomberg K and Bradley A. Repair of double- (SSLP) markers and genetically defined rat strains. strand breaks induced by CRISPR-Cas9 leads to large BMC Genet 2006; 7: 19. deletions and complex rearrangements. Nat Biotechnol 40. Petkov PM, Ding Y, Cassell MA, et al. An efficient SNP 2018; 36: 765–771. system for mouse genome scanning and elucidating strain 24. Rezza A, Jacquet C, Le Pillouer A, et al. relationships. Genome Res 2004; 14: 1806–1811. Unexpected genomic rearrangements at targeted loci 41. Zimdahl H, Nyakatura G, Brandt P, et al. A SNP map of associated with CRISPR/Cas9-mediated knock-in. Sci the rat genome generated from cDNA sequences. Science Rep 2019; 9: 3486. 2004; 303: 807. 25. Nolan PM, Peters J, Strivens M, et al. A systematic, 42. Taft RA, Davisson M and Wiles MV. Know thy mouse. genome-wide, phenotype-driven mutagenesis programme Trends Genet 2006; 22: 649–653. for gene function studies in the mouse. Nat Genet 2000; 43. Myakishev MV, Khripin Y, Hu S, et al. High-throughput 25: 440–443. SNP genotyping by allele-specific PCR with universal 26. Bonaparte D, Cinelli P, Douni E, et al. FELASA guide- energy-transfer-labeled primers. Genome Res 2001; 11: lines for the refinement of methods for genotyping geneti- 163–169. cally-modified rodents: A report of the Federation of 44. Hartl DL. Genetic Management of Outbred Laboratory European Laboratory Animal Science Associations Rodent Populations.Wilmington, MA: Charles River Working Group. Lab Anim 2013; 47: 134–145. Genetic Literature, 2001. 27. Rulicke T, Montagutelli X, Pintado B, et al. FELASA 45. Linder CC. The influence of genetic background on spon- guidelines for the production and nomenclature of trans- taneous and genetically engineered mouse models of com- genic rodents. Lab Anim 2007; 41: 301–311. plex diseases. Lab Anim (NY) 2001; 30: 34–39. 28. Peters H, Reifenberg K and Wedekind D. Substrains of 46. Doetschman T. Influence of genetic background on gen- Inbred Strains. GV-SOLAS Specialist Information, 2013. etically engineered mouse phenotypes. Methods Mol Biol 29. Mekada K, Hirose M, Murakami A, et al. Development 2009; 530: 423–433. of SNP markers for C57BL/6N-derived mouse inbred 47. Hummel KP, Coleman DL and Lane PW. The influence strains. Exp Anim 2015; 64: 91–100. of genetic background on expression of mutations at the 30. Hermsen R, de Ligt J, Spee W, et al. Genomic landscape diabetes locus in the mouse. I. C57BL-KsJ and C57BL- of rat strain and substrain variation. BMC Genomics 6J strains. Biochem Genet 1972; 7: 1–13. 2015; 16: 357. 48. Threadgill DW, Dlugosz AA, Hansen LA, et al. Targeted 31. Kenneth NS, Younger JM, Hughes ED, et al. An inacti- disruption of mouse EGF receptor: effect of genetic back- vating caspase 11 passenger mutation originating from ground on mutant phenotype. Science 1995; 269: 230–234. the 129 murine strain in mice targeted for c-IAP1. 49. Freeman D, Lesche R, Kertesz N, et al. Genetic back- Biochem J 2012; 443: 355–359. ground controls tumor development in PTEN-deficient 32. Stevens JC, Banks GT, Festing MF, et al. Quiet muta- mice. Cancer Res 2006; 66: 6492–6496. tions in inbred strains of mice. Trends Mol Med 2007; 13: 50. Bourdi M, Davies JS and Pohl LR. Mispairing C57BL/6 512–519. substrains of genetically engineered mice and wild-type 33. Specht CG and Schoepfer R. Deletion of the alpha-synu- controls can lead to confounding results as it did in stu- clein locus in a subpopulation of C57BL/6J inbred mice. dies of JNK2 in acetaminophen and concanavalin A liver BMC Neurosci 2001; 2: 11. injury. Chem Res Toxicol 2011; 24: 794–796. 34. Specht CG and Schoepfer R. Deletion of multimerin-1 in 51. Simpson EM, Linder CC, Sargent EE, et al. Genetic vari- alpha-synuclein- deficient mice. Genomics 2004; 83: ation among 129 substrains and its importance for tar- 1176–1178. geted mutagenesis in mice. Nat Genet 1997; 16: 19–27. 35. Freeman HC, Hugill A, Dear NT, et al. Deletion of nico- 52. Fernandez A, Josa S and Montoliu L. A history of tinamide nucleotide transhydrogenase: a new quantitative genome editing in mammals. Mamm Genome 2017; 28: trait locus accounting for glucose intolerance in C57BL/ 237–246. 6J mice. Diabetes 2006; 55: 2153–2156. 53. Wakeland E, Morel L, Achey K, et al. Speed congenics: 36. Mattapallil MJ, Wawrousek EF, Chan CC, et al. The A classic technique in the fast lane (relatively speaking). Rd8 mutation of the Crb1 gene is present in vendor Immunol Today 1997; 18: 472–477. lines of C57BL/6N mice and embryonic stem cells, and 54. Dobrowolski P, Fischer M and Naumann R. Novel confounds ocular induced mutant phenotypes. Invest insights into the genetic background of genetically mod- Ophthalmol Vis Sci 2012; 53: 2921–2927. ified mice. Transgenic Res 2018; 27: 265–275. Benavides et al. 147

55. Markel P, Shu P, Ebeling C, et al. Theoretical and empir- CD147. Biochem Biophys Res Commun 2004; 324: ical issues for marker-assisted breeding of congenic 147–153. mouse strains. Nat Genet 1997; 17: 280–284. 58. Berry MM and Linder CC. Breeding systems: 56. Gurumurthy CB, Joshi PS, Kurz SG, et al. Validation of Considerations, genetic fundamentals, genetic back- simple sequence length polymorphism regions of commonly ground and strain types. In: Fox JGBS, Davisson MT, used mouse strains for marker assisted speed congenics Newcomer CE, et al, (eds) The Mouse in Biomedical screening. Int J Genomics 2015; 2015: 735845. Research. Vol. 1: History, Wild Mice and Genetics. 57. Chen S, Kadomatsu K, Kondo M, et al. Effects of flank- Cambridge, MA: Academic Press, 2007, pp. 53–78. ing genes on the phenotypes of mice deficient in basigin/

Re´sume´ L’assurance qualite´ (AQ) geńe´tique, dont la surveillance geńe´tique (SG) des souches consanguines et la caracte´risation des souches (CS) de mode`les animaux geńe´tiquement modifie´s (GM) devrait constituer un aspect essentiel de tout programme d’AQ dans les installations de recherche animale. Le controˆle qualite´ geńe´tique est aussi important que la surveillance sanitaire ou microbiologique pour assurer la validite´ du mode`le animal. Il devrait eˆtre obligatoire que les e´tudes utilisant des rongeurs de laboratoire, principalement des souris et des rats, utilisent des animaux geńe´tiquement de´finis. Le document pre´sente´ par le groupe de travail FELASA sur l’assurance qualite´ et la surveillance geńe´tiques des rats et souris de laboratoire, dećrit les objectifs et les me´thodes disponibles pour les programmes d’AQ mene´s dans les installations utilisant des rongeurs. Les objectifs principaux d’un programme d’AQ sont les suivants : (i) ve´rifier l’authenticite´ et l’uniformite´ des souches consanguines et des sous-souches, assurant ainsi la maintenance d’une colonie geńe´tiquement fiable; (ii) de´tecter les contaminations geńe´tiques e´ventuelles; et (iii) dećrire prećise´ment la composition geńe´tique des ligneés GM. Bien que cette publication se concentre sur l’AQ geńe´tique des souris et des rats, les principes s’appliqueront a` d’autres espe`ces de rongeurs, dont certaines sont brie`vement mentionneés dans le contexte des animaux issus de ligneés consanguines ou croiseés.

Abstract Die genetische Qualita¨tssicherung (QA), einschließlich des genetischen Monitorings (GeMo) von Inzuchtsta¨mmen und der Hintergrundcharakterisierung (BC) von genetisch vera¨nderten (GA) Tiermodellen, sollte generell ein wesentlicher Bestandteil aller QA-Programme in Versuchstiereinrichtungen sein. Die genetische Qualita¨tskontrolle ist zur Gewa¨hrleistung der Validita¨t von Tiermodellen ebenso wichtig wie die U¨berwachung ihrer Gesundheit und mikrobiologischen Qualita¨t. Fu¨r Studien mit Labornagern, hauptsa¨chlich betrifft es Ma¨use und Ratten, sollte ausschließlich die Verwendung von genetisch definierten Tieren vorge- sehen werden. Dieses Dokument, das von der FELASA Arbeitsgruppe u¨ber genetische Qualita¨tssicherung und genetisches Monitoring von Laborma¨usen und -ratten pra¨sentiert wird, beschreibt die Ziele und verfu¨gbaren Methoden fu¨r genetische QA-Programme in Labortierhaltungen. Die Hauptziele eines jeden genetischen QA Programms sind: (i) U¨berpru¨fung der Authentizita¨t und Uniformita¨t von Inzuchtsta¨mmen und deren Substa¨mme, um so eine genetisch zuverla¨ssige Erhaltung der Kolonie zu gewa¨hrleisten, (ii) Erkennung mo¨glicher genetischer Kontaminationen, und (iii) pra¨zise Beschreibung der genetischen Beschaffenheit von GA-Linien. Diese Vero¨ffentlichung konzentriert sich hauptsa¨chlich auf die genetische QA von Maus und Ratte, wobei die Prinzipien auch fu¨r andere Nagetierarten, von denen einige im Zusammenhang mit Inzucht- und Auszuchtsta¨mmen kurz erwa¨hnt werden, zutreffen.

Resumen La garantıá de calidad gene´tica (QA), incluidos el monitoreo gene´tico (GeMo) de las cepas consanguıńeas y la caracterizacioń de fondo gene´tico (BC) de los animales gene´ticamente modificados (GA), deberıá ser un componente esencial de cualquier programa de QA en los animalarios de roedores. El control de la calidad gene´tica es tan importante para asegurar la validez del modelo animal como lo es el control de calidad sanitaria y microbiolo´gica. Deberıá exigirse que los estudios que utilicen roedores de laboratorio, principalmente ratones y ratas, utilicen exclusivamente animales gene´ticamente definidos. Este manuscrito, presen- tado por FELASA Working Group on Genetic Quality Assurance and Genetic Monitoring of Laboratory Murines, describe los objetivos y me´todos disponibles para los programas de calidad gene´tica en instalaciones de 148 Laboratory Animals 54(2) roedores de laboratorio. Los principales objetivos de cualquier programa de calidad gene´tica son: (i) verificar la autenticidad y uniformidad de las cepas (y sub-cepas) consanguıńeas; (ii) detectar una posible contamina- cioń gene´tica; y (iii) describir con precisioń la composicioń gene´tica de las lıńeas gene´ticamente modificadas. Si bien esta publicacioń se centra principalmente en los controles de calidad gene´tica de ratones y ratas, los mismos principios se aplican a otras especies de roedores de laboratorio, algunas de las cuales se mencio- nan brevemente en el contexto de las cepas consanguıńeas y los grupos exocriados de ratones y ratas. Editorial Laboratory Animals 2020, Vol. 54(2) 149 ! The Author(s) 2020 Article reuse guidelines: Collection on score sheets, severity sagepub.com/journals-permissions DOI: 10.1177/0023677220906399 assessment and humane end points: journals.sagepub.com/home/lan Invitation to submit

Paulin Jirkof1, Gavin Jarvis2 and Beat Riederer3

Laboratory Animals publishes original papers and the development and evaluation of severity assessment reviews on the care and use of animals in biomedical methods, along with example score sheets from differ- research with the aim of promoting the welfare of ent scientific fields (https://journals.sagepub.com/page/ laboratory animals. An important tool for evaluating lan/collections/score-sheets). We hope that this collec- and promoting animal welfare in experiments is severity tion will encourage further research in this field, monitoring. EU Directive 2010/63/EU introduced improving sharing of the knowledge gained from their requirements for classifying the severity of procedures, use and subsequent improvements in laboratory animal not only for project authorisation but also for reporting welfare. actual severity experienced by the individual animals.1 We invite scientists to submit original papers on Score sheets standardise and formalise severity severity assessment methods and humane end-point cri- assessment. This makes it possible to record the effect teria, reviews on the use and suitability of current of scientific procedures on animals, and therefore to assessment methods and scoring schemes and short inform decisions on remedial measures or experiment reports on example score sheets to Laboratory Animals. termination. Additionally, well-defined humane end points reduce suffering and minimise unnecessary Declaration of Conflicting Interests deaths of animals. Data collected in score sheets can The author(s) declared no potential conflicts of interest with direct the development of refinements, for example respect to the research, authorship and/or publication of this optimised analgesic protocols, and facilitate their effect- article. ive implementation, thereby improving the welfare of animals used in research.2 Funding As part of the implementation of EU Directive 2010/ The author(s) received no financial support for the research, 63/EU, extensive efforts are being made to improve authorship and/or publication of this article. methods for severity assessment (see Special Issue ‘‘Severity assessment in animal-based research’’ References Laboratory Animals 2020; 54(1); https://journals.sage- 1. Smith D, Anderson D, Degryse AD, et al. Classification pub.com/toc/lana/54/1). However, there is no one-size- and reporting of severity experienced by animals used in fits-all approach to scoring severity and animal welfare, scientific procedures: FELASA/ECLAM/ESLAV and practical, evidence-based, established and available Working Group report. Lab Anim 2018; 52: 5–57. score sheets are absent for many experimental proced- 2. Golledge H and Jirkof P. Score sheets and analgesia. Lab ures. Further development is therefore essential to Anim 2016; 50: 411–413. increase the accuracy, specificity and effectiveness of scoring and grading systems for severity and animal welfare. 1Department Animal Welfare and 3R, University of Zurich, As a contribution to this process, Laboratory Switzerland 2Department of Physiology, Development and Neuroscience, Animals aims to collate recently published articles on University of Cambridge, UK 3Platform for Morphology and Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Switzerland

Corresponding author: Paulin Jirkof. Department Animal Welfare and 3R, University of Zurich, Zurich, Switzerland. Email: [email protected] Severity Assessment in Animal-based Research Laboratory Animals 2020, Vol. 54(2) 150–158 ! The Author(s) 2020 Impulse for animal welfare outside Article reuse guidelines: the experiment sagepub.com/journals-permissions DOI: 10.1177/0023677219891754 journals.sagepub.com/home/lan Lars Lewejohann1,2 , Kerstin Schwabe3, Christine Ha¨ger4 and Paulin Jirkof5

Abstract Animal welfare is a growing societal concern and the well-being of animals used for experimental purposes is under particular scrutiny. The vast majority of laboratory animals are mice living in small cages that do not offer very much variety. Moreover, the experimental procedure often takes very little time compared to the time these animals have been bred to the desired age or are being held available for animal experimentation. However, for the assessment of animal welfare, the time spent waiting for an experiment or the time spent after finishing an experiment has also to be taken into account. In addition to experimental animals, many additional animals (e.g. for breeding and maintenance of genetic lines, surplus animals) are related to animal experimentation and usually face similar living conditions. Therefore, in terms of improving the overall welfare of laboratory animals, there is not only a need for refinement of experimental conditions but especially for improving living conditions outside the experiment. The improvement of animal welfare thus depends to a large extent on the housing and maintenance conditions of all animals related to experimentation. Given the current state of animal welfare research there is indeed a great potential for improving the overall welfare of laboratory animals.

Keywords 3Rs, animal use, environmental enrichment, housing, laboratory animal welfare

Date received: 7 April 2019; accepted: 10 November 2019

Introduction million individuals in Germany and 2.57 million procedures carried out with mice in the UK. For the year There is no reliable, let alone official, number of labora- 2017, for the first time, not only the number of experi- tory animals used worldwide. Even for highly regulated mental animals but also the number of animals used for areas like the EU, comprehensive data is published with breeding and maintenance as well as the number of a considerable delay. Moreover, the data available comprises a fuzziness as there is no uniform convention 1German Federal Institute for Risk Assessment (BfR), German on what exactly to count. The latest numbers available Centre for the Protection of Laboratory Animals (Bf3R), Berlin, for the EU are for the year 2017, with 9.4 million ani- Germany mals used for animal experimentation.1 Many countries 2Institute of Animal Welfare, Animal Behavior and Laboratory provide yearly statistics on animal use, allowing some Animal Science, Freie Universita¨t Berlin, Germany 3Department of Neurosurgery, Hannover Medical School, Germany general conclusions to be drawn and enabling more 4Institute for Laboratory Animal Science, Hannover Medical current estimates. For example, the numbers for 2018 School, Germany published by the German Federal Ministry of Food 5Department Animal Welfare and 3Rs, University of Zurich, and Agriculture,2 as well as the numbers of procedures Switzerland in the UK in 2018,3 can help to conclude some general Corresponding author: directions. Overall, the latest numbers published were Lars Lewejohann, Freie Universita¨t Berlin, Ko¨nigsweg 67, Berlin, comparable to preceding years, and again, the most 14163 Germany. widely used species was the mouse with roughly 1.54 Email: [email protected] Lewejohann et al. 151 animals bred but not used was assessed throughout order to prevent poor animal welfare rather than the EU. Overall almost 14 million additional animals defining what constitutes animal welfare per se. were counted in 28 EU countries indicating that for Consequently, the later literature emphasizes subjectiv- every two experimental animals additionally three sur- ity for animal welfare and incorporates the view of ani- plus animals have to be counted.1 Assuming that the mals as perceived animal welfare.12,13 The five freedoms majority of the additional animals are mice as well, we lately have been advanced to the ‘‘five provisions’’ will focus on mice for the most part of this impulse focusing on what should be provided to achieve good paper, but the discussion that is being fostered here welfare.10 It was also recognized that animal welfare is will of course also apply to other experimental animal not static and thus concepts include adaptive capabil- species. ities in terms of coping with environmental challenges Additional animals are held available for breeding and/or being able to achieve certain goals.9,14–16 and maintenance of certain genetic lines, are killed for Today’s view on animal welfare also largely includes organ or tissue samples, or are considered surplus ani- a quality of living approach,17,18 focusing on good mals, which will not be used for experimental purposes living rather than mere avoidance of unfavorable con- due to wrong sex, age, or genotype (Table 1). In prin- ditions. Quality of living inherently reflects a much ciple, the same legislative rules for housing and main- more holistic view over an animal’s life and is thus tenance apply for the additional animals as for the less affected by short timed events (e.g. pleasurable experimental animals. Animal experimentation is con- moments, few minutes of fear).17 Our definition of ducted for a wide range of different scientific purposes,4 animal welfare is based on the current literature and and in many studies animals spend the longest time of deliberately reflects the difficulties that come along their lives not in the respective experiment itself. This is with defining as well as with assessing animal welfare. especially true for laboratory mice which are often con- Animal welfare describes (objectively verifiable) the sidered as ‘‘disposable goods’’ in science,5 and new status of a subjectively perceived quality of life of an experiments are usually carried out with new animals individual at a given period and is measured on an being bred in sufficient numbers in local facilities as well ordinal (nonlinear), multidimensional scale. as by commercial breeders. While waiting for the The core unit of animal welfare is the subjective per- experiment or after finishing a non-lethal experiment, ception of an individual. This makes welfare especially laboratory mice are usually living in customary stan- hard to measure, as individual perception seems to dardized laboratory housing conditions. In most coun- notoriously elude scientific quantification. However, tries these housing conditions fall under a variety of recent advances in theoretical concepts and method- restrictions with regard to meeting the minimum ology increasingly allow to quantify for example affect- requirements (e.g. for mice in the EU the minimum ive states.16,19–21 The multidimensionality results from cage floor size is 330 cm2, bedding, nesting material, the different levels on which animal welfare can be and social company shall be provided; see EU guide- affected: An individual might suffer from an injury lines 2010/63/EU). Above the minimum requirements, (an obvious indicator for bad welfare) but might be according to the 3Rs – which are anchored in many engaged in positive social interaction (indicator for statutory provisions – refinement of living conditions positive welfare) at the same time. Naturally, this com- shall be taken into consideration. Therefore, maximiz- plicates the assessment of animal welfare as calculation ing potential welfare by improving living conditions of of potential compensatory and/or additive effects experimental animals is not limited by legislations, but between different dimensions is inherently difficult. rather by experimental and economic reasons. Moreover, the difference between categories such as poor and very poor welfare is not necessarily the What is animal welfare? same as the difference between good and moderate welfare, thus the scale has to be considered ordinal. In the Animal welfare much alike human welfare is a term same vein, nonlinearity owes to the fact that physical that is notably hard to access and disentangle and and physiological parameters as well as descriptors of there is no unambiguous consensus.6–10 An early affective states do not follow simple mathematical addi- approach in defining animal welfare was raised by a tive rules in relation to their impact on animal welfare. press release of the Farm Animal Welfare Council in Although physiological parameters like heart rate or 1979, tracing back to the ‘‘Brambell Committee stress hormonal levels are measured on an interval 1965,’’11 and is referred to as the five freedoms (i.e. scale they are not linearly related to animal welfare.15,22 freedom from 1. hunger and thirst, 2. discomfort, Although the status of animal welfare is usually 3. pain, injury, and disease, 4. fear and distress, and assessed at a given point in time, the measurement 5. restrictions to express normal behavior). The five reflects a period of unknown length preceding the freedoms, however, state what has to be avoided in assessment. In addition, preceding lifetime events Table 1. Subjects categorized with regard to their relation to animal experimentation (rows) and different measures and constraints related to animal welfare (columns) that might be applicable.

