Estrogen Increases Locomotor Activity in Mice throughEstrogen Receptor �: Specificity for the Type of Activity
SONOKO OGAWA, JOHNNY CHAN, JAN-ÅKE GUSTAFSSON, KENNETH S. KORACH, AND
DONALD W. PFAFF
Laboratory of Neurobiology and Behavior, Rockefeller University (S.O., J.C., D.W.P.), New York, New York 10021;Department of Medical Nutrition, Karolinska Institute (J.-Å.G.), S-14186 Huddinge, Sweden; and Laboratory ofReproductive and Developmental Toxicology, National Institute of Environmental and Health Sciences (K.S.K.),Research Triangle Park, North Carolina 27709
Estrogens are known to increase running wheel activity ofrodents primarily by acting on the medial preoptic area(mPOA). The mechanisms of this estrogenic regulation of run-ning wheel activity are not completely understood. In partic-ular, little is known about the separate roles of two typesof estrogen receptors, ER� and ER�, both of which are ex-pressed in mPOA neurons. In the present study the effects ofcontinuous estrogen treatment on running wheel activitywere examined in male and female mice specifically lackingeither the ER� (�ERKO) or the ER� (�ERKO) gene. Mice weregonadectomized and 1 wk later implanted with either a lowdose (16 ng/d) or a high dose (160 ng/d) of estradiol benzoate(EB) or with a placebo control pellet. Home cage runningwheel activity was recorded for 9 d starting 10 d after EBimplants. The same mice were also tested for open field ac-tivity before and after EB implants. In both female and male�ERKO mice, running wheel activity was not different fromthat in corresponding wild-type (�WT) mice in placebo controlgroups. In both females and males it was increased by EB only
in �WT, not �ERKO, mice. In �ERKO mice, on the other hand,both doses of EB equally increased running wheel activity inboth sexes just as they did in �WT mice. Absolute numbers ofdaily revolutions of EB-treated groups, however, were signif-icantly lower in �ERKO females compared with �WT females.Before EB treatment, gonadectomized �ERKO female weresignificantly less active than �WT mice in open field tests,whereas �ERKO females tended to be more active than �WTmice. In male mice there were no effect of ER� or ER� geneknockout on open field activity. Unlike its effect on runningwheel activity, EB treatment induced only a small increase inopen field activity in female, but not male, mice. These find-ings indicate that 1) in both sexes estrogenic regulation ofrunning wheel activity is primarily mediated through theER�, not the ER�; and 2) hormone/genotype effects are spe-cific to the type of locomotor activity (i.e. home cage runningwheel activity and open field activity) measured. (Endocri-nology 144: 230–239, 2003)
IT IS WELL established that estrogen regulates runningwheel activity in female and male rats. Gonadally intact
female rats show the highest activity during proestrus, whenplasma levels of estrogen are elevated (1). In gonadectomizedfemale and male rats, it has been shown that estrogen treat-ment increases running wheel activity (2–5), although estro-gen failed to do so in ferrets of both sexes (6). A number oflesion (7), electrophysiological (8), and site-specific steroidimplant studies (9) revealed that the medial preoptic area(mPOA) is the primary brain site responsible for this behav-ioral effect of estrogen, although the anterior hypothalamicarea just posterior to mPOA may also be involved. Two typesof nuclear estrogen receptors, ER� and ER�, are both local-ized in the mPOA as shown by in situ hybridization as wellas immunocytochemical studies (10–14). It is not known,however, which ER is responsible for estrogenic regulationof running wheel activity. In the present study we examinedthe effects of estrogen treatment on running wheel activityin male and female mice specifically lacking either ER�(�ERKO) or ER� (�ERKO) genes.
Previous studies have shown that estrogen also regulatesopen field activity in female rats (15). It is generally assumed
that estrogen may control open field activity and runningwheel activity through different mechanisms. Open fieldactivity measured in an unfamiliar testing apparatus repre-sent an animal’s response to a new environment, whereasrunning wheel activity is the animal’s home cage activity.Brain site-specific estrogen implants effective in increasingrunning wheel activity in gonadectomized female rats failedto increase open field activity (9). These findings in ratssuggest that estrogenic regulation of running wheel activitymay be primarily localized in the mPOA, whereas that ofopen field activity may involve more than one specific brainsite. Furthermore, a recent study in C57BL/6J mice showedthat systemic estrogen treatments stimulating running wheelactivity did not necessarily increase or could even decreasecertain measurements of open field activity (16). It is as-sumed that activity measured in open field tests may beconfounded by a number of factors, such as fear and emo-tionality, depending on testing conditions, e.g. lighting in-tensity and size of the apparatus (17). Indeed, it is reportedthat open field activity levels were highly correlated withbehavioral responses in other fear-related behavioral testparadigms, such as fear conditioning and dark/light tran-sition tests, rather than with home cage running wheel ac-tivity levels (16). Taken together, these findings suggest thattwo types of activity may be differentially regulated by es-trogen via two types of ERs.
