Foxp1 expression is essential for sex-specific murine neonatal ultrasonic vocalization

Abstract

Autism and speech and language deficits are predominantly found in boys, however the causative mechanisms for this sex bias are unknown. Human FOXP1 is associated with autism, intellectual disability and speech and language deficits. Its closely related family member FOXP2 is involved in speech and language disorder and Foxp2 deficient mice have demonstrated an absence of ultrasonic vocalizations (USVs). Since Foxp1 and Foxp2 form heterodimers for transcriptional regulation, we investigated USV in neonatal brain-specific Foxp1 KO mice. Foxp1 KO pups had strongly reduced USV and lacked the sex-specific call rate from WT pups, indicating that Foxp1 is essential for normal USV. As expression differences of Foxp1 or Foxp2 could explain the sex-dimorphic vocalization in WT animals, we quantified both proteins in the striatum and cortex at P7.5 and detected a sex-specific expression of Foxp2 in the striatum. We further analyzed Foxp1 and Foxp2 expression in the striatum and cortex of CD1 mice at different embryonic and postnatal stages and observed sex differences in both genes at E17.5 and P7.5. Sex hormones, especially androgens are known to play a crucial role in the sexual differentiation of vocalizations in many vertebrates. We show that Foxp1 and the androgen receptor are co-expressed in striatal medium spiny neurons and that brain-specific androgen receptor KO (Ar^NesCre) mice exhibit reduced Foxp1 expression in the striatum at E17.5 and P7.5 and an increased Foxp2 level in the cortex at P7.5. Thus, androgens may contribute to sex-specific differences in Foxp1 and Foxp2 expression and USV.

Introduction

Neurodevelopmental disorders such as autism spectrum disorder (ASD) and speech and language deficits are much more prevalent in males than in females. The male to female ratio is estimated to be 4:1 in ASD and can rise to 10:1 in Asperger syndrome. A higher ‘genetic burden’ of deleterious variants in females led to the assumption that the male brain requires milder genetic alterations to develop ASD. The ‘female protective effect’, sex-specific regulation of distinct genes, and environmental effects have been proposed to promote a male susceptibility to these disorders (1,2). ASD is diagnosed when impairments in social interaction and communication are found together with restricted and repetitive behaviors during early childhood (3). In many cases, patients with ASD also present intellectual disability (ID).

FOXP1 and FOXP2 are members of the Forkhead Box P family. Mutations in the human FOXP1 gene have been associated with ASD, ID and speech and language deficits predominantly in male patients (4). FOXP2, the closest relative of FOXP1, is associated with language disorder (5,6). FOXP1 and FOXP2 share 64% amino acid sequence identity and 89% sequence identity in the DNA-binding Forkhead domain and are highly conserved among vertebrates. The two proteins form heterodimers for transcriptional regulation. They are co-expressed in the GABAergic medium spiny neurons of the striatum, a brain region critically involved in human language, mouse ultrasonic vocalization (USV) and zebra finch vocal imitation. This suggests that both proteins co-operate in common pathways and that heterodimer formation is important for cognitive and language development (7,8).

During murine brain development, expression of Foxp1 and Foxp2 starts at E14.5 and E12.5, respectively (9). Stages between E15 and P23 cover an important developmental period which includes synapse formation, which in rodents occurs during the first three postnatal weeks of life (10). Brain regions with known expression of Foxp1 and Foxp2 include the striatum, cortex, hippocampus, thalamus and cerebellum. Foxp1 and Foxp2 are co-expressed in the striatum and thalamus, but are expressed in different layers of the cortex (Foxp1: III-V; Foxp2: VI). Foxp1 is exclusively expressed in the CA1 region of the hippocampus and Foxp2 is exclusively expressed in the Purkinje cells of the cerebellum (9).

We have specifically deleted Foxp1 in neural tissues using the Cre-Lox system (Nestin-Cre/Foxp1flox/flox), hereafter referred to as Foxp1 KO mice (11). The mice are viable and show pronounced disruption of the developing striatum and subtle alterations in the hippocampus. Foxp1 KO mice are hyperactive and show increased repetitive behaviors, deficits in learning and memory and reduced social interaction (11). However, vocal communication, an important measure for social interaction, has not yet been investigated in these mice.