Minimum Nesting Social Enrichment Space Treats Maximum Constraints providing food, water, ‘‘positive’’ e.g. climbing providing special maximizing welfare bedding, hygiene nesting material, social contact frames, toys, providing treats (e.g. fruits, without limitations status, health shelter, a place to in social objects to additional seeds, sweets, juice, related to an constraints related to Category Description monitoring, etc. sleep comfortable species manipulate space almond milk) experiment planned experiment

Breeding Breeding pairs, pregnant yes yes yes yes yes yes (restrictions might no yes stock females, males used apply) (e.g. separation, for breeding that are weaning age, being separated from attempts to stand- their breeding partner ardize parental influences) Breeding Animals held available as yes yes yes yes yes yes yes no reserve breeding reserve but not currently being paired Biological Animals of wrong sex or yes yes yes yes yes yes yes no surplus genotype for the experimental purpose Managed Animals of wrong age (too yes yes yes yes yes yes yes no surplus old) or weight for the experimental purpose Sentinel Sentinel animals for health yes yes yes yes yes yes/no yes/no (restricted by yes monitoring (same conditions as the purpose and (e.g. co-housing for animals being mode of use of the direct contact sen- monitored) sentinels) tinels, exposure to dirty bedding, or air exhaust in IVCs, social restrictions) Accompanying E.g. castrated males or yes yes yes yes yes yes no yes animals ovariectomized females (if not conflicting (if not conflicting (if not conflicting (if not conflicting (same housing used as social inter- with experimental with experimental with experimental with experimental conditions as action partners procedure) procedures) procedures) procedures) experimental animals) Waiting (Young) experimental ani- yes yes yes yes yes yes no yes mals waiting for reach- (if not conflicting (if not conflicting ing the desired age of with experimental with experimental experimentation procedures) procedures) Experimental Experimental animals yes yes yes yes yes no no yes animals currently in an ongoing (certain (if not conflicting (certain restrictions (certain (certain experiment restrictions with experimental might apply) restrictions restrictions might apply) procedures) might apply) might apply) Post- Animals not being killed yes yes yes yes yes yes yes no experimental after finishing the (should be moni- animals experiment; no subse- tored if switching quent experiments from single to planned social housing) Lewejohann et al. 153 affect a current state of animal welfare differentially (within and outside the experiment) has to be con- depending on frequency of occurrence, length, and sidered as well. intensity, for example. Specific conditions in animal experiments Assessing animal welfare For many years, it has been fostered to try keeping It is obvious from the above that measuring animal wel- external physical, social, and internal physiological fare is not an easy endeavor. It is, however, feasible to states as constant as possible. This was meant not measure animal welfare on an ordinal scale and to anno- only in order to standardize experimental conditions tate labels ranging from very poor to very good welfare but also to guarantee the fulfillment of animal needs. with reasonable precision. Poor welfare can be measured Unfortunately, this approach falls short in regard to by evaluating to what extent the first four freedoms are animal welfare as biological systems have evolved to met. In a broad sense, being free from pain, discomfort, cope with transience of external stimuli and therefore hunger and thirst, fear, and disease can be considered a allostasis (‘‘stability through change’’) rather than minimum standard that should be expected to be the homeostasis is a key element of animal welfare.15,34 normal state a laboratory animal is in. Although not This has been partially addressed by improving housing always overly trivial, these parameters are generally con- conditions especially by introducing environmental sidered to be measurable reasonably well.10,14,23 Above enrichment over the last decades. On the other hand, obvious signs for poor welfare like sickness behavior, experimental set ups naturally require a standardization wounds, signs of starvation or dehydration, physio- strategy (including, e.g., systematic variation to logical parameters (e.g. stress hormones, heart rate) increase external validity35) to minimize animal use may be indicative for how well an animal is able to and maximize test sensitivity. Still, boredom as a nat- cope with the challenges introduced by the environment ural consequence of under stimulation should be con- provided under laboratory conditions. The fifth freedom sidered a major concern with regard to animal welfare to be free to show normal behavior is far more difficult of laboratory animals.36 Sensation seeking is reflecting to assess. A wide spectrum of species specific behavior is such a need for change and has been measured for related to coping with challenging situations in the wild example as proneness to sensory stimuli in mice,37 that one can reasonably assume not to be applicable for self-administration of glucocorticoids in rats,38 or seek- laboratory animals (e.g. extensive foraging, predator ing even aversive stimuli in mink.39 Consequently, bal- avoidance, exaggerate aggressive encounters). ancing standardization against boredom along with the Therefore, it is not unequivocally established what con- animals’ ‘‘need for change’’ will remain a challenge in stitutes ‘‘normal behavior’’ in a laboratory animal. future experimental designs. It would be fallacious to Nevertheless, monitoring day-to-day behavior of labora- expect that any individual animal (or human) could be tory animals and comparing time budgets allocated to in a superior welfare state at all times. Thus fluctuation different behavioral domains is a feasible approach to in welfare states is an inherent part of an animal’s life analyze normal behavior in laboratory animals. In a and also contributes to a life worth living.14 Overall, similar vein, disturbed circadian rhythm or other behav- transience between welfare states within the range of ioral deviations such as stereotypic behavior or hair pull- very good, good, neutral, and even lightly aversive is ing are considered to be associated with impaired animal most likely part of an interesting life worth living. This, welfare.24–27 Finally, post mortem analysis (e.g. ulcers, however, is not at all easy to be realized for laboratory adrenal weights) can also help to retrospectively assess animals. Even if we assume that there were no restric- poor animal welfare.28 tions with regard to financial shortage, qualified per- Measuring good animal welfare on the other hand, is sonnel, and available space, at least some categories generally considered to be more complicated although of laboratory animals will be less eligible for the full not impossible.29 Play behavior and affiliative behav- range of possible welfare enhancement (see Table 1). iors, as well as some vocalizations,30 appear to be pro- mising measurable indicators for assessing positive Enhancing animal welfare animal welfare.31 In addition, recently newly developed approaches in human animal interaction were also Although assessing animal welfare is coming along with related to positive emotions, e.g. clicker-training and a number of problems with regard to accuracy, specifi- tunnel handling might indeed be perceived as positive city, and generalizability, there is also a pragmatic interactions by the animals.32,33 Finally, with regard to approach when the goal is to increase animal welfare the quality of living, which would reflect a more holistic of laboratory animals. A positive welfare state can be view of animal welfare, an ideal assessment should take derived by being able to engage in activities that are into account that welfare throughout an animal’s life perceived as rewarding. Such behaviors are expected to 154 Laboratory Animals 54(2) be capable to elicit positive affects which are related to work in order to get access to water or food). In addition, anticipation of achieving goals, achieving the goal itself, home environments could be improved by providing and retrospectively eliciting the memory of having pre- better opportunities for play behavior. Although play viously achieved a goal.10 Consequently, any measures behavior is most prominent in juveniles and adolescents, that enable laboratory animals to engage in rewarding adults of many species, including mice and rats,45–47 do activities, as well as states associated with anticipation or also play. Play behavior is usually considered to be an memory of rewarding activities are likely to enhance indicator for positive animal welfare,48 but sometimes animal welfare. A classical reward is the provision of even elicited when coping with negative affective treats, which is very common, e.g. in companion ani- states.49 Nevertheless, the absence of play behavior in mals. In laboratory animals, however, treats are often an otherwise playful species certainly is an example for restricted to experiments of operant and classical condi- a deviation from ‘‘normal’’ behavior and thus should tioning where special food items are provided as a generally considered to be an indicator for disturbed reward to increase their performance. If paying attention animal welfare. Adult mice engage frequently in loco- to nutritional needs, there should be no principle objec- motor play if provided with enough space.45,46 Indeed, tions against providing special treats to other laboratory more than 85% of play behavior in mice involves loco- animals as well. However, one should bear in mind that motor play.50 Therefore providing more space (e.g. larger if anticipated rewards are suspended the mismatch cages, connecting several small cages with tubes) or other between expected reward and the reality check possibly opportunities to engage in locomotory activity should be leads to frustration.40 Therefore, withholding treats or taken into account to improve housing conditions for other positive stimuli that previously have been granted laboratory mice. Noteworthy, there is an ongoing can also negatively affect animal welfare. debate with regard to the costs and benefits of changing Positive affective states are also elicited in positive ‘‘established’’ housing conditions with regard to size, type social interactions. Social interactions concern the entire of nesting and bedding material, or different forms of life of social mammals and incisive experiences in early enrichment. For example, excessive usage of running life also affect later social behavior. For example, it has wheels might resemble stereotypic behavior in some indi- been shown that delayed weaning increased social behav- viduals,51,52 but in group housed mice no signs of stereo- ior later in life.41 However, the weaning age is usually typic running wheel behavior were found.53 Moreover, designed to maximize breeding success and does not several behavioral as well as morphological, and physio- necessarily correspond to the natural breeding behavior logical parameters can be affected by introducing envir- of the species. Social housing for laboratory animals later onmental enrichment.54 Also it is known that housing in life is generally recommended except for solitary spe- conditions can have interaction effects with pharmaco- cies. However, group housing for animals that frequently logical treatments.55 However, concerns that enrichment engage in aggressive encounters, e.g. as observed in male generally increase variation in experimental results could groups of many mouse strains,42,43 is sometimes not feas- not be substantiated.56 Overall, possible interferences of ible. This is something that has to be taken into account improved as well as of restricted housing conditions with when planning experiments and choosing the right model the experimental design, reproducibility, and external val- species, strain, or sex. As already outlined above, bore- idity should be kept in mind. Enrichment is generally dom due to a lack of stimuli and missing opportunities to thought to enhance animal welfare although sex differ- engage in rewarding activities in laboratory housing sys- ences might apply (i.e. aggressive behavior42), and it is tems is a growing concern.36 This can be partially ame- not always clear how different items are perceived by the liorated by means of environmental enrichment and animals themselves and thus animal centric strategies like providing materials to perform species typical behavior preference tests will help to assess and rate different (e.g. for rodents, nesting material, burrowing and gnaw- items.57 ing substrate). For laboratory mice nesting material and shelters were slowly introduced over the last three dec- Improving animal welfare in- and outside ades and can nowadays be found in almost all European the experiment animal facilities as this is required by the EU directive 2010/63/EU. Still there is much room for improvement Generally speaking, we should aim to maximize animal with regard to entertaining enrichment and providing welfare of laboratory animals owing to the fact that we opportunities to engage in rewarding behaviors. This are responsible for their well-being. As animal experi- can be realized by providing novel stimuli (e.g. new mentation is under special scrutiny there is a high eth- enrichment items that can be explored44), by introducing ical standard ruling animal experimentation and it has cognitive training (e.g. puzzle boxes, clicker training32) become mandatory to consider refinement measures in into the home environment, or by measures of occupa- the experimental design. Moreover, it is widely tional therapy (e.g. running wheels, or letting the animals accepted that above the ethical concerns there is also Lewejohann et al. 155 a scientific need for improving the welfare of laboratory 2. BMEL. BMEL - U¨bersicht: BMEL informiert u¨ber animals.58 However, restrictions affecting the welfare of Tierschutz - Verwendung von Versuchstieren im Jahr experimental animals cannot always be overcome if 2018, https://www.bmel.de/DE/Tier/Tierschutz/_texte/ they are directly related to the experimental aims. In Versuchstierzahlen2018.html (2019, accessed 2 January our view, relating to historical data or established hous- 2020). ing conditions alone does not suffice to refuse enhan- 3. Fire, Licensing and Public Order Analysis Unit of the HOU. Annual statistics of scientific procedures on cing the living conditions. One of the main lessons to be living animals in Great Britain 2018, https://assets.pub- learnt from the reproducibility crisis should be that lishing.service.gov.uk/government/uploads/system/ only data that can be replicated in other contexts are 35 uploads/attachment_data/file/835935/annual-statistics- truly biologically meaningful. Experimental animals scientific-procedures-living-animals-2018.pdf (2019, spend much of their lives outside the experiment and accessed 27 November, 2019). a large number of animals are not even used for experi- 4. Bert B, Do¨rendahl A, Leich N, et al. Rethinking 3R stra- ments but held available for breeding or other mainten- tegies: digging deeper into AnimalTestInfo promotes ance related purposes. Table 1 summarizes the potential transparency in in vivo biomedical research. PLoS Biol of enhancing animal welfare for different categories of 2017; 15: e2003217. laboratory animals. For each category of animals one 5. Brust V, Schindler PM and Lewejohann L. Lifetime should ask what can reasonably be done to maximize development of behavioural phenotype in the house their welfare. Experimental animals and animals wait- mouse (Mus musculus). Front Zool 2015; 12: S17. ing for an experiment are probably more restricted with 6. Fraser D. Science, values and animal welfare: exploring regard to maximizing their welfare as many measures the ‘inextricable connection’. Anim Welf 1995; 4: 103–117. potentially counteract the experimental purpose. 7. Stafleu F, Grommers F and Vorstenbosch J. Animal wel- Nevertheless, quality and size of measures to increase fare: evolution and erosion of a moral concept. Anim welfare depend on the experimental design and should Welf 1996; 5: 225–234. be evaluated accordingly. On the other hand, for post- 8. Fisher MW. Defining animal welfare: does consistency experimental animals there is basically no limit on what matter? NZ Vet J 2009; 57: 71–73. could be done to increase their welfare, even if it might 9. Ohl F and van der Staay FJ. Animal welfare: at the inter- only be for a short time compared to the life expect- face between science and society. Vet J 2012; 192: 13–19. ancy.5 All in all, the time outside the experiment can be 10. Mellor DJ. Updating animal welfare thinking: moving considered a special opportunity to improve the overall beyond the ‘‘five freedoms’’ towards ‘‘a life worth welfare of laboratory animals. living’’. Animals 2016; 6: 21. 11. Brambell F. Report of the technical committee to enquire into the welfare of animals kept under intensive livestock Declaration of Conflicting Interests husbandry systems, https://www.worldcat.org/title/ The author(s) declared no potential conflicts of interest with report-of-the-technical-committee-to-enquire-into-the- respect to the research, authorship, and/or publication of this welfare-of-animals-kept-under-intensive-livestock-hus- article. bandry-systems/oclc/475844139 (1965, accessed 1 March 2019). Funding 12. Dawkins MS. From an animal’s point of view: motivation, fitness, and animal welfare. Behav Brain Sci 1990; The author(s) disclosed receipt of the following financial sup- 13: 1–9. port for the research, authorship, and/or publication of this 13. Bracke MBM and Hopster H. Assessing the importance article: This work was funded by the DFG (grant numbers of natural behavior for animal welfare. J Agric Environ FOR2591, LE 2356/5-1, and JI 276/1-1). Ethics 2006; 19: 77–89. 14. Broom DM. Indicators of poor welfare. Br Vet J 1986; ORCID iDs 142: 524–526. Lars Lewejohann https://orcid.org/0000-0002-0202-4351 15. Korte SM, Olivier B and Koolhaas JM. A new animal Christine Ha¨ger https://orcid.org/0000-0002-6971-9780 welfare concept based on allostasis. Physiol Behav 2007; Paulin Jirkof https://orcid.org/0000-0002-7225-2325 92: 422–428. 16. Gygax L. Wanting, liking and welfare: the role of affective states in proximate control of behaviour in verte- References brates. Ethology 2017; 123: 689–704. 1. European Commission. 2019 report on the statistics on 17. Broom DM. Quality of life means welfare: how is it the use of animals for scientific purposes in the Member related to other concepts and assessed? Anim Welf 2007; States of the European Union in 2015-2017, https://ec. 16: 45–53. europa.eu/transparency/regdoc/rep/1/2020/EN/COM- 18. Green TC and Mellor DJ. Extending ideas about animal 2020-16-F1-EN-MAIN-PART-1.PDF (2020, accessed 6 welfare assessment to include ‘quality of life’ and related February 2020). concepts. NZ Vet J 2011; 59: 263–271. 156 Laboratory Animals 54(2)

19. Paul ES, Harding EJ and Mendl M. Measuring emo- 37. Olsen CM and Winder DG. Operant sensation seeking tional processes in animals: the utility of a cognitive engages similar neural substrates to operant drug seek- approach. Neurosci Biobehav Rev 2005; 29: 469–491. ing in C57 mice. Neuropsychopharmacology 2009; 34: 20. Bethell EJ, Holmes A, MacLarnon A, et al. Cognitive bias 1685. in a non-human primate: Husbandry procedures influence 38. Piazza PV, Deroche V, Deminie` re JM, et al. cognitive indicators of psychological well-being in captive Corticosterone in the range of stress-induced levels rhesus macaques. Anim Welf 2012; 21: 185–195. possesses reinforcing properties: implications for sensa- 21. Kloke V, Schreiber RS, Bodden C, et al. Hope for the tion-seeking behaviors. Proc Natl Acad Sci 1993; 90: best or prepare for the worst? Towards a spatial cognitive 11738–11742. bias test for mice. PLoS One 2014; 9: e105431. 39. Meagher RK and Mason GJ. Environmental enrichment 22. Hau M and Goymann W. Endocrine mechanisms, behav- reduces signs of boredom in caged mink. PLoS One 2012; ioral phenotypes and plasticity: known relationships and 7: e49180. open questions. Front Zool 2015; 12: S7. 40. Burokas A, Gutie´rrez-Cuesta J, Martı´n-Garcı´a E, et al. 23. Broom DM. Assessing welfare and suffering. Behav Operant model of frustrated expected reward in mice. Processes 1991; 25: 117–123. Addict Biol 2012; 17: 770–782. 24. Spa¨ni D, Arras M, Ko¨nig B, et al. Higher heart rate of 41. Curley JP, Jordan ER, Swaney WT, et al. The meaning of laboratory mice housed individually vs in pairs. Lab Anim weaning: influence of the weaning period on behavioral 2003; 37: 54–62. development in mice. Dev Neurosci 2009; 31: 318–331. 25. Reinhardt V. Hair pulling: a review. Lab Anim 2005; 39: 42. Kappel S, Hawkins P and Mendl M. To group or not to 361–369. group? good practice for housing male laboratory mice. 26. Mason G and Latham N. Can’t stop, won’t stop: is Animals 2017; 7: 88. stereotypy a reliable animal welfare indicator? Anim 43. Weber EM, Dallaire JA, Gaskill BN, et al. Aggression in Welf 2004; 13: 57–69. group-housed laboratory mice: why can’t we solve the 27. Wu¨rbel H, Stauffacher M and Holst D. Stereotypies in Lab Anim laboratory mice: quantitative and qualitative description problem? 2017; 46: 157–161. 44. Herring A, Lewejohann L, Panzer AL, et al. Preventive of the ontogeny of ‘wire-gnawing’ and ‘jumping’ in and therapeutic types of environmental enrichment coun- Zur:ICR and Zur:ICR nu. Ethology 1996; 102: 371–385. 28. David JM, Knowles S, Lamkin DM, et al. Individually teract beta amyloid pathology by different molecular ventilated cages impose cold stress on laboratory mice: a mechanisms. Neurobiol Dis 2011; 42: 530–538. source of systemic experimental variability. J Am Assoc 45. Marashi V, Barnekow A, Ossendorf E, et al. Effects of Lab Anim Sci 2013; 52: 738–744. different forms of environmental enrichment on behav- 29. Scott E, Nolan A, Reid J, et al. Can we really measure ioral, endocrinological, and immunological parameters in animal quality of life? Methodologies for measuring qual- male mice. Horm Behav 2003; 43: 281–292. ity of life in people and other animals. Anim Welf 2007; 46. Freund J, Brandmaier AM, Lewejohann L, et al. 16(Suppl): 17–24. Association between exploratory activity and social indi- 30. Panksepp J and Burgdorf J. ‘Laughing’ rats and the evo- viduality in genetically identical mice living in the same lutionary antecedents of human joy? Physiol Behav 2003; enriched environment. Neuroscience 2015; 309: 140–152. 79: 533–547. 47. Reinhold AS, Sanguinetti-Scheck JI, Hartmann K, et al. 31. Boissy A, Manteuffel G, Jensen MB, et al. Assessment of Behavioral and neural correlates of hide-and-seek in rats. positive emotions in animals to improve their welfare. Science 2019; 365: 1180–1183. ˇ Physiol Behav 2007; 92: 375–397. 48. Held SDE and Spinka M. Animal play and animal wel- 32. Leidinger C, Herrmann F, Tho¨ne-Reineke C, et al. fare. Anim Behav 2011; 81: 891–899. Introducing clicker training as a cognitive enrichment 49. Ahloy-Dallaire J, Espinosa J and Mason G. Play and for laboratory mice. J Vis Exp 2017; 121: e55415. optimal welfare: does play indicate the presence of posi- 33. Clarkson JM, Dwyer DM, Flecknell PA, et al. Handling tive affective states? Behav Processes 2018; 156: 3–15. method alters the hedonic value of reward in laboratory 50. Pellis SM and Pasztor TJ. The developmental onset of a mice. Sci Rep 2018; 8: 2448. rudimentary form of play fighting in C57 mice. Dev 34. McEwen BS and Wingfield JC. The concept of allostasis Psychobiol 1999; 34: 175–182. in biology and biomedicine. Horm Behav 2003; 43: 2–15. 51. Sherwin CM. Voluntary wheel running: a review and 35. Richter SH, Garner JP, Zipser B, et al. Effect of popula- novel interpretation. Anim Behav 1998; 56: 11–27. tion heterogenization on the reproducibility of mouse 52. Richter SH, Gass P and Fuss J. Resting is rusting: a crit- behavior: a multi-laboratory study. PLoS One 2011; 6: ical view on rodent wheel-running behavior. Neuroscience e16461. 2014; 20: 313–325. 36. Burn CC. Bestial boredom: a biological perspective on 53. Weegh N, Fu¨ner J, Jahnke O, et al. Wheel running behav- animal boredom and suggestions for its scientific investi- iour in group-housed female mice indicates disturbed gation. Anim Behav 2017; 130: 141–151. wellbeing due to DSS colitis. Lab Anim. Epub ahead of Lewejohann et al. 157

print 1 November 2019. DOI: 10.1177/002367721987 56. Bailoo JD, Murphy E, Boada-San˜ a M, et al. Effects of 9455). cage enrichment on behavior, welfare and outcome vari- 54. Bayne K. Environmental enrichment and mouse models: ability in female mice. Front Behav Neurosci 2018; 12: current perspectives. Anim Model Exp Med 2018; 1: 232. 82–90. 57. Habedank A, Kahnau P, Diederich K, et al. Severity 55. Hoffmann LC, Schu¨tte SRM, Koch M, et al. Effect of assessment from an animal’s point of view. Berl Munch ‘‘enriched environment’’ during development on adult rat Tierarztl Wochenschr 2018; 18007. behavior and response to the dopamine receptor agonist 58. Poole T. Happy animals make good science. Lab Anim apomorphine. Neuroscience 2009; 158: 1589–1598. 1997; 31: 116–124.

Re´sume´ Le bien-eˆtre des animaux est une preóccupation socie´tale croissante et le bien-eˆtre des animaux utilise´sa` des fins expe´rimentales fait l’objet d’une attention particulie`re. La grande majorite´ des animaux de laboratoire sont des souris qui vivent dans de petites cages n’offrant pas beaucoup de varie´te´. En outre, la proce´dure expe´rimentale prend souvent tre`s peu de temps par rapport a` la pe´riode d’e´levage de ces animaux jusqu’a` l’aˆge de´sire´ ou leur mise a` disposition pour l’expe´rimentation animale. Toutefois, pour l’e´valuation du bien-eˆtre animal, le temps passe´ en attente d’une expe´rience ou le temps passe´ apre`s avoir termine´ une expe´rience doit e´galement eˆtre pris en compte. En plus des animaux de laboratoire, de nombreux animaux (par exemple, pour l’e´levage et l’entretien des lignes geńe´tiques, les animaux exce´dentaires) sont lie´sa` l’expe´rimentation animale et font geńe´ralement face a` des conditions de vie similaires. Par conse´quent, en termes d’ame´lioration du bien-eˆtre geńe ´ral des animaux de laboratoire, il n’existe pas seulement un besoin d’ame´lioration des conditions expe´rimentales, mais surtout d’ame´lioration des conditions de vie en dehors de l’expe´rience. L’ame´lioration du bien-eˆtre des animaux de´pend donc dans une large mesure des conditions de logement et d’entretien de tous les animaux lie´sa` l’expe´rimentation animale. E´tant donne´ l’e´tat actuel de la recherche sur le bien-eˆtre des animaux, il existe en effet un grand potentiel d’ame´lioration du bien-eˆtre geńe´ral des animaux de laboratoire.

Abstract Der Tierschutz ist ein zunehmend wichtiges gesellschaftliches Anliegen, und das Wohlergehen von fu¨r Versuchszwecke dienenden Tieren muss besonders strenger U¨berpru¨fung unterzogen werden. Die u¨berwie- gende Mehrheit von Versuchstieren sind Ma¨use, die in kleinen Ka¨figen ohne viel Abwechslung leben. Hinzu kommt, dass die Dauer der Zu¨chtung der Tiere bis zum erforderlichen Alter bzw. der Haltung in Vorbereitung auf Versuche oft lang ist, wa¨hrend die Versuche selbst nur sehr wenig Zeit in Anspruch nehmen. Bei der Bewertung des Tierschutzes ist jedoch auch die den eigentlichen Versuchen vorausgehende Zeit ebenso wie jene nach Abschluss eines Experiments zu beru¨cksichtigen. Neben den Versuchstieren selbst sind viele weitere Tiere (z. B. solche zur Zu¨chtung und Erhaltung von genetischen Linien, u¨berza¨hlige Tiere) mit Tierversuchen verbunden, die in der Regel unter a¨hnlichen Lebensbedingungen gehalten werden. Im Hinblick auf die Verbesserung des allgemeinen Wohlergehens von Versuchstieren besteht daher nicht nur die Notwendigkeit einer Verbesserung der Versuchsbedingungen, sondern insbesondere auch der Lebensbedingungen außerhalb der Versuche. Die Verbesserung des Tierschutzes ha¨ngt daher in hohem Maße von den Haltungs- und Unterbringungsbedingungen aller im Zusammenhang mit Versuchen stehenden Tieren ab. Nach dem derzeitigen Stand der Tierschutzforschung gibt es in der Tat großes Potenzial, das zur Verbesserung des allgemeinen Wohlergehens von Versuchstieren ausgescho¨pft werden sollte.

