182 Brain Research, 425 (1987) 182-185 Elsevier

BRE22568

Radial-maze performance and structural variation of the hippocampus in mice: a correlation with mossy fibre distribution

Wim E. Crusio 1, Herbert Schwegler I and Hans-Peter Lipp 2 lInstitut fiir Humangenetik und Anthropologie, Universitiit Heidelberg, Heidelberg (F. R. G.) and 2Anatomisches Institut,

Ziirich (Switzerland)

(Accepted 28 July 1987)

Key words: Learning; Radial-maze; Hippocampus; Mossy fiber; Mouse

Twenty-four male mice, belonging to 8 different inbred strains, were tested in an 8-arm radial maze. Clear strain differences were found for performance on the third day of training, which correlated very strongly with the size of the hippocampal intra- and infrapy- ramidal mossy fibre (iip-MF) terminal fields. These results, combined with those from earlier experiments, indicate that genetic varia- tions of the iip-MF projection influence processes that determine behavioural abilities of mice.

In recent years, the problem of hippocampal func-

tioning has produced a blossoming variety of theo- ries 16. Among the most influential are Olton's work- ing-memory 12 and O'Keefe and Nadel's cognitive-

map 11 theories. However, although different theo-

ries offer different explanations, most of them ac- knowledge a critical role of hippocampal circuitry for

learning and memory processes.

To a large extent, all theories of hippocampal func-

tion in rodents are based on the results of lesion stud-

ies. Such studies suffer from methodological prob- lems (see ref. 14 for a more detailed discussion), and often equate the dysfunction of the residual brain with hippocampal function. From our point of view,

the study of the behavioural consequences of natural-

ly occurring and heritable variations of hippocampal structures 2'5 offers a different and rather powerful

approach.

Several correlations between hipp0campal varia- tion and behaviour have been found, the first report-

ing a negative co-variation between the size of the in-

tra- and infrapyramidal mossy fibre (iip-MF) termi- nal field and two-way active-avoidance learning in mice and rats 18. Recently, we demonstrated co-varia-

tions between the size of the iip-MF and two spatial behaviours: performance in water mazes 17'22 and ha-

bituation in an open field 3'7. Based on these observa- tions, we hypothesised 17 that variations of the iip-MF

projections may influence hippocampal processing and, through this, spatial learning 11, working-mem- ory 12, timing 14, or memory consolidation 21.

Radial mazes are frequently used to study hippo-

campal function, and different procedures permit to

dissociate spatial and non-spatial memory capabil-

ities in rodents 9. Initially, mostly rats were subjected to these procedures. Conflicting evidence has been presented concerning the performance of mice in a radial maze: Mizumori et al. s reported that even af-

ter 20 days of training mice showed no improvement

at all, whereas other studies reported a satisfactory performance 13'15. Therefore, in this first, exploratory

experiment, we studied the performance of 8 differ-

ent inbred strains of mice in a relatively simple radial- maze learning task.

Three male mice, aged 12-16 weeks, were taken

from each one of the highly inbred strains BA, C57BL/6J, C57BR/cdJ, BALB/cJ, DBA/2J ,

C3H/HeJ, CPB-K, and NZB/BiNJ 2°. Rearing and

Correspondence: W.E. Crusio, Institut fiir Humangenetik und Anthropologie, Universitfit Heidelberg, Im Neuenheimer Feld 328, D-6900 Heidelberg, F.R.G.

0006-8993/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

housing conditions have been descr ibed elsewhere 17.

A n 8-arm radial-maze was used. The circular central

part measured 22 cm in diameter . The end of each

arm (25 cm long, 6 cm high, 6 cm wide) held a perfo-

ra ted par t i t ion behind which some fresh food pellets

were deposi ted. In this way, the presence or absence

of a reward could not be smelled by the animals. The

reward (pieces of food pellets of approximate ly 10

mg) was placed behind a small barr ier , prevent ing

the animal from seeing whether a specific arm was

still ba i ted or not. Al l 8 arms were baited. The walls

and floor of the maze were made of opaque polyvi-

nylchloride. The maze was covered by t ransparent

plexiglas, presumably enabling the mice to perceive

spatial cues from the surrounding, in which the maze

was always posi t ioned in the same way. Twenty-four

h pr ior to training, animals received a 10-min habit-

uat ion trial (with unbai ted arms). Subsequent ly , the

animals were food, but not water , deprived. Dur ing

training body weight was kept at 8 0 - 9 0 % of the

pre-test body weight. Animals were t ra ined on 3 con-

secutive days. On the first day, trials were te rminated

after 15 min or after the animal had eaten all 8 re-

wards, whatever came first. On the second and third

day no t ime limit was set. In order to have an arm en-

trance scored, animals had to walk into the arm for

over one half of its length. A n error was noted if an

animal en tered an arm in which the bait had been ea-

ten before (or if the animal did not eat the pellet) .

