Effects of estradiol on neural responses to social signals in female túngara frogs

In many female vertebrates, estradiol modulates sexual behavior by influencing sensory, motor and motivational brain regions such that sexual behavior occurs when gonadal activity is high. A principal way that estradiol affects sexual behavior is by regulating neural responses in the basal forebrain, which serves to integrate hormonal and sensory information and, in turn, influences motor circuits underlying sexual behavior. For example, the preoptic area in female rats is estradiol sensitive and enhances lordosis by reducing the excitatory neural impulses that project from the preoptic area to the ventral tegmental area (Sakuma, 2015). Estradiol can also impact sensory systems directly. For example, estradiol shapes auditory tuning in midshipmen fish so that, during the breeding season, females are more sensitive to the harmonics in the calls of courting males, which enhances the adaptive coupling of the sender and the receiver for reproductive success (Sisneros, 2009a,,b,,c; Sisneros et al., 2004). Similar mechanisms of estradiol-dependent modulation of acoustic communication circuits are present in birds in response to plasma estradiol (Maney et al., 2006,, 2008; Maney and Pinaud, 2011; Sanford et al., 2010) and locally produced neuroestrogens (Krentzel and Remage-Healey, 2015; Pawlisch and Remage-Healey, 2015; Remage-Healey, 2012,, 2014; Remage-Healey et al., 2010,, 2008; Remage-Healey and Joshi, 2012).

Anurans serve as excellent models to explore hormonal effects on auditory plasticity because of the wealth of data available on their courtship behaviors and their associated hormone levels (see reviews by Arch and Narins, 2009; Caras, 2013). In túngara frogs [Physalaemus pustulosus (Cope 1864), also known as Engystomops pustulosus], a primary component of female sexual behavior is phonotaxis, or the approach toward a calling male for the purpose of initiating a mating clasp (amplexus). The expression of selective phonotaxis depends on estradiol (Chakraborty and Burmeister, 2009) and estradiol receptors are expressed widely in auditory, motor and motivational brain regions (Chakraborty and Burmeister, 2010). Furthermore, in the auditory midbrain and parts of the limbic forebrain, estradiol receptor expression is sexually dimorphic (Chakraborty and Burmeister, 2010), suggesting that estradiol is an important modulator of neural circuits underlying female-typical behavior. In addition, injections of human chorionic gonadotropins (hCGs), which stimulate production of gonadal steroids (Lynch et al., 2006; Wu et al., 2001), increase neural responses to conspecific social signals in the auditory midbrain (i.e. torus semicircularis) (Lynch and Wilczynski, 2008), suggesting that one way that estradiol modulates behavioral responses to conspecific calls is through modulation of auditory responses (Lynch and Wilczynski, 2008) or through integration of auditory and motor systems (see Hoke and Pitts, 2012). In fact, Hoke and Pitts (2012) propose that a portion of the auditory midbrain (laminar nucleus of the torus semicircularis) is a critical node for initiating behavioral responses to socially relevant sounds in a reproductive context – a finding also supported by another study (Mangiamele and Burmeister, 2011). Additionally, in female green treefrogs (Hyla cinerea), auditory-evoked multi-unit responses are stronger in unmated females compared with females that have recently mated, suggesting that periods of sexual receptivity are accompanied by enhanced sensitivity of the anuran auditory midbrain to facilitate processing of male advertisement calls (Miranda and Wilczynski, 2009).

To better understand how estradiol promotes phonotaxis in túngara frogs – whether through effects on auditory, motor and/or motivational brain regions – we manipulated estradiol levels by injecting estradiol or an estradiol synthesis blocker (fadrozole) and measured neural responses to conspecific or heterospecific calls throughout the brain. We measured neural responses by measuring expression of the transcription factor egr-1 (early growth response 1; synonymous with zif268, NGFI-A, ZENK). Expression of egr-1 is linked to changes in membrane depolarization (via MAP kinase pathways) and it controls expression of many plasticity-related genes (Clayton, 2000). Because of its links to membrane depolarization and synaptic plasticity, the expression of egr-1 is a useful tool for investigating neural responses to behaviorally relevant cues. It also allows measurement of multiple brain regions simultaneously.