Resumen El bienestar animal es una creciente preocupacioń social y el bienestar de los animales utilizados para experimentos esta´ siendo analizado detenidamente. La gran mayorıá de animales de laboratorio son roedores que viven en jaulas pequenãs que no ofrecen gran variedad. Asimismo, el procedimiento experimental a menudo es de poca duracioń en comparacioń con el tiempo en que estos animales han sido criados hasta alcanzar la edad deseada o estań disponibles para experimentar con ellos. Sin embargo, para la evaluacioń del bienestar animal, el tiempo de espera para un experimento o el tiempo transcurrido tras finalizar un 158 Laboratory Animals 54(2) experimento son factores que tambień tienen que considerarse. Adema´s de los animales para experimentos, muchos otros animales (p. ej., para criar y mantener lıńeas gene´ticas o para tener un excedente de ejem- plares) estań relacionados con la experimentacioń animal y a menudo viven en condiciones similares. Por tanto, en lo referente a la mejora del bienestar general de los animales de laboratorio, no solo hay una necesidad de refinar las condiciones experimentales sino especialmente de mejorar las condiciones de vida ma´s alla´ del experimento. La mejora del bienestar animal, por tanto, depende en gran medida de las condiciones de mantenimiento y de las jaulas de los animales de experimentacioń. Dado el estado actual de la investigacioń sobre el bienestar animal, existe gran potencial para mejorar el bienestar general de los animales de laboratorio. Original Article Laboratory Animals 2020, Vol. 54(2) 159–170 ! The Author(s) 2019 Diet-regulated behavior: FVB/N mice Article reuse guidelines: sagepub.com/journals-permissions fed a lean diet exhibit increased nocturnal DOI: 10.1177/0023677219834582 journals.sagepub.com/home/lan bouts of aggression between littermates

Mandi M Murph1 , Shuying Liu2,3, Wei Jia1, Ha Nguyen1, Megan A MacFarlane1, Susan S Smyth4, Sudeepti S Kuppa1 and Kevin K Dobbin5

Abstract The hyperactive FVB/N inbred mouse strain is widely used for transgenic research applications, although rarely for behavioral studies. These mice have visual impairments via retinal degeneration, but are considered highly intelligent and rely largely on olfaction. While investigating diet-induced obesity in autotaxin transgenic FVB/N mice, we observed an increase in the necessity for male, but not female, cage separations. Based on the observations, we hypothesized that feeding FVB/N mice a lean diet increases nocturnal bouts of aggression between male littermates. The diets of adult littermates were switched from normal chow to either ad libitum high-fat (45% fat) or lean (10% fat) matched diets for 27 weeks, whereby the mice reached an average of 43 g versus 35 g, respectively. Then, cage separations due to nocturnal bouts of aggression became mandatory, even though littermates peacefully cohabitated for 10–16 weeks previously. Since the data was of an unusual nature, it required uncommon statistical methods to be engendered to evaluate whether and where significance existed. Therefore, utilizing the randomization and population models, we established a methodology and postulated that either testosterone, the autotaxin transgene or diet alteration was the causal factor. Statistical evaluation demonstrated a significant correlation between cage separations and aggressive behavior associated with the lean-diet-fed mice, not autotaxin. Biochemical data did not appear to explain the behavior. In contrast, energy metabolism highlighted differences between the groups of normally hyperactive mice by diet. This characteristic makes FVB/N male mice unsuitable subjects for long-term studies with lean-diet modifications.

Keywords FVB/N, behavior, mathematical computation, energy, lean diet

Date received: 21 August 2018; accepted: 6 February 2019

1Department of Pharmaceutical and Biomedical Sciences, The University of Georgia, College of Pharmacy, Athens, GA, USA The FVB/N inbred mouse strain is widely used in 2Department of Breast Medical Oncology, The University of Texas, laboratory research applications. It is spontaneously Houston, TX, USA hyperactive, only slightly more anxious when compared 3Department of Systems Biology, The University of Texas, Houston, to C57BL/6 mice, which are also widely used for TX, USA 4 research.1,2 FVB/N mice possess unique characteristics Division of Cardiovascular Medicine and Department of Pharmacology, The University of Kentucky, Lexington, KY, USA; that make them suitable to generate transgenic animals Department of Veterans Affairs Medical Center, Lexington, KY, for experiments, including their high mating frequency, USA their consistent ability to produce large, healthy litters 5College of Public Health Epidemiology and Biostatistics, The and their zygote survivability.3 Female FVB/N mice University of Georgia, Athens, GA, USA exhibit strong maternal behavior whereby moms dis- Corresponding author: play exceedingly low rates of infanticide or cannibal- Mandi M Murph, University of Georgia, 240 West Green Street, ism, which allows robust numbers of pups to survive, Athens, GA 30602, USA. 4 even with a litter size of 12 and an average of 9.5. Email: [email protected] 160 Laboratory Animals 54(2)

Although FVB/N mice are described as ‘unsuitable cohabitating groups of adult male FVB/N mice for behavioral studies’ because they have visual impair- through diet modification to lean chow with 10% fat. ments via retinal degeneration, that conclusion is based To our knowledge, this is the first description of such largely on the reliance of commonly used tests that a scenario in FVB/N, largely due to the non-reliance require acute vision to perform the function, not olfac- on the strain for behavioral research. Therefore, the tion.5,6 For example, Morris water navigation tasks to behavior of this model makes the males unsuitable for find a visible platform and novel object recognition are long-term studies using lean-diet modifications; how- significantly easier without impaired vision. However, ever, this model could be ideal to study male mouse more recent studies show FVB/N mice are superior in aggression that is not reliant on visual stimulation. learning and memory performance over C57BL/6 mice using an olfactory tube maze test.7 This suggests that vision, not intelligence, has led to false impressions Materials and methods about what constitutes normal behavior in mice. Animals There are somewhat conflicting studies surrounding the degree of aggressive behavior in FVB/N males. One All protocols for the use of vertebrate animals at the study concluded that FVB/N males have only ‘some- University of Georgia were approved by The what higher levels of aggressiveness’ when compared to Institutional Animal Care and Use Committee. The C57BL/6 males.1 In contrast, another study concluded Guide for the Care and Use of Laboratory Animals is that in comparison to C57BL/6 male mice, FVB/N endorsed in this study. Transgenic mice were the kind male mice display a higher duration of offensive behav- gift of Drs Gordon B Mills and Susan S Smyth and, as ior and significantly longer attack durations against described previously, generated on an FVB/N genetic intruder mice, which might result from their inability background.11 Pups were weaned after 21 days and fed to maintain regular circadian rhythm patterns.8 Studies ad libitum water and standard rodent chow containing on aggression typically focus on behavior among males, kcal% of 30% protein, 57% carbohydrate and 13% fat not females. (Lab Diet, #5001, St Louis, MO) prior to experiments Initially, we set out to conduct a study about employing the diet modifications. There were no restric- the interaction between diet-induced obesity with auto- tions on the quantity of food or water available and taxin transgene expression. Transgenic mice overex- cages were kept well stocked with pellets. Polymerase pressing autotaxin have larger adipocytes, biological chain reaction (PCR) analysis was used to confirm the differences with obesity and develop spontaneous genotyping and expression from ear clips with an cancers after 12 months of age.9,10 Herein, we observed Agilent 2100 bioanalyzer (Agilent, Santa Clara, CA) an unexpected and striking increase in FVB/N male and the following primers: 5’-GATCCCAGCCAG aggression among certain cages, which necessitated per- TGGACTTA-30 (forward) and 50-TCTGACACGACT manent separations. This required incessant cage GGAA CGAG-30 (reverse). In total, 140 male mice patrolling and routine separations throughout the (106 AT-ATX transgenic and 34 wild-type) were fed course of the experiment. with standard rodent chow, for which the data are Due to the unusual nature of the data, we required included in this study. the creation of a new statistical method to evaluate After approximately 10–16 weeks of age, transgenic whether statistical significance existed within these (39 male) and wild-type (17 male) mice were then observations. Several aspects of these data combine re-assigned to a different diet, from initially eating to prevent it from fitting neatly into any existing data standard chow. Mice were fed either a high-fat diet analysis paradigm. The data are both clustered and with kcal% of 20% protein, 35% carbohydrate and partially missing. The clustering is due to the treatment 45% fat (Research Diet D12451, New Brunswick, NJ) assignment being received by cages of animals, rather or the matched control, lean diet that included 17% than individual animals. The resulting data structure sucrose, lean diet of 20% protein, 70% carbohydrate is similar to a group randomized clinical trial. The par- and 10% fat (Research Diet D12450H). Male cages tial missing-ness is due to the way aggressive behavior is required significant management and cage separations objectively measured by cage separation – aggressive for fighting and wounding, as governed by the Animal animals are isolated in individual cages; however, Use Protocol and recommendations of the veterinarian animal separation is an imperfect surrogate for attendant. The diet study in male mice lasted 27 weeks. aggressiveness. The mouse facility automates the lighting in the Herein we describe the study and the statistical rooms on a 12-hour cycle, whereby the lights come on computations along with the data that inspired the at 7:00 am EST and then go off at 7:00 pm EST. mathematical exploration. Based upon these param- The temperature is maintained at 72F 4F. The eters, we conclude that violent behavior ensues among mice are housed in Individually Ventilated Cages Murph et al. 161

(IVC) mouse DGM (digital ready green line mouse) theorem-based test), so that reliable p-values can be racks and Sealsafe Plus GM500 cages measuring generated with this approach. On the other hand, the 15.40 7.83 6.30 in (Tecniplast, West Chester, PA). population model has the advantage of permitting a The area of the cage floor is 77.66 in2. The mouse cages more detailed description of the effect size and uncer- include Bed-o’CobsÕ bedding and one Bed-r’NestÕ tainty estimates for comparisons of treatment groups portion-controlled nesting material (The Andersons, and control. Inc, Maumee, OH). For enrichment, each cage includes one plastic red mouse house that allows only red wave- The unusual data structure length light to pass through. Aggressive behavior was flagged by our veterinary Unlike all the humane endpoints governing this proto- staff. The animal cages are checked daily for health col, the endpoint of aggression has an unusual statis- and hygiene monitoring. Tail handling is the main tical structure. First, aggressiveness cannot be observed type of handling. The facility also performs full changes in isolation, but is a behavior that requires the presence of the cages every other week. During the daily moni- of other mice in the same cage. When no other mice are toring, they observed heightened and unexpected fight- present, then it is no longer possible to observe aggres- ing without cessation, wounding, chasing and biting siveness for the remaining mouse. This gives rise to a within the cages. Most often this was concurrent with specific type of right censoring that must be taken into bloody bedding, visual wounds and open cuts on vic- consideration during the analysis. Second, the data are tims, whereas perpetrators perched on top of mouse also clustered by cages, so that proper analysis requires houses and lacked extensive bleeding wounds. Cages taking into account the clustering structure and pos- were cleaned on Monday mornings, every other week, sible correlations of observations in the same cage. whereas separations and deaths from aggressive bouts Third, the number of cage separations can be modeled occurred randomly throughout the week. All of these as a discrete random variable taking on values between observations required intervention and cage separ- 0 and the total number of animals in a cage minus one. ations per guidelines for care of the animals. As these The induced distribution due to the right censoring will results were unexpected, there was no blinding. be non-standard and must be mathematically calcu- Humane endpoints were based on an overlapping lated (see Supplemental Figure 2 for calculations). study.10 Prior to this study, we did not foresee aggressive bouts occurring to this degree. The humane endpoints Randomization model methodology governing our animal use protocol included: excessive tumor burden, inability to move, seizures, abdominal dis- The sharp null hypothesis used for the randomization tension that interferes with mobility, inability to feed, model states that there is no effect on the response severe lethargy with little/no spontaneous movement, dif- (aggressiveness) for any of the treatments. Any observed ficult or labored breathing, blue mucus membranes, differences are, therefore, attributable to the randomiza- bleeding from deep wounding or severe ulceration tion itself – that is, to the proverbial ‘luck of the draw.’ where there is an absence of an epidermal layer overlying Therefore, every other possible randomization would the skin that is >50% of the wound size. At the conclu- have resulted in the exact same set of observations. The sion of the study, animals were euthanized with a method ‘unusualness’ of the observed data can, thus, be quanti- consistent with the guidelines of the American Veterinary fied by a permutation test. For this experiment, there is Medical Association. one control condition and four different treatment conditions. The number of cages per condition were 20 for Statistical methodology the control, and 4, 4, 3, 3 for the different treatments, for a total of 34 cages. Therefore, the total number of dis- The treatments and control in this experiment were ran- tinct permutations of the cages is as follows: domly assigned to the cages. The effects of the treatments on aggressiveness can, therefore, be evaluated 34 14 10 6 6 1015 using two different statistical modeling approaches. 20 4 4 3 We examine both approaches to analyze these data. The first approach is based on a randomization model This number suggests that we can obtain a very and a sharp null hypothesis.12 The second approach is small p-value for the test. based on a population model and a null hypothesis 13 about population parameters. For our purposes, Population model analysis the randomization model has the advantage that the limited sample size will not invalidate a randomiza- For the population model approach, the natural popu- tion-based permutation test (unlike a central limit lation parameter of interest is the proportion of animals 162 Laboratory Animals 54(2) in the target population of mice that turn aggressive nitrocellulose membranes on ice for 90 min at 100 V. when exposed to the control or a specific treatment. Membranes were blocked in milk buffer at room tem- Let p0 be the population proportion for the control, perature for 1 h and incubated on a shaker plate with and p1,p2,p3,p4 be the population proportions for Androgen receptor AN4441 (MA5-13426, Thermo each of the active treatments. We assume that being Fisher Scientific, Rockford, IL) overnight at 4 C. separated is an adequate surrogate for turning aggres- Membranes were washed with Tris-buffered saline with sive.14 As noted previously, the data are partially miss- Tween 20 (TBST) three times, 10 min each, then incu- ing because if all the mice end up separated one does bated with secondary for 120 min at room temperature not know whether all turned aggressive, or all but one and washed again, as stated previously. turned aggressive. Also, there may be cage conditions Chemiluminescent substrate Dura (SuperSignal West that impact the proportion of animals that turn aggres- Dura extended duration substrate, Thermo Fisher sive, so different cages in the same condition are per- Scientific, Rockford, IL) was applied to the membrane mitted to have different proportions. for 5 min and then imaged for 6 min. A glyceraldehyde 3- phosphate dehydrogenase (GAPDH) image was Biochemical analysis obtained the same way with a regular chemiluminescent substrate (SuperSignal West Pico, Thermo Fisher Testosterone content in serum was determined using a Scientific, Rockford, IL) applied for 3 min and imaged colorimetric competitive enzyme immunoassay ELISA for 2 min. Imaging was obtained with a Li-cor Odyssey kit (Enzo Life Sciences, Farmingdale, NY) following Fc (Li-cor Biosciences, Lincoln, NE). the protocol according to the manufacturer’s instructions. A calibration curve was created using the testos- Software and calculations terone standard included in the kit, which has a range from 2000 pg/mL to 7.81 pg/mL. The absorbance was Calculations were carried out in RStudio with R ver- read at 405 Nm using a SpectraMax M2 plate reader sion 3.4.1. GraphPad Prism was used to create graph- (Molecular Devices, Sunnyvale, CA). ical presentations and assess the statistical significance The adenosine diphosphate/adenosine triphosphate among some of the data, where indicated. (ADP/ATP) ratio was determined using single-cell suspensions with a luciferin–luciferase bioluminescent assay kit (Abcam, Cambridge, MA) according to Results manufacturer’s instructions. The ratio was calculated Diet-induced aggression in male based on a calibration plot made with a standard FVB/N mice ATP solution with a range from 1 nM to 10 nM (ATP Disodium Trihydrate, Amresco, Solon, OH). Single-cell During a diet-induced obesity study using FVB/N suspensions were created using 200 mg of flash-frozen inbred mice, we observed an unexpected manifestation tissues and dissociated with collagenase type I (Stemcell of aggression among male, but not female, cages that Technologies, Vancouver, Canada) and suspended in encumbered the investigation. In the colony containing phosphate buffered saline (PBS). Luminescence was several hundred FVB/N mice, cages were provided with determined using a Molecular Devices SpectraMax a diet of standard rodent chow ad libitum. The female M5 plate reader (Sunnyvale, CA). The data are plotted cages required no separations for aggressive behavior as a scatterbox using GraphPad Prism to display the (fighting, wounding, chasing, biting, blood discovered average and range of the data points. in the cage, etc.). During the same period, the male cages required only a few separations (Figure 1(a)). Immunoblotting Diets were altered in a small group of the animals after 10–16 weeks to measure weight changes leading to adult- For protein resolution, liver tissue from mice was onset, diet-induced obesity. In this regard, some male homogenized in 500 mL of lysis buffer (RIPA buffer cages were provided ad libitum high-fat (45%) or lean with a protease inhibitor in a 1:100 ratio). Buffered (10% fat) diets over a period of 27 weeks. Males were tissue was sonicated, centrifuged and the supernatant assigned a cage based on their genotype (Supplementary extract was obtained. Protein concentrations were Figure 1) and littermate status before the food was altered, determined by a bovine serum albumin (BSA) protein with a stringent prerequisite to keep siblings together and assay (Pierce BSA protein assay kit, Thermo Fisher encourage peaceful cohabitation (Figure 1(b)). Scientific, Rockford, IL). Approximately 30 mg per Over the 27 weeks of the diet experiment, the weight sample was resolved using sodium dodecyl sulfate-poly- of the males in the respected cages began to increase or acrylamide gel electrophoresis (SDS-PAGE) for stabilize, based on the food provided (Figure 2(a)). 120 min at 110V. Proteins were then transferred to The final average weights of the lean-diet-fed cages Murph et al. 163

Figure 1. FVB/N males, but not females, are more likely to require cage separations. (a) Bar graphs showing the number of total cages required overall with a standard, chow diet. This is reflective of typical cage separations over a four-month time period with FVB/N mice and standard conditions of sterile bedding, supplemental nesting material, mouse houses, ad libitum food and water. Males began with 5 littermates in each of the 20 cages (n ¼ 100) and ended with 26 cages. Females began with 5 littermates in each of the 28 cages (n ¼ 140) and ended without any changes. (b) Four male littermates fed a high-fat diet peacefully cohabitating in their cage. The fifth mouse is just outside of the image and slightly away from the group. were approximately 35 g compared with 43 g for high- The most striking reduction in the number of mice fat diet cages (Figure 2(b); comparing lean with high-fat per cage occurred in the transgenic lean diet cages, diets with the same genotype). The changes in weights where the average number started at 5.0 and ended ranged from 5 g to 18 g (Figure 2(c)). This was reflected with 1.6 (p < 0.01). Although we hypothesized that in the fur of high-fat-diet-fed mice as it changed in the transgene, autotaxin, may be related to the aggres- appearance due to the oily red food and the propensity sive behaviors observed, the following statistical of the males to eat and nest in it, creating a greasy coat data did not support that. Among the high-fat diet (Figure 2(d)). The study was terminated when one of transgenic cages, the average number of mice per cage the largest males, who exceeded 60 g, suddenly died for started at 4.75 mice and finished at 3.2 mice. Similarly, unknown reasons, although obesity and/or fatty liver the wild-type lean diet cages started with an average of disease were suspected by the veterinarian upon nec- 3.4 mice and finished with 1.3 mice. This suggested that ropsy (data not shown). diet was the causal factor. Initially the experiment commenced with peaceful cohabitation among the groups, and this continued in Randomization model results the male high-fat diet cages (Figure 3(a)). Unexpectedly, males in the lean diet cages began to Thus, it became necessary to design a strategy to quan- show wounds during morning observations (Figure tify the effect of each group. In the beginning of the 3(b)). Cage separations were necessary to stop fighting, study, six lean and seven high-fat diet cages were cre- with some unfortunate victims of nocturnal bouts of ated, and at the end of the study, 17 lean and nine high- aggression, during the time of day when rodents are fat diet cages were recorded. To obtain the p-value, a active. Although there were more victims of aggression statistic must be found that gives some measure of the in the lean diet-fed cages, over the course of the experi- differences between the treatment groups and the con- mental period, there was not a significant difference in trol. A natural cage-level response variable is the overall survival between groups (Figure 3(c)). number of separations in a cage divided by the total At first we attributed this to male FVB/N behavior number possible. A corresponding statistic typically and hypothesized that either the transgene, colony size used is an F-test statistic. Note that in this case the or diet was the source. Initially, each male cage had an data need not be normal, and the F-statistic need not average of 3–5 littermates in it when the diet study have an F distribution, because our results will not rely commenced, dependent on the litter size and sex at on any asymptotic, large-sample theory. There is a total birth. However, the number per cage dropped to an of 156 mice, with an average of 4.5 per cage. average of 1.4 (among wild-type, lean diet cages) to The maximum number of possible separations for a an unchanged 3.3 (among wild-type, high-fat diet cage is the number of animals in that cage minus cages) by the study’s end (Figure 3(d)). Once fighting one. The observed F-statistic based on the proportion had occurred exclusively in nearly every lean cage, it of separations was 3.72 and the p-value from became apparent that the behavior was abnormal. 100,000 permutations was 0.025 (estimated 164 Laboratory Animals 54(2)

Figure 2. Male FVB/N mice show significant increases in weight with high-fat diets. (a) Males were fed either a lean or high-fat diet for 27 weeks and then weighed. Box plots shows the range of (b) the final average weight or (c) the change in weight, in grams, per group at the end of the study. The association herein is related to diet, not the transgene, which has no affect. (d) Image comparing a mouse fed a lean diet (bottom) with a mouse fed a high-fat diet (top). Note the difference in coat appearances between the mice due to the fat content of the diets. ***p < 0.001. **p < 0.01; one-way ANOVA, multiple comparisons test, followed by Bonferroni’s post-test. standard error 0.0005). The histogram of the permu- estimator for the proportion of aggressive animals tations is shown in Figure 4 and can also be found (Supplementary Figure 2). We chose to use the max- at: https://github.com/dobbinke/Rmarkdown4cagesep. imum likelihood estimator, which for cage i in condi-

Since the p-value is less than 0.05, we reject the null tion k with mki separations, is hypothesis at the 0.05 signiﬁcance level and conclude that the separations diﬀer by treatment assignment for mki=nki : mki nki 2 p^ki ¼ at least some of the treatment groups and control. 1 : mki ¼ nki 1 P 1 rk Then p^k ¼ p^ki, where rk is the number of Population model results rk k¼1 cages assigned to condition k. A one-way analysis of Therefore, for cage i assigned to condition variance results in an F-statistic with four numerator k 2 fg0,1,2,3,4 , we denote the probability of turning and 29 denominator degrees of freedom, a value of 4.18 aggressive as pki, where Ep½¼ki pk. We mathematically and a p-value of 0.009. Since the p-value is below 0.05, derived (Supplementary Figure 2) both the method of we reject the null hypothesis. But, because there are moments estimator and the maximum likelihood only three or four replicates per active treatment, one Murph et al. 165

Figure 3. Males fed a lean diet significantly increase aggressive behaviors. (a) Bar graphs displaying the overall number of cage separations over the entire diet experiment. The study was required to be terminated due to spontaneous death among an obese male mouse. (b) Image of mice from a lean-diet-fed cage showing a victim (left) that suffered a serious wound as a result of the perpetrator (right) from nocturnal activities that were discovered the next morning. The perpetrator is seemingly sniffing the opponent, even after inflicting horrific wounds overnight. The veterinarians requested that both animals be euthanized using a method consistent with the guidelines of the American Veterinary Medical Association. (c) A Kaplan–Meier statistic was used to estimate the overall survival of the mice in the study, based on diet. The solid red line indicates those on the high-fat diet versus the dashed line for the lean diet. Results were not statistically significant. (d) Group cage separations that were required during the diet study. **p < 0.01; preliminary permutation p-value of 0.00432, based on 100,000 Monte Carlo. may have concern that these small replicate numbers strongly against the null hypothesis. The maximum mean that the distribution of this F-statistic may not likelihood estimates p^k are p^0 ¼ 0:06, p^1 ¼ 0:10, p^2 ¼ be close to normal. A more robust alternative is the 0:00, p^3 ¼ 0:60, p^4 ¼ ,0:33. One can see that group 3 non-parametric Kruskal Wallis test. A non-parametric appears to have a higher probability of aggression Kruskal Wallis one-way analysis of variance produces a than the others. chi-square statistic of 9.0172 with 4 degrees of freedom The Dunnett test15 is appropriate for comparing mul- and a p-value of 0.06. In this case, the p-value is not tiple treatments to a single control. For the Dunnett test, significant at the 0.05 significance level. However, the only group 3 is significantly different than the control, Kruskal Wallis test has low power and that may explain with a corrected t-test statistic of 3.759 and a p-value the discrepancy. Moreover, since the permutation test of 0.0031. The Dunnett test output for all compari- used for the randomization model did not rely on sons is shown in Table 1. The estimate of p^3 p^0 is asymptotics, the p-value from that analysis argues 0.54 and a 95% confidence interval for the difference is 166 Laboratory Animals 54(2)

with this strain. Immunoblotting of the tissues from the animals showed no apparent diﬀerences in the expression of the androgen receptor between groups (Figure 5(c)).