Although the effects of estrogen are not yet known, ourprevious studies suggested that ER� gene disruption mightmodify baseline open field activity in a gender-dependentfashion. In gonadally intact male mice, ER�, but not ER�,gene disruption significantly increased open field activity(18, 19). In females, on the other hand, ER� gene disruptiontended to decrease both open field activity in gonadallyintact mice and chamber transitions in the dark/light tran-sition tests in both gonadally intact and gonadectomizedmice (20). In the present study we first pretested gonadec-tomized mice of both sexes for open field activity to assesssex- and/or gene-dependent effects of gene knockout. Wethen assigned mice to one of three treatment groups basedon activity levels and examined whether estrogen mightaffect running wheel and open field activities separately byusing the same mice in both tests.
Materials and MethodsMice
A total of 172 mice deficient in the ER� (�ERKO) or ER� (�ERKO)gene and their respective wild-type (WT) littermates of both sexes wereused. They were obtained from the �ERKO and �ERKO breeding col-onies maintained at Rockefeller University by mating heterozygousmale and female mice. Original breeding pairs (mixed background ofC57BL/6J and 129) were obtained from the NIEHS. Animal rooms weremaintained on a 12-h light, 12-h dark cycle at a constant temperature (22C). Experimental mice were group-housed with the same sex and mixedgenotype cage mates from weaning (21 d old) until used for the exper-iment. Food and water were available ad libitum throughout the studies.
Procedure
At the age of 9–11 wk, mice were gonadectomized (d 1; see Table 1).On the sixth day all mice were tested once for open field behavior (seebelow) and assigned to 1 of 3 treatment groups based on their activitylevels (see below) and body weights. The numbers of animals in eachtreatment group of each genotype are indicated in parentheses in Tables2 and 3. Two days later the mice were implanted with either a �-estradiol3-benzoate (EB) capsule with a dose of 16 or 160 ng/d or a placebocapsule (Innovative Research of America, Toledo, OH). At the sametime, they were weighed, single-housed in plastic cages (27 � 16 � 13cm), and given a phytoestrogen-free diet (AIN76A, Ralston Purina Co.,
St. Louis, MO). On the 15th day (1 wk after the implant) mice weretransferred to running wheel activity cages (Mini Mitter, Bend, OR) andmonitored daily for the number of wheel revolutions (d 16–25; datastarting on d 17 were included for the analysis). The day after the lastrecording, they were transferred again to regular single-house cages andtested for open field behavior twice 1 wk later (d 33 and 34).
Open field behavior tests
Mice were tested for 5 min in an open field apparatus (40.5 � 40.5 cm,30-cm high wall), which was illuminated with red light from the top atthe center of the apparatus. At the beginning of the test a mouse wasplaced gently in a corner square with its head facing the corner. Activitywas monitored automatically with infrared beams, and data were an-alyzed and stored using a Digiscan Analyzer and Digiscan software(Digiscan model RXYZCM, Accuscan Instruments, Columbus, OH). Thetotal horizontal and vertical moving distances (total activity), total hor-izontal moving distance (total distance), cumulative duration of hori-zontal moving (moving time), moving distance in the center area (centerdistance), and time spent in the center area (center time) were recordedfor each mouse. The first three measurements primarily indicated gen-eral exploratory activity levels in a novel environment, whereas the twomeasurements in the center area were interpreted to also be related tothe animals’ anxiety and fear levels. The center area was defined as thearea more than 1 in. away from the wall. After completion of the openfield test, mice were weighed and assigned to one of three treatmentgroups balanced in terms of total distance and body weight.
Running wheel activity
Mice were housed individually in plastic cages (32 � 17 � 14 cm)equipped with a running wheel (25-cm diameter; Mini Mitter). Eachwheel revolution was registered by a magnetic switch, which was con-nected to a counter. The number of revolutions was recorded daily for9 d as described above.
Statistics
Data from pretreatment behavioral tests were analyzed in each sexwith two-way ANOVAs for the main effects of gene (ER� vs. ER�) andgene knockout (KO vs. WT) and their interactions. Data from posttreat-ment behavioral tests were analyzed in each sex by three-way ANOVAsfor the main effects of gene, knockout, treatment group, and their in-teractions. Repeated measurement data were analyzed in each sex andgene with three-way ANOVA for the main effects of gene knockout(KO vs. WT) treatment group, test block (as a repeated measure), andtheir interactions. If applicable, post hoc one- or two-way ANOVAs wereperformed. Turkey’s test was used for post hoc pairwise comparisons.