We used adult female túngara frogs from a laboratory stock maintained at the University of North Carolina that was originally derived from natural populations near Rio Píro on the Osa Peninsula in Costa Rica. Photoperiod (12 h:12 h, light:dark) and temperature (28°C) approximated natural conditions. Because túngara frogs are small, we manipulated hormones through injection rather than surgery, which meant that females had intact gonads that could produce endogenous steroids. Therefore, we began by inducing breeding so that our subjects would start the experiment with low levels of gonadal activity. To do this, females were injected with 500 IU g⁻¹ body mass of hCG and paired with a male and we provided a water dish to stimulate breeding. We collected 48 females that made nests over a period of 7 days for further hormone manipulations. Ten days after the females had made nests, a time when endogenous gonadal steroids are low (Lynch and Wilczynski, 2005), we raised estradiol levels in half of the females and blocked estradiol synthesis in the other half (Fig. 1) as follows. To block estradiol synthesis, females were injected twice with the estradiol synthesis blocker fadrozole (50 μg each injection in 50 µl), 24 h apart. To raise estradiol, females were injected with saline (50 µl of 0.9% sodium chloride) followed 24 h later by estradiol (0.07 μg g⁻¹ body mass in 50 µl); the purpose of the initial saline injection was to match the number of injections in the fadrozole group. All injections were given in the afternoon.

We were not able to collect blood plasma from the females in this experiment to document the effects of our hormone manipulations because blood collection is incompatible with brain collection in this species owing to the small blood volume (i.e. they do not bleed enough following decapitation). However, these hormone manipulations have been validated previously for túngara frogs. Specifically, two injections of fadrozole are effective at inhibiting estradiol synthesis in hCG-treated females (Chakraborty and Burmeister, 2009; Lynch and Wilczynski, 2005). In addition, when measured 24 h later, a single injection of estradiol at this dose is sufficient to raise estradiol levels above fadrozole-treated females and increases phonotaxis behavior (Chakraborty and Burmeister, 2009). It is important to note that fadrozole would affect both neural and gonadal synthesis of estradiol. Likewise, our estradiol injections likely influence pituitary gonadotropins through negative feedback.

After the final injection, we placed pairs of females, one from each hormone treatment into acoustic isolation chambers; individual females in a pair were maintained separately in individual mesh containers. After an 18 h acclimation period in the dark, we presented the female pairs with a single exemplar of a conspecific (N=24) or a heterospecific call (N=24) for 30 min (see below for stimulus details). We interspersed females assigned to each acoustic treatment across days and chambers. Females were rapidly decapitated 1 h after onset of stimuli, which corresponds to peak accumulation of acoustically induced egr-1 mRNA expression (Burmeister et al., 2008) but occurs before habituation of the egr-1 response is evident (R. M. Glaeser, L. A. Mangiamele and S.S.B., unpublished results). After decapitation, we opened the skull to fix the brains (10 min in 4% paraformaldehyde) before removing them. We froze the brains in liquid nitrogen in 2 ml tubes containing Tissue-Tek OCT compound (Sakura, Finetek, Torrance, CA) and stored at −80°C until sectioning. The University of North Carolina Institutional Animal Care and Use Committee (IACUC) approved the experimental procedures.

Hormonally manipulated females were presented with a conspecific call recorded in the Osa Peninsula or a heterospecific (Physalaemus enesefae) call from Venezuela (Fig. 2). The túngara frog produces a two-note advertisement call (whine-chuck) (Ryan, 1985) whereas P. enesefae only produces a single-note call (whine). The whines of both species are characterized by a downward frequency sweep that spans similar frequencies but they do so with very different time signatures (Fig. 2). Female túngara frogs do not recognize the P. enesefae whine as an acceptable sexual signal (e.g. Ryan et al., 2003). In both cases, the call was repeated every 2 s to reflect the calling rate of P. pustulosus and stimulus amplitude was 82 dB SPL at a distance of 20 cm from the speaker. The speaker was a midlevel woofer (Misco JC46W-4A, Minneapolis Speaker Company, Minneapolis, MN, USA) that was connected to an AMP Five series monoblock amplifier (Audiosource, Portland, OR, USA) and a Macintosh computer. We controlled playback using ProTools audio software (v7.3; Digidesign, Daly City, CA, USA). Using these exemplars, we have previously demonstrated that the conspecific call, but not the heterospecific call, induces significant expression of egr-1 compared with no sound in the auditory midbrain, thalamus and limbic forebrain of túngara frog females (Chakraborty et al., 2010).

Brains were sectioned in the transverse plane at 16 µm in three series on a cryostat. To localize egr-1 mRNA, radioactive in situ procedures were used as previously described (Burmeister et al., 2008). Briefly, radioactively labeled sense and antisense probes were generated from reverse transcription of a 309-nucleotide subclone of P. pustulosus egr-1. Before hybridization, the tissue was fixed for 10 s in 4% paraformaldehyde and washed in phosphate-buffered saline, triethanolamine, acetic anhydride, 2× SSC and a series of ethanols. We hybridized the tissue with 90 µl of 3.0×10⁵ cpm ml⁻¹ hybridization buffer at 65°C overnight. We removed unbound probe with a series of 65°C washes, first in 50% formamide and 2× SSC (1.25 h) followed by two washes in 0.1× SSC (30 min each). We visualized the bound probe as silver grains by exposing the slides to NTB emulsion diluted 1:1 in distilled water for 14 days at 4°C, and we visualized cell bodies by staining with thionin. Tissue incubated with the sense probe showed no significant binding above background.