Discussion In this study, we observed unexpected, heightened manifestations of aggressive nocturnal bouts among FVB/N male cages fed a lean diet. To evaluate this phenomenon, we developed statistical modeling and computation methods to analyze the data using ran- Figure 4. F-statistic histogram of the data permutations. domization and population models with a censored The F-statistic shows the ratio of the variances. Because likelihood. We conclude that male cage separations the p-value is less than 0.05, this resolves that it is differ significantly due to a lean diet (i.e. treatment appropriate to reject the null hypothesis at the 0.05 sig- assignments) and are unrelated to transgene, autotaxin, nificance level. Further, the statistics conclude that the status and the testosterone level of the mice. separations differ by treatment, for at least some of the The logical conclusion of our study is that energy treated groups and control. The vertical line is observed. among FVB/N males is intensified by feeding males a lean diet because this is a normally hyperactive strain,1,2 which then has amplified activity for nocturnal bouts of aggression. The FVB/N strain is defined in some circumstances by their aggressive nature, Table 1. although this is almost exclusively assigned to males, Multiple comparison of means with Dunnett 8 contrasts. not females. Previous studies with crosses of inbred CBA and outbred CFLP mouse strains determined Standard T-test that lean mice exhibit more nocturnal wheel running Estimate error statistic p-value and higher food intake, compared to fat mice, with Group 1 vs control 0.0400 0.1436 0.278 0.997 lean mice expending substantially more energy towards 18 Group 2 vs control –0.0600 0.1624 –0.370 0.992 activity. Group 3 vs control 0.5400 0.1436 3.759 0.003 The opposite is also true – feeding hyperactive FVB/N males a high-fat diet mitigates active behavior Group 4 vs control 0.2733 0.1624 1.683 0.338 because they become sedentary and are less likely to Dunnett contrasts used for comparing the proportions in each engage in vigorous nocturnal movements. Fat inbred group to the control. CBA and outbred CFLP mice are also less likely to par- Group 0 is the control. ticipate in activities and will steadily decrease participation in nocturnal behaviors, such as wheel running (0.26,0.82). Despite the statistically significant p-value, and vertical rearing.19 Taken together, a lean diet with more experiments would be needed to narrow down four-fold less fat enhances active behaviors and keeps this range. the mice agile so that they are more likely to engage in nocturnal bouts of aggression. Biochemical analysis Therefore, the contributing factor that causes the hyperactive, aggressive FVB/N mice to increase To biochemically explain the reason why lean-diet-fed nocturnal activity is a lean diet with 17% sucrose and mice exhibit more aggression towards littermates, we twice the amount of carbohydrates. This diet heightens initially hypothesized that the concentration of testos- nocturnal behaviors, whereby FVB/N male energy is terone was variable between the groups. However, a expended inappropriately on aggressive behavior. To comparison of lean-diet-fed mice versus high-fat-diet- this point, FVB/N mice have a significantly higher fed mice demonstrated no significant difference in tes- total glucose production compared with C57BL/6J, tosterone (Figure 5(a)). In contrast, when the ratio of 129S1/SvImJ and ICR mouse strains.20 Since FVB/N ADP/ATP was assessed in tissue, the lean-diet-fed mice mice exhibit reduced anxiety levels in comparison to demonstrated a significant decrease in comparison other strains,21 it is unlikely that anxiety contributes to high-fat-diet-fed mice (Figure 5(b)). This suggests to this phenomenon. that energy status and metabolism are the crux of the We also tested testosterone levels and androgen enhanced nocturnal activity,16–18 which is consistent receptor expression between groups, but the results Murph et al. 167

Figure 5. Biochemical analysis of male mouse tissue. (a) Serum was collected at necropsy and tested for the concentration of testosterone (ng/mL). (b) The ratio of ADP/ATP was measured in the male tissues to assess the energy status of the groups. (c) Immunoblotting of androgen receptors did not reveal any striking trends between the groups and quantification of the bands revealed no differences. *p < 0.05 using student’s t-test to measure lean versus high-fat diet groups. were not significant (Figure 5(a) and (c)), suggesting diets. To control the fighting, cages with wounded mice these were not contributing factors for this study. required separation to avoid future occurrences. By the However, this is in contrast with a study showing that end of the study, we had an entire housing rack filled the highly aggressive CD1 strain ceases fighting upon with singly housed males in the facility, which requires castration.22 Nevertheless, castration has major ethical written justification for each circumstance and commu- considerations and is an unnatural scenario; and, yet, nication with the veterinary staff. Studies have shown an adult male living among a group of other adult male male mice prefer conspecific housing with another male mice is an unnatural scenario, too. Most wild mice ter- versus an empty cage, even in the presence of aggres- ritories have one adult male, several females and their sion.27 Therefore, group housing is our standard prac- pups, which is not practical in the laboratory.23 tice and deviations must be explicitly justified. We When planning the study, we sought to mitigate avoid solitary housing unless it is absolutely necessary, FVB/N aggressive tendencies by housing males with a position which coincides with the general consensus in littermates, which facilitates stable social groups and favor of group housing for male mice.23 Recently, Weber hierarchy from an early age, and by maintaining et al. strongly recommended limiting necessary male group housing since periods of single housing can group housing to three animals in a standard cage and exacerbate aggression.24 Other studies show that not providing shelter enrichments under scenarios that stress can be induced in male mice with olfaction expos- are known to increase aggression.28 ure to human male experimenters25; however, all of the In summary, we developed a novel statistical method human experimenters in this study were female. The for analyzing an abnormal set of biological data, stem- only male in the research group was the statistician ming from a study about diet-induced obesity. who did not handle the mice. Although unexpected, we observed a rise in male noc- On the other hand, cages did have structural enrich- turnal bouts of aggression among cages with ad libitum ment, in the form of a mouse house, that was notably lean food. Using the randomization and population used as a male marking post. Sometimes these types of models, we concluded significance was present in structures provide opportunities for dominant perpet- some of the data. Cage separations were not random; rators to ambush a victim or attack them as they enter rather, they were resultant from a lean diet consumed and exit the mouse house.24 Although nesting material by FVB/N adult males that enhanced energy status and was provided and is supposed to decrease aggression,26 resulted in nocturnal bouts of aggression. In conclu- this did not prevent aggression among the FVB/N sion, the hyperactive and subsequent aggressive behav- males fed a lean diet, but we did not test what occurs ior of the FVB/N adult males them an unsuitable model in the absence of nesting material. for long-term studies requiring lean-diet modifications; Prior to this study, we did not anticipate an elevation however, they could be utilized to study male mouse in overnight wounding to arise in male cages fed lean aggression that only requires a modification in diet. 168 Laboratory Animals 54(2)

Acknowledgement hemostasis and thrombosis. J Biol Chem 2009; 284: Our thanks to Charnel Byrnes for proofreading this 7385–7394. manuscript. 12. Fisher RA. Statistical methods for research workers. Edinburgh: Oliver & Boyd, 1925. 13. Neyman J. On the application of probability theory of Declaration of Conflicting Interests agricultural experiments: essay on principles. Section 9. The author(s) declared no potential conflicts of interest with Statistical Science 1990; 5: 465–480. respect to the research, authorship and/or publication of this 14. Prentice RL. Surrogate endpoints in clinical trials: defini- article. tion and operational criteria. Stat Med 1989; 8: 431–440. 15. Dunnett CW. A multiple comparison procedure for com- Funding paring several treatments with a control. J Amer Statistical Assoc 1955; 50: 1096–1121. The author(s) disclosed receipt of the following financial sup- 16. Julien SG, Kim SY, Brunmeir R, et al. Narciclasine port for the research, authorship and/or publication of this attenuates diet-induced obesity by promoting oxidative article: The research was supported by a grant (to MMM) metabolism in skeletal muscle. PLoS Biol 2017; 15: from the National Institutes of Health (grant number e1002597. 1R15CA176653). 17. Terai S, Tsujimura T, Li S, et al. Effect of oxygenated perfluorocarbon on isolated islets during transportation. ORCID iD J Surg Res 2010; 162: 284–289. Mandi M Murph http://orcid.org/0000-0002-5152-4977 18. Comelli M, Pretis I, Buso A, et al. Mitochondrial energy metabolism and signalling in human glioblastoma cell lines with different PTEN gene status. J Bioenerg References Biomembr 2018; 50: 33–52. 1. Mineur YS and Crusio WE. Behavioral and neuroana- 19. Simoncic M, Horvat S, Stevenson PL, et al. Divergent tomical characterization of FVB/N inbred mice. Brain physical activity and novel alternative responses to high Res Bull 2002; 57: 41–47. fat feeding in polygenic fat and lean mice. Behav Genet 2. Voikar V, Koks S, Vasar E, et al. Strain and gender dif- 2008; 38: 292–300. ferences in the behavior of mouse lines commonly used in 20. Burgess SC, Jeffrey FM, Storey C, et al. Effect of murine transgenic studies. Physiol Behav 2001; 72: 271–281. strain on metabolic pathways of glucose production after 3. Taketo M, Schroeder AC, Mobraaten LE, et al. FVB/N: brief or prolonged fasting. Am J Physiol Endocrinol an inbred mouse strain preferable for transgenic analyses. Metab 2005; 289: E53–E61. Proc Natl Acad Sci USA 1991; 88: 2065–2069. 21. Kim S, Lee S, Ryu S, et al. Comparative analysis of the 4. Silver LM. Mouse genetics, concepts and applications. Bar anxiety-related behaviors in four inbred mice. Behav Harbor, ME: Oxford University Press, 1995. Processes 2002; 60: 181–190. 5. Wong AA and Brown RE. Visual detection, pattern dis- 22. Lofgren JL, Erdman SE, Hewes C, et al. Castration elim- crimination and visual acuity in 14 strains of mice. Genes inates conspecific aggression in group-housed CD1 male Brain Behav 2006; 5: 389–403. surveillance mice (Mus musculus). J Am Assoc Lab Anim 6. Errijgers V, Van Dam D, Gantois I, et al. FVB.129P2- Sci 2012; 51: 594–599. Pde6b(þ) Tyr(c-ch)/Ant, a sighted variant of the FVB/N 23. Kappel S, Hawkins P and Mendl MT. To group or not to mouse strain suitable for behavioral analysis. Genes Brain group? Good practice for housing male laboratory mice. Behav 2007; 6: 552–557. Animals (Basel) 2017; 7(12). pii: E88. doi: 10.3390/ 7. Girard SD, Escoffier G, Khrestchatisky M, et al. The ani7120088. FVB/N mice: a well suited strain to study learning and 24. Charles River Laboratories International Inc. Reducing memory processes using olfactory cues. Behav Brain Res aggression in mice: technical sheet. 2012, p.3. 2016; 296: 254–259. 25. Sorge RE, Martin LJ, Isbester KA, et al. Olfactory expos- 8. Pugh PL, Ahmed SF, Smith MI, et al. A behavioural ure to males, including men, causes stress and related characterisation of the FVB/N mouse strain. Behav analgesia in rodents. Nat Methods 2014; 11: 629–632. Brain Res 2004; 155: 283–289. 26. Armstrong KR, Clark TR and Peterson MR. Use of 9. Federico L, Ren H, Mueller PA, et al. Autotaxin and its corn-husk nesting material to reduce aggression in product lysophosphatidic acid suppress brown adipose caged mice. Contemp Top Lab Anim Sci 1998; 37: differentiation and promote diet-induced obesity in 64–66. mice. Mol Endocrinol 2012; 26: 786–797. 27. Van Loo PLP, de Groot AC, Van Zutphen BFM, et al. 10. Kuppa SS, Jia W, Liu S, et al. Autotaxin exacerbates Do male mice prefer or avoid each other’s company? tumor progression by enhancing MEK1 and overriding Influence of hierarchy, kinship, and familiarity. Journal the function of miR-489-3p. Cancer Lett 2018; 432: of Applied Animal Welfare Science 2001; 4: 91–103. 84–92. 28. Weber EM, Dallaire JA, Gaskill BN, et al. Aggression in 11. Pamuklar Z, Federico L, Liu S, et al. Autotaxin/lysopho- group-housed laboratory mice: why can’t we solve the lipase D and lysophosphatidic acid regulate murine problem? Lab Anim (NY) 2017; 46: 157–161. Murph et al. 169 Re´sume´ La souche de souris consanguine hyperactive FVB/N est largement utiliseé pour les applications de recherche transgeńique, bien que rarement pour des e´tudes comportementales. Ces souris ont une de´ficience visuelle due a` leur de´geńe´rescence re´tinienne, mais sont conside´reés comme tre`s intelligentes et se basent en grande partie sur l’olfaction. Lors d’une enqueˆte sur l’obe´site´ induite par le re´gime alimentaire meneé chez des souris FVB autotaxines transgeńiques, nous avons observe´ une nećessite´ accrue de se´parer les maˆles, mais pas les femelles, en utilisant des cages se´pareés. Sur la base de nos observations, nous avons e´mis l’hypothe`se que l’alimentation maigre des souris FVB/N augmentait les e´pisodes d’agression nocturne entre les maˆles de la meˆme porteé. L’alimentation des animaux adultes d’une porteée´tait changeédela nourriture normale a` des aliments a` haute teneur en matie`res grasses (45% de matie`re grasse) ad libitum ou maigres (10% de matie`res grasses) pendant 27 semaines, les souris parvenant ainsi a` un poids moyen de 43 g contre 35 g, respectivement. Du fait des combats nocturnes qui ećlataient, la se´paration des animaux en les mettant dans des cages diffe´rentes s’est imposeémeˆme si les souris de la porteé avaient cohabite´ de manie`re pacifique pendant les 10-16 semaines preće´dentes. Les donneés e´tant de nature inhabituelle, elles nećessitaient d’e´laborer des me´thodes statistiques peu communes afin d’e´valuer si elles reveˆtaient de l’importance et si cette importance e´tait significative. Par conse´quent, nous avons e´tabli une me´thodologie en nous appuyant sur la randomisation et les mode`les de population, et e´mit l’hypothe`se que la testoste´rone, l’autotaxine transgeńique ou alte´ration du re´gime alimentaire constituait le facteur causal. Les e´valuations statistiques ont de´montre´ une corre´lation significative entre le comportement agressif et la se´paration dans des cages diffe´rentes associe´s aux souris suivant une alimentation faible en lipides, pas avec l’autotaxine. Les donneés biochimiques ne semblent pas expliquer le comportement. En revanche, le me´tabolisme de l’eńergie a mis en lumie`re des diffe´rences entre les groupes de souris normalement hyperactives selon l’alimentation. Cette caracte´ristique fait des souris FVB/N maˆles des sujets inapproprie´s pour les e´tudes a` long terme impliquant des modifications alimentaires baseés sur un faible taux de lipides.

Abstract Der hyperaktive FVB/N Inzucht-Mausstamm wird haüfig in der transgenen Forschung verwendet, selten jedoch fu¨r Verhaltensstudien. Diese Maüse sind sehbehindert aufgrund von Netzhautdegeneration, gelten aber als hochintelligent und sind weitgehend auf ihren Geruchssinn angewiesen. Bei der Untersuchung von erna¨hrungsbedingter Adipositas bei transgenen Autotaxin-FVB/N-Maüsen beobachteten wir eine versta¨rkte Notwendigkeit der Ka¨figtrennung von ma¨nnlichen, nicht jedoch von weiblichen Tieren. Aufgrund der Beobachtungen vermuteten wir, dass die Fu¨tterung der FVB/N-Maüse mit magerer Kost na¨chtliche Aggressionsschu¨be zwischen ma¨nnlichen Wurfgeschwistern versta¨rkt. Die Erna¨hrung erwachsener Wurfgeschwister wurde ad libitum 27 Wochen lang von normalem Futter auf fettreiche (45% Fett) oder magere (10% Fett) Erna¨hrung umgestellt, wobei die Maüse durchschnittlich 43 g bzw. 35 g erzielten. Daraufhin wurden Ka¨figtrennungen aufgrund na¨chtlicher Aggressionsschu¨be unvermeidlich, obgleich die Wurfgeschwister 10-16 Wochen zuvor friedlich zusammengelebt hatten. Da die Daten ungewo¨hnlich waren, mussten ungewo¨hnliche statistische Methoden entwickelt werden um festzustellen, ob und inwiefern Signifikanz bestand. Deshalb entwickelten wir unter Einsatz der Randomisierungs- und Populationsmodelle eine Methodik und postulierten, dass entweder Testosteron, das Autotaxin-Transgen oder die Erna¨hrungsumstellung der Kausalfaktor war. Statistische Auswertungen zeigten eine signifikante Korrelation zwischen Ka¨figtrennungen und aggressivem Verhalten bei den mit Magerkost gefu¨tterten Maüsen, nicht Autotaxin. Biochemische Daten schienen das Verhalten nicht zu erkla¨ren. Im Gegensatz dazu zeigte der Energiestoffwechsel Unterschiede zwischen den Gruppen von normalerweise hyperaktiven Maüsen je nach Erna¨hrung auf. Diese Eigenschaft macht ma¨nnliche FVB/N-Maüse ungeeignet fu¨r Langzeitstudien mit Erna¨hrungsumstellung auf Magerkost.

Resumen La cepa consanguıńea hiperactiva del ratoń FVB/N se utiliza ampliamente para aplicaciones de investigacioń transgeńica, aunque pocas veces para estudios del comportamiento. Estos ratones tienen dificultades de visioń a trave´s de la degeneracioń de la retina, pero se consideran muy inteligentes y dependen mucho del olfato. Al investigar la obesidad por dieta en ratones FVB/N transgeńicos autotaxin, observamos un aumento 170 Laboratory Animals 54(2) de la necesidad de separar ratones macho, no ası´ con los ratones hembra. Basańdonos en la observacioń, hemos teorizado sobre el hecho de que el dar una dieta saludable a los ratones FVB/N aumenta las peleas nocturnas entre los machos de la camada. Las dietas de los machos adultos de la camada pasaron de comida normal a dietas altas en grasas a voluntad (45% de grasa) o dietas saludables equiparables (10% de grasas) durante 27 semanas, durante las cuales los ratones alcanzaron una media de 43 g frente a 35 g, respecti- vamente. Entonces, la separacioń de las jaulas debido a las peleas nocturnas tuvo que ser obligatorio, a pesar de que los miembros de la camada habitaron pacı´ficamente durante las 10-16 semanas anteriores. Ya que los datos eran poco usuales, se requirieron me´todos estadı´sticos distintos para poder evaluar si habıá datos significativos y dońde los habıá, en caso de existir. Por tanto, utilizar los modelos de poblacioń y aleatoriedad, establecimos una metodologıá y postulamos que el factor causal era la testosterona, el transgene autotaxin o la alteracioń de la dieta. La evaluacioń estadı´stica demostro´ una correlacioń significativa entre las separ- aciones de jaulas y el comportamiento agresivo asociado a una dieta saludable de los ratones, no a la autotaxin. Los datos bioquı´micos no parecıán explicar el comportamiento. Por contraste, el metabolismo energe´tico destaco´ las diferencias entre los grupos de ratones con una hiperactividad normal seguńla dieta. Esta caracterı´stica hace que el ratoń macho FVB/N sea un modelo inadecuado para estudios a largo plazo con modificaciones de dietas saludables. Original Article Laboratory Animals 2020, Vol. 54(2) 171–182 ! The Author(s) 2019 Combination of ketamine and xylazine Article reuse guidelines: sagepub.com/journals-permissions with opioids and acepromazine in rats: DOI: 10.1177/0023677219850211 journals.sagepub.com/home/lan Physiological changes and their analgesic effect analysed by ultrasonic vocalization

Jilma Alema´n-Laporte1,2 , Luciana A Bandini1, Mariana SA Garcia-Gomes1, Dennis A Zanatto1, Denise T Fantoni3, Marco A Amador Pereira3, Pedro E Navas-Sua´rez1, Thiago Berti Kirsten4, Randall R. Jimenez5, Gilbert Alvarado1,6 and Claudia Cabrera Mori1

Abstract In this study, the effect of four anaesthetic protocols that included the combination of xylazine (X) and ketamine (K) with acepromazine (A) and opioids (methadone (Me), morphine (Mo) or tramadol (T)) was evaluated in laboratory rats of both sexes. Ultrasonic vocalization (USV) was used as an indicator of pain during the recovery period. The objective was to evaluate the physiological parameters and the analgesic effect of each protocol to determine which protocol was the safest and fulfil the requirements of a balanced anaesthesia. The better protocols were the XKA protocol for both sexes and the XKMe protocol for females because the combinations achieve surgical plane of anaesthesia in rats. However, pain assessment during the formalin test revealed that rats anaesthetized with XKA produced more numbers of USV, suggesting that it is not a good protocol for the control of immediate postoperative pain. All protocols produced depression in body temperature and respiratory and heart rates, and had important effects, such as micturition and maintenance of open eyes. Only rats anaesthetized with XKA protocol did not present piloerection. These results demonstrated that good monitoring and care during anaesthesia must be included to prevent complications that compromise the life of the animal and to ensure a good recovery. The inclusion of analgesia in anaesthesia protocols must be used routinely, ensuring minimal presence of pain and thus more reliable results in the experimental procedures.

