Mapping key neuropeptides involved in the melanocortin system in Atlantic salmon (Salmo salar) brain

Abstract

The melanocortin system is a key regulator of appetite and food intake in vertebrates. This system includes the neuropeptides neuropeptide y (NPY), agouti‐related peptide (AGRP), cocaine‐ and amphetamine‐regulated transcript (CART), and pro‐opiomelanocortin (POMC). An important center for appetite control in mammals is the hypothalamic arcuate nucleus, with neurons that coexpress either the orexigenic NPY/AGRP or the anorexigenic CART/POMC neuropeptides. In ray‐finned fishes, such a center is less characterized. The Atlantic salmon (Salmo salar) has multiple genes of these neuropeptides due to whole‐genome duplication events. To better understand the potential involvement of the melanocortin system in appetite and food intake control, we have mapped the mRNA expression of npy, agrp, cart, and pomc in the brain of Atlantic salmon parr using in situ hybridization. After identifying hypothalamic mRNA expression, we investigated the possible intracellular coexpression of npy/agrp and cart/pomc in the tuberal hypothalamus by fluorescent in situ hybridization. The results showed that the neuropeptides were widely distributed, especially in sensory and neuroendocrine brain regions. In the hypothalamic lateral tuberal nucleus, the putative homolog to the mammalian arcuate nucleus, npya, agrp1, cart2b, and pomca were predominantly localized in distinct neurons; however, some neurons coexpressed cart2b/pomca. This is the first demonstration of coexpression of cart2b/pomca in the tuberal hypothalamus of a teleost. Collectively, our data suggest that the lateral tuberal nucleus is the center for appetite control in salmon, similar to that of mammals. Extrahypothalamic brain regions might also be involved in regulating food intake, including the olfactory bulb, telencephalon, midbrain, and hindbrain.

npy, agrp, cart, and pomc were expressed in feeding‐related regions, including hypothalamus of Atlantic salmon (Salmo salar). In the hypothalamic lateral tuberal nucleus, the putative homolog to the mammalian arcuate nucleus, npya, agrp1, cart2b, and pomca were predominantly localized in distinct neurons; however, some neurons coexpressed cart2b/pomca.

Abbreviations

CC

cerebellar corpus

Ce

nucleus centralis lobi inferioris hypothalami

Cho

optic chiasm

Cp

commissura posterior

D

dorsal telencephalon

Dc

central zone of dorsal telencephalon

Dd

dorsal zone of dorsal telencephalon

Dl‐d

dorsal part of lateral zone of dorsal telencephalon

Dl‐v

ventral part of lateral zone of dorsal telencephalon

Dm

medial zone of dorsal telencephalon (dorsal pallium)

DTN

dorsal tegmental nucleus

ent

nucleus entopeduncularis

EW

Edinger–Westphal nucleus

FLM

medial longitudinal fasciculus

fMth

fiber of Mauthner cell

Ggl

stratum ganglionare (cerebellum)

Gran

stratum granulare (cerebellum)

Hab

habenula

inf

infundibulum

Lcoer

locus coeruleus

Lih

inferior hypothalamic lobe

Mc

layer of mitral cells (bulbi olfactori)

Mcba

tractus mesencephalo‐cerebellaris anterior

Mfb

medial forebrain bundle

mol

stratum moleculare (cerebellum)

NAT

nucleus anterior tuberis

NDILm

medial part of the diffuse nucleus of inferior lobe

nIII

nucleus oculomotorius

NLT

nucleus lateralis tuberis

NLTa

anterior nucleus lateralis tuberis

NLTp

posterior nucleus lateralis tuberis

NLTv

ventral nucleus lateralis tuberis

NLV

nucleus lateralis valvulae

NMFL

nucleus medial longitudinal fasciculus

NMH

nucleus magnocellularis hypothalami

NPP

nucleus posterioris periventricularis

NPT

nucleus posterior tuberis

Nrl

nucleus recessi lateralis

nV

nervi trigemini

nVm

nucleus motorius nervi trigemini

OT

Optic tectum

Pit

pituitary

Ppa

preoptic area—anterior parvocellular preoptic nucleus

Ppp

preoptic area—posterior parvocellular preoptic nucleus

Psp

nucleus pretectal superficialis magnocellularis

Pt

posterior tuberculum

PTN

nucleus posterior tuberis

PVO

paraventricular organ

Retm

formatio reticularis pars medialis

Rets

formatio reticularis pars superior

RF

reticular formation

rl

recessi lateralis

Rpo

recessus preopticus

SAC

stratum album centrale (tecti mesencephali)

SGC

stratum griseum centrale (tecti mesencephali)

SM

stratum marginale (tecti mesencephali)

SO

stratum opticum (tecti mesencephali)

SOC

supraoptic/suprachiasmatic nucleus

SPV

stratum periventriculare (tecti mesencephali)

Stgr

stratum granulare (bulbi olfactori)

SV

saccus vasculosus

Tbc

tractus tecto‐bulbaris cruciatus

Thd

dorsal thalamus

Thv

ventral thalamus

TL

Torus longitudinalis

Tlat

torus lateralis

TLw

white matter region of torus longitudinalis

Tod

tractus opticus dorsalis

Toll

tractus olfactorius lateralis

TS

Torus semicircularis

Valv

Valvula cerebelli

Vd

dorsal nucleus of ventral telencephalon

Ve4

fourth ventricle (rhombencephali)

vHab

ventral habenula

Vl

lateral nucleus of ventral telencephalon

Vv

ventral nucleus of ventral telencephalon

1. INTRODUCTION

In fish, like in other vertebrates, appetite and food intake are controlled by endocrine signals and neuropeptides released in neural pathways in the brain (Comesaña et al., 2018; Rønnestad et al., 2017; Soengas et al., 2018; Volkoff, 2016). The neuronal network receives continuous feedback from peripheral tissues, especially the gastrointestinal tract, liver, and pancreas, where nutrients and endocrine and neuronal signals interact to regulate food intake and energy balance (Rønnestad et al., 2017). Food intake is also controlled by sensory and hedonic inputs, such as liking and wanting, that drive hunger and satiety. These inputs originate from a motivation/reward center with dopaminergic neurons (Palmiter, 2007; Soengas et al., 2018); the neurons assist in the underlying mechanism of food intake through conditioning, chemosensory stimulation from the smell of food (Rossi & Stuber, 2018), or nutrient sensing in the brain (Comesaña et al., 2018).

The melanocortin system is a key player in neuronal appetite control. In mammals, it is well characterized and comprises two major neuronal circuits in the arcuate nucleus of the hypothalamus (Elias et al., 1998; Hahn et al., 1998; Schwartz et al., 2000). These neurons are known to either stimulate (orexigenic) or inhibit (anorexigenic) appetite. Both orexigenic and anorexigenic neurons are competitively interacting with the melanocortin receptors (Nuzzaci et al., 2015). Coexpression of neuropeptide y (NPY) and agouti‐regulated peptide (AGRP) increases orexigenic activity resulting in an anabolic response. In contrast, neurons coexpressing cocaine‐ and amphetamine‐regulated transcript (CART) and pro‐opiomelanocortin (POMC) act as anorexigenic, providing a catabolic response.

The key neuropeptides involved in the melanocortin system described in mammals have also been identified in teleosts (Delgado et al., 2017; Rønnestad et al., 2017; Soengas et al., 2018; Volkoff, 2016; Volkoff et al., 2005). However, teleosts typically possess multiple paralogs of these genes compared to mammals due to the additional teleost‐specific whole‐genome duplication, and some teleost families, like salmonids, have additional copies because of the salmonid‐specific fourth round of whole‐genome duplication event (Allendorf & Thorgaard, 1984; Lien et al., 2016). For most of these paralogs, their effects on appetite and food intake control remain unclear.

NPY is a potent and abundant orexigenic factor in the brain and plays a key role in energy homeostasis and food intake in mammals (Loh et al., 2015), as well as in several teleosts (Volkoff, 2016). Earlier studies on Atlantic salmon, Salmo salar, supported the involvement of npy in food intake control (Murashita et al., 2009; Valen et al., 2011). Recently, three npy paralogs have been identified in the Atlantic salmon, named npya1, npya2, and npyb (Tolås et al., 2021). Tolås et al. (2021) showed that neither of the npy paralogs was significantly affected by feeding status in the hypothalamus, albeit a trend of increased npya2 mRNA expression following 4 days of fasting was observed. AGRP is also a key player in the orexigenic melanocortic pathway (Morton & Schwartz, 2001). In Atlantic salmon, Murashita et al. (2009) identified two Agouti‐like sequences, named agrp1 and agrp2 (also named asip2b, see Braasch and Postlethwait (2011) and NCBI GenBank ¹ ). The orexigenic effect of Atlantic salmon agrp1 seems to be in line with those reported in mammals (Kalananthan, Murashita, et al., 2020). However, agrp2 seems to not be directly involved in appetite control in Atlantic salmon (Kalananthan, Lai, et al., 2020) but may play other functional roles, as demonstrated in the zebrafish, Danio rerio (Shainer et al., 2017, 2019; Zhang et al., 2010).

