Hormones and Behavior 49 (2006) 587 – 597 www.elsevier.com/locate/yhbeh

Steroid hormone mediation of limbic brain plasticity and aggression in free-living tree lizards, Urosaurus ornatus David Kabelik ⁎, Stacey L. Weiss 1 , Michael C. Moore School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA Received 14 September 2005; revised 2 December 2005; accepted 6 December 2005 Available online 25 January 2006

Abstract The neural mechanisms by which steroid hormones regulate aggression are unclear. Although testosterone and its metabolites are involved in both the regulation of aggression and the maintenance of neural morphology, it is unknown whether these changes are functionally related. We addressed the hypothesis that parallel changes in steroid levels and brain volumes are involved in the regulation of adult aggression. We examined the relationships between seasonal hormone changes, aggressive behavior, and the volumes of limbic brain regions in free-living male and female tree lizards (Urosaurus ornatus). The brain nuclei that we examined included the lateral septum (LS), preoptic area (POA), amygdala (AMY), and ventromedial hypothalamus (VMH). We showed that the volumes of the POA and AMY in males and the POA in females vary with season. However, reproductive state (and thus hormonal state) was incompletely predictive of these seasonal changes in males and completely unrelated to changes in females. We also detected male-biased dimorphisms in volume of the POA, AMY, and a dorsolateral subnucleus of the VMH but did not detect a dimorphism between alternate male morphological phenotypes. Finally, we showed that circulating testosterone levels were higher in males exhibiting higher frequency and intensity of aggressive display to a conspecific, though brain nucleus volumes were unrelated to behavior. Our findings fail to support our hypothesis and suggest instead that plasma testosterone level covaries with aggression level and in a limited capacity with brain nucleus volumes but that these are largely unrelated relationships. © 2005 Elsevier Inc. All rights reserved. Keywords: Testosterone; Corticosterone; Estradiol; Reptile; Preoptic area; Amygdala; Ventromedial hypothalamus; Seasonal changes; Brain volume; Aggressive behavior

Introduction Testosterone's involvement in mediating aggressive behavior in adult males has long been known, though the neural mechanisms underlying its actions remain unclear. Both laboratory and field studies of males of various taxa show that aggression levels covary with testosterone level (Beeman, 1947; Moore, 1987; Nelson and Chiavegatto, 2001; Schwabl and Kriner, 1991), although the mechanisms underlying this association may be manifold. Indeed, adult testosterone has been implicated in many processes including the modulation of neuropeptide and neurotransmitter levels (Brot et al., 1993; McCarthy and Pfaus, 1996), receptor abundance (McIntyre et ⁎ Corresponding author. Fax: +1 480 965 7599. E-mail address: [email protected] (D. Kabelik). 1 Present address: Biology Department, University of Puget Sound, Tacoma, WA 98416, USA. 0018-506X/$ - see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.yhbeh.2005.12.004

al., 2002), enzymatic activity (Fusani et al., 2001), neurogenesis and cell survival (Tanapat et al., 1999), neuronal structure (Cooke and Woolley, 2005), and gross neural morphology (Tramontin et al., 2003). Neuroanatomical studies often demonstrate a positive association between the volumes of certain limbic brain nuclei and levels of testosterone or its metabolites (Ball et al., 2002; Cooke et al., 1999; Tramontin and Brenowitz, 2000). Limbic brain nuclei are also involved in the regulation of aggressive and reproductive behaviors (Greenberg et al., 1984; Nelson and Chiavegatto, 2001). In the present study, we therefore examined testosterone's role in mediating the volumes of several limbic brain nuclei and determined whether these changes parallel changes in aggressive behavior. The effects of testosterone on neural morphology have been studied most extensively in songbirds. The brain nuclei that control song learning and production fluctuate in size between the breeding and non-breeding season (Nottebohm, 1981) and steroid hormones play a role in the regulation of this seasonal

588

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

plasticity (Smith et al., 1997). Various experiments, however, have demonstrated that testosterone's role in this process is not well understood and varies with species (Ball et al., 2002), age of individuals (Absil et al., 2003), as well as social and environmental conditions (Tramontin et al., 1999, 2001). Limbic brain nuclei have also been shown to vary in volume with testosterone level or season (reviewed in Tramontin and Brenowitz, 2000) in taxa including mammals (Commins and Yahr, 1984; Cooke et al., 1999) and reptiles (Wade and Crews, 1991; Wade et al., 1993). However, studies relating changes in limbic neural structure to changes in behavior, especially aggressive behavior, are limited. Using lizards as a model system provides benefits over the study of mammals and birds because lizards have very highly stereotyped aggressive displays (Carpenter, 1978) which are easily quantified and limbic regions comprise of larger portions of their brains (MacLean, 1978), therefore presenting a more easily understood model of the aggression-mediating circuit. We used the tree lizard, Urosaurus ornatus, to examine interactions among steroid hormones, neural morphology, and aggressive behavior. Tree lizards are small iguanid lizards abundant in southwestern North America. Male tree lizards can be divided into morphological phenotypes (or morphs) that differ in aggression levels and territoriality, thus allowing for powerful within-sex comparisons (Moore et al., 1998). The morphs also differ in dewlap (throat-fan) coloration. Males with a bicolored orange-blue dewlap (hereafter referred to as orangeblue males) tend to be more aggressive and territorial than males with unicolored orange dewlaps (orange males) (Thompson and Moore, 1991, 1992). High testosterone and progesterone levels during early development have been shown to be involved in inducing development of the orange-blue phenotype during development (Hews and Moore, 1996; Hews et al., 1994), but morphs do not differ in adult hormone levels (Moore et al., 1998). Furthermore, morph type is irreversible in adulthood (Moore et al., 1998). This study was designed to investigate connections between hormone-induced changes in neural structure and aggressive behavior in a free-living species undergoing natural seasonal hormone fluctuations. The vast majority of studies examining limbic plasticity are conducted in laboratory species, and few relate such neural changes to aggression. Studies of free-living organisms are extremely important because laboratory studies do not always mimic the effects of natural seasonal changes in hormone levels and environment (Ball et al., 2002; Smulders, 2002). We addressed the hypothesis that parallel changes in both steroid levels (testosterone, corticosterone, estradiol) and brain nucleus volumes are involved in the regulation of aggression in the tree lizard. This hypothesis generates three main correlational predictions that can be tested in the current field study of free-living animals. (1) Differences in adult or developmental steroid levels should parallel differences in the volumes of adult brain nuclei. We tested this prediction by comparing the volumes of brain nuclei in tree lizards across male morph types, between sexes, and across seasons and reproductive states. (2) Frequency and intensity of aggressive display should vary with

plasma steroid hormone levels. We tested this prediction by comparing aggression scores with steroid levels in male tree lizards (females of this species show very little aggressive behavior). Furthermore, we compared aggression scores across seasons and reproductive states. (3) Any observed differences in brain nucleus volumes among comparison groups (i.e. season, reproductive state) should be associated with parallel differences in aggression level. Furthermore, a correlation between aggression and brain nucleus volumes should be present. We tested this prediction by comparing observed differences in brain nucleus volumes with observed differences in aggression level and by directly comparing brain nucleus volumes among groups of males exhibiting varying levels of aggression. Methods All procedures described herein were approved by the Arizona State University Institutional Animal Care and Use Committee and were in accordance with NIH Guidelines for the Care and Use of Laboratory Animals. Appropriate scientific collecting permits were obtained from the Arizona Game and Fish Department.

