Introduction

Activity-dependent neuroprotective protein (ADNP) discovered by Gozes and colleagues is a well-conserved gene in Chordata (Bassan et al. 1999; Zamostiano et al. 2001; Gozes et al. 2015). In humans, de novo mutations in ADNP result in the ADNP syndrome, characterized by intellectual (cognitive), auditory, and social impediments as well as motor and speech developmental delays/disabilities (Helsmoortel et al. 2014; Hacohen Kleiman et al. 2015; Gozes et al. 2017a; Arnett et al. 2018; Van Dijck et al. 2019). In mice, Adnp haploinsufficiency, the Adnp+/− model, developed by the Gozes laboratory (Pinhasov et al. 2003, Vulih-Shultzman et al. 2007), causes significant multi-system irregularities, bearing close resemblance to the phenotype of the ADNP syndrome (Malishkevich et al. 2015; Amram et al. 2016; Gozes et al. 2017b; Hacohen-Kleiman et al. 2018, 2019). More specifically, Adnp+/− mice inhere a diverse set of deficiencies including hearing abnormalities and motor, social, learning, and memory deficits, all essentially underlied by significant synaptic and gene regulation abnormalities (Malishkevich et al. 2015; Amram et al. 2016; Hacohen-Kleiman et al. 2018, 2019). ADNP expression in mouse and human (hippocampus) is sexually dichotomous, also appearing in estrous cycle–synchronized fluctuations in the mouse arcuate nucleus (hypothalamus) (Bassan et al. 1999; Furman et al. 2005; Malishkevich et al. 2015). With speech acquisition being a major impediment in ADNP syndrome patients (Gozes et al. 2017a; Van Dijck et al. 2019), ADNP potential effect on mouse vocal production was tested. A significant sex-dependent decline in the number of Adnp+/− mouse pup calls was reported (Hacohen-Kleiman et al. 2019), whereas treatment with the ADNP snippet NAP corrected this abnormality (CP201) (Hacohen-Kleiman et al. 2018).

In contrast to rodents, songbirds are one of the few species able to learn vocalizations, alongside humans (Doupe and Kuhl 1999; Panaitof 2012). A previous study in the zebra finch songbird (ZF; Taeniopygia guttata) brain revealed greater ADNP transcript levels in the cerebrum of young males, compared with females and all other brain regions (Hacohen Kleiman et al. 2015). These findings were suggested to potentially correspond with male-exclusive singing ability and a fully developed cerebral song system (reviewed in Barnea and Pravosudov 2011).

Another highly conserved transcription factor is the Forkhead box protein P2 (FoxP2), known to affect speech in human and song learning in songbirds (Lai et al. 2001; Haesler et al. 2004, 2007; Teramitsu et al. 2004; MacDermot et al. 2005; Feuk et al. 2006; Vernes et al. 2011). FoxP2 has been associated with autism through interactions with known ASD-related genes, including contactin-associated protein-like 2 (CNTNAP2) (Larsen et al. 2016; Adam et al. 2017; Li and Pozzo-Miller 2019). In mice, cortical Foxp2 was suggested to be vital for social behavior, causing aberrations in autism-related gene expression and ultrasonic vocalization production (Medvedeva et al. 2019). In Adnp+/− mice, FoxP2 levels were previously shown to increase in the male hippocampus, compared with Adnp+/+ littermates (Hacohen-Kleiman et al. 2019), proposing FoxP2 as a potential target for ADNP regulation.

To further our research, we strive to reveal the roles of ADNP in the songbird cerebrum of both sexes. Despite high resemblance in the ADNP sequence (predicted protein, 98.5% identical to ZF), canaries are characterized with subtler sex-dependent variations in song and volume of song nuclei (comparison presented in Table 1). Consequently, a study in canaries could subject sex-dependent roles in vocal production to scrutiny, thus allowing to potentially unravel ADNP underlying trajectory for vocal alteration.

Table 1 A comparison between canary and zebra finch
Full size table

Here, quantitative real-time PCR (qRT-PCR) and RNA in situ hybridization (ISH) spatial expression patterns of ADNP were tested in brains of male and female canaries. Corresponding with ADNP qRT-PCR data, transcript expression patterns for additional three genes, related to ADNP or autism/social behavior, were also examined: the ADNP regulator, vasoactive intestinal peptide (VIP) (Bodner et al. 1985; Gozes et al. 1989a; Giladi et al. 1990; Bassan et al. 1999), the ADNP sister gene (ADNP2) (Zamostiano et al. 2001; Kushnir et al. 2008; Dresner et al. 2011, 2012), and FoxP2. Correlation analyses between these gene transcripts were performed for potential significant gene-gene correspondence.

