- Email: [email protected]
PBI creams
Physiology & Behavior 74 (2001) 621 – 628
PBI creams A spontaneously mutated mouse strain showing wild animal-type reactivity Colin A. Hendriea,*, Katja S. Van Drielb, Janet C. Tallingb, Ian R. Inglisb a b
School of Psychology, University of Leeds, Leeds LS2 9JT, UK Central Science Laboratory, Sand Hutton, York YO41 1LZ, UK
Received 12 February 2001; received in revised form 15 February 2001; accepted 20 August 2001
Abstract PBI creams are mice derived from warfarin-resistant wild stock that has been maintained under laboratory conditions since the 1970s. This study compares their behaviour to that of laboratory mice and wild house and wood mice. Animals were tested in a black/white box and a 2.64 1.4 m runway. In the black/white box, the behaviour of PBI creams was not significantly different from that of house mice and differed most from that of laboratory mice. Notably, the PBI creams showed the greatest activity and escape-orientated behaviours. When animals were approached by the experimenter in the open runway test, the PBI creams had higher flight speeds than both house and wood mice, whilst laboratory mice failed to respond. In the closed runway test where the animals could not escape, the PBI creams, house mice and wood mice all turned and attempted to run past the approaching experimenter, whilst the laboratory mice again failed to react. At the end of this test session, the time taken to catch each animal was recorded. It took less than 5 s to catch laboratory mice but significantly longer to catch the wild strains and the PBI creams (90 – 100 s for the latter). In these tests, the PBI creams showed wild animal-type reactivity, and as this behaviour has been retained in the laboratory colony for over 30 years, these animals may be useful in the study of the physiological and genetic basis of fear/anxiety in mice. D 2001 Elsevier Science Inc. All rights reserved. Keywords: PBI creams; Wood mice; House mice; Fear; Anxiety; Genetic mutation
1. Introduction In recent times much attention has been given to producing mutagenesis in mice with the aim of modelling human genetically based disorders in the laboratory [2,6,19,24]. Mutation is, of course, the basis for evolutionary development [10], but it is equally clear that in the wild selection pressures operate at the level of the phenotype. Behaviour must therefore be considered the most important factor influencing the survival of genetic mutations from generation to generation. This point is particularly applicable to mice. In the wild, mice are aggressively territorial [3,8] and live in demes, or small breeding groups usually consisting of a single male and several females. Whilst these structures may be temporarily broken down by an over* Corresponding author. Tel.: +44-113-233-5736; fax: +44-113-2335749. E-mail address: [email protected] (C.A. Hendrie).
abundance of food, such as those found in grain stores [9], these changes in social behaviour are transient. Evidence of such territoriality can also be seen in laboratory mice where food supplies are permanently available in excess [13]. Therefore, where food supplies are stable, territoriality remains the norm. In the wild there is little genetic interchange between these small breeding units [11] to the extent that, even on small islands, introduced mutant alleles fail to spread [1]. Further, territory sizes are small, ranging between 50 and 90 f t 2 [21,26], meaning that distinct subpopulations can coexist within geographical locations as small as a single barn [20]. Data, such as these, graphically demonstrate that the social structure of mice serves to isolate these small local populations and so increases the likelihood of allelic variation within them. With such a high potential for variation, it is unsurprising that populations of wild mice occasionally emerge that have unique or exaggerated behavioural characteristics produced in response to altered selection pressures
0031-9384/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 1 ) 0 0 6 0 4 - 7
622
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
within their environment. The spontaneous emergence of such animals has clear parallels with the controlled studies concerned with the induction of genetic mutations in the laboratory and may have direct use in addressing questions relating to functionality. These wild animals have, by definition, survived and reproduced outside the laboratory and hence overcome the demands of the natural environment. This paper describes the behaviour of ‘PBI creams’, mice derived from animals captured during infestation control at the Plant Breeding Institute, Cambridge in the 1970s [23]. Warfarin was used to control the infestation and over time the incidence of dark-eyed, cream-coloured animals, not previously seen in this population, increased. The gene responsible for this cream colour, the extreme chinchilla (che), was found to be linked to warfarin resistance [22]. Hence, in this wild population, the numbers of creamcoloured mice increased in response to the exposure to warfarin. The present colony, held at the Central Science Laboratory since 1971, was derived from this stock: as far as is known, such animals no longer exist in the wild. The most striking behavioural feature of the PBI creams is their extreme reactivity, and the current studies were carried out in order to characterise their behaviour in the context of the behaviour of laboratory mice, wild wood mice and wild house mice. Two models were employed that had previously been used to examine the behaviour of wild animals. The black/white box has been used to test the behaviour of wild-trapped voles [14], and the runway test has been used as part of a test battery to examine the behaviour of wild-trapped rats (e.g. [5]).
2.2. Procedures 2.2.1. Black/white box The black/white box consists of an open-topped box (45 27 27 cm) that is painted two-fifths black and threefifths white. These two sections are divided by an opaque partition with a small opening (7.5 7.5 cm) positioned in the centre of the box at floor level. The floor area of each section was lined to produce 9 cm squares. The black section was illuminated by a 60-W red light set to give light levels of 50 lx, whilst the white section was illuminated by two 60-W white lights, set to give light levels of 500 lx. Lights were carefully positioned so as not to produce areas of shadow. At the start of each session, individual animals were placed into the middle square of the white section. The experimenter then left the room and the behaviour of the animal was recorded on videotape for the next 5 min. At the end of each session, any faecal boli were removed, and the arena was cloth-wiped and dried before the next animal was introduced. The ethogram used for analysing behaviour in this situation was as follows: Latency to enter black section Time in each section Line crossing Rearing Grooming
2. Methods Sniffing substrate 2.1. Animals and housing Immobile All animals were male and approximately 6 months old at the time of testing. Stocks of outbred albino laboratory mice were originally obtained from the Medical Research Council Laboratory Animals Centre, Carshalton, Surrey, UK and had been held for 71 generations prior to testing. The PBI creams had been held for 52 generations, and the house mice (Mus musculus) and wood mice (Apodemus sylvaticus) were four generations removed from the original wild-trapped stock. All animals were housed in groups of three in polycarbonate RB3 cages (45 28 11 cm) under a 12-h light cycle (lights on: 06:00 h) in a temperature (20 ± 2 °C) and humidity (55 ± 15%) controlled environment. The sawdust substrate was replaced once a week, as were small cardboard containers that served both as shelters and sources of nesting material. Food (GRE3K lab chow) and water were available ad libitum. Animals (n’s = 8 –9 per mouse type) were tested between 10:00 and 15:00 h — initially in the black/white box and 14 days later in the runway test (see below).
Jumps Jump out
Other
Time taken for the animal to move from its original starting point in the white section into the black section. Timing in one section began when all four paws were in that section and ended when all four paws left it. Head breaks the plane of a line drawn on floor. Standing on rear legs with straight back and forepaws off the floor. Using forepaws and licking to groom body. Olfactory examination of walls or floor. Cessation of movement with or without intermittent head or body adjustments. All four paws leaving contact with the floor. As jump but with upward movement resulting in animal escaping from test arena. Any other behaviour.