Field sites and animals Our fieldwork was carried out in the Tonto National Forest, approximately 67 km northeast of Arizona State University, along Arizona State Highway 87 (111°30′ W, 33°42′ N). This area consists of large granite boulder fields and Upper Sonoran Desert vegetation. We captured a total of 174 animals of adult body size and with clearly demarked dewlaps across three seasons in 2002: 68 in the early breeding season (February–March), 59 in the peak breeding season (May–June), and 47 in the late breeding season (August–September). Of these, 45 were females, 62 were orange males, and 67 were orange-blue males. We captured animals by noosing, immediately took blood samples for hormone analysis, and recorded dewlap coloration and body measurements. In addition, the behavioral profiles of a subset of males (33 orange and 43 orange-blue) were rapidly assessed prior to capture (see below). A further 70 animals of both sexes (some overlapping with those whose behavior was measured) were sacrificed, their gonads and adipose tissue were weighed, and their brains were fixed and sectioned for Nissl staining (see below). Dissected females could be divided into reproductive states by examining qualitative differences in egg development. Females were scored as either nonreproductive, vitellogenic (eggs were being yolked), or gravid (eggs were shelled and ovulated). Males were assigned to a reproductive state based upon combined testes mass: regressed (b0.02 g) or developed (N0.02 g). This cutoff was chosen because all early breeding animals had a combined testes mass below 0.017 g, and all peak breeding season animals except one had a combined testes mass greater than 0.023 g. Late breeding season animals spanned both ranges, but with a majority of animals in the regressed testes group.

Behavioral trials Behavioral profiles of male tree lizards were obtained by introducing a tethered orange-blue stimulus male to within a meter of the resident and scoring the resident's behavioral response for 3 min (method adapted from Moore, 1987). Stimulus males were captured in areas at least 300 m away from resident males and were used no more than thrice before being returned to their area of capture. Behavioral trials in which resident and stimulus males were more than 5 mm snout-vent length of each other were excluded from analyses. Analyses were verified by also excluding males size-matched beyond 1 mm (as in Thompson and Moore, 1992), but results did not differ. Behavioral displays were categorized by frequency observed within the 3-min trial and maximal intensity (adapted from Weiss and Moore, 2004). Frequency groups were none, few (1–3), and many (4–7). Intensity groups were low (no display, approach, pushup without lateral compression of body), medium (fullshows and fullshow

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597 holds—consisting of extension of dewlap and lateral compression of body to display blue belly patches), and high (chasing and biting).

Hormone assays At the end of each 3-min behavioral trial, the resident male was immediately captured and bled from the infraorbital sinus using a heparinized capillary tube. Blood samples were obtained within 3.52 ± 0.15 min (mean ± SE) of the end of the 3-min behavioral trial. Females and males not involved in behavioral trials were captured and bled within 4.03 ± 0.18 min (mean ± SE) of being located. The blood was stored on ice until our return to the laboratory several hours later, where plasma was separated by centrifugation and stored at −20°C. Steroid hormone levels were determined by radioimmunoassay following established methods (Moore, 1986). Three assays were run because of the large number of samples. Briefly, samples were equilibrated overnight with 2000 cpm of radioactive hormone for determination of individual recoveries. Steroids were ether-extracted from plasma and separated by stepwise elution using celite chromatography. Collected fractions were then dried down, resuspended in phosphate buffer, and assayed. Resulting values were corrected for plasma volume, losses in extraction and chromatography steps, and assay accuracy relative to known standards. Intra-assay coefficients of variation for testosterone were 17.4%, 16.2%, and 4.6%, while inter-assay variation was 5.4%. Intra-assay variations for corticosterone were 19.7%, 24.5%, and 5.4%, while inter-assay variation was 12.1%. Intra-assay variation for estradiol (females only) was 4.7%. No inter-assay variation is determined for estradiol because all of the female plasma was run together in one assay. Assays were run using testosterone antibody from RDI (cat #WLI-T300301916), corticosterone antibody from MP Biomedicals (formerly ICN, cat #07120016), and estradiol antibody from Endocrine Sciences (no longer available). Minimum detectable values were 0.4 ng/ml for testosterone, 0.3 ng/ml for corticosterone, and 0.3 ng/ml for estradiol.

589

then multiplied by twice the thickness of the measured alternating sections and summed in order to obtain an approximate volume measure for that area. Brain nuclei were measured on both hemispheres. All measurements were conducted blind to treatment group. Saved images depicting measurement outlines were later double-checked for measurement consistency among brains to ensure standardized selection of focal regions. Furthermore, ten brains were remeasured, and the coefficients of variation between the two sets of measurements ranged from 2.7 to 4.6% across brain nuclei.

Brain nuclei Fig. 1 displays the seven limbic brain nuclei that we examined. Except for the habenula (HAB), these nuclei were chosen because of their previous implication in the control of aggression (Greenberg et al., 1984; Nelson and Chiavegatto, 2001; Sugerman and Demski, 1978; Tarr, 1977) and/or the presence of steroid receptors (Kabelik and Moore, unpublished; Moga et al., 2000; Morrell et al., 1979; Rosen et al., 2002; Young et al., 1994). Habenula The HAB was chosen as an easily quantifiable control region devoid of steroid receptors that we assumed would not be affected by changes in circulating steroid hormone levels. The tree lizard HAB, as in other species, is a very distinctive, darkly Nissl staining region found adjacent to the third ventricle at the dorsal end of the diencephalon. We measured this nucleus on all brain sections where it was present. Brain atlases from other lizard species were used to verify identification of this brain region (Figs. 3, 4, 5 in Butler

Brain fixation and staining Following bleeding, subjects were rapidly decapitated and their brains fixed, frozen, cryosectioned, and Nissl-stained. This procedure entailed making incisions into the decapitated skull to allow for penetration of fixative and then submerging the entire head within 5% acrolein in 0.1 M sodium phosphate buffer. The head was kept in this solution for 5 h while on ice, after which it was rinsed in buffer and the brain dissected from the skull. Brains were then embedded in 8% gelatin and stored overnight at 4°C. The gelatin block was trimmed and post-fixed for 16 h in 4% paraformaldehyde followed by 72h cryoprotection in 30% sucrose. The gelatin block was then frozen under powdered dry ice and stored at −80°C until sectioned on a cryostat. Every second floating section (40 μm thick) was collected, mounted, and Nissl-stained for 5 min using 0.5% Cresyl Violet dye. The remaining sections were kept for a separate immunocytochemistry study.