Materials and Methods

Nomenclature

For avian brain regions, we used the revised nomenclature proposed by the Avian Brain Nomenclature Forum (http://avianbrain.org) (Reiner et al. 2004b; Jarvis et al. 2013) as well as the brain atlas by Nixdorf-Bergweiler and Bischof (2007) and online atlas (http://www.zebrafinchatlas.org). For gene nomenclature, we followed the convention proposed by the NCBI and the HUGO Gene Nomenclature Committee database (i.e., ADNP, ADNP2, VIP, and FOXP2 in human and canary (excluding FoxP2 in canary (Kaestner et al. 2000)) and Adnp, Adnp2, Vip, and Foxp2 in mouse). Proteins are in roman type, and genes and RNA in italics (https://www.genenames.org/).

Animals

Domesticated canary brains (n = 13, 2–6-year-old males; and n = 12, 2–5-year-old females) were obtained according to availability from a breeding colony in Professor Nottebohm laboratory at the Field Research Center in Rockefeller University, NY, USA. Canaries were previously kept in a co-sex aviary, housed in groups of two (breeding pair, 18 × 10 × 9 in.) or 8–12 birds per cage (32 × 27 × 22 in.), subjected to 15.25 light hours/24 h a day. Birds were provided with seeds, water, and a high-protein egg diet available ad libitum. Prior to sacrifice, all canaries were in breeding mode. Preliminary analysis revealed that the selected gene expression levels did not change within the range of tested ages in both sexes (2–6 years old, Supplemental Fig. 1). Bird of ages ranging 2–6 years were therefore grouped together and analyzed accordingly. Two to 6 days prior to killing, birds were placed in groups of 3 birds per cage (23.5 × 20 × 14 in.; males and females together) in a 13-h lights-on/11-h lights-off cycle. For exclusion of song effect on gene expression, lights were kept “off” during morning hours of day of sacrifice, minimizing bird singing (up to 1 h prior to sacrifice). Canaries were decapitated following CO2 inhalation overdose (Hacohen Kleiman et al. 2015).

Tissue Preparations for mRNA Analyses

Brains were rapidly removed and snap-frozen in liquid nitrogen either separately cut into two hemispheres or dissected to cerebrum, cerebellum, and brain stem (for RNA ISH or qRT-PCR, respectively) (Hacohen Kleiman et al. 2015). Frozen samples were stored at − 80 °C until use. For qRT-PCR mRNA analysis, RNA extraction was done as previously described (Hacohen Kleiman et al. 2015). Cerebrum, cerebellum, and brain stem RNA transcripts were extracted using the TRI reagent (T9424, Sigma–Aldrich, St. Louis, MO). A total of 0.2 ml of chloroform was added for phase separation (pelleted at 12,000×g for 15 min, 4 °C). A total of 0.5 ml of 2-propanol was added for total RNA pellet (pelleted at 12,000×g for 10 min, 4 °C). The total RNA pellet was washed by adding 1 ml of 75% ethanol and then subjected to centrifugation at 7500×g for 5 min, 4 °C. Total RNA was dissolved in distilled water (Tamar, Mevaseret Zion, Israel) (Hacohen Kleiman et al. 2015). Total RNA purity and concentration were determined using a spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). All materials were added calculated in accordance with 1 ml of TRI Reagent.

Quantitative Real-time PCR

qRT-PCR was performed as previously described (Hacohen Kleiman et al. 2015) (males (n = 10, 2–6 years old) and females (n = 9, 2–5 years old)). Equal amounts of total RNA (1 μg RNA/sample, obtained from each bird) were subjected to reverse transcription (RT) using qScript cDNA Synthesis Kit (Quanta Biosciences, Gaithersburg, MD, USA). Reactions were all carried out in duplicates in 384-well plates. For sample variability exclusion, several samples were repeated between plates as controls. Real-time PCR was performed using PerfeCTa SYBR Green FastMix, Low ROX (Quanta Biosciences, Gaithersburg, MD, USA) and QuantStudio 12K Flex Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). Messenger RNA expression levels were determined using specific canary primers presented in Table 2 (Mukai et al. 2009; Olias et al. 2014). The comparative Ct method was used for quantification of transcripts. As previously performed, data were normalized with those from ribosomal protein L13 (RPL-13; control) (Hacohen Kleiman et al. 2015). Results are presented as mean (± SEM) relative gene expression (2−ΔCT; Fig. 1). Analyses for potential significant correlations between the genes were performed (presented in Table 3).

