Expression of oxytocin receptors in the zebra finch brain during vocal development

Abstract

Like human language, song in songbirds is learned during an early sensitive period and is facilitated by motivation to seek out social interactions with vocalizing adults. Songbirds are therefore powerful models with which to understand the neural underpinnings of vocal learning. Social motivation and early social orienting are thought to be mediated by the oxytocin system; however, the developmental trajectory of oxytocin receptors in songbirds, particularly as it relates to song learning, is currently unknown. This gap in knowledge has hindered the development of songbirds as a model of the role of social orienting in vocal learning. In this study, we used quantitative PCR to measure oxytocin receptor expression during the sensitive period of song learning in zebra finches (Taeniopygia guttata). We focused on brain regions important for social motivation, attachment, song recognition, and song learning. We detected expression in these regions in both sexes from post-hatch day 5 to adulthood, encompassing the entire period of song learning. In this species, only males sing; we found that in regions implicated in song learning specifically, oxytocin receptor mRNA expression was higher in males than females. These sex differences were largest during the developmental phase when males attend to and memorize tutor song, suggesting a functional role of expression in learning. Our results show that oxytocin receptors are expressed in relevant brain regions during song learning, and thus provide a foundation for developing the zebra finch as a model for understanding the mechanisms underlying the role of social motivation in vocal development.

Introduction

In social species, including humans, early social learning depends on preferential orienting to social cues. For example, humans learn language by attending to speaking adults during an early sensitive period. Disruption of social orienting during this period is thought to delay language development (Chevallier et al., 2012; Mundy & Burnett, 2005). Thus, understanding the mechanisms underlying early social processes, specifically their neural bases, may shed light on how these processes contribute to early vocal learning. Animal models are needed, but the most common models in neuroscience, which are primarily rodents and non-human primates, are not known to engage in vocal production learning. Songbirds, for whom song is learned from adults during an early sensitive period, thus present an important opportunity to develop relevant models. Juvenile male zebra finches (Taeniopygia guttata) memorize and learn to sing the song of their male caregiver, or “tutor”, during the first ~100 days of life (Fig. 1; reviewed by Gobes et al., 2019). Multiple lines of evidence indicate that social orienting, recognition, and motivation are important for song learning outcomes. For example, young males preferentially learn the song of their caregiver over that of other males (Immelmann, 1969; Eales, 1987). They will press keys to hear their caregiver’s song in the absence of any unconditioned reward (Adret 1993; Rodriguez-Saltos et al., 2021; Tchernichovski et al., 1999), and the degree to which they prefer to hear that song, over other songs, predicts how well they learn it (Rodriguez-Saltos et al., 2021).

Fig. 1. Developmental trajectory of song learning in zebra finches.

Zebra finches fledge the nest at about post-hatch day 20 (P20); before this time, there is little evidence that they are able to form memories of tutor song (see Gobes et al., 2019 for review). By P25, the critical period of auditory learning has opened, and the young males begin to attend to and memorize male songs. They begin practicing song-like vocalizations, called subsong, a few days later (Arnold, 1975). By P35, males will perform work for access to song (particularly the tutor’s song); they will key-press to hear it (Tchernichovski et al., 1999). Preference for hearing tutor song peaks, on average, before P50; by P55, young males shift their auditory preference from tutor song to the songs of other males (Rodriguez-Saltos et al., 2021; see also Fujii et al., 2021). At about the same time, they enter the “plastic song” phase, during which their song rate increases dramatically (Johnson et al., 2002) and their song begins to resemble tutor song (Kollmorgen et al., 2020). After about P77, a male’s song is highly similar to its final form (Johnson et al., 2002), which will not change substantially in adulthood; it is becoming “crystallized”.

Several independent groups of researchers, including ours, have proposed that motivation to seek out and attend to tutor song is mediated by nonapeptide receptors (Baran, 2017; Maney & Rodriguez-Saltos, 2016; Syal & Finlay, 2011; Theofanopoulou et al., 2017). These include the oxytocin receptor (OTR), the vasopressin 1a receptor (V1aR), and the vasopressin 2 receptor (V2R), all of which have clear homologs in birds (Gubrij et al., 2005; Kelly & Goodson, 2014; Leung et al., 2011; note that the latter two are also known as VTR1A and VTR2A, respectively, see Theofanopoulou et al., 2021). These molecules are widely understood to mediate social motivation and attachment (Gordon et al., 2011), for example in the context of pair bonding, across many vertebrate taxa including birds (Klatt & Goodson, 2013; Pedersen & Tomaszycki, 2012). Further, there is evidence that in both birds and mammals, nonapeptide receptors act as a substrate through which the early social environment shapes social development during early sensitive periods (Baran, 2017; Baran et al., 2016; 2017; Hammock, 2015; 2018; Miller & Caldwell, 2015; Vaidyanathan & Hammock, 2017; Veenema, 2012).