Image analyses Brain images were digitally captured from a camera (Panasonic GP-US502) attached to a light microscope (Olympus BX40) and imported into an image analysis program (Image-Pro Plus, version 4.0, Media Cybernetics, Silver Springs, MD). All microscope, camera, and computer settings were standardized across samples. All images were captured through a 4× objective producing an image on screen similar to a 40× (ocular × objective) magnification as seen through the microscope. Brain nuclei were located according to neural landmarks as no stereotaxic atlas exists for this species. However, our brain nuclei delineations were determined by reference to other lizard atlases (Butler and Northcutt, 1973; Greenberg, 1982; Northcutt, 1967; Smeets et al., 1986). Furthermore, our delineations agree with other published research involving lizard brain nuclei (e.g. Moga et al., 2000; Morrell et al., 1979; Rosen et al., 2002). Consistency in orientation, start and end points, and extent of nuclei was scrupulously ensured by comparison to landmarks. Neural measurements were obtained by manually outlining Nissl-stained regions of dense cell clustering. The image analysis program generated area measurements based on a conversion factor for the magnification. The brain area measurements were

Fig. 1. The brain nuclei quantified in this study: habenula (HAB), lateral septum (LS), preoptic area (POA), nucleus sphericus (NS), amygdala (AMY), ventromedial hypothalamic nucleus (VMH), and dorsolateral subnucleus of ventromedial hypothalamic nucleus (VMHs).

590

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

and Northcutt, 1973; Figs. 12–14 in Greenberg, 1982; Figs. 18–21 in Smeets et al., 1986). Lateral septum (LS) Because the anterior end of the LS was difficult to consistently determine, we measured the lateral septum across six subsequent alternate sections, near its posterior end, ending prior to the section on which it became longer along its medial–lateral axis than its dorsal–ventral axis. This nucleus is also referred to as the nucleus dorsolateralis of the septum in Figs. 9–12 of Greenberg (1982) and the nucleus parolfactorius lateralis in Figs. 6–7 of Northcutt (1967). Smeets et al. (1986, Figs. 13–14,16) also display the lateral septum at a level similar to our measurements. Preoptic area (POA) We quantified the entire POA from its anterior region at the start of the third ventricle to a point where it clearly transitioned to the periventricular nucleus. At this point, the nucleus became less spherical, and the dorsal end began to expand laterally and merge with the dorsolateral hypothalamic nucleus. Examples of this region at similar levels in other lizard species include Smeets et al. (1986, Figs. 14, 16–18) and Greenberg (1982, Figs. 9–11). Although some lizard atlases treat this area as one continuous region (Butler and Northcutt, 1973, Fig. 1; Northcutt, 1967, Figs. 6–9), our measures may have also included the bed nucleus of the stria terminalis and parts of the suprachiasmatic nucleus, both of which are continuous with the POA in this and several other reptilian species (Crosby and Showers, 1969). Although these adjoining regions are small relative to the volume of the POA (Greenberg, 1982, Figs. 9–11), their inclusion may add some variation to our measures of the POA. Nucleus sphericus (NS) The NS is a very distinctive spherical nucleus found adjacent to the amygdala on the dorsal ventricular ridge in reptilian brains and involved in the relay of vomeronasal information (Lanuza and Halpern, 1997). We quantified the NS from its first appearance as a complete oval on a brain section to its posterior end where it merges with the ventromedial wall of the dorsal ventricular ridge, and no longer forms a distinct circular formation. The NS is also a very prominent landmark in other lizard species (Butler and Northcutt, 1973, Figs. 1, 2, 3; Northcutt, 1967, Figs. 8–9; Smeets et al., 1986, Figs. 18–21). Amygdala (AMY) Amygdaloid nuclei cannot be consistently distinguished from one another in this or related species (Northcutt, 1967), and so the entire amygdaloid region excluding the NS was measured as a whole and is referred to as the amygdala. This region is also known as the posterior dorsal ventricular ridge (Greenberg et al., 1979). We started our measurements of the AMY when a dense clustering of cells was first apparent and distinct from surrounding regions. This clustering always coincided with a fissure forming to distinguish the dorsal ventricular ridge of the telencephalon from the diencephalon. We terminated AMY measurements at the posterior end of the dorsal ventricular ridge, when dense AMY staining was no longer present. The tokay gecko AMY, subdivided into component nuclei, is presented in Figs. 13–14, 16–19 of Smeets et al. (1986). Only one subnucleus (ventromedial nucleus of the posterior dorsal ventricular ridge) is clearly demarked and prominent in the green anole lizard (Greenberg, 1982). Ventromedial hypothalamic nucleus (VMH) We measured the entire ventromedial hypothalamic nucleus (VMH). The anterior region of the VMH was identified as a small oval clustering of cells on the ventral end of the diencephalon, immediately posterior to the disappearance of the optic chiasm. The VMH ended when the clustering of cells that defined its borders disappeared along the edge of the third ventricle. Our delineation of the VMH matched those of Smeets et al. (1986, Figs. 21– 22) and Butler and Northcutt (1973, Figs. 3–7, nucleus ventralis hypothalami). VMH subnucleus (VMHs) A region of the dorsolateral VMH that is particularly dense in cells (Fig. 1) and steroid receptors (Kabelik and Moore, unpublished; Morrell et al., 1979;

Rosen et al., 2002; Young et al., 1994) was quantified separately and termed the VMH subnucleus (VMHs)—it is possible that this region is homologous to the lateral zone of the VMH, an area that shows activity during social stress (Goodson and Evans, 2004; Kollack-Walker and Newman, 1995). The VMHs were measured from its anterior appearance as a dense cluster of cells within the VMH to its posterior end, when it could be measured as separate and dorsolateral to the receding VMH.

Brain volume Because male brains were larger than female brains (t = 3.66, df = 74, P b 0.001) and to standardize for overall brain size, a measure of brain volume was included as a covariate in all analyses of nucleus volumes. To estimate brain volume, we multiplied the area of a brain section at the level of a prominent centrally located septal landmark (the nucleus septalis impar, as depicted in Fig. 14 of Smeets et al. (1986)) by a measure of brain length. To be as consistent as possible in our measure of brain length, we chose to measure an area between two prominent centrally located landmarks to designate start and end points instead of trying to measure the entire brain. We therefore measured distance from the start of the habenula to the midpoint of the posterior commissure as brain length. The use of central landmarks minimized possible variability due to sectioning asymmetry.

Data analyses Hormone values were log-transformed to meet assumptions of normality and were analyzed using one- or two-way analysis of variance (ANOVA) with season (early, mid, late) and group (female, orange male, orange-blue male), or male and female reproductive states as main variables. Post hoc analyses were conducted with the Student–Newman–Keuls (SNK) tests with α b 0.05. Hormonal differences between behavior groups were analyzed using one-way ANOVA with SNK post hoc analyses. Behavioral differences among seasons and between male reproductive states were analyzed using one-way ANOVAs for fullshow frequency and Kruskal–Wallis and Mann–Whitney U tests for display intensity. Morph differences in behavioral frequency and intensity were analyzed using Mann– Whitney U tests. Neural measurements were analyzed using multivariate analysis of covariance (MANCOVA). If Wilks' Lambda tests showed a significant multivariate effect, subsequent individual one- or two-way ANCOVAs were performed for each dependent variable (brain region). Because estimated brain volume was used as a covariate, standard post hoc tests could not be performed, and so we instead conducted Bonferroni pairwise comparisons. We included brain volume as a covariate to control for differences in brain nucleus volumes resulting from brain volume differences between sexes—this was a conservative correction that did not result in significant results beyond those found without correction. All resulting figures are presented as relative brain nucleus volumes because they display estimated marginal means (generated from the absolute measurements but corrected so that the covariate (estimated brain volume) is at its mean value).