2.2.2. Runway test The runway was a 264 140 70 cm rectangle, constructed of rigid aluminium sheeting with the width of the floor marked off with thin white lines every 33 cm. In the open runway format a 185 20 70 cm divider was placed in the centre of arena 60 cm from the longest side and positioned to leave a 40-cm gap at each end. In the closed runway format, the gap at one end of the apparatus was closed off. The room housing the runway was lit by four 60 W bulbs.
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Prior to testing, animals were individually placed into the runway and left undisturbed for 5 min. At the end of this period, the experimenter entered the middle of the runway and proceeded to move clockwise. Two circuits were made at approximately 0.3 m s 1, two at 0.6 m s 1 and two at 1.2 m s 1, i.e., slow walking pace, fast walking pace and jogging pace. Following this, one end of the divider was closed off to create the closed runway format. For the closed runway test, the experimenter stood at the furthest line, 231 cm from the closed end, and remained motionless for 30 s. The experimenter then slowly moved forward to the next line, which was 33 cm nearer the closed end, and remained motionless again for another 30 s. This process was repeated every 30 s until the experimenter reached the end zone, the last line 33 cm away from the wall of the closed end. At this point, the experimenter slowly bent down and attempted to pick up the animal. An attempt was also made to pick up the animal if it ran towards and past the experimenter prior to this end point being reached. At the end of each session, Longworth live animal traps were placed in each corner of the apparatus, and the experimenter, using a housing box, attempted to catch the animal as quickly as possible. At the end of each session, any faecal boli were removed, and the arena cloth-wiped and dried before the next animal was introduced. Measures taken during this test were as follows: Open runway Flight Time taken for the animal to move from speed the point where it started to run away from the experimenter to the point where it reached the end of the divider and moved into the other section of the runway. Different measures of this were taken for each of the approach speeds produced by the experimenter. Stops Animal briefly ceases running and orientates towards the approaching experimenter before beginning to run again. Pickup Single attempt to bend down and pick up the animal if the distance between experimenter and animal closes to zero. Closed runway Flight Distance at which the animal ceases distance ongoing activity and runs from the end zone towards and past the experimenter. Pickup Single attempt to bend down and pick the animal up when distance between experimenter and animal is reduced to zero as the animal runs towards and past the experimenter or the experimenter reaches the end zone. Catchtime Time taken to capture each animal at the end of the test session.
623
2.3. Statistical analysis Data were analysed using principal component analysis (PCA) and single or two-factor analysis of variance (ANOVA) employing repeated measures where appropriate. Planned pair-wise follow-up tests were conducted using orthogonal contrasts.
3. Results 3.1. Black/white box The latency to enter the black section from the white section revealed a significant effect of mouse type [ F(3,31) = 3.3, P = .033]. Follow-up comparisons found that this effect resulted from the laboratory mice having significantly longer latencies than the other mice. In rank order, the PBI creams had the shortest latency followed by house mice, wood mice and laboratory mice (Fig. 1). However, analysis of the percentage of trial time spent in the white section revealed no significant type differences [ F(3,31) = 0.56, ns]. Some mice jumped out of the equipment and hence had behavioural records less than 5 min. Approximately 66% of PBI creams, 33% of wood mice and 25% of house mice jumped at least once, and 33% of PBI creams and 25% of house mice succeeded in jumping out of the test apparatus before the end of the test session. No laboratory mice jumped upwards or out of the test arena. As a result, there were significant mouse type differences in the total time spent in the box [Fig. 2; F(3,31) = 3.32, P = .032]. Followup comparisons revealed that this result stemmed from a significantly shorter duration for the PBI creams than for the other species. The behavioural data were, therefore, divided by the total time spent in the black and white sections of the box to give standardised rate scores, which
Fig. 1. Latencies to enter the black section of the black/white box. Wood mice (Wood), house mice (House) and PBI cream mice had significantly shorter latencies than laboratory mice (Lab) to enter the black section from their initial starting point in the white section ( * P < .05 compared to Lab mice). See text for further details.
624
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Fig. 2. Total time spent in the black/white box. Although no laboratory mice (Lab) showed this behaviour, both species of wild mouse (House, Wood) and the PBI creams made attempts to jump out of the test apparatus. Approximately 33% of PBI creams and 25% of house mice were successful. In consequence, PBI creams and house mice spent less than the full 300 s in the black/white box. ( * P < .05 from time laboratory mice spent in test apparatus). See text for further details.
were then analysed for differences between the four types of mice. Of the behavioural categories shown in the black section, significant mouse type effects were found only in the rates of rearing [ F(3,31) = 5.4, P = .005] and sniffing [ F(3,31) = 8.0, P = .001]. Follow-up comparisons revealed that these effects were solely due to the laboratory mice showing significantly less rearing and more sniffing in this section than the other types. Significant differences between the mouse types were also found in the frequencies of rearing [ F(3,31) = 12.5, P = .000] and sniffing [ F(3,31) = 5.2, P = .005] in the white section of the box. Follow-up comparisons showed that the differences in sniffing resulted from the PBI creams having significantly lower scores than the laboratory mice, whilst the rearing differences were caused by both the PBI creams and the house mice showing significantly higher frequencies of rearing than the laboratory mice. Significant activity differences between mouse types were found only in the white section of the box [ F(3,31) = 5.6, P = .004]. This resulted from both the PBI creams and the house mice having similar and significantly higher line cross scores than both the wood mice and the laboratory mice. The relative effects of the white and black sections on the different mouse types were compared by subtracting the black section scores for a given behaviour from the corresponding white section scores and then looking for significant type effects in these indices. The four mouse types were significantly differentiated by two of the resulting indices. There was a significant difference between the mouse types in the relative rates of line crosses in the two sections of the box [ F(3,31) = 3.4, P = 033], with the white section eliciting a significantly greater increase in activity in the PBI creams and the house mice (that were not significantly different from each other) than in the wood
mice and laboratory mice (that were not significantly different from each other). The other index showing significant differences between mouse types involved rearing [ F(3,31) = 5.2, P = .006] where both the PBI creams and the house mice (that were not significantly different from each other) had a significantly higher incidence of rearing in the white section than in the black section, whereas both the wood mice and laboratory mice (that were not significantly different from each other) showed a significant difference in the opposite direction. The frequency scores for the behavioural categories in the black and white sections were analysed using PCA. As shown in Table 1, the first component explained the majority of the variance (36% as compared with 13% for the second and third components and less than 10% for each of the remaining components). This component was an activity index, as it was strongly correlated with the frequencies of line crosses, rearing and jumping in both white and black sections and negatively correlated with the frequency of immobility, grooming and sniffing in both sections. An ANOVA on the first component scores revealed highly significant differences between the mouse types [ F(3,31) = 10.7, P = .0001], which resulted from the PBI creams and house mice having similar and significantly greater first component scores from both the wood mice and the laboratory mice and also the wood mice having significantly greater scores than the laboratory mice. 3.2. Open runway test As the laboratory mice failed to react to the approach of the experimenter in the runway by running away, only the
Table 1 Black/white box: component matrix for principal components with eigenvalues greater than 1 Component Variable Rear white Sniff white Groom white Immobile white Jumps white Line crosses white Rear black Sniff black Groom black Immobile black Jumps black Line crosses black Initial eigenvalues total Percent of variance Cumulative variance
1
2
3
.89 .48 .05 .24 .51 .86 .83 .61 .11 .43 .58 .82
.05 .48 .76 .57 .26 .27 .06 .16 .43 .31 .29 .06
.07 .18 .04 .31 .54 .07 .37 .52 .41 .62 .41 .3
4.32 36.1 36.1
1.67 13.8 49.9
1.63 13.6 63.5
The first component, which explains the majority of the variance, is positively correlated with rearing, jumping and line crossing. It is negatively correlated with immobility and sniffing. This component may therefore represent an index of both activity and escape.