Keywords Anaesthetic protocol, ultrasonic vocalization, analgesia, physiological measures, refinement

Date received: 27 June 2018; accepted: 18 April 2019

4Environmental and Experimental Pathology, Paulista University, Brazil Rats are commonly used in research, and the use of 5Institute of Evolutionary Ecology and Conservation Genomics, anaesthesia is required and recommended for many University of Ulm, Germany experimental procedures. Well-controlled anaesthesia 6Laboratory of Experimental and Comparative Pathology is essential to reduce potential surgical complications (LAPECOM), School of Biology, University of Costa Rica, Costa Rica Corresponding author: Jilma Alemań Laporte, Department of Pathology, School of 1Department of Pathology, University of Saõ Paulo, Brazil Veterinary Medicine and Animal Science, University of Saõ Paulo 2Laboratorio de Docencia en CirugıáyCańcer (DCLab), University (USP), Av. Prof. Dr Orlando Marques de Paiva, 87, Cidade of Costa Rica, Costa Rica Universita´ria, CEP 05508-270, Saõ Paulo Brazil. 3Department of Surgery, University of Saõ Paulo, Brazil Email: [email protected] 172 Laboratory Animals 54(2) and ensures safety and success in the experimental pro- temperature and clinical signs) and their intraoperatory cedures.1 Anaesthesia in laboratory rodents is particu- analgesic power to determine the most effective and safe larly challenging due to the problems related to their protocols for surgical procedures and to ensure the size, for example, accelerated metabolism, easy devel- well-being of laboratory rats. opment of hypothermia and difficulty in determining cardiorespiratory functions.2 Therefore, the goal is to establish a balanced anaesthetic protocol that includes Materials and methods a combination of several drugs to obtain good control Animals of unconsciousness, analgesia and myorelaxation to minimize anaesthetic risks.3 In total, 64 SPF Wistar-Han rats (Rattus norvegicus), The anaesthetic protocols that mix two or three 28 females and 36 males aged from eight to 12 weeks components have better effects compared to a single were used. Animals were obtained from the animal anaesthetic because one product cannot produce all facility of the Institute of Biomedical Science of the the effects that are sought in anaesthesia (hypnosis, University of Sa˜ o Paulo, and they were free of ecto- muscle relaxation and analgesia).4 Ketamine is one of and endoparasites, Mycoplasma pulmonis, Pasteurella the most commonly used parenteral drugs in laboratory pneumotropica, Bordetella bronchiseptica, Helicobacter rats and is normally combined with other drugs such as spp., Klebsiella oxytoca, Klebsiella pneumoniae, xylazine.5,6 The use of this combination with acepro- Pasteurella multocida, Pseudomonas aeruginosa, mazine has been described in the literature,3,5 but it is Staphylococcus aureus, Streptococcus -hemolytic spp., not commonly used in many laboratories. In addition, Streptococcus pneumoniae and Salmonella spp., Kilham there is not much information on anaesthetic protocols Rat Virus, Pneumonia Virus of Rat and Reovirus. At used in laboratory rodents that include the combin- most, 4 animals were housed per cage on corn-cob bed- ation of these anaesthetics with opioids, such as trama- ding (Granja R.G., SP, Brazil) in polypropylene cages dol, morphine and methadone. The preoperative (41 34 16 cm) that were changed once a week. Paper administration of opioids simultaneously with the rolls were added as environmental enrichment. All the anaesthetic protocols has advantages such as the possi- animals were maintained under climate-controlled con- bility of reducing the required dose of anaesthetics to ditions of 12:12-h light/dark cycle, temperature achieve surgical anaesthesia as well as reducing the need range from 22 to 24C, relative humidity of 45–65%, for postoperative analgesics.7 In view of these benefits ad libitum access to drinking water, and a standard the present study is important to clarify the interaction food-pellet diet (NuvitalÕ-Quimtia, PR, Brazil). of the combination of XK with different opioids in an Before the beginning of the experiments, the animals injectable anaesthetic protocol. were housed in the animal facility of the Department Vocalization serves as an indicator of emotional and of Pathology of the School of Veterinary Medicine and aversive states in rats.8 Juvenile and adult rats emit Animal Science, University of Sa˜ o Paulo, where they basically two types of ultrasonic vocalizations (USV) stayed for at least one week for adaptation and hand- that are distinguished on the basis of the frequency: ling before the start of the experiment. the 22 kHz vocalizations produced in response to aver- All animal procedures were performed in accord- sive and dangerous situations and the 50 kHz vocaliza- ance with the guidelines of the Ethics Committee of tions that are produced in response to appetitive the School of Veterinary Medicine and Animal situations, such as copulation and play behavior.9 Science of the University of Sa˜ o Paulo, Brazil Jourdan et al. also demonstrated that monitoring (no.9635260116). rodent USV is a potential method of measuring the negative affective component of pain by a brief elec- Anaesthetic procedure trical pulse applied to the tail.10 The effects of opioids have been extensively studied Four different protocols were established for this using the formalin test, and it has been shown that study16–18 (see Supplementary Material 1). morphine, methadone and tramadol influence pain con- For males: trol in rats.11,12,13 However, there are no studies that have been performed yet with the opioids that were 1. XKA: Xylazine 7.5 mg/kg (AnasedanÕ-Ceva, 20 mg/ml, used in this experiment in the context of anaesthesia SP, Brazil) þ Ketamine 60 mg/kg (DopalenÕ-Ceva, in rats. For this reason, it is important to evaluate 100 mg/ml, SP, Brazil), þAcepromazine 2 mg/kg intraoperative pain and recovery from anaesthe- (AcepranÕ-Vetnil, 2 mg/kg, SP, Brazil); sia.10,14,15 The objective of this study was to evaluate 2. XKMe: Xylazine 5 mg/kg þ Ketamine 60 mg/kg þ the effect of different anaesthetic protocols on the Methadone 5 mg/kg (MytedomÕ-Crista´lia, 10 mg/ml, physiological parameters (heart and respiratory rate, SP, Brazil); Alemań-Laporte et al. 173

3. XKMo: Xylazine 7 mg/kg þ Ketamine 60 mg/kg, 100% oxygen using a nose cone until the end of the þ Morphine 1 mg/kg (DimorfÕ-Crista´lia, 10 mg/ml, anaesthesia (Supplementary Material 2). SP, Brazil); and After the induction time, a pulse-oximeter 4. XKT: Xylazine 7.5 mg/kg þ Ketamine 65 mg/kg, (NT1A-VÕ-Solaris Medical Technology, Inc., CA, Tramadol 5 mg/kg (TramadonÕ-Crista´lia, 50 mg/ml, USA) was positioned on the pad of the left hind limb, SP, Brazil). and then the physiologic parameters and reflexes were measured every 10 minutes until the righting reflex For females: return. The heart rate was determined by the pulse- oximeter and the respiratory rate was counted by obser- 1. XKA: Xylazine 5 mg/kg þ Ketamine 60 mg/kg þ ving thoracic or abdominal movements. The body Acepromazine 1 mg/kg; temperature was measured by a digital thermometer 2. XKMe: Xylazine 5 mg/kg þ Ketamine 60 mg/kg þ placed into the animal’s rectum. The reflexes were eval- Methadone 5 mg/kg; uated as follows: the pedal withdrawal reflex by press- 3. XKMo: Xylazine 5 mg/kg þ Ketamine 60 mg/kg þ ing the paw with an atraumatic forceps, alternating Morphine 1.5 mg/kg; and both hind limbs; the palpebral reflex by a slight touch 4. XKT: Xylazine 5 mg/kg þ Ketamine 60 mg/kg, with the atraumatic forceps in the margin of the eyelid; Tramadol 5 mg/kg. and the tail pinch reflex by a compression of the tail with the atraumatic forceps;19 any negative or positive Rats were previously weighed to determine the exact response was classified as the loss or the return of the dose of each drug. To avoid any interference of circa- reflex, respectively. dian rhythm, the experiment was always performed from 6.00 a.m. to 12.00 p.m. The administration of the Pain evaluation by ultrasonic vocalization anaesthetic protocol was done by intraperitoneal route. The drugs were mixed in a sterile plastic tube immedi- When the rats recovered the pedal reflex or their whis- ately prior to the administration and were given in one kers started to move, a subcutaneous injection of 60 ml injection to minimize handling stress with a maximum of formalin solution (formaldehyde 37%, wt/wt, injection volume of 0.5–0.7 ml. Then, the rat was placed diluted to 10% in 0.9% saline wt/wt)8 was performed in a cage alone to observe its behaviour, and when the in the dorsal surface of the right hind paw of each rat.14 animal lost the righting reflex, it was laid in dorsal After 5 minutes of the formalin administration, the recumbency on a preheated thermal blanket (between USV were recorded using a USV detector monitor 35 and 37C) to minimize the loss of body temperature. (Ultravox 2-0, Noldus Professional System for Sterile ocular lubricant (VidisicÕ-Gel-BauchþLomb, Automatic Monitoring of Ultrasonic Vocalization, 2 mg/kg, SP, Brazil) was administered to both eyes. USA). Each rat was tested individually by placing During the anaesthesia, the following periods were the USV detector 2 cm above the head of the ani- recorded: mal in dorsal recumbency without any kind of restric- tion. The USV detector was set to detect frequencies 1. Induction time: the period from the administration of 22 kHz with an amplitude filter setting of 4 to of the anaesthesia until the loss of the righting reflex. minimize background noise. To avoid environmental 2. Non-surgical anaesthesia: the period between the noise, the experiment was performed in a soundproof loss of the righting reflex and the loss of most or room. all reflexes (blink, pedal and tail withdrawal reflex). The total number and duration of USV for each 3. Surgical anaesthesia: the period between the loss of rat were recorded during four sessions, starting 5 min- all the reflexes (considering surgical tolerance the utes after the application of formalin, for a loss of the pedal withdrawal reflex) and their total period of 25 minutes or until the righting reflex recuperation. recovery. Each session consisted of 3 minutes of 4. Recuperation time: the period between the recuper- recordings with 2-minute lapses between them. ation of all the reflexes and the return of the ability During the recording, the rats received a stimulus to walk. every 20 seconds by compressing the area where the formalin was injected using a bulldog clamp. The total duration time of the anaesthesia, the The intensity of the clamp tightening was always the period from the administration of the anaesthetic same, letting it completely close for a second and protocol until the withdrawal of the ability to walk, immediately open. was also measured. At the end of the study all animals were The rats initially breathed room air but 7 minutes euthanized by an overdose of X (30 mg/kg) and K after losing the righting reflex, they were supplied with (300 mg/kg) IP. 174 Laboratory Animals 54(2) Statistical analysis statistics to describe the patterns of vocalization dur- A power calculation was carried out with a total power ation due to small size to fit a statistical model. The of 0.8 and a two-sided significance of <0.05. data are shown as the mean SE, unless otherwise Estimates of variability for power calculation were stated. We decided not to report statistical significance based on data from a pilot project. The study was (p values) in order to focus attention on the bio- appropriately powered to detect a mean difference of logical relevance of effect sizes.22 Statistical analyses 20% in the total duration of surgical anaesthesia, since were conducted in Rv.3.3.2 (R Development Core this was considered clinically relevant. In males, we Team 2017). used up to 10 rats because the administration of formalin caused an immediate recovery from anaesthesia in some animals, and this did not allow the USV record- Results ing of them. The rats of each treatment were distributed Physiological parameters randomly in different cages by sex. Linear mixed effects models (LMMs) were used to The best fit and well-supported model for explaining test the effect of drug treatment and time (predictor variation in the body temperature and heart and variables) on the response variables: body temperature, respiratory rates in male and female rats was the time cardiac frequency and respiratory frequency in male only model (Supplementary material 3). The significant and female rats. The rat identity was used as a differences between the parameters over time using random factor to account for repeated measures in mean differences and 95% CIs can be observed in each rat. Linear models (LMs) were used to investigate Supplementary material 3–5. the effect of drug treatment on the response variables: Considering all the anaesthetic protocols, the mean induction time, non-surgical anaesthesia, recovery and body temperature of the males ranged from 35.66 total duration of anaesthesia. Generalized linear mixed 0.19C (at 10 minutes) using XKA to 37.62 0.24C models with Poisson distribution were conducted to (at 60 minutes) using XKMe (Figure 1(a)). In females, investigate whether the number of vocalizations the mean body temperature ranged from 36.30 (response variable) differed according to the predictor 0.07C (at 10 minutes) using XKMe to 37.78 variables: drug treatment and time. The random factor 0.54C (at 50 minutes) using XKMo (Figure 1(b)). In was rat identity. A candidate set of models were con- males, all protocols presented a decrease in the mean structed for each analysis previously mentioned and heart rate, which reached the lowest rate at minute 20 used an information-theoretic model selection (I-T) (approximately 258 bpm); after this time, there was an to determine which model(s) was best fit and well sup- increase in the mean heart rate in the four anaesthetic ported by the data.21 The I-T model selection provides protocols (Figure 2(a)). In the case of the females, in a strength of evidence for a set of a priori models which the protocol of XKMe was administered, after based on Akaike’s Information Criterion values 10 minutes, the mean heart rate remained approxi- (AIC) and do not rely on null hypothesis testing.21 mately 250 bpm until minute 40 (Figure 2(b)). All The best fit model of each candidate set was deter- protocols for both sexes presented means lower than mined based on AIC adjusted for small sample size 70 rpm at minute 10 in the respiratory rate. However, (AICc); and delta AICc (ÁAICc) and Akaike weights females had a marked increase in respiratory rate after (AICcw) were used to evaluate the support of the 20 minutes, except in the XKMo and XKMe protocols models (for a complete treatment of I-T, see (Figure 3). In both males and females, there was a Burnham and Anderson).20 Here, we considered a decrease in all the physiological parameters evaluated model having the lowest AICc, ÁAICc <2 and in the first 10 minutes. AICcw approaching 0.90 or higher the best fit and well-supported model for the containing predictor(s). Clinical signs, reflexes and We took the best fit model in each case and reported anaesthetic times the effects of predictor(s). When a main effect was detected we conducted post-hoc pairwise comparisons. In both sexes, more than 86% of the rats presented Specifically, we used mean difference and Cohen’s d micturition in all the protocols, and 100% of the rats effect size for conducting post-hoc tests on LMMs and presented opening of the eyes during the anaesthesia. LMs, respectively, and 95% confidence intervals of Only the rats that were anaesthetized with XKA did not estimates (95% CIs) to measure precision and ‘statis- present piloerection (Table 1). tical significance’. If 95% CIs overlapped, 0 indicates The protocol in which acepromazine was used had ‘statistical non-significance’ in a pairwise comparison. better surgical tolerance in both sexes, and females also The rules of thumb defined in Sawilowsky21 were used had good tolerance with the protocol XKMe. The rats for interpreting Cohen’s d values. We used descriptive from both sexes from protocol XKT did not achieve Alemań-Laporte et al. 175

Figure 1. Variation in body temperature (C) in male (a) and female (b) rats submitted to different anesthetic associations. (X: Xylazine, K: Ketamine; A: Acepromazine; Me: Methadone; Mo: Morphine; and T: Tramadol.) The points represent fits to the model’s predicted values, and the lines indicate the SE of model fits.

Figure 2. Variation in heart rate (beat per minute (bpm)) in male (a) and female (b) rats submitted to different anaesthetic associations. (X: Xylazine; K: Ketamine, A: Acepromazine, Me: Methadone, Mo: Morphine, and T: Tramadol.) The points represent fits to the model’s predicted values, and the lines indicate the SE of model fits. surgical tolerance, as well as males from protocol duration of anaesthesia, the model with drug treatment XKMo (Table 1). effect was not well supported but provided the best fit The drug treatment model was the best fit and well (Supplementary material 7). The protocols XKA and supported model for explaining variation in the dur- XKMo showed a higher duration compared to XKMe ation of nonsurgical anaesthesia and recovery of male and XKT, and their differences were large according to rats (Supplementary material 7). Male rats anaesthe- Cohen’s d (Supplementary material 8). tized with XKA showed a lower duration of non- In females, the duration of non-surgical anaesthesia surgical anaesthesia compared to the other protocols differed between protocols (Supplementary material 7). (Table 1), and differences were large between them The protocols XKA and XKMe showed a lower dur- according to Cohen’s d effect sizes (Supplementary ation compared to the protocols XKMo and XKT material 8). In relation to duration of recovery, there (Table 1). Drug treatment models were not well were large and medium differences between protocols, supported but provided the best fit to examine the dur- except between XKT and XKMe, where the difference ation of recovery and total duration (Supplementary was negligible (Supplementary material 8). Regarding material 6). Recovery was the lowest in rats 176 Laboratory Animals 54(2)

Figure 3. Variation in the respiratory rate (breaths per minute) in male (a) and female (b) rats submitted to different anaesthetic associations. (X: Xylazine; K: Ketamine; A: Acepromazine; Me: Methadone; Mo: Morphine; and T: Tramadol.) The points represent fits to the model’s predicted values, and the lines indicate the SE of model fits. anaesthetized with XKMe (Table 1) and showed some drugs. Because the combinations of XK with opioids significant differences with other protocols according to used in this study were not described in the literature, Cohen’s d effect sizes (Supplementary material 8). Total the doses were based on a pilot study conducted prior to duration was the lowest in female rats anaesthetized. this experiment (unpublished data). There is evidence with XKT (Table 1) with some significant medium that rats female and male can have differences in the and very large differences with other protocols response to the same anaesthetics,23 so the sex of the ani- (Supplementary material 8). mal is a variable that should be considered for the establishment of the doses to be used. Although the doses Ultrasonic vocalization were not identical for both sexes, the similar behaviour in the dynamics of the physiological parameters of each For the number of male vocalizations, the best fit and protocol allows us to conclude that the doses used in this well-supported model of the four possible models study were adequate. For this same reason, the primary examined through AICc criteria was the drug treatment objective was not to compare the effect between sexes. by time interaction model; the same result was observed in female rats (Supplementary material 9). These inter- Physiological parameters actions indicate that the effect of the protocol on the number of times rats conducted a vocalization was not The reduction in all physiological parameters in the first constant through time. The painful stimulus increased minutes observed in all anaesthetic protocols has been the number of vocalizations with a longer duration pat- previously described with the XK protocol.6,24,25 The tern in the XKA protocol for both sexes (Figure 4 reduction in body temperature observed in rats may be and 5). In the case of male rats, after the application related to the depressant effect of XK on thermoregula- of formalin some animals recovered the righting reflex tory mechanisms, in addition to the fact that small before the USV recordings started, so to evaluate the rodents easily lose heat when anaesthetized, given the formalin test, the number of animals was increased in high proportion of body surface in relation to their order to reach at least seven animals per group. Only in weight.25,26 the 20–23 interval, in males, was an increase in the The heart rate depression observed in all protocols in number of vocalizations observed with the XKMo male rats and in the XKMe protocol in females can be protocol, but with a lower duration pattern in compari- produced mainly by the action of xylazine and metha- son to the XKA protocol (Figure 4(a)). done. Ketamine usually causes a stimulation of cardiovascular function causing an increase in heart rate and Discussion blood pressure. Thus, the reduction in heart rate may be a side effect of a2-agonists as a result of an increase The doses recommended in the literature for ketamine in vagal reflex activity and decrease in norepinephrine and xylazine are very varied and depend on whether release in the sympathetic nervous system.26 they are applied alone or in combination with other Methadone can cause calcium channel blockage Alema ´ -aot tal. et n-Laporte

Table 1. Comparison of safety margin, clinical signs, reflexes and anaesthesia times (minutes) of male and female rats submitted to different anaesthetic associations.

Male rats Female rats

XKA XKMe XKMo XKT XKA XKMe XKMo XKT Anaesthetic drugs (n ¼ 7) (n ¼ 10) (n ¼ 9) (n ¼ 10) (n ¼ 7) (n ¼ 7) (n ¼ 7) (n ¼ 7)

Safety margin Surgical tolerance (%) 86 (6/7) 50 (5/10) 0 (0/9) 0 (0/10) 86 (6/7) 86 (6/7) 29 (2/7) 0 (0/7) Death rate (%) 14 (1/7) 0 (0/10) 0 (0/9) 0 (0/10) 0 (0/7) 0 (0/7) 0 (0/7) 0 (0/7) Clinical signs Micturition (%) 100 (7/7) 90 (9/10) 89 (8/9) 100 (10/10) 86 (6/7) 100 (7/7) 86 (6/7) 86 (6/7) Opening eyes (%) 100 (7/7) 100 (10/10) 100 (9/9) 100 (10/10) 100 (7/7) 100 (7/7) 100 (7/7) 100 (7/7) Piloerection (%) 0 (0/7) 80 (8/10) 78 (7/9) 60.00 (6/10) 0 (0/7) 100 (7/7) 86 (6/7) 43 (3/7) Loss of reflexes Pedal withdrawal reflex (%) 86 (6/7) 50 (5/10) 0 (0/9) 0 (0/10) 86 (6/7) 86 (6/7) 29 (2/7) 0 (0/7) Tail pinch reflex (%) 100 (7/7) 100 (10/10) 78 (7/9) 0 (0/10) 86 (6/7) 100 (7/7) 100 (7/7) 100 (7/7) Blink reflex (%) 100 (7/7) 100 (10/10) 78 (7/9) 0 (0/10) 100 (7/7) 100 (7/7) 100 (7/7) 100 (7/7) Righting reflex (%) 100 (7/7) 100 (10/10) 78 (7/9) 0 (0/10) 100 (7/7) 100 (7/7) 100 (7/7) 100 (7/7) Anaesthesia times Induction 3.00 0.50 2.90 0.30 3.30 0.60 3.80 0.50 2.30 0.30 3.90 0.90 3.10 0.30 2.90 0.30 Non-surgical anaesthesia 10.90 2.20 26.20 6.60 44.90 6.70 46.70 2.50 9.40 3.20 7.90 0.90 37.70 7.60 29.40 5.50 Surgical anaesthesia 27.70 5.70 34.80 6.60 – – 31.30 5.20 42.30 8.70 45.00 1.00 – Recovery 31.00 6.70 9.30 2.90 22.20 5.90 10.90 4.00 38.10 13.70 7.10 3.60 29.60 11.30 24.10 4.30 Total duration 90.30 15.60 62.90 10.40 87.60 10.60 62.10 5.50 85.90 16.50 87.30 8.52 99.90 16.50 64.00 5.10

X: Xylazine; K: Ketamine; A: Acepromazine; Me: Methadone; Mo: Morphine; T: Tramadol. 177 178 Laboratory Animals 54(2)

Figure 4. Number of ultrasonic vocalizations during the formalin test in male (a) and female (b) rats at different time intervals (A: 5–8 minutes; B: 10–13 minutes; C: 15–18 minutes; and D: 20–23 minutes), submitted to different anaesthetic associations. (X: Xylazine; K: Ketamine; A: Acepromazine; Me: Methadone; Mo: Morphine; and T: Tramadol.) The bars represent fits to the model’s predicted values, and the lines indicate the 95% confidence interval of model fits.

