PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of elifeeLifeRecent contentAbout eLifeFor authorsSign up for alerts
 
eLife. 2017; 6: e21012.
PMCID: PMC5384829

Embryonic transcription factor expression in mice predicts medial amygdala neuronal identity and sex-specific responses to innate behavioral cues

Carol A Mason, Reviewing editor
Carol A Mason, Columbia University, United States;

Abstract

The medial subnucleus of the amygdala (MeA) plays a central role in processing sensory cues required for innate behaviors. However, whether there is a link between developmental programs and the emergence of inborn behaviors remains unknown. Our previous studies revealed that the telencephalic preoptic area (POA) embryonic niche is a novel source of MeA destined progenitors. Here, we show that the POA is comprised of distinct progenitor pools complementarily marked by the transcription factors Dbx1 and Foxp2. As determined by molecular and electrophysiological criteria this embryonic parcellation predicts postnatal MeA inhibitory neuronal subtype identity. We further find that Dbx1-derived and Foxp2+ cells in the MeA are differentially activated in response to innate behavioral cues in a sex-specific manner. Thus, developmental transcription factor expression is predictive of MeA neuronal identity and sex-specific neuronal responses, providing a potential developmental logic for how innate behaviors could be processed by different MeA neuronal subtypes.

DOI: http://dx.doi.org/10.7554/eLife.21012.001

Research Organism: Mouse

eLife digest

Within the brain, a set of interconnected structures called the limbic system is involved in emotion, motivation and memory. This system – and in particular a structure called the medial amygdala – also contributes to behavioral drives that help an animal to survive and reproduce. These include the drive to avoid predators, to defend territory, and to find a mate. Such behaviors are thought to be inborn or innate. This means that animals display them instinctively whenever specific triggers are present, without the need to learn them beforehand.

However, just as a computer must be programmed to perform specific tasks, these innate behavioral responses must also be programmed into the brain. Given that animals do not learn these behaviors, Lischinsky et al. reasoned that specific events during the development of the brain must provide the animal’s brain with the necessary instructions. To test this idea, they studied how the development of the medial amygdala in mouse embryos may give rise to differences in innate mating behavior seen between male and female mice.

The medial amygdala contains many subtypes of neurons, which show different responses to sex hormones such as estrogen and androgen. Lischinsky et al. show that two sets of cells give rise to some of the different neurons of the adult medial amygdala. One set of these precursor cells makes a protein called Dbx1 and the other makes a protein called Foxp2. These two sets of precursors generate medial amygdala neurons with different arrays of sex hormone receptors in male and female mice. Moreover, while the two sets of medial amygdala neurons are activated during aggressive encounters, they show different patterns of activation in male and female animals during mating.

These findings suggest that the development of Dbx1-derived and Foxp2+ neurons in the medial amygdala helps program innate reproductive and aggressive behaviors into the brain. The new findings also provide insights into why these behaviors differ in male and female mice. The next challenge is to identify the inputs and outputs of these two distinct subpopulations of medial amygdala neurons. This should make it possible to work out exactly how these populations of cells control innate behaviors in male and female animals.

DOI: http://dx.doi.org/10.7554/eLife.21012.002

Introduction

One of the major functions of the limbic system is to integrate conspecific and non-conspecific environmental cues with social and survival salience to generate appropriate behavioral responses (Sokolowski and Corbin, 2012; Stowers et al., 2013). The medial subnucleus of the amygdala (MeA) serves as a hub in this function, residing only two synapses away from sensory neurons in the vomeronasal organ (Dulac and Wagner, 2006; Sokolowski and Corbin, 2012). The MeA along with the bed nucleus of the stria terminalis (BNST) and multiple nuclei of the hypothalamus including the ventromedial hypothalamus, form a core limbic circuit largely dedicated to processing innate behaviors (Dulac and Wagner, 2006; Gross and Canteras, 2012; Choi et al., 2005). Classical studies investigating patterns of neuronal activation in response to behavioral or olfactory cues (Kollack and Newman, 1992; Erskine, 1993), lesioning studies (Vochteloo and Koolhaas, 1987; Takahashi and Gladstone, 1988; Kondo, 1992) and more recent optogenetic approaches (Hong et al., 2014) have revealed a central role for the MeA in the regulation of innate behaviors such as aggression, mating and predator avoidance.

In addition to the processing of innate cues, the MeA is one of many known sexually dimorphic regions of the brain, with differences in numbers of neurons, glia, and synaptic organization between males and females (Cooke and Woolley, 2005; Johnson et al., 2008; McCarthy and Arnold, 2011). The critical role that the MeA plays in regulating sex-specific behaviors is reflected in the high expression levels of steroid hormone pathway proteins such as aromatase, estrogen receptor and androgen receptor (Wu et al., 2009; Juntti et al., 2010; Unger et al., 2015). MeA neuronal subpopulations expressing different combinations of these proteins have been shown to regulate aggression or mating behaviors in male and female mice (Juntti et al., 2010; Hong et al., 2014; Unger et al., 2015). Nonetheless, understanding how developmental programs are linked to behavioral processing in the MeA remains unknown.

As unlearned behaviors are largely inborn, we reasoned that there must be embryonic developmental programs that guide the formation of sub-circuitry dedicated for different innate behaviors. Previous studies of MeA development revealed that progenitors located at the telencephalic-diencephalic border are a major source of MeA neuronal populations (Zhao et al., 2008; Hirata et al., 2009; Soma et al., 2009; García-Moreno et al., 2010). Our previous work revealed that one of these progenitor populations is defined by the transient expression of the developmentally regulated transcription factor, Dbx1, which in turn generates a subclass of MeA putative inhibitory projection neurons (Hirata et al., 2009). However, the MeA is also comprised of diverse populations of local interneurons and both excitatory and inhibitory output neurons (Bian, 2013; Keshavarzi et al., 2014). This suggests the contribution of other progenitor subpopulations, perhaps also originating from the POA to MeA neuronal diversity, populations which may in turn play different roles in innate behavioral processing.

Here, we demonstrate that in addition to Dbx1 expression, a subset of MeA embryonic progenitors are complementarily marked by expression of Foxp2, a forkhead transcription factor implicated in the development and function of neurons and required in the motor coordinating centers of the brain for the appropriate production of speech (French and Fisher, 2014). We find this embryonic parcellation interestingly persists into postnatal stages where Dbx1-derived and Foxp2+ MeA neurons are separate, non-overlapping inhibitory output neuronal subpopulations. Strikingly, both subpopulations are activated during specific innate behaviors in a sex-specific manner. Thus, our findings link developmental patterning to innate behavioral processing and further provide an embryonic developmental framework for how these behaviors may emerge.

Results

Dbx1 and Foxp2 expression segregates embryonic and postnatal MeA subpopulations

Our previous work along with the work of others revealed that the telencephalic-diencephalic border is a major source of neurons that will populate the MeA (Zhao et al., 2008; Hirata et al., 2009; Soma et al., 2009; García-Moreno et al., 2010). Our previous studies (Hirata et al., 2009) revealed that the preoptic area (POA), which lies on the telencephalic side of this border (Flames et al., 2007), is a source of Dbx1+ progenitors fated to generate a subpopulation of MeA inhibitory output neurons. Our previous studies further revealed that progenitors arising from ventral telencephalic Shh+ and Nkx2.1+ domains also contributed to diverse neuronal subpopulations of the MeA (Carney et al., 2010). Thus, while a molecular map of MeA embryonic niche diversity is beginning to emerge, the diversity of mature neurons derived from this niche and whether there is a link between embryonic identity, mature identity and function remains unknown. Moreover, as these previously identified subpopulations only generate a subset of MeA neurons, we reasoned that there must be other transcription factor marked progenitor populations within the telencephalic-diencephalic niche.

Here, in addition to Dbx1+ progenitors, we observed a progenitor population comprised of Foxp2+ cells, residing primarily in the putative subventricular zone (SVZ) of the POA (Figure 1a–f,s). Interestingly, during embryogenesis, Dbx1-derived and Foxp2+ progenitor populations were non-overlapping (Figure 1a–l). Both populations were also generally distinct from OTP+ progenitors (Figure 1—figure supplement 1a–i), a population previously shown to define a subset of MeA-fated progenitors (García-Moreno et al., 2010). We next investigated whether Foxp2+ progenitors overlapped with ventral telencephalic populations derived from Shh or Nkx6.2 lineages, which also encompass the POA (Carney et al., 2010; Fogarty et al., 2007). We found embryonic Foxp2+ cells were not derived from either lineage (~5% overlap) (Figure 1—figure supplement 2) further expanding our knowledge of the molecular diversity of the MeA niche.

Figure 1.
Embryonic and postnatal segregation of Dbx1-derived and Foxp2+ cells.

Interestingly, this embryonic molecular parcellation persisted into adulthood as Dbx1-derived and Foxp2+ cells remained non-overlapping across the rostro-caudal extent of the postnatal MeA (Figure 1m–r; Figure 1—figure supplement 3). Similarly, postnatally, the Dbx1-derived and Foxp2+ neurons remained distinct from OTP+ cells (Figure 1—figure supplement 1j–x). Taken together, these findings reveal that Dbx1-derived and Foxp2+ populations, although appearing to derive from the same embryonic niche, remain distinct subpopulations from embryonic development to adulthood (Figure 1t), a novel finding that we hypothesize has implications for later subtype identity and function, explored in the next sets of experiments.

Foxp2+ neurons in the postnatal amygdala are inhibitory

Our previous work revealed that MeA Dbx1-derived neurons are a subclass of inhibitory neurons, likely projection as opposed to local interneurons (Hirata et al., 2009). However, the identity of MeA Foxp2+ neurons remains unknown. Therefore, we next examined whether adult MeA Foxp2+ cells were neurons or glia. Our analysis revealed that 81% ± 2.6 of Foxp2+ cells expressed NeuN, a pan neuronal marker (Mullen et al., 1992) (Figure 2a–d), with none co-expressing the oligodendrocyte marker, CC1 (Koenning et al., 2012) (Figure 2e–h). We next wanted to determine if Foxp2+ neurons were excitatory or inhibitory. We found that only 3% ± 1.2 of Foxp2+ cells were derived from the Emx1-lineage, a broad marker of excitatory neurons (Gorski et al., 2002) (Figure 2i–l). We further found that 22% ± 6.6 of Foxp2+ cells expressed the inhibitory marker Calbindin (Figure 2m–p), with a smaller percentage of Foxp2+ cells (15% ± 2.5) expressing nNOS (Figure 2q–t), or somatostatin (5% ± 0.6) (Figure 2u–x), inhibitory markers that mark a subset of MeA output neurons (Tanaka et al., 1997) and interneurons (Ascoli et al., 2008), respectively. Collectively, although we did not fully assess all putative inhibitory markers, these data reveal that Foxp2+ MeA are not excitatory and are likely inhibitory (Figure 2y).

Figure 2.
Identity of Foxp2+ medial amygdala neurons.

As neither Dbx1-derived nor Foxp2+ cells expressed OTP+, we therefore investigated the identity of this population. We found that 1% ± 0.4 of OTP+ cells were derived from the Emx1-lineage, and 13.5% ± 9.1 co-expressed CAMKIIα (Jones et al., 1994), both excitatory markers. In contrast, we observed 56% ± 18.9 of OTP+ cells co-expressed calbindin, while none (0% ± 0.1) co-expressed somatostatin (Figure 2—figure supplement 1). Therefore, similar to Dbx1-derived and Foxp2+ neurons, the majority of OTP+ cells appear to be inhibitory.