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
625
Fig. 3. Flight speed in the open runway. Dark grey bars = flight speed when experimenter approached at 0.3 m s 1; grey bars = flight speed when experimenter approached at 0.6 m s 1; white bars = flight speed when experimenter approached at 1.2 m s 1. Laboratory mice (Lab) failed to match their behaviour to the approach of the experimenter. Hence, their flight speed could not be determined. House mice (House), wood mice (Wood) and PBI cream mice all showed increased flight speed at the fastest approach speed compared to their flight speed at the slowest approach speed ( * P < .05). The flight speed of PBI creams was faster than that of wood mice and house mice at 0.3 and 0.6 m s 1 ( + P < .05 as compared to wood mice and house mice). See text for further details.
data from the two wild types and the PBI creams could be analysed. ANOVA performed on data from these animals revealed a significant effect of both type [ F(2,23) = 8.19, P = .002] and approach speed [ F(2,46) = 12.25, P = .0001; Fig. 3]. Follow-up tests revealed these effects to be due to significantly increased flight speeds when approached at 1.2 m s 1 as compared with being approached at 0.3 m s 1 for house mice [ F(1,23) = 8.9, P = .006], wood mice [ F(1,23) = 12.2, P = .0001] and PBI creams [ F(1,23) = 4.6, P = .04]. The flight speed of PBI creams was significantly greater than house mice at 0.3 m s 1 [ F(1,23) = 6.65, P = .016] and at 0.6 m s 1 [ F(1,23) = 5.86, P = .02] but not at 1.2 m s 1. A similar effect was seen when comparing the flight speeds of PBI creams and wood mice at 0.3 m s 1 [ F(1,23) = 14.0, P = .001] and at 0.6 m s 1 [ F(1,23) = 4.5, P = .04]. These findings indicate that whilst flight speeds change in response to increasing approach speed by the experimenter, the flight speeds of PBI creams are faster than that of both wood and house mice at submaximum speeds.
ANOVA revealed a significant effect of type on attempted pickups [ F(3,31) = 26.4, P = .00001], which was found to be due to more attempts with laboratory mice as compared to house mice [ F(1,31) = 35.3, P = .0000001], wood mice [ F(1,31) = 46.9, P = .0000001] and PBI creams [ F(1,31) = 65.66, P = .0000001]. Statistical analysis was not performed on the number of successful pickups because only laboratory mice and one wood mouse were successfully picked up. 3.3. Closed runway test Flight distance is defined as the distance at which an animal stops ongoing behaviour and uses the only escape route available, that is, running towards and past the approaching experimenter. There was a significant effect of type on flight distance [Fig. 4; F(3,31) = 23.53, P = .00001], which was due to house mice [ F(1,31) = 27.1, P = .00001] and PBI cream mice [ F(1,31) = 57.2, P = .0000001] having
Fig. 4. Flight distance in the closed runway. Flight distance is defined as the distance at which an animal stops ongoing behaviour and uses the only escape route available, that is, running towards and past the approaching experimenter. Laboratory mice (Lab) failed to show this behavioural strategy, whilst wood mice (Wood) favoured a freeze and stay strategy ( * P < .05 from Lab mice). See text for further details.
626
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Fig. 5. Catchtime in the closed runway. Laboratory mice (Lab) were the easiest to catch, followed by wood mice (Wood) and house mice (House). PBI creams showed the most vigorous escape attempts and took the longest time to catch ( * P < .05 from time taken to catch Lab mice). See text for further details.
greater flight distances than laboratory mice and to house mice [ F(1,31) = 13.03, P = .001] and PBI creams [ F(1,31) = 35.1, P = .000002] also having greater flight distances than wood mice (demonstrating this species’ preference to freeze when faced with a potential predator). Statistical analysis was not performed on the number of successful pickups because only laboratory mice were successfully picked up at the end of the closed runway test. An ANOVA on the time taken to catch each animal (catchtime) once the test session was over revealed a significant effect of type [Fig. 5; F(3,30) = 8.98, P = .0002], which was due to increased catchtime seen for house mice [ F(1,30) = 10.67, P = .0003], wood mice [ F(1,30) = 5.8, P = .02] and PBI creams [ F(1,30) = 25.97, P = .00002] as compared to laboratory mice. Catchtime was also significantly longer in PBI creams compared to wood mice [ F(1,30) = 6.42, P = .02].
4. Discussion In its standard use in behavioural pharmacology, a marked avoidance of the white section in the black/white box is viewed as being correlated with high levels of anxiety. The loss of this preference in drug-treated animals is taken as an indication of reduced levels of anxiety and is indicative of anxiolytic action [7,15]. Although in our experiments there were no significant differences between the mouse types in the percentage of time spent in the white section, there were highly significant differences in the latency to leave the white section and in the incidence of escape behaviour shown in this section. The behaviour shown in the black/white box by the PBI creams was closest to that of the house mice and very different from that of the laboratory mice. This pattern is clearly revealed in the PCA results. The mean first component scores of the PBI creams and the house mice do not significantly differ, and both are significantly greater than the scores for the laboratory mice. Whilst this component
(which explained 36% of the variance) appears to be an index of activity (i.e., highly positively weighted by line crosses, rearing and jumping and negatively weighted by grooming, sniffing and immobile), the high loadings with jumping behaviour suggests that it also measures the strength of an escape tendency. Although 66% of PBI creams, 33% of wood mice and 25% of house mice attempted to escape from the black/white box by jumping, no laboratory mice showed this behaviour. Therefore, the results from the black/white box test indicate that the PBI creams are far more like the wild mice than the laboratory mice in both their activity levels and escape tendencies. This distinction is reinforced by data from the runway tests. In the open runway, laboratory mice showed very few signs of reacting to the presence of a human in the same arena and failed to match their behaviour to the approach speed of the experimenter. Further, the distance between the experimenter and laboratory mice closed to zero on several occasions, as the experimenter caught up with the animal. In almost every instance where this happened, it was possible to pick up the mouse. In sharp contrast, the distance between the experimenter and PBI creams, wood mice or house mice rarely closed to zero, and when it did, this was caused by animals rapidly changing their direction of flight and running back towards and past the experimenter. On no occasion was it possible to pick up any of these species. The typical profile seen in PBI creams, wood mice and house mice was for these animals to maintain a distance of more than 2 m from the experimenter at all times and where possible to keep out of visual contact. Consequently, all these animals showed increased flight speed in response to increased approach speed by the experimenter. PBI creams tended to run immediately to the far end of the runway as soon as the experimenter turned the corner from one side of the runway to the other and so came into visual contact. By contrast, house mice and, in particular, wood mice, tended briefly to stop once or twice and orientate themselves towards the experimenter before continuing their run to the far end of the runway. At the highest approach speed
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
