The Olfactory Amygdala in Amniotes: An Evo-Devo Approach

Authors

  • Antonio Abellán,

    1. Laboratory of Brain Development and Evolution, Department of Experimental Medicine, Faculty of Medicine, University of Lleida, Institute of Biomedical Research of Lleida (IRBLLEIDA), Lleida, Spain
    Search for more papers by this author
  • Ester Desfilis,

    1. Laboratory of Brain Development and Evolution, Department of Experimental Medicine, Faculty of Medicine, University of Lleida, Institute of Biomedical Research of Lleida (IRBLLEIDA), Lleida, Spain
    Search for more papers by this author
  • Loreta Medina

    Corresponding author
    • Laboratory of Brain Development and Evolution, Department of Experimental Medicine, Faculty of Medicine, University of Lleida, Institute of Biomedical Research of Lleida (IRBLLEIDA), Lleida, Spain
    Search for more papers by this author

Correspondence to: Loreta Medina, Ph.D., Department of Medicina Experimental, Faculty of Medicine, University of Lleida, Edificio Biomedicina I - IRBLLEIDA, Avda. Alcalde Rovira Roure 80, Lleida 25198, Catalonia, Spain. E-mail: loreta.medina@mex.udl.cat

Abstract

In tetrapods, the medial amygdala is a forebrain center that integrates olfactory and/or vomeronasal signals with the endocrine and autonomic systems, playing a key role in different social behaviors. The vomeronasal system has undergone important changes during evolution, which may be behind some interspecies differences in chemosensory-mediated social behavior. These evolutionary changes are associated with variations in vomeronasal-recipient brain structures, including the medial amygdala. Herein, we employed an evolutionary developmental biology approach for trying to understand the function and evolution of the medial amygdala. For that purpose, we reviewed published data on fate mapping in mouse, and the expression of orthologous developmental regulatory genes (Nkx2.1, Lhx6, Shh, Tbr1, Lhx9, Lhx5, Otp, and Pax6) in embryos of mouse, chicken, emydid turtles, and a pipid frog. We also analyzed novel data on Lhx9 and Otp in a lacertid lizard. Based on distinct embryonic origin and genetic profile, at least five neuronal subpopulations exist in the medial amygdala of rodents, expressing either Nkx2.1/Lhx6, Shh, Lhx9, Otp/Lhx5, or Pax6. Each neuronal subpopulation appears involved in different functional pathways. For example, Lhx6 cells are specifically activated by sex pheromones and project to preoptic and hypothalamic centers involved in reproduction. Based on data in nonmammals, at least three of these neuronal subtypes might have been present in the medial amygdala of the amniote common ancestor. During mammalian evolution, the downregulation of Nkx2.1 in the alar hypothalamus may have been a driving force for an increment of the Otp/Lhx5 subpopulation. Anat Rec, 296:1317-1332, 2013. © 2013 Wiley Periodicals, Inc.

Abbreviations used
AAv

anterior amygdala, ventral part

ACo

anterior cortical amygdalar area

AHN

anterior hypothalamic nucleus

AOB

accessory olfactory bulb

AON

anterior olfactory nucleus

APH

parahippocampal area (chicken)

Arc

arcopallium (part of the pallial amygdala in chicken)

AStr

amygdalo-striatal transition area

BAOT

bed nucleus of the accessory olfactory bulb

Bas

basal nucleus (Meynert)

BC

basal amygdalar complex

BL

basolateral amygdalar nucleus

BLA

basolateral amygdalar nucleus, anterior part

BLV

basolateral amygdalar nucleus, ventral part

BM

basomedial amygdalar nucleus

BMA

basomedial amygdalar nucleus, anterior part

BSTia

bed nucleus of the stria terminalis, intra-amygdaloid part

BSTM

bed nucleus of the stria terminalis, medial part

BSTM1

dorsolateral subnucleus of BSTM (chicken)

BSTM2

ventromedial subnucleus of BSTM (chicken)

CDL

dorsolateral corticoid area (chicken)

Ce

central amygdala (central amygdalar nucleus)

Co

cortical amygdalar area (chicken, comparable to ACo of mammals)

CPu

caudate-putamen

CxA

cortico-amygdalar transition area

DC

dorsal cortex (reptile)

DEn

dorsal endopiriform nucleus

EMT

prethalamic (or thalamic) eminence

Ent

entorhinal cortex

Gaba+

γ-aminobutyric acid (GABA)-ergic cells

Glu+

glutamatergic cells

GP

globus pallidus

GPL

globus pallidus, lateral part

GPM

globus pallidus, medial part

Hb

basal hypothalamus

Hip

hippocampal formation, rostral olfacto-recipient part

I

intercalated cell groups of the amygdale

ic

internal capsule

IM

intercalated cell groups of the amygdala, main division

L

lateral amygdalar nucleus

LC

lateral cortex (reptile)

LGE

lateral ganglionic eminence

LOT

nucleus of the lateral olfactory tract

LOT1

nucleus of the lateral olfactory tract, layer 1

MC

medial cortex (reptile)

Me

medial amygdala in nonmammals

MeA

medial amygdala, anterior part

MeAV

medial amygdala, anteroventral part

MePD

medial amygdala, posterodorsal part

MePV

medial amygdala, posteroventral part

MePVc

medial amygdala, posteroventral part, central subdivision

MePVs

medial amygdala, posteroventral part, superficial subdivision

MGEd

medial ganglionic eminence, dorsal part

MGEcvmedial ganglionic eminence, caudoventral part (previously named or included as part of the anterior peduncular area or AEP); MOB

main olfactory bulb

MPN

medial preoptic nucleus

MS

medial septum

NC

nidopallium caudale or caudal nidopallium (chicken)

NCL

caudolateral nidopallium (chicken)

NCx

neocortex

NSph

nucleus sphericus (reptile, it appears comparable to PMCo of mammals)

ot

optic tract

PDVR

posterior dorsal ventricular ridge (reptile)

Pe

periventricular hypothalamic nucleus

Pir

piriform cortex

PLCo

posterolateral cortical amygdalar area

PMCo

posteromedial cortical amygdalar area

PMv

ventral premammillary nucleus

POC

commissural preoptic area (or commissural septo-preoptic area)

POH

preopto-hypothalamic boundary

PTh

prethalamus

PVN

paraventricular hypothalamic nucleus

PVT

paraventricular thalamic nucleus

S

subpallium

Se

septum

sm

stria medullaris

SO

supraoptic nucleus

SPV

supraopto-paraventricular hypothalamic domain

svz

subventricular zone

Th

thalamus

Tu

ofactory tuberculum

VAA

ventral anterior amygdala (reptile, comparable to avian Co and mammalian ACo)

ve

ventricle

VMHdm

ventromedial hypothalamic nucleus, dorsomedial part

VMHvl

ventromedial hypothalamic nucleus, ventrolateral part

VP

ventral pallium

vz

ventricular zone

In most vertebrates, olfaction is essential for behaviors that are necessary for survival and reproduction, such as fear responses, feeding, sexual behavior, social recognition, and aggression/defense. Olfaction is now thought to play an important role for emotions, motivation, and several social behaviors even in animals with a reduced olfactory system, such as birds and primates (including humans), in which this chemical sense was classically considered irrelevant, (Shepherd, 2004; Heymann, 2006; Balthazart and Taziaux, 2009). Not surprisingly, the neural substrates underlying olfactory processing in different vertebrates share some basic features, including: (1) odor/pheromone detection by highly conserved multigene families of olfactory/vomeronasal receptors expressed by sensory neurons located in specific organs of the nasal and/or vomeronasal cavities; (2) projections to specific glomeruli of the olfactory bulbs, where an odor-map is formed; and (3) secondary pathways to several higher-order centers of the telencephalon involved in odor discrimination, learning and memory formation, and control of behavior (Buck, 2000; Mombaerts, 2006; Martínez-García et al., 2007; Martínez-Marcos, 2009; Yoshihara, 2009).

