Physiology and Behavior, Vol. 14, pp. 833--838. Brain Research Publications Inc., 1975. Printed in the U.S.A.
Genetic Analysis of Water Maze Discrimination Learning for Mus musculus: Polygenes and Albinism M E R R I L L F. E L I A S 3
Department o f Psychology and All-University, Gerontology Center, Syracuse University Syracuse, N Y 13210 AND B A S I L E. E L E F T H E R I O U
The Jackson Laboratory, Bar Harbor, ME 04609
(Received 15 O c t o b e r 1974) ELIAS, M. F. AND B. E. ELEFTHERIOU. A genetic analysis of water maze discrimination learning for Mus musculus: polygenes and albinism. PHYSIOL. BEHAV. 14(6) 833-838, 1975. - Recombinant inbred strains, their progenitor strains and reciprocal F 1 hybrids were given thirty spatial discrimination learning trials in the water maze. The pattern of RI strains in relation to the reciprocal F 1 hybrids and the progenitor strains, and differences among strains, suggested that different sets of genes affect the various components of water maze discrimination learning. There was no evidence that the albino gene influenced either inferior or superior performance. Pigmented strains performed as well as, or better than, some albino strains, and albino strain BALB/cBy differed significantly from albino strains CXBG and CXBI. The two reciprocal F 1 hybrids differed for Blocks 2 and 3 of discrimination trials when swimming time was used a dependent variable. Thus, it appears that performance of the F 1 hybrids was influenced by paternal effects, maternal effects, or paternal maternal interactions. Albino gene
Discrimination learning
Single gene hypothesis
R E S U L T S of m a n y e x p e r i m e n t s (e.g. [3, 8, 17, 2 7 ] ) with inbred strains o f mice indicate that differences in g e n o t y p e are associated with differences in maze learning ability. More recent e x p e r i m e n t s have been c o n c e r n e d w i t h the m o d e o f inheritance o f maze learning ability. Comparisons of F 1 hybrids and their parental strains have provided evidence for a d o m i n a n t or o v e r d o m i n a n t m o d e o f inheritance depending on the specific parental strains which were crossed to p r o d u c e the F 1 hybrids [4, 18, 3 3 ] . In a d d i t i o n to use of F 1 hybrids for the study o f m o d e o f inheritance, m u t a n t mice have been used as a means o f evaluating the c o n t r i b u t i o n o f single genes to maze learning ability. F o r example, coat color m u t a n t s and neurological mutants segregating on a C 5 7 B L background have been tested for maze learning with tas .ks involving f o o d reward and water escape [ 2 1 ] . Albino (cJ) and u n d e r w h i t e (ww)
Maze learning
Recombinant inbred strains
mice m a d e fewer errors than normal C 5 7 B L / 6 J mice. This finding p r o m p t e d the conclusion that single genes m a y contribute to a considerable e x t e n t to additive and d o m i n a n c e genetic variance for maze learning. However, a question has been raised regarding the possibility that single gene effects observed for coat color m u t a n t and neurological m u t a n t m i c e may reflect trivial pleiotropic effects involving peripheral mechanisms [30] or may constitute the behavi o r a l manifestation of severe central nervous system disorder [ 10]. The value o f single gene m u t a n t research and the issue of triviality of peripheral nervous system effects has been debated [25,30] and the value of single gene research has been aptly d e f e n d e d [ 2 5 ] . Nevertheless, alternative m e t h o d s for testing single gene h y p o t h e s e s and for establishing linkage between a behavioral p h e n o m e n o n and a
1This research was supported in part by NIH research grants HD 05860 to BEE and HD 08220 to MFE, both from the National Institute of Child Health and Human Development. We gratefully acknowledge the use of computer services provided by Mr. Robert Rupp (deceased) of the Jackson Laboratory. 2The Jackson Laboratory is fully accredited by the American Association for Accreditation of Laboratory Animal Care. 3 Performed while MFE was a Visiting Investigator at the Jackson Laboratory. 833