CART is a neuropeptide involved in several processes in the brain, including appetite control. Mammals have one cart gene that plays an anorexigenic role (Akash et al., 2014). However, there are 10 cart paralogues in Atlantic salmon with varying and differential expressions in different brain regions, and their full physiological function(s) are not fully established (Kalananthan et al., 2021). POMC is a precursor peptide that is post‐translationally cleaved into several peptides with a wide range of functions, including α‐ and β‐melanocyte‐stimulating hormones, adrenocorticotropic hormone, and β‐endorphin (Takahashi & Mizusawa, 2013). Three pomc paralogs (pomca1, pomca2, and pomcb) have been previously identified in Atlantic salmon (Murashita et al., 2011; Valen et al., 2011) and are primarily expressed in the pituitary and hypothalamus (Kalananthan, Lai, et al., 2020).

The topology of central neuropeptides of the melanocortin system has been mapped in the whole brain or in specific brain regions in several teleost species, including catfish Clarias batrachus (Gaikwad et al., 2004; Singru et al., 2008; Subhedar et al., 2011), Indian major carp Cirrhinus cirrhosus (Saha et al., 2015), goldfish Carassius auratus (Cerdá‐Reverter & Peter, 2003; Cerdá‐Reverter, Schiöth, et al., 2003; Kojima et al., 2010; Matsuda et al., 2009), zebrafish (Akash et al., 2014; Forlano & Cone, 2007; Jeong et al., 2018; Kaniganti et al., 2021; Koch et al., 2019; Mukherjee et al., 2012; Shainer et al., 2017, 2019), sea bass Dicentrarchus labrax (Agulleiro et al., 2014; Cerdá‐Reverter et al., 2000), Atlantic cod Gadus morhua (Le et al., 2016), and the African cichlid fish Astatotilapia burtoni (Porter et al., 2017). However, to our knowledge, coexpression of npy/agrp and cart/pomc in the hypothalamus has never been observed in a teleost species. In salmonids, pomc and agrp have been identified in the hypothalamus of rainbow trout Oncorhynchus mykiss (Otero‐Rodino et al., 2019). Npy expression has been documented in the brown trout Salmo trutta fario brain including the dorsal and ventral telencephalon, habenula, periventricular and tuberal hypothalamus, saccus vasculosus, tectum, tegmentum, and the rhombencephalon (Castro et al., 1999). In Atlantic salmon and Gambusia affinis brain, Npy expression was found in the ventral telencephalon, tectum, tegmentum, and rhombencephalon (Garcia‐Fernandez et al., 1992). However, the spatial distribution of these melanocortin neuropeptides has not been fully explored in the whole brain of salmonids. Atlantic salmon is an important aquaculture species and understanding the systems that control appetite and food intake is central to optimize their feeding regimes. Taking into consideration that appetite is controlled by neuronal circuits in the brain, mapping the various neuroendocrine cell clusters in the different brain regions is key to uncover the melanocortin system contribution.

In this study, we have described the mRNA expression of npy, agrp, cart, and pomc genes in the Atlantic salmon parr brain by in situ hybridization (ISH). Next, to identify potential key neural circuits involved in appetite control, we investigated the possible coexpression of putative anorexigenic and orexigenic neuropeptides in the Atlantic salmon tuberal hypothalamus.

2. MATERIALS AND METHODS

2.1. Ethical statement

The Atlantic salmon were obtained from the Industrial and Aquatic Laboratory (Bergen High Technology Center, Norway) which has all the necessary approvals for running trials on fish. Atlantic salmon were reared following the Norwegian Veterinary Authorities’ standard protocols. The fish did not undergo any treatment or handling except for euthanasia; thus, special approval from the food authorities and ethics committee was deemed unnecessary according to Norwegian National legislation via the Norwegian Animal Welfare Act (LOV‐2015‐06‐09‐16‐65) and Regulations on the Use of Animals in Experiments (FOR‐2017‐04‐05‐ 451), as required in the European Union (Directive 2010/63/EU) for animal experiments. All fish used were euthanized with an overdose of MS‐222 (MS‐222™; MSD Animal Health, the Netherlands) on site, before further handling.

2.2. Sampling

For RNA extraction and cloning, the brain and pituitary were dissected from one Atlantic salmon (weight = 900 g, standard length = 38.5 cm), stored in RNAlater (Invitrogen, Carlsbad, USA) at 4°C overnight, and then transferred to −80°C. For ISH, 18 Atlantic salmon parr (weight = 33.7 ± 3.5 g, standard length = 14.1 ± 0.5 cm) were killed with an overdose of 200 mg/L MS‐222. An incision was made mid‐ventral to expose the heart for whole‐animal perfusion fixation with 4% paraformaldehyde (PF) in phosphate‐buffered saline (PBS) pH 7.4 (4% PF). Thereafter, brains were carefully dissected out of the skull and post‐fixed in 4% PF for 48 h, rinsed in 1× PBS, and immersed in 25% sucrose/25% OCT (CellPath, UK) for 24 h as described in Eilertsen et al. (2021). The brains were embedded in 100% OCT and coronal parallel cryosectioned across the entire extent of the brain at 10 $\mu$m using Leica CM 3050s cryostat (Leica Biosystems, Germany) and collected on SuperFrost Ultra Plus glasses (Menzel Glaser, Germany). Sections were dried at 65°C for 30 min and stored at −20°C until analyzed by ISH.

2.3. RNA extraction and cDNA synthesis

Total RNA was isolated from both Atlantic salmon brain and pituitary using TRI reagent (MilliporeSigma, St. Louis, USA) following the manufacturer's instructions, and further treated with TURBO DNA‐free (ThermoFisher Scientific, Indianapolis, USA). First‐strand cDNA was synthesized from 1.5 $\mu$g of DNase‐treated total RNA using oligo(dT)20 primer from SuperScript III First‐Strand Synthesis system for RT‐PCR kit (ThermoFisher Scientific).

2.4. Molecular cloning

Primer design was done in ApE‐A plasmid editor (http://biologylabs.utah.edu/jorgensen/wayned/ape/, RRID: SCR_014266). Primers and product sizes are listed in Table 1. Atlantic salmon npya, npyb, agrp1, agrp2, pomca, pomcb, cart1b, and cart2b amplification was performed with Advantage 2 PCR kit (Clontech, Mountain View, CA, USA) using Advantage SA buffer. PCR amplification was performed using a BIO‐RAD C1000 Touch Thermal Cycler (Bio‐Rad, Germany) with an initial step of 95°C for 3 min, and 34 cycles of 30 s denaturation at 95°C, 30 s annealing at 58–60°C, and 1 min extension at 68°C ending with a final extension at 68°C for 10 min.

TABLE 1.

List of primers used for molecular cloning of genes involved in the melanocortin system in Atlantic salmon

Target gene	Accession number	Probe length (bp)	Primer sequence (5′−3′)
npya	npya1 (NM_001146681)	359	F: GCCTGAGGACAACTTCTATC
	npya2 (XM_014178359)		R: GACACTATTACCACAACGACG
npyb	npyb (XM_045697117 and XM_014184208)	423	F: GCGAGCACAGAACAGTCATTC
			R: GTGGTGTTGTGACAAACAGGC
agrp1	agrp1 (NM_001146677)	612	F: GAAGCGCTTTGTTGCATCAGC
			R: GTACACCCAACGTAACATCCATC
	T3 and T7 primers for agrp1 probe		T3: CATTAACCCTCACTAAAGGGAAGAAGCGCTTTGTTGCATCAGC
			T7: TAATACGACTCACTATAGGGCTATAGGCCCCACCTCATGGA
agrp2/asip2b	agrp2 (NM_001146678)	479	F: GAGCGAGAACATTCTGAGCTG
			R: GTCTAGGTCTTCTTGGGGCAG
cart1a	cart1a (XM_014149393)	482	F: CGTATAAAACCTTGGTCCAGG
			R: CATACAACATTGAGTCATCCCG
cart1b	cart1b1 (XM_014150559)	618	F: CTGTATCTCCATCCCTTCTG
	cart1b2 (XM_014151634)		R: GACAACAAACCCTCCATTAC
cart2a	cart2a (ENSSSAG00000015472)	894	F: ATGGAGAGCTCTAAACTGTGGA
			R: CACAAGCACTTCAACAGAAAGAAG
cart2b	cart2b1 (NM_001146680)	567	F: CGGGACCTTTTGGAGACGAAA
	cart2b2 (XM_014183838)		R: TGGGGTTTGGACAATCTCTCAG
cart3a	cart3a1 (XM_014177116)	585	F: GAACTGCAAATTAGAGAGGGAG
	cart3a2 (NM_001141227)		R: TCAAGACAGTCATACATGCAG
cart3b	cart3b (XM_014127320)	389	F: CATTGGGAAGCTCTGTGAC
			R: GCTGTAAATGCTTTCTGGG
cart4	cart4 (XM_014141614)	811	F: GCCTACAGCTTGTGTCAACC
			R: GACGTACTGGGAAAGTGTTCAT
pomca	pomca1 (NM_001198575)	689	F: GTTCTGACCTCACCGCCAAA
	pomca2 ( NM_001198576 )		R: GAGCTAACTGGCTCTAAGTCCT
pomcb	pomcb (NM_001128604)	624	F: AGGTAGTCCCCAGAACCCTC
			R: CAGTACGGTTCTCCGCTTCTT

Note: Accession numbers from GenBank or Ensembl are provided for the different target genes. In bold are the genes on which the probe synthesis was based.

cart1a, cart3a, cart3b, and cart4 were amplified with Q5 High Fidelity 2X polymerase (New England Biolabs, Ipswich, MA, USA) using the following conditions: 98°C for 30 s; 34 cycles of 98°C for 10 s, 60°C for 20 s, 72°C for 30 s; and a final step at 72°C for 2 min.