ResultsPreestrogen treatment levels of open-field activity andbody weights
Gene disruption of ER� and ER� had opposite effects onopen field activity in gonadectomized female, but not male,
TABLE 1. Experimental procedure
Days Treatment/behavioral tests
1 GDX6 Pre-OFT8 Implant/Individual Housing
//15 Moved to RW apparatus1617 RW d 118 RW d 219 RW d 320 RW d 421 RW d 522 RW d 623 RW d 724 RW d 825 RW d 926 Removed from the RW apparatus//33 Post-OFT134 Post-OFT2
GDY, Gonadectomy; OFT, open field test; RW, running wheelactivity.
a Values are the mean � SEM; the number of animals is given inparentheses.
Ogawa et al. • Activity in �ERKO and �ERKO Mice Endocrinology, January 2003, 144(1):230–239 231
mice (Fig. 1). In four open field activity measurements, butnot in the measurement of center time, there were significantinteractions of gene and knockout [total activity: F(1,84) �9.05; P � 0.004; Fig. 1A; total distance: F(1,84) � 6.31; P �0.014; Fig. 1B; moving time: F(1,84) � 6.61; P � 0.012; Fig. 1C;center distance: F(1,84) � 8.11; P � 0.006; Fig. 1D]. �ERKOfemales were less active than �WT females, whereas �ERKOfemale mice tended to be more active than �WT females.
In males, on the other hand, there were no significanteffects of gene knockout in either ER� or ER� genes, exceptthat �ERKO mice tended to show shorter moving times than�WT. Instead, there were significant main effects of geneticbackground [total activity: F(1,81) � 5.21; P � 0.025; Fig. 1A;total distance: F(1,81) � 2.97; P � 0.089; Fig. 1B; centerdistance: F(1,81) � 5.05; P � 0.027 (Fig. 1D); center time:F(1,81) � 11.43; P � 0.001; Fig. 1E]. �WT and �ERKO wereless active than �WT and �ERKO mice.
Body weights of female mice were not different among thefour genotype groups. In males, there were significant maineffects of genetic background [F(1,81) � 14.60; P � 0.0003;Fig. 1F]. Overall, �WT and �ERKO were heavier than �WTand �ERKO mice.
Based on the results of open field tests, mice were assignedto one of three treatment groups, balanced in terms of totalmoving distance and body weight (Tables 2 and 3). Tables 2and 3 also indicate the numbers of animals in each treatmentgroup of each genotype.
Effects of estrogen on overall (running wheel, d 1–9)running wheel activity
EB treatment significantly increased running wheel activ-ity in both �WT [F(2,17) � 5.54; P � 0.014] and �WT[F(2,20) � 6.14; P � 0.009] female mice, although in theformer group only the higher dose of EB (160 ng/d) wassignificantly different from the placebo control (Fig. 2A).Deletion of ER� completely abolished the facilitatory actionof estrogen on running wheel activity. For �ERKO femalemice, there were no differences between the EB-treatedgroups and the placebo control group.
In �ERKO female mice, on the other hand, both doses ofEB treatment increased running wheel activity comparedwith the placebo control treatment [F(2,21) � 14.94; P �0.0002] as found in �WT mice. However, absolute numbersof daily revolutions in the EB-treated groups, but not theplacebo control group, were significantly lower in �ERKOcompared with �WT [effect of knockout in two EB groups
combined: F(1,28) � 4.36; P � 0.046] female mice. As a sidepoint, in placebo control groups there were no genotypedifferences in running wheel activity between �ERKO and�WT female mice (Fig. 2A). This is in contrast to the genotypedifferences found in pretreatment open field tests (Fig. 1).
In males EB treatment significantly increased runningwheel activity in both groups of WT mice (�WT and �WT;Fig. 2B), although overall �WT mice were more active than�WT mice [F(1,34) � 12.02; P � 0.0014]. Effects of geneknockout on the estrogen-inducible activity increase weredependent on which gene was deleted, as revealed by three-way ANOVA [gene � knockout � treatment: F(2,73) � 8.10;P � 0.0007]. Deletion of the ER� gene affected estrogenicregulation of running wheel activity [knockout � treatment:F(1,37) � 14.19; P � 0.0001]. That is, in �ERKO mice estrogenfailed to increase running wheel activity [treatment: P � NS],whereas in �WT mice estrogen significantly increased it[treatment: F(2,15) � 24.60; P � 0.0001].