(i.e., 1.2 m s 1), such orientating strategies were rendered more difficult and nearly all animals ran at maximum speed from one end of the runway to the other. In consequence, the flight speeds of all three species were similar at this point. In the closed runway, all species (except laboratory mice which were placed there) moved to the far end of the runway in response to the presence of the experimenter. The experimenter then slowly approached the animal in order to determine its flight distance, which is the distance within which an animal will not tolerate a predator or other potential threat to come [12,13,17]. At this distance, animals usually break and run, using any escape route that is available. In the closed runway, the only possible escape route lies behind the slowly approaching experimenter. Hence, flight distance is defined as the distance at which the animal stops ongoing behaviour and runs towards and past the experimenter. The flight distance of PBI creams was around 90 cm: in direct contrast, the experimenter was able to approach laboratory mice to within a few centimetres and then pick them up. The high reactivity of the PBI creams was also highlighted by the time taken to catch each animal in the runway in order for it to be returned to its home cage. Even using the assistance of a holding box and Longworth live mouse traps, it still took in the order of 1½ min to catch these animals, whilst the time taken to catch each laboratory mouse was less than 5 s. Our data indicate that when PBI creams are placed in an environment where they are visually exposed, like the black/ white box, they show a much stronger motivation to escape than do laboratory mice. Furthermore, when PBI creams are placed in the same environment as a potential predator, as in the runway test, they are more reactive to the presence of a human than either wild wood mice or house mice. Physiological studies are now underway to determine the involvement of HPA axis activation because the marked contrast between the wild animal-type behaviour shown by the PBI creams and the behaviour of laboratory mice is of particular importance in the context of studying anxiety/fear in the laboratory. Current models of anxiety/fear are based on the assumption that a given situation induces this psychological state in animals. In many instances (i.e., all models not based on conditioning paradigms), the evidence for this is weak. The behavioural changes seen in standard laboratory mice are slight and prone to being greatly influenced by relatively minor procedural or environmental variations [4,16,25]. Further, the source of the proposed anxiety/fear in these models remains the subject of speculation [13]. Finally, its presence can only be inferred through the actions of drugs that have anxiety-reducing properties in humans [18]. Current data provide strong indications that PBI creams may be of immense value in overcoming these shortcomings. The extreme reactivity of PBI creams is in all probability a consequence of the selection pressures the original wild stock were placed under whilst being exposed to warfarin. The animals that developed resistance to this rodenticide
627
also developed a lighter coat [23]. Therefore, whilst warfarin resistance was clearly an advantage, it was linked to the disadvantage of being more easily detected by predators. The behavioural strategy developed to compensate for this seemingly took the form of accentuated flight responses. The combined effect of these two selection pressures was, thus, to produce a strain of light-coloured house mice that were more reactive to predators than normal house mice. The present colony of PBI creams, derived from this original wild stock, has been maintained under laboratory conditions for the past 30 years. It is remarkable that these accentuated flight responses have been retained over all this time, and it is unlikely that this could have happened outside an institution where techniques and training in animal husbandry are focused on the maintenance of wild-trapped animals. The combination of factors that led to the emergence and maintenance of the PBI creams represents a unique phenomenon.
References [1] Anderson PK. The role of breeding structure in the evolutionary process of Mus musculus populations. In: Honcariv R, editor. Mutation in populations. Prague: Academia, 1965. pp. 17 – 21. [2] Babinet C, Morello D, Renard JP. Transgenic mice. Genome 1989;31: 938 – 49. [3] Berry RJ. The natural history of the house mouse. Field Stud 1970;219 – 62. [4] Belzung C, Le Pape G. Comparison of different behavioural test situations used in psychopharmacology for the measurement of anxiety. Physiol Behav 1994;56:623 – 8. [5] Blanchard DC, Hori K, Rodgers RJ, Hendrie CA. Attenuation of defensive threat and attack wild rats (Rattus rattus) by benzodiazepines. Psychopharmacology 1989;97:392 – 401. [6] Capecchi MR. The new mouse genetics: altering the genome by gene targeting. Trends Genet 1989;5:70 – 6. [7] Costall B, Hendrie CA, Kelley ME. The anxiolytic effects of the substituted benzamides. Br J Pharmacol 1986;88:322. [8] Crowcroft P. Territoriality in wild house mice Mus musculus L. Mammal 1955;36:299 – 301. [9] Crowcroft P, Rowe FP. Social organisation and territorial behaviour in the wild house mouse (Mus musculus L). Proc Zool Soc London 1963;140:517 – 31. [10] Darwin C. On the origin of species by means of natural selection, or, the preservation of favoured races in the struggle for life. London: J. Murray, 1859. [11] DeFries JC, McClearn GE. Behavioural genetics and the fine structure of mouse populations: a study in microevolution. In: Dobzhansky T, Hecht MK, Steere WC, editors. Evolutionary biology. New York (NY): Appleton-Century-Crofts, 1972. pp. 279 – 91. [12] Gallup GG. Animal hypnosis: factual status of a fictional concept. Psychol Bull 1974;81:836 – 53. [13] Hendrie CA, Weiss SM, Eilam D. Exploration and predation models of anxiety: evidence from laboratory and wild species. Pharmacol, Biochem Behav 1996;54:13 – 20. [14] Hendrie CA, Eilam D, Weiss SM. The effects of diazepam and buspirone on the behaviour of wild voles (Microtus socialis) in two models of anxiety. Pharmacol, Biochem Behav 1997;58:573 – 6. [15] Hendrie CA, Neil JC, Shepherd JK, Dourish CT. The effects of CCKA and CCK-B antagonists on activity in the black/white exploration model of activity in mice. Physiol Behav 1993;54:689 – 93.
628
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
[16] Hogg S. A review of the validity and variability of the elevated-plus maze as an animal model of anxiety. Pharmacol, Biochem Behav 1996;54:21 – 30. [17] Ratner SC. Immobility in invertebrates: what can we learn? Psychol Rev 1977;1:1 – 14. [18] Rodgers RJ, Cole JC. The elevated plus maze: pharmacology, methodology and ethology. In: Cooper SJ, Hendrie CA, editors. Ethology and psychopharmacology. Chichester: Wiley, 1994. pp. 9 – 44. [19] Rogers DC, Fisher EMC, Brown JDM, Peter J, Hunter AJ, Martin JE. Behavioural and functional analysis of the mouse phenotype: SHIRTA, a proposed protocol for comprehensive phenotype assessment. Mamm Genome 1997;8:711 – 3. [20] Selander RK. Behavior and genetic variation in natural populations. Am Zool 1970;10:53 – 66.
[21] Southern HN, Laurie HM. The house-mouse (Mus musculus) in corn ricks. J Anim Ecol 1946;15:134 – 49. [22] Wallace ME, MacSwiney FJ. Warfarin resistance: obesity. Mouse News Lett 1974;50:28. [23] Wallace ME, MacSwiney FJ. A major gene controlling warfarinresistance in the house mouse. J Hyg 1976;76:173 – 81. [24] Weber JS, Salinger A, Justice MJ. Optimal N-ethyl-N-nitrosourea (ENU) doses for inbred mouse strains. Genesis 2000;26:230 – 3. [25] Weiss SM, Lightowler S, Stanhope KJ, Kennett GA, Dourish CT. Measurement of anxiety in transgenic mice. Rev Neurosci 2000;11: 53 – 74. [26] Young H, Strecker RL, Emlen JT. Localization of activity in two indoor populations of house mice, Mus musculus. J Mammal 1950;31:403 – 10.