In zebra finches, nonapeptide receptors are known to be expressed in regions of the brain critical for social attachment, song recognition, and song learning (Leung et al., 2009; 2011). Studies of the distributions of these receptors, however, have been conducted only in adults. We therefore do not know where, when, or the extent to which these receptors are expressed during the sensitive period of song learning, nor do we understand how their developmental trajectories relate to the milestones of vocal development (Fig. 1). This lack of knowledge hinders the development of this species as a model to study, on a molecular level, how these receptors contribute to vocal learning. Quantifying their expression in juveniles, with temporal resolution sufficient to characterize their developmental trajectory, is a crucial first step in understanding how nonapeptide systems may influence periods of behavioral change in this potentially powerful model.

In this study, we quantified the expression of nonapeptide receptor mRNA in zebra finches from post-hatch day 5 (P5) to adulthood (P95), which covers the entire period of song learning (Fig. 1). We quantified mRNA expression in four brain regions, chosen partly on the basis of their roles in social motivation and attachment, song recognition, and song learning. Because of the clear role of oxytocin in social reward and attachment across species, we focused on OTR in particular; therefore, our regions of interest were also chosen partly on the basis of known OTR expression in adult zebra finches (Leung et al., 2009; 2011) and other songbirds (Leung et al., 2009; 2011; Ondrasek et al., 2018; Voorhuis et al., 1990; Wilson et al., 2016). Below, we explain our rationale for the selection of these four regions for this study.

Social motivation and attachment.

In birds, the OTR population in the lateral septum (LS) is thought to be important for social memory and social motivation (Goodson et al., 2009; Lukas et al., 2013). In zebra finches, administration of an oxytocin antagonist into this region reduced gregariousness, eliminating preferences for large group sizes (Goodson et al., 2009). This treatment also disrupted pair bond formation between males and females (Klatt & Goodson, 2013; Pedersen & Tomaszycki, 2012). OTR in LS is therefore a top candidate for mediating motivation to interact socially, for example with song tutors, early in life.

Song recognition.

Recent evidence from rodents indicates that oxytocin acts within sensory systems to enhance the perceptual salience of conspecific communication signals (reviewed by Hammock, 2018). In songbirds, the auditory forebrain has been hypothesized to store acoustic memories of song (Bolhuis & Moorman, 2015; Yanagihara & Yazaki-Sugiyama, 2016) as well as to encode its attractiveness (Maney, 2013; Maney & Rodriguez-Saltos, 2016; Rodriguez-Saltos et al., 2018). Thus, the dense OTR population in this region (Leung et al., 2009; 2011) could be an important substrate by which social cues shape learning.

Song learning.

OTR expression is found in several regions that regulate the learning and production of song (Leung et al., 2009; 2011). Among these is the cortical nucleus HVC (used as a proper name, see Reiner et al., 2004). HVC projects directly into the major motor pathway for song production as well as into a pathway necessary for song learning (reviewed by Brenowitz et al., 1997). The latter pathway is directly interconnected with a second region of dense OTR, the dorsal arcopallium (Ad; Leung et al., 2009; 2011; Voorhuis et al., 1990). Ad is thought to contribute to song learning by comparing memories of tutor song with attempts to mimic it (Bottjer & Altenau, 2010; Doupe, 1993). OTR populations in these areas could thus participate in the formation and maintenance of memories of tutor song.

Hypotheses and approach.

It is currently unknown where, when, and to what extent nonapeptide receptors are expressed in juvenile zebra finches during the sensitive period of song learning. Our main objective in this study was to quantify OTR expression and characterize its developmental trajectory in order to establish this species as a potential model for the development of social motivation and its contributions to vocal learning. We hypothesized that nonapeptide receptors would be expressed in juvenile zebra finches, particularly during the processes of tutor choice and early memorization of song. Because in this species only males learn to sing, we quantified expression in both sexes in order to test for sex differences that might be relevant to that learning. We expected to find higher expression in males in brain regions involved specifically in vocal production learning—namely HVC and Ad. In contrast, we had no reason to expect sex differences in other regions, since the processes of social attachment and song recognition are likely similar in both sexes (Adret, 1993; Riebel, 2003).

We quantified the expression of mRNA using quantitative real-time PCR in micropunched brain tissue, which does not allow visualization of the distribution of binding within the samples. In order to document the pattern of receptor expression in our regions of interest and confirm the expression of the protein in those regions, we also performed receptor autoradiography on a small number of animals at each age. The latter approach allowed us to compare, qualitatively, the general distribution of labeling between juveniles and adults (see Leung et al., 2009), and to determine the approximate age at which adult-like patterns of binding emerge.