Dbx1-derived and Foxp2+ neurons possess distinct electrophysiological properties

To determine if the Dbx1-derived and the Foxp2+ populations are functionally distinct subclasses, we next examined their electrophysiological properties. Previous studies (Bian, 2013; Keshavarzi et al., 2014) revealed a significant diversity in intrinsic electrophysiological properties of MeA local and projection neurons. Here, we found that the majority (19/28) of Dbx1-derived neurons were characterized by a regular, tonic spiking pattern with 3–4 spikes at rheobase (Figure 3a). In contrast, the majority (15/23) of Foxp2+ neurons were distinguished by a phasic firing pattern and displayed a single or double spike upon repolarization after hyperpolarization, a profile characteristic of inhibitory neurons (Llinás, 1988) (Figure 3b). Dbx1-derived and Foxp2-derived neurons (confirmed Foxp2+ by immunohistochemistry) also displayed significant differences in resting membrane potential, input resistance, capacitance, and action potential frequency but not in rheobase (Figure 3c–g). In addition, the presence of spines in Foxp2+ neurons (Figure 3—figure supplement 1) suggested that similar to Dbx1-derived neurons, Foxp2+ neurons are projection neurons. This reveals that despite both populations being inhibitory, the Dbx1-derived and the Foxp2+ populations possess distinct firing patterns.

Figure 3.
Dbx1-derived and Foxp2+ MeA neurons possess distinct electrophysiological properties.

We further analyzed spontaneous excitatory post-synaptic currents (sEPSCs), a measure of excitatory inputs. Dbx1-derived neurons received significantly more frequent and greater amplitude of sEPSCs than Foxp2+ neurons (Figure 3h–l). This suggests that Dbx1-derived MeA neurons receive a greater number and/or stronger excitatory inputs than Foxp2+ neurons. In summary, a combination of neuronal marker expression (Figure 2) and electrophysiological (Figure 3) analyses, combined with our previous analysis (Hirata et al., 2009) revealed that Dbx1-derived and Foxp2+ neurons are distinct subclasses of inhibitory, and are likely projection neurons.

Molecular identity of Dbx1-derived and Foxp2+ postnatal MeA cells

Based on the above analyses revealing that Dbx1-derived and Foxp2+ neurons are separate subclasses, we next wanted to determine whether these two populations express different combinations of steroid pathway proteins previously associated with MeA function such as estrogen receptor-alpha (ERα), aromatase and androgen receptor (AR) (Wu et al., 2009; Juntti et al., 2010; Unger et al., 2015). As the MeA is a sexually dimorphic nucleus (Cooke and Woolley, 2005; McCarthy and Arnold, 2011; Johnson et al., 2008), we characterized the expression of these markers in both male and female mice (Figure 4, Figure 4—figure supplement 1, Figure 4—figure supplement 2). We found that Dbx1-derived and Foxp2+ cells in males expressed ERα to the same extent (28.4% ± 4.8 in Dbx1-derived cells; 24.0% ± 7.2 in Foxp2+ cells). However, Dbx1-derived and Foxp2+ cells in females showed significant differences in ERα expression (45% ± 3.4 in Dbx1-derived cells; 24.8% ± 5.8 in Foxp2+ cells) (Figure 4a–g). The majority of Dbx1-derived cells expressed aromatase both in males (61.7% ± 7.6) and females (52.4% ± 5.2), which was at a significantly higher percentage than in Foxp2+ cells in males (0.12% ± 0) and females (7.2% ± 6.0) (Figure 4h–n). There were no subpopulation differences in AR expression as both Dbx1-derived and Foxp2+ neurons in both males (26.8% ± 4.1 in Dbx1-derived cells; 16.2% ± 3.2 in Foxp2+ cells) and females (8.4% ± 3.6 in Dbx1-derived cells; 7.0% ± 1.8 in Foxp2+ cells) co-expressed AR at the same levels (Figure 4o–u). However, there were sex-specific differences observed as a greater percentage of Dbx1-derived cells in males (26.8% ± 4.1) expressed AR than Dbx1-derived cells in females (8.4% ± 3.6).

Figure 4.
Expression of sex hormone pathway markers in Dbx1-derived and Foxp2+ cells.

We also analyzed the contribution of the Dbx1-derived and Foxp2+ populations to the total ERα, Aromatase and AR MeA populations (Figure 4—figure supplement 2). We observed that both Dbx1-derived and Foxp2+ cells comprised between ~10% to 22% of the total ERα population in male and female mice. The Dbx1-derived population contributed between ~30–40% of the total aromatase population in males and females, which was significantly greater than the contribution of the Foxp2+ population in both males and females. Dbx1-derived (males only) and Foxp2+ cells (males and females) comprised ~20–38% of the total AR+ population. In contrast, the Dbx1-derived population contributed to only 3% of the total AR+ population in females, which was significantly less than the Foxp2+ contribution in females and less than the contribution of Dbx1-derived cells in males. Collectively, these data reveal that Dbx1-derived and Foxp2+ cells contributed differentially to the aromatase and AR populations, but not to the ERα population.

Sex-specific subtype activation patterns during innate behaviors

The MeA receives direct inputs from the accessory olfactory bulb (AOB) (Scalia and Winans, 1975; Martel and Baum, 2009; Bergan et al., 2014) and integrates this chemosensory information to regulate innate behaviors including territorial aggression, maternal aggression, mating, and predator avoidance (Dulac and Wagner, 2006; Kim et al., 2015). Previous data revealed that at least aggressive and mating behaviors are controlled by MeA GABAergic neurons (Choi et al., 2005; Hong et al., 2014). However, whether different subsets of inhibitory neurons are activated during these behaviors, or if neuronal subtype activation is generalizable across behaviors remains unknown. To directly test these possibilities we performed well characterized aggression, mating and predator odor avoidance behavioral tests in both male and female resident mice and examined the patterns of activation of Dbx1-derived and Foxp2+ cells using c-fos as a readout of neuronal activity in resident mice (Figure 5—figure supplement 1).

Dbx1-derived and Foxp2+ neurons are activated during aggressive encounters

To evaluate activation of Dbx1-derived and Foxp2+ neurons during male conspecific aggression, we performed a territorial aggression assay in which an intruder mouse was placed in a resident cage. Concordant with previous literature (Wang et al., 2013), we found a significant increase in the number of c-fos+ cells in the MeA in comparison to naïve mice (Figure 5a–c). In addition, both the number and proportion of activated Dbx1-derived and Foxp2+ subpopulations in males were significantly increased during territorial aggression compared to control (naïve) mice (Figure 5d–k,x). We next evaluated aggression in female mice by conducting a maternal aggression assay in which pups were removed from a nursing female and a sexually naïve male intruder was introduced (Haney et al., 1989). In addition to the naïve control, we established a second control in which pups were removed and no intruder was presented. With this control, we could compare levels of neuronal activation during maternal aggression (pups removed and presence of intruder) to levels of neuronal activation in response to a strong stressor (pups removed but no intruder) and to a naïve condition (with pups and without intruder). We found a significant increase in the number of c-fos+ cells in the MeA in the maternal aggression condition compared to both stressed and naïve controls (Figure 5l–o). When examining subtype-specific levels of activation, we found that both the number and proportion of activated Dbx1-derived and the Foxp2+ subpopulations significantly increased during maternal aggression in comparison to the naïve condition (Figure 5p–w,y). Thus, during an aggressive encounter with a conspecific, both Dbx1-derived and Foxp2+ MeA subpopulations were activated to a greater extent in both male and female mice.

Figure 5.
Dbx1-derived and Foxp2+ neurons are activated during aggressive encounters.

Dbx1-derived and Foxp2+ neurons are activated in a sex-specific manner during mating

We next conducted male and female mating assays and monitored the animals for mating (mounting and intromission followed by the presence of a vaginal plug). Consistent with previous studies (Rowe and Erskine, 1993), we observed an increase in c-fos+ cells in the MeA during both male and female mating (Figure 6a–c). Intriguingly, we found subpopulation specific differences in activation patterns across sexes. While the number and proportion of the activated Dbx1-derived subpopulation was significantly increased during both male and female mating when compared to naïve controls (Figure 6d–g,l), the number and proportion of activated Foxp2+ cells increased only during male but not during female mating (Figure 6h–k,m). Thus, while Dbx1-derived and Foxp2+ MeA subpopulations were both activated during male mating, only Dbx1-derived neurons were activated during female mating.

Figure 6.
Sex-specific subpopulation responses during mating.

Activation of Dbx1-derived and Foxp2+ neurons during predator odor exposure

We next sought to determine whether Dbx1-derived and Foxp2+ MeA subpopulations were activated by a strong innate stressor. To accomplish this, we exposed mice to rat odor, a well-characterized predator cue that evokes a strong aversive response in mice (Apfelbach et al., 2005; Carvalho et al., 2015Sokolowski et al., 2015). Mice were exposed to soiled bedding from a rat cage (predator) or clean bedding (benign) as a control (Figure 7). Consistent with previous studies (Canteras et al., 2015), in response to predator odor we observed a significant increase in the number of c-fos+ cells in the MeA in both male and female mice (Figure 7a–c). Interestingly, the Dbx1-derived cells in male and female mice were not activated in response to predator odor exposure in comparison to the benign bedding (Figure 7d–g). In contrast, we observed a significant activation of Foxp2+ cells in comparison to controls in female mice, but not in males (Figure 7h–k). When we assessed the percentage of both Dbx1-derived (YFP+;c-fos+/total YFP+) and Foxp2+ (Foxp2+;c-fos+/total Foxp2+) cells co-labeled with c-fos over the total population there was no significant difference in either subpopulation nor in male or female mice compared to control (Figure 7m–l). This suggests that despite an increase in the number of activated Foxp2+ cells in female mice in the presence of rat bedding, this increase might not be physiologically relevant. Therefore, to investigate which MeA subpopulation may be responding to predator odor, we next examined the activation patterns of the OTP+ population. We observed that a greater number of OTP+ cells in both males and females were activated in response to predator odor (Figure 7—figure supplement 1). When analyzing the proportion of OTP+ cells activated, we found no differences between the benign and the predator avoidance groups in males, but we did find a significant increase in the OTP+ population in females. Therefore, in the female brain, the OTP+ population, in contrast to Dbx1-derived or Foxp2+ cells, were activated at a level above control in response to a strong aversive innate olfactory cue.

Figure 7.
Dbx1-derived and Foxp2+ MeA subpopulation activation patterns in response to predator odor.

In summary, our analyses of activation patterns in response to a battery of innate behavior tasks revealed that Dbx1-derived and Foxp2+ cells in the MeA were differentially activated depending on the instinctive behavioral task (Figure 8). The most striking of these differences occurred during mating behaviors. While both subpopulations were activated during male mating, in the female MeA only Dbx1-derived neurons were activated, with no activation of the Foxp2+ population. In contrast, the OTP+ population in females appears more tuned to a predator odor cue. Importantly, sex differences in patterns of neuronal activation during mating and predator odor were not due to overall differences in the activation of the MeA as these cues activated the MeA in both sexes. Taken together with our electrophysiological and molecular findings of the Dbx1-derived and Foxp2+ populations, our data reveal that the developmental parcellation of MeA progenitors predicts mature neuronal identity and sex-specific innate behavioral activation patterns.

Figure 8.
Summary of findings.

Discussion

Across a variety of species, innate behaviors such as aggression, mating and avoidance of predators are initiated by sensory cues primarily detected by olfaction (Dulac and Wagner, 2006; Stowers et al., 2013). Here, focusing on the medial amygdala (MeA), a critical brain region for the processing of olfactory-based sensory cues for unlearned behavior in vertebrates, we took a neural developmental approach to shed light on how innate behavioral information may be encoded in the male and female brain. Integrating genetic fate-mapping, patch clamp electrophysiology and animal behavioral assays we uncover a fundamental link between embryonic patterning and brain responses to innate behavioral cues at two levels: (1) differential transcription factor expression within the embryonic MeA progenitor niche predicts mature output neuronal subtype identity and (2) further predicts subpopulation sex-specific responses to mating and predator odor avoidance cues. Our findings further suggest that transcription factor expression at the progenitor stage may be instructive for the establishment of neuronal populations and sub-circuits regulating sex-specific behaviors.