In contrast, deletion of the ER� gene did not affect estro-genic regulation of running wheel activity [treatment:F(2,36) � 16.30; P � 0.0001; knockout and knockout � treat-ment: P � NS]. In �ERKO mice, both doses of EB treatmentsignificantly increased running wheel activity [treatment:F(2,17) � 14.97; P � 0.0002] as found in �WT mice [treat-ment: F(2,19) � 5.10; P � 0.017]. Unlike in females, there wasa tendency in the three genotypes of male mice (�WT, �WT,and �ERKO) in which EB increased activity for the lower EBdose (16 ng/d) to be slightly more effective than the higherEB dose (160 ng/d), although there was no statistical differ-ence between these two dose groups.
Time course of estrogen effects on running wheel activity
Effects of ER� and ER� gene knockout on estrogen-inducible running wheel activity were further analyzed ineach sex and gene by dividing the data into three blocks of3 d each (Figs. 3 and 4).
Female �ERKO vs. aWT
Regardless of treatment group and genotype, there was asteady increment in running wheel activity along days[block: F(2,68) � 25.45; P � 0.0001; block � treatment: P �NS; block � knockout � treatment: P � NS; Fig. 3A]. Overallgenotype differences were detected throughout three blocks[knockout: F(1,34) � 6.65; P � 0.014; block � knockout: P �NS]. Estrogen did not modify running wheel activity in
TABLE 3. Preimplant open field activity (total distance)
a Values are the mean � SEM; the number of animals is given in parentheses.
232 Endocrinology, January 2003, 144(1):230–239 Ogawa et al. • Activity in �ERKO and �ERKO Mice
�ERKO mice in all three blocks. In �WT female mice EBtreatment effectively increased running wheel activity in allthree blocks, although with the lower dose of EB, activity wasnot different from that in controls during the second block.
Female �ERKO vs. �WT
Overall, �ERKO mice were less active than �WT micethroughout the experiment [knockout: F(1,41) � 4.27; P �
0.045; block � knockout: P � NS; Fig. 3B]. Regardless ofgenotype, estrogen treatment became more effective towardthe end of the testing period [block � treatment: F(4,82) �2.98; P � 0.024; block � knockout � treatment: P � NS].
Male �ERKO vs. �WT
Changes in running wheel activity during all three testblocks varied according to the genotype and treatment group
FIG. 1. Effects of ER� and ER� gene disruption on open field activity in gonadectomized mice of both sexes. A, Total activity; B, total movingdistance; C, cumulative moving time in seconds; D, moving distance in the center area; E, cumulative time spent in the center area; F, bodyweight. *, P � 0.05 vs. respective WT; †, P � 0.10 vs. respective WT.
Ogawa et al. • Activity in �ERKO and �ERKO Mice Endocrinology, January 2003, 144(1):230–239 233
as revealed by the three-way interaction of knockout, treat-ment, and test block [F(4,74) � 3.66; P � 0.009; Fig. 4A]. In�ERKO mice all three groups of mice showed a steady in-crease in running wheel activity along test days (treatment �block: P � NS). In �WT mice, only EB-treated groups, not thecontrol group, showed an increase in running wheel activityalong test days [treatment � block: F(4,30) � 5.03; P � 0.003].
Male �ERKO vs. �WT
There were no genotype differences in the time course forestrogenic regulation of running wheel activity (blocks �knockout � treatment: P � NS; Fig. 4B). However, in bothgenotypes two doses of EB treatment produced their effectsin different time course [blocks � treatment: F(4,72) � 3.81;P � 0.0073]. During the first block the higher dose of EB wassignificantly less effective than the lower dose of EB in bothgenotypes (at � � 0.05), whereas during the second and thirdblocks the two doses were equally effective.
Effects of estrogen on open field activity
After the completion of running wheel tests, mice weretested twice for open field activity. Although mice weresignificantly more active on the first day than on the secondday, there were no interactions of test days with genotypes
and/or treatment groups. Therefore, mean values of the twotests were used for further analyses.
In female mice there were significant gene � knockoutinteractions in total activity [F(1,60) � 4.44; P � 0.039; datanot shown], total distance [F(1,60) � 11.91; P � 0.01; Fig. 5A],moving time [F(1,60) � 6.39; P � 0.014; data not shown], andcenter distance [F(1,60) � 5.99; P � 0.017; Fig. 5B], but not incenter time (data not shown). �ERKO mice were less activethan �WT, whereas �ERKO mice were not different from�WT mice. There were also overall treatment effects on totalactivity [F(2,60) � 4.00; P � 0.023; data not shown], totaldistance [F(2,60) � 3.79; P � 0.028; Fig. 5A], and moving time[F(2,60) � 3.35; P � 0.042; data not shown]. Mice treated withEB at a dose of 160 ng/d were significantly more active thanmice treated with either the low dose of EB (16 ng/d) orplacebo regardless of genotype (gene � knockout � treat-ment: P � NS).
