Nucleus accumbens subregions: hodological and immunohistochemical study in the domestic chick (Gallus domesticus)
- First Online:
- Cite this article as:
- Bálint, E. & Csillag, A. Cell Tissue Res (2007) 327: 221. doi:10.1007/s00441-006-0295-0
The nucleus accumbens was identified in avian species some time ago. However, the precise localization and extent of this nucleus is still a matter of controversy. We have used immunolabeling against calbindin, neuropeptide Y, and DARPP-32 (dopamine- and adenosine-related phosphoprotein, 32 kDa) for the selective marking of putative accumbens subdivisions and have followed the anterograde transport of biotinylated dextran amine injected to the nucleus tractus solitarii region of 7-day-old domestic chicks. The nucleus accumbens extending between rostrocaudal atlas coordinates A 10.6 and A 8.8 can be subdivided into the core and shell, the core corresponding to the ventromedial and juxtaventricular medial striatum laterodorsal to the bed nucleus of stria terminalis, and the shell representing an arched region situated ventrally and ventrolaterally to the core. Immunoreactivity to both calbindin and neuropeptide Y is more intense in the shell than in the core division. DARPP-32 immunolabeling does not differ in the two divisions but is markedly weaker in the bed nucleus of stria terminalis, enabling the separation of this nucleus from the surrounding accumbens subdivisions. Fibers from the nucleus solitarius predominantly terminate in the shell division, similar to the situation described in mammals. Whereas the suggested core lies entirely within the boundary of the medial striatum, the shell seems partially to overlap the ventral pallidum. We have been unable to subdivide the remaining part of accumbens lying rostral to A 10.6 into a putative shell and core by the methods employed in the present study. This region probably corresponds to the rostral pole of the nucleus accumbens.
KeywordsBrainStriatumPallidumCalcium-binding proteinsNeuropeptide YChick (Hunnia)
The mammalian nucleus accumbens (Ac) is a ventral forebrain structure associated with the striatal complex and plays a crucial role in limbic neural circuits. It is responsible for motivated goal-directed behavior, reward mechanisms, and emotionality (Kelley 1999; Groenewegen and Uylings 2000) and has been envisaged as an interface between the limbic and motor systems (Nauta and Domesick 1976; Mogenson et al. 1980; Kelley 1999; Groenewegen and Uylings 2000; Heimer 2003) based on its input from the limbic forebrain, viz., basolateral amygdala, hippocampal formation, anterior cingulate cortex, and medial prefrontal cortex, and an output to the ventral pallidum (VP; Heimer et al. 1997). Similar landmarks have been reported to characterize the avian Ac, although the precise position of this nucleus has only been established recently (see below). The region designated as the Ac receives input from a variety of evidently limbic forebrain structures: the hippocampus of zebra finch (Székely and Krebs 1996) and pigeon (Atoji et al. 2002; Atoji and Wild 2004), the septum of pigeon (Atoji and Wild 2004), the piriform cortex (Veenman et al. 1995) and caudal VP of pigeon (Medina and Reiner 1997), and parts of the arcopallium of pigeon (Davies et al. 1997) and mallard (Dubbeldam et al. 1997).
In mammals, Ac has three distinct subdivisions. Whereas the rostral aspect of the nucleus is of largely uniform appearance (Zahm and Brog 1992), the caudal two-thirds of the Ac is divided into two subregions: the Ac core (AcC) and Ac shell (AcS; Záborszky et al. 1985; Heimer et al. 1997). The shell subregion has connections with limbic structures such as the VP, ventral tegmental area (VTA), lateral hypothalamus, bed nucleus of the stria terminalis (BST), and periaqueductal gray. The core subregion is similar to the surrounding striatal areas and has strong connections with the structures of the basal ganglia including the nucleus subthalamicus, globus pallidus, and substantia nigra pars compacta (Heimer et al. 1997). The AcS and AcC are histochemically separable by the differential distribution of neurochemical markers, such as immunoreactivity to calbindin (Meredith et al. 1996; Brauer et al. 2000), calretinin, neuropeptide Y (NPY), and tyrosine hydroxylase (TH; Brauer et al. 2000). Overall, the core can be regarded as similar to the caudatoputamen, and the shell as being more closely related to the extended amygdala and other limbic structures.