Methods

Animals

All procedures were approved by the Emory University Institutional Animal Care and Use Committee and were in compliance with the US National Research Council's Guide for the Care and Use of Laboratory Animals, the US Public Health Service's Policy on Humane Care and Use of Laboratory Animals, and Guide for the Care and Use of Laboratory Animals. Juvenile zebra finches were produced by a breeding colony housed in a mixed-sex aviary (12h light, 12h dark) at Emory. All birds were free to engage in social interactions, and parents had constant access to their offspring.

We used 138 birds in total. Brains were harvested from juveniles at post-hatch days P5, P15, P25, P35, P45, P55, P65, P75, P85, and P95. For qPCR, five males and five females were used for each age group except for P65 females (n=4) and P95 males (n=3). For the autoradiography, two males and two females were used for each age group; a single male at age P40 was also included. Birds at least 45 days old were sexed by plumage and the presence of an ovary or testis. Younger birds were sexed using PCR analysis of a feather or liver sample (Griffiths et al., 1998). All birds were sacrificed by isoflurane overdose and rapidly decapitated. The brains of P5 birds were left in the skull; brains of all older birds were dissected from the skulls. Brains were flash-frozen in powdered dry ice and stored at −80°C until sectioning.

Quantitative real-time PCR

Brains were sectioned on a cryostat in the coronal plane. We took alternating 60μm and 300μm sections. The 60μm sections were Nissl-stained to provide a map of each 300μm section. We used the Palkovits punching technique (Palkovits & Brownstein, 1983) to extract brain tissue from the frozen 300μm sections under a dissecting microscope. All punches were taken with punch tools chilled on dry ice, and punches were immediately ejected into microtubes on dry ice. We sampled four brain regions (see Introduction): the septal area (targeting LS but also likely containing a portion of the medial septum), the auditory forebrain, the dorsal telencephalon (dTel) containing the HVC area, and Ad (Fig. 2). For LS, two 0.5mm punches were collected from the septal area in two adjacent sections, on either side of the midline. For the punches of auditory forebrain, we targeted the region that responds maximally to playback of conspecific song (see Mello & Clayton, 1994). Our sample excluded L2, which does not express appreciable OTR (Leung et al., 2011), but likely contained portions of L3 and the caudal nidopallium (see Vates et al., 1996). A 1mm punch was collected from the auditory forebrain from the same hemisphere of two consecutive sections. For each tissue sample of dTel and Ad, three side-by-side 0.5mm punches were collected, again from the same hemisphere of two consecutive sections. For regions that were sampled from one hemisphere only (auditory forebrain, dTel and Ad), the hemisphere for which the borders of each region were most clearly visible in the Nissl-stained sections was selected to punch.

Fig. 2. Locations of micropunches for quantitative PCR.

Nissl-stained sections are shown to indicate the locations of our punches of fresh frozen 300μm sections. The punch of auditory forebrain was 1mm; other punches were 0.5mm. Ad, dorsal arcopallium. dTel, dorsal telencephalon.

Because of the small size of the P5 brains, micropunches of P5 birds were obtained using slightly different methods, as follows: for the septal area, a single 0.5mm punch was obtained directly on the midline in two adjacent sections. For the auditory forebrain, a 1mm punch was obtained from each hemisphere of the same section. For dTel and Ad in P5 birds, two side-by-side 0.5mm punches were obtained from each hemisphere of the same section. All punches were stored at −80°C for a maximum of 2 days before RNA isolation.

We extracted RNA from each punch using the Qiagen Allprep RNA/DNA micro kit with modifications (Zinzow-Kramer et al., 2015). To produce cDNA, we performed reverse transcription using the Transcriptor First Strand cDNA synthesis kit with random hexamer primers. We designed exon-spanning primers (Table S1) for use with probes from the Roche Universal Probe Library for OTR, V1aR, and V2R, then verified the amplified sequences via cloning and sequencing. qPCR was performed using a Roche LightCycler 480 Real-Time PCR System, in triplicate for each sample on a 384-well plate as previously described (Zinzow-Kramer et al., 2014). The target genes (OTR, V1aR, and V2R) and reference genes were always run on the same plate for each sample, and sex and age were balanced across plates. The reference genes were glyceraldehyde 3-phosphate dehydrogenase (GADPH) and peptidylprolyl isomerase A (PPIA), which have been previously validated for use in zebra finch brain tissue and are stably expressed across sex (Zinzow-Kramer et al., 2014). Using the LightCycler 480 Software Version 1.5.0, we calculated crossing point values using the Abs Quant/2nd Derivative Max method. We normalized the expression of each gene of interest to the geometric mean of the two reference genes (Pfaffl, 2001; Vandesompele et al., 2002). For samples with a coefficient of variance above 21.8% within triplicates, indicating technical error, we identified and removed one outlying point within the triplicate (Bookout et al., 2006) using a modified Dixon’s q test (Dean & Dixon, 1951).