In males only the main effect of gene, but not the maineffect of knockout or gene � knockout interaction, was sta-tistically significant in all five measurements of open fieldtests [total activity: F(1,64) � 5.59; P � 0.021; data not shown;total distance: F(1,64) � 5.46; P � 0.023; Fig. 5C; moving time:F(1,64) � 4.14; P � 0.046; data not shown; center distance:F(1,64) � 5.10; P � 0.027; Fig. 5D; center time: F(1,64) � 5.95;P � 0.018; data not shown]. �WT and �ERKO mice were
FIG. 2. Effects of ER� and ER� gene disruption onestrogenic regulation of running wheel activity. Meandaily revolutions during the 9-d test period wereshown for females (A) and males (B). , Placebo controlgroup; z, low dose of EB (16 ng/d) treatment group; f,high dose of EB (160 ng/d) treatment group. *, P � 0.05vs. the respective placebo control group; a, P � 0.05 vs.�WT in the respective treatment group.
234 Endocrinology, January 2003, 144(1):230–239 Ogawa et al. • Activity in �ERKO and �ERKO Mice
more active than �WT and �ERKO mice. Furthermore, un-like females, there were no significant main effects of treat-ment or interactions with gene and/or knockout in anymeasurement.
DiscussionRoles of ER� and ER� in estrogenic regulation of runningwheel activity
It was found in this study that estrogen greatly potentiatedrunning wheel activity in mice, as previously reported in rats(2–5). The present study demonstrates for the first time adifferential dependence on the ER� vs. the ER� gene in theestrogenic regulation of running wheel activity. In both sexesof mice ER� gene disruption completely abolished theestrogen-inducible increase in running wheel activity,whereas ER� gene disruption did not affect it. These resultssuggest that estrogen may facilitate running wheel activityprimarily by acting through ER�, presumably in the mPOA(7–9). As it is known that responsiveness to EB for the po-tentiation of running wheel activity is not affected by neo-
natal steroid manipulation (4), it is likely that ER� activationat the time of testing in adulthood is critical for estrogenicregulation of this behavior. The present study, however, doesnot completely rule out the possibility that the lack of EBeffects in �ERKO mice may also be due to the lack of ER�stimulation during perinatal development regardless of sex.
The present findings also suggest that activation of ER� byitself is not sufficient to increase running wheel activity inmice. This idea is consistent with the results of our studiesin which an ER�-specific compound (Ogawa et al., unpub-lished data) failed to increase running wheel activity inSwiss-Webster mice. However, this does not rule out thepossibility that ER� participates in the regulation of runningwheel activity. In fact, in female �ERKO mice, estrogen wasslightly less effective than in �WT female mice in terms ofabsolute number of daily revolutions after EB treatment.Therefore, it is possible that simultaneous activation of ER�and ER� might be necessary for complete estrogenic regu-lation of running wheel activity at least in female mice. Itshould be noted that ER� and ER� are both localized in the
FIG. 3. Time course of estrogenic regulation of running wheel activity in �WT and �ERKO (A) and �WT and �ERKO (B) female mice. Meandaily revolutions of three time blocks of 3 d each are presented for each genotype and treatment group. , Placebo control group; z, low doseof EB (16 ng/d) treatment group; f, high dose of EB (160 ng/d) treatment group. *, P � 0.05 vs. the respective placebo control group; a, P �0.05 vs. �WT of the respective treatment group.
Ogawa et al. • Activity in �ERKO and �ERKO Mice Endocrinology, January 2003, 144(1):230–239 235
mPOA (10–14), although potential synergistic actions of thetwo receptors could well be localized in different parts of thebrain.
It should be noted that the doses of estrogen found to bequite effective to increase running wheel activity in wild-typeand �ERKO mice in the present study were much lower thanthe doses normally used to induce female sexual behavior. Bythe last day of the running wheel activity test, mice in thelower and higher dose estrogen-treated groups received atotal of 0.27 or 2.7 �g, respectively. Serum estradiol levelsdetermined in some of the mice used in the present study on26 d after pellet implant were indeed less than the detectablelevel of the assay (�5 pg/ml) even in the higher estrogendose group (160 ng/d). Nevertheless, we found that stimu-latory effects of estrogen on running wheel activity levels in�WT, �WT, and �ERKO mice of both sexes were more ob-vious during the second and third blocks compared with thefirst block. Except in �WT female mice, in which the inter-action between block and treatment was not statistically sig-nificant, the placebo groups of these genotypes did not show
such an increment in running wheel activity during the day.In the present study we used EB to ensure stimulatory actionof systemically administered estrogen on running wheel ac-tivity, particularly in �ERKO mice, in which most of theestrogen-inducible behaviors were severely disrupted. As EBhas a longer half-life in the body than 17�-estradiol, accu-mulation of the steroid over days might cause a gradualincrease in running wheel activity in estrogen-treatedgroups. In �ERKO mice, on the other hand, all three treat-ment groups, including the placebo groups, showed a grad-ual increase in running wheel activity in both sexes. Thesefindings suggest that running experience could itself stim-ulate running wheel activity in �ERKO mice even thoughestrogen failed to modify it.