PBI creams A spontaneously mutated mouse strain showing wild animal-type reactivity Colin A. Hendriea,*, Katja S. Van Drielb, Janet C. Tallingb, Ian R. Inglisb a b
School of Psychology, University of Leeds, Leeds LS2 9JT, UK Central Science Laboratory, Sand Hutton, York YO41 1LZ, UK
Received 12 February 2001; received in revised form 15 February 2001; accepted 20 August 2001
Abstract PBI creams are mice derived from warfarin-resistant wild stock that has been maintained under laboratory conditions since the 1970s. This study compares their behaviour to that of laboratory mice and wild house and wood mice. Animals were tested in a black/white box and a 2.64 1.4 m runway. In the black/white box, the behaviour of PBI creams was not significantly different from that of house mice and differed most from that of laboratory mice. Notably, the PBI creams showed the greatest activity and escape-orientated behaviours. When animals were approached by the experimenter in the open runway test, the PBI creams had higher flight speeds than both house and wood mice, whilst laboratory mice failed to respond. In the closed runway test where the animals could not escape, the PBI creams, house mice and wood mice all turned and attempted to run past the approaching experimenter, whilst the laboratory mice again failed to react. At the end of this test session, the time taken to catch each animal was recorded. It took less than 5 s to catch laboratory mice but significantly longer to catch the wild strains and the PBI creams (90 – 100 s for the latter). In these tests, the PBI creams showed wild animal-type reactivity, and as this behaviour has been retained in the laboratory colony for over 30 years, these animals may be useful in the study of the physiological and genetic basis of fear/anxiety in mice. D 2001 Elsevier Science Inc. All rights reserved. Keywords: PBI creams; Wood mice; House mice; Fear; Anxiety; Genetic mutation
1. Introduction In recent times much attention has been given to producing mutagenesis in mice with the aim of modelling human genetically based disorders in the laboratory [2,6,19,24]. Mutation is, of course, the basis for evolutionary development [10], but it is equally clear that in the wild selection pressures operate at the level of the phenotype. Behaviour must therefore be considered the most important factor influencing the survival of genetic mutations from generation to generation. This point is particularly applicable to mice. In the wild, mice are aggressively territorial [3,8] and live in demes, or small breeding groups usually consisting of a single male and several females. Whilst these structures may be temporarily broken down by an over* Corresponding author. Tel.: +44-113-233-5736; fax: +44-113-2335749. E-mail address: [email protected] (C.A. Hendrie).
abundance of food, such as those found in grain stores [9], these changes in social behaviour are transient. Evidence of such territoriality can also be seen in laboratory mice where food supplies are permanently available in excess [13]. Therefore, where food supplies are stable, territoriality remains the norm. In the wild there is little genetic interchange between these small breeding units [11] to the extent that, even on small islands, introduced mutant alleles fail to spread [1]. Further, territory sizes are small, ranging between 50 and 90 f t 2 [21,26], meaning that distinct subpopulations can coexist within geographical locations as small as a single barn [20]. Data, such as these, graphically demonstrate that the social structure of mice serves to isolate these small local populations and so increases the likelihood of allelic variation within them. With such a high potential for variation, it is unsurprising that populations of wild mice occasionally emerge that have unique or exaggerated behavioural characteristics produced in response to altered selection pressures
0031-9384/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 1 ) 0 0 6 0 4 - 7
622
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
within their environment. The spontaneous emergence of such animals has clear parallels with the controlled studies concerned with the induction of genetic mutations in the laboratory and may have direct use in addressing questions relating to functionality. These wild animals have, by definition, survived and reproduced outside the laboratory and hence overcome the demands of the natural environment. This paper describes the behaviour of ‘PBI creams’, mice derived from animals captured during infestation control at the Plant Breeding Institute, Cambridge in the 1970s [23]. Warfarin was used to control the infestation and over time the incidence of dark-eyed, cream-coloured animals, not previously seen in this population, increased. The gene responsible for this cream colour, the extreme chinchilla (che), was found to be linked to warfarin resistance [22]. Hence, in this wild population, the numbers of creamcoloured mice increased in response to the exposure to warfarin. The present colony, held at the Central Science Laboratory since 1971, was derived from this stock: as far as is known, such animals no longer exist in the wild. The most striking behavioural feature of the PBI creams is their extreme reactivity, and the current studies were carried out in order to characterise their behaviour in the context of the behaviour of laboratory mice, wild wood mice and wild house mice. Two models were employed that had previously been used to examine the behaviour of wild animals. The black/white box has been used to test the behaviour of wild-trapped voles [14], and the runway test has been used as part of a test battery to examine the behaviour of wild-trapped rats (e.g. [5]).
2.2. Procedures 2.2.1. Black/white box The black/white box consists of an open-topped box (45 27 27 cm) that is painted two-fifths black and threefifths white. These two sections are divided by an opaque partition with a small opening (7.5 7.5 cm) positioned in the centre of the box at floor level. The floor area of each section was lined to produce 9 cm squares. The black section was illuminated by a 60-W red light set to give light levels of 50 lx, whilst the white section was illuminated by two 60-W white lights, set to give light levels of 500 lx. Lights were carefully positioned so as not to produce areas of shadow. At the start of each session, individual animals were placed into the middle square of the white section. The experimenter then left the room and the behaviour of the animal was recorded on videotape for the next 5 min. At the end of each session, any faecal boli were removed, and the arena was cloth-wiped and dried before the next animal was introduced. The ethogram used for analysing behaviour in this situation was as follows: Latency to enter black section Time in each section Line crossing Rearing Grooming
2. Methods Sniffing substrate 2.1. Animals and housing Immobile All animals were male and approximately 6 months old at the time of testing. Stocks of outbred albino laboratory mice were originally obtained from the Medical Research Council Laboratory Animals Centre, Carshalton, Surrey, UK and had been held for 71 generations prior to testing. The PBI creams had been held for 52 generations, and the house mice (Mus musculus) and wood mice (Apodemus sylvaticus) were four generations removed from the original wild-trapped stock. All animals were housed in groups of three in polycarbonate RB3 cages (45 28 11 cm) under a 12-h light cycle (lights on: 06:00 h) in a temperature (20 ± 2 °C) and humidity (55 ± 15%) controlled environment. The sawdust substrate was replaced once a week, as were small cardboard containers that served both as shelters and sources of nesting material. Food (GRE3K lab chow) and water were available ad libitum. Animals (n’s = 8 –9 per mouse type) were tested between 10:00 and 15:00 h — initially in the black/white box and 14 days later in the runway test (see below).
Jumps Jump out
Other
Time taken for the animal to move from its original starting point in the white section into the black section. Timing in one section began when all four paws were in that section and ended when all four paws left it. Head breaks the plane of a line drawn on floor. Standing on rear legs with straight back and forepaws off the floor. Using forepaws and licking to groom body. Olfactory examination of walls or floor. Cessation of movement with or without intermittent head or body adjustments. All four paws leaving contact with the floor. As jump but with upward movement resulting in animal escaping from test arena. Any other behaviour.
2.2.2. Runway test The runway was a 264 140 70 cm rectangle, constructed of rigid aluminium sheeting with the width of the floor marked off with thin white lines every 33 cm. In the open runway format a 185 20 70 cm divider was placed in the centre of arena 60 cm from the longest side and positioned to leave a 40-cm gap at each end. In the closed runway format, the gap at one end of the apparatus was closed off. The room housing the runway was lit by four 60 W bulbs.