Potential transcription factor codes for establishment of MeA neuronal diversity

The generation of neuronal diversity across amygdala subnuclei has been posited to occur in a compartmentalized manner with amygdala inhibitory neurons generated in the subpallial ganglionic eminences and excitatory neurons arising from the cortical pallial region (Swanson and Petrovich, 1998). In this model, the amygdala and cerebral cortex develop by a similar mechanism with neurons in both structures originating in shared progenitor domains. However, more recent studies have revealed that the generation of amygdala neuronal diversity is more complex with large populations of neurons originating in progenitor niches dedicated for limbic structures (Remedios et al., 2007; Hirata et al., 2009; Soma et al., 2009; Waclaw et al., 2010; García-Moreno et al., 2010). One of these major niches encompasses the region at the telencephalic-diencephalic border, an origin of MeA output neurons (Hirata et al., 2009; García-Moreno et al., 2010). Our previous studies revealed that the homedomain encoding transcription factor, Dbx1, marks a subpopulation of progenitors within the POA, which will later generate a subset of MeA inhibitory output neurons (Hirata et al., 2009). Here, we significantly extend this work by revealing the presence of a complementary population of progenitors within this niche marked by expression of Foxp2. Thus, our findings, combined with previous work, suggest a model of MeA development in which distinct progenitor populations at the telencephalic-diencephalic border defined by differential transcription factor expression (e.gs. Dbx1, Foxp2, OTP) are a major source for MeA neuronal diversity.

The function of combinatorial sets of transcription factors in neural progenitors has been shown across the neuraxis as the mechanism for the generation and specification of distinct subclasses of neurons (Kepecs and Fishell, 2014; Stepien et al., 2010; Shirasaki and Pfaff, 2002). In addition to specification of neuronal subtype identity, recent studies in the spinal cord and globus pallidus (Dodson et al., 2015; Bikoff et al., 2016) have revealed that different combinatorial codes in progenitor pools predict neuronal subtype connectivity patterns, neuronal firing properties and in the case of the globus pallidus, distinct functions in regulating voluntary movements (Dodson et al., 2015). Thus, transcription factor expression at the earliest stages of neuronal development likely represents the beginning of an instructive continuum for the establishment of not only neuronal identity, but also development of sub-circuitry regulating different components of motor behaviors. Here, we show that in the MeA, complementary transcription factor expression marks subsets of progenitors and predicts neuronal subtype identity as defined by molecular and electrophysiological signatures. At the molecular level, Dbx1-derived and Foxp2+ neurons express different combinations of the sex steroid hormone pathway protein aromatase and estrogen receptor alpha (ERα). At the electrophysiological level, these two populations possess distinct intrinsic electrophysiological profiles. Thus, our study generates a novel cell-specific transcription factor-based means to predict postnatal MeA neuronal identity.

Behavioral activation of Dbx1-derived and Foxp2+ neurons

The central importance of the MeA for processing innate behaviors such as aggression, mating and predator avoidance is well-established (Dulac and Wagner, 2006; Sokolowski and Corbin, 2012). Despite this knowledge, the question of which MeA neuronal subtypes encode instinctive behavioral information has only recently begun to be addressed. Recent optogenetic manipulation of the MeA revealed that glutamatergic neurons mediate repetitive self-grooming behaviors while in contrast GABA-ergic neurons regulate either aggressive or mating behaviors, depending on the level of neuronal activity driven by light stimulation (Hong et al., 2014). Moreover, MeA neuronal subclasses expressing different components of the stress response system control appropriate behavioral responses to social cues (Shemesh et al., 2016). Here, we contribute to the understanding of amygdala cell-specific regulation of behavior by generating a transcription factor based map of MeA subtype responsiveness to innate-behavioral cues. Thus, our findings provide a developmental molecular context in which to further dissect neuronal subtype control of MeA-driven behaviors.

In addition to playing a central role in processing sensory information required for instinctive behaviors, the MeA is one of the known sexually dimorphic structures of the brain (Cooke and Woolley, 2005; McCarthy and Arnold, 2011; Johnson et al., 2008). Recent studies employing in vivo recording techniques revealed that a significant number of MeA neurons are dedicated to processing olfactory cues from the opposite sex rather than the same-sex, thus providing a direct demonstration of sex-specific differences in sensory processing (Bergan et al., 2014). However, the identity of MeA neurons in males and females that differentially process olfactory-based sensory information has not been delineated. Here, we reveal that while Dbx1-derived and Foxp2+ neurons are broadly activated by mating, aggressive and predator cues, we found stark differences in Dbx1-derived, Foxp2+ and OTP+ cell-specific responses in the male and female brain to mating and predator odor cues. A similar sex-specific control of innate behavior has been directly demonstrated in the ventromedial hypothalamus (VMH), where progesterone receptor-expressing neurons while required for male aggressive and mating behaviors, are only required for female mating behavior (Yang et al., 2013). Collectively, our studies in combination with previous studies point to a larger picture in which there are neuronal subpopulations in the MeA and VMH that are involved in the regulation of different innate behaviors in a sex-specific manner.

Although our data do not reveal the neuronal and/or circuit mechanisms underlying our observation of sex-specific subpopulation responses to mating behavior and predator odor presentation, some of our findings do provide potential insight as to how this differential processing may occur. The two most straightforward and non-exclusive potential mechanisms are: 1) with regard to mating, intrinsic differences in Dbx1-derived and Foxp2+ neurons and/or 2) subpopulation specific patterns of local and/or long-range connectivity.

Regarding the first potential mechanism, sex steroid hormones and receptors such as aromatase, ERα and AR have been extensively characterized as critical for the output of distinct components of male and female aggressive and mating behaviors (Yang and Shah, 2014). For example, deletion of AR resulted in alterations in attack duration during territorial aggression and initiation of male mating (Juntti et al., 2010), while ablation of aromatase neurons led to impairments in the production of distinct components of aggression in male and female mice (Unger et al., 2015). Consistent with the critical role that these factors play in components of innate behaviors, we found that aromatase is expressed solely in the Dbx1-derived lineage. Across species, aromatase has a masculinizing effect (Wu et al., 2009; Balthazart et al., 2011). Thus, it will be interesting to explore how Dbx1-derived aromatase expressing neurons may control male behavioral displays such as mounting and territorial aggression.

Furthermore, both Dbx1-derived and Foxp2+ neurons possess distinct electrophysiological profiles, another potential mechanism to control different components of behaviors. Previous work in both vertebrates and invertebrates has revealed that the timing of AP spiking is directly linked to specific behavioral actions. For example, in vertebrates timescale firing differences are associated with dopamine (DA) release for the determination of reward behaviors (Schultz, 2007; Zhang et al., 2009). It will therefore prove interesting to explore if and how firing patterns of Dbx1-derived and Foxp2+ MeA neurons may control different components of innate behaviors across sexes.

The second potential and perhaps more intriguing mechanism that may account for Dbx1-derived and Foxp2+ subtype specific male versus female patterns of neuronal activation during mating are sex specific local and/or long-range patterns of connectivity. Although currently not yet observed in a brain circuit in mammals, such a mechanism has recently been uncovered in c. elegans in which shared male and female circuits show differences in connectivity that is established during wiring (Oren-Suissa et al., 2016). Consistent with this, there is some suggestion of sex-specific differences in olfactory-MeA projections in rodents (Kang et al., 2011). Although we did not differentiate according to the sex of the animal, we found differences in both the amplitude and frequency of EPSCs between lineages, indicating differences in the strength and/or number of inputs between Dbx1-derived and Foxp2+ MeA neurons. Although the source of input cannot be determined from our analysis, there are direct excitatory inputs to the MeA emanating from the mitral/tufted neurons of the accessory olfactory bulb (AOB) (Martel and Baum, 2009; Bergan et al., 2014). Thus, determination of input/output wiring patterns of male and female Dbx1-derived and Foxp2+ MeA neurons will be highly informative.

In summary, although the precise instructive developmental mechanisms programming innate behaviors remain to be elucidated, we reveal that differential transcription factor expression during development is predictive of neuronal identity based on molecular and electrophysiological criteria and sex-specific patterns of neuronal activation during innate behaviors.

Materials and methods

Animals

Mice were housed in the temperature and light-controlled (12 hr light-dark cycle) Children’s National Medical Center animal care facility and given food and water ad libitum. All animal procedures were approved by the Children’s National Medical Center’s Institutional Animal Care and Use Committee (IACUC) and conformed to NIH Guidelines for animal use. Mice used were Dbx1cre+/- (kindly provided by A. Peirani, Institut Jacques Monod, Paris), Shhcre+/- (Jackson Labs strain B6.Cg-Shhtm1(EGFP/cre)Cjt/J), Emx1cre+/- (Jackson Labs strain B6.129S2-Emx1tm1(cre)Krj/J), Nkx6.2cre+/-(Jackson Lab strain Tg Nkx6-2-icre)1Kess/SshiJ), Foxp2cre+/-(kindly provided by R. Palmiter, University of Washington) (Rousso et al., 2016), all crossed to Rosa26YFP+/+mice (Jackson lab strain R26R-EYFP). For analysis of aromatase expression, we used Aromatase LacZ reporter mice (kindly provided by N. Shah, University of California-San Francisco) (Wu et al., 2009). Mice were genotyped by Transnetyx Inc. Genotyping Services. Adult mice were considered between 3–7 months of age. The sample size was based on previous experiments and published data. No statistical methods were used to determine sample sizes.

Immunohistochemistry

Postnatal mice were transcardially perfused with 4% paraformaldehyde (PFA) and brains post fixed overnight at 4°C, embedded in 4% agarose (Invitrogen) and sectioned at 50 μm with a vibrating microtome (Leica VT1000S). Embryos were fixed in 4% PFA overnight at 4°C degrees, cryoprotected in 30% sucrose, embedded in O.C.T. compound (Tekka) and sectioned at 20 μm on a cryostat (Leica CM1850). For IHC, tissue sections were incubated overnight with primary antibody, then washed with PBST and 10% normal donkey serum and incubated for 4 hr with the corresponding secondary antibodies, and mounted with DAPI Fluoromount (SouthernBiotech 0100–20, Birmingham, AL). Primary antibodies used were rat anti-GFP (to detect YFP expression, (1:1000, Nacalai 04404–84, San Diego, CA), goat anti-Foxp2 (1:200; Santa Cruz sc-21069, Dallas, TX), rabbit anti-Foxp2 (1:500; abcam ab16046, Cambridge, UK), rabbit anti-OTP (1:2000; kind provided by F. Vaccarino, Yale University), rabbit anti-androgen receptor (1:750; Epitomics AC-0071, Cambridge, UK), rabbit anti-estrogen receptor α (1:6000; Millipore 06–935); mouse anti-NeuN (1:200; Millipore MAB-377), mouse anti-CC1 (1:250; Calbiochem OP80-100), goat anti-calbindin (1:200; Santa Cruz sc-7691), rat anti-somatostatin (1:100; Millipore MAB354, Billerica, MA), rabbit anti-cfos (1:500; Santa Cruz sc-52, Dallas, TX), goat anti-cfos (1:300; Santa Cruz sc-52G, Dallas, TX), anti-rabbit nNOS (1:8000; ImmunoStar 24287, Hudson, WI), mouse anti-CAMKIIα (1:500 Biomol ARG22260.50, Farmingdale, NY) and chicken anti-βGal (1:2000; abcam 9361, Cambridge, UK). Secondary antisera used were donkey anti-rat or anti-goat Alexa 488 (1:200; Life Technologies, Waltham, MA), anti-rabbit or anti-goat Cy5 (1:1000; Jackson ImmunoResearch, Westgrove, PA), anti-rabbit or anti-mouse Cy3 (1:1000; Jackson ImmunoResearch, Westgrove, PA), anti-mouse dylight 649 (1:500; Jackson ImmunoResearch, Westgrove, PA), or anti-chicken dylight 549 (1:500; Jackson ImmunoResearch, Westgrove, PA).