Results Hormones and brain nucleus volumes: morph, sex, season, and reproductive state Testosterone levels varied across both group (F(2,164) = 137.23, P b 0.001) and season (F(2,164) = 3.12, P = 0.047), and there was a significant group × season interaction (F(4,164) = 2.49, P = 0.045). Post hoc tests revealed no difference in testosterone levels between male morph types, but both morphs had higher testosterone levels than females. Furthermore, mean testosterone levels were higher in both early and mid-season animals than in late season animals (P b 0.001). Supplementary

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

591

Fig. 2. (A) Testosterone levels across groups (female, orange male, orange-blue male) and seasons (early, mid, and late breeding season). Sample sizes were: females 13, 11, 21; orange males 29, 21, 12; orange-blue males: 25, 27, 14 across early, mid, and late breeding season respectively. (B) Corticosterone levels across groups and seasons. Sample sizes same as above except orange-blue males, mid-season: 26. (C) Brain region volumes between orange and orange-blue male morphs. (D) Brain region volumes between male and female tree lizards. (E) Brain region volumes among early, mid, and late breeding seasons. A–B denote untransformed mean values ± SE, and common letters above symbols represent homogeneous subsets at P N 0.05. C–E are based on estimated marginal means including overall brain volume as a covariate. Sample sizes per group are listed within brackets within legends. * denotes a difference between groups at P b 0.05, ** at P b 0.01. Lines over bars represent homogeneous subsets at P N 0.05. See text and Table 1 for further statistics.

one-way ANOVAs run separately for each group and season revealed that the interaction term was due to a decrease in late season male testosterone levels (Fig. 2A). Corticosterone levels also varied across groups (F(2,165) = 4.50, P = 0.013) but not across seasons (F(2,165) = 2.63, P = 0.075) and were higher than levels previously observed (Knapp and Moore, 1996; Lacy et al., 2002; Moore et al., 1998). No interaction between group and season was present (F(4,165) = 0.446, P = 0.78). Post hoc analyses revealed significantly higher corticosterone levels in females than in males of either morph type (Fig. 2B). Estradiol levels were measured only in females and were found to differ among seasons (F(2,41) = 3.47, P = 0.040, data not shown). Estradiol levels were higher in late season (1.66 ± 0.63 ng/ml) than in early season (0.31 ± 0.07 ng/ml), with mid-season animals possessing intermediate levels (1.70 ± 1.29 ng/ml; all values mean ± SE). All sample sizes are denoted in figure captions or legends. There was no overall effect of morph type on brain nucleus volumes (F(7,37) = 0.35, P = 0.93, Fig. 2C), and no morph differences were found within any individual brain regions (P N 0.50 for all), so, to increase statistical power, we collapsed

both morph types into the single group “males” for subsequent brain volume analyses. For verification purposes, we also ran analyses with morphs separated and obtained similar results. Multivariate tests revealed differences in the volumes of brain nuclei both between sexes (F(7,57) = 3.70, P = 0.004) and among seasons (F(14,114) = 1.96, P = 0.027). No interaction between sex and season was detected (F(14, 114) = 1.24, P = 0.26). Individual two-way ANCOVAs revealed significant differences for specific brain nuclei (Table 1). Briefly, males possessed a larger POA, AMY, and VMHs than did females (Fig. 2D), and the volume of the POA, NS, and AMY varied with season, being largest during the peak breeding season (Fig. 2E). Even without a significant interaction term between sex and season, we present seasonal neural plasticity separately for males and females (Fig. 3) since seasonal hormone patterns differ between sexes. The seasonal change in POA volume was robust in both males (F(2,42) = 3.46, P = 0.040) and females (F(2,20) = 3.85, P = 0.039), but the seasonal differences in AMY (F(2,42) = 3.65, P = 0.035) volumes seen in males were not present in females. Surprisingly, our control nucleus (HAB)

Table 1 Individual two-way ANCOVAs examining the effect of sex (male, female) and season (early, mid, late) on the volumes of examined brain nuclei (HAB, LS, POA, NS, AMY, VMH, VMHs)

Sex Season

df

HAB

LS

F

P

F

P

F

P

F

P

F

P

F

P

F

P

1, 63 2, 63

0.03 2.41

0.87 0.10

3.09 2.44

0.08 0.10

5.58 5.90

0.02 0.004

3.46 3.87

0.07 0.03

18.47 3.92

0.001 0.03

3.10 2.40

0.08 0.10

5.70 1.88

0.02 0.16

Significant P values are noted in bold.

POA

NS

AMY

VMH

VMHs

592

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

testes. There were no differences in brain nucleus volumes in females among any reproductive states (overall MANCOVA: F(14,28) = 1.18, P = 0.34; Fig. 6C). Hormones and aggression: morph, sex, season, and reproductive state There were no differences between orange and orange-blue males in either display frequency or intensity (Table 2). This lack of a difference in display frequency was true when examining all males (U = 669.5, n = 76, P = 0.63) or just those with high levels of testosterone (N30 ng/ml) (U = 170.5, n = 38, P = 0.78). Likewise, male morphs did not differ in display

Fig. 3. Brain nucleus volumes across seasons from Fig. 2E plotted separately for males and females in order to demonstrate seasonal changes within sexes. Figures are based on estimated marginal means including overall brain volume as a covariate. Sample sizes are displayed in brackets within the legends. * denotes a difference between groups at P b 0.05, † denotes P = 0.056.

was significantly smaller in late season females (F(2,20) = 5.79, P = 0.010), and a trend for a peak in VMH volume during peak breeding season was also marginally significant (F(2,20) = 3.34, P = 0.056). Although seasonal plasticity is apparent from our data within both male and female brains, it is unclear whether these volumetric changes are triggered by changes in hormone level or other seasonally varying factors. Therefore, we reexamined our data across male and female reproductive states (Fig. 4). Male tree lizards with fully developed testes had higher testosterone levels than males with regressed testes (t = 2.23, df = 88, P = 0.03, Fig. 4A); corticosterone levels did not differ (t = 0.26, df = 88, P = 0.80). Females of differing reproductive states had different testosterone (F(3,39) = 11.53, P b 0.001) and estradiol levels (F(3,39) = 10.37, P b 0.001), but not corticosterone levels (F(3,39) = 2.27, P = 0.10). Vitellogenic females had higher testosterone and estradiol levels than females of all other reproductive states (Fig. 4A). Interestingly, vitellogenic female testosterone levels rose to the low end of male-typical levels, suggesting a possible physiological role in female tree lizards for circulating testosterone. There was no overall difference in brain nucleus volumes among males with partially regressed versus fully developed testes (overall MANCOVA: F(7,37) = 1.75, P = 0.13, Fig. 4B). However, trends within the POA and AMY were present, and individual ANCOVAs indicated that males with larger testes had larger POA volumes (F(1,43) = 5.58, P = 0.02), though not AMY volumes (F(1,43) = 2.04, P = 0.16), than males with regressed