Figure 5. Duration of ultrasonic vocalizations (seconds) during the formalin test in male (a) and female (b) rats at different time intervals (A: 5–8 minutes; B: 10–13 minutes; C: 15–18 minutes; and D: 20–23 minutes), submitted to different anaesthetic associations. (X: Xylazine; K: Ketamine; A: Acepromazine; Me: Methadone; Mo: Morphine; and T: Tramadol.) The bars represent mean values, and the lines indicate SE. producing bradycardia,27 an effect that was observed in All protocols in this study follow the same pattern of both sexes with the XKMe protocol. physiological parameter decrease at the first 10 minutes. In addition to cardiac depression, respiratory The clinical support of the animal is critical mainly in depression was observed in all protocols. Opioids can this period of time. A better recovery and a reduction of produce this effect in high doses or in combination with mortality can be enhanced by monitoring, heating and other CNS depressants.17 Opioid receptors are abun- oxygenating the animals. dant in the respiratory control center. Moreover, m-opioid receptor agonists, such as morphine and Clinic signs, reflexes and anaesthetic times methadone, bind to these receptors, activating them and causing respiratory rate depression.28 Xylazine A marked micturition (86% or more) was observed in can also cause CNS depression by stimulation of all protocols, which was also recorded in rats treated a2-adrenoreceptors.29 with xylazine.25,26,30 This drug produces inhibition of Alemań-Laporte et al. 179 the antidiuretic hormone release, which is the reason for surgical procedures. As in males with the XKT why the use of combinations with xylazine are not rec- protocol, no female reached anaesthesia, which indi- ommended in animals with urinary tract obstruction, cates that in this dose interval, it is not a good protocol dehydration or hypovolemia.31 Only the rats that were for use in surgery either. anaesthetized with XKA did not present piloerection. The only animal that died in the experiment (with Xylazine can cause piloerection as a side effect, but the XKA protocol), presented hyperthermia and when combined with acepromazine, which has relaxant increased heart rate. This could be associated to an properties, this effect can be counteracted.32 individual reaction to anaesthesia. All animals anaesthetized in this study kept their eyes open during anaesthesia; besides the loss of blink Ultrasonic vocalization reflex, the combination of XK induces a mild proptosis of the globe and the retraction of the eyelid.33 Although there was great individual variability among According to Turner and Albassam, it is possible that the treated individuals, in both sexes, the XKA proto- the blink reflex is depressed to a greater degree in rats col presented the highest number of ultrasonic vocal- anaesthetized with the combination of XK versus other izations with a frequency of 22 kHz with the longest injectable or inhaled anaesthetic combinations.34 duration, but these vocalizations did not exceed 0.3 Corneal lesions predisposed by the use of this anaes- seconds classified as short call. Two biologically signifi- thetic combination may occur due to trans-corneal cant subtypes of 22 kHz vocalizations have been iden- water loss due to aqueous humour alteration when con- tified, both expressing negative emotional sates: long tinuous corneal exposure occurs,35 or by the vasocon- calls (more than 300 ms) that serve as alarm calls and striction of the ciliary and iridial vessels, and the local signal external danger or potential danger; short calls and systemic hypoxemia caused by xylazine, which may that express a state of discomfort or distress without culminate in cell death.34 For this reason, the use of external source of danger.37 Thus, although XKA protocols with XK is not recommended for ocular stu- protocol seemed to be a safe anaesthetic protocol, we dies in which corneal evaluation is important. However, showed that these rats may be feeling pain. if there is no other option to use a different anaesthetic The great difference of vocalization observed from combination, eye drops, 100% oxygen supplementation the protocol XKA with the other protocols can be and the administration of an adequate dose of yohim- because acepromazine is a phenothiazine that can pro- bine (to reverse the action of xylazine) are recom- duce moderate sedation but has no analgesic effect mended to minimize the incidence of corneal lesions. alone.27 On the other hand, opioids are analgesics The XKA protocol resulted in a better anaesthetic that interact with m-opioid receptors that inhibit pain. effect in males. This protocol allows procedures with a It has been demonstrated that the emission of 22 kHz duration time of approximately 25 minutes with a good USVs by rats subjected to experimental pain can be margin of safety. Based on Guedel’s stages of anaesthe- attenuated by the administration of drugs that possess sia, the loss of the pedal withdrawal reflex indicates that clinically relevant analgesic properties.8 the animal achieved the second plane of the third stage The increase in the number of vocalizations in the of anaesthesia.36 This means that those rats that lost 20–23-minute interval with the XKMo protocol in this reflex achieved surgical tolerance. The XKMe males could be because morphine can generate an exci- protocol had a longer duration of surgical anaesthesia tatory effect in rats39 that could generate the animals compared to XKA protocol having a recovery time vocalizing more. three times faster; however, fewer animals achieved Contrary to what was described by Wallace et al.,40 this level of anaesthesia. None of the animals achieved these results suggest that the USV can be used as a surgical anaesthesia with the XKMo and XKT proto- complementary behavioural parameter in experimental cols. Therefore, the use of these doses for surgical pro- models of pain. Nevertheless, this study was a pioneer cedures is not recommended. For females, XKMe and in using USV to assess pain in anaesthetized rats in a XKA were the most effective anaesthetic protocols. The recovery phase, so additional research is required to XKMe protocol had a longer surgical anaesthetic dur- verify possible anaesthesia influences on the USV pat- ation and a much faster recovery compared to XKA. tern of rats. Since methadone is an opioid that has great analgesic In conclusion, the XKA anaesthetic protocol was the effect, the XKMe combination may be a good option best protocol for male rats and was also suitable for for procedures that are invasive and painful. females, considering only the intraoperative period. When using the XKMo protocol, only two out of Indeed, the analgesic effect of this protocol is low, seven animals achieved surgical anaesthesia and the and the use of an intraoperative analgesic is recom- recovery period was very prolonged. These results sug- mended. The XKMe protocol resulted in better results gest that this protocol is not the best recommendation for females due to its high analgesic effect that can be 180 Laboratory Animals 54(2) used for more invasive surgical procedures. USV meas- 7. Hedenqvist P, Roughan JV and Flecknell PA. Effects of urement was revealed to be a useful tool to evaluate the repeated anaesthesia with ketamine/medetomidine and of potential analgesic effect of drugs used in rats subjected pre-anaesthetic administration of buprenorphine in rats. to anaesthesia for surgical procedures, and it could Lab Anim 2000; 34: 207–211. be used in other behavioural models used in 8. Oliveira AR and Barros HMT. Ultrasonic rat vocaliza- neuropharmacology. tions during the formalin test: A measure of the affective dimension of pain? Anesth Analg 2006; 102: 832–839. 9. Portfors CV. Types and functions of ultrasonic vocaliza- Acknowledgements tions in laboratory rats and mice. J Am Assoc Lab Anim We thank Maria Martha Bernardi, Aline Magalha˜ es and Sci 2007; 46: 28–34. Yameˆ Miniero for earlier discussions about the topic. We 10. Jourdan D, Ardid D, Chapuy E, et al. Audible and thank the anonymous reviewers for their thoughtful com- ultrasonic vocalization elicited by single electrical noci- ments that improved the manuscript. We would like to ceptive stimuli to the tail in the rat. Pain 1995; 63: thank Carol Valenzuela for the English language review. 237–249. 11. Sevostianovaa N, Zvartaub E, Bespalovb A, et al. Effects Declaration of Conflicting Interests of morphine on formalin-induced nociception in rats. Eur J Pharmacol 2003; 462: 109–113. The author(s) declared no potential conflicts of interest with 12. Shimoyama N, Shimoyama M, Elliott KJ, et al. respect to the research, authorship, and/or publication of this d-Methadone is antinociceptive in the rat formalin test. article. J Pharmacol Exp Ther 1997; 283: 648–652. 13. Zegre CC, Kissling GE, Goulding DR, et al. Analgesic Funding effects of tramadol, carprofen or multimodal analgesia in The author(s) disclosed receipt of the following financial sup- rats undergoing ventral laparotomy. Lab Animal 2011; port for the research, authorship, and/or publication of this 40: 85–93. article: this work was supported by the Coordenaça˜ ode 14. Miczek KA and Vivian JA. Effects of and opioid Aperfeiçoamento de Pessoal de Nı´vel Superior-Brazil agonists and antagonists on affective vocal and reflexive (CAPES) Finance Code 001. pain responses during social stress in rats. J Psychopharmacol 1998; 139: 364–375. 15. Zubieta JK, Smith YR, Bueller JA, et al. Regional mu- ORCID iD opioid receptor regulation of sensory and affective Jilma Alemań-Laporte https://orcid.org/0000-0003-4466- dimensions of pain. Science 2001; 293: 311–315. 512X 16. Abreu M, Aguado D, Benito J, et al. Reduction of the sevoflurane minimum alveolar concentration induced by References methadone, tramadol, butorphanol and morphine in rats. 1. Albrecht M, Henke J, Tacke S, et al. Effects of isoflurane, Lab Anim 2012; 46: 200–206. ketamine-xylazine and a combination of medetomidine, 17. Flecknell P. Laboratory Animal Anesthesia, 3rd ed. midazolam and fentanyl on physiological variables con- Newcastle: Elsevier Inc, 2009, p.300. tinuously measured by telemetry in Wistar rats. BMC 18. Santos M, Kunkar V, Garcıá-Iturral de P, et al. Vet Res 2014; 10: 198–212. Meloxicam, a specific COX-2 inhibitor, does not enhance 2. Gargiulo S, Greco A, Gramanzini M, et al. Mice anesthe- the isoflurane minimum alveolar concentration reduction sia, analgesia and care, part I: anesthetic considerations in produced by morphine in the rat. Anesth Analg 2004; 98: preclinical research. Ilar J 2012; 53: 55–69. 359–363. 3. Bougherara H and Bouaziz O. Effects of the anaesthetic/ 19. Ferrari L, Turrini G, Rostello C, et al. Evaluation of two tranquillizer treatments (Ketamine, Ketamine þ combinations of Domitor, Zoletil 100, and Euthatal to Acepromazine, Zoletil) on selected plasma biochemical obtain long-term non-recovery anesthesia in Sprague- parameters in laboratory rats. Cent Eur J Exp Biol 2014; Dawley rats. Comp Med 2005; 55: 256–264. 3: 1–5. 20. Burnham KP and Anderson DR. Model Selection and 4. Arras M, Autenried P, Rettich A, et al. Optimization of Multimodel Inference: A practical information-theoretic intraperitoneal injection anesthesia in mice: drugs, dos- approach, 2nd ed. New York: Springer-Verlag, 2002, ages, adverse effects, and anesthesia depth. Comp Med p.488. 2001; 51: 443–456. 21. Sawilowsky SS. New Effect Size Rules of Thumb. J Mod 5. Welberg LA, Kinkead B, Thrivikraman K, et al. Appl Stat Methods 2009; 8: 597–599. Ketamine-xylazine-acepromazine anesthesia and post- 22. Nakagawa S and Cuthill IC. Effect size, confidence inter- operative recovery in rats. J Am Assoc Lab Anim Sci val and statistical significance: a practical guide for biolo- 2006; 45: 13–20. gists. Biol Rev 2007; 82: 591–605. 6. Wellington D, Mikaelian I and Singer L. Comparison of 23. Zambricki EA and D’Alecy LG. Rat sex differences in ketamine–xylazine and ketamine–dexmedetomidine anes- anesthesia. Comp Med 2004; 54: 49–53. thesia and intraperitoneal tolerance in rats. J Am Assoc 24. Rodrigues SF, De Oliveira MA, Martins JO, et al. Lab Anim Sci 2013; 52: 481–487. Differential effects of chloral hydrate- and ketamine/ Alemań-Laporte et al. 181

xylazine-induced anesthesia by the S.C. route. Life Sci 33. Zhang F, Lo¨fgren S and So¨derberg PG. Interaction of 2006; 79: 1630–1637. anaesthetic drugs and UV-B irradiation in the anterior 25. Wixson SK, White WJ, Hughes HC, et al. The effects of segment of the rat eye. Acta Ophthalmol 2007; 85: pentobarbital, fentanyl–droperidol, ketamine–xylazine 745–752. and ketamine–diazepam on core and surface body tem- 34. Turner PV and Albassam MA. Susceptibility of rats to perature regulation in adult male rats. Lab Anim Sci corneal lesions after injectable anesthesia. Comp Med 1987; 37: 743–749. 2005; 55: 175–182. 26. Thurmon JC, Tranquilli WJ and Benson GJ. Lumb and 35. Calderone L and Grimes PSM. Acute reversible cataract Jones’ Veterinary Anesthesia and Analgesia, 3rd ed. induced anesthesia by xylazine and by in rats and mice. Baltimore: Williams and Wilkins, 1996, pp.210–240, Exp Eye Res 1986; 42: 331–337. 686–735. 36. Grimm KA, Lamont LA, Tranquili WJ, et al. 27. Seyler DE, Borowitz JL and Maickel RP. Calcium chan- Lumb and Jones’ Veterinary Anesthesia and Analgesia, nel blockade by certain opioids. Fundam Appl Toxicol 5th ed. Chichester: John Wiley & Sons, Ltd, 2015, 1983; 3: 536–542. pp.3–6. 28. Wamsley JK. Opioid receptors: autoradiography. 37. Barker DJ, Root DH, Ma S, et al. Dose-dependent dif- Pharmacol Rev 1983; 35: 69–83. ferences in short ultrasonic vocalizations emitted by rats 29. Sinclair MD. A review of the physiological effects of during cocaine self-administration. Psychopharmacol a2-agonists related to the clinical use of medetomidine 2010; 211: 435–442. in small animal practice. Can Vet J 2003; 44: 885–897. 38. Brudzynski SM, Bihari F, Ociepa D, et al. Analysis of 30. Hsu WH, Bellin SI, Dellmann HD, et al. Xylazine- 22 kHz ultrasonic vocalization in laboratory rats: long ketamine induced anesthesia in rats and its antagonism and short calls. Physiol Behav 1993; 54: 215–221. by yohimbine. J Am Assoc Lab Anim Sci 1986; 189: 39. Ayhan IH and Randrup A. Behavioral and pharmaco- 1040–1043. logical studies on morphine-induced excitation of 31. Talukder H and Hikasa Y. Diuretic effects of medetomi- rats. Possible relation to brain catecholamines. dine compared with xylazine in healthy dogs. Can J Vet Psychopharmacol 1973; 29: 317–328. Res 2009; 73: 224–236. 40. Wallace VCJ, Norbury TA and Rice ASC. Ultrasound 32. Suckow MA, Stevens KA and Wilson RP. The vocalization by rodents does not correlate with behav- Laboratory Rabbit, Guinea Pig, Hamster and Other ioral measures of persistent pain. Eur J Pain 2005; 9: Rodents. London: Academic Press, 2012, pp.34–53. 445–452.

Re´sume´ Dans cette e´tude, l’effet de quatre protocoles anesthe´siques dont l’0association de xylazine (X) et de ke´tamine (K) avec l’ace´promazine (A) et les opiace´s [me´thadone (Me), morphine (Mo) ou tramadol (T)] a e´te´ e´value´ chez des rats de laboratoire des deux sexes. La vocalisation ultrasonique (USV) a e´te´ utiliseé comme indicateur de la douleur pendant la pe´riode de re´tablissement. L’objectif e´tait d’e´valuer les parame`tres physiologiques et l’effet analge´sique de chaque protocole pour de´terminer le protocole le plus suˆrquire´ponde aux exigences d’une anesthe´sie e´quilibreé. Le protocole XKA s’est ave´re´ le meilleur pour les deux sexes, tandis que le SKMe e´tait pre´fe´rable pour les femelles parce que ces associations permettaient d’atteindre un niveau d’anesthe´sie chir- urgicale chez les rats. Cependant, l’e´valuation de la douleur au cours du test a` la formaline a re´ve´le´ que les rats anesthe´sie´sa` l’XKA produisaient davantage d’USV, sugge´rant qu’il ne s’agissait pas d’un bon protocole pour controˆler la douleur postope´ratoire imme´diate. Tous les protocoles ont induit une baisse de la tempe´rature corporelle et du de´bit respiratoire et cardiaque et ont eu des effets importants, les rats urinant et gardant les yeux ouverts. Seuls les rats anesthe´sie´s selon le protocole XKA n’a pas pre´sente´ d’horripilation. Ces re´sultats ont de´montre´ qu’une bonne surveillance et des soins de qualite´ au cours de l’anesthe´sie doivent eˆtre pre´vus pour pre´venir les complications qui compromettent la vie de l’animal et lui assurer un bon re´tablissement. L’inclusion de l’analge´sie dans les protocoles d’anesthe´sie doit eˆtre utiliseédefaçon routinie`re en minimisant la douleur et en assurant donc des re´sultats plus fiables au cours des proce´dures expe´rimentales.

Abstract In dieser Studie wurde die Wirkung von vier Ana¨sthesieprotokollen, die die Kombination von Xylazin (X) und Ketamin (K) mit Acepromazin (A) und Opioiden (Methadon (Me), Morphin (Mo) oder Tramadol (T)) umfassten, bei Laborratten beiderlei Geschlechts untersucht. Ultraschallvokalisation (USV) wurde als Indikator fu¨r Schmerzen wa¨hrend der Erholungsphase benutzt. Ziel war es, die physiologischen Parameter und die analge- tische Wirkung der einzelnen Protokolle zu bewerten, um zu ermitteln, welches Protokoll am sichersten ist und die Anforderungen an eine ausgewogene Ana¨sthesie erfu¨llt. Die besseren Protokolle waren das XKA-Protokoll fu¨r beide Geschlechter und das XKMe-Protokoll fu¨r weibliche Tiere, da die Kombinationen 182 Laboratory Animals 54(2) chirurgisches Narkosestadium bei Ratten erreichen. Die Schmerzbeurteilung wa¨hrend des Formalin-Tests ergab jedoch, dass Ratten, die mit XKA beta¨ubt wurden, mehr USV produzierten, was darauf hindeutet, dass dies nicht als gutes Protokoll zur Kontrolle unmittelbarer postoperativer Schmerzen zu werten ist. Alle Protokolle verursachten eine Absenkung der Ko¨rpertemperatur sowie der Atem- und Herzfrequenzen und hatten wichtige Auswirkungen, wie z. B. Miktion und Augen-Offen-Halten. Nur mit dem XKA-Protokoll beta¨ubte Ratten zeigten keine Piloarrektion. Diese Ergebnisse belegen, dass eine gute U¨berwachung und Betreuung wa¨hrend der Narkose erfolgen muss, um das Leben des Tieres beeintra¨chtigende Komplikationen zu vermeiden und eine gute Erholung zu gewa¨hrleisten. Analgesie muss in Ana¨sthesieprotokollen routinema¨ßig beru¨cksichtigt werden, um eine minimale Schmerzbelastung und damit zuverla¨ssigere Ergebnisse in den experimentellen Verfahren zu gewa¨hrleisten.

Resumen En este estudio, el efecto de cuatro protocolos aneste´sicos que incluıán la combinacioń de xilacina (X) y ketamina (K) con acepromacina (A) y opioides [metadona (Me), morfina (Mo) o tramadol (T)] se evaluo´ en ratas de laboratorio de ambos sexos. La vocalizacioń ultrasońica (USV) fue utilizada como indicador de dolor durante el periodo de recuperacioń. El objetivo era evaluar los para´metros fisiolo´gicos y el efecto analge´sico de cada protocolo para determinar que´ protocolo era el ma´s seguro y cumplıá con los requisitos de una anestesia equilibrada. Los mejores protocolos fueron XKA para ambos sexos y XKMe para las hembras porque las combinaciones conseguıán un plano quiru´rgico de anestesia en ratas. Sin embargo, la evaluacioń de dolor durante la prueba de formol revelo´ que las ratas anestesiadas con XKA producıán ma´snu´meros de USV, lo que sugerıá que no es un buen protocolo para el control del dolor postoperativo inmediato. Todos los protocolos producıán un descenso de la temperatura corporal y de la frecuencia cardıáca y respiratoria y, asimismo, tuvieron efectos importantes como miccioń y un mantenimiento de los ojos abiertos. Solo las ratas anestesiadas con el protocolo XKA no presentaron piloereccioń. Estos resultados demostraron que se debe incluir un buen control y cuidado durante la anestesia para evitar complicaciones que pongan en peligro la vida del animal y para garantizar una buena recuperacioń. La inclusioń de analgesia en los protocolos de anestesia debe usarse de forma rutinaria, garantizando una presencia de dolor mıńima y, por tanto, unos resultados ma´s fiables en los procedimientos experimentales. Original Article Laboratory Animals 2020, Vol. 54(2) 183–191 ! The Author(s) 2019 Examining compliance with ethical Article reuse guidelines: sagepub.com/journals-permissions standards for animal research: DOI: 10.1177/0023677219841080 journals.sagepub.com/home/lan is there a need for refinement? A qualitative study from northern Europe

Aurora Brønstad1 and Peter Sandøe2

Abstract Ethical guidelines for research on animals such as the 3Rs (Replacing, Reducing, Refining) and positive harm- benefit evaluations are anchored in EU Directive 2010/63. In this qualitative study we investigated how ethical guidelines interact and/or compete with other considerations when animal research is planned. Four focus groups consisting mainly of researchers involved in animal use were conducted in four Northern European countries and findings were analysed thematically with the support of NVIVO. Practical issues and the importance of doing good science were dominant topics. Practical issues could not easily be separated from the goal of good science. Participants expressed concerns which accord with the core-values of the 3Rs, but in one group they explicitly referred to the 3Rs as a concept. Conflicts between reductions in animal numbers and the risk of creating unreliable results were addressed. They also criticized the practice of using more animals to improve statistical figures to get results published in highly ranked journals – a finding we believe is new. The main conclusion of this study is that ethical values could not easily be separated from the goal of producing good science. Whereas policy makers seem to expect researchers to explicitly take ethical considerations into account, we found that their ethical thinking is mainly manifested as an implicit part of methodology and design. We don’t see this as a problem as long as the underlying core values are implicitly respected, or promoted, in the relevant experimental practice.

Keywords 3Rs, animal use, ethics, ethics and welfare

Date received: 19 July 2018; accepted: 10 March 2019

Introduction their number, or by Refining procedures to minimize or eliminate harm to them.6 Directive 2010/638 article 1 Animal research raises ethical issues. In most societies explicitly employs the language of the 3Rs: there are conflicting views about under what conditions it is ethically justifiable.1–7 In Europe, this kind of This Directive establishes measures for the protection research is regulated through a common directive of animals used for scientific or educational purposes. implemented in different national legislations that defines minimum standards for animal use in research, 1Department of Clinical Medicine, University of Bergen, Norway teaching and testing and is designed to reflect and 2Department of Veterinary and Animal Sciences and Department of enforce widely accepted ethical standards. Food and Resource Economics, University of Copenhagen, Among the criteria defined in EU Directive 2010/638 Denmark are the principles that all experiments must implement Corresponding author: the 3Rs6 and be based on a positive harm-benefit ana- 5 Aurora Brønstad, Dyreavdelingen, University of Bergen Vivarium, lysis. The 3Rs aim to reduce negative impacts on sen- Haukeland sykehus Bergen, No-5021 Norway. tient animals either by Replacing animals, Reducing Email: [email protected] 184 Laboratory Animals 54(2)