Microscopy

Fluorescent photographs were taken using an Olympus FX1000 Fluoview Laser Scanning Confocal Microscope (1 um optical thickness).

Quantification

Molecular marker analysis

For embryonic analysis, the average of 2–3 sections encompassing the POA were imaged and quantified (Figure 1; Figure 1—figure supplements 12). For adult analyses (Figures 17; Figure 1—figure supplements 12; Figure 2—figure supplement 1; Figure 4—figure supplement 1; Figure 5—figure supplement 1; Figure 7—figure supplement 1), every sixth serial coronal section encompassing the anterior to posterior extent of the MeA (Bregma −1.30 to −1.90, see Figure 1—figure supplement 3) was immunostained with antibodies to Foxp2, YFP and markers of interest. Quantification was done by counting single and double-labeled cells encompassing the entire domain of expression within the POA or MeA (Figures 12 and and4;4; Figure 1—figure supplements 12; Figure 2—figure supplement 1; Figure 4—figure supplement 1).

Neuronal activation

For analyses of neuronal activation (Figures 57, Figure 7—figure supplement 1), a single MeA section with the highest number of c-fos+ cells corresponding to the presence of Dbx1-derived, Foxp2+ or OTP+ cells was chosen for quantification. c-fos, YFP and Foxp2 triple immunostaining was conducted on the same section and single and double-labeled cells counted. c-fos and OTP double immunostaining was conducted and single and double cells counted.

Statistical evaluation

Unless otherwise stated, data were analyzed using GraphPad six statistical software. We first tested the distribution of the data with the Shapiro-Wallis test for normality. Data that were normally distributed was analyzed using an unpaired two-tailed t-test for analysis of experiments involving two groups (Figure 1—figure supplement 1; Figure 1—figure supplement 2; Figure 5x; Figure 6c females, k, m, l; Figure 7c,g males, k, l, m Foxp2+ subpopulation; Figure 7—figure supplement 1c,d,h,i) and a one-way ANOVA (Figure 5o,w,y Foxp2+ subpopulation) followed by Tukey-Kramer multiple comparison test was used for analysis of experiments involving three or more groups. Data with a non-normal distribution were analyzed by using the non-parametric test Mann-Whitney when comparing two groups (Figure 5c,g,k,y Dbx1-derived subpopulation; Figure 6c males, g; Figure 7g females, m Dbx1-derived subpopulation; Figure 7—figure supplement 1j,k) and Kruskal-Wallis with Dunn’s post-hoc corrections for data with three groups (Figure 5—figure supplement 5s). For the analysis of data shown in Figure 4 and Figure 4—figure supplement 1 we performed the following statistical analysis: when the data met the normality assumption or could be transformed to meet the normality assumption, generally two-way analysis of variance models were implemented to evaluate the evidence of differences in mean effects of the two experimental factors on cell activation (Figure 4g,u; Figure 4—figure supplement 2a,b). In the situation where no data transformation could be found to achieve an acceptable level of normality, quantile regression was performed, which does not require the normality assumption, to evaluate comparable differences in median effects (Figure 4n; Figure 4—figure supplement 2c). In each case, the initial models included a cross-products term to assess evidence of effect modification or interaction. When there was no interaction, it was taken as evidence of the absence of effect modification and the cross-products term was removed leaving only a model that assessed independent effects of the each factor separately, while holding constant any effects of the other factor. Depending on the underlying model, either mean or median effects ± 95% confidence intervals were derived to reflect the differences that were consistent with statistically meaningful differences in the final model. Under consideration of protecting the experiment-wise error rate, as long as the evaluation of differences focused only on identifying the nature of effects deemed statistically meaningful in the final model, there was no correction made for multiple comparisons. Analysis of data meeting the normality assumption was based on GraphPad Prism six and analysis based on quantile regression was implemented in Stata 13. As mice had to be sacrificed after each behavioral assay in order to conduct c-fos immuno-analysis, technical repeats were not available. Measurements from different mice were considered biological repeats to determine sample size. Data points were considered outliers and excluded if they were two standard deviations away from the mean.

Electrophysiology and biocytin filling

Mice (P25-40) were anaesthetized with isoflurane and sacrificed. Brains were removed and immediately immersed in ice-cold oxygenated (95% O2/5% CO2) sucrose solution (234 mM sucrose, 11 mM glucose, 26 mM NaHCO3, 2.5 mM KCl,1.25 mM NaH2PO4.H2O, 10 mM MgSO4 and 0.5 mM CaCl2). Coronal slices of 300 µm in thickness were cut. Slices with amygdala were collected and placed in oxygen-equilibrated artificial cerebral spinal fluid (ACSF) as previously described (Hirata et al., 2009). Either Dbx1-derived or Foxp2-derived neurons were then visualized using a fluorescent lamp (Nikon) with a 450-490λ filter. Whole-cell patch-clamp recordings from YFP-positive fluorescent cells were performed at room temperature with continuous perfusion of ACSF (Multiclamp 700A, DigiDATA1322, Molecular Devices). Intracellular solution (in mM): 130 Kgluconate, 10 KCl, 2 MgCl2, 10 HEPES, 10 EGTA, 2 Na2-ATP, 0.5 Na2-GTP. All measurements of intrinsic and synaptic properties were analyzed off-line using Clampfit Software (V.10.2, Molecular Devices) and graphing software (OriginPro 9.1). At the end of each recording, biocytin (1%) was injected with the depolarizing current (1nA) for post-hoc morphology analysis. All slices were then fixed with paraformaldehyde overnight at 4°C and processed for Fluorescein-conjugated Avidin-D (1:200, Vector Laboratories), YFP IHC (Dbx1-derived and Foxp2+ recordings) or Foxp2 IHC (for Foxp2+ recordings) as described above.

Neuronal reconstruction

Neurons were filled with biocytin and imaged using an Olympus FX1000 Fluoview Laser Scanning Confocal Microscope (0.5 um optical thickness). VIAS software was used to align confocal images taken at 40x and 60x in the same plane (x,y,z). Neurons were then traced using neuTube software, which uses fixed radii small tubes to estimate the dendritic branches length and thickness (Feng et al., 2015).

Behavioral assays

Dbx1cre+/-;Rosa26YFP+/+ male and female mice 3–7 months old were used for the behavioral assays. One week prior to testing, animals were single housed. Testing was performed between the hours of 18:00 and 20:00 corresponding to the beginning of the dark cycle for all assays except the maternal aggression assay which took place from 13:00 to 15:00pm corresponding to the light cycle.

Mating

Sexually naïve hormonally primed females were analyzed during female mating. Mating was assessed by placing a mouse of the opposite sex inside the resident’s cage and checking for plugs every 30 min. When a plug was observed, noting successful intromission, female or male mate was removed from the cage and the experimental animal was left inside the cage for an additional 30 min before being sacrificed. Females with no plugs were excluded from the analysis as no mating occurred.

Territorial and maternal aggression

Male territorial aggression was assessed by performing a resident-intruder assay in which an unfamiliar male mouse (‘intruder’) was placed inside the resident’s cage for 10 min. During this period, the homecage male displayed typical aggressive behaviors including attacking and biting. The intruder was removed and then after an additional 50 min the resident male was sacrificed. For the maternal aggression assay, female mice were single-housed after a plug was observed. The following experimental conditions were run: (1) for maternal aggression pups age between P5-P8 were removed and 2 min later a sexually naïve male was introduced into the cage for 10 min, (2) for ‘no pups’ condition, pups were removed but no intruder presented and (3) for naïve control, the cage was undisturbed in which pups were not removed and no intruder presented. In conditions (1) and (2) the female mice were sacrificed after 50 min after pup removal. If female mice did not actively attack the intruder at least two times, they were excluded from the analysis. All male mice attacked the intruder for the 10 min period.

Predator avoidance

Predator avoidance was assessed by introducing a petri dish containing rat bedding to the homecage for 1 hr. The control predator group was presented with a petri dish containing clean mouse bedding (benign). Mice were sacrificed after 1 hr presentation of rat bedding or benign bedding. Mice that did not show escaping responses after the presentation of the petri dish were excluded.

Acknowledgements

We thank V Gallo, J Triplett and I Zohn for constructive input and/or critical reading of the manuscript. We also thank members of the Corbin and Triplett labs for input during the course of this study and Robert McCarter, Director of the DC- IDDRC Biostatistics and Informatics Core for his expert biostatistics assistance. We thank A Pierini for Dbx1cre mice, R Palmiter for Foxp2cre mice sent prior to publication (Rousso et al., 2016), N Shah for AromataseLacZ reporter mice, and F Vaccarino for the OTP antibody. This work was partially supported by NIH grants R01 NIDA020140 (JGC), and R01 DC012050 (JGC). Core support was provided by the CNMC DC-IDDRC Imaging, Biostatistics and Informatics and Animal Neurobehavior Evaluation Cores (NIH IDDRC P30HD040677). JEL is a predoctoral student in the Molecular Medicine Program of the Institute for Biomedical Sciences at The George Washington University. This work is from a dissertation to be presented to the above program in partial fulfillment of the requirements for the Ph.D. degree.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

  • http://dx.doi.org/10.13039/100000026National Institute on Drug Abuse F32 DA035754 to Katie Sokolowski.
  • Goldwin Foundation Grant for Pediatric Epilepsy to Judy Liu.
  • http://dx.doi.org/10.13039/100000026National Institute on Drug Abuse R01 NIDA020140 to Joshua G Corbin.
  • http://dx.doi.org/10.13039/100000055National Institute on Deafness and Other Communication Disorders R01 DC012050 to Joshua G Corbin.

Additional information

Competing interests

The authors declare that no competing interests exist.

Author contributions

JEL, Conceptualization, Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing.

KS, Data curation, Formal analysis, Supervision, Validation, Investigation, Methodology.

PL, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology.

SE, Data curation, Formal analysis, Supervision, Validation, Investigation, Methodology.

YK, Data curation, Formal analysis, Investigation, Methodology.

MG, Data curation, Formal analysis, Supervision, Validation, Investigation, Methodology, Writing—review and editing.

LO, Data curation, Formal analysis, Validation, Investigation, Methodology, Writing—review and editing.

TRH, Data curation, Investigation.

MK, Data curation, Investigation.

DF, Data curation, Investigation.

MH, Conceptualization, Supervision, Methodology.

JL, Supervision, Funding acquisition.

JGC, Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing.

Ethics

Animal experimentation: All animal procedures were approved by the Children's National Medical Center's Institutional Animal Care (Animal Welfare Assurance Number: A3338-01) and Use Committee (IACUC) protocols (#00030435) and conformed to NIH Guidelines for animal use. All surgery was performed under ketamine/xylazine cocktail anesthesia, and every effort was made to minimize suffering.