Fig. 4. (A) Testosterone, corticosterone, and estradiol levels across male and female reproductive states. Sample sizes: regressed testes (reg): 38, fully developed testes (dev): 52, pre-vitellogenic (pre-vit): 25, vitellogenic (vit): 8, gravid (grav): 8, post-reproductive (post-rep): 2. Brain nucleus volumes between (B) males and (C) females of different reproductive states. Because of a low sample sizes in the post-reproductive female group during analyses of brain nucleus volumes, this group was combined with pre-reproductive females as both had similar hormone levels. Figures are based on estimated marginal means including overall brain volume as a covariate. Sample sizes are displayed in brackets within the legends. * denotes a difference between groups at P b 0.05. See text for further statistics.

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597 Table 2 Counts of orange and orange-blue males expressing indicated levels of aggression frequency and intensity Morph type

Fullshow frequency None (0)

Orange Orange-blue Orange (high T only) Orange-blue (high T only)

Few (1–3)

Behavior intensity Many (4–7)

Low

Medium

High

8 (24%) 20 (61%) 5 (15%) 8 (24%) 12 (36%) 13 (39%) 8 (19%) 28 (65%) 7 (16%) 8 (19%) 19 (44%) 16 (37%) 3 (15%) 13 (65%) 4 (20%) 3 (15%) 8 (40%) 9 (45%)

3 (17%) 10 (56%) 5 (28%) 3 (17%)

6 (33%)

9 (50%)

High T denotes solely animals with high breeding levels of testosterone (N30 ng/ml). All morph comparisons are non-significant (see text for statistics).

593

levels than males exhibiting no fullshows and that males displaying high intensity behaviors had higher testosterone levels than males exhibiting low intensity behaviors. Corticosterone level was unrelated to either display frequency (F (2,60) = 1.39, P = 0.26, Fig. 5A) or intensity (F(2,60) = 1.10, P = 0.34, Fig. 5B). Aggression level, however, was not found to change with breeding season or reproductive state. Neither display intensity (χ2 = 0.29, df = 2, P = 0.87, Kruskal–Wallis test) nor frequency of fullshows (F(2,76) = 0.20, P = 0.82, one-way ANOVA) varied with season. Furthermore, neither aggression intensity (U = 65, n = 31, P = 0.59, Mann–Whitney U test) nor number of fullshows (F(1,31) = 0.25, P = 0.62, one-way ANOVA) varied with male reproductive state. Brain nuclei and aggressive behavior

intensity (U = 697, n = 76, P = 0.89), even if examining only high testosterone males (U = 174, n = 38, P = 0.85). Hence, we collapsed both morph types into the single group “males” for subsequent behavior analyses. Testosterone levels were higher in males displaying higher frequencies of aggressive displays within our 3-min observation period (F(2,60) = 3.24, P = 0.046, Fig. 5A), as well as males exhibiting a higher maximal display intensity (F(2,60) = 4.69, P = 0.013, Fig. 5B). Post hoc analyses revealed that males exhibiting a high frequency of fullshows had higher testosterone

Fig. 5. (A) Testosterone and corticosterone levels (mean ± SE) among males exhibiting no (0), few (1–3), and many (4–7) fullshow displays to an intruder. Sample sizes are 13, 40, and 10 respectively for groups none, few, and many. (B) Testosterone and corticosterone levels (mean ± SE) across males exhibiting low (none, approach, pushup), medium (fullshow, fullshow hold), and high (chase, bite) intensity displays to an intruder. Sample sizes are 13, 25, and 25 respectively for groups low, medium, and high. Letters denote groups different at the P b 0.05 significance level. See text for further statistics.

We compared brain nucleus volumes among males exhibiting no, few, and many fullshow displays, as well as among males exhibiting low, medium, and high maximal intensity of displays to a conspecific intruder. However, multivariate analyses revealed no difference in any brain nucleus volumes among behavioral frequency (F(14,20) = 0.99, P = 0.50, Fig. 6A) or behavioral intensity groups (F(14,20) = 0.81, P = 0.65, Fig. 6B).

Fig. 6. (A) Brain region volumes among males exhibiting no (0), few (1–3), and many (4–7) fullshow displays to an intruder. (B) Brain region volumes among males exhibiting low (none, approach, pushup), medium (fullshow, fullshow hold), and high (chase, bite) intensity displays to an intruder. Sample sizes per group are listed within brackets within legends. No significant differences in brain nucleus volumes were found among groups. See text for further statistics.

594

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

Discussion This study demonstrated that seasonal plasticity in the volumes of several limbic brain regions occurs in free-living animals under natural conditions. Although POA volume was greater in males with fully developed testes (an indicator of long-term testosterone level), other male brain nuclei changed with season but not male reproductive state. Hence, although testosterone may play a role in inducing seasonal changes in brain nucleus volumes, other variables that also change with season are likely involved in mediating this neural plasticity. Furthermore, although seasonal changes in brain nucleus volumes were also observed in females, female reproductive state (an indicator of long-term estradiol and testosterone levels) was not indicative of this neural plasticity. A male-biased sexual dimorphism was also observed in several brain regions, but whether this difference is due to developmental or adult hormone levels is unknown. This study also demonstrated that aggressive behavior is related to circulating testosterone levels in a free-living animal. However, no differences in brain nucleus volumes were detected between groups differing in aggression frequency or intensity nor did aggression levels parallel the seasonal profile of neural changes. These results therefore do not support our hypothesis that parallel changes in both steroid levels and brain nucleus volumes are involved in the regulation of aggression in a free-living animal. Instead, we conclude that the observed changes in adult brain nucleus volumes are independent of testosterone-induced changes in aggression level. Hormones and brain nucleus volumes: morph, sex, season, and reproductive state The first prediction from our hypothesis states that differences in adult or developmental steroid levels should parallel differences in the volumes of adult brain nuclei. When examining the effects of developmental hormone levels, we found no differences in adult brain nucleus volumes between male morphs, even though morphs differ in developmental hormone levels (Hews and Moore, 1996; Hews et al., 1994). We did detect a sexual dimorphism within several brain nuclei with males on average possessing a larger POA, AMY, and VMHs than females. Currently, we are not sure if these sexual dimorphisms are organized during development because adult hormone levels also differ between the sexes and have already been shown to maintain sexual dimorphisms in the medial amygdala of the rat (Cooke et al., 1999) and (via manipulations of steroid coactivators) the preoptic medial nucleus of the quail (Charlier et al., 2005). Furthermore, both the POA and AMY have been shown to differ in volume across seasons within this study, demonstrating adult plasticity within these brain nuclei in the tree lizard. In partial agreement with our prediction of parallel steroid levels and brain nuclei volumes, we observed robust seasonal variation in the volumes of the POA and AMY in males and the POA in females. The volumes of these brain nuclei were greatest in mid-breeding season when male testosterone levels