All statistical analyses of qPCR data were conducted in SPSS v26. For the data within each gene and region, we removed extreme outliers using the interquartile range with 3.0 as a multiplier. We then performed univariate ANOVAs for each gene, with region, sex, and age as between-subjects factors. When three-way interactions were found (interactions between sex and age that depended on region), we then conducted univariate ANOVAs within each region to test for main effects of age and sex and interactions between them. When sex x age interactions were found, we followed this ANOVA with post-hoc comparisons (t-tests) to test for effects of sex at each age. To control for qPCR batch effects, qPCR plate was included as a factor in all ANOVAs.

Because in this species only males sing, we were interested in the extent to which sex explained variation in OTR expression at key milestones of vocal development. In order to increase our power to detect sex differences relevant to song learning, we pooled ages within the learning phases and made planned comparisons between males and females. Our groupings (see Fig. 1 for our rationale) were “nestling” (P5, P15), “auditory learning” (P25, P35, P45), “plastic song” (P55, P65, P75) and “crystallization” (P85, P95). We performed t-tests for effects of sex within each phase.

Receptor autoradiography

Four brains from each age group (one of each sex in the coronal plane and one of each sex in the sagittal plane) were sectioned on a cryostat at a thickness of 20μm. One P40 male was sectioned on the coronal plane. Sections were thaw-mounted onto microscope slides in six parallel series. One series from each brain was used for receptor autoradiography as described by Leung et al. (2009). The ligand was ¹²⁵I-ornithine vasotocin (d(CH2)5[Tyr(Me)2,Thr4,Orn8,[¹²⁵I]Tyr9-NH2]; catalog No. NEX254050UC; Perkin Elmer), which was developed for use in rodent brain tissue (Elands et al., 1988) and has been used to label OTR in several species of primate (reviewed by Freeman & Young, 2016). Although the ligand is highly selective for OTR in rodents, it was more recently shown to bind to both OTR and V1aR in primates (see Freeman & Young, 2016). Songbirds have at least three different nonapeptide receptors that are expressed in brain: OTR, V1aR, and V2R (Leung et al., 2011; see also Theofanopoulou et al., 2021 for slightly different nomenclature). In adults, the pattern of ¹²⁵I-OVTA labeling strongly resembles that of OTR mRNA, not V1aR or V2R mRNA, in our regions of interest (Leung et al., 2011; Fig. S1). Therefore, we interpret ¹²⁵I-OVTA labeling in these regions as predominantly binding OTR.

Labeled sections were exposed to Kodak BioMax maximum resolution film for seven days. The film was then developed in a Konica SRX101-A developer and scanned at 2500 dpi using a digital Epson V700 scanner. A second series of sections from each brain was Nissl-stained in order to visualize major landmarks and verify the locations of label in the autoradiographic films.

Results

Developmental trajectories of receptor expression

Because of the well-known role of OTR and related receptors in social motivation (Gordon et al., 2011), particularly during early sensitive periods of development (Hammock, 2018), we hypothesized that these receptors are expressed in juvenile zebra finches during times when they seek out and attend to song tutors (Fig. 1). We tested this hypothesis by characterizing the developmental trajectory of OTR mRNA expression over the course of vocal learning. Toward this goal, we used qPCR to measure OTR mRNA expression in four brain regions with known roles in social motivation, song recognition, and vocal learning (see Introduction). We performed this analysis at ten different ages, spanning a developmental window from five days post-hatch (P5) to near-adulthood (P95), a period that encompasses the entire sensitive period during which song is learned in this species (see Fig. 1).

OTR mRNA was detectable in each of our regions of interest at each age. An ANOVA that included age, sex and region showed a significant interaction between region and age (F_{27, 18} = 3.337; p = 0.005), indicating that the developmental trajectories of receptor expression differed among the regions of interest (Fig. 3). We then proceeded to test within each region for a main effect of age, that is, a significant change in expression over time (Table 1). We detected significant main effects of age for the auditory forebrain and Ad. These effects were likely driven by a steep decrease in expression at P55 in the auditory forebrain (Fig. 3B) and peaks at P25 and P75 in Ad (Fig. 3D). Thus, we have evidence that the expression of OTR changes over the course of development in two regions associated with tutor song memory. In each of these regions, expression was relatively high prior to the transition from auditory learning to plastic song, then decreased during the transition to rise again during the plastic song phase. This finding is consistent with a role for OTR in enhancing selective responses to conspecific vocalizations (see Kanwal & Rao, 2002; Marlin et al., 2015).

Fig. 3. Developmental trajectories of oxytocin receptor mRNA expression in four brain regions of interest.

(A-D) Means and SEM are shown for males (black) and females (white) at ten ages post-hatch. The background colors correspond to the developmental phases shown in the lower panels. (E–H) Means and SEM are shown for males and females pooled within four developmental phases: nestling phase (P5, P15), early auditory learning/subsong (P25, P35, P45), plastic song (P55, P65, P75) and crystallization (P85, P95). See Fig. 1 for details on how the phases were defined. For all panels, expression was normalized to two reference genes for analysis, then normalized to the overall mean expression within each brain region for visualization. Thus, 1.0 on the Y axis corresponds to the average across both sexes and all ages for each region, and region-to-region variation is not represented here. dTel, dorsal telencephalon (includes HVC when HVC is present). Ad, dorsal arcopallium. *p < 0.05.