Table 2 Canary-specific primers utilized in qRT-PCR
Full size table
Fig. 1
figure 1

Gene expression and positive correlations in the canary brain. Results are presented as relative gene expression (2−ΔCT). Male results are presented in blue; female results are presented in magenta. Results were normalized to RPL-13 (control). Two-way ANOVA with Tukey’s post hoc test was performed for statistical analyses. a For ADNP, a significant interaction between brain regions and sex was found (F(2,51) = 4.366, p = 0.018). Significantly higher ADNP expression was observed in the male cerebrum, compared with female and the other brain regions (***p < 0.001). In females, ADNP expression levels were significantly higher in the cerebrum, compared with the brain stem (*p < 0.05). b For VIP, main effect for brain region was found (F(2,49) = 36.842, p < 0.001). VIP was significantly less expressed in the cerebellum compared with other brain regions, in both sexes (***p < 0.001). c For ADNP2, main effect for brain region was found (F(2,51) = 15.194, p < 0.001). ADNP2 expression was significantly reduced in the brain stem of both sexes (*p < 0.05 in males, ***p < 0.001 in females), with even lower expression in females (*p < 0.05). d For FoxP2, a statistically significant interaction between brain region and sex was found (F(2,51) = 4.174, p = 0.021). In males, significantly higher expression was observed in both the cerebrum and brain stem, compared with females (**p < 0.01, ***p < 0.001 respectively). For both sexes, FoxP2 was less expressed in the cerebellum compared with other brain regions (***p < 0.001). Males (n = 10, 2–6 years old) and females (n = 9, 2–5 years old). Significant p values appear in asterisks

Full size image
Table 3 Gene correlations in the canary brain within three brain regions: cerebrum, cerebellum, brain stem. Male results are presented in blue; female results are presented in magenta. Data were normalized with those from RPL-13 (control). Multiple statistically significant positive correlations were found between the genes within the different brain regions, in a sex-dependent manner. Correlation tests for all genes were performed using either the Pearson correlation coefficient method or Spearman’s rank correlation coefficient, if at least one of the data sets was not normally distributed. Table cells which stand for repeated comparisons between genes were disregarded and marked in gray color. Males (n = 10, 2–6 years old) and females (n = 9, 2–5 years old). Significant p values appear in asterisks
Full size table

ADNP cDNA Cloning from Canary Brain

Desired ADNP cDNA sequence (260 bp) was amplified through PCR using ADNP primers (sense 5′-TCAGTGGCAACACAGCTG-3′, antisense 5′-AGGACCTGTAGCAGCCAC-3′) and a canary cDNA sample. The PCR product was examined on a 2% agarose gel, cleaned using Wizard® SV Gel and PCR Clean-Up System (Promega, WI, USA), and cloned into a pGEM T Easy vector (Promega, Madison, WI), as previously done (Haesler et al. 2004; Dresner et al. 2012). ADNP sense and antisense riboprobes were in vitro transcribed from T7 and SP6 promoter sides of the pGEM T easy cloning vector containing the ADNP cDNA clone, using digoxigenin (DIG) labeling (Roche, Germany).

RNA In Situ Hybridization

ISH was performed as previously described (Mendoza et al. 2015). Canary brains (males (n = 3, 4–5 years old) and females (n = 3, 2–5 years old)) were cut frozen on the sagittal plane using a rotary cryostat (14 μm). Sections were then fixed in 4% paraformaldehyde for 5 min, washed in 0.025× PBS, dehydrated in 70%, 95%, and 100% ethanol, and air-dried. Sections underwent acetylation for 10 min using triethanolamine (Carl-Roth, Germany) and acetic anhydride (Sigma–Aldrich, St. Louis, MO). Slides were washed twice with 2× SSC (0.3 M NaCl, 0.03 M tri-Natriumcitrat Dihydrat, pH = 7), re-dehydrated with the same increasing concentrations of ethanol, and air-dried. Sections were pre-hybridized in hybridization buffer (5× SSC, 2% blocking reagent, 50% formamide) for at least 1 h at 60 °C. For hybridization, 7 μl of probe/100 μl hybridization solution was applied on each section and was left overnight at 55 °C in an oil bath (330779 light Mineral oil, Sigma–Aldrich, St. Louis, MO). The following day, sections were repeatedly rinsed using chloroform and 2× SSC and washed for 20–30 min in the following solutions, preheated to 55 °C: 1× SSC/50% formamide, 2× SSC, and 0.2× SSC (twice). Slides were then washed (RT) in MABT (100 mM malic acid, 150 mM NaCl, 0.1% Tween 20, pH 7.5), blocked with TNB (TSA Blocking Reagent, FP1020, PerkinElmer, USA) for 2 h, and incubated overnight with alkaline phosphatase–conjugated goat anti-DIG Fab’ antibody (11093274910, Roche, Germany), 4 °C. Colorimetric detection was performed by an immunoalkaline phosphatase reaction with BM purple solution (11442074001, Roche, Germany) as the substrate. Antisense and sense probes were always run in parallel. Ten sections of each brain, in 280-μm intervals, were mapped along the sagittal axis of the brain for ADNP-expressed regions, with Olympus × 4/0.16 Objective (Fig. 2a, b). Images were taken using a light microscope (Olympus bx-65) equipped with a color camera (Micro-BrightField Ltd., USA), a motorized stage, and a computerized brain-mapping system (Neurolucida; Stereo Investigator version 9; Micro-BrightField Ltd., USA). Settings for microscope lighting and camera exposure were kept constant for all images obtained.