However, important variations in the olfactory epithelium and in the olfactory brain centers and pathways are observed between different species, indicating that several major changes have taken place in evolution. For example, while teleost fishes possess a single type of olfactory organ (the olfactory rosette) for general odor and pheromone detection (Fujita et al., 1991; Dulka, 1993; Yoshihara, 2009), lungfishes, and tetrapods possess two major olfactory organs, the nasal olfactory epithelium and the vomeronasal organ,1 and segregated central pathways to either the main olfactory bulb or the accessory olfactory bulb, respectively (Swaney and Keverne, 2009; González et al., 2010). These two major systems were thought to be involved in detection of different olfactory chemicals and participate in different functions. Thus, the main olfactory system was thought to detect general volatile odors, acting as a general analyzer subserving feeding, fear responses, and identification of conspecifics (Dulka, 1993; Restrepo et al., 2004; Keller et al., 2009). In contrast, the accessory olfactory system was thought to detect sex pheromones and be involved in control of sexual behavior (Dulka, 1993; Restrepo et al., 2004; Keller et al., 2009). The dual hypothesis was reinforced by the finding of segregated projections from each bulb to distinct brain centers (Scalia and Winans, 1975).

In contrast to this classical view, current knowledge indicates that the main olfactory system of mammals, reptiles, amphibians, and possibly birds can also detect some pheromones2 (such as species-specific sex pheromones, including those involved in courtship), while the accessory olfactory system also appears to be involved in odor discrimination of conspecifics, preys, and predators (Graves and Halpern, 1990; Johnston, 1998; Restrepo et al., 2004; Wysocki and Preti, 2004; Baxi et al., 2005; Balthazart and Taziaux, 2009; Keller et al., 2009; Swaney and Keverne, 2009; Woodley, 2010; Shine et al., 2012). Notably, the main olfactory epithelium of rodents and humans, in addition to a majority of ciliated sensory neurons expressing general odorant receptors (OR) of a highly conserved multigene family (1,000 OR genes in the mouse), includes small subpopulations of sensory cells expressing distinct types of receptors able to detect putative pheromones, such as vomeronasal receptors (subtype 1 or VR1) and trace amine-associated receptors3 (Liberles and Buck, 2006; Hagino-Yamagishi, 2008; Munger et al., 2009; Ma, 2010). In contrast to the high conservation of the OR family, vomeronasal receptors (VR1 and VR2 families) vary considerably from species to species (Swaney and Keverne, 2009). Moreover, it appears that most pheromones are able to activate both vomeronasal and main olfactory sensory neurons in mice (Rodriguez and Boehm, 2009). In addition, OR receptors are also present in the vomeronasal epithelium (Wysocki and Preti, 2004; Munger et al., 2009). It now appears that both olfactory systems can detect different pheromone signals (Swaney and Keverne, 2009), and are likely involved in complementary aspects of mate recognition and sexual behavior (Keller et al., 2009).

Interestingly, the vomeronasal system has undergone important modifications in different lineages during tetrapod evolution, and these changes are correlated with modifications in its telencephalic targets. Thus, while the vomeronasal olfactory system and related brain centers are highly developed in some mammals (such as rodents, felines, canids, and bovids), most reptiles (such as scleroglossa lizards and snakes) and amphibians (Martínez-García et al., 1991; Lohman and Smeets, 1993; Schwenk, 1993; Lanuza and Halpern, 1998; Martínez-Marcos et al., 2002; Halpern and Martínez-Marcos, 2003; Moreno and González, 2007; Martínez-Marcos, 2009), this system has undergone regression in other mammals (such as bats, aquatic mammals and primates), other reptiles (such as turtles and crocodiles), and birds (Reiner and Karten, 1985; Eisthen, 1997; Reep et al., 2007; Swaney and Keverne, 2009). Since in tetrapods the vomeronasal system is involved in species-specific communication (Swaney and Keverne, 2009), evolutionary variations in this system may have contributed to some of the species differences in social behavior mediated by chemosensory signals.

One of the main telencephalic targets of secondary olfactory and/or vomeronasal projections is the amygdala, which appears to mediate the effects of odorants on emotion, motivation, and social behavior (Swanson and Petrovivch 1998; Halpern and Martínez-Marcos, 2003; Martínez-García et al., 2007, 2012; Moreno and González, 2007; Moreno et al., 2008; Roth and Laberge, 2011). In mammals, olfactory and vomeronasal projections are partially convergent and overlap in some parts of the amygdala (Halpern and Martínez-Marcos, 2003; Martínez-Marcos, 2009; Kang et al., 2009). Convergence of olfactory and vomerolfactory projections has also been reported in the amygdala of salamanders, suggesting that it may be more common than previously thought (Roth and Laberge, 2011). However, evolutionary variations in the olfactory and vomeronasal systems correlate with important modifications in the amygdala, and we are still far from understanding the organization and evolution of this telencephalic structure. This is essential if we want to decipher the neurobiological basis of the effects of odorants on emotion and social behavior.

The goal of this review is to analyze and try to understand the organization and evolution of the amygdala using an evolutionary developmental biology (evo-devo) approach, by studying the combinatorial expression patterns of developmental regulatory genes. These genes, encoding transcription factors or signaling proteins, are expressed during development in specific and combinatorial patterns that help to delineate the basic progenitor subdivisions of the neural tube as well as their derivatives (reviewed by Puelles and Medina, 2002; Medina, 2007). At least the coding sequences of many of these regulatory genes are highly conserved across different vertebrates, and so are their expression patterns in the brain and spinal cord. Evolutionary variations at the level of the regulatory regions of these genes (including enhancers) or at the level of factors or cofactors that bind to them and regulate their expresion are thought to be behind changes in expression patterns and evolutionary divergence (Carroll et al., 2001; Davidson, 2006). The comparative analysis of these genes is extremely important to: (1) dissect molecularly the different cell subpopulations of the olfactory/vomeronasal amygdala, establishing a correlation between specific embryonic origins, transcription factors, and functional pathways; (2) investigate the presence of the same cell subtypes in different vertebrates by analyzing different transcription factors during development; and (3) detect possible developmental variations that may be behind the evolution of the olfactory/vomeronasal amygdala.