Table 1.

Results of univariate F-tests within gene and region, showing main effects of age (P5 – P95) and sex on expression of oxytocin receptor, and interactions between age and sex, in the four regions of interest. Significant effects are shown in bold and marked by asterisks. Ad, dorsal arcopallium

	Age		Sex		Age x Sex
Region	F	p	F	p	F	p
Septum	0.991	0.482	0.769	0.393	0.940	0.510
Auditory forebrain	4.441	0.003*	1.980	0.176	1.076	0.423
Dorsal telencephalon	2.055	0.112	8.078	0.012*	1.394	0.269
Ad	17.509	<0.001*	1.482	0.247	8.639	0.001*

Because nonapeptides are likely to bind somewhat promiscuously to OTR and the related receptors V1aR, and V2R (reviewed by Theofanopoulou, 2021), we also quantified expression of V1aR and V2R mRNA in the same tissue samples (Figs. S2 and S3). We emphasize that OTR is likely the predominant nonapeptide receptor in our regions of interest (see Fig. S1 and the autoradiographic binding results below). We could not detect effects of region on the developmental trajectory of expression for either receptor (age x region interactions not significant) and therefore did not proceed to tests for age effects within region. We did note a potential decline in expression of V1aR mRNA in the auditory forebrain in both sexes from P5 to P15, with expression remaining low in this region throughout the rest of the developmental period (Fig. S2B). The role of V1aR expression in auditory development, prior to fledging, needs further investigation.

Sex differences in receptor expression

In this species, only males sing. We therefore included sex as a factor in all ANOVAs and looked in particular for sex-specific changes in expression over time, that is, interactions between sex and age. We hypothesized that males would express more OTR than females, particularly in regions of interest related specifically to learning to sing (i.e., HVC and Ad). Because females also form social bonds, learn to recognize particular songs, and are attracted to song as juveniles (Riebel, 2003), we had no a priori reason to predict sex differences in the septal or auditory areas.

For OTR mRNA we found a significant sex x age x region interaction (F_{24, 18} = 2.911; p = 0.012), meaning that there were sex-specific effects of age that depended on region. We therefore proceeded to test within each region for interactions between sex and age, which would indicate a sex difference at particular age(s). We found such an interaction only for Ad (Table 1). Post-hoc tests showed that males expressed significantly higher OTR mRNA in this region than females at P25 (p = 0.033), early during the auditory learning phase (Fig. 3D). This result is consistent with a role for Ad in the formation of auditory memories early during song learning (Bottjer & Altenau, 2010).

For the other two receptors, V1aR and V2R, we could detect no sex-specific effects of age that depended on region. In other words, the three-way interaction between sex, age, and region was not significant for the expression of either mRNA. We therefore did not proceed to test within region for sex differences in their developmental trajectories. We did, however, note a significant main effect of sex for V1aR (F_{1, 13}; p = 0.019), with expression higher in males. This finding could be relevant to the well-known sex difference in immunoreactivity for the endogenous ligand of this receptor, vasopressin (also called vasotocin). In most avian species yet studied, males have higher levels in the bed nucleus of the stria terminalis, one of the major vasopressinergic cell groups (Grossman et al., 2002).

One of our main objectives in this study was to test for sex differences in receptor expression at time points relevant to the developmental trajectory of vocal learning. For example, during the auditory learning phase (Fig. 1), young males actively seek out tutoring opportunities, and neurons in HVC develop selective responses to tutor song (Adret et al., 2012; Nick & Konishi, 2005). Later, in the plastic song phase, motivation to hear tutor song diminishes (Rodriguez-Saltos et al., 2021; Fujii et al., 2021) and males begin practicing singing (reviewed by Gobes et al., 2019; see also Johnson, 2002). Our strategy of sampling every ten days, although systematic, did not take such milestones into consideration and did not allow powerful tests of our a priori hypotheses about sex. Therefore, we pooled ages within the phases of vocal development to test the extent to which birds that learn to sing (males) differed from those who do not (females) during each phase. Using this strategy, we detected a significant sex difference in OTR mRNA in dTel (containing HVC), with expression higher in males, during auditory learning (p = 0.041; Fig. 3G). We detected a sex difference during auditory learning in Ad as well (p = 0.029; Fig. 3H). These findings are consistent with a model in which expression of OTR supports song learning in males. We did not detect significant sex differences in OTR mRNA in the septal or auditory regions, nor did we find differences in any of the regions for the other two receptors (Figs. S2, S3).