Fig. 2
figure 2

ADNP expression shows a robust signal in the mesopallium (M). 14-μm sagittal sections of male and female canary brains were tested for RNA in situ hybridization (ISH) of ADNP expression. In all images, arrows and dotted lines indicate the examined regions. Images of ADNP expression in mesopallium (M) are presented above with visible borders. All images were taken under the same illumination and camera settings (0.6 Lux), with × 4/0.16 Objective. a Three representative images of male and female brains are presented similarly (upper images, males; lower images, females). Higher ADNP mRNA expression was implied by strong signal observed in M (marked on atlas, left side of the figure). Males (n = 3, 4–5 years old) and females (n = 3, 2–5 years old). b ADNP antisense probe specificity was verified by the lack of signal with the ADNP sense probe (representative pictures of ADNP antisense and sense probes appear respectively from left to right). c For signal quantification, borders were drawn around each area using ImageJ. ADNP antisense probe was compared with signal from the ADNP sense probe in adjacent sections in each bird. In graph, data are presented as 255-mean of histogram gray value. Results for ADNP antisense probe signal are presented in blue; ADNP sense probe signal are presented in turquoise. ADNP antisense probe signal was significantly stronger compared with that of ADNP sense probe (**p = 0.007, Student’s t test). Males (n = 1, 5 years old) and females (n = 3, 2–5 years old). The atlas image was reproduced from the ZEBrA database with permission from Mello laboratory (Oregon Health & Science University, Portland, OR 97239; http://www.zebrafinchatlas.org) (Lovell et al. 2020). Directions are indicated by arrows: dorsal (D); lateral (L); caudal (C); rostral (R). Significant p values appear in asterisks (*p < 0.05, **p < 0.01, ***p < 0.001). Scale bar = 150 μM

Full size image

ADNP mRNA Signal Quantification in ISH

For all images, borders were drawn around desired brain region prior to quantification as mean gray value, using the ImageJ software (Schneider et al. 2012). Significance of signal intensity is presented for mesopallium (M), compared with the ADNP sense probe in adjacent sections (Fig. 2b, c). Results are presented as 255-mean of signal intensity ± SEM (RGB; Fig. 2c) (males (n = 1, 5 years old) and females (n = 3, 2–5 years old)).

Statistics

For qRT-PCR and ISH quantifications, results are presented as mean ± SEM. Data were tested for normal distribution using the Shapiro–Wilk test and outliers were excluded (https://graphpad.com/quickcalcs/Grubbs1.cfm). For two different categorically independent variables, a 2-way ANOVA followed by Tukey’s post hoc test was performed. For correlation testing, Pearson’s correlation analysis was performed if both plotted data sets were normally distributed. Spearman’s rank correlation coefficient method was performed if at least one of the data sets was not normally distributed. Student’s t test analysis was performed when needed. p values ≤ 0.05 were considered statistically significant, and all tests were 2-tailed. All statistical analyses were conducted using the SigmaPlot software (version 11 Windows) or GraphPad Prism (version 6 Windows).