To address the impact of Foxp1 on social and vocal communication, we investigated USV. To explore sex-specific effects, we compared male and female Foxp1 KO and male and female wild type animals. USVs are whistle-like sounds that are emitted by mouse pups (Supplementary Material, Fig. S1) following isolation from the nest (isolation-induced USV), and during nonaggressive interactions and mating behaviors in adult mice. In ASD research, USV in mice constitutes a valuable readout for social and communicative behavior (12). However it is not directly comparable to human speech and language, as vocal behavior in mice is structurally largely innate and not learned by auditory feedback (13,14). In contrast to mouse USV, human language is characterized by its symbolic nature and certain rules (syntax) that produce novel meanings by systematic composition of the units that make up the language. Both mechanisms largely depend on learning, as symbolism and syntax are based on conventionalization (15).

Social behavior is strongly regulated by testosterone in a sex-specific manner in rodents and other vertebrates. Interestingly, previous work on zebra finch (Taeniopygia guttata) has demonstrated that Foxp2 mRNA and protein levels are acutely downregulated within area X, the specific region of the songbird striatum dedicated to song, when adult males sing (16). This finding indicates a steroid impact, as testosterone is known to play a decisive role in the regulation of song behavior (17). We therefore raised the question whether androgens regulate murine Foxp1 and Foxp2 expression during specific periods of brain development and whether this might have an influence on social and communicative behavior.

Results

Foxp1 KO pups show strongly reduced isolation-induced USV and lack sex-specific differences in peak frequency and call rate

As Foxp1 and Foxp2 form heterodimers in the striatum and regulate mutual striatal target genes (8), we hypothesized that Foxp1 similar to Foxp2 (18) may also be involved in USV. To investigate this, we compared isolation-induced USV in male and female Foxp1 KO pups with WT littermates at P4.5, P7.5 and P12.5. Usually, the rate of calls in mice peaks around the eighth day after birth and decreases to zero approximately 2 weeks after birth.

Foxp1 KO pups made significantly fewer calls at all three stages when isolated from the mother (Fig. 1A). In WT animals, the number of calls increased between P4.5 and P7.5, and then declined between P7.5 and P12.5. In contrast, the calls made by Foxp1 KO pups already declined between P4.5 and P7.5 and remained constantly low until P12.5. Interestingly, male WT pups called significantly more often than female WT pups at P4.5 (P = 0.0311) and P7.5 (P = 0.005), but this sex-specific behavior was not observed in Foxp1 KO mice (Fig. 1A).

Figure 1

Isolation-induced ultrasonic vocalization (USV) recorded at postnatal days 4.5, 7.5 and 12.5. (A) Number of emitted calls in male and female WT and Foxp1 KO pups. (B) Latency to start calling of WT and Foxp1 KO pups. (C) Peak frequencies of male and female WT calls at P7.5. Data are expressed as mean, stars indicate significant difference (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, t-test).

The latency to the onset of calling was higher in Foxp1 KO compared to WT mice (Fig. 1B). At P7.5, WT males emitted calls with significantly higher peak frequencies (pitch) compared to females (P = 0.017). In Foxp1 KO animals, the situation was reversed, with females showing significantly higher peak frequencies (P = 0.029). Furthermore, the calls made by WT females had significantly lower peak frequencies than Foxp1 KO females, and this difference was not observed in male animals (Fig. 1C). Other USV parameters including the average call duration were not altered in Foxp1 KO mice (data not shown).

To identify clusters of isolation-induced USV emitted by Foxp1 KO pups and WT pups at P7.5, we used density plots illustrating the distribution of individual isolation-induced USV calls depending on peak frequencies in kHz (maximum pitch) and call peak amplitudes in dB (maximum loudness) (Fig. 2). Calls made by male and female WT pups formed two distinct clusters characterized by peak frequencies between 50–70 kHz and 73–110 kHz. Although the number of calls was strongly reduced in male and female Foxp1 KO mice, the overall cluster formation was not affected. Together our findings indicate that Foxp1 is important for normal USV. Sex-specific differences were observed in the number of calls made by WT mice at P4.5 and P7.5, which was not observed in Foxp1 KO mice suggesting that sex-specific differences in USV in WT mice are mediated by Foxp1.

Figure 2

Distribution of individual isolation-induced ultrasonic vocalizations (USV) in WT and Foxp1 KO mouse pups emitted at P7.5. Density plots depicting the distribution of individual isolation-induced USV depending on call peak frequency in kilohertz (kHz) and call peak amplitude in decibel in WT control mice and Foxp1 KO mice, with color coding reflecting frequencies as percentages. n = 8.