Herein, we will review published data on the development of the amygdala in different tetrapods, including species with different degrees of development of the olfactory/vomeronasal systems. In particular, we will review the expression of highly conserved developmental regulatory genes (such as Emx1, Tbr1, Pax6, Lhx2, Lhx9, Dlx1/2, Nkx2.1, Shh, Lhx5, and Otp) in the embryonic amygdala in different tetrapods (mouse, chicken, turtle, and frog). We will also include novel data obtained in embryos of a lacertid lizard (Psammodromus algirus), a sauropsid with well developed olfactory and vomeronasal systems, and thus a better choice than chicken and turtles for comparison with frogs and rodents aiming to understand how the amygdala was in the amniote common ancestor.

THE OLFACTORY AND VOMERONASAL AMYGDALA IN RODENTS: DIFFERENT TRANSCRIPTION FACTORS DELINEATE DISTINCT NEURONAL SUBPOPULATIONS

Recent data in rodents using double tract-tracing and electron microscopy have shown that olfactory and vomeronasal projections overlap in many parts of the amygdala, but are completely segregated in the posterior cortical amygdala (Fig. 1; Pro-Sistiaga et al., 2007; Martínez-Marcos, 2009; see also previous reviews by Halpern, 1987; Halpern and Martínez-Marcos, 2003). Overlapping projections occur in some parts of the amygdala classically considered purely olfactory (nucleus of the lateral olfactory tract, amygdalo-cortex transition zone, anterior cortical amygdala), and other parts of the amygdala classically considered purely vomeronasal (ventral anterior amygdala, bed nucleus of the accessory olfactory tract (BAOT), anteroventral medial amygdala, and posterodorsal medial amygdala; von Campenhausen and Mori, 2000; Mohedano-Moriano et al., 2007; Pro-Sistiaga et al., 2007; Martínez-Marcos, 2009; Kang et al., 2009, 2011; Thompson et al., 2012). Nevertheless, olfactory terminals appear to predominate in the classical olfactory part of the amygdala, while vomeronasal terminals predominate in the classical vomeronasal amygdala (Pro-Sistiaga et al., 2007). In contrast, the posterolateral cortical amygdalar area only receives olfactory input, while the posteromedial cortical amygdalar area only receives vomeronasal input (Pro-Sistiaga et al., 2007; Martínez-Marcos, 2009). Some data also point to segregation of olfactory and vomenasal projections in other parts of the amygdala or the extended amygdala complex. For example, only vomeronasal projections appear to reach the dorsal anterior amygdala and the medial bed nucleus of the stria terminalis (BSTM, mainly its principal or posteromedial part; von Campenhausen and Mori, 2000; Mohedano-Moriano et al., 2007).

Figure 1.

Diagram showing the projections of the main and accessory olfactory bulbs in a rodent, a lizard, and a pigeon. Rodents and most lizards possess well developed olfactory and vomeronasal systems. In contrast, the pigeon (as other birds) lacks a vomeronasal organ and an accessory olfactory bulb. Many of the projections target some nuclei or areas of the amygdala, a key center for integration of olfactory/vomeronasal signals and the endocrine and autonomic systems, and for control of social behavior. The nuclei and cortical areas of the amygdala are included within a blue box. In mammals, olfactory and vomeronasal projections converge in some amygdalar nuclei. This does not appear to occur in nonmammals, although in reptiles both types of information converge in the medial amygdala and olfactory cortex due to interconnections between olfactorecipient and vomerorecipient nuclei or areas. In nonmammals, the projections of the main olfactory bulb are bilateral and cross to the contralateral hemisphere through the stria medullaris. For abbreviations, see list.

In rodents, the vomeronasal projections include two distinct types of axons: (1) axons from mitral/tufted cells of the rostral (or anterior) accessory olfactory bulb, transmitting pheromonal information from subtype 1 vomeronasal receptors (V1R) related to male-female social interactions, and (2) axons from mitral/tufted cells of the caudal (or posterior) accessory olfactory bulb, transmitting pheromonal information from subtype 2 vomeronasal receptors (V2R) related to male-male social interactions (Kumar et al., 1999; von Campenhausen and Mori, 2000; Mohedano-Moriano et al., 2007). Projections of rostral and caudal parts of the accessory olfactory bulb converge in layer I of the BAOT, medial amygdala, and the posteromedial cortical amygdalar area (Mohedano-Moriano et al., 2007). However, at least in some species such as the rat, it appears that these different types of vomeronasal information segregate at some levels of the anterior amygdala, BAOT, and medial amygdala (Mohedano-Moriano et al., 2007). The dorsal anterior amygdala and the deep layers of both the BAOT and the anteroventral medial amygdala of the rat receive exclusively vomeronasal input from the caudal part of the accessory olfactory bulb, whereas the BSTM receives input only from the rostral part of the accessory olfactory bulb (Mohedano-Moriano et al., 2007). A similar segregation has also been described in the oppossum (Martínez-Marcos and Halpern, 1999), but such a segregation has not been found in the mouse (von Campenhausen and Mori, 2000), suggesting the existence of inter-species variations (Mohedano-Moriano et al., 2007).

Within the amygdala, the medial amygdala is particularly interesting because it integrates olfactory/vomeronasal signals with the neuroendocrine system, playing a critical role in different aspects of social behavior (Canteras et al., 1995; Swanson, 2000). In rodents, mammals that rely highly on olfactory/vomeronasal signals for reproduction and other social interactions, the medial amygdala is very complex and contains at least four different subdivisions: anterodorsal (MeAD), anteroventral (MeAV), posterodorsal (MePD), and posteroventral (MePV) (de Olmos et al., 1985, 2004; Canteras et al., 1995; Swanson, 2000; Martínez-García et al., 2012). All of these subdivisions receive vomeronasal input, which overlap with olfactory input at least in the anteroventral and posterodorsal subdivisions (Simerly, 2004; Mohedano-Moriano et al., 2007; Pro-Sistiaga et al., 2007; Kang et al., 2009, 2011). Based on projections to preoptic and/or hypothalamic centers involved in either defense4 or reproduction,5 and in functional studies of c-fos activation following exposure to predator or opposite sex-related odors, it appears that the anterior and posteroventral parts of the medial amygdala are involved in aggression/defense (Canteras et al., 1995; Dielenberg et al., 2001; McGregor et al., 2004; Fendt et al., 2005), whereas the posterodorsal medial amygdala is involved in neuroendocrine control (Canteras et al., 1995; Simerly, 2004) and sexual behavior (Fernandez-Fewell and Meredith, 1994; Bressler and Baum, 1996; Heeb and Yahr, 1996; Kollack-Walker and Newman, 1997; Swanson, 2000; Simerly, 2004; Choi et al., 2005). However, each major subdivision of the medial amygdala contains additional molecularly distinct subdomains (Remedios et al., 2004; Hertel et al., 2012) and distinct cell subpopulations (for a review see Table 1 in de Olmos et al., 2004), making it very difficult to understand the specific function of each subdomain and neuronal subpopulation.