We followed the nomenclature of O'Connell et al. (2011,, 2010). egr-1 was measured in brain regions that are known to receive significant auditory input as well as those that are part of the anuran social behavior network (Newman, 1999; O'Connell et al., 2010): the auditory midbrain (principal and laminar nuclei of the torus semicircularis), parts of the thalamus that receive ascending auditory input (posterior, central and anterior thalamic nuclei) and parts of the limbic forebrain that either receive auditory input and/or are directly implicated in modulating phonotaxis (preoptic area, striatum, lateral septum, dorsal part of the medial pallium and nucleus accumbens). Estradiol receptors are expressed at detectable levels in all these brain regions, except for the posterior thalamus (Chakraborty and Burmeister, 2010).

Our quantification method was described previously (Burmeister et al., 2008; Chakraborty et al., 2010) and was derived from Mize (1994), Masseroli et al. (1993) and Lucas et al. (1994). Briefly, for each brain region, an individual's mean was calculated from three brain sections. For each section, we captured three images at 630× magnification: a color image of Nissl-stained tissue in the region of interest, a blue-filtered image of only the silver grains in the same field of view (grains image) and a blue-filtered image of an area of the slide adjacent to the tissue to represent local background silver grain density (background image). We measured local background silver grain density to account for emulsion thickness, which varies across the slide and influences silver grain density. In the blue-filtered images that we used to measure silver grain density, exposure, brightness and contrast settings were the same for each picture of a given section. ImageJ (http://rsbweb.nih.gov/ij/) was used to convert the grains and background images to binary and to count the silver grains in each image using the ‘analyze particles’ feature. The number of background silver grains was subtracted from the number of silver grains in the region of interest to obtain the number of silver grains above background per image. We then used the point selection tool in ImageJ to mark and count all visible cells in the color image of the region of interest. Cells were identified visually by their clearly stained soma and the cell counting was performed by experienced individuals who were blind to treatment. Our final measure of egr-1 mRNA expression for each section was the number of silver grains above background per cell. The quality of the brain sections influenced the sample sizes reported for each brain region, as we were unable to obtain data from all individuals in the study for all brain regions sampled.

To test whether hormone treatment, call type or their interaction influenced egr-1 expression, we conducted separate two-way analysis of variances (ANOVAs) for each brain region with call type, hormone treatment and their interaction as between-subjects factors. Although brain regions are not independent of one another, we could not account for covariation among brain regions with a multivariate analysis (e.g. repeated-measures ANOVA) because missing values for individual brain regions would result in the exclusion of most subjects. For the same reason, missing samples prevented us from including ‘pair’ as a variable in the analyses.

The conspecific call elicited higher egr-1 expression than the heterospecific call in the auditory midbrain, thalamus and limbic forebrain (main effects for call type, see Table 1, Figs 3–5), demonstrating that conspecific social signals are highly salient elicitors of robust responses throughout the brain, largely independent of hormone status. When compared with fadrozole, injections of estradiol increased egr-1 expression in the auditory midbrain and the limbic forebrain, but not in the thalamus (main effects of hormone treatment, see Table 1, Figs 3–5). That is, estradiol increased egr-1 expression in the midbrain and forebrain independently of the acoustic environment, suggesting that estradiol increases the basal activity of these brain regions.

For most brain regions, the main effects of call type and hormone were independent of one another, and both conspecific calls and estradiol increased egr-1 expression, creating an additive effect. In the preoptic area, however, the effect of call type depended on the hormonal status of the animal (F_1,32=4.1, P=0.05). That is, estradiol enhanced the effect of the conspecific call, suggesting that the preoptic area may have a particularly important role in integrating hormonal and social cues. We observed a similar pattern for the septum, but with less robust statistical support (F_1,29=3.5, P=0.07).

We have previously shown that estradiol is necessary and sufficient for the expression of phonotaxis in the túngara frog (Chakraborty and Burmeister, 2009) and that estradiol receptors show sexually dimorphic expression in the auditory midbrain and parts of the limbic forebrain (Chakraborty and Burmeister, 2010), indicating that estradiol could mediate changes in phonotaxis through direct effects on these brain regions. Consistent with this, we now show that estradiol and conspecific calls both increase egr-1 in the auditory midbrain and parts of the limbic forebrain, including motor (striatum) and motivational (preoptic area, septum, and nucleus accumbens) brain regions. In the preoptic area, the effect of conspecific calls depended on the hormonal status of the animal. Such an effect is consistent with behavioral responses to conspecific calls, which only occur in the presence of estradiol (Chakraborty and Burmeister, 2009), raising the possibility that the preoptic area acts as a hormonal gatekeeper to reproductive behavior in the túngara frog.