PCR amplicons were purified from agarose gel using the MinElute Gel Extraction kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol before being cloned into a StrataClone PCR cloning vector (Agilent Technologies, CA, USA). The products were sequenced at the University of Bergen Sequencing Facility using BigDye protocol (BigDye™ Terminator v3.1, ThermoFisher Scientific). Sequence identity was confirmed in Vector NTI software to ensure 100% sequence identity with the public available sequences.

2.5. Riboprobe

The cloned inserts were used to synthesize antisense and sense (control) digoxigenin (DIG)‐labeled (MilliporeSigma) and fluorescein (FITC)‐labeled (Roche Diagnostics, Germany) riboprobes. Synthesis was carried out as described in Thisse and Thisse (2008) using PCR products with added T3 (5′‐CATTAACCCTCACTAAAGGGAA‐3′) and T7 (5′‐TAATACGACTCACTATAGGG‐3′) on the forward and reverse cloning primers for sense and anti‐sense riboprobes, respectively (Table 1). The one exception was the T7 primer for agrp1 (5′‐TAATACGACTCACTATAGGGCTATAGGCCCCACCTCATGGA‐3′). The RNA probes were precipitated using 4 M LiCl, 1 $\mu$g/$\mu$l tRNA (Roche Diagnostics), and 100% EtOH. When two paralogue genes shared a high sequence identity of the targeted region (> 92% identity level), one template was used in the riboprobe synthesis as the probe will label both genes. Note that for the paralogs of pomca, the overall sequence similarity of the target region was 81%; however, large fragments of the probe will hybridize to both genes (100% identity between 240–543 bp and 613–928 bp).

2.6. In situ hybridization

ISH was carried out using a modified protocol from Sandbakken et al. (2012) and replacing 50% deionized formamide with 4 M urea. In summary, sections were dried at room temperature for 30 min, and then at 65°C for 30 min before being rehydrated using an ethanol series (95%–50%). Afterward, sections were permeabilized with 10 $\mu$g/ml proteinase K (MilliporeSigma), post‐fixed in 4% PF, and treated with 0.25% acetic anhydride (MilliporeSigma) in 0.1 M triethanolamine (MilliporeSigma), ending with dehydration using an ethanol series (50%−100%).

The hybridization was carried out with a DIG‐labeled RNA probe overnight at 65°C. After hybridization, sections were washed and treated with RNase A (0.02 mg/ml, MilliporeSigma) before being incubated with sheep polyclonal anti‐DIG antibody (anti‐digoxigenin‐alkaline phosphatase FAB‐fragment, 1:2000, cat. # 11093274910, Roche Diagnostics, RRID: AB_514497) in 1× blocking solution (MilliporeSigma) overnight at room temperature. The result was visualized using 4‐Nitro blue tetrazolium chloride and 5‐Bromo‐4‐chloro‐3‐indolyl‐phosphate system (NBT/BCIP Ready‐to‐use tablets, MilliporeSigma). Parallel sections were Nissl stained with cresyl fast violet (Chroma‐Gesellschaft, Germany). Sections were rehydrated in an ethanol series (96%–50%), dipped in staining solution (0.35% cresyl violet), differentiated in 70% ethanol, and dehydrated in 100% ethanol (2 × 5 min) ending with clearing in xylene. For all genes, sense probes were applied as a control for nonspecific staining.

2.7. Double labeling fluorescence ISH

To investigate the coexpression of npy and agrp or cart and pomc in the tuberal hypothalamus, fluorescence double labeling ISH was done as described in Eilertsen et al. (2018), and replacing 50% deionized formamide with 4 M urea. DIG‐labeled riboprobes were incubated with sheep polyclonal anti‐DIG antibody (anti‐digoxigenin‐alkaline phosphatase FAB‐fragment, 1:2000, cat. # 11093274910, Roche Diagnostics, RRID: AB_514497) and detected with either Fast Red tablet (Roche Diagnostics) dissolved in 0.1 M Tris‐HCl pH 8.2 and 0.1% Tween‐20 or with 100 mg/ml Fast Blue BB salt (MilliporeSigma) and 100 mg/ml naphthol AS‐MX phosphate (MilliporeSigma) in 0.1 M Tris‐HCl pH 8.2, 50 mM MgCl₂, 0.1 M NaCl, and 0.1% Tween‐20 (MilliporeSigma). A 2% blocking solution (MilliporeSigma) in 2× saline‐sodium citrate buffer was used for blocking the sections, followed by the visualization of FITC‐labeled riboprobes using sheep polyclonal anti‐FITC (anti‐fluorescein conjugated with horseradish peroxidase, Fab fragments, cat. # 1142636910, Roche, RRID: AB_840257) and TSA™ Fluorescein (Akoya Biosciences Marlborough, USA) according to the manufacture's protocol. Sections were mounted with ProLong Glass antifade medium with NucBlue (Invitrogen).

2.8. Microscopy

Whole sections were scanned at 20×/0.8 objective with ZEISS Axio Scan.Z1 Slide scanner (Zeiss, Germany, RRID: SCR_020927) and ZEN software (Zeiss). The setting for NBT/BCIP was in TL brightfield (BF) using Hitachi HV‐F202SCL. Fluorescent sections were scanned with DAPI, AF488 (TSA staining for npya and cart2b), AF568 (FastRed for pomca), and AF647 (FastBlue for agrp1), and using the Hamamatsu Orca Flash imaging device.

Confocal images were acquired by a laser‐scanning confocal microscope (Olympus FV3000, Olympus, Japan, RRID: SCR_017015) with 10× and 60× silicon‐immersion oil objective (UPLSAPO 40XS, Olympus), using DAPI, AF488 (TSA staining for npya and cart2b), AF568 (FastRed for pomca), and AF647 (FastBlue for agrp1). Image stacks from each channel were imported into Fiji (https://fiji.sc/, RRID: SCR_002285; Schindelin et al., 2012) to create z‐projections based on maximum intensity.

Figures were prepared using Adobe Photoshop (version 22.1.1, Adobe Systems, San Jose, CA, RRID: SCR_014199). The background was removed, and brightness and contrast were adjusted if necessary. The rainbow trout, O. mykiss, was used for reference and nomenclature of the brain regions in this study (Folgueira et al., 2004a, 2004b; Meek & Nieuwenhuys, 2014).

3. RESULTS

To map the expression of the neuropeptides involved in the melanocortin system in Atlantic salmon parr brain, npy, agrp, cart, and pomc mRNA were examined across the entire rostrocaudal extent of the brain by ISH in coronal parallel cryosections. A summary of the results is shown in Table 2.

TABLE 2.

Summary of the mRNA expression of npy, agrp, cart, and pomc in the Atlantic salmon brain

Gene	Telencephalic regions	Diencephalic regions	Pituitary	Midbrain	Hindbrain
npya	MC, Vd, Vl	Ppa, SOC, vHab, Thv, Thd, Pt, NAT, NLTa, NLTv, NMH, Lih NLTp/NPT		SPV, SGC, EW*	FLM*, nV*
npyb	Vd, Vl	Thd		FLM*
agrp1		NLTv, NLTp/NPT
agrp2	Dm, Dc, Dl‐v	Thv			nV*, RF*
cart1a				NFLM*, nIII*
cart1b				NFLM*
cart2a	Vv,	Tod, Thd, PTN		SPV, SAC, SO
cart2b	Stgr, Vv, Vd, Vl, Dm, Dl, Dd	Ppp, NAT, NLTa, SPV, TS, Tlat, Ce, NLTv, NMH, Lih, PVO, PRN, NRL, NDILm, SV		Gran, lcoer, Rets
cart3a	Vv, Ent	SOC, Thd, NMH/PVO		TLw, EW*	FLM*, lcoer*, nV*
cart3b		Ppp, Thd		tlat, TS	FLM*, nV*, rets*
cart4		ppp
pomca		NLTa, NLTv, NLTp/NPT	pit
pomcb			pit

^*Cells expressing this gene near the indicated brain region. See Section 3 more details.