Sex-specific expression of FOXP2 in the striatum at P7.5 in C57/BL6 mice

To test whether Foxp1 and Foxp2 are expressed in a sex-specific manner at P7.5, where sex-differences in neonatal ultrasonic call number in WT animals were observed, we quantified Foxp1 and Foxp2 protein in the striatum and cortex of C57/BL6 WT mice. In both tissues expression levels between male and female animals did not reach statistical significance for Foxp1. However, we detected differential sex-specific expression of Foxp2 in the striatum with males expressing 25% less Foxp2 than female littermates (P = 0.03) (Fig. 3A).

Figure 3

Sex-specific expression of Foxp1 and Foxp2 in the developing striatum and cortex in WT mouse. Relative Foxp1 and Foxp2 protein and mRNA expression as quantified by western blot and real-time PCR, respectively. (A) Sex-specific Foxp2 protein expression in the striatum of C57/BL6 mice at P7.5. (B–D) Foxp1 and Foxp2 expression levels measured in CD1 animals. (B) Foxp1 and Foxp2 mRNA expression in the striatum at E17.5 and Foxp1 and Foxp2 protein levels. (C) Foxp2 mRNA and protein expression in the cortex at E17.5. (D) Foxp1 and Foxp2 mRNA expression in the striatum at P7.5. Data are expressed as mean, significant sex difference (*P ≤ 0.05, **P ≤ 0.01, two-way ANOVA).

The CD1 mouse brain exhibits sex-specific expression of Foxp1 and Foxp2 at E17.5 and P7.5

To further test whether Foxp1 and Foxp2 are expressed in a sex-specific manner during brain development, we quantified Foxp1 and Foxp2 in the striatum and cortex at different developmental stages (E15.5, E17.5, P1.5, P7.5, P12.5 and P23.5) using real-time PCR. For this investigation we used CD1 mice, as this line—in contrast to C57/BL6—has large litters and is therefore more suitable for such a broad approach. We observed distinct sex-specific differences in the expression of both genes at E17.5 and P7.5 (Fig. 3B–D) in the striatum and cortex, but not at other time points investigated. At E17.5, Foxp1 mRNA expression was 30% higher in the male striatum compared to female tissue, yet it did not reach significance (Fig. 3B). However, Foxp1 protein expression was significantly higher (61%) in the male striatum at E17.5 (P = 0.029) (Fig. 3B). At P7.5, Foxp1 expression was significantly lower (20%) in the male striatum compared to female tissue (P = 0.048) (Fig. 3D).

At E17.5, Foxp2 expression was 30% higher in the male striatum and 35% higher in the male cortex compared to female tissue (striatum: P = 0.004; cortex: P = 0.035), which was confirmed on protein level (striatum: 43% higher, P = 0.01; cortex: 41% higher, P = 0.004) (Fig. 3B and C). At P7.5, Foxp2 expression was significantly higher (20%) in the male striatum (P = 0.021) (Fig. 3D). Taken together, these findings demonstrate that Foxp1 and Foxp2 expression differs between males and females in the striatum and cortex at E17.5 and P7.5 while differences in expression were not observed at the other time points analyzed (data not shown). Moreover these results confirm the lower Foxp2 expression in the male striatum at P7.5, which was detected in C57/BL6 pups.

Foxp2 is more abundant than Foxp1 during embryonic development and declines postnatally

Foxp1 and Foxp2 bind to target DNA and regulate transcription as homo- or hetero-dimers (19). To investigate the likelihood of Foxp1/2 homo- or heterodimer formation based on expression levels, we compared Foxp1 and Foxp2 expression in male and female striata of CD1 mice at E15.5, E17.5, P1.5, P7.5, P12.5 and P23.5. Foxp2 mRNA levels were approximately 150-times higher than Foxp1 mRNA levels at E17.5 (Fig. 4A). Postnatally, Foxp2 expression declined rapidly, reaching similar levels to Foxp1 expression by P23.5. Similarly, Foxp2 expression levels also declined in the cortex (Fig. 4B). These data suggest that in the striatum Foxp2 predominantly forms homodimers during embryonic stages and heterodimers with Foxp1 during postnatal stages.

Figure 4

Foxp1 and Foxp2 expression level in the striatum during brain development. Expression level of Foxp1 and Foxp2 at different developmental stages measured in CD1 (ICR) mice by quantitative real-time PCR. Note that the y-axis is logarithmical. Data are expressed as mean, (n = 10–20/developmental stage).