References

  • Apfelbach R, Blanchard CD, Blanchard RJ, Hayes RA, McGregor IS. The effects of predator odors in mammalian prey species: a review of field and laboratory studies. Neuroscience & Biobehavioral Reviews. 2005;29:1123–1144. doi: 10.1016/j.neubiorev.2005.05.005. [PubMed] [Cross Ref]
  • Ascoli GA, Alonso-Nanclares L, Anderson SA, Barrionuevo G, Benavides-Piccione R, Burkhalter A, Buzsáki G, Cauli B, Defelipe J, Fairén A, Feldmeyer D, Fishell G, Fregnac Y, Freund TF, Gardner D, Gardner EP, Goldberg JH, Helmstaedter M, Hestrin S, Karube F, Kisvárday ZF, Lambolez B, Lewis DA, Marin O, Markram H, Muñoz A, Packer A, Petersen CC, Rockland KS, Rossier J, Rudy B, Somogyi P, Staiger JF, Tamas G, Thomson AM, Toledo-Rodriguez M, Wang Y, West DC, Yuste R, Petilla Interneuron Nomenclature Group Petilla terminology: nomenclature of features of GABAergic interneurons of the cerebral cortex. Nature Reviews Neuroscience. 2008;9:557–568. doi: 10.1038/nrn2402. [PMC free article] [PubMed] [Cross Ref]
  • Bale TL, Epperson CN. Sex differences and stress across the lifespan. Nature Neuroscience. 2015;18:1413–1420. doi: 10.1038/nn.4112. [PMC free article] [PubMed] [Cross Ref]
  • Balthazart J, Charlier TD, Cornil CA, Dickens MJ, Harada N, Konkle AT, Voigt C, Ball GF. Sex differences in brain aromatase activity: genomic and non-genomic controls. Frontiers in Endocrinology. 2011;2:34 doi: 10.3389/fendo.2011.00034. [PMC free article] [PubMed] [Cross Ref]
  • Bergan JF, Ben-Shaul Y, Dulac C. Sex-specific processing of social cues in the medial amygdala. eLife. 2014;3:e02743 doi: 10.7554/eLife.02743. [PMC free article] [PubMed] [Cross Ref]
  • Bian X. Physiological and morphological characterization of GABAergic neurons in the medial amygdala. Brain Research. 2013;1509:8–19. doi: 10.1016/j.brainres.2013.03.012. [PubMed] [Cross Ref]
  • Bikoff JB, Gabitto MI, Rivard AF, Drobac E, Machado TA, Miri A, Brenner-Morton S, Famojure E, Diaz C, Alvarez FJ, Mentis GZ, Jessell TM. Spinal inhibitory interneuron diversity delineates variant motor microcircuits. Cell. 2016;165:207–219. doi: 10.1016/j.cell.2016.01.027. [PMC free article] [PubMed] [Cross Ref]
  • Canteras NS, Pavesi E, Carobrez AP. Olfactory instruction for fear: neural system analysis. Frontiers in Neuroscience. 2015;9:276 doi: 10.3389/fnins.2015.00276. [PMC free article] [PubMed] [Cross Ref]
  • Carney RS, Mangin JM, Hayes L, Mansfield K, Sousa VH, Fishell G, Machold RP, Ahn S, Gallo V, Corbin JG. Sonic hedgehog expressing and responding cells generate neuronal diversity in the medial amygdala. Neural Development. 2010;5:14 doi: 10.1186/1749-8104-5-14. [PMC free article] [PubMed] [Cross Ref]
  • Carvalho VM, Nakahara TS, Cardozo LM, Souza MA, Camargo AP, Trintinalia GZ, Ferraz E, Papes F. Lack of spatial segregation in the representation of pheromones and kairomones in the mouse medial amygdala. Frontiers in Neuroscience. 2015;9:283 doi: 10.3389/fnins.2015.00283. [PMC free article] [PubMed] [Cross Ref]
  • Choi GB, Dong HW, Murphy AJ, Valenzuela DM, Yancopoulos GD, Swanson LW, Anderson DJ. Lhx6 delineates a pathway mediating innate reproductive behaviors from the amygdala to the hypothalamus. Neuron. 2005;46:647–660. doi: 10.1016/j.neuron.2005.04.011. [PubMed] [Cross Ref]
  • Cooke BM, Woolley CS. Sexually dimorphic synaptic organization of the medial amygdala. Journal of Neuroscience. 2005;25:10759–10767. doi: 10.1523/JNEUROSCI.2919-05.2005. [PubMed] [Cross Ref]
  • Dodson PD, Larvin JT, Duffell JM, Garas FN, Doig NM, Kessaris N, Duguid IC, Bogacz R, Butt SJ, Magill PJ. Distinct developmental origins manifest in the specialized encoding of movement by adult neurons of the external globus pallidus. Neuron. 2015;86:501–513. doi: 10.1016/j.neuron.2015.03.007. [PMC free article] [PubMed] [Cross Ref]
  • Dulac C, Wagner S. Genetic analysis of brain circuits underlying pheromone signaling. Annual Review of Genetics. 2006;40:449–467. doi: 10.1146/annurev.genet.39.073003.093937. [PubMed] [Cross Ref]
  • Erskine MS. Mating-induced increases in FOS protein in preoptic area and medial amygdala of cycling female rats. Brain Research Bulletin. 1993;32:447–451. doi: 10.1016/0361-9230(93)90289-N. [PubMed] [Cross Ref]
  • Feng L, Zhao T, Kim J. neuTube 1.0: a new design for efficient neuron reconstruction software based on the SWC format. eNeuro. 2015;2:pii: ENEURO.0049-14.2014 doi: 10.1523/ENEURO.0049-14.2014. [PMC free article] [PubMed] [Cross Ref]
  • Flames N, Pla R, Gelman DM, Rubenstein JL, Puelles L, Marín O. Delineation of multiple subpallial progenitor domains by the combinatorial expression of transcriptional codes. Journal of Neuroscience. 2007;27:9682–9695. doi: 10.1523/JNEUROSCI.2750-07.2007. [PMC free article] [PubMed] [Cross Ref]
  • Fogarty M, Grist M, Gelman D, Marín O, Pachnis V, Kessaris N. Spatial genetic patterning of the embryonic neuroepithelium generates GABAergic interneuron diversity in the adult cortex. Journal of Neuroscience. 2007;27:10935–10946. doi: 10.1523/JNEUROSCI.1629-07.2007. [PubMed] [Cross Ref]
  • French CA, Fisher SE. What can mice tell us about Foxp2 function? Current Opinion in Neurobiology. 2014;28:72–79. doi: 10.1016/j.conb.2014.07.003. [PubMed] [Cross Ref]
  • García-Moreno F, Pedraza M, Di Giovannantonio LG, Di Salvio M, López-Mascaraque L, Simeone A, De Carlos JA. A neuronal migratory pathway crossing from diencephalon to telencephalon populates amygdala nuclei. Nature Neuroscience. 2010;13:680–689. doi: 10.1038/nn.2556. [PubMed] [Cross Ref]
  • Gorski JA, Talley T, Qiu M, Puelles L, Rubenstein JL, Jones KR. Cortical excitatory neurons and Glia, but not GABAergic neurons, are produced in the Emx1-expressing lineage. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience. 2002;22:6309–6314. doi: 10.1016/b0-12-227210-2/00147-3. [PubMed] [Cross Ref]
  • Gross CT, Canteras NS. The many paths to fear. Nature Reviews Neuroscience. 2012;13:651–658. doi: 10.1038/nrn3301. [PubMed] [Cross Ref]
  • Haney M, Debold JF, Miczek KA. Maternal aggression in mice and rats towards male and female conspecifics. Aggressive Behavior. 1989;15:443–453. doi: 10.1002/1098-2337(1989)15:6<443::AID-AB2480150605>3.0.CO;2-U. [Cross Ref]
  • Hirata T, Li P, Lanuza GM, Cocas LA, Huntsman MM, Corbin JG. Identification of distinct telencephalic progenitor pools for neuronal diversity in the amygdala. Nature Neuroscience. 2009;12:141–149. doi: 10.1038/nn.2241. [PMC free article] [PubMed] [Cross Ref]
  • Hong W, Kim DW, Anderson DJ. Antagonistic control of social versus repetitive self-grooming behaviors by separable amygdala neuronal subsets. Cell. 2014;158:1348–1361. doi: 10.1016/j.cell.2014.07.049. [PMC free article] [PubMed] [Cross Ref]
  • Johnson RT, Breedlove SM, Jordan CL. Sex differences and laterality in astrocyte number and complexity in the adult rat medial amygdala. The Journal of Comparative Neurology. 2008;511:599–609. doi: 10.1002/cne.21859. [PMC free article] [PubMed] [Cross Ref]
  • Jones EG, Huntley GW, Benson DL. Alpha calcium/calmodulin-dependent protein kinase II selectively expressed in a subpopulation of excitatory neurons in monkey sensory-motor cortex: comparison with GAD-67 expression. Journal of Neuroscience. 1994;14:611–629. [PubMed]
  • Juntti SA, Tollkuhn J, Wu MV, Fraser EJ, Soderborg T, Tan S, Honda S, Harada N, Shah NM. The androgen receptor governs the execution, but not programming, of male sexual and territorial behaviors. Neuron. 2010;66:260–272. doi: 10.1016/j.neuron.2010.03.024. [PMC free article] [PubMed] [Cross Ref]
  • Kang N, McCarthy EA, Cherry JA, Baum MJ. A sex comparison of the anatomy and function of the main olfactory bulb-medial amygdala projection in mice. Neuroscience. 2011;172:196–204. doi: 10.1016/j.neuroscience.2010.11.003. [PMC free article] [PubMed] [Cross Ref]
  • Kepecs A, Fishell G. Interneuron cell types are fit to function. Nature. 2014;505:318–326. doi: 10.1038/nature12983. [PMC free article] [PubMed] [Cross Ref]
  • Keshavarzi S, Sullivan RK, Ianno DJ, Sah P. Functional properties and projections of neurons in the medial amygdala. Journal of Neuroscience. 2014;34:8699–8715. doi: 10.1523/JNEUROSCI.1176-14.2014. [PubMed] [Cross Ref]
  • Kim Y, Venkataraju KU, Pradhan K, Mende C, Taranda J, Turaga SC, Arganda-Carreras I, Ng L, Hawrylycz MJ, Rockland KS, Seung HS, Osten P. Mapping social behavior-induced brain activation at cellular resolution in the mouse. Cell Reports. 2015;10:292–305. doi: 10.1016/j.celrep.2014.12.014. [PMC free article] [PubMed] [Cross Ref]
  • Koenning M, Jackson S, Hay CM, Faux C, Kilpatrick TJ, Willingham M, Emery B. Myelin gene regulatory factor is required for maintenance of myelin and mature oligodendrocyte identity in the adult CNS. Journal of Neuroscience. 2012;32:12528–12542. doi: 10.1523/JNEUROSCI.1069-12.2012. [PMC free article] [PubMed] [Cross Ref]
  • Kollack SS, Newman SW. Mating behavior induces selective expression of fos protein within the chemosensory pathways of the male syrian hamster brain. Neuroscience Letters. 1992;143:223–228. doi: 10.1016/0304-3940(92)90270-H. [PubMed] [Cross Ref]
  • Kondo Y. Lesions of the medial amygdala produce severe impairment of copulatory behavior in sexually inexperienced male rats. Physiology & Behavior. 1992;51:939–943. doi: 10.1016/0031-9384(92)90074-C. [PubMed] [Cross Ref]
  • Llinás RR. The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science. 1988;242:1654–1664. doi: 10.1126/science.3059497. [PubMed] [Cross Ref]
  • Martel KL, Baum MJ. A centrifugal pathway to the mouse accessory olfactory bulb from the medial amygdala conveys gender-specific volatile pheromonal signals. European Journal of Neuroscience. 2009;29:368–376. doi: 10.1111/j.1460-9568.2008.06564.x. [PMC free article] [PubMed] [Cross Ref]
  • McCarthy MM, Arnold AP. Reframing sexual differentiation of the brain. Nature Neuroscience. 2011;14:677–683. doi: 10.1038/nn.2834. [PMC free article] [PubMed] [Cross Ref]
  • Mullen RJ, Buck CR, Smith AM. NeuN. A neuronal specific nuclear protein in vertebrates. Development. 1992;116:201–211. [PubMed]
  • Oren-Suissa M, Bayer EA, Hobert O. Sex-specific pruning of neuronal synapses in Caenorhabditis elegans. Nature. 2016;533:206–211. doi: 10.1038/nature17977. [PMC free article] [PubMed] [Cross Ref]
  • Remedios R, Huilgol D, Saha B, Hari P, Bhatnagar L, Kowalczyk T, Hevner RF, Suda Y, Aizawa S, Ohshima T, Stoykova A, Tole S. A stream of cells migrating from the caudal telencephalon reveals a link between the amygdala and neocortex. Nature Neuroscience. 2007;10:1141–1150. doi: 10.1038/nn1955. [PubMed] [Cross Ref]
  • Rousso DL, Qiao M, Kagan RD, Yamagata M, Palmiter RD, Sanes JR. Two pairs of ON and OFF retinal ganglion cells are defined by intersectional patterns of transcription factor expression. Cell Reports. 2016;15:1930–1944. doi: 10.1016/j.celrep.2016.04.069. [PMC free article] [PubMed] [Cross Ref]
  • Rowe DW, Erskine MS. c-Fos proto-oncogene activity induced by mating in the preoptic area, hypothalamus and amygdala in the female rat: role of afferent input via the pelvic nerve. Brain Research. 1993;621:25–34. doi: 10.1016/0006-8993(93)90294-W. [PubMed] [Cross Ref]
  • Scalia F, Winans SS. The differential projections of the olfactory bulb and accessory olfactory bulb in mammals. The Journal of Comparative Neurology. 1975;161:31–55. doi: 10.1002/cne.901610105. [PubMed] [Cross Ref]
  • Schultz W. Multiple dopamine functions at different time courses. Annual Review of Neuroscience. 2007;30:259–288. doi: 10.1146/annurev.neuro.28.061604.135722. [PubMed] [Cross Ref]
  • Shemesh Y, Forkosh O, Mahn M, Anpilov S, Sztainberg Y, Manashirov S, Shlapobersky T, Elliott E, Tabouy L, Ezra G, Adler ES, Ben-Efraim YJ, Gil S, Kuperman Y, Haramati S, Dine J, Eder M, Deussing JM, Schneidman E, Yizhar O, Chen A. Ucn3 and CRF-R2 in the medial amygdala regulate complex social dynamics. Nature Neuroscience. 2016;19:1489–1496. doi: 10.1038/nn.4346. [PubMed] [Cross Ref]
  • Shirasaki R, Pfaff SL. Transcriptional codes and the control of neuronal identity. Annual Review of Neuroscience. 2002;25:251–281. doi: 10.1146/annurev.neuro.25.112701.142916. [PubMed] [Cross Ref]
  • Sokolowski K, Corbin JG. Wired for behaviors: from development to function of innate limbic system circuitry. Frontiers in Molecular Neuroscience. 2012;5:55 doi: 10.3389/fnmol.2012.00055. [PMC free article] [PubMed] [Cross Ref]
  • Sokolowski K, Esumi S, Hirata T, Kamal Y, Tran T, Lam A, Oboti L, Brighthaupt SC, Zaghlula M, Martinez J, Ghimbovschi S, Knoblach S, Pierani A, Tamamaki N, Shah NM, Jones KS, Corbin JG. Specification of select hypothalamic circuits and innate behaviors by the embryonic patterning gene dbx1. Neuron. 2015;86:403–416. doi: 10.1016/j.neuron.2015.03.022. [PMC free article] [PubMed] [Cross Ref]
  • Soma M, Aizawa H, Ito Y, Maekawa M, Osumi N, Nakahira E, Okamoto H, Tanaka K, Yuasa S. Development of the mouse amygdala as revealed by enhanced green fluorescent protein gene transfer by means of in utero electroporation. The Journal of Comparative Neurology. 2009;513:113–128. doi: 10.1002/cne.21945. [PubMed] [Cross Ref]
  • Stepien AE, Tripodi M, Arber S. Monosynaptic Rabies virus reveals premotor network organization and synaptic specificity of cholinergic partition cells. Neuron. 2010;68:456–472. doi: 10.1016/j.neuron.2010.10.019. [PubMed] [Cross Ref]
  • Stowers L, Cameron P, Keller JA. Ominous odors: olfactory control of instinctive fear and aggression in mice. Current Opinion in Neurobiology. 2013;23:339–345. doi: 10.1016/j.conb.2013.01.007. [PMC free article] [PubMed] [Cross Ref]
  • Swanson LW, Petrovich GD. What is the amygdala? Trends in Neurosciences. 1998;21:323–331. doi: 10.1016/S0166-2236(98)01265-X. [PubMed] [Cross Ref]
  • Takahashi LK, Gladstone CD. Medial amygdaloid lesions and the regulation of sociosexual behavioral patterns across the estrous cycle in female golden hamsters. Behavioral Neuroscience. 1988;102:268–275. doi: 10.1037/0735-7044.102.2.268. [PubMed] [Cross Ref]
  • Tanaka M, Ikeda T, Hayashi S, Iijima N, Amaya F, Hisa Y, Ibata Y. Nitrergic neurons in the medial amygdala project to the hypothalamic paraventricular nucleus of the rat. Brain Research. 1997;777:13–21. doi: 10.1016/S0006-8993(97)00948-7. [PubMed] [Cross Ref]
  • Unger EK, Burke KJ, Yang CF, Bender KJ, Fuller PM, Shah NM. Medial amygdalar aromatase neurons regulate aggression in both sexes. Cell Reports. 2015;10:453–462. doi: 10.1016/j.celrep.2014.12.040. [PMC free article] [PubMed] [Cross Ref]
  • Vochteloo JD, Koolhaas JM. Medial amygdala lesions in male rats reduce aggressive behavior: interference with experience. Physiology & Behavior. 1987;41:99–102. doi: 10.1016/0031-9384(87)90137-5. [PubMed] [Cross Ref]
  • Waclaw RR, Ehrman LA, Pierani A, Campbell K. Developmental origin of the neuronal subtypes that comprise the amygdalar fear circuit in the mouse. Journal of Neuroscience. 2010;30:6944–6953. doi: 10.1523/JNEUROSCI.5772-09.2010. [PMC free article] [PubMed] [Cross Ref]
  • Wang Y, He Z, Zhao C, Li L. Medial amygdala lesions modify aggressive behavior and immediate early gene expression in oxytocin and vasopressin neurons during intermale exposure. Behavioural Brain Research. 2013;245:42–49. doi: 10.1016/j.bbr.2013.02.002. [PubMed] [Cross Ref]
  • Wu MV, Manoli DS, Fraser EJ, Coats JK, Tollkuhn J, Honda S, Harada N, Shah NM. Estrogen masculinizes neural pathways and sex-specific behaviors. Cell. 2009;139:61–72. doi: 10.1016/j.cell.2009.07.036. [PMC free article] [PubMed] [Cross Ref]
  • Yang CF, Chiang MC, Gray DC, Prabhakaran M, Alvarado M, Juntti SA, Unger EK, Wells JA, Shah NM. Sexually dimorphic neurons in the ventromedial hypothalamus govern mating in both sexes and aggression in males. Cell. 2013;153:896–909. doi: 10.1016/j.cell.2013.04.017. [PMC free article] [PubMed] [Cross Ref]
  • Yang CF, Shah NM. Representing sex in the brain, one module at a time. Neuron. 2014;82:261–278. doi: 10.1016/j.neuron.2014.03.029. [PMC free article] [PubMed] [Cross Ref]
  • Zhang L, Doyon WM, Clark JJ, Phillips PE, Dani JA. Controls of tonic and phasic dopamine transmission in the dorsal and ventral striatum. Molecular Pharmacology. 2009;76:396–404. doi: 10.1124/mol.109.056317. [PubMed] [Cross Ref]
  • Zhao T, Szabó N, Ma J, Luo L, Zhou X, Alvarez-Bolado G. Genetic mapping of Foxb1-cell lineage shows migration from caudal diencephalon to telencephalon and lateral hypothalamus. European Journal of Neuroscience. 2008;28:1941–1955. doi: 10.1111/j.1460-9568.2008.06503.x. [PMC free article] [PubMed] [Cross Ref]
2017; 6: e21012.