3.1. neuropeptide y (npy)

3.1.1. npya

npya was widely distributed throughout the brain (Figure 1). In the olfactory bulb, npya expression was found lateroventrally in the mitral cell layer (mc, Figure 1b(1–2)). A high density of npya is further seen in the lateral nucleus of the ventral telencephalon (Vl). Medial scattered neuronal clusters in the dorsal nucleus of ventral telencephalon (Vd, Figure 1c(1–2)), and scattered cells in the posterior dorsal telencephalon (Figure 1(d1)) also expressed npya. Cells expressing npya were identified in the preoptic region, including ventral to the recessus preopticus (rpo, Figure 1(d2)), and supraoptic/suprachiasmatic nucleus (SOC, Figure 1e(1–2)). npya was expressed in the ventrolateral habenula (vHab) and thalamic regions including the ventral (Thv) and dorsal thalamus (Thd), and the posterior tuberculum (Pt, Figure 1f(1–2) and Figure 1g(1–3)).

FIGURE 1.

npya mRNA expression in Atlantic salmon parr brain. (a) Schematic representation of the brain indicating the position of each transverse section. (b1–l1) Nissl‐staining compared to schematic drawing illustrating npya expression by green dots. (b2–l2, g3–j3) Nissl‐staining and corresponding npya expression along with neuroanatomical structures. (b) npya expression in the mc of the olfactory bulb. (c) npya expression in the Vd, and Vl of the telencephalon. (d) npya expression in dorsal telencephalon and preoptic area—ppa. (e) npya expression in the SOC. (f) npya expression in the vHab and Thv. (g) npya expression in the SPV, Thv, Pt, NAT, and NLTv. (h) npya expression in the SPV, SGC, and NMH. (h) npya expression in the Thd and NMH. (j) npya expression in the dorsal tegmentum near EW. (k) npya expression near FLM. (l) npya expression near nV. Abbreviations; Cho, optic chiasm; D, dorsal telencephalon; EW, Edinger–Westphal nucleus; FLM, fasciculus longitudinalis medialis; Hab, habenula; Lih, lobus inferior hypothalami; mc, layer of mitral cells; NAT, nucleus anterior tuberis; NLT, nucleus lateralis tuberis; NLTv, ventral nucleus lateralis tuberis; NMH, nucleus magnocellularis hypothalami; nV, nervus trigemini; Ppa, preoptic area—anterior parvocellular preoptic nucleus; Pt, posterior tuberculum; Rpo, recessus preopticus; SGC, stratum griseum centrale; SM, stratum marginale; SOC, supraoptic/suprachiasmatic nucleus; SPV, stratum periventriculare of the optic tectum; stgr, stratum granulare (bulbi olfactory); Thd, dorsal thalamus; Thv, ventral thalamus; Ts, torus semicircularis; Valv, valvula cerebelli; Vd, dorsal nucleus of ventral telencephalon; vHab, ventral habenula; Vl, lateral nucleus of ventral telencephalon; Vv, ventral nucleus of ventral telencephalon. Scale bar (if no other indication) = 500 $\mu$m

Hypothalamic npya expression was observed at the pituitary stalk in ventral nucleus lateralis tuberis (NLTv), nucleus anterior tuberis (NAT), and nucleus magnocellularis hypothalami (NMH, Figure 1g(1–3) (h3), and (i3)). In the optic tectum, npya was expressed in the periventricular layer (SPV) and in a few cells in the griseum layer (SGV, Figure 1(h2)). Positive npya expression was found in a cluster of dorsal tegmentum nucleus (DTN), potentially near the Edinger–Westphal nucleus (EW, Figure 1j(1–3)). Two rhombencephalic npya expressions were found, one in a small cell cluster located lateral to the fasciculus longitudinalis medialis (FLM, Figure 1k(1–2)), and a larger cluster located ventral to the Ve4 close to the trigeminal nerve (nV, Figure 1l(1–2)).

3.1.2. npyb

npyb was detected in the telencephalon and diencephalon brain regions (Figure 2). In the telencephalon, npyb mRNA expression was found in the dorsal nucleus of the ventral telencephalon (Vd) toward the telencephalic ventricle in the mid telencephalon (rostral‐caudal direction) about 600 $\mu$m (Figure 2a(1–2)). Light staining for npyb was found in a few cells of the lateral nucleus of ventral telencephalon (Vl, Figure 2a(1–3)). Ventral to the optic tectum, one npyb expressing cell cluster was observed in the ventral thalamus region (Thv) (Figure 2b(1–3)) toward the hypothalamic NMH near the nucleus posterioris periventricularis (NPPv, Figure 2(b3)). A cell cluster expressing npyb was observed just ventral to the cerebellar valvula adjacent to the medial longitudinal fasciculus (FLM, Figure 2c(1–3)).

FIGURE 2.

npyb mRNA expression in Atlantic salmon parr brain from rostral (a) to caudal (c) brain regions. (a1–c1) Nissl‐staining and schematic drawing illustrating npyb expression by purple dots. Lower left corner represents a schematic drawing of the salmon brain indicating the position of the section (a2–c2, a3–c3). Nissl‐staining and corresponding npyb expression with neuroanatomical structures. (a) npyb expression in the Vd, Vv and Vl parts of ventral telencephalon. (b) npyb expression near the Thv and NPPv. (c) npyb expression near the FLM and nIII. Abbreviations: FLM, fasciculus longitudinalis medialis; nIII, nervus oculomotorius; NMH, nucleus magnocellularis hypothalami; NPPv, nucleus posterioris periventricularis; Thv, ventral thalamus; Valv, valvula cerebelli; Vd, dorsal nucleus of ventral telencephalon; Vl, lateral nucleus of ventral telencephalon; Vv, ventral nucleus of ventral telencephalon. Scale bar (if no other indication) = 500 $\mu$m

3.2. agouti‐related peptide 1 (agrp1)

Analysis by ISH of agrp1 showed labeled neurons in the hypothalamic NLT, including the ventral NLT (NLTv, Figure 3). agrp1‐expressing neurons located at the pituitary stalk were situated medially toward the infundibulum as infundibular cerebrospinal‐fluid contacting cells, and a few neurons laterally from cerebrospinal fluid, connecting the posterior pituitary and caudal hypothalamus. agrp1 mRNA expression was not detected in any other brain region.

FIGURE 3.

agrp1 mRNA expression in Atlantic salmon parr brain. (a1) Nissl‐staining at equivalent level to the schematic drawing illustrating agrp1 expression by red dot. Lower left corner represents a schematic brain indicating the position of the section. (a2) Nissl‐staining and corresponding agrp1 expression in the ventral nucleus lateralis tuberis (NLTv) at the pituitary stalk. Abbreviation: Pit, pituitary. Scale bar (if no other indication) = 500 $\mu$m

3.3. agouti‐related peptide 2 (agrp2)

agrp2 was mainly expressed in the telencephalon, but a scattered expression was also found in the diencephalon and rhombencephalon (Figure 4). In the telencephalon, neurons expressing agrp2 demonstrated a specific pattern from the medial region (Dm) and central region (Dc) toward the ventral part of the lateral zone (Dl‐v) of the dorsal telencephalon (Figure 4a(1–3)). A few agpr2‐positive neurons were located in the ventral thalamus (Thv, Figure 4b(1–3)). In the rhombencephalon, agrp2 mRNA expression was identified in cells lateroventral to the rhombencephalic ventricle in small nuclei near large nuclei of the nucleus motorius nervi trigemini (nV, Figure 4c(1–3)) and laterally to the FLM in the reticular formation (RF, Figure 4d(1–3)).

FIGURE 4.

agrp2 mRNA expression in Atlantic salmon parr brain. (a1–d1) Nissl‐staining compared to schematic drawing illustrating agrp2 expression by yellow dots. Lower left corner represents a schematic brain indicating the position of the section. (a2–d2, a3–d3) Nissl‐staining and corresponding agrp2 expression with neuron anatomical structures. (a) agrp2 expression in the Dm, Dc, and Dl‐v parts of the telencephalon. (b) agrp2 expression in the Thv. (c) agrp2 expression near the FLM and (d) the RF. Abbreviations: Dc, central zone of dorsal telencephalon; Dd, dorsal part of dorsal telencephalon; Dl‐d, dorsal part of lateral zone of dorsal telencephalon; Dl‐v, ventral part of lateral zone of dorsal telencephalon; Dm, medial zone of dorsal telencephalon; FLM, fasciculus longitudinalis medialis; OT, optic tectum; RF, reticular formation; Thv, ventral thalamus. Scale bar (if no other indication) = 500 $\mu$m

3.4. cocaine‐ and amphetamine‐regulated transcript (cart)

The cart paralogs mRNA distributions in the brain of Atlantic salmon parr were analyzed by ISH using seven distinct RNA probes (Table 1). The results are presented in a rostrocaudal direction from the most abundant, cart2b, to the lowest abundant cart paralogs (cart4, 1a, and 1b).