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Prior to testing, animals were individually placed into the runway and left undisturbed for 5 min. At the end of this period, the experimenter entered the middle of the runway and proceeded to move clockwise. Two circuits were made at approximately 0.3 m s 1, two at 0.6 m s 1 and two at 1.2 m s 1, i.e., slow walking pace, fast walking pace and jogging pace. Following this, one end of the divider was closed off to create the closed runway format. For the closed runway test, the experimenter stood at the furthest line, 231 cm from the closed end, and remained motionless for 30 s. The experimenter then slowly moved forward to the next line, which was 33 cm nearer the closed end, and remained motionless again for another 30 s. This process was repeated every 30 s until the experimenter reached the end zone, the last line 33 cm away from the wall of the closed end. At this point, the experimenter slowly bent down and attempted to pick up the animal. An attempt was also made to pick up the animal if it ran towards and past the experimenter prior to this end point being reached. At the end of each session, Longworth live animal traps were placed in each corner of the apparatus, and the experimenter, using a housing box, attempted to catch the animal as quickly as possible. At the end of each session, any faecal boli were removed, and the arena cloth-wiped and dried before the next animal was introduced. Measures taken during this test were as follows: Open runway Flight Time taken for the animal to move from speed the point where it started to run away from the experimenter to the point where it reached the end of the divider and moved into the other section of the runway. Different measures of this were taken for each of the approach speeds produced by the experimenter. Stops Animal briefly ceases running and orientates towards the approaching experimenter before beginning to run again. Pickup Single attempt to bend down and pick up the animal if the distance between experimenter and animal closes to zero. Closed runway Flight Distance at which the animal ceases distance ongoing activity and runs from the end zone towards and past the experimenter. Pickup Single attempt to bend down and pick the animal up when distance between experimenter and animal is reduced to zero as the animal runs towards and past the experimenter or the experimenter reaches the end zone. Catchtime Time taken to capture each animal at the end of the test session.
623
2.3. Statistical analysis Data were analysed using principal component analysis (PCA) and single or two-factor analysis of variance (ANOVA) employing repeated measures where appropriate. Planned pair-wise follow-up tests were conducted using orthogonal contrasts.
3. Results 3.1. Black/white box The latency to enter the black section from the white section revealed a significant effect of mouse type [ F(3,31) = 3.3, P = .033]. Follow-up comparisons found that this effect resulted from the laboratory mice having significantly longer latencies than the other mice. In rank order, the PBI creams had the shortest latency followed by house mice, wood mice and laboratory mice (Fig. 1). However, analysis of the percentage of trial time spent in the white section revealed no significant type differences [ F(3,31) = 0.56, ns]. Some mice jumped out of the equipment and hence had behavioural records less than 5 min. Approximately 66% of PBI creams, 33% of wood mice and 25% of house mice jumped at least once, and 33% of PBI creams and 25% of house mice succeeded in jumping out of the test apparatus before the end of the test session. No laboratory mice jumped upwards or out of the test arena. As a result, there were significant mouse type differences in the total time spent in the box [Fig. 2; F(3,31) = 3.32, P = .032]. Followup comparisons revealed that this result stemmed from a significantly shorter duration for the PBI creams than for the other species. The behavioural data were, therefore, divided by the total time spent in the black and white sections of the box to give standardised rate scores, which
Fig. 1. Latencies to enter the black section of the black/white box. Wood mice (Wood), house mice (House) and PBI cream mice had significantly shorter latencies than laboratory mice (Lab) to enter the black section from their initial starting point in the white section ( * P < .05 compared to Lab mice). See text for further details.
624
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Fig. 2. Total time spent in the black/white box. Although no laboratory mice (Lab) showed this behaviour, both species of wild mouse (House, Wood) and the PBI creams made attempts to jump out of the test apparatus. Approximately 33% of PBI creams and 25% of house mice were successful. In consequence, PBI creams and house mice spent less than the full 300 s in the black/white box. ( * P < .05 from time laboratory mice spent in test apparatus). See text for further details.
were then analysed for differences between the four types of mice. Of the behavioural categories shown in the black section, significant mouse type effects were found only in the rates of rearing [ F(3,31) = 5.4, P = .005] and sniffing [ F(3,31) = 8.0, P = .001]. Follow-up comparisons revealed that these effects were solely due to the laboratory mice showing significantly less rearing and more sniffing in this section than the other types. Significant differences between the mouse types were also found in the frequencies of rearing [ F(3,31) = 12.5, P = .000] and sniffing [ F(3,31) = 5.2, P = .005] in the white section of the box. Follow-up comparisons showed that the differences in sniffing resulted from the PBI creams having significantly lower scores than the laboratory mice, whilst the rearing differences were caused by both the PBI creams and the house mice showing significantly higher frequencies of rearing than the laboratory mice. Significant activity differences between mouse types were found only in the white section of the box [ F(3,31) = 5.6, P = .004]. This resulted from both the PBI creams and the house mice having similar and significantly higher line cross scores than both the wood mice and the laboratory mice. The relative effects of the white and black sections on the different mouse types were compared by subtracting the black section scores for a given behaviour from the corresponding white section scores and then looking for significant type effects in these indices. The four mouse types were significantly differentiated by two of the resulting indices. There was a significant difference between the mouse types in the relative rates of line crosses in the two sections of the box [ F(3,31) = 3.4, P = 033], with the white section eliciting a significantly greater increase in activity in the PBI creams and the house mice (that were not significantly different from each other) than in the wood
mice and laboratory mice (that were not significantly different from each other). The other index showing significant differences between mouse types involved rearing [ F(3,31) = 5.2, P = .006] where both the PBI creams and the house mice (that were not significantly different from each other) had a significantly higher incidence of rearing in the white section than in the black section, whereas both the wood mice and laboratory mice (that were not significantly different from each other) showed a significant difference in the opposite direction. The frequency scores for the behavioural categories in the black and white sections were analysed using PCA. As shown in Table 1, the first component explained the majority of the variance (36% as compared with 13% for the second and third components and less than 10% for each of the remaining components). This component was an activity index, as it was strongly correlated with the frequencies of line crosses, rearing and jumping in both white and black sections and negatively correlated with the frequency of immobility, grooming and sniffing in both sections. An ANOVA on the first component scores revealed highly significant differences between the mouse types [ F(3,31) = 10.7, P = .0001], which resulted from the PBI creams and house mice having similar and significantly greater first component scores from both the wood mice and the laboratory mice and also the wood mice having significantly greater scores than the laboratory mice. 3.2. Open runway test As the laboratory mice failed to react to the approach of the experimenter in the runway by running away, only the
Table 1 Black/white box: component matrix for principal components with eigenvalues greater than 1 Component Variable Rear white Sniff white Groom white Immobile white Jumps white Line crosses white Rear black Sniff black Groom black Immobile black Jumps black Line crosses black Initial eigenvalues total Percent of variance Cumulative variance
1
2
3
.89 .48 .05 .24 .51 .86 .83 .61 .11 .43 .58 .82
.05 .48 .76 .57 .26 .27 .06 .16 .43 .31 .29 .06
.07 .18 .04 .31 .54 .07 .37 .52 .41 .62 .41 .3
4.32 36.1 36.1
1.67 13.8 49.9
1.63 13.6 63.5
The first component, which explains the majority of the variance, is positively correlated with rearing, jumping and line crossing. It is negatively correlated with immobility and sniffing. This component may therefore represent an index of both activity and escape.