834 specific chromosome segment may be of considerable value. This is particularly true if these alternate techniques do not depend on the use of mutant stocks which display obvious structural or physiological defects of the central or peripheral nervous system. Bailey [1] has provided a method for exploring the possibility of single gene (or major gene) effects and, under appropriate circumstances, for establishing linkage. The Recombinant Inbred (RI) Strains have been utilized quite effectively in the analysis of gene systems involved in endocrine, behavioral, and immunological phenomena [2, 6, 19]. The purpose of the present experiment was to apply this method of genetic analysis to the study of water maze discrimination learning. Water maze learning was chosen for several reasons. It is a highly popular learning task for the mouse. It was designed specificially for studies of motivation and learning in the mouse [28], and it has been used in studies dealing with effects of lesions [9], neurological deficit [ 10], brain weight [ 7, 31 ], motivation [ 28,29 ], and sustained high blood pressure on behavior [ 11 ]. Significant positive genetic correlations between water maze learning and shock avoidance have been reported for F 3 mice and their F 2 parents [18], and for two single gene mutants [21]. Based on this finding, Oliverio and Messeri [21] suggest that pleiotropic effects of the same gene(s) influence both maze learning and shock avoidance learning or that, alternatively, both behavioral phenomena are affected by a common underlying physiological mechanism. In view of the correlation between avoidance learning and maze learning and the recent demonstration of a major, or single gene effects, for avoidance learning in a number of studies [6, 23, 24], it seemed reasonable to explore the possibility of a major gene effect for water maze discrimination learning with the genetic analysis technique made possible by Bailey's RI strains. In addition to the questioning of a major gene effect on maze learning with a new method of genetic analysis, the RI strains may be used to determine the extent to which the albino gene is, in fact, directly responsible for performance extremes observed for inbred albino strains. Oliverio and Messeri's [21 ] finding of better learning for albino mutant mice (c i) as opposed to normal C57BL/6J controls is not in agreement with the general observation that albino mice are inferior to pigmented mice on a variety of learning and performance tasks [ 13,16]. Winston and Lindzey [34] suggest that action of the c allele [ 14] contributes to inferior water escape and avoidance learning for albino mice. They also point out that it has not been conclusively demonstrated that behavioral correlates of albinism are reflective of multiple phenotypic effects of the c allele rather than effects of other closely linked alleles. The battery of RI strains contains both pigmented and albino mice. Uniformly poor or superior performance for the albino mice would provide additional evidence for the influence of the albino gene on performance extremes. The examination of single gene, or major gene, effects on behavior and attempts to establish linkage are of particular importance to the ultimate understanding of the relationship between genes and physiological correlates of behavior. The establishment of linkage is a first step in the tracing of physiological and structural characteristics which mediate the influence of major genes on behavioral phenomena.
ELIAS AND ELEFTHERIOU METHOD
Animals Animals were 218 experimentally naive male mice ranging in age from 8 to 10 weeks for each group. Table 1 shows the number of animals and mean body weight for each group. All mice were maintained on food and water ad lib during the course of the experiment and the day night cycle was 12 hours light and 12 hours dark. The seven RI strains will be designated as follows: CXBD (D), CXBE (E), CXBG (G), CXBH (H), CXBI (I), CXBJ (J), and CXBK (K). Derivation o f the R I strains. The mice used in the experiment represented a sample of the population of RI strains maintained and bred in BEE's animal colony at the Jackson Laboratory. Derivation of the population of RI strains is illustrated in Fig. 1. Two unrelated, but highly inbred progenitor strains, BALB/cBy (C) and C57BL/6J (B6) were crossed to produce F 1 hybrids. Beginning with the F 2 (F 1 X F 1) generation, the RI strains were maintained under a strict regimen of inbreeding for more than 40 generations. This procedure progressively fixes the chance recombination of genes as inbreeding continues and as full homozygosity is approached. The battery of strains which results from this procedure may be considered a replicable recombinant population. A battery of congenic lines has been developed for linkage testing under circumstances in which the RI strain distribution pattern suggests that a single gene may influence the physiological or behavioral phenotype of interest [1]. These lines were not utilized in the present e x p e r i m e n t because strain distribution patterns were inconsistent with any single gene hypothesis.