Sex comparisons across the estrogenic regulation of runningwheel activity
In the present study, there were no obvious sex differencesin estrogen effects on running wheel activity. Both male and
FIG. 4. Time course of estrogenic regulation of running wheel activity in �WT and �ERKO (A) and �WT and �ERKO (B) male mice. Mean dailyrevolutions of three time blocks of 3 d each are presented for each genotype and treatment group. , Placebo control group; z, low dose of EB(16 ng/d) treatment group; f, high dose of EB (160 ng/d) treatment group. *, P � 0.05 vs. the respective placebo control group; a, P � 0.05 vs.�WT of the respective treatment group.
236 Endocrinology, January 2003, 144(1):230–239 Ogawa et al. • Activity in �ERKO and �ERKO Mice
female �ERKO failed to respond to EB treatment, whereas�WT, �WT, and �ERKO mice of both sexes showed re-sponses to EB. These findings are consistent with a previousstudy in rats showing no sex differences in the response toestrogen after gonadectomy (4). The study by Gentry andWade (4) also revealed that neonatal androgenization of fe-male rats did not change the magnitude of the response toestrogen later in adulthood for the potentiation of runningwheel activity, although they needed a longer latency torespond. Taken together, these findings suggest that facili-tation of running wheel activity by estrogen is not a sexuallydimorphic trait.
A number of studies in golden hamsters and rats haveshown that there is a clear sex difference in estrogenic reg-ulation of free running circadian activity rhythms (21–24). Inthese studies it is implied that responsiveness to estrogen (e.g.shortened free running period, �) in adulthood is regulatedby the neonatal steroid condition. Both males and neonatallyandrogenized females failed to show shortened � in responseto estrogen in adulthood during free running circadianrhythm tests. In the present study we measured only dailyactivity under a normal 12-h light, 12-h dark entrained con-dition. Therefore, it remains to be determined in future stud-ies whether circadian systems measured by the use of run-ning wheel activity are also affected by ER� and/or ER� gene
disruption, and if so, if there is any sex difference. Potentialroles of ER� and ER� genes in the regulation of the clock geneexpression (25, 26) in the suprachiasmatic nucleus as well asother brain regions receiving input from it (e.g. paraven-tricular nucleus) need to be further investigated.
Although there was no sex difference in EB effects onrunning wheel activity, there was sex/knockout interactionin placebo groups of �WT and �ERKO mice. Unlike in openfield activity, as discussed below, there was no genotypedifference in running wheel activity in the placebo groups offemale mice. On the other hand, the placebo control groupof male �ERKO mice tended to be more active than �WTmice in running wheel tests, particularly during the last 3 dof 9-d tests. These results suggest that running wheel activityin gonadectomized mice might be modulated by ligand-independent effects of ER� gene disruption in male mice.
Roles of two types of ERs in the regulation of openfield activity
In contrast to a great potentiation of running wheel activity(up to 2- to 2.5-fold) by estrogen in both sexes, open fieldactivity was only slightly affected by estrogen in females. Inmales there was no effect of estrogen in any genotype of mice.Differential effects of estrogen on running wheel activity and
FIG. 5. Effects of estrogen treatment on open field activity. Means of 2-d tests are presented for the total moving distance (A for females andC for males) and the center distance (B for females and D for males). , Placebo control group; z, low dose of EB (16 ng/d) treatment group;f, high dose of EB (160 ng/d) treatment group. a, P � 0.05 vs. �WT of the respective treatment group.
Ogawa et al. • Activity in �ERKO and �ERKO Mice Endocrinology, January 2003, 144(1):230–239 237
open field activity were also described with brain site-specific implants of estrogen in female rats (9) and systemicestrogen administration in female mice (16).