The exper imenta l variable was the number of errors

made on the third day ra ther than the more usual

number of novel arms visited on the first 8 choices

(e.g. ref. 15). The rat ionale was that , for the la t ter

variable, chance performance will lead to a score

around 5.3 (as, indeed, was realised by our worst per-

forming strains), whereas a maximal score of only 8 is

possible. These floor and ceiling effects great ly limit

the possible range of variat ion, diminishing the

chance of finding co-variat ions between per formance

and some other variable. As can be seen in Table I,

the mean number of errors made on the third day of

training ranges from 1.3 to 8.7. The best per forming

strain was C3H/HeJ, CPB-K was the worst.

In order to relate the behavioural variat ion with

the heri table variat ion in the size of the i ip-MF pro-

ject ion, we used strain means ob ta ined in an ear l ier

study 17 in which o ther animals belonging to the same

strains were used. Full procedura l details can be

183

TABLE I

Means + S. E.M. of the behavioural and hippocarnpal variables measured in 8 inbred mouse strains

Strain Errors iip -MF*

C3H/HeJ 1.3 + 0.3 4.09 + 0.62 BALB/cJ 3.0 + 1.7 2.58 + 0.06 C57BL/6J 5.0 + 1.7 3.86 + 0.30 C57BIUcdJ 5.0 + 0.6 2.39 + 0.15 BA 6.7 + 0.9 2.32 + 0.13 DBA/2J 8.0 + 2.3 1.54 + 0.30 NZB/BiNJ 8.0 + 1.7 1.49 + 0.22 CPB-K 8.7 + 3.5 1.63 + 0.34

* Area of the hippocampal intra- and infrapyramidal mossy fi- bre terminal fields (expressed as percentage of the regio in- ferior).

found elsewhere TM. Briefly, 4 males of each strain

were perfused and processed for T imm's silver-sul-

fide stain for heavy metals , according to the method

of Danscher and Z i m m e r 4, visualizing the zinc-con-

taining mossy fibre terminals. For morphomet ry , 5

horizontal sections were sampled randomly from

either the right or the left h ippocampus. Sampling at

the midsep to tempora l level s tar ted immedia te ly be-

low the most ventral extension of the septal pole of

the fascia dentata , taking every second section. The

hippocampal fields of regio infer ior and the hilus of

the denta te area (= CA3/CA4) were drawn by means

of a project ion microscope. Areas were measured on

a graphics tablet connected to a desk computer . The

size of the i ip-MF pro jec t ion was expressed as per-

centages of C A 3 + C A 4. Inspect ion of Table I reveals

a strong negative correla t ion be tween error number

and the size of the i ip-MF pro jec t ion (r s = -0 .88 , df

= 6, P < 0.01; Spearman rank correla t ion correc ted

for ties19). If the number of novel arms visi ted on the

first 8 choices was used as behavioural variable, the

same trend was revealed (r s = 0.62, df = 6, P <

0.06).

Our data suggest that in mice her i table variat ions

of the i ip-MF projec t ions corre la te with per formance

in our radial -maze task. However , it is not clear yet

what exactly has been learned by the mice. We sug-

gest that our mice used a non-spat ial , kinesthetic

strategy to solve the present radia l -maze task. First ,

in our exper iment , no guillotine doors were present

to enforce an interchoice interval upon the experi-

mental subjects. Bolhuis et al. 1 showed that the pres-

ence or absence of such doors de te rmined which kind

184

%

7 0 -

60

50

40

30

2 0 -

1 0 -

0 -

45

I

I

90 135 180

INTERCHOICE ANGLE

Fig. 1. Occurrence of various angles between successively cho- sen arms of the radial-maze on the third day of training. Data presented are the results of all mice (n = 24) pooled over strains, because similar distributions of interchoice-angle fre- quencies were found for all strains. Bars represent one S.E.M.

of behavioural strategies rats used to solve the maze

problem. With doors present , an apparent ly spatial ly

or iented strategy was employed , whereas with doors

absent a clear preference existed for adjacent arms.

Thus, we also de te rmined , for each choice in the final

trial, the distance to the previously chosen arm (in

terms of the angle be tween successive chosen arms;

Fig. 1). The results indicate that in the absence of a

confinement p rocedure mice, too, show a preference

for adjacent arms. Second, the use of a closed instead

of an open radial -maze may also facilitate kinesthetic

strategies 13. This p robab ly explains the differences

between t h e results of the present s tudy for the

strains C57BR/cd, C57BL/6, and BALB/c , and those

repor ted by others xS. Third, our bes t -performing

strain, C3H/HeJ, carries the retinal degenera t ion

(rd) allele 2°. Al though not causing blindness, the vis-

ual acuity of animals homozygous for rd is severely

impaired 1°. It is therefore highly unlikely that these

mice used extramaze spatial cues in solving the maze-

task. These animals, as do bl inded rats 23, most proba-

bly used a kinesthetic strategy to solve the radial-

maze problem. Thus, the degree of spontaneous re-

sponse chaining appears to be an impor tant determi-

nant of the radial-maze per formance of our mice.