3.4.1. cart2b

cart2b was the most abundant cart paralog and showed a wide distribution in several brain regions (Figure 5), being continuously observed from the olfactory bulb to the thalamus. In the olfactory bulb, numerous cells expressing cart2b were found in the granular cell layer (strg, Figure 5b(1–2)). cart2b‐positive cells were found in the subpallium in the dorsal nucleus of ventral telencephalon (Vd) near the telencephalic ventricle (Figure 5c(1–3)). In the lateral telencephalon, a cluster of cells expressing cart2b was observed in the lateral nucleus of the ventral telencephalon (Vl). Scattered cells expressing cart2b were observed in the ventral nucleus of ventral telencephalon (Vv), lateral (Dl‐d and Dl‐v), medial (Dm), and dorsal (Dd) zone of the dorsal telencephalon (Figure 5(c3)). The cart2b mRNA expression from strg and Vd could be followed continuously to the periventricular preoptic region, as shown in nucleus posterior periventricularis (Ppp, Figure 5d(1–3)). cart2b was detected from the optic tectum in the periventricular layer (SPV) toward the stratum album centrale (SAC) border (Figure 5(e2)), and from the dorsal‐most region adjacent to torus longitudinalis until the base of the optic tectum near torus semicircularis (Figure 5(e1) (f1) (g1) (h1), and (i1)). In the midbrain, cart2b was present in the dorsal thalamus (Thd), posterior tuberculum (Pt) toward the diencephalic ventricle, and in the hypothalamus (Figure 5e,h).

FIGURE 5.

cart2b mRNA expression neurons in Atlantic salmon parr brain. (a) Schematic representation of the brain indicating the position of each transverse section. (b1–l1) Nissl‐staining compared to schematic drawing illustrating cart2b expression by blue dots. (b2‐i2, c3‐i3) Nissl‐staining and corresponding cart2b expression along with neuroanatomical structures. (b) cart2b expression in the olfactory bulb stgr. (c) cart2b expression in the Dm, Dd, and Dl zones of dorsal telencephalon as well as the Vd, Vl and Vv nucleus of ventral telencephalon. (d) cart2b expression in the preoptic region—Ppp. (e) cart2b expression in the SPV, Thd, Pt, NAT, and NLTv. (f) cart2b expression in the SPV, Thd, NMH, NAT, tlat, and Lih. (g) cart2b expression in the SPV, scattered neurons in the Ts, large cluster near NMH and PVO, scattered neurons in NAT, nrl, Ce, tlat, and Lih. (h) cart2b expression in the SPV, and dorsal midbrain tegmentum toward Ts and the nlv, in the NDILm and nrl, and SV. (i) cart2b expression in the SPV, gran of the cerebellum, and in the lcoer and rets. Abbreviations: Ce, nucleus centralis lobi inferioris hypothalamic; Cho, optic chiasm; D, dorsal telencephalon; Dd, dorsal zone of dorsal telencephalon; Dl‐d, dorsal part of lateral zone of dorsal telencephalon; Dl‐v, ventral part of lateral zone of dorsal telencephalon; Dm, medial zone of dorsal telencephalon; Ggl, stratum ganglionare—cerebelli; Gran, stratum granulare—cerebelli; Hab, habenula; Lcoer, locus coeruleus; Lih, lobus inferior hypothalami; Mcba, tractus mesencephalo‐cerebellaris anterior; Mol, stratum moleculare—cerebelli; NAT, nucleus anterior tuberis; NDILm, medial part of the diffuse nucleus of inferior lobe; NLT, nucleus lateralis tuberis; NLTv, ventral nucleus lateralis tuberis; NLV, nucleus lateralis valvulae; NMH, nucleus magnocellularis hypothalami; NRL, nucleus recessi lateralis; OT, optic tectum; Ppp, posterior parvocellular preoptic nucleus; Pt, posterior tuberculum; PVO, paraventricular organ; Rets, formatio reticularis pars superior; SAC, stratum album centrale; SPV, stratum periventriculare of the optic tectum; Stgr, stratum granulare—bulbi olfactory; SV, saccus vasculosus; Thd, dorsal thalamus; Thv, ventral thalamus; Tlat, torus lateralis; Toll, tractus olfactorius lateralis; Ts, torus semicircularis; Valv, Valvula cerebelli; Vd, dorsal nucleus of ventral telencephalon; Vl, lateral nucleus of ventral telencephalon; Vv, ventral nucleus of ventral telencephalon; Scale bar (if no other indication) = 500 $\mu$m

cart2b mRNA was abundant in the hypothalamus, in a dorsoventral direction from the NAT toward the ventral NLT (NLTv, Figure 5(e3)). Scattered cells expressing cart2b were observed laterally in torus lateralis (tlat) and lobus inferior hypothalami (Lih) toward the cerebrospinal fluid (Figure 5(f3)). A cluster of cells expressing abundantly cart2b were also observed in the paraventricular organ (PVO), nucleus posterior tuberis (PTN) as well as in NMH (Figure 5g(1–3)). From NAT, cart2b expression continued to be detected dorsolateral to the infundibulum, into nucleus recessi lateralis (NRL) and medial part of the diffuse nucleus of Lih (NDILm, Figure 5(h1) and (h3)). Scattered cells expressing cart2b were observed in saccus vasculosus (SV, Figure 5(h3)).

Ventral to the valvula in dorsal tegmentum, scattered neurons expressing cart2b were observed in torus semicircularis (Ts) toward nucleus lateralis valvulae (nlv, Figure 5(g2) and (h2)). In the rhombencephalon, cart2b was observed in stratum ganglionare (ggl) of corpus cerebelli (Figure 5i(1–2)) and ventrolateral to the fourth ventricle near locus coeruleus (lcoer) and formatio reticularis pars superior (rets, Figure 5(i1) and (i3)).

3.4.2. cart3a

cart3a mRNA expression was identified in the telencephalon, midbrain, and rhombencephalon (Figure 6). In the telencephalon, neurons expressing cart3a were identified in the ventral nucleus of ventral telencephalon (Vv, Figure 6b(1–b2)), in the ventrolateral telencephalon in nucleus entopeduncularis (ent), and within SOC of the preoptic region (Figure 6c(1–3)). The cart3a in the optic tectum was expressed in the less densely populated neurons in the torus longitudinalis (TL) toward the white matter of the torus longitudinalis (TLw, Figure 6(d1) (d2) (e1), and (e2)). cart3a expression was observed in scattered cells in the dorsal thalamus (Figure 6(d3)), and in the hypothalamus dorsal to the paraventricular organ (PVO) in PTN and NMH (Figure 6(e3)). cart3a expression was also found in the dorsal mesencephalic tegmentum (DTN)—possibly near the EW (Figure 6f(1–3)), dorsomedial to fasciculi longitudinalis medialis (FLM, Figure 6g(1–2)), and scattered neurons laterally to the rhombencephalic ventricle near lcoer (Figure 6(g3)). In rostral rhombencephalon, cart3a mRNA was found ventral of nervus trigeminus (nV) in nervus motorius nervi trigemini (nVm, Figure 6h(1–3)).

FIGURE 6.

cart3a mRNA expression in Atlantic salmon parr brain. (a) Schematic representation of the brain indicating the position of each transverse section. (b1–h1) Nissl‐staining compared to schematic drawing illustrating cart3a expression by red dots. (b2–h2, c3–h3) Nissl‐staining and corresponding cart3a expression along with neuroanatomical structures. (b) cart3a expression in the Vv. (c) cart3a expression in the ent and in the SOC. (d) cart3a expression in the TLw of the optic tectum and in the Thd. (e) cart3a expression in the TL, and in the NMH and PVO. (f) cart3a expression in the TL and dorsal tegmentum near EW. (g) cart3a expression near the FLM and lcoer. (h) cart3a expression near the nV and nVm. Abbreviations: Cho, optic chiasm; cp, commissura posterior; Ent, nucleus entopeduncularis; EW, Edinger–Westphal nucleus; FLM, fasciculus longitudinalis medialis; fMth, fiber of Mauthner cell; Lcoer, locus coeruleus; NAT, nucleus anterior tuberis; NMH, nucleus magnocellularis hypothalami; nV, nervi trigemini; nVm, nucleus motorius nervi trigemini; OT, optic tectum; PVO, paraventricular organ; SOC, supraoptic/suprachiasmatic nucleus; Thd, dorsal thalamus; TL, torus longitudinalis; TLw, white matter of the torus longitudinalis; Valv, valvula cerebelli; Vv, ventral nucleus of ventral telencephalon. Scale bar (if no other indication) = 500 $\mu$m

3.4.3. cart3b

The localization of cart3b by ISH demonstrated its presence in the preoptic region, thalamus, tegmentum, and rhombencephalon (Figure 7). cart3b mRNA expression was detected in the caudal preoptic region, specifically in the posterior part of the parvocellular preoptic nucleus (Ppp, Figure 7b(1–2)). Ventral to the optic tectum in the diencephalon, cart3b probe labeled a neuronal line from the dorsal thalamus (Thd) to the ventrolateral direction of the hypothalamic torus lateralis (tlat, Figure 7c(1–3)). Ventral to the valvula, cart3b was expressed in scattered neurons of the central and ventral torus semicircularis (TS, Figure 7d(1–3)). The cart3b probe also hybridized scattered neurons located ventrolateral to the FLM (Figure 7e(1–3)), neurons near nervus trigemini (nV, Figure 7f(1–3)), and neurons in the formatio reticularis pars superior (rets) as well as dorsal cells to rets (Figure 7g(1–3)).