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
625
Fig. 3. Flight speed in the open runway. Dark grey bars = flight speed when experimenter approached at 0.3 m s 1; grey bars = flight speed when experimenter approached at 0.6 m s 1; white bars = flight speed when experimenter approached at 1.2 m s 1. Laboratory mice (Lab) failed to match their behaviour to the approach of the experimenter. Hence, their flight speed could not be determined. House mice (House), wood mice (Wood) and PBI cream mice all showed increased flight speed at the fastest approach speed compared to their flight speed at the slowest approach speed ( * P < .05). The flight speed of PBI creams was faster than that of wood mice and house mice at 0.3 and 0.6 m s 1 ( + P < .05 as compared to wood mice and house mice). See text for further details.
data from the two wild types and the PBI creams could be analysed. ANOVA performed on data from these animals revealed a significant effect of both type [ F(2,23) = 8.19, P = .002] and approach speed [ F(2,46) = 12.25, P = .0001; Fig. 3]. Follow-up tests revealed these effects to be due to significantly increased flight speeds when approached at 1.2 m s 1 as compared with being approached at 0.3 m s 1 for house mice [ F(1,23) = 8.9, P = .006], wood mice [ F(1,23) = 12.2, P = .0001] and PBI creams [ F(1,23) = 4.6, P = .04]. The flight speed of PBI creams was significantly greater than house mice at 0.3 m s 1 [ F(1,23) = 6.65, P = .016] and at 0.6 m s 1 [ F(1,23) = 5.86, P = .02] but not at 1.2 m s 1. A similar effect was seen when comparing the flight speeds of PBI creams and wood mice at 0.3 m s 1 [ F(1,23) = 14.0, P = .001] and at 0.6 m s 1 [ F(1,23) = 4.5, P = .04]. These findings indicate that whilst flight speeds change in response to increasing approach speed by the experimenter, the flight speeds of PBI creams are faster than that of both wood and house mice at submaximum speeds.
ANOVA revealed a significant effect of type on attempted pickups [ F(3,31) = 26.4, P = .00001], which was found to be due to more attempts with laboratory mice as compared to house mice [ F(1,31) = 35.3, P = .0000001], wood mice [ F(1,31) = 46.9, P = .0000001] and PBI creams [ F(1,31) = 65.66, P = .0000001]. Statistical analysis was not performed on the number of successful pickups because only laboratory mice and one wood mouse were successfully picked up. 3.3. Closed runway test Flight distance is defined as the distance at which an animal stops ongoing behaviour and uses the only escape route available, that is, running towards and past the approaching experimenter. There was a significant effect of type on flight distance [Fig. 4; F(3,31) = 23.53, P = .00001], which was due to house mice [ F(1,31) = 27.1, P = .00001] and PBI cream mice [ F(1,31) = 57.2, P = .0000001] having
Fig. 4. Flight distance in the closed runway. Flight distance is defined as the distance at which an animal stops ongoing behaviour and uses the only escape route available, that is, running towards and past the approaching experimenter. Laboratory mice (Lab) failed to show this behavioural strategy, whilst wood mice (Wood) favoured a freeze and stay strategy ( * P < .05 from Lab mice). See text for further details.
626
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
Fig. 5. Catchtime in the closed runway. Laboratory mice (Lab) were the easiest to catch, followed by wood mice (Wood) and house mice (House). PBI creams showed the most vigorous escape attempts and took the longest time to catch ( * P < .05 from time taken to catch Lab mice). See text for further details.
greater flight distances than laboratory mice and to house mice [ F(1,31) = 13.03, P = .001] and PBI creams [ F(1,31) = 35.1, P = .000002] also having greater flight distances than wood mice (demonstrating this species’ preference to freeze when faced with a potential predator). Statistical analysis was not performed on the number of successful pickups because only laboratory mice were successfully picked up at the end of the closed runway test. An ANOVA on the time taken to catch each animal (catchtime) once the test session was over revealed a significant effect of type [Fig. 5; F(3,30) = 8.98, P = .0002], which was due to increased catchtime seen for house mice [ F(1,30) = 10.67, P = .0003], wood mice [ F(1,30) = 5.8, P = .02] and PBI creams [ F(1,30) = 25.97, P = .00002] as compared to laboratory mice. Catchtime was also significantly longer in PBI creams compared to wood mice [ F(1,30) = 6.42, P = .02].
4. Discussion In its standard use in behavioural pharmacology, a marked avoidance of the white section in the black/white box is viewed as being correlated with high levels of anxiety. The loss of this preference in drug-treated animals is taken as an indication of reduced levels of anxiety and is indicative of anxiolytic action [7,15]. Although in our experiments there were no significant differences between the mouse types in the percentage of time spent in the white section, there were highly significant differences in the latency to leave the white section and in the incidence of escape behaviour shown in this section. The behaviour shown in the black/white box by the PBI creams was closest to that of the house mice and very different from that of the laboratory mice. This pattern is clearly revealed in the PCA results. The mean first component scores of the PBI creams and the house mice do not significantly differ, and both are significantly greater than the scores for the laboratory mice. Whilst this component
(which explained 36% of the variance) appears to be an index of activity (i.e., highly positively weighted by line crosses, rearing and jumping and negatively weighted by grooming, sniffing and immobile), the high loadings with jumping behaviour suggests that it also measures the strength of an escape tendency. Although 66% of PBI creams, 33% of wood mice and 25% of house mice attempted to escape from the black/white box by jumping, no laboratory mice showed this behaviour. Therefore, the results from the black/white box test indicate that the PBI creams are far more like the wild mice than the laboratory mice in both their activity levels and escape tendencies. This distinction is reinforced by data from the runway tests. In the open runway, laboratory mice showed very few signs of reacting to the presence of a human in the same arena and failed to match their behaviour to the approach speed of the experimenter. Further, the distance between the experimenter and laboratory mice closed to zero on several occasions, as the experimenter caught up with the animal. In almost every instance where this happened, it was possible to pick up the mouse. In sharp contrast, the distance between the experimenter and PBI creams, wood mice or house mice rarely closed to zero, and when it did, this was caused by animals rapidly changing their direction of flight and running back towards and past the experimenter. On no occasion was it possible to pick up any of these species. The typical profile seen in PBI creams, wood mice and house mice was for these animals to maintain a distance of more than 2 m from the experimenter at all times and where possible to keep out of visual contact. Consequently, all these animals showed increased flight speed in response to increased approach speed by the experimenter. PBI creams tended to run immediately to the far end of the runway as soon as the experimenter turned the corner from one side of the runway to the other and so came into visual contact. By contrast, house mice and, in particular, wood mice, tended briefly to stop once or twice and orientate themselves towards the experimenter before continuing their run to the far end of the runway. At the highest approach speed
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