Apparatus and Procedure The maze has been described in detail in a previous paper [28]. It was constructed of sheet metal painted flat grey. It had rounded arms containing escape ladders hidden from view from the start and choice points. The maze had no top or bottom and it was immersed to half its depth in a tank of water maintained at room temperature (24 -+ 2°C). The top-bottom and front-back orientation of the maze was reversed randomly between blocks of trials in order to control for visual cues. There were two stages of testing: habituation and discrimination learning. Nine habituation trials were given in order to familiarize the mouse with the maze prior to introduction of the discrimination task, to determine spontaneous turn preference, and to obtain an indication of swimming times-to-escape prior to introduction of the discrimination task. For habituation the animals could escape from either arm of the maze and a record was kept of turn direction and escape time. For the discrimination learning task, the correct arm of the maze was separately defined for each animal as that which was chosen on less than half of the habituation trials. The ladder was removed from the incorrect arm and replaced with a dummy ladder which did not permit escape. Mice were given one trial per day until they had completed thirty trials. The correction method was used for training on the spatial discrimination task. For each daily session, the animal was placed in the maze at the choice point and permitted to make incorrect turns until it made the correct turn, found the ladder, and escaped. Each turn into the
GENETICS OF MAZE DISCRIMINATION LEARNING
835 TABLE 1
MEANS AND STANDARD ERRORS OF THE MEAN AND CONTRASTS BETWEEN MEANS FOR BODY WEIGHT* (G) Group
CB6F l
B6CFt
G
K
C
J
Mean
30.3
27.7
26.1
26.0
25.5
24.9
SEM
-+0.4
-+0.8
-+0.7
-+0.6 -+0.6 -+0.8 -+0.4 -+0.3 -+0.5 -+0.4 -+0.5
8
10
30
Nt
26
20
B6
H
2 4 . 3 24.0
24
21
16
I
D
E
22.8
22.5
22.0
24
21
18
*Lines connect nonsignificant differences 07>0.05). tNumber of subjects per group
P
B A LB/c By
100'
C57BL/6By
iI
9oi
cf
FI
¢"sI
x
l xALBINO
80~ _
"~
7oi
l
/
PIGMENTED
RECIPROCALFI HYBRIDS
/
/
Fa
50 cL
o 401 of
oj °
30 2OI 16CF,
F39
CXBD
C X B E CXBG CXBH CXBI
G X B J CXBK
B1 B~ B'a
~ Bi B'2 ~3
~
Bi a~ B~
HABITUATION(H) AND BLOCKS(B) OF TRIALS
FIG. 1. A schematic diagram of the deprivation of the RI strains by D. W. Bailey (1972). incorrect arm from the choice point during a single daily session was recorded as an error. Performance measures were mean time-to-escape and mean errors. Uniform illumination (3,8 ft. -c) of the testing area was provided by a fluorescent light mounted 6.5 feet above the apparatus.
Design and Data Analyses In order to reduce trial-to-trial performance variability and for clarity of presentation, trials were divided into three blocks of 10 each. This blocking of trials was based on results of previous experiments [ 11] which indicated that performance improves after 10 sessions, and reaches asymptote after 20 sessions when only one trial (session) per day is given. An I 1 (Groups) X 3 (Blocks) weighted means analysis of variance (ANOVAs) were done for the discrimination scores: latency of escape and errors. These o v e r a l l a n a l y s e s w e r e followed by between-groups (ANOVAs) for each block of trials and between-blocks (ANOVAs) for each group of animals. These single factor ANOVAs were then followed by multiple contrasts performed with the Tukey test [32] and the Student Newman Keuls multiple range test [32].