This present result agrees with previous findings

that variat ions of the i ip-MF project ions corre la te

with learning in a complex water-maze 17, reversal

learning in a Morris- type water-maze 22, choice opti-

misation in a Y-maze discrimination task 6, and be-

haviour in an open-f ield 3. Taken together , these

studies strongly suggest that variat ions of the i ip-MF

project ion influence processes that de termine behav-

ioural abilities of mice.

We thank Profs. F. Vogel and M. Schachner for

reading the manuscript and Prof. W. Buselmaier for

providing all necessary facilities. W.E .C . was sup-

por ted by an Alexander -von-Humbold t s t ipend,

H.S. by the Deutsche Forschungsgemeinschaf t

(Schw 252), and H.P .L . by Gran t SNF 3.041-0.84.

1 Bolhuis, J.J., Bijlsma, S. and Ansmink, P., Exponential de- cay of spatial memory of rats in a radial maze, Behav. Neu- ral Biol., 46 (1986) 115-122.

2 Crusio, W.E., Genthner-Grimm, G. and Schwegler, H., A quantitative-genetic analysis of hippocampal variation in the mouse, J. Neurogenet., 3 (1986) 203-214.

3 Crusio, W.E. and Schwegler, H., Hippocampal mossy fiber distribution covaries with open-field habituation in the mouse, Behav. Brain Res., in press.

4 Danscher, G. and Zimmer, J., An improved Timm sulfide silver method for light and electron microscopic localization of heavy metals in biological tissues, Histochemistry, 55 (1978) 27-40.

5 Heimrich, B., Schwegler, H. and Crusio, W.E., Hippocam- pal variation between the inbred mouse strains C3H/HeJ and DBA/2: a quantitative-genetic analysis, J. Neuroge- net., 2 (1985) 389-401.

6 Lipp, H.-P., Schwegler, H. and Hausheer-Zarmakupi, Z., The infrapyramidal mossy fiber projection in the hippocam- pus of the mouse: a positive correlation with Y-maze avoid- ance learning, Soc. Neurosci. Abstr., 12 (1986) 751.

7 Lipp, H.-P., Schwegler, H., Heimrich, B., Cerbone, A. and Sadile, A.G., Strain-specific correlations between hip- pocampal structural traits and habituation in a spatial nov- elty situation, Behav. Brain Res., 24 (1987) 111-123.

8 Mizumori, S.J.Y., Rosenzweig, M.R. and Kermisch, M.G., Failure of mice to demonstrate spatial memory in the radial maze, Behav. Neural Biol., 35 (1982) 33-45.

9 Nadel, L. and MacDonald, L., Hippocampus: cognitive map or working memory?, Behav. Neural Biol., 29 (1980) 405-409. Nagy, Z.M. and Misanin, J.R., Visual perception in the ret- inal degenerate C3H mouse, J. Comp. Physiol. Psychol., 72 (1970) 306-310.

10

11 O'Keefe, J. and Nadel, L., The Hippocampus as a Cogni- tive Map, Clarendon, Oxford, 1978, 570 pp.

12 Olton, D.S., Handelmann, G.E. and Walker, J.A., Hippo- campus, space and memory, Behav. Brain Sci., 2 (1979) 313-365.

13 Pico, R.M. and Davis, J.L., The radial maze performance of mice: assessing the dimensional requirements for serial order memory in animals, Behav. Neural Biol., 40 (1984) 5-26.

14 Rawlins, J.N.P., Associations across time: the hippocam- pus as a temporary memory store, Behav. Brain Sci., 8 (1985) 479-496.

15 Reinstein, D.K., DeBoissiere, T., Robinson, N. and Wurt- man, R.J., Radial maze performance in three strains of mice: role of the fimbria/fornix, Brain Research, 263 (1983) 172-176.

16 Schmajuk, N.A., Psychological theories of hippocampal function, Physiol. Psychol., 12 (1984) 166-183.

17 Schwegler, H., Crusio, W.E., Lipp, H.-P. and Heimrich, B., Water-maze learning in the mouse correlates with vari- ation in hippocampal morphology, Behav. Genet., 18

185

(1988) (in press). 18 Schwegler, H. and Lipp, H.-P., Hereditary covariation of

neuronal circuitry and behavior: correlation between the proportions of hippocampal synaptic fields in the regio infe- rior and two-way avoidance in mice and rats, Behav. Brain Res., 7 (1983) 1-38.

19 Siegel, S., Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill, New York, 1956, 312 pp.

20 Staats, J., Standardized nomenclature for inbred strains of mice: eighth listing, Cancer Res., 45 (1985) 945-977.

21 Teyler, T.J. and DiScenna, P., The hippocampal memory indexing theory, Behav. Neurosci., 100 (1986) 147-154.

22 Wolfer, D.P., Lipp, H.-P., Leisinger-Trigona, M.-C. and Hausheer-Zarmakupi, Z., Correlations between hippo- campal circuitry and behavior: infrapyramidal mossy fibers and navigation in a Morris water maze, Acta Anat., in press.

23 Zoladek, L. and Roberts, W.A., The sensory basis of spa- tial memory in the rat, Anim. Learn. Behav., 6 (1978) 77-81.