To that end, it lays down rules on the following: (a) the researchers and three as senior researchers. In addition, replacement and reduction of the use of animals in pro- one identified him/herself as a research animal veterin- cedures and the refinement of the breeding, accommo- arian, one as a technician and one as a bioethical assist- dation, care and use of animals in procedures. ant. The contributions of the bioethical assistant did not change the dynamics or the content of the discus- The 3Rs – and especially refinement – have had a cen- sion as compared with the other focus groups. Different tral place in the evaluation by the competent bodies levels and types of competence were represented in all (IACUC, Ethical committees).9,10 Although the direct- focus groups. At the start of the interviews the partici- ive defines the role of advisory bodies on the 3Rs, the pants introduced themselves, as they did not always researchers are mainly responsible for ensuring that the know each other, and described their research projects. 3Rs are properly applied. The 3Rs have been investi- The interviews lasted for 60–90 minutes and were gated in earlier studies. The findings were mixed,3,11–13 recorded and transcribed verbatim. Two of them were and differences between scientists’ views and policy conducted in English and two in Scandinavian goals were identified.14 languages. Harm–benefit assessment is another import- The participants were first asked what they thought ant,5,9,15–19 mandatory8 part of project evaluation.20–23 was important, second about their motivation, and The harm–benefit analysis brings the particular project finally about difficult issues when planning animal into a context where the harm to animals is justified by research. The question about important factors was expected benefits.5 Harm–benefit evaluations resonate asked because we assumed that the comments partici- with the public1,2,24 as well as policymakers,8,25–27 pants spontaneously made in response to it would be although they can be difficult to compare and weigh likely to indicate the matters to which they gave most against each other.15,18 attention in planning their research. For the question Research activities face time and resource pressures, about motivation, it was assumed that these would also and researchers must define priorities when they plan receive much attention. The final question about diffi- experiments. There is a risk that the aspiration to apply culties was asked to explore if there were any ambiva- high ethical standards will compete with other interests lences, or conflicts, in the planning of the animal or needs. Guidelines requiring researchers to continue experiments. As a follow-up question, interviewees to adhere to ethical standards have been defined both were asked how they solved their difficulties. No for the planning phase28 and the reporting of animal attempt was made to lead the discussion towards eth- research.29,30 Several bodies have defined codes of con- ical issues, but relevant issues came up spontaneously in duct describing acceptable research practice.31–33 all interviews, and all participants contributed by refer- Compliance with these codes is crucial if scientific integ- ring to their own experiences or concerns. Additional rity and credibility is to be maintained.31 follow-up questions were put to the informants to The aim of this qualitative study, based on inter- encourage them to further elaborate their contribu- views with researchers in four Northern European tions. All participants contributed to the open discus- countries, was to deepen our understanding of the sions, bringing in new elements or commenting on ways in which, in practice, ethical guidance, and in par- other participants’ contributions. There was no need ticular that given by the 3Rs, interacts and/or competes to interrupt or rein in dominant informants. with other considerations when research involving We assumed that ethical standards must be a part of animal use is being planned. the planning of an experiment if they are to be taken into consideration in the performance of subsequent Methods research. Further, we assumed that spontaneously referring to ethical norms reflects moral sensitivity Four focus group interviews were performed in four about the use of animals in research34 which empha- countries (the Netherlands, Denmark, Sweden and sizes the researcher’s commitment to practices that Norway) between 2011 and 2013. The interviewees maintain ethical standards. were recruited through contacts who volunteered to The transcribed interviews were coded and themat- organize the events, and no incentives to participate ically analysed using NVIVO11 software. A provisional were offered. In total 21 interviewees participated, list of codes35 included ‘3R’, ‘replacement’, ‘reduction’ 11 males and 10 females. The interviewees’ contribu- and ‘refinement’. In applying these we were guided by tions were anonymized. Their age was between 25 and interpretations of the meaning of the comments, that is 56 years (average 35, median 34 years) with between 0 we did not require explicit matches with the wording and 30 years of experience in animal research (average of the comments. The codes were then sub-coded as 8.4, median 5 years). Eight participants identified them- ‘3R explicit’ or ‘3R values’. In an inductive approach, selves as PhD students, two as post-docs, five as additional codes were added during the analysis to Brønstad and Sandøe 185 systematize issues raised during the discussion.35 The You got a p value different or journal ...The question secondary codes also included ‘other ethical concerns’. is, of course, is it ethical? (1F2) This sub-code, which covered ethical topics that could not be directly related to the context of the 3Rs, was Reducing animal numbers by excluding pilots or con- subdivided into animal related or non-animal related trols was also discussed. Pilots can provide important issues. The secondary codes also included ‘operational information for further studies. Controls are important issues’ and ‘good science’, both of which turned out to elements of experimental design. Some respondents be major topics. wondered whether historical controls should be used Participants signed informed consent forms before more often. the interviews. Demographic data was recorded in other documents and stored in a different folder to you can use historical data actually as a control. It assure anonymity. Studies such as the present one depends on what type of study you’re [doing]. (2M3) does not require ethical approval in Norway, but they must be and were notified to the Norwegian Social Better sharing of organs and tissues from animals to Science Data Service. reduce animal numbers, was also discussed. Respect for the interest of healthy animals to stay alive was Results conveyed in another interview, and frustration that research regulations require healthy animals to be Solutions to practical issues and the importance of killed was also expressed. doing good science were dominant topics in all focus groups. The practical issues could not easily be sepa- That is the basic rule here – all animals come in – they rated from the goal of doing good science in the ana- don’t come out ...so therefore we have to kill it. That is lysis, because these themes are interrelated. stupid – a healthy animal! (1M2) The practical issues the respondents considered included the acquisition of ethical permissions, the The dilemma between using fewer animals and refining availability of resources, the choice of animal proced- the housing conditions was raised in one of the inter- ures, logistics and information or communication chal- views. The respondents appeared to think that this lenges. The discussions of good science addressed the dilemma has no clear solution, but they demonstrated planning and designing of experiments, the choice of ethical awareness and the ability to identify and reflect models, pilots, statistical considerations revolving on such dilemmas. around p values, and the reporting of animal experiments. There was agreement that proper design is I think sometimes that there is too much focus on the necessary to obtain valid, reproducible results, and reduction bit and too little on the refinement ...To me that animal use can be justified only if the design is in it makes more sense to look at the refinement, to look order. This may, however, create a tension with the at how do we reduce the stress on the animals ...(2F1) demand for a reduction in the number of animals used. Solving practical tasks was a recurrent topic in all of the Which means that if you are very concerned about the interviews. The collection of information enabling good ethics, or the large number of animals you are using, practical solutions while taking animal-welfare concerns trying to downsize it [animal number] in some way the into consideration, was an important focus. Although study is unethical, because you are decreasing your such strategies fall into the refinement category, the chances to actually get the correct answer. (2M3 – word ‘refinement’ was not often used. However, the interview/sex/informant) word ‘stress’ was frequently used when negative impacts on animals were being described. The importance of On the other hand, criticism was voiced about the idea treating animals well to avoid stress of the kind that that more animals should be used simply to improve may bias an experiment was frequently raised. statistical power – with the aim of getting results published in a more highly ranked journal. you ensure that the animals are not stressed, so you’re getting something that is physiologically meaningful ... but then there is more needed to get it published, to get you want the animal to be happy, and to be as close to it published in a better journal. People they look at its normal [way of life] as possible, to be sure that ...I your output, in what kind of journals you publish. mean, both to reduce the data variability, but also to, I (1M1) mean, to be sure that the data you are getting actually So it depends on impact factors, you will increase how reflects like a normal physiological regulation, and not many animals you will need. (1M2) a regulation in response to stress ...(2M2) 186 Laboratory Animals 54(2)

Accurate information about tolerance level for an Consideration of alternatives was not limited to the animal – for example by injection – makes a significant planning phase. According to one participant, alterna- difference to the welfare of that animal. Searches for tives should be reconsidered as the research project information on how to improve or refine procedures progresses. were important in the planning of experiments. At a certain point you have to say we cannot see this in When we start with new tumors, we look up in the an animal – maybe we should go back to the lab and literature – What can we inject in these mice, and see if we can do some in vitro [investigations]. (1F3) how should we inject it? (1F2) Although the informants raised concerns about alter- Web-based literature searches were identified as useful natives, animal numbers and strategies to refine and methods for collecting information. However, talking improve procedures, only a few occasions they referred to experienced colleagues seemed to be a preferred explicitly to the 3Rs. When they did so they treated the strategy for collecting information, especially when it 3Rs as a tool for continuous improvement and change came to refinement. of practices.

There are a lot of things written down, but other things can we do it in a diﬀerent way? And having that not, and sometimes you go through the hard way, policy put into the system, also with the 3Rs award asking people, asking people – and calling people. and so on, I think that’s a good instrument actually And then you ﬁnd out. (1M2) to try to making it ...eh ...make it more natural, you I have also talked with people – I mean, a lot of people can say – to put your ideas actually into a format that is already have experience, and with this type of model, also put into reality. (2M2) and what are the pitfalls. (1M3)

Talking with experts also served the purpose of build- Discussion ing good collaborative relationships – for example, with the animal care staff. The aim of this study was to explore researchers’ attention to ethical standards and concerns when they are My experience is that the more contact, the more they planning animal experiments, with a special focus on look after the animals ...(1M1) the 3Rs. During the analysis the practical issues anchored in Replacing animal research with experimental ethical values could not easily be separated from the approaches using alternative methods is a goal of goal of producing good science because these two elem- policy makers.8 However, the respondents voiced con- ents are interdependent. This created some methodo- cerns about the consequences for science if this goal is logical problems for the analysis. On the other hand, enforced when no full replacements are available. On it demonstrates that the identification of practical solu- the other hand, they expressed disapproval of the situ- tions is a prerequisite of good science, and there is a ation where animals are still used, and where regula- room for the application of ethical standards in achiev- tions may prohibit the use of available alternatives. ing this. While the present study confirmed that scientists put actually the in-vitro test is so much more sensitive than a strong emphasis on scientific standards, it also the rabbit test, eh? So it’s quite astonishing that the demonstrated that researchers address values in accord- FDA has been so reluctant to accept that. ance with ethical standards for research, including the It’s the health authorities that actually have the last 3Rs. When the 3R issues were addressed, they were word ...(2F1) usually connected with practical solutions. On only a few occasions, and in only one of the interviews, were When the necessity of animal models was discussed, the 3Rs explicitly referred to in their own right. The respondents often focused on the idea that there was relevant EU directive, however, requires researchers no need to use animals. Already, in the planning of an to prepare an explicit statement explaining how com- animal experiment and the choice of model, the ques- pliance with the 3Rs has been ensured. It requires: A tion of alternatives was raised. demonstration of the compliance with the requirement of replacement, reduction and refinement (article 43 1.b).8 Does it [the animal model] tell you what you want to The spirit of the 3Rs is mentioned in several places in know? Is it really needed? I mean some basic questions the directive (Article 1.1 a, Article 4, Article 13.2 a-c, can sometimes be done in vitro. (1M2) Article 27.1.b, ANNEX V.10, ANNEX VI.2).8 Brønstad and Sandøe 187

In ANNEX VI.2 the 3Rs are explicitly referred to: we interviewed. This still accords with Russell and ‘2. Application of methods to replace, reduce and Burch’s definition of reduction: refine the use of animals in procedures’. ANNEX VI. 3–5 states that the researcher must describe any meth- Reduction in the number of animals used to obtain odological strategies that have been employed to avoid information of a given amount and precision.6 pain, suffering and to achieve humane endpoints – all of these are examples of refinement. Similarly, ANNEX The use of animals for a certain experimental purpose, VI.6 requires researchers to describe strategies they or in a certain context, brings the benefit domain into have used to minimize or reduce the number of animals play. Some benefits are more legitimate than others. used. The ANNEX VI guidance may indeed encourage Increasing animal numbers to improve the statistical researchers to perceive the instruction to demonstrate validity of a study is acceptable; using more animals compliance with the 3Rs as a detached element rather to strengthen a study statistically in order to get studies than an integrated part of the methods they select in published in higher ranked journals as a personal bene- order to maximize scientific benefits or reduce negative fit was regarded unacceptable – although, of course, impact on animals. there is a fine line between these two goals. To our A recent paper by Kirk37 stresses that Russell and knowledge, the issue of increasing animal numbers to Burch framed the 3Rs, not as an ethical tool, but as a improve statistical parameters in order to get results strategy to maximize high-quality science. The 3Rs are published in higher ranked journals, or journals with only one pathway among others to high-quality science, higher impact scores, has not been raised before. Such a and this might explain the informants’ tendency to practice has troubling implications for animals and can focus on ‘good science’ without explicitly referring to also be criticized for bad data management and violat- the 3Rs. ing principles of good science.31–33 Young scientists As the 3Rs were launched as a concept for humane especially may be under pressure from senior colleagues experimental technique by Russell and Burch,6 or journal reviewers asking for more experiments. Wu¨ rbel19 launched the 3Vs as a principle to maximize The preference for talking with experts rather scientific validity. According to Wu¨ rbel and others, the than conducting a literature search is at odds with an assessment of scientific quality should be part of the earlier study from Canada, where web-based searches ethical evaluation of animal experiments.19,21,23,31,38,39 were the preferred method of collecting information.11 Some authors regard the scientific quality of experi- Information collection was one of the priorities men- ments as a part of the harm domain.21,23,39 Others tioned in connection with the planning of experiments, treat it as part of the benefit domain.40–42 By contrast, but the participants in this study preferred to talk to Bateson38 and Wu¨ rbel19 assess quality as a separate colleagues rather than search databases. This accords domain. with van Luijk et al.,13 a study showing that 3Rs data- In a previous study of researchers’ attitudes to the bases were rarely used and that other forms of know- 3Rs, only a minority were found to be aware of the 3Rs ledge exchange were preferred.13 and able to properly name the principles.3 This accords A web-based study by Fenwick et al.11 discovered with findings of the present study, that is only in one that 3Rs assistance should be ‘constructive’ and ‘neu- interview were the 3Rs explicitly mentioned. However, tral’ and provided by an ‘expert’. The same study also it might be asked whether it is important to be able to suggested that such 3Rs assistance should not be man- name 3Rs correctly as long as the core value of limiting datory but be offered when problems are identified. harm are implicitly respected, or promoted, in the rele- Results from our qualitative study indicated that vant experimental practice, as they seem to be in researchers are very aware of their need to seek this study. advice, they seem to have a good relationship with The same study3 showed that researchers favoured their 3Rs advisors, and that advice was provided in a refinement over reduction. The dilemma of reduction collegial, collaborative manner. Studies have shown versus refinement was spontaneously raised in one of that the collaborative approach is more efficient, the focus groups. Here, the view that refinement was when it comes to compliance with the advice given, more important was again expressed. In our study, than a ‘paternal’ relationship where a superior instruct however, reduction issues were addressed more fre- on the right solution. Hence, the collaborative quently than refinement issues. Concerns about using approach has been described as the best way of building too few animals to generate reliable data, were voiced a culture in which the 3Rs are governing values.36 and regarded as an unnecessary and unethical animal Lack of reproducibility, and the validity of experi- use. The inclusion of a sufficient number of animals to ments more generally, is a major concern in preclinical assure statistical power was very important, and neces- science.43,44 A strong commitment to principles of good sary to justify animal use, according to the researchers science, as addressed in this study, reflects a serious 188 Laboratory Animals 54(2) emphasis on achieving this. However, the attention The results of this study show that researchers do given to principles of good science expressed in this indeed address values and work in accordance with study seems to be at odds with the existence of a repro- the 3Rs. However, terms such as ‘necessity’ or ‘stress’ ducibility crisis.43,44 One explanation may be that good are rather used than the official terminology. This has intentions in the planning phase are not necessarily fol- implications for the way we should examine the appli- lowed up in practice. This may be because animal cation of the 3Rs. For example, instead of asking what experiments are complex, with many pitfalls, so that replacement alternatives have been considered, it may good intentions drown in practicalities. be more productive to ask a researcher: why are ani- Both the finding that researchers preferred to talk mals necessary for this study? Similarly, instead of with, and collect information from, colleagues and the asking what refinement strategies have been applied, point about the complexity and pitfalls of animal stu- we could ask: what decisions have been made which dies suggest that it is necessary to establish strong, com- reduce unnecessary stress in the animals? Talking the petent, updated support functions, as well as nurturing same ‘language’ will probably nurture a more fruitful a collaborative culture around animal research activ- and collaborative collegial relationship, and it may ities to better facilitate both good science and high also stimulate reflection on current practices and standards of animal welfare. changes to these. A qualitative method was chosen for this study, The respondents in this study were never asked dir- since the aim was to explore the role of ethical stand- ectly what place ethical standards had in the planning ards in the planning of animal experiments, and to of their experiments. Nevertheless, they spontaneously investigate how these standards are addressed and com- brought ethical issues into the discussion, and in doing plied with – all with a focus on the 3Rs. A qualitative this they demonstrated an ethical awareness of the approach opens up a more nuanced picture than can be issues raised by the use of animals in research, and obtained by quantitative investigation11,35,45 or by empathy for their animals. using questionnaires of the sort adopted in earlier stu- 3,11,46 dies. A qualitative approach provides a better Acknowledgements understanding of the way in which concepts like those We are grateful to Merel Ritskes-Hoitinga, Lars Friis of replacement, reduction and refinement are perceived, 45 Mikkelsene and Anders Forslid for assistance in recruiting interpreted, phrased and applied. It provides insight participants in this study; to the University of Bergen for into the researcher’s perspective that is useful in plan- ‘Sma˚ forsk’ funding and Charles River Laboratories for ning as well as reporting and evaluating studies. travel grants; to Jesper Lassen and Thomas Bøker Lund for Qualitative studies are not suitable for quantifying, or advice on methodology and Paul Robinson for language pol- discovering the distribution of certain opinions and ishing; and to Joseph for being a patient listener, challenging their results should not be generalized. The strength our ideas and giving constructive feedback from a lay of the focus group method is that participants in such perspective. groups share, compare and discuss different views, and give feedback on each other’s opinions.45 There is Declaration of Conflicting Interests always a danger that some individuals will dominate The author(s) declared no potential conflicts of interest with the discussion and obscure other group-members’ opin- respect to the research, authorship, and/or publication of this ions, but this was not seen in the present study. Four article. northern European counties were chosen, as we assumed that they are rather similar in terms of their Funding levels of legislation and more generally in the attention they give to animal welfare. The author(s) received no financial support for the research, authorship, and/or publication of this article. The data in this study were collected during the period of transition to full implementation of the new directive. Although the 3Rs as such are referred to for ORCID iDs the first time in the 2010 directive, they are also present, Aurora Brønstad http://orcid.org/0000-0002-5511-5425 albeit without being explicitly set out, in the previous Peter Sandøe http://orcid.org/0000-0003-0397-3273 1986 directive (Articles 7 2, 3 and 4). We suspect that stronger explicit emphasis on the 3Rs in the 2010 direct- References ive has affected the way researchers talk about changes 1. Lund TB, Lassen J and Sandoe P. Public attitude and improvement in terms of the 3Rs, even if the formation regarding animal research. Anthrozoo¨s 2012; changes are part of a process that takes time. We there- 25: 475–490. fore believe that our findings still give valuable insight 2. Lund TB, Mørkbak MR, Lassen J, et al. Painful dilem- into the thinking of researchers working with animals. mas: A study of the way the public’s assessment of animal Brønstad and Sandøe 189

research balances costs to animals against human 18. Grimm H. Turning apples into oranges? The harm-ben- benefits. Public Underst Sci 2014; 23: 428–444. efit analysis and how to take ethical considerations into 3. Franco NH and Olsson IA. Scientists and the 3Rs: atti- account. ATLA 2015; 43: 22–24. tudes to animal use in biomedical research and the effect 19. Wu¨ rbel H. More than 3Rs: the importance of scientific of mandatory training in laboratory animal science. Lab validity for harm-benefit analysis of animal research. Lab Anim 2014; 48: 50–60. Anim 2017; 46: 164–166. 4. Lund TB, Sørensen TI, Olsson IA, Hansen AK and 20. Laber K, Newcomer CE, Decelle T, Everitt JI, Guillen J Sandøe P. Is it acceptable to use animals to model and Brønstad A. Recommendations for addressing obese humans? A critical discussion of two arguments harm–benefit analysis and implementation in ethical against the use of animals in obesity research. J Med evaluation: Report from the AALAS–FELASA Ethics 2014; 40: 320–324. Working Group on harm–benefit analysis, Part 2. Lab 5. Bronstad A, Newcomer CE, Decelle T, Everitt JI, Guillen Anim 2016; 50: 21–42. J and Laber K. Current concepts of harm-benefit analysis 21. Stafleu FR, Tramper R, Vorstenbosch J and Joles JA. of animal experiments: Report from the AALAS- The ethical acceptability of animal experiments: A pro- FELASA Working Group on Harm-Benefit Analysis. posal for a system to support decision-making. Lab Anim Part 1. Lab Anim 2016; 50(1 Suppl): 1–20. 1999; 33: 295–303. 6. Russell WMS and Burch RL. The Principles of Humane 22. Mellor D and Reid C. Concepts of animal well-being and Experimental Technique. London: Methuen, 1959. predicting the impact of procedures on experimental ani- 7. Degrazia D. On the ethics of animal research. mals. In: Baker RM, Jenkin G and Mellor DJ (eds) In: Ashcroft RE, et al. (eds) Principles of Health Care Improving the Well-Being of Animals in the Research Ethics. 2nd ed. Chichester: Wiley, 2007, pp. 689–695. Environment. Urrbrae, Australia: Australian and New 8. European Commission. Directive 2010/63/EU of the Zealand Council for the Care of Animals in Research European Parliament and of the council of 22 September and Teaching, 1994, pp. 3–18. 2010 on the protection of animals used for scientific 23. Porter DG. Ethical scores for animal experiments. Nature purposes, https://eur-lex.europa.eu/LexUriServ/ 1992; 356: 101–102. LexUriServ.do?uri=OJ:L:2010:276:0033:0079:en:PDF 24. Swedish Research Council. Public Opinion in Sweden on the EuropeanCommission (2010). Use of Animals Research. 2008, https://www.vr.se/english/ 9. Bout HJ, Fentener van Vlissingen JM and Karssing ED. analysis-and-assignments/we-analyse-and-evaluate/all- Evaluating the ethical acceptability of animal research. publications/publications/2008-06-10-public-opinion-in- Lab Anim (NY) 2014; 43: 411–414. sweden-on-the-use-of-animals-in-research.html. 10. Silverman J, Lidz CW, Clayfield JC, Murray A, Simon 25. OIE. Terrestrial Animal Health Code, section 7.8 Use of LJ and Rondeau RG. Decision making and the animals in research and education. W. O. f. A. Health, IACUC: Part 1 – protocol information discussed at http://www.oie.int/en/standard-setting/terrestrial-code/ full-committee reviews. J Am Assoc Lab Anim Sci 2015; access-online/ (2018). 54: 389–398. 26. Haywood JR, et al. CIOMS-ICLAS International guiding 11. Fenwick N, Danielson P and Griffin G. Survey of principles for biomedical research involving animals, Canadian animal-based researchers’ views on the Three CIOMS-ICLAS. International Council for Laboratory Rs: replacement, reduction and refinement. PLoS One Animal Science https://www.google.com/url?sa¼t&rct¼ 2011; 6: e22478. j&q¼&esrc¼s&source¼web&cd¼1&ved¼2ahUKEwj3x– 12. Views on the Three Rs. National Centre for Replacement, 73u3hAhVKalAKHW92BocQFjAAegQIBBAC&url¼ht- Refinement and Reduction of Animals in Research, tps%3A%2F%2Folaw.nih.gov%2Fsites%2Fdefault% 2008, Faunalytics, https://faunalytics.org/views-on-the- 2Ffiles%2FGuiding_Principles_2012.pdf&usg¼AOvVaw 3rs-survey-report-2008/#. 0uKueEV5Iw6Bauh9F-ODhg 2012. 13. van Luijk J, Cuijpers Y, van der Vaart L, de Roo TC, 27. Garber JC, et al. Guide for the Care and Use of Leenaars M and Ritskes-Hoitinga M. Assessing the Laboratory Animals. 8th ed. Washington, DC: National application of the 3Rs: A survey among animal welfare Academies Press, 2010. officers in The Netherlands. Lab Anim 2013; 47: 210–219. 28. Smith AJ, Clutton RE, Lilley E, Hansen KEA and 14. Fenwick N and Griffin G. Scientists’ Views on the Three Brattelid T. PREPARE: guidelines for planning animal Rs: Comparison of Canadian and UK scientists in World research and testing. Lab Anim 2018; 52: 135–141. Congresses on alternatives and animal use in the life sci- 29. Hooijmans C, de Vries R, Leenaars M and Ritskes- ences. Montreal: ALTEX, 2011. Hoitinga M. The Gold Standard Publication Checklist 15. Balancing harm and benefit, ANZCCART, http://www. (GSPC) for improved design, reporting and scientific adelaide.edu.au/ANZCCART/. quality of animal studies GSPC versus ARRIVE guide- 16. Buning T de C. Practical difficulties in balancing harms lines. Lab Anim 2011; 45: 61. and benefits in the modern use of laboratory animals: 30. Kilkenny C, Browne WJ, Cuthill IC, Emerson M and Biotechnology. Altern Lab Anim 2004; 32: 459–463. Altman DG. Improving bioscience research reporting: 17. Dolan K. Cost-Benefit: The balancing act. In: Dolan K The ARRIVE guidelines for reporting animal research. (ed.) Ethics, Animals and Science. Blackwell and Keeley, J Pharmacol Pharmacother 2010; 1: 94–99. Iowa State University Press, Iowa State University, 1999, 31. European Science Foundation. The European code of con- pp. 211–235. duct for research integrity. Berlin: ALLEA, 2011. 190 Laboratory Animals 54(2)

32. European Science Foundation. The European code of con- 40. Rickard MD. The use of animals for research on animal duct for research integrity. Rev. ed. Berlin: ALLEA, 2017. diseases: Its impact on the harm-benefit analysis. Altern 33. OECD. Best Practices for Ensuring Scientific Integrity Lab Anim 2004; 32: 225–227. and Preventing Misconduct. http://www.oecd.org/sci- 41. Schuppli CA, Fraser D and McDonald M. Expanding the ence/inno/40188303.pdf (2007). three Rs to meet new challenges in humane animal experi- 34. Tynes-Pedersen LJ. See no evil: moral sensitivity in the mentation. Altern Lab Anim 2004; 32: 525–532. formulation of business problems. Business Ethics 2009; 42. Smith JA, van den Broek FA, Martorell JC, et al. 18: 335–348. Principles and Practice in Ethical Review of Animal 35. Miles BM and Huberman AM. Early steps in analysis. Experiments Across Europe. Summary of the report of a In Qualitative Data Analysis: An expanded sourcebook. FELASA working group on ethical evaluation of animal Thousand Oaks, California: SAGE, 1994, pp. 50–88. experiments. 2005. Lab Anim 41: 143–160. 36. Brønstad A and Berg AG. The role of organizational 43. Baker M. Is there a reproducibility crisis? Nature 2016; culture in compliance with the principles of the 3Rs. 533: 452–454. Lab Anim (NY) 2011; 40: 22–26. 44. Begley CG and Ellis LM. Drug development: Raise 37. Kirk RGW. Recovering the principles of humane experi- standards for preclinical cancer research. Nature 2012; mental technique: The 3Rs and the human essence of 483: 531–533. animal research. Sci Technol Hum Values 2018; 43: 45. Morgan DL and Krueger RA. Why should you use focus 622–648. groups? In Focus Group Kit1. USA: SAGE Publications, 38. Bateson P. When to experiment on animals. New Sci 1997, pp. 9–15. 1986; 109: 30–32. 46. van Luijk J, Leenaars M, van Dongen AM, van der Vaart 39. Voipio H-M, et al. Nordic-European workshop on eth- L and Ritskes-Hoitinga M. Outcomes of a Dutch work- ical evaluation og animal experiments. Scand J Lab Anim shop on improvements for the 3Rs in daily practice. Sci 2004; 4: 251–267. ALTEX 2012; 29: 440–443.