were highest. However, females do not show hormonal peaks at this time point, suggesting that reproductive hormone levels may not be the primary determinants of brain nucleus volume within this sex. We are confident that the observed changes in female POA volumes are valid, and this finding is supported by previous work demonstrating a change in cell size within the POA and AMY of both male and female green anole lizards between the breeding and non-breeding season (O'Bryant and Wade, 2002). Interestingly, our results differed from Wade and Crews (1991), who also found seasonal changes in limbic brain nucleus volumes in males but not in female whiptail lizards and who also reported seasonal differences in the volume of the VMH that we did not find. Interestingly, the VMH delineations of Wade and Crews (1991) were very similar to our own, whereas their POA delineations were slightly more extensive on the ventral–dorsal axis. It must also be noted that, since we could not completely distinguish the POA from the bed nucleus of the stria terminalis or the suprachiasmatic nucleus in our study, potential changes in these nuclei may also be contributing to our POA results. However, consistency between our findings and those of Wade and Crews (1991), even allowing for slight delineation differences, supports the robustness of the POA as a plastic brain nucleus in adults. Although we have demonstrated adult plasticity in several limbic brain nuclei, we are unsure if these are hormonally induced changes since season may not be a complete predictor of hormonal state. Hence, we reexamined our brain data measurement according to hormonal state. However, because steroid hormone levels fluctuate greatly throughout the day, our hormone measurements are only a snapshot of an individual's current state and may not reflect the average hormone levels being experienced over a multi-day period. Hence, we conducted analyses of brain nucleus volumes by male and female reproductive states, which should reflect average hormone levels across a longer time period. These analyses revealed that the observed seasonal variation in brain nucleus volumes could not be completely explained by reproductive state as male reproductive state only affected POA volume and female reproductive state was completely unrelated to brain nucleus volumes. It is thus unclear what role steroids play in influencing the observed seasonal changes in brain nucleus volumes. Much of this volumetric change may be attributable to other hitherto unknown environmental and physiological factors. This uncertainty over testosterone's role in mediating changes in brain nucleus volumes is reflected in the avian literature where a number of studies observed an effect of testosterone (Ball et al., 2002; Panzica et al., 1996; Smith et al., 1997; Tramontin and Brenowitz, 2000), whereas others have demonstrated partial to complete independence of neural plasticity and circulating testosterone levels (Ball et al., 2002; Deviche and Gulledge, 2000; Leitner et al., 2001; Tramontin et al., 2001). Hormones and aggression: morph, sex, season, and reproductive state The second prediction from our hypothesis states that the frequency and intensity of aggressive displays performed by

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

male tree lizards should vary with steroid hormone levels. Indeed, testosterone level was higher in groups of animals showing higher frequency and intensity of aggression. We assume that elevated concentrations of testosterone facilitate the behavior, rather than vice-versa, for two reasons. First, an earlier hormone manipulation study revealed a direct relationship between testosterone level and aggression (Weiss and Moore, 2004). Second, blood samples were obtained rapidly (mean ± SE: 3.52 ± 0.15 min) following the 3-min behavior trials, and previous studies on this species suggest that steroid hormone levels do not change this rapidly following an aggressive encounter (Knapp and Moore, 1995). We did not, however, detect differences in aggressive behavior across seasons or male reproductive states. Since a portion of the males from early and late seasonal time points had high testosterone levels, similar to levels of most males in the mid-season, seasonal variation in aggression may have been obscured by high levels of within-season variation. However, it is surprising that we detected no behavioral differences among male reproductive states, given the hormonal difference between these states. It is possible that, since testosterone levels fluctuate on the order of hours, correlations between aggression level and testosterone may be stronger than between aggression and reproductive state. We also did not detect differences in aggression level between male morph types. Based upon the work of Hover (1985), Thompson and Moore (1991, 1992), and CramerMeldrum (2000), we expected to observe greater aggression levels from orange-blue males. A multi-year drought leading to harsh and stressful environmental conditions, a low resultant population density and changed social structure, or the brief duration of our behavioral trials may all account for the lack of a difference in aggression level between morphs. Unfortunately, this behavioral finding precludes us from drawing firm conclusions regarding the accompanying lack of a neural dimorphism between morph types. Brain nuclei and aggressive behavior The third prediction from our hypothesis states that any observed differences in brain nucleus volumes among comparison groups (i.e. season, reproductive state) should be associated with parallel differences in aggression level. Furthermore, a correlation between aggression and brain nucleus volumes should be present. However, from the above two sections, it is apparent that, although seasonal changes in brain nucleus volumes are present, these are not paralleled by corresponding seasonal changes in aggression level. Furthermore, male reproductive state is only partly related to changes in brain nuclei and unrelated to aggression level. Finally, no difference in nucleus volumes was observed among groups varying in aggression frequency or intensity. These results suggest that changes in brain nucleus volumes are therefore not involved in the regulation of aggressive behavior as was predicted from our hypothesis. Although our results are correlational in nature, correlations between brain nucleus

595

volumes and aggression level would be expected if these variables were related causally. It is possible that changes in brain nuclei volumes may be more closely related to changes in reproductive behaviors than to aggressive behaviors. For instance, studies in male rats have demonstrated adult plasticity in the medial amygdala and positive correlations between this brain region and the frequency of displayed noncontact erections in physical proximity to a partitioned-off estrous female (Cooke et al., 2003; Cooke et al., 2000). Furthermore, Riters et al. (2000) demonstrated a positive correlation between medial preoptic area volume and singing behavior in starlings. However, other studies have failed to detect correlations between brain regions that show adult plasticity in volume and behavior. For instance, Wade et al. (1993) demonstrated that steroid hormone manipulations can alter both courting intensity to a receptive female as well as POA and VMH volumes in male whiptail lizards. However, the authors found no correlation between the volume of either nucleus and behavioral intensity. Furthermore, other studies show an effect of steroid manipulation of behavior but fail to demonstrate changes in brain morphology. For instance, O'Bryant and Wade (2002) detected effects of testosterone treatment on reproductive display behavior following testosterone treatment but failed to observe an effect of this treatment on cell size in various limbic regions, even though these cells change in size seasonally with reproductive state. A recent study has also shown that, whereas testosterone treatment affects both song nucleus volumes and singing behavior in canaries, singing behavior precedes alteration of some song control brain nuclei, thus suggesting at least a partial dissociation between structure size and behavior in those brain regions (Sartor et al., 2005). Hence, the literature on structure–function relationships between brain nuclei and reproductive behaviors suggests that these are complex relationships that may vary with sex, species, season, or experimental conditions. Similarly, the lack of correlation between brain nucleus volumes and aggressive behavior in the present study should not be taken as an absolute as the results may vary under different experimental conditions. To further determine whether testosterone-induced changes in limbic brain volumes do influence aggression in our model system, we are conducting a further experiment involving steroid hormone manipulations in this same model system. Conclusion The functional relevance and mechanisms underlying seasonal fluctuations in POA and AMY volumes in the tree lizard remain unknown and require further investigation. Our study suggests that steroid hormones play only a limited role in determining volume changes in male brain nuclei in a natural setting, and perhaps none in females. Other factors that may trigger these seasonal changes require investigation. Furthermore, our study suggests that, in nature, aggressive behavior seems to vary more with current testosterone level than either season or reproductive state. Thus, a clear dissociation seems present between seasonal changes in brain nuclei volumes and