Interestingly, a recent study of the medial amygdala of adult rat has shown the presence of three distinct subpopulations of neurons, each one expressing a different transcription factor (Lhx6, Lhx9, or Lhx5), having a different set of projections and being involved in a different function (Fig. 2; Choi et al., 2005). Thus, neurons expressing Lhx6 are preferentially located in the MePD, are specifically activated by reproduction-related odors, and project to preoptic and hypothalamic targets involved in reproduction, such as the medial preoptic nucleus (MPN), the ventrolateral part of the ventromedial hypothalamic nucleus (VMHvl), and the ventral premammillary nucleus (PMv; Choi et al., 2005). The MePD projection to these targets is direct and indirect mainly by way of the principal or posteromedial part of the BSTM, which is also preferentially enriched in cells expressing Lhx6 and appears to be targeted by axons of Lhx6 expressing cells of MePD (Choi et al., 2005). In contrast, Lhx5 and Lhx9 were expressed in different neurons subpopulations of MeA and MePV, and the Lhx5 expressing cells of MeA (preferentially located in MeAV) are activated by aggressive encounters with conspecifics and project primarily to hypothalamic targets involved in defense such as the anterior hypothalamic nucleus (AHN), and the dorsomedial part of the ventromedial hypothalamic nucleus (VMHdm; Choi et al., 2005). MePV contained many cells that are specifically activated by predator odors, some of which project to defensive hypothalamic targets (such as VMHdm) while others project to reproductive targets (VMHvl), but none of these cells appear to express Lhx5, Lhx9, or Lhx6 (Choi et al., 2005).

Figure 2.

Diagram representing the different neuronal subpopulations present in the subdivisions of the medial amygdala of rodents, based on their distinct embryonic origin and genetic profile (Lhx6 cells derived from the caudoventral part of the medial ganglionic eminence; Shh cells derived from the commissural preoptic subdivision; Lhx9 cells derived from the ventral pallium; Otp/Lhx5 cells derived from the supraopto-paraventricular domain of the alar hypothalamus; Pax6 cells derived from the prethalamic eminence; see text for references and details). The connections of some of these cells are known, suggesting that each major subtype is involved in specific functional pathways. For example, the neurons expressing Lhx6 of the medial amygdala are specifically activated by sex odors and project to preoptic and hypothalamic nuclei involved in reproduction (Choi et al., 2005). The inhibitory or excitatory nature of the projections is specified. See text for more details. For abbreviations, see list.

In conclusion, the transcription factor Lhx6 appears to delineate a neural pathway from the MePD to the hypothalamus specifically involved in reproduction (activated by sex pheromones), while Lhx5 is expressed in cells of the medial amygdala that appear related to aggressive encounters with conspecifics (Choi et al., 2005). More studies are needed to understand the role of Lhx9 expressing cells of the medial amygdala (perhaps related to habituation in chronic stress, as explained below), and the molecular features of medial amygdala neurons activated by predator odors.

Development of the olfactory/vomeronasal amygdala in rodents: from specific embryonic divisions and distinct gene expression profile to specific cell types and functional pathways

Recent developmental studies based on gene expression patterns and fate mapping in mice have provided evidence indicating that different neuron subpopulations of the amygdala originate in molecularly distinct embryonic domains of the forebrain (Fig. 3; García-López et al., 2008; Hirata et al., 2009; Abellán et al., 2010a,b; Carney et al., 2010; García-Moreno et al., 2010; Waclaw et al., 2010; Bupesh et al., 2011a,b; reviewed by Medina et al., 2011; and Sokolowski and Corbin, 2012). In particular, neurons of the amygdala expressing Lhx6 appear to originate in the caudoventral part of the medial ganglionic eminence (MGEcv; García-López et al., 2008; Bupesh et al., 2011b); neurons expressing Lhx9 appear to originate in the ventral pallium (VP; García-López et al., 2008; Bupesh et al., 2011b); and neurons expressing Lhx5 originate in the supraopto-paraventricular domain of the alar hypothalamus (SPV; Abellán et al., 2010a,b; García-Moreno et al., 2010; Bupesh et al., 2011b). As explained below, other medial amygdala cell types originate in the preoptic area (García-López et al., 2008; Hirata et al., 2009; Carney et al., 2010; Bupesh et al., 2011b) or the prethalamic eminence (also known as thalamic eminence) during development (Puelles et al., 2000; Bupesh et al., 2011a). Moreover, these new developmental results together with previous data on neurochemistry and connections point to the existence of multiple and partially overlapped cell corridors in the extended amygdala, each one consisting of neurons that share the same embryonic origin and gene expression profile, that are interconnected, project to similar targets in the hypothalamus or brainstem, and are likely involved in similar functions (Bupesh et al., 2011a,b; Medina et al., 2011).

Figure 3.

A, B: Schematic drawings of frontal sections of the mouse embryonic forebrain at intermediate (A) or caudal (B) levels, showing in distinct colors the major subdivisions that produce neurons for the medial amygdala. The genetic profile of these dictinct neuronal subpopulations is also indicated, as well as their GABAergic or glutamatergic phenotype. C, D: Schematic drawings of details of frontal sections of the forebrain of adult mouse, at the level of the anterior (C) or posterior (D) medial amygdala, representing the most important neuronal subpopulations present in each subdivision (a color code is used for representing their distint embryonic origin). For abbreviations, see list.

Regarding the cell corridor of Lhx6 expressing neurons, these cells originate in MGEcv and populate the medial amygdala (where they are densely concentrated in MePD, and more loosely arranged in MeA), the ventral anterior amygdala, part of the sublenticular extended amygdala and the BSTM (primarily its principal or posteromedial subdivision: BSTMpm; Figs. 3, 4A,D; García-López et al., 2008; Bupesh et al., 2011b). These cells also express Nkx2.1 (as typical in MGE derivatives; Xu et al., 2008; Carney et al., 2010), are GABAergic (Choi et al., 2005; Carney et al., 2010), and most of them contain calbindin (Bupesh et al., 2011b). As noted above, data in adult rat have shown that Lhx6 cells of MePD project to BSTMpm, and are specifically activated by reproduction-related odors (i.e., chemicals present in urine of opposite sex; Fig. 2; Choi et al., 2005). Moreover, Lhx6 cells of both MePD and BSTMpm project to preoptic and hypothalamic targets involved in reproduction control (Choi et al., 2005). Other cells of the same cell corridor may be engaged in similar connections and functions, but more studies are needed to investigate it.

Figure 4.

Details of frontal (A–G, I) or horizontal (H) sections of the mouse embryonic forebrain (E14.5 to E18.5) at the level of the anterior (A, B, G, H) or posterior amygdala (C–F, I) hybridized for Lhx6, Shh, Lhx9, Lhx5, or Pax6. Note the trend for segregation of the expression of different genes in the posterior amygdala, where the majority of the Lhx6-expressing cells locate in the MePD, the majority of the Shh-expressing cells locate in the central part of the MePV, and the majority of the Lhx9-expressing cells locate in the superficial part of the MePV. In contrast, most cell subtypes appear to be intermingled in the anterior medial amygdala, although most Lhx5 cells appear to concentrate in the MeAV subdivision. For abbreviations, see text. Scales: bar in A = 300 µm (applies to A–F, I, H) bar in G = 500 µm.

On the other hand, the Lhx5 expressing neurons originate in the SPV domain of the alar hypothalamus, coexpress the transcription factor Otp and appear to be glutamatergic (Figs. 3, 4G,H; Abellán et al., 2010a,b; García-Moreno et al., 2010; see also Bardet et al., 2008; and Bupesh et al., 2011b). Most of the SPV-derived Otp cells and Lhx5 cells appear to concentrate in the anterior medial amygdala, especially its MeAV part (Abellán et al., 2010a,b; García-Moreno et al., 2010), although some Otp cells also reach posterior amygdalar levels (García-Moreno et al., 2010). SPV-derived Otp cells are also present in the BST, including BSTM and intra-amygdaloid BST (Bardet et al., 2008; García-Moreno et al., 2010). These cells almost completely disappear in Otp-knockout mice, in which the SPV domain is severely affected (Wang and Lufkin, 2000; García-Moreno et al., 2010). The SPV produces vasopressinergic and oxytocinergic neurons for the supraoptic and paraventricular hypothalamic nuclei, and small subpopulations of these cells are also observed in the medial amygdala and/or BST of some rodents (de Vries and Buijs, 1983; Caffe et al, 1987; Wang, 1995; Wang and De Vries, 1995; Wang and Lufkin, 2000; de Olmos et al., 2004). Vasopressin- and oxytocin-expressing cells fail to differentiate in the paraventricular and supraoptic hypothalamic nuclei of Otp-knockout mice, and in these mutants vasopressinergic cells also disappear in the medial extended amygdala (Wang and Lufkin, 2000). Based on this evidence, it has been suggested that at least part of the SPV-derived cells of the medial amygdala/BST contain vasopressin (Bupesh et al., 2011b; Medina et al., 2011). In the mouse, the SPV appears to produce cells expressing the enzyme tyrosine hydroxylase for the medial amygdala, and such cells may co-contain vasopressin (Bupesh et al., 2013). As noted above, in the medial amygdala of adult rat, Lhx5 expressing cells (also containing Otp; García-Moreno et al., 2010) are activated by aggressive encounters with conspecifics and project primarily to hypothalamic targets involved in aggression/defensive behavior, such as AHM and VMHdm (Fig. 2; Choi et al., 2005). It is possible that Lhx5/Otp (SPV-derived) cells of the medial amygdala and BST are interconnected and project to similar hypothalamic targets involved in defense/aggression and/or other aspects of social behavior (see next section).

However, Lhx9 expressing cells of the medial amygdala appear to originate in the ventral pallium, as do those of the anterior and posteromedial cortical amygdalar areas and part of the basal amygdalar complex (Figs. 3; 4B,C,E; García-López et al., 2008; Bupesh et al., 2011b). As typical of the pallium, these Lhx9 neurons of the medial amygdala are likely glutamatergic (García-López et al., 2008; Bupesh et al., 2011b). In the MePV, these Lhx9 expressing cells show a trend to locate at the surface, while Lhx9 cells are scattered in MeA. The connections and function of Lhx9 cells of the medial amygdala are unclear, since they were not overactivated by sex- or predator-related odors, and were not retrogradely labeled following tract-tracer injections in defense or reproduction-related nuclei of the hypothalamus (Choi et al., 2005). Nevertheless, Lhx9-related fibers were observed in defensive nuclei of the hypothalamus, AHN, and VMHdm, but these fibers appear to be axons of Lhx9-projection neurons located in the paraventricular thalamic nucleus (PVT; Fig. 2; Choi et al., 2005), a center involved in habituation to prolonged exposure to stressors (Bhatnagar et al., 2002; Jaferi et al., 2003). Projections from the medial amygdala to the PVT have been described in rats (Pitkänen et al., 2000), and it is possible that Lhx9 delineates a medial amygdala-PVT-hypothalamic pathway related to different aspects of defense and/or other responses under chronic stress or anxiety.

For reproduction, defense/aggression, and other social behaviors, the modulation of both the hypothalamo-hypophyseal system (for hormone release) and the autonomic system is essential. Different parts of the medial amygdala, including MeA, MePD, and MePV, are known to project to the paraventricular hypothalamic nucleus, and this projection appears to target both the parvocellular (neuroendocrine-related) and the autonomic parts (Canteras et al., 1995; Tanaka et al., 1997; Prewitt and Herman, 1998; Simerly, 2004). Through these direct projections, and also indirectly by way of the BST, the medial amygdala may influence hormone release by the adenohypophysis and the activation of the autonomic nervous system (Armstrong, 2004; Simerly, 2004). It is likely that different cell types of the medial amygdala are involved in these responses.

The medial amygdala is also essential for hypothalamo-hypophyseal-adrenal (HHA) responses to predator odors (Masini et al., 2009). As noted above, many cells of the medial amygdala (in particular, in MePV) are overactivated in response to predator odors, but they do not correspond to the cells expressing either Lhx6 nor Lhx5 (Choi et al., 2005). The MePV (and the MeA) also contains cells derived from the preoptic area, which express the signaling protein Sonic hedgehog (Shh; Figs. 3, 4F; García-López et al., 2008; Carney et al., 2010; Bupesh et al., 2011b). These neurons are GABAergic, and many of them are nitrergic (Hirata et al., 2009; Carney et al., 2010), and appear to represent a previously described neuron subtype projecting to the paraventricular hypothalamic nucleus (Fig. 2), perhaps including the neuroendocrine-related part (Tanaka et al., 1997). Therefore, this particular GABAergic/nitrergic cell type may be able to modulate the HHA. However, the predator odor-activated cells of MePV are glutamatergic and project to other hypothalamic targets (such as VMHdm, related to defense; Choi et al., 2005), and therefore cannot be the Shh-expressing cells. Glutamatergic cells of the medial amygdala appear to include three major subpopulations (Figs. 3, 4): Lhx9-expressing cells that derive from the ventral pallium (García-López et al., 2008; Bupesh et al., 2011b), Lhx5/Otp-expressing cells that derive from SPV (Bardet et al., 2008; Abellán et al., 2010a,b; García-Moreno et al., 2010), and Pax6 cells that appear to derive from the prethalamic eminence (Bupesh et al., 2011a). It is possible that the Pax6 cells of the medial amygdala (Fig. 4I) are the glutamatergic neurons reported by Choi et al. (2005) to be activated by predator odors, since they do not appear to colocalize with Lhx5 (Choi et al., 2005), and do not overlap with the preferential position of Lhx9 cells in MePV (primarily at the surface of MePV; Choi et al., 2005; García-López et al., 2008; Bupesh et al., 2011b). Thus, different cell types of the amygdala may be modulating distinct, complementary aspects of the defense/aggressive behavior. This may be similar for reproduction, since this appears to be modulated by a double-gate system, including opposite sex-odor activated Lhx6 cells of MePD, which have inhibitory projections to VMHvl,6 as well as by predator odor activated cells of MePV, which show excitatory projections to the same target (Choi et al., 2005).

Evolution of different neuronal subpopulations of the olfactory and vomeronasal amygdala: comparative study by way of expression of developmental regulatory genes

As noted above, at least five different neuron subpopulations are present in the medial amygdala of rodents based on their distinct embryonic origin and molecular profile (Fig. 3): (1) Nkx2.1/Lhx6 neurons, derived from MGE of the telencephalic subpallium; (2) Shh neurons, derived from the preoptic area of the telencephalic subpallium; (3) Lhx9 neurons, derived from the telencephalic ventral pallium; (4) Otp/Lhx5 neurons, derived from the hypothalamic SPV; and (5) Pax6 neurons apparently derived from the prethalamic eminence. The first two neuronal subtypes are GABAergic, while the other three appear to be glutamatergic. Each neuron subtype additionally express specific markers (neuropeptides, enzymes, etc.), and based on partial data it appears that each neuron subtype shows specific connections and may be involved in specific functions.

Are these different neuron subpopulations present in the medial amygdala in other amniotes? This is an important question if we want to understand the evolution of these different cell subpopulations, and whether they were present in the medial amygdala of the common ancestor of mammals and/or amniotes. As noted in the Introduction, the vomeronasal system has undergone important changes during evolution, which are correlated with variations in the vomeronasal brain centers, such as the medial amygdala, and likely had an impact on aspects of social behavior that these centers control or modulate. Herein, we will try to understand how the medial amygdala was in the common ancestor of amniotes by way of expression of developmental regulatory genes in sauropsids with either reduced or well developed vomeronasal systems. As models of sauropsids with a reduced vomeronasal system we will present data from chicken (including material from our own publications) and the emydid turtles Trachemys scripta (previously named Pseudemys scripta) and Emys orbicularis (published by other groups), and as a sauropsid with a highly developed vomeronasal system we will present some novel data in the lacertid lizard Psammodromus algirus. To understand better how the medial amygdala of the amniote common ancestor was, we will compare these data in amniotes with published information in a pipid frog: the African clawed frog Xenopus laevis.

The expression of highly conserved developmental regulatory genes, such as Emx1, Tbr1, Pax6, Lhx2, Lhx9, Dlx1/2, Nkx2.1, Shh, Lhx5, and Otp, have revealed the existence of the same basic subdivisions in the developing forebrain of amphibians, reptiles, birds, and mammals (Fernandez et al., 1998; Puelles et al., 2000; Bachy et al., 2002, 2001; González et al., 2002a,b; Brox et al., 2003, 2004; Moreno et al., 2004, 2010, 2012; Abellán and Medina, 2009; Abellán et al., 2009, 2010a,b; Domínguez et al., 2010). Importantly, the same progenitor domains that produce neurons for the amygdala in mouse are also present in sauropsids and amphibians: in the telencephalon, there is a pallium including a ventral pallial subdivision, and a subpallium including striatal, pallidal and preoptic subdivisions, comparable to those of mouse; in the alar hypothalamus, there is a SPV subdivision; and in the diencephalon proper, there is a prethalamic eminence located in the dorsal aspect of prosomere 3 (Fernández et al., 1998; Puelles et al., 2000; Bachy et al., 2002, 2001; González et al., 2002a,b; Brox et al., 2003, 2004; Moreno et al., 2004, 2010, 2012; Abellán and Medina, 2009; Abellán et al., 2009, 2010a,b; Domínguez et al., 2010). Below, we will analyze whether in sauropsids these different forebrain subdivisions produce the same neuronal subpopulations that they produce for the medial amygdala in the mouse.

Nkx2.1/Lhx6 Neurons Derived from the Pallidal (MGE) Subdivision of the Subpallium

In the subpallium of mouse embryos, the pallidal subdivision (medial ganglionic eminence or MGE) expresses Nkx2.1 in the ventricular zone (vz), subventricular zone (svz) and mantle (Sussel et al., 1999). Nkx2.1 is upstream of Lhx6, which also starts to be expressed from early stages in the svz and mantle of MGE (Sussel et al., 1999). The caudoventral subdivision of MGE produces Nkx2.1/Lhx6-expressing cells for MeA and MePD (García-López et al., 2008; Carney et al., 2010; Bupesh et al., 2011b). In turtle and chicken, the pallidal subdivision also expresses Nkx2.1 and produces Nkx2.1-expressing cells for the medial amygdala7 (Puelles et al., 2000; Abellán et al., 2009; Moreno et al., 2010). In the chicken, the pallidal subdivision also expresses Lhx6, and it appears that its caudoventral subdivision produces Lhx6 cells for the medial amygdala (Fig. 5B; Abellán and Medina, 2009). However, turtles have a reduced vomeronasal system (Eisthen, 1997; Halpern and Martínez-Marcos, 2003), and birds appear to lack a vomeronasal system (Reiner and Karten, 1985; Eisthen, 1997). In birds, the medial amygdala is quite small; in pigeons and chicken it has been shown to receive input from the main olfactory bulb (Reiner and Karten, 1985; Striedter et al., 1998; Patzke et al., 2011), and may be comparable only to the anterior part of the medial amygdala of mammals, where olfactory and vomeronasal inputs overlap (discussed in Yamamoto et al., 2005; Abellán and Medina, 2009; Medina et al., 2011). This nucleus in chicken, quail, and other birds projects to the BSTM and MPN, and is involved in reproduction (reviewed by Reiner et al., 2004; Balthazart and Taziaux, 2009; Medina et al., 2011). In reptiles, this nucleus is larger and adjacent to the BSTM, and both receive vomeronasal input, similarly to mammals (Martínez-García et al., 1991; Lohman and Smeets, 1993; Lanuza and Halpern, 1998; reviewed by Martínez-García et al., 2007). In contrast to birds and mammals, the medial amygdala of reptiles does not appear to receive direct olfactory input, but only vomeronasal (Martínez-García et al., 1991; Lohman and Smeets, 1993; Lanuza and Halpern, 1998). Nevertheless, convergence of both types of information appears to occur in this reptilian structure, because it receives input from the olfactory cortex (Martínez-Marcos et al., 2002). The reptilian medial amygdala projects to the medial preoptic area and ventromedial hypothalamus, and also appears involved in reproductive behavior (reviewed by Martínez-García et al., 2007). The Nkx2.1/Lhx6-expressing cells located in both the medial amygdala and BSTM of sauropsids are likely involved in such function (as those in the mouse). In the frog, the pallidal subdivision also expresses Nkx2.1 and produces some (rather few) cells for the medial amygdala and many cells for BST (Moreno and González, 2007), and this neuron subpopulation was likely present in the medial amygdala-BST of the amniote ancestor (Medina et al., 2011).

Figure 5.

A–C: Oblique sections of the chicken embryonic forebrain (E12-E16) at the level of the medial amygdala and BSTM, hybridized for Lhx9, Lhx6, or Shh. C′ shows a schematic drawing of the section shown in A. D–F: Frontal (D) or oblique sections (E,F) of the embryonic forebrain (developmental stages 37–40) of the lizard P. algirus, at the level of the medial amygdala and BSTM, hybridized for Lhx9 or Otp. Note that, although the medial amygdala and BSTM are distant nuclei in mammals, they are contiguous in chicken and lizard. In the chicken, the medial amygdala-BSTM shows expression of Lhx6 and Shh. In chicken and lizard, Lhx9 expression is strong in the ventral pallial amygdala, and a few cells expressing Lhx9 are present in the medial amygdala. In lizard, Otp expression is strong in the supraopto-paraventricular domain (SPV) of the alar hypothalamus, but a few Otp-expressing cells appear to extend (arrow) from this domain into the BSTM-medial amygdala. See text for more details. For abbreviations, see list. Scales: bar in A = 1 mm; bar in B = 1 mm; bar in C = 1 mm; bar in D = 300 µm (applies to D–F).

Shh Neurons Derived from the Preoptic Area of the Subpallium

In the subpallium of mouse embryos, the preoptic subdivision is characterized by vz expression of Nkx2.1 and Shh (Flames et al., 2007; García-López et al., 2008). In particular, the commissural preoptic subdivision produces Shh-expresing neurons for the MeA and MePV (García-López et al., 2008; Carney et al., 2010; Bupesh et al., 2011b). In rats, many Shh-expressing cells of MePV are nitrergic (Carney et al., 2010) and appear to project to the parvocellular part of the paraventricular hypothalamic nucleus, thus being able to modulate hormone release by HHA (Tanaka et al., 1997). In chicken, the commissural preoptic area also expresses Nkx2.1 and Shh in the vz, and produces Shh-expressing cells for the medial amygdala (Fig. 5C), where they overlap with Lhx6 cells coming from the pallidal subdivision (Fig. 5B; Abellán and Medina, 2009). In reptiles, there is no data on the expression of Shh in the brain. In the frog X. laevis, the preoptic area also expresses Nkx2.1 and Shh in the vz (van den Akker et al., 2008; Domínguez et al., 2010). However, Shh-expressing cells are not observed in the medial amygdala in these animals (Domínguez et al., 2010), and it is unknown whether such cells are present in the BST. Therefore, in the absence of data in reptiles, it is unclear whether these cells were present in the medial amygdala and BST of the amniote common ancestor (Medina et al., 2011).

Lhx9 Expressing Neurons Derived from the Ventral Pallium

In mouse embryos, the amygdalar derivatives of the ventral pallium are characterized by expression of Tbr1, Lhx2, and Lhx9 (Puelles et al., 2000; Medina et al., 2004; Tole et al., 2005; García-López et al., 2008). These include part of the basal amygdalar complex, and the anterior and posteromedial cortical amygdalar areas. Based on Lhx2/Lhx9 expression and fate mapping, the ventral pallium also produces cells for the MeA and MePV (García-López et al., 2008; Bupesh et al., 2011b). Data in adult rat have shown that the Lhx9 cells of the medial amygdala may project to the paraventricular thalamic nucleus, which in turn projects to defense-related centers of the hypotalamus and appears involved in habituation under chronic stress situations (Choi et al., 2005). In chicken, the caudal part of the ventral pallium and its derivatives (arcopallium, caudal nidopallium, and superficial olfactory areas) also express Tbr1, Lhx2, and Lhx9 (Abellán et al., 2009; Fig. 5A). A few cells from here also appear to invade the medial amygdala (Fig. 5A; Abellán et al., 2009). Recent data shows that a similar situation occurs in embryos of the lacertid lizard P. algirus (Fig. 5D), and in the frog X. laevis (Brox et al., 2004; Moreno et al., 2004; Moreno and González, 2007). It is likely that this cell subpopulation was present in the medial amygdala of the amniote common ancestor.

Otp/Lhx5 Neurons Derived from the Hypothalamic SPV

In the developing mouse, the SPV domain of the alar hypothalamus expresses Otp and Lhx5 and produces Otp/Lhx5-expressing cells for the medial amygdala (MeA and apparently MePV; Bardet et al., 2008; Abellán et al., 2010a,b; García-Moreno et al., 2010). In adult rat, these cells project to defense-related centers of the hypothalamus and appear involved in defensive/aggressive behaviors (Choi et al., 2005). In chicken (Bardet et al., 2008), emydid turtle (Moreno et al., 2010), lacertid lizard (present results), and pipid frog (Bardet et al., 2008; Domínguez et al., 2011), the SPV domain also expresses Otp and appears to produce cells for the medial amygdala-BST (Fig. 5E,F). In the chicken, this domain also expresses Lhx5, and some Lhx5 cells are observed in the medial amygdala, but their origin is unclear, since they may also originate in the prethalamic eminence, which also expresses Lhx5 (Abellán et al., 2010a,b). In any case, it is likely that the Otp cell subpopulation was present in the medial amygdala-BST of the amniote common ancestor (although it is possible that in the ancestor these cells primarily invaded the BST; Medina et al., 2011). Interestingly, the SPV domain may have changed in size during evolution of amniotes due to variations in the expression of Nkx2.1 in the alar hypothalamus (Medina, 2008; van der Akker et al., 2008), and this may have affected the SPV-derived cell subpopulation migrating to the medial amygdala (Medina et al., 2011). This is based on the fact that Nkx2.1, expressed in the alar hypothalamus8 in frogs and turtles (van der Akker et al., 2008; Dominguez et al., 2011; Moreno et al., 2012), appears to be downregulated in this site in the mammalian lineage (van der Akker et al., 2008). In the developing mouse, the SPV is enlarged in size under lack of function of Nkx2.1 (in Nkx2.1-knockout mice), possibly as a consequence of a severe malformation and reduction of the basal hypothalamus (Marín and Rubenstein, 2002), and it is possible that the SPV has undergone a similar enlargement during mammalian evolution following downregulation of Nkx2.1 in the alar hypothalamus (Medina et al., 2011). If, as noted above, the SPV produces vasopressinergic and oxytocinergic cells for the BSTM and/or vasopressinergic cells for the medial amygdala in mammals, an enlargment in this domain may have produced more cells of these types for the medial extended amygdala. In the mammalian brain, vasopressin and oxytocin play important roles in different aspects of social behavior, including social recognition, maternal care and pair-bonding (Hammock and Young, 2006), and vasopressin cells of the medial amygdala and BSTM projecting to the septum have been implicated in monogamous behavior (Wang, 1995; de Vries and Miller, 1998; Wang et al., 1998). Therefore, the downregulation of Nkx2.1 may have acted as a driving force for medial extended amygdalar evolution, and the acquisition of a more sophisticated control of social behavior by the amygdala. However, this suggestion needs be taken with caution because: (1) vasopressin (vasotocin) cells are present in the BSTM of most species (from amphibians to mammals) without apparent interspecies variations regarding their abundance (for example, Caffé et al., 1987; Wang et al., 1997; Hilscher-Conklin et al., 1998; Smeets and González, 2001); and (2) the distribution of AVP-expressing cells in the medial amygdala shows variations within amphibians (Hilscher-Conklin et al., 1998; Smeets and González, 2001) and between different mammalian species (Caffé et al., 1987; Wang et al., 1997; Rood and De Vries, 2011). Instead of the number of cells, the different regulation in the expression of vasopressin and oxytocin receptors appears to be important for the differences in social behavior in mammals (Hammock and Young, 2006), but such data are missing in nonmammals.