(i.e., 1.2 m s 1), such orientating strategies were rendered more difficult and nearly all animals ran at maximum speed from one end of the runway to the other. In consequence, the flight speeds of all three species were similar at this point. In the closed runway, all species (except laboratory mice which were placed there) moved to the far end of the runway in response to the presence of the experimenter. The experimenter then slowly approached the animal in order to determine its flight distance, which is the distance within which an animal will not tolerate a predator or other potential threat to come [12,13,17]. At this distance, animals usually break and run, using any escape route that is available. In the closed runway, the only possible escape route lies behind the slowly approaching experimenter. Hence, flight distance is defined as the distance at which the animal stops ongoing behaviour and runs towards and past the experimenter. The flight distance of PBI creams was around 90 cm: in direct contrast, the experimenter was able to approach laboratory mice to within a few centimetres and then pick them up. The high reactivity of the PBI creams was also highlighted by the time taken to catch each animal in the runway in order for it to be returned to its home cage. Even using the assistance of a holding box and Longworth live mouse traps, it still took in the order of 1½ min to catch these animals, whilst the time taken to catch each laboratory mouse was less than 5 s. Our data indicate that when PBI creams are placed in an environment where they are visually exposed, like the black/ white box, they show a much stronger motivation to escape than do laboratory mice. Furthermore, when PBI creams are placed in the same environment as a potential predator, as in the runway test, they are more reactive to the presence of a human than either wild wood mice or house mice. Physiological studies are now underway to determine the involvement of HPA axis activation because the marked contrast between the wild animal-type behaviour shown by the PBI creams and the behaviour of laboratory mice is of particular importance in the context of studying anxiety/fear in the laboratory. Current models of anxiety/fear are based on the assumption that a given situation induces this psychological state in animals. In many instances (i.e., all models not based on conditioning paradigms), the evidence for this is weak. The behavioural changes seen in standard laboratory mice are slight and prone to being greatly influenced by relatively minor procedural or environmental variations [4,16,25]. Further, the source of the proposed anxiety/fear in these models remains the subject of speculation [13]. Finally, its presence can only be inferred through the actions of drugs that have anxiety-reducing properties in humans [18]. Current data provide strong indications that PBI creams may be of immense value in overcoming these shortcomings. The extreme reactivity of PBI creams is in all probability a consequence of the selection pressures the original wild stock were placed under whilst being exposed to warfarin. The animals that developed resistance to this rodenticide
627
also developed a lighter coat [23]. Therefore, whilst warfarin resistance was clearly an advantage, it was linked to the disadvantage of being more easily detected by predators. The behavioural strategy developed to compensate for this seemingly took the form of accentuated flight responses. The combined effect of these two selection pressures was, thus, to produce a strain of light-coloured house mice that were more reactive to predators than normal house mice. The present colony of PBI creams, derived from this original wild stock, has been maintained under laboratory conditions for the past 30 years. It is remarkable that these accentuated flight responses have been retained over all this time, and it is unlikely that this could have happened outside an institution where techniques and training in animal husbandry are focused on the maintenance of wild-trapped animals. The combination of factors that led to the emergence and maintenance of the PBI creams represents a unique phenomenon.
References [1] Anderson PK. The role of breeding structure in the evolutionary process of Mus musculus populations. In: Honcariv R, editor. Mutation in populations. Prague: Academia, 1965. pp. 17 – 21. [2] Babinet C, Morello D, Renard JP. Transgenic mice. Genome 1989;31: 938 – 49. [3] Berry RJ. The natural history of the house mouse. Field Stud 1970;219 – 62. [4] Belzung C, Le Pape G. Comparison of different behavioural test situations used in psychopharmacology for the measurement of anxiety. Physiol Behav 1994;56:623 – 8. [5] Blanchard DC, Hori K, Rodgers RJ, Hendrie CA. Attenuation of defensive threat and attack wild rats (Rattus rattus) by benzodiazepines. Psychopharmacology 1989;97:392 – 401. [6] Capecchi MR. The new mouse genetics: altering the genome by gene targeting. Trends Genet 1989;5:70 – 6. [7] Costall B, Hendrie CA, Kelley ME. The anxiolytic effects of the substituted benzamides. Br J Pharmacol 1986;88:322. [8] Crowcroft P. Territoriality in wild house mice Mus musculus L. Mammal 1955;36:299 – 301. [9] Crowcroft P, Rowe FP. Social organisation and territorial behaviour in the wild house mouse (Mus musculus L). Proc Zool Soc London 1963;140:517 – 31. [10] Darwin C. On the origin of species by means of natural selection, or, the preservation of favoured races in the struggle for life. London: J. Murray, 1859. [11] DeFries JC, McClearn GE. Behavioural genetics and the fine structure of mouse populations: a study in microevolution. In: Dobzhansky T, Hecht MK, Steere WC, editors. Evolutionary biology. New York (NY): Appleton-Century-Crofts, 1972. pp. 279 – 91. [12] Gallup GG. Animal hypnosis: factual status of a fictional concept. Psychol Bull 1974;81:836 – 53. [13] Hendrie CA, Weiss SM, Eilam D. Exploration and predation models of anxiety: evidence from laboratory and wild species. Pharmacol, Biochem Behav 1996;54:13 – 20. [14] Hendrie CA, Eilam D, Weiss SM. The effects of diazepam and buspirone on the behaviour of wild voles (Microtus socialis) in two models of anxiety. Pharmacol, Biochem Behav 1997;58:573 – 6. [15] Hendrie CA, Neil JC, Shepherd JK, Dourish CT. The effects of CCKA and CCK-B antagonists on activity in the black/white exploration model of activity in mice. Physiol Behav 1993;54:689 – 93.
628
C.A. Hendrie et al. / Physiology & Behavior 74 (2001) 621–628
[16] Hogg S. A review of the validity and variability of the elevated-plus maze as an animal model of anxiety. Pharmacol, Biochem Behav 1996;54:21 – 30. [17] Ratner SC. Immobility in invertebrates: what can we learn? Psychol Rev 1977;1:1 – 14. [18] Rodgers RJ, Cole JC. The elevated plus maze: pharmacology, methodology and ethology. In: Cooper SJ, Hendrie CA, editors. Ethology and psychopharmacology. Chichester: Wiley, 1994. pp. 9 – 44. [19] Rogers DC, Fisher EMC, Brown JDM, Peter J, Hunter AJ, Martin JE. Behavioural and functional analysis of the mouse phenotype: SHIRTA, a proposed protocol for comprehensive phenotype assessment. Mamm Genome 1997;8:711 – 3. [20] Selander RK. Behavior and genetic variation in natural populations. Am Zool 1970;10:53 – 66.
[21] Southern HN, Laurie HM. The house-mouse (Mus musculus) in corn ricks. J Anim Ecol 1946;15:134 – 49. [22] Wallace ME, MacSwiney FJ. Warfarin resistance: obesity. Mouse News Lett 1974;50:28. [23] Wallace ME, MacSwiney FJ. A major gene controlling warfarinresistance in the house mouse. J Hyg 1976;76:173 – 81. [24] Weber JS, Salinger A, Justice MJ. Optimal N-ethyl-N-nitrosourea (ENU) doses for inbred mouse strains. Genesis 2000;26:230 – 3. [25] Weiss SM, Lightowler S, Stanhope KJ, Kennett GA, Dourish CT. Measurement of anxiety in transgenic mice. Rev Neurosci 2000;11: 53 – 74. [26] Young H, Strecker RL, Emlen JT. Localization of activity in two indoor populations of house mice, Mus musculus. J Mammal 1950;31:403 – 10.