Distribution of oxytocin antagonist binding

Our qPCR analysis allowed us to quantify mRNA expression and compare those levels across age and sex during development. Such an analysis does not, however, provide information about the distribution of receptors, which was expected to be heterogeneous within our regions of interest, nor does it indicate protein expression at those locations. Therefore, we performed receptor autoradiography with a radiolabeled OTR antagonist, ¹²⁵I-OVTA, in coronal and sagittal sections. For this part of our study, we were interested in observing the spatial pattern of binding, not quantifying the binding, so we did not perform quantitative analyses. We used four brains at each of the ten ages from the qPCR study, one of each sex in each plane of section.

The pattern of ¹²⁵I-OVTA labeling in juveniles largely resembled that of adults (Figs. 4–6; compare with Leung et al., 2009) even at the earliest phases of development we sampled, and closely followed the distribution of OTR mRNA as seen using in situ hybridization (Leung et al. 2011; Fig S1). Below, we have summarized our impressions of the binding patterns in each of our regions of interest and how those patterns changed over time. These results are organized by region; please refer to the Introduction for our rationale for focusing on these regions a priori. Following those results, we present a short discussion of notable binding in regions outside our a priori regions of interest.

Fig. 4. Binding of ¹²⁵I-OVTA in the lateral septum.

The upper right photo shows ¹²⁵I-OVTA binding in an adult male zebra finch. Dense labeling can be seen in the lateral septum (LS) as well as in the periventricular area of the hyperpallium (red arrow in the drawing). On the left, the same labeling is shown at several ages post-hatch. The labeling was similar in males and females. All images are of coronal sections.

Fig. 6. Binding of ¹²⁵I-OVTA in HVC and the arcopallium.

On the left, labeling is shown in song control nucleus HVC in male zebra finches over the course of vocal development. Coronal and sagittal views are shown; in the sagittal images, rostral is to the right. On the right, labeling is shown in coronal images of the arcopallium (A) in both sexes. The arcopallium contains the robust nucleus of the arcopallium (RA), which is a major target of HVC. Dense label is seen in the dorsal arcopallium and the capsular region surrounding RA.

Lateral septum:

In adults, some of the most intense binding of ¹²⁵I-OVTA is found in LS (Leung et al., 2009), and we found this region densely labeled in juveniles as well. Levels of binding were high in LS by post-hatch day 5 (P5) (Fig. 4). The pattern of labeling was similar to what has been reported in adults (Leung et al. 2009), which was shown to overlap closely with expression of OTR mRNA (Leung et al., 2011; Fig. S1). Thus, the OTR population in this region was largely adult-like both with respect to its distribution pattern and level of expression (Fig. 3A).

Auditory forebrain:

We found dense ¹²⁵I-OVTA labeling in both primary and secondary auditory cortex (Fig. 5; see Calabrese & Woolley, 2015 for definitions and boundaries of these regions). This labeling was apparent as early as P5. At all ages, the signal appeared darkest in regions of Field L, primarily L1 and L3, and in the caudal nidopallium. Field L2, which receives auditory thalamic input, was unlabeled compared to surrounding regions. The contrast between Field L2 and the surrounding regions was barely discernible at P5 but was obvious by P15. In adult zebra finches, this distinctive pattern of ¹²⁵I-OVTA labeling overlaps with OTR mRNA expression more closely than it does with V1aR or V2R mRNA expression (Fig. S1). Given the high degree of selectivity in this region for conspecific song (Mello & Clayton, 1994), this pattern of labeling is consistent with a model in which OTR contributes to the recognition of tutor song.

Fig. 5. Binding of ¹²⁵I-OVTA in the auditory forebrain.

Labeling is shown in sagittal sections on the left, and coronal on the right. Dense labeling can be seen surrounding but not including Field L2 at all ages. Label is also notable in the periventricular area of the hyperpallium (red arrow). These images represent a mix of males and females; we noted no sex differences. NC, caudal nidopallium. CM, caudal mesopallium. N, nidopallium. Cb, cerebellum.

Song control nuclei HVC and Ad:

In our previous studies in adults, labeling in the song nucleus HVC was highly variable in a sample of young, adult males (Leung et al., 2011). In the current study, we observed ¹²⁵I-OVTA labeling in HVC in males, but not females, from P35 to P95 (Fig. 6). The intensity of the labeling was variable and not distinct in every animal, similar to what we reported previously for adult males (Leung et al., 2009). In Ad, we found dense, adult-like labeling at each age (Fig. 6). Overall, these results confirm that OTR is present in the song system, in an adult-like pattern, during vocal development.

Other notable binding.