FIG. 2. Mean escape times for habituation trials and for three blocks of discrimination trials for the 11 strains used. RESULTS Figure 2 shows mean escape time for habituation trials and for the three blocks of discrimination trials. There was a significant difference among groups (strains and hybrids) for the habituation stage of learning, F(10,270) = 8.32, p
Genetic Analysis of Water Maze Discrimination Learning for Mus musculus: Polygenes and Albinism M E R R I L L F. E L I A S 3
Department o f Psychology and All-University, Gerontology Center, Syracuse University Syracuse, N Y 13210 AND B A S I L E. E L E F T H E R I O U
The Jackson Laboratory, Bar Harbor, ME 04609
(Received 15 O c t o b e r 1974) ELIAS, M. F. AND B. E. ELEFTHERIOU. A genetic analysis of water maze discrimination learning for Mus musculus: polygenes and albinism. PHYSIOL. BEHAV. 14(6) 833-838, 1975. - Recombinant inbred strains, their progenitor strains and reciprocal F 1 hybrids were given thirty spatial discrimination learning trials in the water maze. The pattern of RI strains in relation to the reciprocal F 1 hybrids and the progenitor strains, and differences among strains, suggested that different sets of genes affect the various components of water maze discrimination learning. There was no evidence that the albino gene influenced either inferior or superior performance. Pigmented strains performed as well as, or better than, some albino strains, and albino strain BALB/cBy differed significantly from albino strains CXBG and CXBI. The two reciprocal F 1 hybrids differed for Blocks 2 and 3 of discrimination trials when swimming time was used a dependent variable. Thus, it appears that performance of the F 1 hybrids was influenced by paternal effects, maternal effects, or paternal maternal interactions. Albino gene
Discrimination learning
Single gene hypothesis
R E S U L T S of m a n y e x p e r i m e n t s (e.g. [3, 8, 17, 2 7 ] ) with inbred strains o f mice indicate that differences in g e n o t y p e are associated with differences in maze learning ability. More recent e x p e r i m e n t s have been c o n c e r n e d w i t h the m o d e o f inheritance o f maze learning ability. Comparisons of F 1 hybrids and their parental strains have provided evidence for a d o m i n a n t or o v e r d o m i n a n t m o d e o f inheritance depending on the specific parental strains which were crossed to p r o d u c e the F 1 hybrids [4, 18, 3 3 ] . In a d d i t i o n to use of F 1 hybrids for the study o f m o d e o f inheritance, m u t a n t mice have been used as a means o f evaluating the c o n t r i b u t i o n o f single genes to maze learning ability. F o r example, coat color m u t a n t s and neurological mutants segregating on a C 5 7 B L background have been tested for maze learning with tas .ks involving f o o d reward and water escape [ 2 1 ] . Albino (cJ) and u n d e r w h i t e (ww)
Maze learning
Recombinant inbred strains
mice m a d e fewer errors than normal C 5 7 B L / 6 J mice. This finding p r o m p t e d the conclusion that single genes m a y contribute to a considerable e x t e n t to additive and d o m i n a n c e genetic variance for maze learning. However, a question has been raised regarding the possibility that single gene effects observed for coat color m u t a n t and neurological m u t a n t m i c e may reflect trivial pleiotropic effects involving peripheral mechanisms [30] or may constitute the behavi o r a l manifestation of severe central nervous system disorder [ 10]. The value o f single gene m u t a n t research and the issue of triviality of peripheral nervous system effects has been debated [25,30] and the value of single gene research has been aptly d e f e n d e d [ 2 5 ] . Nevertheless, alternative m e t h o d s for testing single gene h y p o t h e s e s and for establishing linkage between a behavioral p h e n o m e n o n and a
1This research was supported in part by NIH research grants HD 05860 to BEE and HD 08220 to MFE, both from the National Institute of Child Health and Human Development. We gratefully acknowledge the use of computer services provided by Mr. Robert Rupp (deceased) of the Jackson Laboratory. 2The Jackson Laboratory is fully accredited by the American Association for Accreditation of Laboratory Animal Care. 3 Performed while MFE was a Visiting Investigator at the Jackson Laboratory. 833