We found that open field activity, both total and in thecenter area, before estrogen implants was significantly re-duced in �ERKO females, whereas it tended to increase in�ERKO females compared with that in the respective WTcontrol mice. A similar trend was noticed in the placebocontrol groups during the postimplant open field tests. Onthe other hand, no such genotype differences were observedin running wheel activity in the placebo control groups offemale mice regardless of gene. We found previously that�ERKO female mice tended to be less active, particularly inthe light compartment, during the dark/light transition testsregardless of gonadal state (20). Taken together, these find-ings suggest that the lack of ER� gene expression might byitself ligand-independently modify baseline activity in anovel environment (open field and dark/light transitiontests), but not home cage activity (running wheel), in femalemice. Furthermore, genotype differences between �ERKOand �WT female mice in both open field and dark/lighttransition (20) tests were consistent regardless of gonadalstates, i.e. gonadally intact, gonadectomized, and gonadec-tomized plus estrogen treatment. These findings suggest thatthe reduction of activity in �ERKO female mice in these novelenvironment tests may be due to a combination of the lackof responsiveness to estrogen in adulthood as well as de-velopmental effects of ER� gene disruption. During preim-plant open field tests in the present study, �ERKO femalemice resembled male �WT and �ERKO mice more thanfemale �WT mice in terms of total activity, total distance, andmoving time. We speculate that this type of behavioral mas-culinization in �ERKO female mice might be due to elevatedstimulation of AR during neonatal development. Starting onpostnatal d 9–21, the number of ligand-bound AR-immu-noreactive cells in �ERKO female mouse brains was signif-icantly higher than that in �WT females in the mPOA amongother brain regions and was almost equivalent to that in malemouse brains (27).
ER� gene disruption, on the other hand, tended to increaseopen field activity in female mice. This is in marked contrastto a great reduction of open field activity and elevation ofanxiety reported in ER� knockout female mice (28, 29), whichwere developed separately from those used in the presentstudy. There are three important factors that might influencethe differences in open field activity between these two stud-ies. In the present study mice were tested under red light,whereas in the study by Krezel et al. (29), they were testedunder brightly illuminated white light conditions. It is as-sumed that the latter is more sensitive than the former toreveal traits such as anxiety or emotionality. Secondly, wetested the mice after gonadectomy, but Krezel et al. testedgonadally intact mice. However, this may not be a majorfactor in explaining the behavioral differences, as the effectsof ER� gene disruption might also be independent of thegonadal state at the time of testing, as we found in the caseof ER� (see above). Thirdly, a more important differencebetween the two studies may be the age of the mice. Mice inthe present study were tested at 10–12 wk, whereas those inKrezel’s study were tested when they were more than 7
months old. In fact, we found that in older �ERKO females(�30 wk of age), open field activity and radial maze activityduring a spatial learning paradigm, both of which wereperformed under bright light illumination, were greatly re-duced (30). It is possible, therefore, that the effects of ER�gene disruption may best appear in older female mice.
Acknowledgments
The authors are thankful to Dr. C. J. Krebs for his molecular biologicalexpertise, and Dr. C. Pavlides and Ms. L. Frank for critical and editorialreading of the manuscript.
Received May 15, 2002. Accepted September 23, 2002.Address all correspondence and requests for reprints to: Sonoko
Ogawa, Ph.D., Laboratory of Neurobiology and Behavior, RockefellerUniversity, Box 275, 1230 York Avenue, New York, New York 10021.E-mail: [email protected].
This work was supported by National Science Foundation GrantIBN-9728579 and National Institute of Mental Health Grant 62147(to S.O.).
References
1. Wollnik F, Turek FW 1988 Estrous correlated modulations of circadian andultradian wheel-running activity rhythms in LEW/Ztm rats. Physiol Behav43:389–396
2. Rodier III WI 1971 Progesterone-estrogen interactions in the control ofactivity-wheel running in the female rat. J Comp Physiol Psychol 74:365–373
3. Roy EJ, Wade GN 1975 Role of estrogens in androgen-induced spontaneousactivity in male rats. J Comp Physiol Psychol 89:573–579