Re´sume´ Les lignes directrices e´thiques concernant la recherche meneé sur les animaux, tels que les 3Rs et les e´valuations dommage-avantage, sont ancreés dans la Directive de l’UE 2010/63. Dans cette e´tude qualitative nous avons e´tudie´ comment les lignes directrices d’e´thique interagissent avec et/ou concurrencent d’autres e´le´ments a` prendre en compte lorsqu’une recherche animale est pre´vue. Quatre groupes de discussion compose´s principalement de chercheurs implique´s dans l’utilisation des animaux ont e´te´ organise´s dans quatre pays d’Europe du Nord et les conclusions ont e´te´ analyseés de façonthe´matique avec l’aide de NVIVO. Les questions pratiques et l’importance de pratiquer une bonne science e´taient des sujets dominants. Les questions pratiques ne pouvaient pas eˆtre facilement se´pareés de l’objectif d’une bonne science. Les participants ont exprime´ des preóccupations en accord avec les valeurs fondamentales des 3R, mais dans un groupe, ils ont expresse´ment fait re´fe´rence aux 3R en tant que concept. Les conflits entre la re´duction du nombre d’animaux et le risque de creér des re´sultats peu fiables a e´te´ aborde´. Les participants ont e´galement critique´ la pratique consistant a` utiliser davantage d’animaux pour ame´liorer les donneés statistiques et obtenir la publication des re´sultats dans des revues tre`s coteés - un fait qui nous semble nouveau. La principale conclusion de cette e´tude est que les valeurs e´thiques ne pouvaient pas eˆtre facilement se´pareés de l’objectif de produire une bonne science. Alors que les responsables politiques semblent s’attendre a` ce que les chercheurs tiennent explicitement compte des conside´rations e´thiques, nous avons constate´ que leur re´flexion e´thique se manifestait principalement de manie`re implicite dans le cadre de la me´thodologie et de la conception. Nous ne voyons pas cela comme une proble`me tant que les valeurs fondamentales sous-jacentes sont implicitement respecteés, ou promues, dans la pratique expe´rimentale concerneé.

Abstract Ethische Richtlinien fu¨r die Tierforschung wie das 3R-Prinzip und positive Schaden-Nutzen-Bewertungen sind in der EU-Richtlinie 2010/63 verankert. In der vorliegenden qualitativen Studie haben wir untersucht, wie ethische Richtlinien bei der Planung von Tierversuchen mit anderen U¨berlegungen zusammenwirken und/ oder konkurrieren. Die Untersuchungen wurden mit vier Fokusgruppen, die hauptsa¨chlich aus mit der Tierverwendung befassten Forschern bestanden, in vier nordeuropa¨ischen La¨ndern durchgefu¨hrt und die Ergebnisse mit Unterstu¨tzung von NVIVO thematisch analysiert. Praktische Fragen und die Bedeutung guter Wissenschaft waren dominierende Themen. Dabei ließen sich praktische Fragen nicht leicht vom Ziel einer guter Wissenschaft trennen. Die Teilnehmer a¨ußerten Bedenken im Einklang mit den Kernwerten der 3R, wobei die 3R jedoch von einer Gruppe ausdru¨cklich als Konzept bezeichnet wurden. Konflikte zwischen Brønstad and Sandøe 191 einer Verringerung der Tierzahlen und dem Risiko, unzuverla¨ssige Ergebnisse zu erzielen, wurden ange- sprochen. Man kritisierte auch die Praxis des Einsatzes von mehr Tieren, um statistische Zahlen zu verbes- sern, damit Ergebnisse in hoch angesehenen Zeitschriften vero¨ffentlicht werden – eine Erkenntnis, die wir fu¨r neu halten. Die wichtigste Schlussfolgerung dieser Studie besteht darin, dass ethische Werte schwer vom Ziel der Realisierung guter Wissenschaft zu trennen sind. Wa¨hrend politische Entscheidungstra¨ger offenbar erwarten, dass Forscher ethische U¨berlegungen explizit beru¨cksichtigen, haben wir festgestellt, dass sich ihr ethisches Denken hauptsa¨chlich als impliziter Teil von Methodologie und Design manifestiert. Wir sehen dies nicht als Problem, solange die zugrunde liegenden Kernwerte in der jeweiligen experimentellen Praxis implizit respektiert bzw. gefo¨rdert werden.

Resumen Las directrices e´ticas para la investigacioń animal como las 3R y las evaluaciones positivas danõ/beneficio estań ancladas en la Directiva de la UE 210/63. En este estudio cualitativo, investigamos co´mo las directrices e´ticas interaccionan y/o compiten con otras consideraciones al planificarse una investigacioń animal. Se llevaron a cabo cuatro grupos de debate compuestos principalmente de investigadores involucrados en el uso de animales en cuatro paı´ses de Europa del Norte y las conclusiones se analizaron tema´ticamente con el soporte de NVIVO. Los temas praćticos y la importancia de hacer una ciencia buena fueron los temas dominantes. Los temas praćticos no podıán separarse faćilmente del objetivo de conseguir una ciencia buena. Los participantes expresaron su preocupacioń que concuerdan con los valores principales de las 3R, pero en un grupo se refirieron explıćitamente a las 3R como un concepto. Se trataron los conflictos entre la reduccioń de los nu´meros de animales y el riesgo de crear resultados poco fiables. Tambień criticaron la praćtica de utilizar ma´s animales para mejorar las cifras estadı´sticas a fin de conseguir que sus resultados sean publicados en revistas de prestigio, un hecho que pensamos que es nuevo. La principal conclusioń de este estudio es que los valores e´ticos no podrıán separarse faćilmente del objetivo de producir buena ciencia. Mientras los creadores de polı´ticas parecen esperar que los investigadores tengan explıćitamente en cuenta consideraciones e´ticas, encontramos que su pensamiento e´tico se manifiesta principalmente como una parte implıćita de una metodologıá y disenõ. No pensamos que esto suponga un problema siempre que los valores principales subyacentes se respeten implıćitamente, o se fomenten, en la praćtica experimental relevante. Explore the Laboratory Animals Handbooks

NEW EDITION!

The Design of Animal Experiments Health and Safety in Laboratory Parasites of Laboratory Animals nd (2 Edition) Animal Facilities Dawn G Owen Reducing the Use of Animals Edited by in Research through Better Margery Wood and Experimental Design Maurice W. Smith Michael Festing

Find out more and buy online at uk.sagepub.com/lahandbooks

Laboratory Animals

The international journal of laboratory animal science, medicine, technology and welfare, Laboratory Animals publishes peer-reviewed original papers and reviews on all aspects of the use of animals in biomedical research. The journal promotes improvements in the welfare or well-being of the animals used, particularly focusing on research that reduces the number of animals used or which replaces animal models with in vitro alternatives.

Read the latest content at journals.sagepub.com/home/lan

journals.sagepub.com/home/lan

N8J2049 LAN Flyer Update.indd 1 09/01/19 5:14 PM Expert in the world of research diets

Special Diets Services is the largest supplier of Laboratory Animal diets in Europe and the only dedicated manufacturer in the UK. Special Diets Services has a global reputation for the quality of its diets and manufacturing and storage facilities. Special Diets Services

Special Diets Services PO Box 705, Witham, Essex, England CM8 3AD Telephone: +44 (0) 1376 511260 SDS www.sdsdiets.com Fax: +44 (0) 1376 511247 the essential resource for quality research diets Email: [email protected]

SDS8222 SDS Corporate Ad LAJ 280x210mm.indd 1 19/01/2017 10:40 News Laboratory Animals 2020, Vol. 54(2) 194 ! The Author(s) 2020 Article reuse guidelines: 3Rs – Reduce Reuse Recycle sagepub.com/journals-permissions DOI: 10.1177/0023677220909272 journals.sagepub.com/home/lan Jean-Philippe Mocho

or set up policies to reduce environmental impact. Individuals (researchers, technicians, vets) have developed their own tricks. Strategic discussions now involve all stakeholders with sustainability in mind, for example, when designing a new facility and deciding on ven- tilation or disinfection and sterilisation of a barrier. We would like to hear about all these initiatives: on the bench in the labs, in the animal facility, or even in the staff room! FELASA is collecting tips, good ideas, It is often shocking to see how much plastic is sent for and references to peer-reviewed articles or useful news recycling after a single experiment. Imagine the scale of items that could help the LAS community in its effort it for your whole establishment, a country or the to reduce environmental impact. All relevant received Laboratory Animal Science (LAS) community. Some information will be shared on the FELASA website universities are now taking action to significantly http://www.felasa.eu/, anonymously, for all to access reduce single-use plastics. Labs across the globe rou- easily. Please send your input to me (details below) tinely set their À80C freezer to warmer temperatures and to Penny, the FELASA secretariat, at felasa@ (À60CtoÀ70C) and save over one-third of energy hushmail.com. consumption. Anaesthetists have stopped using gas like Another challenge for you: are you ready to cycle to desflurane or nitrous oxide due to their very high green- the FELASA 2022 congress in Marseille http://www. house effect. felasa2022.eu/? Green initiatives are flourishing in the LAS community. Establishments have joined accreditation schemes

FELASA Honorary Secretary, FELASA, Eye, UK

Corresponding author: J-P Mocho, FELASA, PO Box 372, Eye, IP22 9BR, UK. Email: [email protected]

Contributions to the News section are not subject to peer review and reflect the opinion of your subscribed society.

News Laboratory Animals 2020, Vol. 54(2) 196–197 ! The Author(s) 2020 El XV Congreso de la SECAL (Sociedad Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0023677220906766 Espan˜ola para las Ciencias del Animal de journals.sagepub.com/home/lan Laboratorio) tuvo lugar en Sevilla del 6 al 8 de noviembre

Elena Hevia1 and C. Oscar Pintado2

El evento fue todo un e´xito con ma´s de 300 participantes Sevilla fue un marco incomparable para este encuen- venidos de todo el paı´s y de algunos paı´ses latinoamer- tro. El tiempo agradable invitaba a pasear por sus calles icanos y europeos. 27 ponentes de reconocido prestigio tras las sesiones o a ir a pasear, correr o montar en bici- de varias nacionalidades participaron en las 9 sesiones cleta en los alrededores del rıó. El programa social incluyo´ cientı´ficas que cubrıán un amplio rango de aspectos rela- unavisitaguiadaporlaciudadasıćomo una cena de gala cionados con el animal de laboratorio. (Las sesiones que donde todos pudimos disfrutar de las vistas, la comida y se impartieron en ingle´s contaron con traducciońsimul- del baile que tuvo lugar al terminar la misma. tańea gracias a la ayuda concedida por LAL). En encuentro terminoćon la presentacioń de la sede Las sesiones del congreso versaron sobre: del pro´ximo congreso que nos reunira´de nuevo a todos en la ciudad de Lleida en 2021. Nuevas tecnologıás y su impacto en los animales de laboratorio Motivacioń del personal y resolucioń de conflictos XV Congress of SECAL (Spanish Society for Gestioń de centros de produccioń animal Laboratory Animal Science) took place in Avances en las 3Rs the beautiful city of Seville on 6–8 Desafıós de la investigacioń en ambientes no November convencionales. Bases e´ticas de la proteccioń y experimentacioń animal The event was a great success, with more than 300 par- Reproducibilidad y traslacioń ticipants from all over the country, as well as from Transparencia en la investigacioń animal Latin American and European countries. Twenty- Cuidados veterinarios en procedimientos severos. seven renowned speakers from various nationalities participated in the nine scientific sessions that covered Durante las sesiones el nivel de participacioń activa fue a wide range of laboratory animal science topics. extraordinario lo que determino´que el congreso fuera (Sessions were delivered in English, with simultaneous especialmente enriquecedor. translation available thanks to a LAL grant). Para la mesa redonda, cinco representantes de las The topics included were: autoridades competentes discutieron con los participantes la aplicacioń de la legislacioń en animales de . New technologies related to laboratory animals experimentacioń y en la formacioń del personal que . Staff motivation and conflict resolution trabaja con animales. Se presentaron ma´s 100 posters de alta calidad lo que hizo difıćil la seleccioń por parte del ComiteĆientı´fico de los dos trabajos ganadores. El 1Centro de Biologıá Molecular Severo Ochoa, Madrid, Spain 2 programa cientı´fico se completoćon 4 talleres que se Animal Facility and Transgenic Unit. CITIUS III, University of Seville, Seville, Spain desarrollaron el dıá 5 de noviembre de los que se tuvo unos espleńdidos comentarios. Corresponding author: La exposicioń comercial tambień tuvo gran e´xito y Elena Hevia, Centro de Biologıá Molecular Severo Ochoa, permitioá 24 exhibidores con ma´sde30standsmostrar c/ Nicola´s Cabrera, 1 Campus de la Universidad Autońoma a todos los participantes sus u´ltimas innovaciones en sus Cantoblanco, Madrid 28049, Spain. Email: [email protected] productos o servicios. Adema´s tuvo lugar por primera vez en la SECAL una sesioń comercial para la presenta- Contributions to the News section are not subject to peer review cioń corta de las empresas presentes en el Congreso. and reflect the opinion of your subscribed society. Hevia and Pintado 197

. Efficient management of laboratory animal facilities which took place on 5 November, completed the scien- . Advances in the 3Rs tific programme from which splendid feedback was . Challenges in non-conventional research received. environments The commercial exhibition was also very popular . Philosophical bases of animal protection and and allowed 24 exhibitors with more than 30 stands research to show participants the latest innovations in their . Reproducibility and translation of experiments products or services. Furthermore, a commercial ses- . Transparency of animal research sion took place for the first time in a SECAL . Veterinary care in severe experiments Conference, allowing for flash presentations of the companies that attended the meeting. During the sessions, the level of attendance and Seville was an incomparable setting for this meeting. active participation was extraordinary, making the The lovely weather invited us to walk through the meeting especially enriching. streets at the end of the sessions or to go jogging, walk- For the round table, five representatives of the ing or cycling on the riverbanks. The social programme Competent Authorities discussed with the attendees included a night tour and a gala dinner in which we all the application of legislation in animal experimentation enjoyed the views, the food and the dancing that took and in the training of personnel working with animals. place next. More than 100 high-quality posters were presented, The meeting ended with the presentation of the next making the selection of the two winning works hard Congress that will bring us all together again in the city for the Scientific Committee. Four practical workshops, of Lleida in 2021. Thanks to Reviewers Laboratory Animals 2020, Vol. 54(2) 198–199 ! The Author(s) 2020 Thanks to Reviewers Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0023677220903579 journals.sagepub.com/home/lan

The Editor-in-Chief, Editorial Board and Publisher would like to thank the following reviewers for their contribution to Laboratory Animals in 2019.

Abelson, Klas Czaplik, Michael Hansen, Axel Addo, Phyllis Da Cunha, Daise Hansen, Kristine Adedeji, Temitope Dahlborn, Kristina Hatch, Graham Adjei, Samuel Davies, Ben Hawkins, Penny Aﬁfy, Mamdouh Decrock, Frederic Hecht, Gil Afonson, Ricardo Dennis, Mike Hedrich, Hans Aldrich, Greg Deziel, Robert Hendriksen, Coenraad Allen, Kyle Dhondt, Kevin Hermann, Brian Allen, Sue Diederich, Kai Hiebel, Bernhard Anderson, David Dirnagl, Ulrich Hilken, Gero Antunes, Luis Doe, Brendan Hinkel, Rabea Appelgren, Lars-Erik Dontas, Ismene Holterman, Chet Arras, Margarete Dorsch, Martina Hood, derek Baca, Justin Doube, Michael Horst, Klemens Backhans, annette Drynan, Lesley Hultgren, Jan Bainbridge, David Ekstrand, Carl Idahor, Kingsley Barnard, alun Essex-Lopresti, Angela Illgen-Wilcke, Brunhilde Beckmann, Nicolau Estanislau, Celio Jarvis, Gavin Beekhuijzen, Marion Exner, Conny Jaubert, Jean Behr, Ru¨diger Figueiredo, Isabel Jirkof, Paulin Benavides, Fernando Findlay, amy Jolivet, Genevie` ve Benga, Laurentiu Finnemore, Paul Kalliokoski, Otto Bergadano, Alessandra Flecknell, Paul Kane, Alice Bernard, Rene´ Fonio, Ehud Katsimpoulas, Michalis Bertrand, Henri France, Malcolm Kimmina, Sarah Bessems, Jos Fray, Martin King, Karen Bindelle, Jerome Fruijtier-Po¨lloth, Claudia klarenbeek, Sjoerd Bleich, Andre´ Galatos, Apostolos Klinge, Uwe Bleilevens, Christian Gardiner, Mark Knotek, Zdenek Bourge` s-Abella, Nathalie Garrels, Wiebke Kolbe, Thomas Bradbrook, Carl Gaskill, Brianna Kra¨mer, Stephanie Brossard, Ludovic Getchell, Rodman Kuiken, Thijs Brown, Mark Giamberardino, Maria Langermans, Jan Buckmaster, Cindy Giavedoni, Luis Larsson, Anders Buettner, Manuela Gilbert, Colin Leenaars, Cathalijn Busardo, F Gjendal, Karen Li, zhen Bussell, James Glowka, Tim Lines, Kate Chen, Chih-Cheng Gosselin, Romain-Daniel Lofgren, Jennie Chen, Jichun Gregori, Michela Lutz, Thomas Cho, Jong Ki Griﬃn, Gilly Magara, Fulvio Chourbaji, Sabine Grujic-Milanovic, Jelica Mahabir, Esther Clutton, Eddie Habedank, Anne Manell, Elin Thanks to Reviewers 199

Marinelli, Sara Reichel, Rudolf Swindle, Michael Martinez-Salgado, Carlos Reinhardt, Dieter Sztein, Jorge Masano, Yuuki Richardson, Claire Talbot, Steven Mason, Georgia Riebel, Katharina Tarigan, Bernadetta McAllister, Stacy Riederer, Beat Tho¨ne-Reineke, Christa McCLure, Fiona Robb, David Vergara, Patri Merkus, Daphne Rocio, Rocio Verreck, F Miller, Amy Rodrigues dos Santos, Nuno Viebahn, Christoph Mogil, Jeffrey Ru¨licke, Thomas Visser, Jenny Moreno-del Val, Gonzalo Ryden, Anneli Voelkl, Bernhard Morgan, Michael Saeb-Parsy, Kourosh Vo˜ ikar, Vootele Morton, David Sahine, Erhan Vollert, Jan Muro, Andres Fernando Salvatori, Daniela Wagner, Heike Neumann, Detlef Sanchez-Morgado, Jose´ Wallgren, Per Neunaber, Claudia Sanders, Justin Walski, Tomasz Niimi, Manabu Schenkel, Johannes Wang, Dongfang Olsen, Lena Schiering, Ina Warshaw Funk, Amy Pacharinsak, Cholawat Schofield, Paul Wedekind, Dirk Parker-Thornburg, Jan Schwiening, Christof Wells, Sara Parks, Simon Sena, Emily S Windschnurer, Ines Pekow, Cynthia Sheeley, Heather Wolterbeek, Andre Pierzynowska, Kateryna Slattery, David Wyatt, Jeff Pintado, Belen Smith Richards, Brenda Yoshiyuki, Rieko Ploegaert, Tosca Smith, Adrian Young, Simon Ploj, Karolina Smith, Jennifer Zacharioudaki, Argyro Prins, Jan-Bas Sorzano, Carlos Oscar Zechner, Dietmar Pritchett-Corning, Kathleen Spergser, Joachim Zintzsch, Anne Pryce, Christopher Steinborn, Ralf Raschzok, Nathanael Strecker, Jan 200...... Calendar of events

In light of concerns over COVID-19, some of the events below may be cancelled or postponed. Please check individual websites for further information. 2020 28 May 3R Symposium-Alternatives to CO2, Bern, Switzerland. For further information visit https://www.nc3rs.org.uk/events/3r-symposium-%E2%80%93-alternatives-co2 22–26 June FELASA Laboratory Animal Science Course on Primates, Gottingen, Germany. For further information visit https://www.nc3rs.org.uk/sites/default/files/documents/ LASCourse_Registration_Juni2020_form.pdf 2–3 July 2nd Lung In Vitro event (LIVe2020), Nice, France. For further information visit https:// www.epithelix.com/support/LIVe2020 5 July The science of animal sentience: refining experimental biology, Prague, Czech Republic. For further information contact [email protected]. 23–27 August 11th World Congress on Alternatives and Animal Use in the Life Sciences, Maastricht, The Netherlands. For further information see http://wc11maastricht.org/ 6–9 September Eurotox 2020, Copenhagan, Denmark. For further information visit http://www. eurotox-congress.com/2020/ 16–18 September ANZLAA, Brisbane, Australia. For further information see http://www.anzlaa.org/ 16–18 September GV-SOLAS, Wo´rzburg, Germany. For further information see http://www.gv-solas. de/ 25–29 October AAALAS National meeting, Charlotte, NC. Further information to follow. 1–3 December AFSTAL Annual Conference, Marseille, France. For further information visit https:// www.colloque-afstal.com/2020/

2021 13–15 April ScandLAS 50th Symposium, Tallin, Estonia. For further information visit https://www. scandlas2020.ee/ 16–19 March IAT Congress, UK. For further information visit www.iat.org.uk

Index to Advertisers APRIL 2020

AALAS 130 Granovit AG / Kliba Nafag 123 Tecniplast SpA IBC Altromin International OBC GVG Diagnostics GmbH 134 AnLab Ltd 195 ZOONLAB 127 AVID plc 129 Laboratory Animals Ltd (LAL) 192, 195 LBS 131 Charles River Laboratories IFC North Kent Plastics 132 Datesand Ltd 128 Special Diets Services 193 Fine Science Tools GmbH 124 ssniff Spezialdia¨ten GmbH 133

Your global operating experts in lab animal diets.

$VDQΖ62FHUWLȴHGFRPSDQ\ ZLWKRYHU\HDUVRIH[SHULHQFH ZHR΍HU\RX VWDQGDUGL]HGDQLPDOQXWULWLRQDOVROXWLRQV DQGFXVWRPL]HGVSHFLDOGLHWV

altromin Spezialfutter GmbH & Co. KG Im Seelenkamp 20 D- 32791 Lage Made in Germany www.altromin.com International