Decision letter

Carol A Mason, Reviewing editor
Carol A Mason, Columbia University, United States;

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Embryonic lineage predicts medial amygdala neuronal identity and sex-specific responses to innate behavioral cues" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Margaret M McCarthy (Reviewer #1); Dayu Lin (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The reviewers were enthusiastic about your study linking embryonic patterning of the medial amygdala, neuronal identity and amygdala-directed instinctive behaviors, calling it "exciting" and "elegant". The study nicely charts the lineage of two distinct cell types that originate in the POA and then populate the medial amygdala, and shows that these cell populations do not overlap, have distinct electrophysiological profiles, and respond to varying degrees in males versus females in tests of aggressive, sexual and fear behavior.

Essential revisions:

A major concern of reviewers 2 and 3 is that the quantification of the different populations (Dbx1, OTP, Foxp2) and the modes of representation are confusing. For example, 15% overlap of all 3 markers is unclear, especially if, as reviewer 2 says, 85% OTP cells are neither Foxp2+ nor Dbx1 derived, this reviewer asks why this population seems to be ignored in subsequent electrophysiological and behavioral analyses? Please clarify or defend this decision.

Similarly, reviewer 2 appreciated that during mating behavior, the Dbx1 population showed increased c-fos expression in males and females but Foxp2 cells increased their c-fos in males only. Are you able to address which populations account for the increase in c-fos in predator behavior in males, and whether these cells could be the OTP cells?

Reviewer 3 believes that it would be helpful to indicate the percentage of OPT+ and Dbx1+ cells, and the expected overlap by random chance, even though Foxp2+ and OPT+ do seem to be largely non-overlapping as they occupy virtually non-overlapping spatial locations.

Other comments that could be addressed textually:

Reviewer 1 requests being mindful of the concept of sex differences (in c-fos expression), but not terming these sexual dimorphisms, even though reviewer 2 speaks of "demonstrate sexually dimorphic activation of specific lineage-derived populations in the medial amygdala".

This reviewer also called for citing a previous study on the dissonance between estrogen receptor and ligand, along with different expression of aromatase.

The reviewers thought that Figure 8 was too speculative and that Figure 8—figure supplement 1 be used instead.

Reviewer #1:

This is an exciting paper. The authors trace the lineage of two distinct cell types that originate in the POA and then populate the medial amygdala. Both cells are inhibitory but one is derived from the cells that originally expressed the brain development gene Dbx1, while the other expresses the transcription factor Foxp2 throughout life. These cells do not overlap, show distinct electrophysiological profiles and respond to varying degrees in males versus females in tests of aggressive, sexual and fear behavior. The anatomy is exquisite and the execution of the behavioral tests and conclusions reached are well grounded in the data. Only modest changes are needed, mostly in terminology, and a few points for pondering are offered.

1) The authors make the common and frequent mistake of referring to even a modest sex difference as a sexual dimorphism. It may seem like semantics but the term sexual dimorphism should be reserved for those sex differences which involve two (di-) forms (morphs). At this point, what the authors have discovered is a sex difference in c-fos expression, not a sex dimorphism. This is no less interesting. I suggest the authors comb through the manuscript and at each point where they have used the term sexual dimorphism ask if sex difference would serve the same purpose.