596

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597

testosterone's mediation of aggression. We conclude that the observed effects of testosterone on aggression level and brain nucleus volumes are likely independent effects of the hormone and that changes in brain nucleus volumes are not necessary for changes in aggression level to occur. Acknowledgments We would like to thank D.F. Denardo, P. Deviche, M. Orchinik, R.L. Rutowski, and two anonymous reviewers for comments on the manuscript and also T. Crombie and T.A. Sultani for help in the laboratory. This research was funded by NIMH grants (5R01MH048564-10) to MCM and (MH12112) to SLW, as well as NSF DDIG (0408009) to DK. References Absil, P., Pinxten, R., Balthazart, J., Eens, M., 2003. Effect of age and testosterone on autumnal neurogenesis in male European starlings (Sturnus vulgaris). Behav. Brain Res. 143, 15–30. Ball, G.F., Riters, L.V., Balthazart, J., 2002. Neuroendocrinology of song behavior and avian brain plasticity: multiple sites of action of sex steroid hormones. Front. Neuroendocrinol. 23, 137–178. Beeman, E.M., 1947. The effect of male hormone on aggressive behaviour in mice. Physiol. Zool. 20, 373–405. Brot, M.D., De Vries, G.J., Dorsa, D.M., 1993. Local implants of testosterone metabolites regulate vasopressin mRNA in sexually dimorphic nuclei of the rat brain. Peptides 14, 933–940. Butler, A.B., Northcutt, R.G., 1973. Architectonic studies of the diencephalon of Iguana iguana (Linnaeus). J. Comp. Neurol. 149, 439–462. Carpenter, C.C., 1978. Ritualistic social behaviors in lizards. In: Greenberg, N., MacLean, P.D. (Eds.), Behavior and Neurology of Lizards. NIMH, Rockville, MD, pp. 253–267. Charlier, T.D., Ball, G.F., Balthazart, J., 2005. Inhibition of steroid receptor coactivator-1 blocks estrogen and androgen action on male sex behavior and associated brain plasticity. J. Neurosci. 25, 906–913. Commins, D., Yahr, P., 1984. Adult testosterone levels influence the morphology of a sexually dimorphic area in the Mongolian gerbil brain. J. Comp. Neurol. 224, 132–140. Cooke, B.M., Woolley, C.S., 2005. Gonadal hormone modulation of dendrites in the mammalian CNS. J. Neurobiol. 64, 34–46. Cooke, B.M., Tabibnia, G., Breedlove, S.M., 1999. A brain sexual dimorphism controlled by adult circulating androgens. Proc. Natl. Acad. Sci. U. S. A. 96, 7538–7540. Cooke, B.M., Chowanadisai, W., Breedlove, S.M., 2000. Post-weaning social isolation of male rats reduces the volume of the medial amygdala and leads to deficits in adult sexual behavior. Behav. Brain Res. 117, 107–113. Cooke, B.M., Breedlove, S.M., Jordan, C.L., 2003. Both estrogen receptors and androgen receptors contribute to testosterone-induced changes in the morphology of the medial amygdala and sexual arousal in male rats. Horm. Behav. 43, 336–346. Cramer-Meldrum, E., 2000. Environmental Effects on Differences in Aggressive Behavior Among Tree Lizards, Urosaurus ornatus. MNS, Arizona State University. Crosby, E.C., Showers, M.J.C., 1969. Comparative anatomy of the preoptic and hypothalamic areas. In: Haymaker, W., Anderson, E., Nauta, W.J.H. (Eds.), The Hypothalamus. Thomas, Springfield, Ill, pp. 61–135. Deviche, P., Gulledge, C.C., 2000. Vocal control region sizes of an adult female songbird change seasonally in the absence of detectable circulating testosterone concentrations. J. Neurobiol. 42, 202–211. Fusani, L., Hutchison, J.B., Gahr, M., 2001. Testosterone regulates the activity and expression of aromatase in the canary neostriaturn. J. Neurobiol. 49, 1–8.

Goodson, J.L., Evans, A.K., 2004. Neural responses to territorial challenge and nonsocial stress in male song sparrows: segregation, integration, and modulation by a vasopressin V1 antagonist. Horm. Behav. 46, 371–381. Greenberg, N., 1982. A forebrain atlas and stereotaxic technique for the lizard, Anolis carolinensis. J. Morphol. 174, 217–236. Greenberg, N., MacLean, P.D., Ferguson, J.L., 1979. Role of the paleostriatum in species-typical display behavior of the lizard (Anolis carolinensis). Brain Res. 172, 229–241. Greenberg, N., Scott, M., Crews, D., 1984. Role of the amygdala in the reproductive and aggressive behavior of the lizard, Anolis carolinensis. Physiol. Behav. 32, 147–151. Hews, D.K., Moore, M.C., 1996. A critical period for the organization of alternative male phenotypes of tree lizards by exogenous testosterone? Physiol. Behav. 60, 425–429. Hews, D.K., Knapp, R., Moore, M.C., 1994. Early exposure to androgens affects adult expression of alternative male types in tree lizards. Horm. Behav. 28, 96–115. Hover, E.L., 1985. Differences in aggressive behavior between two throat color morphs in a lizard, Urosaurus ornatus. Copeia 933–940. Knapp, R., Moore, M.C., 1995. Hormonal responses to aggression vary in different types of agonistic encounters in male tree lizards, Urosaurus ornatus. Horm. Behav. 29, 85–105. Knapp, R., Moore, M.C., 1996. Male morphs in tree lizards, Urosaurus ornatus, have different delayed hormonal responses to aggressive encounters. Anim. Behav. 52, 1045–1055. Kollack-Walker, S., Newman, S.W., 1995. Mating and agonistic behavior produce different patterns of Fos immunolabeling in the male Syrian hamster brain. Neuroscience 66, 721–736. Lacy, E.L., Sheridan, M.A., Moore, M.C., 2002. Sex differences in lipid metabolism during reproduction in free-living tree lizards (Urosaurus ornatus). Gen. Comp. Endocrinol. 128, 180–192. Lanuza, E., Halpern, M., 1997. Afferent and efferent connections of the nucleus sphericus in the snake Thamnophis sirtalis: convergence of olfactory and vomeronasal information in the lateral cortex and the amygdala. J. Comp. Neurol. 385, 627–640. Leitner, S., Voigt, C., Garcia-Segura, L.M., Van't Hof, T., Gahr, M., 2001. Seasonal activation and inactivation of song motor memories in wild canaries is not reflected in neuroanatomical changes of forebrain song areas. Horm. Behav. 40, 160–168. MacLean, P.D., 1978. Why brain research on lizards? In: Greenberg, N., MacLean, P.D. (Eds.), Behavior and Neurology of Lizards. NIMH, Rockville, MD, pp. 1–10. McCarthy, M.M., Pfaus, J.G., 1996. Steroid modulation of neurotransmitter function to alter female reproductive behavior. Trends Endocrinol. Metab. 7, 327–333. McIntyre, K.L., Porter, D.M., Henderson, L.P., 2002. Anabolic androgenic steroids induce age-, sex-, and dose-dependent changes in GABA(A) receptor subunit mRNAs in the mouse forebrain. Neuropharmacology 43, 634–645. Moga, M.M., Geib, B.M., Zhou, D., Prins, G.S., 2000. Androgen receptorimmunoreactivity in the forebrain of the Eastern Fence lizard (Sceloporus undulatus). Brain Res. 879, 174–182. Moore, M.C., 1986. Elevated testosterone levels during nonbreeding-season territoriality in a fall-breeding lizard, Sceloporus jarrovi. J. Comp. Physiol., A Neuroethol. Sens. Neural Behav. Physiol. 158, 159–163. Moore, M.C., 1987. Castration affects territorial and sexual behavior of freeliving male lizards, Sceloporus jarrovi. Anim. Behav. 35, 1193–1199. Moore, M.C., Hews, D.K., Knapp, R., 1998. Hormonal control and evolution of alternative male phenotypes: generalizations of models for sexual differentiation. Am. Zool. 38, 133–151. Morrell, J.I., Crews, D., Ballin, A., Morgentaler, A., Pfaff, D.W., 1979. 3Hestradiol, 3H-testosterone and 3H-dihydrotestosterone localization in the brain of the lizard Anolis carolinensis: an autoradiographic study. J. Comp. Neurol. 188, 201–223. Nelson, R.J., Chiavegatto, S., 2001. Molecular basis of aggression. Trends Neurosci. 24, 713–719. Northcutt, R.G., 1967. Architectonic studies of the telencephalon of Iguana iguana. J. Comp. Neurol. 130, 109–148.