Although we focused our analysis on our a priori regions of interest, we observed binding in three other regions that may be relevant to vocal learning: the capsular region surrounding the robust nucleus of the arcopallium (RA), the periventricular hyperpallium, and the dorsolateral corticoid area. The potential role of each of these regions in vocal learning is described in the Discussion, below. RA capsular labeling is known to be absent in adult zebra finches (Leung et al., 2009); we noted binding in both sexes as early as P5 and it was most pronounced around P25–35. At P55, this labeling was seen in males only, and by P75 it was weaker in both sexes while Ad remained prominently labeled. We also noted prominent labeling in the periventricular region of the hyperpallium, along the lateral ventricles. The distribution of this label is depicted in detail in Fig. 7, but it can be seen also in Figs. 4, 5, 6, and S1A. Nearby but discontinuous labeling was also noted in an area labeled as the core of the dorsolateral corticoid area in the Zebra Finch Expression Brain Atlas (ZEBrA, zebrafinchatlas.org). Our findings in each of these three regions should be considered preliminary.

Fig. 7. Binding of ¹²⁵I-OVTA in the periventricular area of hyperpallium.

This series of photos shows sagittal sections, medialmost (A) to lateralmost (P) in a female zebra finch at P85. Rostral is to the right. All sections are from the same individual and are 120μm apart. Labeling was similar at all ages in both sexes. Dense labeling can be seen in the periventricular area of the hyperpallium (arrow in A, becoming more obvious in D) that extends for several hundred μm medially, ending at J. A second, discontinuous area of label, corresponding to the dorsolateral corticoid area, emerges at H and extends laterally to P.

Discussion

Overview of findings

Social motivation is thought to be an integral and even critical factor contributing to social learning, particularly vocal learning (Chevallier et al., 2012; Mundy & Burnett, 2005; Syal & Finlay, 2011). The mechanisms underlying social motivation in humans are not well-understood. A large literature suggests that across vertebrate taxa, including in humans, nonapeptides such as oxytocin promote attention and attraction to social cues (Gordon, 2011; Hammock, 2015; 2018). Thus, nonapeptide systems may play an important role in vocal learning by facilitating motivation to attend to the cues necessary for that learning.

We propose that the zebra finch, a species for which vocal learning is clearly social, should be developed as a model to understand these processes at the molecular level. In this foundational study, we carried out an important first step toward this goal: to characterize and track the expression of nonapeptide receptors in relevant brain regions, in both sexes, over the entire trajectory of vocal development. We report that the three nonapeptide receptors found in birds, OTR, V1aR, and V2R, are expressed throughout the sensitive period for song learning in regions relevant to that learning. Therefore, these receptor populations are likely available to promote social motivation and social attachment in juveniles of this species.

In zebra finches, all singing is done by males. We therefore hypothesized that males would express higher levels of OTR mRNA in regions related to singing. This hypothesis was supported by our finding that OTR expression was higher in males in dTel, which contains the song nucleus HVC, and in Ad, a region heavily interconnected with other regions that support song learning. These sex differences were detectable only during the auditory learning phase, which is a time during which males are not yet singing, but they are strongly motivated to hear and memorize the songs of a chosen tutor. In contrast, we could not detect sex differences at any age in the septal area or the auditory forebrain. These OTR populations likely function to support social motivation and song recognition in juveniles of both sexes.

Using receptor autoradiography, we identified a number of other regions that might warrant further investigation in the future. One such area is the capsular region surrounding song nucleus RA (Fig. 6). This region is not labeled in adult zebra finches (Leung et al., 2009); however, in juveniles younger than P55, we found the binding pattern to be strikingly similar to that reported for species that undergo seasonal changes in song variability (Leung et al., 2009; Ondrasek et al., 2018; Voorhuis et al., 1990; Wilson et al., 2016). Testosterone treatment, which reverses variability in song (Meitzen et al., 2007; Whaling et al., 1995), reduced OTR mRNA in the arcopallium of white-throated sparrows (Grozhik et al., 2014), showing that OTR in this region is likely associated with song variability. The labeling in this region is compelling because it could represent a substrate underlying song variability and perhaps even new learning.

Also of potential interest is a stripe of dense binding in the periventricular hyperpallium and the nearby dorsolateral corticoid area (Fig. 7). In chickens, these regions are thought to be important for filial imprinting (Aoki et al., 2015; Nakamuri et al., 2010), a form of early learning that, like song learning, takes place during a sensitive period. In mammals as well as in birds, imprinting is part of an over-arching model of experience-dependent brain development in which early attachment, proximity-seeking, and social orienting are regulated by oxytocin (Hammock, 2015; 2018; Loveland et al., 2019). Because song learning in zebra finches depends critically on such processes,(Adret, 1993; Baran, 2017; Rodriguez-Saltos, 2017), this species represents powerful opportunities to understand the role of oxytocin in socially-guided vocal development.