Pax6 Neurons Derived from the Prethalamic Eminence

In the embryonic mouse forebrain, the prethalamic eminence appears to produce Pax6-expressing cells for the medial amygdala (reaching the MePV; Bupesh et al., 2011b; also suggested by Puelles et al., 2000). Based on expression of Tbr1 (Puelles et al., 2000), these cells appear to be glutamatergic (Hevner et al., 2001), and may represent the predator-odor activated cells of MePV that project to defense-related centers of the hypothalamus (see above). In the chicken, the prethalamic eminence also expresses Pax6, Tbr1 and vesicular glutamate transporter 2 (VGLUT2, marker of glutamatergic cells) and produces Pax6/Tbr1/VGLUT cells for the caudolateral telencephalon, including the amygdala (Puelles et al., 2000; Abellán and Medina, 2009).9 In reptiles, it is unknown whether the medial amygdala contains Pax6 cells, or cells derived from the prethalamic eminence (Moreno et al., 2010). Expression of Pax6, markers of glutamatergic neurons (such as Tbr1), and Lhx5 also characterize the prethalamic eminence in the frog X. laevis (Moreno et al., 2004; Domínguez et al., 2011), but it is unknown whether in anamniotes this domain produces cells for the medial amygdala. Interestingly, the prethalamic eminence domain is traversed by the stria medullaris on its way to the habenular commissure, and this tract includes olfactory axons that project to the contralateral medial amygdala and olfactory cortex in amphibians (Scalia et al., 1991), reptiles (Reiner and Karten, 1985; Lohman and Smeets, 1993; Lanuza and Halpern, 1998) and birds (Reiner and Karten, 1985; Patzke et al., 2011). However, the stria medullaris, although present, is not used for contralateral olfactory projections in mammals (Eisthen, 1997). Thus, at least in some vertebrates, the prethalamic eminence seems to constitute a permissive pathway for growing olfactory axons in their way to the habenular commissure and the contralateral medial amygdala and other olfactory targets. The prethalamic eminence appears to have lost its permissive features in mammals.

Summary and Conclusions

In all amniotes (mammals, birds, and reptiles), the medial amygdala acts as an interface integrating olfactory/vomeronasal signals with the neuroendocrine and autonomic systems. In rodents, mammals that rely highly on olfaction, the medial amygdala is very complex, containing four major subdivisions. All of them receive vomeronasal input, and there is overlap with olfactory input at least in two subdivisions. Developmental data have shown that it includes at least five distinct neuronal subpopulations that originate in distinct embryonic domains. Each neuron subpopulation is characterized by a specific gene expression profile, and appears to be involved in a specific functional pathway. The neuronal subpopulations are: (1) Nkx2.1/Lhx6 (GABAergic) neurons derived from the caudoventral MGE, which are activated by sex-related odors and are involved in reproduction; (2) Shh (GABAergic) neurons derived from the preoptic area, many of which are nitrergic and appear able to modulate hormone release by the adenohypophysis; (3) Lhx9 (glutamatergic) neurons derived from the ventral pallium, that may be involved in habituation under chronic stress conditions; (4) Otp/Lhx5 (glutamatergic) neurons derived from the hypothalamic SPV, which appear involved in aggression and perhaps other social interactions; (5) Pax6 (glutamatergic) neurons derived from the prethalamic eminence, which may represent a neuron subtype described to be activated by predator odors and involved in defense. During evolution, birds have undergone regression of the olfactory system and appear to lack a vomeronasal system. The avian medial amygdala is very tiny and only receives input from the main olfactory bulb, but as that of mammals it appears involved in reproduction. Data in chicken indicate that the avian medial amygdala contains the five neuron subtypes described in the medial amygdala of mammals. The medial amygdala of reptiles (data in emydid turtles and lacertid lizards) includes at least three subtypes of neurons of those present in mammals and birds: neurons expressing Nkx2.1, Lhx9, or Otp. These three neuron subtypes are also present in the medial amygdala-BST of anuran amphibians, and were likely present in the amniote common ancestor. However, the medial amygdala in amphibians lacks Shh-expressing neurons (with preoptic origin), and it is unclear whether it contains neurons derived from the prethalamic eminence. Data in reptiles are missing. During evolution to mammals, the number of SPV-derived neurons (expressing Otp and Lhx5) in the medial amygdala has notably increased. In the hypothalamus, SPV-derived neurons (including vasopressin and oxytocin cells) play important roles in different aspects of social behavior, and SPV-derived neurons migrating to the telencephalon likely contributed to a more sophisticated role of the medial amygdala in modulation of social behavior.

  1. 1

    In addition, some mammals (such as rodents) possess other olfactory organs, such as the Grueneberg organ (involved n detection of alarm pheromones) and the septal organ (Fuss et al., 2005; Breer et al., 2006; Brechbühl et al., 2008).

  2. 2

    Pheromone is here used as originally defined by Karlson and Luscher (1959), as “substances which are secreted by an individual and received by a second individual of the same species, in which they release a specific reaction, for example a defintive behavior or a developmental process”.

  3. 3

    In addition, some cells of the olfactory epithelium express transient receptor potential chanels (TRPs, in particular TRPM5) involved in the transduction of semiochemicals (Lin et al., 2007; Ma, 2010).

  4. 4

    MePV/MeA hypothalamic targets, involved in aggression/defense: anterior hypothalamic nucleus (AHN), and dorsomedial part of the ventromedial hypothalamic nucleus (VMHdm) (Canteras et al., 1995; Swanson, 2000; Canteras, 2002; Simerly, 2004; Choi et al., 2005).

  5. 5

    MePD preoptic/hypothalamic targets, involved in reproduction: medial preoptic nucleus (MPN), ventrolateral part of the ventromedial hypothalamic nucleus (VMHvl) and ventral premammillary nucleus (PMv) (Swanson, 2000; Canteras, 2002; Simerly, 2004; Choi et al., 2005).

  6. 6

    This projection is direct and indirect by way of the BSTMpm (in the latter case, the double inhibition produce disinhibition of the final hypothalamic target).

  7. 7

    The medial amygdala of sauropsids and amphibians, comparable to that of mammals, was initially identified based on similarity of location in the caudoventral telencephalon and similarity of connections (olfactory/vomeronasal input and output to the medial parts of the preoptic area and hypothalamus, which are involved in reproduction) (reviewed by Martínez-García et al., 2007; Moreno and González, 2007). The initial proposal was recently supported by data on expression of developmental regulatory genes, indicating an identical embryonic origin of at least part of its neurons (Moreno and González, 2007; Abellán and Medina, 2009).

  8. 8

    In the suprachiasmatic domain, which is adjacent and topologically ventral to SPV.

  9. 9

    The prethalamic eminence also expresses Lhx5 in mouse and chicken, and at least in the chicken it appears to produce Lhx5 cells for the medial amygdala (Abellán et al. 2010a,b). Therefore, some of the Lhx5 cells found in the medial amygdala of chicken may originate in this domain, instead of the SPV.

Ancillary