834 specific chromosome segment may be of considerable value. This is particularly true if these alternate techniques do not depend on the use of mutant stocks which display obvious structural or physiological defects of the central or peripheral nervous system. Bailey [1] has provided a method for exploring the possibility of single gene (or major gene) effects and, under appropriate circumstances, for establishing linkage. The Recombinant Inbred (RI) Strains have been utilized quite effectively in the analysis of gene systems involved in endocrine, behavioral, and immunological phenomena [2, 6, 19]. The purpose of the present experiment was to apply this method of genetic analysis to the study of water maze discrimination learning. Water maze learning was chosen for several reasons. It is a highly popular learning task for the mouse. It was designed specificially for studies of motivation and learning in the mouse [28], and it has been used in studies dealing with effects of lesions [9], neurological deficit [ 10], brain weight [ 7, 31 ], motivation [ 28,29 ], and sustained high blood pressure on behavior [ 11 ]. Significant positive genetic correlations between water maze learning and shock avoidance have been reported for F 3 mice and their F 2 parents [18], and for two single gene mutants [21]. Based on this finding, Oliverio and Messeri [21] suggest that pleiotropic effects of the same gene(s) influence both maze learning and shock avoidance learning or that, alternatively, both behavioral phenomena are affected by a common underlying physiological mechanism. In view of the correlation between avoidance learning and maze learning and the recent demonstration of a major, or single gene effects, for avoidance learning in a number of studies [6, 23, 24], it seemed reasonable to explore the possibility of a major gene effect for water maze discrimination learning with the genetic analysis technique made possible by Bailey's RI strains. In addition to the questioning of a major gene effect on maze learning with a new method of genetic analysis, the RI strains may be used to determine the extent to which the albino gene is, in fact, directly responsible for performance extremes observed for inbred albino strains. Oliverio and Messeri's [21 ] finding of better learning for albino mutant mice (c i) as opposed to normal C57BL/6J controls is not in agreement with the general observation that albino mice are inferior to pigmented mice on a variety of learning and performance tasks [ 13,16]. Winston and Lindzey [34] suggest that action of the c allele [ 14] contributes to inferior water escape and avoidance learning for albino mice. They also point out that it has not been conclusively demonstrated that behavioral correlates of albinism are reflective of multiple phenotypic effects of the c allele rather than effects of other closely linked alleles. The battery of RI strains contains both pigmented and albino mice. Uniformly poor or superior performance for the albino mice would provide additional evidence for the influence of the albino gene on performance extremes. The examination of single gene, or major gene, effects on behavior and attempts to establish linkage are of particular importance to the ultimate understanding of the relationship between genes and physiological correlates of behavior. The establishment of linkage is a first step in the tracing of physiological and structural characteristics which mediate the influence of major genes on behavioral phenomena.
ELIAS AND ELEFTHERIOU METHOD
Animals Animals were 218 experimentally naive male mice ranging in age from 8 to 10 weeks for each group. Table 1 shows the number of animals and mean body weight for each group. All mice were maintained on food and water ad lib during the course of the experiment and the day night cycle was 12 hours light and 12 hours dark. The seven RI strains will be designated as follows: CXBD (D), CXBE (E), CXBG (G), CXBH (H), CXBI (I), CXBJ (J), and CXBK (K). Derivation o f the R I strains. The mice used in the experiment represented a sample of the population of RI strains maintained and bred in BEE's animal colony at the Jackson Laboratory. Derivation of the population of RI strains is illustrated in Fig. 1. Two unrelated, but highly inbred progenitor strains, BALB/cBy (C) and C57BL/6J (B6) were crossed to produce F 1 hybrids. Beginning with the F 2 (F 1 X F 1) generation, the RI strains were maintained under a strict regimen of inbreeding for more than 40 generations. This procedure progressively fixes the chance recombination of genes as inbreeding continues and as full homozygosity is approached. The battery of strains which results from this procedure may be considered a replicable recombinant population. A battery of congenic lines has been developed for linkage testing under circumstances in which the RI strain distribution pattern suggests that a single gene may influence the physiological or behavioral phenotype of interest [1]. These lines were not utilized in the present e x p e r i m e n t because strain distribution patterns were inconsistent with any single gene hypothesis.