FIGURE 7.

cart3b mRNA expression in Atlantic salmon parr brain. (a) Schematic representation of the brain indicating the position of each transverse section. (b1–g1) Nissl‐staining compared to schematic drawing illustrating cart3b expression by pink dots. (b2–g2, c3–g3) Nissl‐staining and corresponding cart3b expression along with neuroanatomical structures. (b) cart3b expression in the preoptic region—Ppp. (c) cart3b expression in a ventrolateral direction from the Thd toward tlat. (d) cart3b expression in dorsal tegmentum near Ts. (e) cart3b expression ventral to the FLM. (f) cart3b expression near nV. (g) cart3b expression ventral to the FLM near the rets. Abbreviations: FLM, fasciculus longitudinalis medialis; fMth, fiber of Mauthner cell; Lih, lobus inferior hypothalami; nV, nervi trigemini; ppp, posterior parvocellular preoptic nucleus; Rets, formatio reticularis pars superior; Tbc, tractus tecto‐bulbaris cruciatus; Thd, dorsal thalamus; Tlat, torus lateralis; Ts, torus semicircularis; Valv, valvula cerebelli; Ve4, fourth ventricle (rhombencephali). Scale bar (if no other indication) = 500 $\mu$m

3.4.4. cart2a

cart2a‐expressing cells were identified in the telencephalon, optic tectum, thalamus, and hypothalamus (Figure 8). One cart2a neuronal cluster was present in the ventral nucleus of the ventral telencephalon (Vv) toward the telencephalic ventricle (Figure 8b(1–2)). Scattered cart2a‐positive cells were detected in tractus opticus pars distalis (tod, Figure 8c(1–2)). In the optic tectum, scattered cells expressing cart2a were observed in stratum marginale (SM), stratum opticum (SO, Figure 8(c3)), in the album layer (SAC), and evenly distributed in the periventricular layer (SPV) toward the SAC (Figure 8d(1–2) and (e1)). In the midbrain, one cart2a cell cluster was observed in the dorsal thalamus (Thd, Figure 8d(1–3)). In the hypothalamus, a cluster of cells expressing cart2a was identified in the PTN (Figure 8e(1–3)).

FIGURE 8.

cart2a mRNA expression in Atlantic salmon parr brain. (a) Schematic representation of the brain indicating the position of each transverse section. (b1–e1) Nissl‐staining compared to schematic drawing illustrating cart2a expression by blue dots. (b2‐e2, c3‐e3) Nissl‐staining and corresponding cart2a expression along with neuroanatomical structures. (b) cart2a expression in Vv of telencephalon. (c) cart2a expression in tod and SO. (d) cart2a expression in SPV, SAC, and SO of the optic tectum and ventrally in the Thd. (e) cart2a expression in the hypothalamus near the PTN. Abbreviations: Cho, optic chiasm; Hab, habenula; inf, infundibulum; Lih, lobus inferior hypothalamic; NLTp, posterior nucleus lateralis tuberis; Ppp, posterior parvocellular preoptic nucleus; PTN, nucleus posterior tuberis; Rl, recessi lateralis; SAC, stratum album centrale (tecti mesencephali); SM, stratum marginale (tecti mesencephali); SO, stratum opticum (tecti mesencephali); SPV, stratum periventriculare of the optic tectum; SV, saccus vasculosus; Thd, dorsal thalamus; Tod, tractus opticus dorsalis; Vd, dorsal nucleus of ventral telencephalon; Vv, ventral nucleus of ventral telencephalon. Scale bar (if no other indication) = 500 $\mu$m

3.4.5. cart4, 1b, and 1a

cart4, 1b, and 1a were expressed in distinct brain regions from the rostral to the caudal direction (Figure 9). cart4 was only expressed in the most rostral area of the diencephalon, in the posterior parvocellular preoptic nucleus (Ppp, Figure 9b(1–2)). In the dorsomedial mesencephalic tegmentum ventral to the valvula near the nucleus medial longitudinal fasciculus (NMFL), cart1a (Figure 9c(1–3)) and cart1b (Figure 9d(1–3)) mRNA expression were identified. cart1b was only observed in one cell cluster, while the cart1a probe identified two separate clusters of neurons adjacent and medial to the FLM and oculomotor nucleus (NIII, Figure 9e(1–3)).

FIGURE 9.

cart4, 1b, and 1a mRNA expression in Atlantic salmon parr brain. (a) schematic representation of the brain indicating the position of each transverse section. (b1–e1) Nissl‐staining compared to schematic drawing illustrating cart4 expression by pink dot, cart1b by green dots, and cart1a by blue dots. (b2–e2, c3–e3) Nissl‐staining and corresponding cart4, 1b, and 1a expression along with neuroanatomical structures. (b) cart4 expression the preoptic region—Ppp. (c) cart1b expression near the NMFL. (d) cart1a expression near the NMFL. (e) cart2a expression near the nIII. Abbreviations: Cho, optic chiasm; Cp, commissura posterior; Ent, nucleus entopeduncularis; FLM, medial longitudinal fasciculus; Lih, inferior hypothalamic lobe; nIII, Nucleus oculomotorius; NMFL, nucleus medial longitudinal fasciculus; OT, optic tectum; Ppp, posterior parvocellular preoptic nucleus; SV, saccus vasculosus; TL, torus longitudinalis; Ts, torus semicircularis; Valv, valvula cerebelli. Scale bar (if no other indication) = 500 $\mu$m

3.5. pro‐opiomelanocortin (pomc)

In the Atlantic salmon parr brain, pomca‐expressing cells were detected in the pituitary (adenohypophysis), and in the NLTv of the hypothalamus (Figure 10a(1–2)). The hypothalamic pomca‐expressing cells were located medially toward the infundibulum. pomcb was strongly expressed in the adenohypophysis of the pituitary (Figure 10b(1–2)), and was not observed in the NLT, or in any other brain regions.

FIGURE 10.

pomc mRNA expression in Atlantic salmon parr brain. (a1–b1) Nissl‐staining compared to schematic drawing illustrating pomca expression by light blue dots, and pomcb expression by green dots. Lower left corner represents a schematic brain indicating the position of the section. (a2–b2) Nissl‐staining and corresponding pomca and pomcb expression with neuroanatomical structures. (a) pomca expression in the NLTv and adenohypophysis of the Pit. (b) pomcb expression in the adenohypophysis of the Pit. Abbreviations: NLTv, ventral nucleus lateralis tuberis. Pit, pituitary. Scale bar (if no other indication) = 500 $\mu$m

3.6. Hypothalamic expression of melanocortin system neuropeptides

To determine whether the Atlantic salmon tuberal hypothalamus coexpress npya/agrp1 and/or cart2b/pomca, double labeling fluorescent ISH was used. The results show that neurons expressing npya did not coexpress agpr1 (Figure 11). In the anterior NLT (NLTa) of the rostral tuberal hypothalamus, few neurons expressed npya mRNA, but no agrp1 expression was found in this region (Figure 11a). Toward the NLTv at the pituitary stalk, both npya and agrp1 were abundantly expressed in neighboring neurons (Figure 11b(1–4)). npya and agrp1 were still present in neighboring neurons of the ventral NLT (NLTv) bordering the nucleus posterior tuberis (NPT) in the caudal tuberal hypothalamus (Figure 11c), but their expression decreased, particularly for agrp1, in comparison to the NLTv at the pituitary stalk.

FIGURE 11.

npya, agrp1, cart2b, and pomca mRNA expression in Atlantic salmon tuberal hypothalamus. Left side: schematic representation of the tuberal hypothalamus indicating the position of each transverse section. (a–c) npya (TSA‐green) and agrp1 (FastBlue‐red) expression. (d–f) cart2b (TSA‐green) and pomca (FastRed‐red) expression. (a) npya expression in the NLTa of the rostral tuberal hypothalamus. (b1) npya and agrp1 expression in the NLTv at the pituitary stalk. (b2–b4) npya and agrp1 expression in neighboring neurons in the NLTv. The absence of yellow staining indicates no coexpression between npya and agrp1 mRNA. (c) npya and agrp1 expression in the NLTp/NPT of the caudal tuberal hypothalamus. (d) cart2b and pomca expression in the NLTa of the rostral tuberal hypothalamus. The presence of yellow staining (white arrows) indicates coexpression of cart2b and pomca mRNA. (e) cart2b and pomca expression in the NLTv of the tuberal hypothalamus. (e2–e4) The presence of yellow staining (white arrows) indicates coexpression of cart2b and pomca mRNA in the NLT. (f) cart2b expression in NAT, and pomca expression in the NLTp/NPT. Sections were mounted with ProLong Glass antifade medium with NucBlue (Invitrogen). Abbreviations: NAT, nucleus anterior tuberis; NLTa, anterior nucleus lateralis tuberis; NLTp, posterior nucleus lateralis tuberis; NLTv, ventral nucleus lateralis tuberis; NPT, nucleus posterior tuberis; Pit, pituitary; SV: saccus vasculosus

cart2b/pomca coexpression was observed in the NLTa and NLTv of the tuberal hypothalamus (Figure 11d,e). However, cart2b and pomca were mainly expressed in distinct neurons of the tuberal hypothalamus. The cart2b‐positive neurons were gradually located dorsally toward the NAT, while pomca expression remained ventrally in the NLT (Figure 11e,f). Thus, no cart2b/pomca coexpression was found in the caudal tuberal hypothalamus.