Foxp1 and the androgen receptor are co-expressed in the striatum and cortex

Functional binding sites for the androgen receptor were identified via ChIP-seq in intronic sequences within the FOXP1 locus (20). These sites also exist in the murine Foxp1 locus (Supplementary Material, Fig. S2). We therefore postulated that higher testosterone levels in males may influence Foxp1 and Foxp2 expression levels, thereby modulating behavior. To address this, we examined androgen receptor (Ar) expression (a transcription factor directly activated by binding testosterone or dihydrotestosterone) in the striatum and cortex at E17.5 and P7.5 where sex-specific Foxp1 and Foxp2 expression was found. We detected expression of the androgen receptor at E17.5 and P7.5 in the analyzed tissues (Fig. 5A). Moreover, we confirmed nuclear co-expression of Foxp1 and the androgen receptor in the striatum and the cortex by immunohistochemistry (Fig. 5B).

Figure 5

Expression of Foxp1, Foxp2 and the androgen receptor in the striatum and the cortex. (A) RT-PCR demonstrating Foxp1, Foxp2, androgen receptor (Ar) and housekeeper Sdha1 and Hprt1 expression in the cortex and striatum at E17.5 and P7.5. At these stages a sex-dimorphic expression in the respective tissue was observed. (B) Specific antibodies against Foxp1 (green) and the androgen receptor (red) were used in the mouse brain at P1.5. Nuclei were stained with DAPI. Foxp1 and the androgen receptor (Ar) show co-expression in the striatum and the cortex at P1.5.

Foxp1 and Foxp2 expression is altered in the striatum and cortex of Ar^NesCre mice

If the androgen receptor regulates Foxp1 and/or Foxp2 expression in the striatum and cortex at P7.5, we would expect altered Foxp1/2 expression in a brain-specific androgen receptor (Ar) KO mouse. To examine this, we specifically deleted the androgen receptor in neural tissues (Nestin-Cre/Ar^−/−), hereafter referred to as Ar^NesCre mice. Site-specific recombination was checked on striatal and cortical cDNA from P7.5 WT and Ar^NesCre brains with specific primers flanking the floxed exon 2 (Fig. 6A). Ar^NesCre animals are viable and fertile with no obvious signs of illness. Foxp1 and Foxp2 expression was quantified in the striatum and cortex at P7.5 using quantitative real-time PCR. In addition, we also examined brain tissue of Ar^NesCre mice at P1.5 as during the perinatal period, testosterone levels are highest in the male brain at around P1 (21). We did not detect any alterations of Foxp1 and Foxp2 mRNA levels in the striatum and the cortex in Ar^NesCre pups at P1.5 (data not shown). However, we could demonstrate that Foxp1 and Foxp2 expression was significantly altered in Ar^NesCre mice compared to WT animals at E17.5 and P7.5 (Fig. 6B). At E17.5, Foxp1 expression was 27% lower in the Ar^NesCre striatum compared to WT controls (P = 0.015) (Fig. 6B). At P7.5, Foxp1 expression was 34% lower in the striatum (P = 0.010) and 31% lower in the cortex of Ar^NesCre mice (P = 0.002). In contrast, Foxp2 expression was 33% higher in the cortex of Ar^NesCre mice (P = 0.041) (Fig. 6B). If sex was considered separately, both male and female Ar^NesCre animals showed a reduced Foxp1 expression in the striatum at E17.5 and P7.5 but the values did not reach significance due to high variation (with the exception of female cortex at P7.5) (Supplementary Material, Fig. S3). These results indicate that androgens influence the expression of Foxp1 in the striatum at E17.5 and P7.5 and the expression of Foxp1 and Foxp2 in the cortex at P7.5.

Figure 6

Altered expression of Foxp1 and Foxp2 in the striatum and cortex of brain-specific Ar^NesCre mice at developmental stages E17.5 and P7.5. (A) RT-PCR on striatal and cortical cDNA showing site-specific recombination in E17.5 Ar^NesCre brains. (B) The relative expression level of Foxp1 and Foxp2 (marked in green) was quantified by real-time PCR. Data are expressed as mean, n = 16/16 for Foxp1 in striatum and Foxp2 in cortex at P7.5, n = 22/14 for Foxp1 in striatum at E17.5 and n = 15/14 for Foxp1 in cortex at P7.5, (*P ≤0.05, **P ≤ 0.01, t-test).