Apparatus and Procedure The maze has been described in detail in a previous paper [28]. It was constructed of sheet metal painted flat grey. It had rounded arms containing escape ladders hidden from view from the start and choice points. The maze had no top or bottom and it was immersed to half its depth in a tank of water maintained at room temperature (24 -+ 2°C). The top-bottom and front-back orientation of the maze was reversed randomly between blocks of trials in order to control for visual cues. There were two stages of testing: habituation and discrimination learning. Nine habituation trials were given in order to familiarize the mouse with the maze prior to introduction of the discrimination task, to determine spontaneous turn preference, and to obtain an indication of swimming times-to-escape prior to introduction of the discrimination task. For habituation the animals could escape from either arm of the maze and a record was kept of turn direction and escape time. For the discrimination learning task, the correct arm of the maze was separately defined for each animal as that which was chosen on less than half of the habituation trials. The ladder was removed from the incorrect arm and replaced with a dummy ladder which did not permit escape. Mice were given one trial per day until they had completed thirty trials. The correction method was used for training on the spatial discrimination task. For each daily session, the animal was placed in the maze at the choice point and permitted to make incorrect turns until it made the correct turn, found the ladder, and escaped. Each turn into the
GENETICS OF MAZE DISCRIMINATION LEARNING
835 TABLE 1
MEANS AND STANDARD ERRORS OF THE MEAN AND CONTRASTS BETWEEN MEANS FOR BODY WEIGHT* (G) Group
CB6F l
B6CFt
G
K
C
J
Mean
30.3
27.7
26.1
26.0
25.5
24.9
SEM
-+0.4
-+0.8
-+0.7
-+0.6 -+0.6 -+0.8 -+0.4 -+0.3 -+0.5 -+0.4 -+0.5
8
10
30
Nt
26
20
B6
H
2 4 . 3 24.0
24
21
16
I
D
E
22.8
22.5
22.0
24
21
18
*Lines connect nonsignificant differences 07>0.05). tNumber of subjects per group
P
B A LB/c By
100'
C57BL/6By
iI
9oi
cf
FI
¢"sI
x
l xALBINO
80~ _
"~
7oi
l
/
PIGMENTED
RECIPROCALFI HYBRIDS
/
/
Fa
50 cL
o 401 of
oj °
30 2OI 16CF,
F39
CXBD
C X B E CXBG CXBH CXBI
G X B J CXBK
B1 B~ B'a
~ Bi B'2 ~3
~
Bi a~ B~
HABITUATION(H) AND BLOCKS(B) OF TRIALS
FIG. 1. A schematic diagram of the deprivation of the RI strains by D. W. Bailey (1972). incorrect arm from the choice point during a single daily session was recorded as an error. Performance measures were mean time-to-escape and mean errors. Uniform illumination (3,8 ft. -c) of the testing area was provided by a fluorescent light mounted 6.5 feet above the apparatus.
Design and Data Analyses In order to reduce trial-to-trial performance variability and for clarity of presentation, trials were divided into three blocks of 10 each. This blocking of trials was based on results of previous experiments [ 11] which indicated that performance improves after 10 sessions, and reaches asymptote after 20 sessions when only one trial (session) per day is given. An I 1 (Groups) X 3 (Blocks) weighted means analysis of variance (ANOVAs) were done for the discrimination scores: latency of escape and errors. These o v e r a l l a n a l y s e s w e r e followed by between-groups (ANOVAs) for each block of trials and between-blocks (ANOVAs) for each group of animals. These single factor ANOVAs were then followed by multiple contrasts performed with the Tukey test [32] and the Student Newman Keuls multiple range test [32].
FIG. 2. Mean escape times for habituation trials and for three blocks of discrimination trials for the 11 strains used. RESULTS Figure 2 shows mean escape time for habituation trials and for the three blocks of discrimination trials. There was a significant difference among groups (strains and hybrids) for the habituation stage of learning, F(10,270) = 8.32, p