Results

ADNP Is Predominantly Expressed in the Canary Male Cerebrum, Similar to Zebra Finch

Given our interest to study ADNP relations to sex-dependent vocal production and circuits, we further examined ADNP transcript expression profile in a different songbird as compared with ZF, the domesticated canary songbird (Serinus canaria domestica) (presented in Table 1). Male and female canary gene expression (qRT-PCR; males (n = 10, 2–6 years old) and females (n = 9, 2–5 years old)) data were analyzed for potential sex or brain region effects in the cerebrum, cerebellum, and brain stem (Fig. 1). Additional correlation measurements were applied for potential gene-gene associations in the canary brain (presented in Table 3). For ADNP, qRT-PCR results present a similar mRNA distribution to that seen in young ZF, in both sexes (Hacohen Kleiman et al. 2015) with highest expression in male cerebrum (2-fold, p < 0.001) and lowest in the female brain stem (p < 0.05) (Fig. 1a). For the ADNP regulator gene, VIP, expression was found to be similar for both sexes with lowest expression in the cerebellum (p < 0.001) (Fig. 1b). For the ADNP sister gene, ADNP2, a significantly lower expression was observed in the brain stem, with even lower expression in females (p < 0.05) (Fig. 1c).

FoxP2, a Known Speech Regulator, Is Grossly Distributed in a Similar Fashion to VIP, with Significant Sex-Dependent Differences in the Cerebrum, Like ADNP and the Brain Stem

The transcript levels of FoxP2, a known language acquisition regulator (Teramitsu et al. 2004; Scharff and Haesler 2005), were also tested (Fig. 1d). qRT-PCR results showcased an overall similar expression pattern to that of VIP, with lower expression in the cerebellum compared with the cerebrum and brain stem in males (p < 0.001). Similarly, lower FoxP2 expression levels were observed in the female cerebellum, compared with the brain stem (p < 0.001). Like ADNP, a comparison between the sexes revealed significantly higher expression levels for FoxP2 in the male cerebrum (Fig. 1a, d). Furthermore, a similar sex-dependent difference in FoxP2 levels was found in the brain stem, comparing male expression levels with females (p < 0.01, p < 0.001, respectively) (Fig. 1d).

Multiple Positive Sex- and Brain Region–Dependent ADNP/ADNP2/VIP/FoxP2 Gene Transcript Correlations in the Canary Brain

Correlations between ADNP, VIP, ADNP2, and FoxP2 transcripts were measured in the three tested brain regions, following qRT-PCR (above). Significant correlations (p < 0.05) are shown in Table 3. In short, high positive correlations were observed for ADNP and all tested transcripts in a sex/tissue-dependent manner as follows. For VIP, ADNP correlations were observed only in males in the cerebrum and brain stem. For ADNP/ADNP2, correlations were solely observed in the female cerebrum, the male cerebellum, and the brain stem of both sexes. The most extensive correlations were observed for FoxP2 and ADNP appearing in all brain areas and in both sexes, except for the female brain stem. VIP also correlated with ADNP2, only in the female cerebrum (like with ADNP) and male brain stem. In the same brain areas, VIP correlated also with FoxP2, only in the male cerebrum and in the brain stem of both sexes.

Extensive ADNP mRNA Expression in the Brain with Robust Signal in Canary Mesopallium

As ADNP was significantly more expressed in the cerebrum (Fig. 1), we focused on this brain area for our ISH evaluations. An overall light and uniform signal was observed across the cerebrum with a robust signal detected in the mesopallium (M) region. Three male and three female panels are shown, indicating overall similar distribution (Fig. 2a). Figure 2b presents a sense and antisense ADNP probe comparison (in female) with a marked difference. Quantification of the ADNP signal in M is shown as 255-mean of histogram signal intensity (RGB) with the ADNP sense probe signal found to be nearly negligent (control; Fig. 2c, including the brain coordinates). These results further indicate the significant labeling of ADNP in M.

Discussion

In the present study, sex/brain area–specific ADNP expression are shown in 2–6-year-old canary brains, with highest expression in the male cerebrum (Fig. 1a). Our previous findings in the ZF songbird demonstrated a similar sexual/brain area dichotomy in the young (6 months old) brain, which changed upon aging (Hacohen Kleiman et al. 2015).