Discussion

Sex-specific differences have been described in the susceptibility to neurodevelopmental disease, however the molecular basis for these disparities have remained largely unexplained. A female ‘protective effect’ has been discussed, as the male brain seems to be more susceptible to develop ASD than females who carry a higher genetic burden. In the present study we used the brain-specific Foxp1 KO mouse as a model for ASD to unravel sex differences in social and communicative behaviors. We investigated early vocal ultrasonic communication and sex-specific expression of Foxp1 and its closest family member and interaction partner Foxp2 during late embryonic and early postnatal development.

While the function of Foxp2 has been previously addressed in great detail (18,22,23), there is still a lack of knowledge on the role of Foxp1. A previous study with body-wide Foxp1 reduction in Foxp1^+/−animals has given first evidence that Foxp1 haploinsufficiency results in reduced mouse USV, however sex-specific features were not addressed (8). Our study showed that Foxp1 KO pups of both sexes exhibit a strong reduction of ultrasonic distress calls compared to WT littermates which is in agreement with observations made in Foxp1^+/−animals (8).

As basal ganglia circuits are involved in USV in rodents and speech production in humans (22,24), defective calling behavior likely reflects the progressive striatal degeneration in Foxp1 KO animals which starts after birth (11). This degeneration process may also explain why the increasing number of calls normally observed in WT mice between P4.5 and P7.5 was not observed in Foxp1 KO mice. Moreover, since basal ganglia are known to interact reciprocally with the cortex via the thalamus, cortical defects may also occur and contribute to reduction of USV in Foxp1 KO mice. This was supported by the finding that lack of Foxp1 in the brain impairs neuronal migration within the developing cortex (25).

Vocal communication is sexually dimorphic in adult rodents and sensitive to sex steroids (26,27). However, sex differences in communication during early postnatal stages, when the brain is still developing, have not been well investigated. Interestingly, we observed that male WT pups call significantly more than females at P4.5 and P7.5 with a higher peak frequency at P7.5. These sex differences are abrogated in Foxp1 KO animals as the call number does not differ between male and female pups and females show the higher peak frequency at P7.5. The lacking sex-dimorphic call number, however, could be also due to a floor effect as the call rate is strongly reduced in both sexes.

Strikingly, the analysis of Foxp1 and Foxp2 expression in the WT striatum and cortex at different time points during development revealed that Foxp2 is expressed much higher than Foxp1 during embryogenesis, but declines postnatally, reaching similar levels to Foxp1 at around P23. As Foxp1 and Foxp2 execute their transcriptional activity as homo- or heterodimers (19), these findings provide important clues regarding the formation of heterodimers in the striatum, where both genes are co-expressed (28). At late embryonic stages when Foxp2 expression is at its peak, it is much more likely that Foxp2 functions through homodimers. When Foxp2 levels decline, the ratio of Foxp1/Foxp2 dramatically changes, with the possible outcome that these proteins predominantly regulate transcription as heterodimers. FOXP2 expression has already been shown to decline rapidly in the postnatal human brain, with sumoylation playing a role (29,30). The down-regulation of Foxp2 mRNA in mice at postnatal stages may imply that in humans as well reduced FOXP2 mRNA levels contribute to the decline of FOXP2 protein levels. Clearly, different abundances of Foxp1 and Foxp2 proteins in regions of co-expression will influence how these proteins dimerize, with potentially far reaching consequences on target gene regulation, brain development and behavior. Therefore it is likely that the sex-dimorphic Foxp2 expression in the striatum at P7.5 also contributes to the increased call number and peak frequency in male WT USV.

Sex differences in gene expression and splicing were previously studied in different brain regions in human post-mortem brains (31). A total of 17,501 genes were examined and only a small proportion of these genes (0.3%; 55/17.501) were found to be expressed in a sex-specific manner. Sex-specific FOXP1 and FOXP2 expression was not detected in the analyzed brain regions, though it is important to note that the examined brain samples derived from adult individuals (age 16–102), and therefore cannot provide information about sex-specific gene expression at early stages of postnatal brain development.

Differences in the onset and severity of ASD symptoms, as well as neuroanatomical differences between males and females, have pointed to an involvement of hormones, most likely steroid hormones. Increasing evidence suggests a complex interaction between sex hormones and gene regulation, which eventually lead to sex-specific differences in brain development (32), and presumably to sex-specific behavior. However, sex effects have not been well investigated in mouse models of ASD, except for exploratory and spontaneous stereotyped behaviors in male mice (33).