Our results for the auditory midbrain parallel a number of studies from other taxa that describe an important role for estradiol in sensory processing of social signals and of audition (Arch and Narins, 2009; Caras, 2013; Earp and Maney, 2012; Maney, 2013; Maney et al., 2006,, 2008; Maney and Pinaud, 2011; Petrulis, 2013; Remage-Healey, 2012,, 2014; Sanford et al., 2010; Sisneros, 2009c; Tremere et al., 2009; Walpurger et al., 2004; Yovanof and Feng, 1983). In the túngara frog, a previous study showed that hCG, which stimulates production of gonadal steroids (Lynch et al., 2006; Wu et al., 2001), increases egr-1 expression in the laminar nucleus of the torus (Lynch and Wilczynski, 2008). Our findings suggest that the effect observed by Lynch and Wilczynski (2008) was mediated by estradiol and not by direct effects of gonadotropins (Yang et al., 2007) or through changes in other steroids, such as progesterone or androgens. The increased expression of egr-1 in response to estradiol might reflect a lower behavioral threshold for approaching calls: when females are nearing oviposition, a time when gonadal hormones are high, they are willing to approach calls that are typically unattractive, although their ability to discriminate among calls remains intact (Lynch et al., 2006). The limited electrophysiological data support this conclusion. That is, both reproductive activity (Miranda and Wilczynski, 2009) and estradiol treatment (Yovanof and Feng, 1983) increase auditory responses to sounds in frogs, including biologically irrelevant sounds (noise burst, tones) as well as conspecific calls. There are also broad correspondences between egr-1 (also known as ZENK) expression and electrophysiological activity in songbirds: ZENK expression is selective for conspecific social signals (song) over tones only in the presence of estradiol in female white-throated sparrows (Zonotrichia albicolis) (Maney et al., 2006,, 2008; Maney and Pinaud, 2011). Estradiol also enhances auditory encoding of song, sometimes through rapid effects of brain-derived estradiol (Krentzel and Remage-Healey, 2015; Remage-Healey, 2012,, 2014; Remage-Healey et al., 2010; Tremere and Pinaud, 2011). In zebra finches, the impact of estradiol extends further to the song control system, as estradiol enhances the neural selectivity for the bird's own song (BOS) but not to conspecific songs and fadrozole injections reduce this selectivity (Remage-Healey and Joshi, 2012). Thus, effects of estradiol on auditory processing of social signals are manifest at multiple levels (i.e. MAPK-dependent gene expression and electrophysiological activity) and across vertebrate classes.

An interesting pattern that we observed was that estradiol affected egr-1 expression in the midbrain and limbic forebrain, but not the thalamus. This was in spite of the fact that estradiol receptors are expressed in parts of the thalamus (central and anterior nuclei; Chakraborty and Burmeister, 2010) and the primary auditory afferent to the thalamus (i.e. the torus) was strongly affected by estradiol. Although intriguing, the implications of these results will remain unclear until we have a more complete understanding of the effects of estradiol on electrophysiological responses to conspecific calls. At this time, however, we can conclude that the thalamus appears to lack estradiol-induced auditory plasticity.

We found that estradiol had potent effects on egr-1 expression in the preoptic area and other parts of the limbic forebrain that are thought to modulate socially relevant behavioral responses. While the preoptic area is central to sexual behavior in all vertebrates, a role of the preoptic area in anuran sexual behavior has been less well studied. It has long been known that the anuran preoptic area receives auditory input and that it is sensitive to conspecific calls (Allison, 1992). We recently demonstrated that the anuran preoptic area expresses estradiol receptors (Chakraborty and Burmeister, 2010) and we now show that estradiol modifies its response to conspecific calls (present study). Together, these data indicate that, like other vertebrates, the anuran preoptic area influences reproductive behaviors by integrating sensory and hormonal information. Furthermore, the fact that the effect of estradiol on the preoptic area mirrors that of estradiol on female phonotaxis suggests that the preoptic area may play a particularly important role in female phonotaxis.

In summary, we found that both conspecific calls and estradiol increased egr-1 expression in the auditory midbrain and limbic forebrain. Together with previous behavioral (Chakraborty and Burmeister, 2009) and electrophysiological studies (Yovanof and Feng, 1983), our results suggest that estradiol could promote selective phonotaxis by lowering the threshold for responses in both the midbrain and limbic forebrain, while highlighting a role for the preoptic area. The effect of estradiol at multiple points in the neural circuit for sexual behavior in túngara frogs is part of a common theme in vertebrates for the influence of steroids on social behavior.

We thank Lauren Eberly, Josh Thompson, Suhana Sarkar, Stuart Jeckel and Mihnea Mangalea for assistance with image analysis, Lisa Mangiamele for the recordings of the P. pustulosus calls and Dr Zaida Tárano for the recordings of the P. enesefae calls. We also thank Christina Lebonville for help with the experiment.