Notably, this study holds potential limitation due to restricted selection of samples. As mentioned in the “Materials and Methods” section, canaries were obtained according to aviary availability, which constituted a great age repertoire (2–6 years of age) with a limited number of samples per age (1–3 or 2–5 samples per age in males and females, respectively). This predicament may have affected the variation of transcript data, possibly distorting result distribution. Importantly, despite given age range, preliminary analysis of qRT-PCR data disclosed constant gene expression for all tested genes, in both sexes (Supplemental Fig. 1). Therefore, while not ideal, analyzed data was treated regardless of age appearing in two groups: males and females. For ADNP transcript data (qRT-PCR), while significant sex-dependent differences were found, sample limitation may have consequently led to unexpected males’ and females’ partial overlap in the cerebrum. This impasse could possibly be settled by data presented in Table 1. Accordingly, canaries are reported with longer maximal lifespan, compared with ZF, potentially indicating a slower aging process. In this case, results in older canaries may stand in equal measure with the younger ZF (6 months old). Furthermore, unlike ZF studied before, the canaries here were kept in a monitored indoor environment (also presented in Table 1). Importantly, stress and environmental changes, including light/darkness conditions, may alter gene expression, encompassing, for example, VIP (Holtzman et al. 1989; Casal and Yanovsky 2005; Tung and Gilad 2013; Dawson 2015; Saunderson et al. 2016; Dominoni et al. 2018). In this respect, ADNP was also reported to increase in response to stress in humans, while Adnp+/− mice were found more affected by light-related stress, compared with Adnp+/+ littermates (Sragovich et al. 2019). Ultimately, these results further amplify the intrigue of ADNP function in the songbird brain, suggesting potential sex-dependent roles. Such sex-dependent roles were previously described in mice with further association to behavioral regulation (Malishkevich et al. 2015; Amram et al. 2016; Hacohen-Kleiman et al. 2018).

Highly significant correlations were discovered among ADNP, VIP, ADNP2, and FoxP2. VIP is an ADNP regulator, important for embryogenesis (Gressens et al. 1993; Bassan et al. 1999; Gozes et al. 1999; Giladi et al. 2007). VIP is highly conserved across species (Bodner et al. 1985; Giladi et al. 1990) and found identical in many mammals (Giladi et al. 1990) with minor differences observed between human and chicken (Nussdorfer and Malendowicz 1998; Nowak et al. 2003). In birds, VIP was previously associated with pair bonding and social and parental behaviors (Kingsbury et al. 2013; Kingsbury and Goodson 2014; Kingsbury et al. 2015; Vistoropsky et al. 2016). In our canaries, VIP expression of both sexes was increased in the cerebrum and brain stem with significantly less expression in the cerebellum (Fig. 1b). This corroborates with previous reports of very low density of VIP binding sites in the cerebellum of the pigeon Columba livia (Hof et al. 1991). ADNP-VIP positively correlated only in males, in both the cerebrum and brain stem (presented in Table 3), probably contingent to specific sex-dependent regulation on the ADNP gene (Malishkevich et al. 2015) and estrogen/steroid sex-dependent regulation of the VIP gene (Gozes et al. 1989b).

ADNP2 is an ADNP paralog (33% identical, 46% similar) and does not include the microtubule-binding neuroprotective NAP motif (Zamostiano et al. 2001)(Fig. 1c). The ADNP family (ADNP, ADNP2) was suggested to play a role in erythropoiesis in vertebrates (Dresner et al. 2012). To our knowledge, ADNP2 has not been studied yet in birds. In mice, VIP reduction modulated ADNP and sister protein ADNP2 expression patterns in the brain, presenting autistic like traits (Giladi et al. 2007). Here, VIP positively correlated with ADNP2 expression in a sex-dependent manner, in the female cerebrum and the male brain stem. For ADNP-ADNP2, ADNP2 was shown to correlate with ADNP mRNA as well as with premorbid intelligence in humans (Malishkevich et al. 2016). Resembling ADNP expression profile in the female canary, ADNP2 transcripts were lowest in the canary brain stem, with significantly lower expression in females, compared with males. Accordingly, ADNP-ADNP2 positively correlated in the brain stem of both sexes, as well as in the female cerebrum and male cerebellum. ADNP2 correlation with VIP and ADNP in the female cerebrum may relate to past findings in mouse and human, associated with brain function importance, in a sex-dependent manner, with deregulated hippocampal ADNP/ADNP2 linked to schizophrenia (Dresner et al. 2011). Furthermore, the VIP receptor VPAC2 (VIPR2) gene copy number variations (CNVs) impact (cause) schizophrenia (Tian et al. 2019) and allow VIP regulation of ADNP (Zusev and Gozes 2004).

FoxP2 is a highly conserved transcription factor (Haesler et al. 2004) implicated in song and speech in both humans and songbirds (Teramitsu et al. 2004). Mutations in FoxP2 were linked to a language comprehension and speech disorder in humans (Scharff and Haesler 2005). In the songbird brain, song production and learning pathways are both distributed in the cerebrum and brainstem (Kubikova et al. 2010). More specifically, FoxP2 was previously reported to be expressed primarily in the striatum of avian and reptiles, as well as in the thalamus, midbrain visual processing regions, the inferior-olive of the medulla, Purkinje cells in the cerebellum, deep cerebellar nuclei, and sensory auditory midbrain structures (Takahashi et al. 2003; Haesler et al. 2004; Teramitsu et al. 2004; Scharff and Haesler 2005; Campbell et al. 2009; Kato et al. 2014; Wohlgemuth et al. 2014). Here, our results coincide with these previous reports, presenting higher FoxP2 mRNA levels in both the cerebrum (including the striatum) and the brain stem (comprising the thalamus, midbrain, and hindbrain (Kubikova et al. 2010)), compared with much lower expression in the cerebellum, in both sexes (Fig. 1d).