In mammals, testosterone production by the testis during fetal development is thought to influence the formation of a ‘male brain’ (34,35) and androgens seem to play a decisive role in vocal communication (36,37). The observation that male WT mice exhibit a lower Foxp2 expression in the striatum at P7.5 compared to female animals further hints to an androgenic regulation of Foxp2. Our additional results from CD1 animals which show that male WT mice compared to females have a higher Foxp1 and Foxp2 expression in the striatum at E17.5 and a lower expression of both genes at P7.5 might reflect the higher testosterone levels of males at E17.5, which strongly decline after P1.5, reaching even lower levels in males compared to females at P7.5 (21). However, it is likely that also other factors such as estrogens or non-hormonal causes such as cofactor expression contribute to the reversed expression. Interestingly, a study with rats showed that dihydrotestosterone (DHT) injection at P0 increased Foxp1 and Foxp2 expression in the striatum but reduced expression in the cerebellar vermis and cortex at P3. Moreover, the administration of DHT to female rat pups eliminated sex-specific differences in isolation-induced USV, with males producing more calls with lower frequency and amplitude (38).

Two functional androgen receptor binding sites within the human FOXP1 gene also exist in the murine Foxp1 and are highly conserved between human and mouse. To investigate a putative effect of androgens on Foxp1 and Foxp2 expression, we analyzed Ar^NesCre animals at developmental stage E17.5 and P7.5, when sex-differential expression of Foxp1 and Foxp2 was observed in WT mice. Remarkably, Foxp1 mRNA levels were reduced in the striatum of male and female Ar^NesCre mice at E17.5 and P7.5 and in the cortex at P7.5. In contrast, Ar^NesCre animals of both sexes showed an increased Foxp2 expression in the cortex at P7.5, which nicely correlates with the observation of androgens decreasing Foxp2 mRNA and protein levels in the rat cortex (38).

In summary, our findings demonstrate a sex-specific spatio-temporal pattern of Foxp1 and Foxp2 expression in the developing mouse brain, and we indicate that androgens contribute to this. Interestingly, this regulation seems to mediate sex-specific differences in USVs, commonly used as readout for altered social behavior. The results of our study therefore provide evidence that an altered regulation of FOXP1 and FOXP2 by testosterone in humans contributes to the development of social and communicative deficits such as those found in patients with ASD.

Materials and Methods

Animals

Mice were kept in the specific pathogen-free Interfacultary Biomedical Facility at the University of Heidelberg, under a 12 h light-dark cycle and given ad libitum access to water and food. All procedures were conducted in strict compliance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and approved by the National Institute of Mental Health animal care and use committee. CD1 (ICR) mice were used for sex-specific expression analyses at different developmental stages, as C57/BL6 litters are not large enough to perform these studies.

Embryonic stage was calculated by vaginal plug check and by controlling the morphological parameters which accord to the respective Theiler stage. The day of birth was considered as postnatal day (P) 0.5. Sex genotyping was performed by PCR detection of the male-specific Sry gene.

Generation of a conditional Foxp1 KO mouse

Female homozygous floxed Foxp1 mice (39) were crossed with male Nestin-Cre deleter mice (40) hemizygous for the floxed Foxp1 allele. In our study, we compared homozygous floxed Foxp1/Nestin-Cre animals (Foxp1 KO) with homozygous floxed Foxp1 mice which served as wild type (WT) control.

Generation of a conditional Ar^NesCre mouse

In brief, female homozygous floxed Ar mice (41) were crossed with male Nestin-Cre deleter mice (40) hemizygous for the floxed Ar allele. Ar flox mice (B6N.129-Ar^tm1Verh/Cnrm) were obtained from the European Mouse Mutant Archive (EMMA) which had been backcrossed into the C57Bl/6 background for over 12 generations prior to the arrival at our laboratory.

cDNA synthesis

Total RNA was prepared from frozen mouse brain tissue samples using peqGOLD TriFast™ (PEQLAB-Life Science). First strand cDNA synthesis was performed with 1.5 µg RNA using a Superscript II reverse transcriptase kit (ThermoFisher Scientific) and oligo dT_12-18-primers (ThermoFisher Scientific) according to the manufacturer’s instructions.