4. Gentry RT, Wade GN 1976 Sex differences in sensitivity of food intake, bodyweight, and running-wheel activity to ovarian steroids in rats. J Comp PhysiolPsychol 90:747–754
5. Ahdieh HB, Wade GN 1982 Effects of hysterectomy on sexual receptivity, foodintake, running wheel activity, and hypothalamic estrogen and progestinreceptors in rats. J Comp Physiol Psychol 96:886–892
6. Stockman ER, Albers HE, Baum MJ 1985 Activity in the ferret: oestradioleffects and circadian rhythms. Anim Behav 33:150–154
7. King JM 1979 Effects of lesions of the amygdala, preoptic area, and hypo-thalamus on estradiol-induced activity in the female rat. J Comp PhysiolPsychol 93:360–367
8. Takeo T, Sakuma Y 1995 Diametrically opposite effects of estrogen on theexcitability of female rat medial and lateral preoptic neurons with axons to themidbrain locomotor region. Neurosci Res 22:73–80
9. Fahrbach SE, Meisel RL, Pfaff DW 1985 Preoptic implants of estradiol increasewheel running but not the open field activity of female rats. Physiol Behav35:985–992
10. Shughrue PJ, Scrimo P, Lane M, Askew R, Merchenthaler I 1997 The dis-tribution of estrogen receptor-� mRNA in forebrain regions of the estrogenreceptor-� knockout mouse. Endocrinology 138:5649–5652
11. Shughrue PJ, Lane MV, Merchenthaler I 1997 Comparative distribution ofestrogen receptor-� and -� mRNA in the rat central nervous system. J CompNeurol 388:507–525
12. Shughrue PJ, Scrimo PJ, Merchenthaler I 1998 Evidence for the colocalizationof estrogen receptor-� mRNA and estrogen receptor-� immunoreactivity inneurons of the rat forebrain. Endocrinology 139:5267–5270
13. Nomura M, Korach KS, Pfaff DW, Ogawa S, Differential regulation of es-trogen receptor (ER)-� immunoreactivity by gonadal steroids in specific brainregions of male wild type and ER-� knockout mice. Mol Brain Res, in press
14. Mitra SW, Wilkinson H, Rohrer S, Liu H, Yuckovitz J, Schaeffer JM, McEwenBS, Pfaff DW, Ogawa S, Hoskin E, Alves SE, Characterization of a novel ER�antibody and localization of ER� immunoreactivity in the mouse brain. Pro-gram of the 83rd Annual Meeting of The Endocrine Society, Denver, CO, 2001,p 160 (Abstract P1-48)
15. Blizard DA, Lippman HR, Chen JJ 1975 Sex differences in open-field behaviorin the rat: the inductive and activational role of gonadal hormones. PhysiolBehav 14:601–608
16. Morgan MA, Pfaff DW 2001 Effects of estrogen level on activity and fear-related behaviors in mice. Horm Behav 40:472-482
17. Nagy ZM, Glaser HD 1970 Open-field behavior of C57BL/6J mice: effects ofillumination, age, and number of test days. Psychon Sci 19:143–145
19. Ogawa S, Chan J, Chester AE, Gustafsson J, Korach KS, Pfaff DW 1999Survival of reproduction-related behaviors in male and female estrogen re-
238 Endocrinology, January 2003, 144(1):230–239 Ogawa et al. • Activity in �ERKO and �ERKO Mice
ceptor � deficient (�ERKO) male and female mice. Proc Natl Acad Sci USA96:12887–12892
20. Ogawa S, Eng V, Taylor JA, Lubahn DB, Korach KS, Pfaff DW 1998 Rolesof estrogen receptor-� gene expression in reproduction-related behaviors infemale mice. Endocrinology 139:5070–5081
21. Thomas EM, Armstrong SM 1989 Effect of ovariectomy and estradiol on unityof female rat circadian rhythms. Am J Physiol 257:R1241–R1250
22. Albers HE 1981 Gonadal hormones organize and modulate the circadiansystem of the rat. Am J Physiol 241:R62–R66
23. Zucker I, Fitzgerald KM, Morin LP 1980 Sex differentiation of the circadiansystem in the golden hamster. Am J Physiol 238:R97–R101
24. Ruiz de Elvira MC, Persaud R, Coen CW 1992 Use of running wheels regulatesthe effects of the ovaries on circadian rhythms. Physiol Behav 52:277–284
25. Takahashi JS 1995 Molecular neurobiology and genetics of circadian rhythmsin mammals. Annu Rev Neurosci 18:531–553
26. King DP, Takahashi JS 2000 Molecular genetics of circadian rhythms inmammals. Annu Rev Neurosci 23:713–742
28. Dupont S, Krust A, Gansmuller A, Dierich A, Chambon P, Mark M 2000 Effectof single and compound knockouts of estrogen receptors � (ER�) and � (ER�) onmouse reproductive phenotypes. Development 127(Suppl):4277–4291
29. Krezel W, Dupont S, Krust A, Chambon P, Chapman PF 2001 Increasedanxiety and synaptic plasticity in estrogen receptor �-deficient mice. Proc NatlAcad Sci USA 98:12278–12282
30. Sacrestano D, McKenna E, Ogawa S, Gustafsson J-Å, Smithies O, Korach KS,Pfaff DW 2000 Spatial learning in gene knockout mice of estrogen receptors� (�ERKO) and � (�ERKO). Soc Neurosci Abstr 26:1746
Ogawa et al. • Activity in �ERKO and �ERKO Mice Endocrinology, January 2003, 144(1):230–239 239