D. Kabelik et al. / Hormones and Behavior 49 (2006) 587–597 Nottebohm, F., 1981. A brain for all seasons: cyclical anatomical changes in song control nuclei of the canary brain. Science 214, 1368–1370. O'Bryant, E.L., Wade, J., 2002. Seasonal and sexual dimorphisms in the green anole forebrain. Horm. Behav. 41, 384–395. Panzica, G.C., Viglietti-Panzica, C., Balthazart, J., 1996. The sexually dimorphic medial preoptic nucleus of quail: a key brain area mediating steroid action on male sexual behavior. Front. Neuroendocrinol. 17, 51–125. Riters, L.V., Eens, M., Pinxten, R., Duffy, D.L., Balthazart, J., Ball, G.F., 2000. Seasonal changes in courtship song and the medial preoptic area in male European starlings (Sturnus vulgaris). Horm. Behav. 38, 250–261. Rosen, G., O'Bryant, E., Matthews, J., Zacharewski, T., Wade, J., 2002. Distribution of androgen receptor mRNA expression and immunoreactivity in the brain of the green anole lizard. J. Neuroendocrinol. 14, 19–28. Sartor, J.J., Balthazart, J., Ball, G.F., 2005. Coordinated and dissociated effects of testosterone on singing behavior and song control nuclei in canaries (Serinus canaria). Horm. Behav. 47, 467–476. Schwabl, H., Kriner, E., 1991. Territorial aggression and song of male European robins (Erithacus rubecula) in autumn and spring: effects of antiandrogen treatment. Horm. Behav. 25, 180–194. Smeets, W.J., Hoogland, P.V., Lohman, A.H., 1986. A forebrain atlas of the lizard Gekko gecko. J. Comp. Neurol. 254, 1–19. Smith, G.T., Brenowitz, E.A., Wingfield, J.C., 1997. Roles of photoperiod and testosterone in seasonal plasticity of the avian song control system. J. Neurobiol. 32, 426–442. Smulders, T.V., 2002. Natural breeding conditions and artificial increases in testosterone have opposite effects on the brains of adult male songbirds: a meta-analysis. Horm. Behav. 41, 156–169. Sugerman, R.A., Demski, L.S., 1978. Agonistic behavior elicited by electrical stimulation of the brain in western collared lizards, Crotaphytus collaris. Brain Behav. Evol. 15, 446–469. Tanapat, P., Hastings, N.B., Reeves, A.J., Gould, E., 1999. Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci. 19, 5792–5801.

597

Tarr, R.S., 1977. Role of the amygdala in the intraspecies aggressive behavior of the iguanid lizard, Sceloporus occidentalis. Physiol. Behav. 18, 1153–1158. Thompson, C.W., Moore, M.C., 1991. Throat color reliably signals status in male tree lizards, Urosaurus ornatus. Anim. Behav. 42, 745–753. Thompson, C.W., Moore, M.C., 1992. Behavioral and hormonal correlates of alternative reproductive strategies in a polygynous lizard: tests of the relative plasticity and challenge hypotheses. Horm. Behav. 26, 568–585. Tramontin, A.D., Brenowitz, E.A., 2000. Seasonal plasticity in the adult brain. Trends Neurosci. 23, 251–258. Tramontin, A.D., Wingfield, J.C., Brenowitz, E.A., 1999. Contributions of social cues and photoperiod to seasonal plasticity in the adult avian song control system. J. Neurosci. 19, 476–483. Tramontin, A.D., Perfito, N., Wingfield, J.C., Brenowitz, E.A., 2001. Seasonal growth of song control nuclei precedes seasonal reproductive development in wild adult song sparrows. Gen. Comp. Endocrinol. 122, 1–9. Tramontin, A.D., Wingfield, J.C., Brenowitz, E.A., 2003. Androgens and estrogens induce seasonal-like growth of song nuclei in the adult songbird brain. J. Neurobiol. 57, 130–140. Wade, J., Crews, D., 1991. The relationship between reproductive state and “sexually” dimorphic brain areas in sexually reproducing and parthenogenetic whiptail lizards. J. Comp. Neurol. 309, 507–514. Wade, J., Huang, J.M., Crews, D., 1993. Hormonal control of sex differences in the brain, behavior and accessory sex structures of whiptail lizards (Cnemidophorus species). J. Neuroendocrinol. 5, 81–93. Weiss, S.L., Moore, M.C., 2004. Activation of aggressive behavior by progesterone and testosterone in male tree lizards, Urosaurus ornatus. Gen. Comp. Endocrinol. 136, 282–288. Young, L.J., Lopreato, G.F., Horan, K., Crews, D., 1994. Cloning and in situ hybridization analysis of estrogen receptor, progesterone receptor, and androgen receptor expression in the brain of whiptail lizards (Cnemidoand C. inornatus). J. Comp. Neurol. 347, 288–300.