4. DISCUSSION

In this study, ISH was utilized to map the spatial distribution of npy, agrp, cart, and pomc in the Atlantic salmon parr brain (summarized in Table 2). The topology of these neuropeptides, particularly in the lateral tuberal nucleus (NLT), supports that the hypothalamic nucleus is associated with appetite and food intake regulation. We also demonstrated the presence of cart2b/pomca coexpression in the anterior and ventral NLT.

As a result of the salmonid‐specific fourth whole‐genome duplication event (Allendorf & Thorgaard, 1984; Lien et al., 2016), several paralogs of npy, cart, and pomc have been identified in Atlantic salmon (Kalananthan et al., 2021; Murashita et al., 2011; Tolås et al., 2021; Valen et al., 2011). Although the fate of the duplicated genes of key players in the melanocortin system are not yet fully understood, one hypothesis is that some genes will still play a role in appetite control by facilitating physiological, sensory, or periprandial responses. Indeed, our results demonstrate that these neuropeptides are expressed in the salmon brain regions known to be related to feeding and energy status. These regions include the hypothalamus, known to be related to regulation of vital homeostatic feeding control in both fish and mammals, as well as the olfactory bulb, telencephalon, optic tectum, and secondary gustatory nucleus, which are linked to feeding (Demski & Knigge, 1971; Volkoff et al., 2005). Indeed, we found npy, agrp2, and cart in the olfactory bulb, telencephalon, and optic tectum (Figures 1, 2, 3, 4, 5, 6, 7, 8, 9). Several of these brain regions are indirectly linked to chemical stimulation of appetite, either through inputs from sensory organs (olfaction and taste) or by hedonic (nonhomeostatic) regulation (Arikawa et al., 2020; Rossi & Stuber, 2018; Volkoff, 2019).

4.1. Hypothalamic expression of melanocortin system neuropeptides

The hypothalamic neuropeptides npya, agrp1, cart2b, and pomca are involved in appetite control as their expression levels responded to a fed/fasted state in Atlantic salmon (Kalananthan et al., 2021; Kalananthan, Murashita, et al., 2020; Murashita et al., 2011, 2009; Tolås et al., 2021; Valen et al., 2011). Here, we show the presence of these neuropeptides in the NLT region of the Atlantic salmon parr, the putative homolog to the mammalian arcuate nucleus ((Cerdá‐Reverter & Peter, 2003; Cerdá‐Reverter, Ringholm, et al., 2003) and reviewed in Biran et al. (2015)), supporting previous evidence that this region and these genes are involved in appetite control. This can be further supported by the presence of a few neurons coexpressing cart2b/pomca, and the expression of agrp1 and npya in neighboring cells in the NLT (Figure 11). There is evidence that the homeostatic control of appetite by the melanocortin system involves the stimulation of hypothalamic arcuate nucleus first‐order orexigenic and anorexigenic neurons, which then project to second‐order hypothalamic neurons which in turn project to autonomic centers in the hindbrain (Morton et al., 2006; Schwartz et al., 2000). The resulting neuronal net output stimulates anabolic or catabolic pathways. Previous studies mapping the neuroanatomical distribution of melanocortin circuits in teleosts have hypothesized possible coexpressions (Delgado et al., 2017; Porter et al., 2017; Soengas et al., 2018); however, this has never been demonstrated. Thus, to our knowledge, this is the first evidence of coexpression between cart2b/pomca in the NLT region of a teleost species. In agreement with the findings of Jeong et al. (2018) in the hypothalamus of zebrafish, no coexpression of npy/agrp1 was observed in Atlantic salmon NLT. Therefore, as previously suggested, coexpression of npy/agrp1 might not be required for the action of these neuropeptides in appetite control of teleost fishes (Jeong et al., 2018). As a contrast, at least 90% of the neurons in the mammalian arcuate nucleus that express Npy or Cart also express Agrp or Pomc, respectively, and play a crucial role in a homeostatic regulation of appetite (Elias et al., 1998; Hahn et al., 1998; Schwartz et al., 2000). The limited number of neurons coexpressing cart2b/pomca in the tuberal hypothalamus of Atlantic salmon, and that there was no coexpression of npya/agrp1 suggest that coexpression might not be required for homeostatic feeding control in Atlantic salmon and other teleost species.

The presence of npy in the NLT region seems to be conserved throughout evolution since it has been observed in several teleost species including sea bass (Cerdá‐Reverter et al., 2000), goldfish (Kojima et al., 2010), Atlantic cod (Le et al., 2016), and African cichlid fish (Porter et al., 2017). The NLT region is considered a site for integrating and releasing neurotransmitters to higher‐order neurons linked to neuroendocrine appetite control and feeding behavior (Rønnestad et al., 2017). In Atlantic salmon, agrp1 was exclusively detected in the hypothalamic NLT (Figure 3). The NLT agrp1 expression is in line with observations for other species like goldfish (Cerdá‐Reverter & Peter, 2003), zebrafish (Forlano & Cone, 2007; Koch et al., 2019; Shainer et al., 2017), sea bass (Agulleiro et al., 2014), African cichlid fish (Porter et al., 2017), and rainbow trout (Otero‐Rodino et al., 2019). Indeed, agrp1 function has been associated with appetite control in Atlantic salmon (Kalananthan, Murashita, et al., 2020; Murashita et al., 2009). Furthermore, in zebrafish, it has been shown that agrp‐neurons are hypophysiotropic, projecting from the NLT to the pituitary (Zhang et al., 2012). The high degree of similarity in the NLT agrp1 population among fish suggests that the involvement of this region in controlling appetite is well conserved. The hypothalamic nuclei expressing cart2b mRNA (Figure 5e,h) in Atlantic salmon are consistent with previous studies of hypothalamic cart expression (Akash et al., 2014; Porter et al., 2017). Additionally, salmon hypothalamic cart2b expression has been shown to respond to a fed/fasted state (Kalananthan et al., 2021). Atlantic salmon hypothalamic neurons also expressed cart3a in the NMH area. Indeed, it has been shown that cart3a expression is upregulated in the hypothalamus after 3 days of fasting, indicating a potential role in modulating appetite control (Kalananthan et al., 2021). In zebrafish, cart2a (previously named cart2) presence in the NRL indicated a role in mediating energy homeostasis (Akash et al., 2014). Thus, the expression of cart2a in the salmon PTN near the infundibulum supports the observations that cart2a might modulate food intake in salmon (Kalananthan et al., 2021). In Atlantic salmon, cart2 (cart2a and 2b) seems to be the only cart gene with similar potential proteolytic sites as its mammalian homolog, based on their sequence alignment (Kalananthan et al., 2021). POMC is a key regulator in the melanocortin system that is post‐transcriptionally cleaved into α‐ and β‐melanocyte‐stimulating hormones, adrenocorticotropic hormone, and β‐endorphin (Takahashi & Mizusawa, 2013). Here, pomca was expressed in the NLT area in the brain of Atlantic salmon parr. This result is in line with previous studies of pomc or α‐melanocyte‐stimulating hormones in goldfish (Cerdá‐Reverter, Schiöth, et al., 2003; Forlano & Cone, 2007; Porter et al., 2017), barfin flounder Verasper moseri (Amano et al., 2005), zebrafish (Zhang et al., 2012), African cichlid fish (Porter et al., 2017), and rainbow trout (Otero‐Rodino et al., 2019). These observations together with the findings of Kalananthan, Murashita, et al. (2020) suggest the involvement of pomca in Atlantic salmon appetite control.