2) The authors also report a fascinating difference in the cellular distribution of aromatase, with about half of Dbx- lineage cells expressing it but Foxp2 cells not at all. There were no cell lineage effects observed for AR or ER. This is consistent with the notion that sexual differentiation of the brain is mediated by a sex difference in ligand, i.e. estradiol, and not by the amount or cellular localization of the receptor. The functional significance of having aromatase in some cells that express ER and others that do not is unknown but the authors might consult work on the Japanese quail brain and the analogous POA (POM) where a similar circumstance is found.

Reviewer #2:

Lischinsky et al. present an elegant story in which they demonstrate sexually dimorphic activation of specific lineage-derived populations in the medial amygdala. Their data are clear and approach logical, and the results are exciting. A couple of major concerns are detailed below.

1) The text is difficult to follow in some important places. For example, looking at Figure 1—figure supplement 1, and the imprecise text associated with it ("generally distinct,", subsection “Dbx1 and Foxp2 expression segregates embryonic and postnatal MeA subpopulations”) one does not immediately grasp whether there are three populations (Foxp2, OTP, and Dbx derived) or two. The graph in f is very confusing- the X axis appears to be the fraction of OTP cells that are Dbx1 derived, plotted against% colocalization with Foxp2, which indicates a 15% overlap of all 3 markers but surely there's an easier way to present this.

This becomes important because if 85% OTP cells are neither Foxp2+ nor Dbx1 derived, then why do the authors ignore this population in their subsequent electrophysiological and behavioural analysis? The study seems incomplete given the presence of this population. If the authors provide an explanation for why they do not examine this population further, that would be sufficient, but it needs to be well justified.

2) The examination of which population is active in different behaviors is beautifully done and the most interesting part of the paper. During mating behavior, c-fos increased in the Dbx1 derived population in both males and females, but in the Foxp2 population, c-fos increase was seen only in males. This is a key result. In predator avoidance, however, the results are puzzling. On the one hand, both males and females show an increased c-fos expression in the MeA. Yet when the Dbx1 derived and Foxp2+ population was examined, it showed an increase only in females and not in males. The authors do not examine which cells account for the c-fos increase in males. Could this perhaps be the OTP population (point 1).

3) The schematic in Figure 8A is highly speculative and goes beyond the data. The schematic seeks to imply that Foxp2+ neurons participate in mounting behaviour, which females do not display, which perhaps explains why this population appears silent during mating in females. This is needlessly speculative- such a hypothesis would need to be examined very carefully and I suggest the schematic be eliminated. In Figure 8B, the authors suggest a sexually dimorphic connectivity such that Foxp2+ neurons are inhibited only in females but not in males. There is no evidence in support of this either. I suggest instead that the summary schematic in Figure 8—figure supplement 1 should be used as Figure 8, it helps the reader.

Reviewer #3:

In Lischinsky et.al. study, the authors identified two nearly exclusive populations in the MeA. One is Dbx1 derived as previously described and the other is Foxp2 derived. These two populations are distinctive at both embryonic and adult stage and have differential electrophysiological properties. The authors also investigated the Fos induction pattern after various social behaviors and concluded that the Foxp2 population is involved in only male mating not female mating, while aggression, predator odor induced similar responses in Dbx-1 and Foxp2 derived cells in both males and females.

The study is overall well performed. The finding is novel and interesting. The developmental origin of diverse MeA cells is poorly understood. This paper represents an important addition to our knowledge. However, the cell counting analysis is incomplete and the images could be improved. As a result, some important information might be missed in the current paper.

1) When understanding the overlap between two types of cells, the absolute number is not very informative. The percentage of cells of double labeled cells over single labeled cells provides some understanding of the overlap but it does not address the real question – are the two populations under investigation preferentially overlapped or not? For example, imagine that Dxb1 population accounts for 50% of the total number of cells and the Esr1 accounts for 50% of the total number of cells, if the double labeled cell is 25% of Dbx1 cells, then the Esr1+ cells are preferentially not derived from the Dbx1 cells. However, if Esr1 account for 10% of the total population, then the Esr1 is preferentially derived from the Dbx1 cells. Without knowing the percentage of the Esr1 cells, it is unclear what the 25% overlap actually means. Thus, throughout the paper, the authors should include information regarding the percentage of cells labeled by gene1, the percentage of cells labeled by gene 2 and the percentage of double labeled cells. Whether the two genes are preferentially labeled or not could be calculated as double%/(Gene1% xGene2%). A value over 1 means that the two genes are preferentially overlapped and a value below one will mean the opposite. With multiple animals, it can be calculated whether the preferential (or non-preferential) overlap is statistically significant or not.

2) While the authors analyzed the sexually dimorphic involvement of the Dbx1 and Foxp2 cells in social behaviors, the basic characterization of the Dbx1 and Foxp2 cells in males and females is lacking. Does the percentage or the absolute number of Dbx1 or Foxp2 cells differ between sexes?

3) Regarding the overlapping of the Dbx1/Foxp2 with behavior induced Fos, there are several questions that could be readily addressed with the data. First, is the Dbx1 or Foxp2 cells preferentially/non-preferentially activated by the behavior (e.g. aggression)? For example, in Figure 5G, the number of Dbx-1 cells expressing aggression induced Fos appears to be quite low. It is about 5% of the total Fos cells based on the graph. Although the Dbx1% is not reported, based on Figure 1, it appears to be quite dense, probably over 20% of the total cells. If so, the Dbx1 cells seem to be preferentially not activated during aggression. This should be analyzed and addressed. Second, to address whether the overlap between the Dbx1/Foxp2 and Fos is sexually dimorphic or not, it is better to compare the value of double%/(Gene1% xGene2%) rather than the double%. The double% is less informative as this value will be affected by the Fos% and the Dbx1% or FoxP2%. If Fos% is sexually dimorphic to begin with, it does not mean much if the double% differs between sexes.

4) Given the different number of Fos cells in Foxp2 and Dbx1 populations (e.g. Figure 5G and K), it is likely that the distribution of Fos is biased topographically in the MeA. It will be helpful to present large view images of the Fos induction patterns. This will help the readers to understand potential reasons for overlap/non-overlap.

5) The statistics in the paper should be double checked or the number of animals needs to be increased to provide sufficient statistical power. For example, in Figure 4G, there is a strong trend (if not yet significant) that the co-labeled% in female Dbx-1 population is higher than the rest of the groups. If one perform a t-test between the male and female Esr1 and Dbx1 double labeled%, it should be significant given that there is no overlap in values and the variability is not huge.

6) Lastly, the authors should provide the Figure number in each figure file to facilitate the review process.

Overall, this is a good study with interesting findings. However, the results can be better analyzed to better understand the potential molecular features and behavioral relevance of Dbx1 and FoxP MeA cells.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Embryonic transcription factor expression in mice predicts medial amygdala neuronal identity and sex-specific responses to innate behavioral cues" for further consideration at eLife. Your revised article has been favorably evaluated by a Senior Editor, a Reviewing Editor, and one reviewer. The second reviewer had added comments that were in agreement with the reviewer 1.

You have done an excellent job of responding to the reviews of your revised manuscript with added additional data and new analyses. The study indeed contributes to an understanding of the developmental logic for processing of innate behaviors by different MeA neuronal subtypes. However, there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) You changed the statistical tests in new Figure 4:

You state in response to Reviewer #3 Critique #5, that some groups in Figure 4 looked significantly different, but no such difference was detected by ANOVA, and then you proceeded to conduct 2-tailed t-tests. The reviewers believe that there is no justification for this approach, as one of the recurrent themes of the manuscript is the presence and absence of sex differences. Neither can be concluded when males and females are analyzed separately, only within sex comparisons can be made. In Figure 4G the bars indicate a significant difference between males and females but this panel alone involves at least 3 t-tests, which is an inappropriate use of multiple comparisons without correction. This is true for most of the data in this figure, and perhaps others. The appropriate analysis is an ANOVA with sex as a factor and only when there is either a main effect of sex or a sex by other factor interaction can post-hoc tests be done.

Further, for use of one way ANOVA, the data need to be normally distributed and exclude outliers. One outlier can throw off the statistics and provide a false negative. While pairwise comparison should not be done without finding significance in ANOVA, it is essential to check whether ANOVA is correctly used here.

Given the very small n's in some of these endpoints, which is understandable given the labor intensiveness of the co-localization analyses, it is imperative that correct statistics are conducted. In this instance it is probably better to miss a sex difference or effect than it is to incorrectly claim one based on a small n and inappropriate statistics. We thus encourage you to seek advice from someone well-versed in statistical analysis to help you work through this, report back to us, and revise the manuscript accordingly.

2) Also in relation to the graphical presentation of statistics: in Figure 5, and others, the use of bars above the histograms is confusing as usually an open-ended bar indicates all groups below it are different from each other. There are clearer ways to indicate post-hoc significant differences between 2 groups, but hard to explain in writing: either a combination of letters, or bars with end feet that point to only the 2 groups. You are encouraged to look at other papers with ANOVA's for some ideas.

3) The reviewers appreciate your removal of the mis-used term "lineage" from your title, changed to: "Embryonic transcription factor expression in mice predicts medial amygdala neuronal identity and sex specific responses to innate behavioral cues".

Essential revisions:

A major concern of reviewers 2 and 3 is that the quantification of the different populations (Dbx1, OTP, Foxp2) and the modes of representation are confusing. For example, 15% overlap of all 3 markers is unclear, especially if, as reviewer 2 says, 85% OTP cells are neither Foxp2+ nor Dbx1 derived, this reviewer asks why this population seems to be ignored in subsequent electrophysiological and behavioral analyses? Please clarify or defend this decision.

In previous Figure 1—figure supplement 1, we mislabeled the y-axis, which lead to some confusion. The axes are now correctly labeled (revised Figure 1—figure supplement 1).

Our focus was on Dbx1-derived and Foxp2+ subpopulations primarily due to our observations that both populations mark the POA, the telencephalic niche that we were most interested in studying.

Additionally, the availability of both Dbx1cre and Foxp2cre mice greatly facilitated our technical ability to conduct patch clamp recordings specifically in these two populations. To our knowledge, OTPcre mice have yet to be generated. However, the reviewers raise an interesting and important point that the OTP population may be of interest. To further investigate the identity of these neurons, following the reviewers’ input we conducted a number of new experiments:

1) To examine in greater detail the potential overlap with the OTP+ population and the Dbx1-derived and Foxp2+ populations we conducted a new series of double and triple immuno-labeling at postnatal stages. We found that Dbx1-derived and Foxp2+ populations were generally not overlapping with the OTP+ cells. This new data is shown in new Figure 1—figure supplement 1.

2) We further conducted a series of double immuno-labeling experiments with a series of cell subtype markers to examine the identity of OTP+ cells. Previous published analyses (Garcia-Moreno et al. Nature Neuroscience, 2010) suggested that a subpopulation of OTP neurons may be excitatory. However, this analysis was cursory and done at embryonic stages. Therefore, to assess the identity of this population in more detail, we performed double immuno-staining with OTP and: YFP+ (in Emx1cre;RYFP mice which mark excitatory neurons), CAMKIIα (excitatory neurons), calbindin (inhibitory neurons) and somatostatin (inhibitory neurons). From these new analyses, we found that the majority of OTP+ neurons are inhibitory. This new data is shown in new Figure 2—figure supplement 1.

Similarly, reviewer 2 appreciated that during mating behavior, the Dbx1 population showed increased c-fos expression in males and females but Foxp2 cells increased their c-fos in males only. Are you able to address which populations account for the increase in c-fos in predator behavior in males, and whether these cells could be the OTP cells?

The reviewers also raised an interesting question as to whether the OTP+ population is the missing population that responds to predator odor in males and whether we could conduct an electrophysiological analyses on these cells.

1) As to the best of our knowledge, there is no OTPcre or OTP reporter mouse line currently available which would have greatly facilitated our ability to identify and patch clamp these neurons as OTP+ cells in the adult MeA represents a very small population.