Changes in OTR expression at a life-history transition

Nonapeptide systems may play important roles in life-history transitions during development, for example in rodents transitioning from juvenile prosocial play behavior to adult competitive behavior (Hammock, 2018; Kelly et al., 2018). In the current study, we noted a remarkable decline in OTR expression in males during the transition from auditory learning to plastic song. In all four regions in which we measured OTR mRNA, this decrease occurred at around P55 (Fig. 7). We observed a similarly-timed decrease in a previous study (Maney & Rodriguez-Saltos, 2016). At around P55, the auditory learning phase is drawing to a close (Gobes et al., 2019) and males are entering the plastic song phase, dramatically increasing their song rate (Johnson et al., 2002) as their own song comes to resemble the tutor’s (Kollmorgen et al., 2020). After this transition, males spend more time practicing singing and less time seeking out access to the tutor. At the same age, social preferences transition from the family unit to unrelated birds (Adkins-Regan & Leung, 2006). This transition is reflected in the motivation to hear tutor song; preferences for the father’s song over that of an unfamiliar male began to wane at around P50-P60 (Rodriguez-Saltos et al., 2021; Fujii et al., 2021); this shift was apparent in males but not in females (Fujii et al., 2021). We hypothesize that the decrease in OTR expression at ~P55 may mediate a shift in attention and motivation away from the family unit and toward other conspecifics, which is consistent with a role for oxytocin in social transitions during development in other species (Hammock, 2018).

Consideration of endogenous ligands

In this study, we quantified mRNA for nonapeptide receptors but not the mRNAs for the nonapeptides themselves. Other work suggests that the endogenous ligands, vasotocin and mesotocin (the homologs of vasopressin and oxytocin, respectively, in birds) are produced during the vocal learning period. In chickens, immunoreactivity for these nonapeptides is detectable by embryonic day 6 and reaches adult-like levels in both magnocellular and parvocellular neurons by the day of hatch (Milewski et al., 1989; Tennyson et al., 1986). In canaries, vasotocin immunoreactive neurons appear in the paraventricular nucleus by P28 (Voorhuis et al., 1991). We have detected both vasotocin and mesotocin mRNA in micropunches of hypothalamus of juvenile zebra finches at P42, P56, and P70, and both mRNAs are expressed at adult-like levels at all three ages (W. M. Zinzow-Kramer, unpublished data). Overall, these results suggest that ligands of nonapeptide receptors are expressed during the period of song learning. We do not know, however, which endogenous ligand is more likely to bind at the sites we labeled. Both vasotocin and mesotocin are likely to bind to avian OTR (see Leung et al., 2011), and both nonapeptides perform oxytocin-like functions (Robinzon et al., 1994; Saito & Koike, 1992; Goodson & Bass, 2001). In zebra finches, binding of ¹²⁵I-OVTA in the LS is displaced equally by mesotocin and vasotocin (Leung et al., 2009). More detailed studies will be required to determine which endogenous ligand is the more relevant to each of the receptor populations we have identified.

Limitations of this study

Our qPCR data showed high between-subject variation. Some of this variation, particularly for V1aR and V2R, is likely related to low levels of mRNA expression. Wide individual variation is a hallmark of nonapeptide receptor expression in both birds and mammals (Francis et al., 2002; Leung et al., 2009; 2011; Phelps & Young, 2003). Our autoradiographic study, which was designed to reveal spatial patterns of binding, did not include enough animals to detect meaningful variation in binding levels. We did not, for example, notice a decrease in signal in the auditory forebrain around P55 in the films, even though our qPCR data show this decrease both in this study (Fig. 3B) and in a previous one (Maney & Rodriguez-Saltos, 2016). We emphasize that our results from autoradiography and qPCR are not necessarily expected to match, because whereas the first method labels binding, the second targets mRNA expression. Given the well-documented individual variation in nonapeptide receptor expression across species, autoradiography should be conducted on a larger number of animals to detect changes in the signal over time.

Conclusion

Social learning requires close attention to conspecifics, thus the neural systems underlying social orienting need to be well-developed for such learning to occur. Our understanding of how these neural systems support learning, particularly vocal learning, has been hindered by a lack of appropriate animal models. In zebra finches, social interactions greatly facilitate vocal development, yet we have little information about the underlying mechanisms. We and others have proposed that nonapeptides, such as oxytocin, contribute critically to the process of vocal learning in songbirds (Baran, 2017; Maney & Rodriguez-Saltos, 2016; Syal & Finlay, 2011; Theofanopoulou et al., 2017). It was previously unknown, however, whether and where OTR is expressed in the brains of young birds as they learn to sing. Here, by mapping and quantifying the expression of OTR in juveniles, we lay the foundation to develop the zebra finch as a model for studying the role of oxytocin in vocal development. Our results show that OTR mRNA is expressed in brain regions relevant to social attachment and song learning, including the auditory forebrain and song system, as early as P5 and continues to be expressed throughout the sensitive period of song learning. Thus, this study demonstrates that this neuropeptide system is already well-developed in juveniles; it is available to facilitate attention to conspecifics and therefore to mediate song learning. Experimental work will be required to explore further its role in that learning.

Data availability statement:

The data that support the findings of this study are available from the corresponding author upon request.