Quantitative real-time PCR

Quantitative real-time PCR was performed using the qTOWER system (Analytic Jena) with an annealing temperature of 60 °C using SYBR Green No-ROX Fast Mix (Bioline) according to manufacturer’s instructions. Each of the samples was analyzed in triplicate and relative mRNA levels were assessed using the Standard Curve Method by normalization to succinate dehydrogenase complex subunit A (Sdha1) and hypoxanthine phosphoribosyltransferase 1 (Hprt1). All primer sequences are listed in Supplementary Material, Table S1.

Protein analysis

Protein isolation was performed using standard protocols. Western blot analysis was executed using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA). We used the following primary antibodies: rabbit anti Foxp1 (Abcam, Cambridge, UK), rabbit anti Foxp2 (Abcam, Cambridge, UK), and mouse anti GAPDH (Abcam, Cambridge, UK). IRDye 800CW and IRDye 680 (LI-COR Biosciences, Lincoln, NE, USA) were used as secondary antibodies according to manufacturer’s instructions. Protein bands were quantified using Image Studio Lite 3.1 software (LI-COR Biosciences).

Pup USV recordings

The litters that were used for USV experiments had a maximum number of nine pups. At P3.5, animals were marked by foot tattoo with non-toxic animal tattoo ink. All measurements were performed in the light period at 22 °C. P4.5, P7.5 and P12.5 pups were isolated from the mother and littermates in random order and placed in the open field (42 cm × 42 cm × 42 cm) for 5 min for USV recordings, then immediately returned to the home cage. USV was recorded using an UltraSoundGate condenser microphone (CM16/CMPA, Avisoft Bioacoustics) placed 30 cm above the testing arena. The microphones were connected to a computer via an Avisoft UltraSoundGate USG416H audio device. USV recordings were analyzed using SASLabPro software (Avisoft Bioacoustics) and a fast Fourier transform (FFT) was conducted (512 FFT length, 100% frame, Hamming window, 75% time window overlap). To exclude software mistakes, the calls were also analyzed manually. The USV analysis was conducted blinded without knowing genotype and sex of the pups.

Immunofluorescence

A P1.5 mouse brain was fixed by perfusion with 4% paraformaldehyde and then dehydrated through an ethanol series and isopropanol. The brain was cleared in toluene prior to infiltration and embedded in paraffin for sectioning. 10 μm paraffin sections were prepared. Sections were deparaffinized and rehydrated through an ethanol series; thereafter slides were incubated in citrate buffer for antigen retrieval. Immunostaining was performed using standard protocols. Brain sections were incubated overnight with primary antibodies at 4 °C and incubated with appropriate fluorescent secondary antibodies for 1h at room temperature. The following primary antibodies were used: mouse anti Foxp1 (Abcam) 1:100, rabbit anti Ar (Abcam) 1:100.

Statistics

IBM SPSS STATISTICS 21 and Microsoft Office Excel software were used to analyze the data. Outliers in the data were determined via IBM SPSS STATISTICS 21 and excluded from further analysis. All data were checked for normal distribution via the Kolmogorov-Smirnov and Shapiro-Wilk test. If appropriate, two-way ANOVA was performed, in which litter was used as covariate. In all other cases, unpaired two-tailed Student’s t-tests served to compare differences between the two groups. P values of ≤0.05 were considered significant. All data are expressed as means + standard error.

Supplementary Material

Supplementary Material is available at HMG online.

Author Contributions

H.F. and G.A.R. designed the study. H.F. carried out and analyzed most experiments. R.R. executed the western blots and did the gene expression analysis in Ar^NesCre animals. N.S. performed the USV measurements. S.A. together with H.F. executed the gene expression analysis in WT animals. H.F. and G.A.R. wrote the manuscript.

Acknowledgements

The authors thank Rolf Sprengel, Thomas Kuner and Beate Niesler for helpful comments. We greatly acknowledge the advice of Claudia Pitzer and support from the Interdisciplinary Neurobehavioral Core for the USV measurements. We also thank Martin Granzow for help with R version 3.2.2 and Christine Fischer for statistical advice.

Conflict of Interest statement. None declared.

Funding

This work was supported by a grant from the Else Kröner-Fresenius-Stiftung (2013_A212) and the Land Baden-Württemberg. R.R. was supported by a DAAD scholarship (PKZ 91541533). G.A.R. is a member of the CellNetworks Cluster of Excellence.