FoxP2 male and female brain distribution patterns were grossly similar; nevertheless, significantly higher levels were reported in males, compared with females (Fig. 1d). The overall resemblance may be attributed to female canary ability to produce songs (Ko et al. 2020), whereas significant differences in song and song system anatomy may stand for the male-specific transcript increase. In this respect, reports of female spontaneous singing are relatively rare and still significantly differ from male or testosterone-induced female song (Ko et al. 2020). Anatomically, the female canary song control system nuclei appear smaller, compared with males (Nottebohm and Arnold 1976) (presented in Table 1). A previous study in black-capped chickadees (Poecile atricapillus) may also correspond with our findings of subtler FoxP2 sex-dependent differences, found in Area X (Phillmore et al. 2015). Likewise, this moderate discrepancy was potentially associated with female ability to produce learned chickadee call songs with lower output, compared with males (Phillmore et al. 2015).

For ADNP-FoxP2, a connection was previously implied with reports of a specific increase in FoxP2 levels in the male Adnp+/−mouse hippocampus, compared with Adnp+/+ littermates (Hacohen-Kleiman et al. 2019) as well as its normalization of expression in the presence of the ADNP snippet drug candidate NAP (NAPVSIPQ; CP201) (Bassan et al. 1999; Oz et al. 2014; Hacohen-Kleiman et al. 2018). Similarly, given the connection of ADNP and VIP to schizophrenia, previous studies have also assessed FoxP2 expression in the disrupted-in-schizophrenia (DISC1) mouse model, with FoxP2 regulating DISC1. Thus, previous results showed that FoxP2 transcript levels were increased in the hippocampus of the DISC1-mutated mice and were significantly lowered after treatment with NAP (Vaisburd et al. 2015), as shown above for the Adnp+/− mouse model.

In the current study, ADNP expression pattern resembled that of FoxP2 with higher male expression in the cerebrum, coupled with significant positive correlations in both sexes. Additional ADNP-FoxP2 positive correlations were observed in the cerebellum in both sexes and solely in the male brain stem. Like FoxP2-ADNP, FoxP2-ADNP2 correlations were also present in the cerebrum of both sexes and in the brain stem of males only. These findings imply once again of sex-dependent importance in the regulation and function for ADNP, ADNP2, and FoxP2 in the brain.

For FoxP2-VIP, to the best of our knowledge, there is no previous description of a potential connection between the genes. Here, FoxP2 and VIP gross expression patterns were found vastly similar, in both sexes with several FoxP2-VIP positive correlations, in the male cerebrum and in the brain stem of both sexes. We presume this could be referred to specific sexual dichotomous differences found in FoxP2 (not in VIP), pointing to higher expression in males versus females in both the cerebrum and brain stem. In this regard, FoxP2, ADNP, and VIP may be connected through dendritic spine density regulation. FoxP2 was found to regulate spine density in Area X in ZF (Haesler et al. 2007; Schulz et al. 2010), as well as CNTNAP2 expression, suggesting a CNTNAP2-mediated FoxP2 effect on spines (Adam et al. 2017; Mendoza and Scharff 2017). Importantly, CNTNAP2 is an ASD-associated gene alongside ADNP (Helsmoortel et al. 2014; Larsen et al. 2016; Adam et al. 2017; Gozes et al. 2017a; Gozes et al. 2017b; Li and Pozzo-Miller 2019; Van Dijck et al. 2019; Grigg et al. 2020; Satterstrom et al. 2020). Like FoxP2, ADNP and VIP were also reported to regulate dendritic spines, possibly acting through ADNP (Hill et al. 1994; Hacohen-Kleiman et al. 2018). Treatment with NAP, in turn, normalized spine density in the Adnp+/− mice (Hacohen-Kleiman et al. 2018).