2) To address the behavioral question, we conducted new experiments to evaluate whether the OTP+ population was activated (c-fos+) during male and female predator exposure behavioral assays. We found that while there were significantly more OTP+/c-fos+ cells during male predator odor exposure as compared to control, when comparing the percent of male OTP+ cells that were activated (c-fos+) we did not observe significant differences between benign and predator avoidance conditions. However, when analyzing activation in female brains, we found an increase in the number of OTP+ cells activated as well as an increase in the percent of OTP+ cells activated vs benign. Therefore, OTP+ cells are being activated in the female MeA; nonetheless, the male predator odor activated population remains to be elucidated. This new behavioral data is shown in new Figure 7—figure supplement 1.

Reviewer 3 believes that it would be helpful to indicate the percentage of OPT+ and Dbx1+ cells, and the expected overlap by random chance, even though Foxp2+ and OPT+ do seem to be largely non-overlapping as they occupy virtually non-overlapping spatial locations.

Embryonically, OTP+ cells are segregated from the Dbx1-derived and Foxp2+ populations. This segregation generally persists into adulthood with overlap of the OTP+ population of approximately 3-5% of Dbx1-derived or Foxp2+ cells (new Figure 1—figure supplement 2). Therefore, as there was almost no-colocalization, we did not consider that there is any possibility of co-localization by random chance.

Other comments that could be addressed textually:

Reviewer 1 requests being mindful of the concept of sex differences (in c-fos expression), but not terming these sexual dimorphisms, even though reviewer 2 speaks of "demonstrate sexually dimorphic activation of specific lineage-derived populations in the medial amygdala".

This reviewer also called for citing a previous study on the dissonance between estrogen receptor and ligand, along with different expression of aromatase.

The reviewers thought that Figure 8 was too speculative and that Figure 8—figure supplement 1 be used instead.

We thank the reviewer for pointing out our misuse of the term sexual dimorphism. As suggested, we have replaced it with the term ‘sex-differences’. In addition, as suggested, we have now included a citation of a review of the studies of the Japanese quail brain (Balthazart et al., Front Endocrinology, 2011) and integrated a brief conceptual understanding of this topic into the Discussion section. Furthermore, as suggested, we have removed the overly speculative Figure 8 and replaced it with the previous supplementary summary schematic, which is also now updated to include the new OTP predator odor data.

Reviewer #3:

[…]

1) When understanding the overlap between two types of cells, the absolute number is not very informative. The percentage of cells of double labeled cells over single labeled cells provides some understanding of the overlap but it does not address the real question – are the two populations under investigation preferentially overlapped or not? For example, imagine that Dxb1 population accounts for 50% of the total number of cells and the Esr1 accounts for 50% of the total number of cells, if the double labeled cell is 25% of Dbx1 cells, then the Esr1+ cells are preferentially not derived from the Dbx1 cells. However, if Esr1 account for 10% of the total population, then the Esr1 is preferentially derived from the Dbx1 cells. Without knowing the percentage of the Esr1 cells, it is unclear what the 25% overlap actually means. Thus, throughout the paper, the authors should include information regarding the percentage of cells labeled by gene1, the percentage of cells labeled by gene 2 and the percentage of double labeled cells. Whether the two genes are preferentially labeled or not could be calculated as double%/(Gene1% xGene2%). A value over 1 means that the two genes are preferentially overlapped and a value below one will mean the opposite. With multiple animals, it can be calculated whether the preferential (or non-preferential) overlap is statistically significant or not.

We now provide new data showing the percent contribution of the Dbx1-derived and Foxp2+ cells to each sex steroid pathway marker (ERα, Aromatase, AR) (new Figure 4—figure supplement 2). We have also now conducted the analysis showing the percent contribution of OTP+ population to the percent total of the Dbx1-derived and Foxp2+ populations (revised Figure 1—figure supplement 1).

2) While the authors analyzed the sexually dimorphic involvement of the Dbx1 and Foxp2 cells in social behaviors, the basic characterization of the Dbx1 and Foxp2 cells in males and females is lacking. Does the percentage or the absolute number of Dbx1 or Foxp2 cells differ between sexes?

We have performed non-stereological analysis of the number of Dbx1-derived and Foxp2+ cells in the MePD and MePV and we did not observe any significant changes in the absolute number of cells between the populations. However, a conclusion as to whether these cells are indeed different between sexes would require a much more rigorous stereological analysis, therefore at this stage we would prefer not to comment in the manuscript on any potential differences, or lack of differences.

3) Regarding the overlapping of the Dbx1/Foxp2 with behavior induced Fos, there are several questions that could be readily addressed with the data. First, is the Dbx1 or Foxp2 cells preferentially/non-preferentially activated by the behavior (e.g. aggression)? For example, in Figure 5G, the number of Dbx-1 cells expressing aggression induced Fos appears to be quite low. It is about 5% of the total Fos cells based on the graph. Although the Dbx1% is not reported, based on Figure 1, it appears to be quite dense, probably over 20% of the total cells. If so, the Dbx1 cells seem to be preferentially not activated during aggression. This should be analyzed and addressed. Second, to address whether the overlap between the Dbx1/Foxp2 and Fos is sexually dimorphic or not, it is better to compare the value of double%/(Gene1% xGene2%) rather than the double%. The double% is less informative as this value will be affected by the Fos% and the Dbx1% or Foxp2%. If Fos% is sexually dimorphic to begin with, it does not mean much if the double% differs between sexes.

We have now performed additional analysis by assessing the percent of Dbx1-derived or Foxp2+ cells that were activated (c-fos+) during aggressive, mating and predator avoidance behaviors and comparing to their respective controls. This new analyses is shown in revised Figures 57. This data is supportive of and strengthens our original findings with regard to numbers of cells activated in mating and aggression, but suggests a more complex interpretation with regard to predator odor, results of which are discussed in the revised Discussion.

4) Given the different number of Fos cells in Foxp2 and Dbx1 populations (e.g. Figure 5G and K), it is likely that the distribution of Fos is biased topographically in the MeA. It will be helpful to present large view images of the Fos induction patterns. This will help the readers to understand potential reasons for overlap/non-overlap.

Images of the c-fos induction pattern are now present in new Figure 5—figure supplement 1.

5) The statistics in the paper should be double checked or the number of animals needs to be increased to provide sufficient statistical power. For example, in Figure 4G, there is a strong trend (if not yet significant) that the co-labeled% in female Dbx-1 population is higher than the rest of the groups. If one perform a t-test between the male and female Esr1 and Dbx1 double labeled%, it should be significant given that there is no overlap in values and the variability is not huge.

We previously used ANOVA for the statistical analysis of Figure 4 to account for multiple group comparisons. As per the reviewer’s suggestions, we have performed t-tests instead and changes are now reflected in revised Figure 4 and changes made to the Materials and methods to reflect this revised analysis.

6) Lastly, the authors should provide the Figure number in each figure file to facilitate the review process.

The figure numbers have been included in every figure.

Overall, this is a good study with interesting findings. However, the results can be better analyzed to better understand the potential molecular features and behavioral relevance of Dbx1 and FoxP MeA cells.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

You have done an excellent job of responding to the reviews of your revised manuscript with added additional data and new analyses. The study indeed contributes to an understanding of the developmental logic for processing of innate behaviors by different MeA neuronal subtypes. However, there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) You changed the statistical tests in new Figure 4:

You state in response to Reviewer #3 Critique #5, that some groups in Figure 4 looked significantly different, but no such difference was detected by ANOVA, and then you proceeded to conduct 2-tailed t-tests. The reviewers believe that there is no justification for this approach, as one of the recurrent themes of the manuscript is the presence and absence of sex differences. Neither can be concluded when males and females are analyzed separately, only within sex comparisons can be made. In Figure 4G the bars indicate a significant difference between males and females but this panel alone involves at least 3 t-tests, which is an inappropriate use of multiple comparisons without correction. This is true for most of the data in this figure, and perhaps others. The appropriate analysis is an ANOVA with sex as a factor and only when there is either a main effect of sex or a sex by other factor interaction can post-hoc tests be done.

Further, for use of one way ANOVA, the data need to be normally distributed and exclude outliers. One outlier can throw off the statistics and provide a false negative. While pairwise comparison should not be done without finding significance in ANOVA, it is essential to check whether ANOVA is correctly used here.

Given the very small n's in some of these endpoints, which is understandable given the labor intensiveness of the co-localization analyses, it is imperative that correct statistics are conducted. In this instance it is probably better to miss a sex difference or effect than it is to incorrectly claim one based on a small n and inappropriate statistics. We thus encourage you to seek advice from someone well-versed in statistical analysis to help you work through this, report back to us, and revise the manuscript accordingly.

We have consulted with our in-house director of the DC-IDDRC Biostatistics Core. In our reanalysis, we now first determined whether the data was normally distributed by performing the Shapiro-Wilk test for normality and we also excluded outliers (any data points two standard deviations away from the mean). For Figure 4G, U, we now performed a two-way ANOVA with interaction for sex and neuronal subpopulation for data that was normally distributed. If there was a main effect or an interaction for sex, then we performed multiple comparisons. If data was not normally distributed, we performed a non-parametric test, in this case a quantile regression analysis (Figure 4N). For Figure 4—figure supplement 2, we also checked for normal distribution and performed a two-way ANOVA for normally distributed data (Figure 4—figure supplement 2A, B) or a quantile regression analysis for non-normally distributed data (Figure 4—figure supplement 2C). A detailed description of the models used for the data analysis of Figure 4 and Figure 4—figure supplement 2 can be found in the Materials and methods section. In addition, for Figures 5, ,66 and and77 we also performed the Shapiro-Wilk test to determine normality. For data normally distributed we performed an unpaired t-test or a one-way ANOVA followed by Tukey’s post-hoc and for data that was not normally distributed we performed either the Mann-Whitney test (data with two groups) or the Krustal-Wallis analysis followed by Dunn’s corrections (data with 3 groups). Details of the new statistical analyses are described in the respective revised Figure legends and the Materials and methods sections.

This new analyses revealed the following differences from our previous analyses:

1) Figure 4G, U: We now find that there are no sex-specific differences in the total Dbx1-derived subpopulation expressing estrogen receptor α. Also, we do now find sex-specific differences in the expression of androgen receptor in the Dbx1-derived subpopulation.

2) Figure 4—figure supplement 2C: Related, we now find sex-specific differences in the co-expression of Dbx1-derived cells with androgen receptor, over the total androgen receptor subpopulation, consistent with the above finding in Figure 4U.

3) Figure 5S, Y: no statistically significant difference is now found between one of the controls (“no pups” condition) and maternal aggression (in the previous version this was significant). Importantly, statistically significant differences remain between all naïve controls and experimental conditions in both sexes.

4) Figure 7G: no statistically significant difference is now found between the number of cfos+;Dbx1- derived cells during predator odor exposure compared to benign control. In our previous and current analyses we did not find any significant difference between the percentage of cfos+;Dbx1-derived cells/cfos+ cells (Figure 7M). Thus, our previous conclusion that neither the Dbx1-derived nor Foxp2+ populations were activated during predator odor remains unchanged.

Importantly these newfound differences do not affect our conclusions in any manner.

2) Also in relation to the graphical presentation of statistics: in Figure 5, and others, the use of bars above the histograms is confusing as usually an open-ended bar indicates all groups below it are different from each other. There are clearer ways to indicate post-hoc significant differences between 2 groups, but hard to explain in writing: either a combination of letters, or bars with end feet that point to only the 2 groups. You are encouraged to look at other papers with ANOVA's for some ideas.

We included in all figures that contained three groups bars with end feet to point out clearly which groups showed significance post-hoc.


Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd