Preferences of mice and rats for types of bedding material - USP

Preferences of mice and rats for types ofbedding materialH. J. M. Blom1, G. VanTintelen 3 , C. J. A. H. V. Van Vorstenbosch 2 ,V. Baumans' & A. C. Beynen'lDepartment of Laboratory Animal Science and 2Department of Functional Morphology, Faculty ofVeterinary Medicine, Utrecht University and 3Laboratory Animals Centre, Agricultural University,Wageningen, The NetherlandsSummaryThe type of bedding material has been reported to affect the environment in both the animalcage and animal room. It has an impact on the health and well-being of the animals and maycause biased experimental results. Requirements for bedding materials, particularly thoseregarding animal comfort are poorly supported by experimental data. In this study, varioustypes of bedding material were evaluated using preference tests with mice and rats. It wasfound that beddings consisting of relatively small particles I,::; 1.2 x 1.6mm 2 ) were generallyavoided, whereas beddings consisting of large fibrous particles were preferred. Thecharacteristics of the bedding materials were further investigated by scanning the size andshape of the particles, and by the assessment of ultrasound produced by the moving of thebeddings. The results seem to indicate that size and manipulability are among the maindeterminants of the appreciation of bedding particles by laboratory mice and rats, and largerparticles are preferred.Keywords Bedding; environment; preference; behaviour; rat; mouseRodent beddings serve to absorb moisturefrom excrements and provide nesting material.The type of bedding material affects notonly the microenvironment of the animalcage (Hirsjiirvi &. Viiliaho 1985, Raynor et al.1983), but also the environment of the animalroom (Sakaguchi et al. 1989). Furthermore,the type of bedding may interact withexperimental treatments and affect the outcomeof certain experiments such as those onenzyme-inducing, cytotoxic and carcinogeniccompounds and anaesthetics (Torronen et al.1989, Potgieter &. Wilke 1992). To enhancecomparability and reproducibility of resultsbetween experiments, the type of beddingCorrespondence to: H. f. M. Blom, Department of LaboratoryAnimal Science, Faculty of Veterinary Medicine, UtrechtUniversity, P.O. Box 80.166, 3508 TD Utrecht,The Netherlands; Tel: 31-30-532033, Fax: 31-30-537997Accepted 2 Novem ber 1995should be standardized (Beynen 1991). Thequestion then arises on how to define andselect proper rodent bedding.In order to sustain the health and wellbeingof laboratory animals and animal caretakers,and to minimize variability in experimentalresults, requirements for beddingmaterials have been listed (Kraft 1980, Wirth1983, Weichbrod et al. 1986). These requirementsinclude accommodating animal comfort,such as not being harmful, beingnestable and allowing natural behaviour.However, it is not known which types ofbedding meet these criteria.The suitability of bedding for nesting bymice has been studied. Mulder (1975)reported that pregnant mice invariablyselected aspen bedding from 10 differentcommercially available bedding materials.Odynets et al. (1991) using C57BL/6, BALB/cLaboratoryAnimals (1996) 30, 234-244

Bedding materials and rodent preferences235and wild mice also found that for nestbuilding aspen bedding was the favourite offive different types of bedding material. Thesestudies indicate that the origin of beddingmaterial is an important determinant in theappraisal by the mice, but other propertiessuch as particle size or dust content may beimportant as well. When offered a choice ofvarious possibilities, rodents may be able toselect the most suitable type of bedding orcage floor covering. On this assumptionpreference experiments were conducted withmice and rats to investigate their appraisal ofdifferent types of cage floor covering. In anattempt to elucidate the basis for theobserved choice behaviours of the animals,selected properties of the beddings used werestudied.Materials and methodsPreferencetest systemsThe test systems used for mice lBlom et a1.1992)and rats (Blom et a1. 1995)have beenvalidated and described in detail. Briefly, weused a multiple-choice housing system consistingof a central cage surrounded by eithertwo or four test cages. The wire-toppedcentral cage was made of transparent perspexand was lacking food and water. Macrolontype I (UNO Roestvaststaal, Zevenaar, TheNetherlands) and type 375 (RUCO Metaalindustrie,Valkenswaard, The Netherlandslwere used as test cages for mice and rats,respectively. The test cages were suppliedwith a known amount of food pellets (RMH-B®,Hope Farms, Woerden, The Netherlands)in the food hopper of the wire top and had adrinking bottle with a known amount of tapwater. The type of floor covering differedbetween the test cages and between experiments.The central cage and test cages wereconnected by passages.Each choice test started with the introductionof a single animal in the central cage of atest system with thoroughly cleaned cagesand passages. During the test period of 48 hthe animal could move freely from one cageto another. In the test system for mice, allpassages between the cages were recordedphoto-electrically. In selected tests, thebehaviour of the C57BL/6J1comouse wasrecorded on video. In the rat test system,passages were recorded by means of electronicbalances which were placed under thetest cages. Continuous turning of the preferencetest system during the testing of eachmouse or turning its position between thetesting of each rat was performed to cancelout any bias of choice behaviour related tothe position of the preference test systemwithin the experimental room. At the end ofeach test period, animal preference wasassessed by calculating (relative) dwellingtimes per cage and occasionally analysinganimal behaviour per cage.Preference testsFour experiments were conducted with miceand one with rats. Table 1 summarizes thetypes of floor covering tested. In the firstseries with mice, the choice behaviour wasstudied in a test system with four test cages.The floor of the test cages was covered eitherwith the wire mesh, aspen wood chips,shredded filter paper or Hahnflock® H9sawdust (Fig. I, panel A). The perspex floor ofthe central cage had no covering throughout.In the experiments with rats there werefour test cages with different types of bedding:Hahnflock® H9, Lignocel® 58/15, Lignocel@3/4 or Hahnflock® Tierwohl. All these beddingswere manufactured of fir/spruce, butdiffered in particle size (Fig. I, panel B).Thefloor of the central cage was covered with awire mesh (stainless steel wire: rod diameter,3.3 mm; mesh size, 40 x 40 mm 2 ).In three more experiments with mice, atest system with two test cages was used. Thetest cages were fitted with wire mesh versuswood chips; wire mesh versus Hahnflock®H9 sawdust; and wood chips versus Hahnflock®H9 sawdust. The floor of the centralcage was again left without covering.Animals and husbandryFor the first experiment with mice, 20 femaleC57BL/6JIcoand 20 female BALB/cBYJlcoanimals, aged 6 to 24 weeks, were used. Nineto 16of these mice of each strain were used inthe three experiments conducted later. Beforeand between the experiments, the mice were

236 Blometal.Table 1 Types of floor covering material used in the test cages during the preference tests with mice andrats. (Fig. 1 shows the materials listed)FloorcoveringType ofmaterial Trade name particles Size SupplierWire mesh not applicable not applicable rod diameter, 2 mm UNO Roestvaststaal,mesh size, 10 x 10mm 2 Zevenaar, The NetherlandsWood chips Finn Tapvei aspen chips 4x4x 1mm 3 Tapvei OY, Kaavi, FinlandFilter paper Whatmann shreds approx. 2 x 5 mm 2 Boom, Meppel, ThechromatographicNetherlands (shreds werepaper typeproduced by cutting sheets3MM Chr.of paper)Sawdust Hahnflock® H9 granulate 0.9 x 1.3mm 2 Hahn & Co, Bredenbek-Kronsburg, GermanySawdust Lignocel® 58/15 granulate 1.0 x 2.5 mm 2 Rettenmaier & 56hne,Ellwangen-Holzmuhle,GermanyWood Lignocel® 3/4 fibres 2.0 x7.0mm 2 Rettenmaier & 56hne,shavingsEllwangen-Holzmuhle,GermanyWood Hahnflock® fibres 2.0 x4.5mm 2 Hahn & Co, BredenbekshavingsTierwohl Kronsburg, Germanyhoused in groups of five animals in a housingsystem consisting of two wire-toppedMacrolon type I cages connected by a passage.This housing system allowed the animals tobecome accustomed to moving through thepassages of the test system. Both cages ofeach system were supplied with food pellets,water and sawdust bedding ILignocel® S8/15).Stock and experimental rooms (conventional)had a controlled photoperiod (stock andexperimental room: white light, approx.200 Ix at 1m above the floor from 07:00-19:00h, dark in stock room, and 700nm, redlight, approx. 5lx at 1m above the floor in theexperimental room from 19:00-07:00 h),temperature (20-22°C), relative humidity(50-60%) and ventilation (15 air changes perhour).In the experiment with rats, 12 female and12 male Wistar (Cpb:WU) animals, bred at theLaboratory Animals Centre, Wageningen,were used. The rats were tested at the ageof 12 to 31 weeks; each rat was tested once.The animals had been born and housed inMacrolon type 375 cages with sawdustbedding (Lignocel® S8/15). until weaning.From weaning to the beginning of theexperiment, the animals were housed in,.. ",.. ..• 1," .J' t• . i:' ~•• \.' ,1'0;".:..... :, "~'l".~. 1./ "".'> :' . ~ . _. " . ..'.~'.;.,.,~. q1 _~ ".• ...... : ..•..•... ',' . :.'P..... . ~' --' .;:::•": ~ t" , _ABFig 1 Types of cage floor covering used in the preference tests with mice (panel A (from left to right):wire mesh, FinnTapvei 4 x 1 xl aspen wood chips, shredded filter paper, Hahnflock® H9 sawdust) and rats(panel B (from left to right): Hahnflock® H9, Lignocel® 58/15,Lignocel® 3/4, Hahnflock®Tierwohl). Table 1documents further characteristics of the floor coverings

Bedding materials and rodent preferences237groups of three in metal cages (type KS/4,RUeO Metaalindustrie) with wire meshfloors (stainless steel wire: rod diameter,2mm; mesh size, 10 x 50mm 2 ) and sawdustbedding (Lignocel® S8/15) at 70mm underneath.The rats had free access to foodpellets (RMH-B®) and tap water. Immediatelybefore testing, each rat went through apre-test (habituation) period of 48 h in thetest system with uniform cages. Each testcage was fitted with a wire mesh floor(stainless steel wire: rod diameter, 2 mm;mesh size, 10 x 30 mm 2 ) with sawdust bedding(Lignocel® 58/15) at 25 mm underneath.The animals were kept and tested inconventional rooms with controlled photoperiod(06:00-18:00h: light, approx. 470lx at1m above the floor), temperature (20-22°C),relative humidity [50-70%) and ventilation(18 air changes per hour).Characterization of bedding materialsThe bedding materials were further characterizedin an attempt to explain the observedchoice behaviours. Actual particle size andshape, as well as dust content, was assessedwith the use of a scanning electron microscope(Stereoscan, Cambridge Inc,Cambridge, United Kingdom). The beddingmaterials were also studied for ultrasoundproduction when moved. For this purpose,50 ml of bedding material was put in a250 ml glass beaker and stirred with a glassrod (approx. 60 rotations min -1). Ultrasoundproduction was detected at 25 cm above thebedding material with a sensitive condensermicrophone (Type 4135, Briiel &. Kjaer,Nieuwegein, The Netherlands), preamplifier(Type 2633, Briiel &. Kjaer),measuring amplifier (Type 2610, Briiel &.Kjaer), and programmable dual filter (TypePDF3700 B, Difa Benelux BV, Breda, TheNetherlands).StatisticalproceduresFor statistical analyses, dwelling times (sec)per cage for the 48-h pre-test or test periodwere transformed into their logarithms. APearson correlation matrix for dwelling timesof different cages was constructed to obtaininformation on competition between cages.Differences between mouse strains or sex ofthe rats were evaluated using univariaterepeated measures analysis. Differencesbetween floor covering and interactions offloor covering and either mouse strains or sexof rats or rat preferences during the pre-testperiod was evaluated using multivariaterepeated measures analysis (Wilks' Lambdatest). The level of significance was pre-set atP

238 810m eta/.80~

Bedding materials and rodent preferences 239Table 2 Relative deposition of faeces and urine and relative consumption of food and water by femaleC57BL/6J1coand BALB/cBYJlco mice on four different types of floor covering in a preference testC57BL/6JlcoType of floor coveringshredded filter sawdustwire mesh wood chips paper (Hahnflock® H9)Faeces 19.6± 1.9 27.6± 1.9 16.1 ±2.1 35.0±2.1(% of tota I bol i)Urine 1.6±0.1 1.7±0.1 1.1 ±0.1 2.0±0.0(mean urine score*)Food 22.2±2.4 23.0± 1.9 26.6±3.0 28.2±2.8(% of tota Ig consumed)Water 23.1 ±2.0 23.8± 1.2 27.6± 1.4 25.6±1.4(% oftotal ml consumed)BALB/cBYJlcoType of floor coveringshreddedsawdustwire mesh wood chips filter paper (Hahnflock® H9)Faeces 28.6±2.2 27.8±2.3 10.3±0.9 31.3±1.8(% oftotal boli)Urine 1.7±0.1 1.9 ± 0.1 1.1 ±0.1 2.0±0.0(mean urine score*)Food 26.8±2.8 21.6±1.9 22.1 ±2.4 29.5±2.5(% of total g consumed)Water 22.2 ± 1.7 26.0±3.0 25.7±1.9 26.1 ±2.0(%oftotal ml consumed)Results are expressed as means ±SEM for 20 mice per strain*Urine was scored as: 1dry; 2 moist; 3 wet; 4 soakingTo assess ultrasound production of thebedding materials when moved by theanimals, the frequency and relative intensityof ultra sounds produced after stirring with aglass rod were measured. Ultrasound productionwas strongest with Hahnflock® Tierwahlwood shavings, whereas Hahnflock® H9sawdust produced the fewest ultrasounds(data not shownl. In general, ultrasoundintensity increased with increasing particlesize of the beddings tested. The highest levelsof the ultrasounds were found in the lowerfrequency ranges (20-40 kHz).DiscussionFigure 2 illustrates that shredded filter paperwas so attractive to female laboratory micethat it masked differential preferences forwood chips, Hahnflock® H9 sawdust and awire mesh floor. In the test system with twotest cages, wood chips were preferred oversawdust and wire mesh. The preferencepatterns of the two inbred strains used weregenerally similar. Apart from relative dwellingtimes, the mice also discriminatedbetween the types of floor covering withrespect to defaecation and urination. Themice tended to deposit less faeces and urinein the cage with shredded filter paper, i.e. thecage where they spent most time on resting.This suggests that to a certain extent miceactively keep their nesting area clean fromexcreta. Under practical conditions, multiplecage compartments usually cannot be offered.When shredded paper is given as the salebedding material it may become wet quickly.Since the absorbing potency of sawdust beddingexceeds that of shredded filter paper, a

240 Blom etal.60Q)E 4001CQ)~uQ) 20.:::iii~oH9D females88/15 3/4 Tierwohltype of beddingmaterialFig4 Distribution of relative dwelling times(means± SEM; n=12 per sex) during the test periodof 48 h for male and female Cpb:WU rats in a preferencetest system with the central cage and fourtest cages fitted with different types of beddingmaterial. Repeated measures analysis: P'ex=0.004;Pbedding=0.01O;Pbedding x 'ex=0.173. Dwell ing times forHahnflock® H9 sawdust differed significantly(P=0.0015 to 0.0025, paired t-test) from those for theother three test cagescombination of both bedding materials maybe favourable when animals are housed in asingle cage.In the preference tests with rats, data onfaeces and urine production were not collected.It was however apparent that at theend of the test periods more faecal boli werepresent in the central cage with a wire meshfloor than in the test cages with various typesof bedding. Possibly, rats also prefer a cleanliving area.Both mice and rats appeared to avoidbedding material with relatively small particles(Hahnflock® H9 sawdust I and preferredbeddings consisting of large particles (aspenwood chipsj or fibres (Lignocel® 3/4 woodshavings, Hahnflock® Tierwohl wood shavings,shredded filter paper). Bedding materialsconsisting of relatively small particles areunsuitable for nest building (Hamalainel1 &Tirkkonen 1991). The observed preferencesfor bedding materials may relate to theirsuitability as nesting material. Analysis ofmouse behaviour revealed that digging, i.e.manipulation of bedding, and resting in a nestwere performed most on the preferred bedding.The preference of the mice for shreddedfilter paper is in accordance with the results_maleswire mesh(cenlral cage)of Nolen and Alexander (1966). In a limitednumber of additional choice tests with fourtest cages we offered mice a cage with a sheetof filter paper fixed to the floor instead ofshredded filter paper. The filter paper wasnow only preferred when the animals managedto remove and tear the sheet (Fig. n Ifthey did not succeed in doing so, the cagewith filter paper floor covering was avoided.Apparently, the structure of filter paper thathas been shredded may determine preferenceof the mice.Small bedding particles and especially dust(particles smaller than 300/lm, Wirth 1983)can be irritating or damaging when inhaled(Sakaguchi et al. 1989, Thigpen et al. 1989).Such particles also irritate the vaginalmucosa (Plank & Irwin 1966) or causepreputial infections IVan Herck,' personalcommunication). In this light it is plausiblethat Hahnflock® H9 sawdust is avoided whencompared with the other bedding materialstested. Nevertheless, the female rats tendedto prefer Hahnflock® Tierwohl wood shavingsas bedding material, despite its highproportion of dust. .Rodents are very sensitive to ultrasoundlRalls 1967, Sales & Milligan 1992). Weassumed that moving of bedding materialsby rodents may produce ultrasonic soundsthat can either be irritating or pleasurable.The bedding materials tested were found toproduce ultrasounds upon stirring, thehighest intensity being caused by Hahnflock®Tierwohl wood shavings. This characteristicof Hahnflock® Tierwohl couldhave contributed to the relatively highdegree of appraisal by the female rats. Thesame may be true for the preferencetowards aspen wood chips (Finn Tapvei)when compared with Hahnflock® H9 sawdust,as displayed by the mice. Thus,relative preference and/or avoidance of agiven bedding material may be associatedwith ultrasound production.The results of this study indicate that sizeand/or shape might be a major determinant inthe appreciation of bedding particles bylaboratory mice and rats. Beddings consistingof large fibres seem to be preferred. It shouldhowever be emphasized that the present datacannot be interpreted unequivocally to show

Bedding materials and rodent preferences 24180 80AB~ ~:>R 60* 600(])

242 Blom eta',Fig 6 Scanning electron micrographs of bedding materials studied in preference tests with mice and rats.The photographs have approximately identical magnification: the bar indicates 1mm

Bedding materials and rodent preferences243Fig 7 Illustration of a test cage with a sheet offilter paper fixed to the floor as used in a preferencetest system for mice consisting of the centralcage and four test cages with different types ofcage flooring (d.Table 1).Thetest cage with the filterpaper was either avoided (left) or preferred (right)depending on the animal being successful in creatinga nest by tearing the sheet of papermay contribute to the adequate adaptation ofexisting guidelines and practices on housinglaboratory animals. Promoting the adaptationof laboratory animal housing systemstowards the animals' needs may be conduciveto animal welfare but does not necessarilysustain scientific interests. The use of preferredbedding material so as to improve theanimals' well-being seems appropriate, but itshould be realized that certain experimentsrequire specific cage floor coverings, such aswire mesh bottoms, that are not compatiblewith the animals' preference. Thus, theintroduction of changes should always beconsidered along with their consequences forthe experiment.Acknowledgments The authors are grateful to BMI,Helmond, distributor of Lignocel® and Hahnflock®beddings in The Netherlands and to Tapvei OY, Kaavi,Finland, for kindly supplying bedding materials, and toDr J. Van den Broek for assisting with the statisticalanalysis.ReferencesBeynen AC (1991) The basis for standardization ofanimal experimentation. Scandinavian Journal ofLaboratory Animal Science 18, 95-9Blom HJM Van Vorstenbosch CJAHV, Baumans V,Hoogervorst MJC, Beynen AC, Van Zuthpen LFM(1992) Description and validation of a preferencetest system to evaluate housing conditions forlaboratory mice. Applied Animal Behaviour Science35,67-82Blom HJM, Van Tintelen G, Baumans V, Van denBroek J, Beynen AC (1995) Development andapplication of a preference test system to evaluatehousing conditions for laboratory rats. AppliedAnimal Behaviour Science 43, 279-90Hiimiiliiinen A, Tirkkonen T (19911 How to choose agood bedding material. Laboratornye Zhyvotnye 1,62-5Hirsjiirvi P, Viiliaho T (1985) Microclimate in twotypes of rat cages. Zeitschrift flir Versuchstierkunde27, 93Mulder JB (1975) Bedding preferences of pregnantlaboratory-reared mice. Behavior Research Methods& Instrumentation 7, 21-2Kraft LM (1980) The manufacture, shipping andreceiving and quality control of rodent beddingmaterials. Laboratory Animal Science 30,366-74Les EP (1980) Comment to Kraft. Laboratory AnimalScience 30, 75Nolen GA, Alexander JC (1966) Effects of diet and typeof nesting material on the reproduction andlactation of the rat. Laboratory Animal Care 16,327-36Odynets A, Simonova 0, Kozhuhov A, et al. (1991)Beddings for laboratory animals: criteria ofbiological evaluation. Laboratornye Zhyvotnye 1,70-6Plank S1, Irwin R (19661 Infertility of guinea pigson sawdust bedding. Laboratory Animal Care16, 9-11Potgieter FJ, Wilke PI (1992) Laboratory animal bedding:a review of wood and wood constituents as apossible source of external variables that couldinfluence experimental results. Animal Technology43,65-88Ralls K (196 7lAuditory sensitivity in mice: Peromyscusand Mus musculus. Animal Behaviour 15, 123-8Raynor TH, Steinhagen WH, Hamm Jr TE (19831Differences in the microenvironment of a polycarbonatecaging system; bedding vs raised wirefloors. Laboratory Animals 17, 85-9Sales GD, Milligan SR (1992) Ultrasound andlaboratory animals. Animal Technology 43,89-98Sakaguchi M, Inouye 5, Miyazawa H, Kamimura H,Kimura M, Yamazaki 5 (19891 Evaluation of dustrespirators for elimination of mouse aeroallergens.Laboratory Animal Science 39, 63-6Thigpen JE, Lebetkin EH, Dawes ML, et al. 11989) Astandard procedure for measuring rodent beddingparticle size and dust content. Laboratory AnimalScience, 39, 60-2

244810m etal.T6rr6nen R, Pelkonen K, Kiirenlampi S 11989) Enzyme-inducingand cytotoxic effects of wood-basedmaterials used as bedding for laboratory animals;comparison by a cell culture study. Life Sciences 45,559-65Weichbord RH, Hall JE, Simmonds RC, Cisar CF119861Selecting bedding material. LaboratoryAnimals 15, 25-9Wirth H \19831 Criteria for the evaluation of laboratoryanimal bedding. Laboratory Animals 17, 81-4