ISH observations revealed ADNP expression in the mesopallium (M) of both sexes (Fig. 2). Notably, VIP mRNA was also similarly described in the bird M (Hof et al. 1991; Kuenzel et al. 1997). M, formerly known as hyperstriatum ventral (HV) (Reiner et al. 2004a), is a telencephalic region, part of the avian pallium (Reiner et al. 2004a). Previous studies associated M with color discrimination in pigeons (Chaves and Hodos 1997) while HV relative size was found to significantly correlate with the rate of feeding innovation in birds (Timmermans et al. 2000). In primates, a similar correlation was observed with the relative size of the neocortex (Timmermans et al. 2000; Michael et al. 2015). The intermediate medial HV was suggested to participate in the memory formation for avoidance learning in chicks (Patterson et al. 1990) and recognition following visual imprinting (Horn 1990). Additional past studies in the caudal medial mesopallium (CMM) and nidocaudal mesopallium (NCM), two important auditory areas, previously associated these brain areas to perceptual processing of song and the formation of auditory memories (Lynch et al. 2013; Haakenson et al. 2019). Respectively, this information could align with social and object recognition (visual-based) impairments and auditory pathway abnormalities following Adnp haploinsufficiency (Vulih-Shultzman et al. 2007; Malishkevich et al. 2015; Amram et al. 2016; Hacohen-Kleiman et al. 2018, 2019), as well as intellectual disability, autism-like traits, and atypical auditory brain stem response (ABR) in the ADNP syndrome patients (Helsmoortel et al. 2014; Hacohen-Kleiman et al. 2019; Van Dijck et al. 2019). Furthering the association to song, functional connections between caudal medial mesopallium (CMM) and other important regions: HVC, and caudal medial nidopallium (CMN) were suggested to take part in song learning and production in males. In parallel, an experiment in female canaries indicates of distinct functions in female songbirds related to song recognition rather than song production (Lynch et al. 2013). For Adnp, haploinsufficiency was implicated in vocal production in mice, corresponding with reports of speech delays in children of the ADNP syndrome (Helsmoortel et al. 2014; Hacohen-Kleiman et al. 2018, 2019; Van Dijck et al. 2019). The notable decrease in Adnp+/− number of calls was completely abolished following NAP treatment (Hacohen-Kleiman et al. 2018).

Finally, the intermediate medial mesopallium is known to partake in learning and memory and suggested to engage in rewarding effects (Csillag 1999; He et al. 2010). In this respect, ADNP was previously associated with the reward system in mice, affecting alcohol consumption in a sex-dependent manner (Ziv et al. 2019). NAP in turn normalized the phenotype presented in Adnp+/− females (Ziv et al. 2019).

Taken all together, this study further strengthens previous findings in ZF, mouse, and human, validating a sexual dichotomous ADNP expression pattern in the brain, with higher levels found in males (Hacohen Kleiman et al. 2015; Malishkevich et al. 2015). For ADNP-VIP-FoxP2, the role in dendritic spine density regulation, as well as VIP and FoxP2 striking correlation patterns, captivates even further in the search of a potential non-direct connection between ADNP and FoxP2, possibly involving VIP. For FoxP2, a gross role in the cerebrum and brain stem is suggested for both sexes with a more significant/additional role in males, potentially attributed to the further-complex singing ability/song plasticity (Phillmore et al. 2015). In the current study, as canary song behavior was limited but not surely avoided prior to sacrifice, the potential existence of male singing towards females in the cage could only further strengthen our explanation for the differences in FoxP2 expression between the sexes. For ADNP ISH, extensive mRNA distribution in the songbird brain may further corroborate with its essential roles as a transcription factor/chromatin remodeler and microtubule/autophagy regulator (Pinhasov et al. 2003; Mandel and Gozes 2007; Mandel et al. 2007; Merenlender-Wagner et al. 2015; Amram et al. 2016; Ivashko-Pachima et al. 2017, 2019a, b; Kaaij et al. 2019; Grigg et al. 2020). In turn, the palpable strong signal in canary M could indicate of a relation to songbird cognition, sense integration, and memory formation. This integrates well with ADNP association with similar functions, portrayed in the haploinsufficient mouse model, ADNP syndrome patients, and Alzheimer’s disease patients who may suffer from ADNP somatic mutations (Ivashko-Pachima et al. 2019a). Thus, additional investigation is necessary to unravel potential roles for ADNP in the establishment of auditory and visual-based memories and recognition in songbirds and humans. For singing behavior and perception, higher expression in M may be translated into different functions for each sex, playing a role in motor song production in males, while possibly implicating auditory song perception in females. Consequently, studies of ADNP in rodents and songbirds may be complimentary, paving the path for a better depiction of ADNP mutations and deficiencies in humans, potentially beneficial for treatment development.