Earlier studies in mammals have demonstrated the presence of noradrenaline (NA)-containing fibers in ventral striatal structures, including the shell subregion, but not in the core subregion of the Ac (Gaspar et al. 1985; Lindwall and Stenevi 1978; Swanson and Hartman 1975). The primary source of these NA-immunoreactive afferents to the Ac shell is the A2 region of nucleus tractus solitarii (NTS; Delfs et al. 1998).
Recent research has shown that many aspects of basal ganglia organization, including the projections and neurotransmitter organization of the dorsal pallidum and dorsal striatum, are similar in mammals, birds, and reptiles (Reiner et al. 1998a). The reptilian Ac seems similar to that of mammals in terms of location, immunohistochemical features, and connections (Guirado et al. 1999). Chemically heterogeneous, the Ac of reptiles has been found to show different connectivities in its rostromedial or caudolateral parts; these resemble, to some extent, the shell and core of the mammalian Ac, respectively (Guirado et al. 1999).
Since the evolution of basal ganglia in amniotes seems highly conservative, the question arises as to whether the Ac of the domestic chick also contains separable subregions, similarly to the caudal Ac of mammals. We have limited information about the borders of the Ac in the domestic chick. The original concept of the location of the chicken Ac (Kuenzel and Masson 1988) has been revised according to more recent data (Veenman et al. 1995; Mezey and Csillag 2002) suggesting that, at rostral levels, the Ac extends to the entire medioventral part of the formerly charted lobus parolfactorius (LPO). At more caudal levels, the ventral tip of the lateral ventricle is surrounded by the bed nucleus of the stria terminalis pars lateralis (BSTl), rather than the Ac, whereas the Ac proper is shifted laterally and dorsally from the BSTl. A hodological difference has been found between the Ac and medial striatum (MSt) in the chick, i.e., the accumbal neurons project to the VTA, whereas the rest of the MSt projects mainly to the substantia nigra pars compacta (Mezey and Csillag 2002). The latter study has suggested that the avian Ac has no distinct boundary with the MSt, the accumbal neurons extending into the anatomically defined medial MSt and colocalizing with MSt neurons.
An earlier investigation (Roberts et al. 2002) has shown that the Ac of parrots can be divided into two different subregions, representing the AcS and AcS, on the basis of histological markers, viz., staining for TH, calbindin (CB), substance P (SP), and acethylcholine esterase (AChE). In the budgerigar Melopsittacus undulatus, CB neurons are present in the AcS but not within the AcC (Roberts et al. 2002). These patterns are similar to those found in marmosets but differ from the pattern described in rats, humans, and rhesus monkeys (Jongen-Rêlo et al. 1994; Meredith et al. 1996) in which staining for CB is more intense in the core than in the shell subregion. The aim of the present study has been to define the subregions of the Ac in domestic chick, based on hodological and cytochemical markers, and to harmonize these observations with the current views on the location of avian Ac and the presumable homology between mammalian and avian Ac subterritories.
Materials and methods
Fourteen 7-day-old Hunnia broiler hybrid domestic chicks were used for these experiments. Food and water were available ad libitum.
Anterograde pathway tracing
Five chicks were used specifically for these experiments. The animals were deeply anesthetized with a mixture containing 34 mg/kg body weight (b.wt.) ketamine and 7 mg/kg b.wt. xylazine (i.m.) and placed in a Kopf stereotaxic apparatus. Body temperature was maintained at 39–41°C with a heating pad. All surgical coordinates were based on the atlas of Kuenzel and Masson (1988).
Biotinylated dextran amine (0.15 μl 20% BDA in distilled water, Molecular Probes; MW: 10 kDa) was unilaterally injected via a Hamilton micropipette into the NTS (coordinates: 0.45 cm caudal from bregma, 0.08 cm lateral from midline, and 0.7 cm ventral from dura; beak bar: −0.5 cm below horizontal). Ten days after this injection, the animals were further processed using the following protocol.
Perfusion and sectioning
Animals were deeply anaesthetized with a mixture containing 34 mg/kg b.wt. ketamine and 7 mg/kg b.wt. xylazine(i.m.) and transcardially perfused first with 50–100 ml saline followed by 250–300 ml 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). The brains were removed from the skull, postfixed at 4°C in 4% paraformaldehyde in 0.1 M PB overnight, and sectioned at 60 μm on a Leica vibratome in the coronal plane. Sections were stored at 4°C in 0.02% sodium azide in 0.1 M PB until processed for the immunohistochemical procedure.
Histochemical detection of BDA
Sections were pretreated in 1% H2O2, in phosphate-buffered saline (PBS, pH 7.4) for 15 min, rinsed in PBS (4×), incubated in avidin-biotinylated horseradish peroxidase complex (ABC, Vector Laboratories) overnight at 4°C, and then rinsed in PBS (4×) and in TRIS buffer (pH 8.0; 1×). The stain was developed with the nickel-enhanced diaminobenzidine (DAB) reaction. In detail, the sections were incubated in TRIS buffer (pH 8.0) containing 0.015% DAB and 0.25% nickel-ammoniumsulphate for 10 min. After stain development, sections were rinsed in TRIS buffer (pH 8.0) for 10 min and in PBS (3×), mounted on gelatin/chromalum-subbed slides, air-dried, dehydrated through graded alcohols, cleared with xylene, coverslipped with Depex, and viewed under an Olympus BX 51 microscope by employing the image-capturing and processing programs Viewfinder Lite and Studio Lite.
Tissue sections were rinsed in 1% H2O2 (PBS, pH 7.4) for 15 min to suppress nonspecific peroxidase activity present in residual blood, rinsed in PBS containing 0.3% Triton X-100 (4×), placed in blocking serum consisting of 5% normal goat serum (NGS) for 60 min at room temperature, and incubated for 48 h at 4°C in PBS containing the primary antibody. The rabbit polyclonal antibody against NPY (raised, characterized, and donated by T. Görcs; batch no. N-1010A, Quality Sera, Budapest; see Csiffáry et al. 1990) was used at a dilution of 1:16,000. The mouse monoclonal antiserum against calbindin D28K (Swant, code no. 300, lot no. 18 F) was diluted 1:500. The mouse monoclonal antiserum against DARPP-32 (dopamine-related and adenosine-related phosphoprotein, 32 kDa; ammonium-sulfate-fractionated; batch no. C24-5a; generous gift from H.C. Hemmings; see Hemmings et al. 1987) was applied at a dilution of 1:15,000. The sections were washed PBS (4×), incubated for 2 h in PBS containing biotinylated anti-rabbit IgG diluted at 1:100, rinsed in PBS (4×), and incubated in avidin-biotin-horseradish peroxidase complex overnight at 4°C, rinsed in PBS (4×) and TRIS buffer (pH 8.0; 1×). The stain was developed by using the nickel-enhanced DAB reaction, as above. After stain development, the sections were rinsed in TRIS buffer (pH 8.0) for 10 min and in PBS (3×), mounted on gelatin/chromalum-subbed slides, air-dried, dehydrated through graded alcohols, cleared with xylene, coverslipped with Depex, and viewed under an Olympus BX 51 microscope with the programs Viewfinder Lite and Studio Lite, as above.
In control sections, the primary antibodies (to NPY, CB, or DARPP-32) were replaced by pre-immune normal serum. No visible signal was detected when these sections were processed simultaneously with experimental sections through identical incubation steps.
Anterograde pathway tracing from NTS
For better separation of the BSTl, particularly its rostral aspect from the putative AcC, we used immunostaining against DARPP-32. Characteristic for dopaminoceptive neurons, this peptide was abundant in the striatal region. In accordance with previous observations, little immunolabeling to DARPP-32 was found in the rostrobasal juxtraventricular area, highlighting the BSTl distinctly from the surrounding rostral Ac (Fig. 3e). However, no clear separation between the Ac and MSt was observed by using DARPP-32 as marker.
Because of its uniform labeling intensity, the rostral part of Ac did not show subdivisions with any of the markers applied in the present study.
Based on the results of tract tracing and CB and NPY immunohistochemistry, the Ac of the domestic chick can be divided into subregions probably representing the core and shell. The shell subregion is located more ventrally than the core and is larger in its lateral and medial extent than the core subregion. The putative AcS is rich in CB immunoreactivity, whereas staining in the putative AcC is much lighter. This pattern is similar to that found in the budgerigar, Melopsittacus undulatus (Roberts et al. 2002), and marmoset but differs from the staining pattern found in rats and humans (Meredith et al. 1996). NPY labeling in the chick brain is also stronger in the AcS (present study). This pattern is similar to that reported in mammals (Brauer et al. 2000). The AcS subregion is continuous with the surrounding ventral striatal areas (TO and BSTl) and the relevant part of the VP, whereas the CB-impoverished AcC subregion is coextensive with part of the MSt, as distinct from other parts of MSt, which contain denser labeling with CB.
The separation of AcS and AcC in the relevant region of chick brain is further specified by our hodological finding, i.e., the presence of anterogradely labeled fibers in the presumed shell (but not core) subregion, arising from the NTS. A similar fiber distribution and termination of NA-ergic afferents has been reported in mammals (Delfs et al. 1998). A projection from NTS to the Ac has also been proposed in an earlier anterograde tracing study in the pigeon by Arends et al. (1988), although the authors apparently defined the BST1 as the Ac (see revision of terminology below). Although the NA-ergic nature of this particular projection has not been specified in birds, the presence of NA-ergic fibers in the region that we define as the AcS has been reported in quail and chicken brains, following immunostaining against NA or the synthetising enzyme dopamine-beta-hydroxylase (Bailhache and Balthazart 1993; Moons et al. 1995).
As mentioned above, one complication in this system is the precise position of the BSTl. This terminology was first used to designate a region previously defined as the Ac (Reiner et al. 1983), whereas the proposed site of the “genuine” Ac was shifted accordingly. Given the novel data concerning the localization of the Ac (Mezey and Csillag 2002), which probably extends over a larger area than previously thought, some authors have suggested that the Ac may even be coextensive with the BSTl (implied by Roberts et al. 2002). However, other observations do not seem to support this assumption. For example, Da Silva et al. (2003) have reported feeding effects by the injection of AMPA (α-amino-3-hydroxy-5-methylisoxazole-4-proprionic acid) antagonist into ventral striatal areas of the pigeon, but no ingestive response has been found when the agent is injected into the site of the “classic” Ac (now officially renamed the BSTl; Reiner et al. 2004). A useful marker for the position of BSTl is DARPP-32, which strongly labels all striatal (and presumed accumbal) components, but the immunolabeling in the BSTl is weak (Reiner et al. 1998b; Durstewitz et al. 1998). The selective impoverishment of immunoreactivity to DARPP-32 at the site of newly defined BSTl (confirmed also in the present study) argues against a possible colocalization between this nucleus and the Ac.
The parcellation of the Ac-relevant region is further supported by physiological observations. Lesions to the caudal MSt of domestic chicks elicit impulsive choice behavior (Izawa et al. 2003). No such effect is found when the lesions are placed in the rostral MSt. Impulsive choice behavior has been found to be a typical consequence of lesions to the AcC of mammals (Cardinal et al. 2001). The caudal striatal regions (A 10.4–A 9.2 of the chicken atlas by Kuenzel and Masson 1988) consistently affected by lesions eliciting impulsive choice in the study of Izawa et al. (2003) correspond to the area of the proposed AcC (present study; A 10.6–A 8.8 in the same atlas). Indeed, we have been able to separate putative core and shell subdivisions within this rostro-caudal extent of striatal tissue. Conversely, the rostral striatal regions described by Izawa et al. (2003) as having no effect on impulsive behavior (A 11.6–A 11) extend to that part of the proposed Ac that does not show separation by the cytochemical or hodological markers applied in the present study.
Neuronal units relevant to food reward have been demonstrated in the ventral striatum of chicks (Yanagihara et al. 2001; Izawa et al. 2005) in the region designated as the Ac in the present study. Lesions of the ventral striatum elicit impulsive choice behavior based on the anticipated spatial proximity of reward (Aoki et al. 2006). In another experiment, the apomorphine-induced pecking response of pigeons has been found to be more intense when the agent is injected into the peripheral, rather than the central, region of the Ac (Acerbo et al. 2002), although these subdivisions are not based on histological markers and may not be fully equivalent to the AcS and AcC, as defined in the present study.
The possibility of the heterogeneous composition of avian MSt has been raised by several authors. In particular, the area X of songbirds has been found to comprise both striatal and pallidal elements (Farries and Perkel 2002; Farries et al. 2005; Carrillo and Doupe 2004). Carrillo and Doupe (2004) have interpreted this finding in the light of the accumbens problem, inferring that the MSt (and also area X) might contain accumbens-like parts.
In summary, an Ac-relevant region can be defined in the ventrobasal forebrain of the domestic chick. Based upon the data of the present report and previous physiological observations, the Ac region extending between coordinates A 10.6 and A 8.8 can be subdivided into a core and a shell, the core corresponding to the ventromedial and juxtaventricular MSt dorsolateral to BSTl, and the shell representing an arched region situated ventrally and ventrolaterally to the core (Fig. 4). Whereas the suggested core lies entirely within the boundary of MSt, the shell seems to partially overlap the VP, in agreement with the assumption of Da Silva et al. (2003). The latter authors have proposed, on physiological grounds (ingestive responses to N-methyl-D-aspartate and AMPA-kainate antagonists), that an Ac-relevant territory of the pigeon forebrain extends over the ventral MSt (then termed the LPO; see Reiner et al. 2004), TO, and VP. In mammals, similar feeding responses to AMPA-kainate receptor antagonists have been selectively evoked in the accumbens shell but not in other striatal subdivisions (Kelley and Swanson 1997).
The results of the present study can be compared with other existing interpretations regarding the position and subdivisions of avian Ac. In agreement with previous observations, the Ac extends to (but has no distinct histological boundary with) the previously defined MSt (Mezey and Csillag 2002). Veenman et al. (1995) place the Ac of the pigeon just lateral to the BSTl, and this position probably corresponds to the AcC of our study. The authors do not mention the existence of a separate shell, which, in our interpretation, encroaches upon part of the VP. In the parrot study (Roberts et al. 2002), the relative position of suggested core and shell subregions is similar to that described here for the domestic chick. However, Roberts et al. (2002) do not show the BSTl, and their AcS does not seem to overlap with the VP (albeit the position of the VP might be different in the two species).
No shell/core parcellation has been found in the MSt lying rostral to A 10.6 in the present study (Fig. 4). However, this region has been reported to contain an Ac-relevant medioventral part (Veenman et al. 1995; Medina and Reiner 1997); this has also been confirmed on the basis of selective connections with the VTA (Mezey and Csillag 2002). This part of the MSt probably corresponds to the rostral pole of the Ac of mammals (Zahm and Brog 1992). Although its border with the other non-accumbens-relevant part of MSt is “fuzzy”, it can clearly be delineated, on the basis of DARPP-32 immunoreactivity, from the “embedded” rostral BSTl.