4.2. Expression of melanocortin system neuropeptides in other brain regions

The widespread distribution of npy and cart in the brain of Atlantic salmon parr indicates various potential functional roles in the central nervous system. The neuropeptides npya and cart2b were the most abundant, as previously demonstrated (Kalananthan et al., 2021; Tolås et al., 2021). Additionally, cart2b expression resembled that of npy in brain areas associated with sensory processing, such as its presence in the olfactory bulb, which is known to be linked with processing chemosensory information, immune responses, and reproduction (Ye et al., 2020). This expression is consistent with other studies in teleosts for npy (Cerdá‐Reverter et al., 2000; Gaikwad et al., 2004; Kaniganti et al., 2021; Pirone et al., 2008; Porter et al., 2017) and cart or cart2b (Akash et al., 2014; Bonacic et al., 2015; Kalananthan et al., 2021; Le et al., 2016). Interestingly, fasted zebrafish express higher levels of npy in the olfactory bulb compared to those fed (Kaniganti et al., 2021), while in Atlantic salmon fasting decreased npya1 and increased cart2b mRNA levels in the olfactory bulb (Kalananthan et al., 2021; Tolås et al., 2021). In fact, npy has been suggested to serve as a neurotransmitter, while cart is involved in modulating the activity that can affect chemosensory processing and food‐seeking behavior (Akash et al., 2014; Gaikwad et al., 2004; Kalananthan et al., 2021; Kaniganti et al., 2021; Singru et al., 2008; Tolås et al., 2021). Taken together, these two peptides (npya and cart2b) may work together or independently in processing and transmitting olfactory sensory information in Atlantic salmon.

npy is expressed in the telencephalon of all teleost species investigated to date (Castro et al., 1999; Cerdá‐Reverter et al., 2000; Gaikwad et al., 2004; Le et al., 2016; Pirone et al., 2008; Porter et al., 2017; Saha et al., 2015; Singru et al., 2008; Tolås et al., 2021). In the ventral telencephalon, npya was abundantly expressed (Figure 1c), while npyb was much less abundant (Figure 2a). Further, cart2a, 2b, and 3a were also expressed in the ventral telencephalon (Figures 5, 6, and 8b). Telencephalon plays a role in sensory input processing connected to various functions such as reproduction (Saha et al., 2015; Uezono et al., 2015), behavior (Comesaña et al., 2018), and appetite control (Ye et al., 2020). Anatomically, the telencephalon has afferent and efferent connections with many brain regions, including the olfactory bulb, preoptic region, and tuberal hypothalamus (Folgueira et al., 2004a, 2004b). Telencephalic cart expression is linked with sensory‐motor function, while npy expression in the telencephalon has been linked to olfactory sensory processing (Singru et al., 2008), suggesting that npy might be involved in the hedonic control of food intake in this brain region. Interestingly, zebrafish cart2 (Akash et al., 2014) and catfish cart (Subhedar et al., 2011) decrease in the entopeduncular nucleus during starvation, while fasting had no impact on Atlantic salmon npy and cart transcripts in the telencephalon (Kalananthan et al., 2021; Tolås et al., 2021). The species‐specific cart responses indicate that more research is needed to understand the role of telencephalic cart in appetite control in teleost species. agrp2 was strongly expressed in the dorsal telencephalon in Atlantic salmon parr (Figure 4a), which is in line with previous findings in salmon (Kalananthan, Lai, et al., 2020). Opposite to that found in zebrafish, no agpr2 expression was observed in the pineal cells of Atlantic salmon pa (Shainer et al., 2017, 2019; Zhang et al., 2010). Zebrafish agrp2 has been found in novel, uncharacterized, nonphotosensitive pineal cells, in addition to a few neurons in the preoptic region that project to the adenohypophysis, indicating that this neuropeptide is linked to the hypophysiotropic stress axis in zebrafish (Shainer et al., 2017). As suggested in zebrafish, agrp2 in salmon might have a functional role in the spatial navigation network or a stress response via cortisol and the medial and lateral zones of the dorsal telencephalic serotonergic system (Rodríguez et al., 2021; Silva et al., 2015).

The preoptic region, located rostral to the hypothalamus, is functionally and neurochemically associated with the hypothalamus—including reproduction and sensory processing (Porter et al., 2017). In fact, the preoptic region functions as a key region for downstream signaling as the neurons from the preoptic region may be connected to several brain regions (Folgueira et al., 2004b), and innervate the pituitary via the hypothalamic‐neurohypophyseal tract (Akash et al., 2014; Forlano & Cone, 2007; Herget et al., 2014). These signals include serotonergic and corticotropin‐releasing factor systems that can affect food intake (Ortega et al., 2013). Preoptic expression of Atlantic salmon npya was observed in several subregions, including the SOC (Figure 1e), as it has previously been shown for other teleost species (Cerdá‐Reverter et al., 2000; Jeong et al., 2018; Le et al., 2016; Perez Sirkin et al., 2013; Pirone et al., 2008; Porter et al., 2017). Moreover, cart2b, 3a, 3b, and 4 were expressed in the preoptic region (Figures 4, 5, 6b, and 8a), similar to that reported for other teleosts (Akash et al., 2014; Le et al., 2016; Mukherjee et al., 2012; Porter et al., 2017), suggesting that cart, like npya, might be involved as preoptic neuroendocrine regulators. npya was expressed ventrally in the left and right habenula (Figure 1f), which is homologous to the mammalian lateral habenula (Amo et al., 2010) where NPY modulates excitatory transmissions (Cheon et al., 2019). Moreover, lateral habenular NPY might be indirectly linked to the hedonic regulation of appetite in primates (reviewed by Rezitis et al. (2022)). The lateral habenula is indeed a central node connecting rostral and caudal brain regions; afferent connections originate from the nucleus entopeduncularis (ENT, homologous to the globus pallidus in primates), and efferent connections to the median raphe nucleus in the ventral tegmentum (Hikosaka et al., 2008; Turner et al., 2016). Furthermore, agrp2, cart2a, 2b, 3a, 3b, and npya were expressed in the thalamus (Figures 1, 4, 6e,d, 7c, and 8d), suggesting an involvement of these neuropeptides in the modulation of sensory inputs to the telencephalon (Folgueira et al., 2004a, 2004b; Singru et al., 2007). In the midbrain, npyb was expressed exclusively near NPPv as well as in proximity to NIII (Figure 2b), with no expression observed ventrally toward the hypothalamus. This suggests that npyb might not have a direct role in appetite control, which agrees with previous studies on Atlantic salmon (Tolås et al., 2021), tiger puffer Takifugu rubripes (Kamijo et al., 2011), and Nile tilapia Oreochromis niloticus (Yan et al., 2017).

In zebrafish, visual information is essential to modulate feeding behavior (Muto et al., 2017), and feeding state modulates the activity of sensory processing involved in fine‐tuning the response to external stimuli, such as prey capture or avoidance behavior (Corradi & Filosa, 2021). Atlantic salmon npya expression in the SPV and scattered cells in SGV of the optic tectum (Figure 1g,l) was in line with previous studies in teleost species (Cerdá‐Reverter et al., 2000; Das et al., 2019; Porter et al., 2017). Together, this indicates that npya may have a role in both signaling feeding status and visual perception, as suggested previously for Atlantic salmon (Tolås et al., 2021) and zebrafish (Filosa et al., 2016), and also supported by the high expression of npy in the salmon eye (Murashita et al., 2009). In the optic tectum, cart2b expressed in the SPV (Figure 5e,a,i) was similar to that of npya, while cart2a and 3a were expressed in the distal layers of SPV and torus longitudinalis, respectively, (Figures 6d–f and 8d,e) suggesting a role in integrating visual information for the later (Filosa et al., 2016).

In mammals, the EW plays a vital role in the integration and modulation of sympathetic outflow affecting stress and energy homeostasis through orexigenic and anorexigenic projections from the hypothalamic arcuate nucleus and paraventricular nucleus (Cano et al., 2021). In contrast, the EW in teleosts is rarely connected to appetite but it is described to be photosensitive in zebrafish (Hang et al., 2014). In this study, npya and cart3a were found near EW (Figures 1, 6), which is in line with previous studies for cart in teleosts, including catfish (Singru et al., 2007) and zebrafish (Akash et al., 2014), and this indicates that further studies are needed to better understand this region. Laminated TS receives inputs from the lateral line and visual system (Pirone et al., 2008), suggesting that cart2b and 3b might be involved in the processing of both visual and lateral stimuli. In the rhombencephalon, npya, b, agrp2, cart1a, 3a, and 3b were observed proximally to the FLM and nV. The npy expression observed here is similar to previous findings in Atlantic salmon and Gambusia affinis by NPY‐immunoreactivity (Garcia‐Fernandez et al., 1992). The expression of these neuropeptides near the nV indicates a possible involvement in food intake and sensory inputs from the oral cavity (Pirone et al., 2008). The expression of cart1a, 3a, and 3b in the rhombencephalic region suggests that these neurons may innervate the spinal cord and, thus, these neuropeptides may play a role in descending control from the brain stem, as speculated for cart in zebrafish (Akash et al., 2014).

The main site for pomc expression was the adenohypophysis, in line with previous observations in Atlantic salmon by qPCR (Kalananthan, Lai, et al., 2020; Kalananthan, Murashita, et al., 2020) and other teleost species (Amano et al., 2005; Forlano & Cone, 2007; Otero‐Rodino et al., 2019; Zhang et al., 2012). Downstream signaling from the adrenocorticotropic hormone, one peptide produced from pomc, is the hypothalamus‐pituitary‐interrenal axis affecting food intake through glucocorticoid production. Interestingly, starved zebrafish have been shown to have lower cortisol levels than fed fish (Filosa et al., 2016). Downstream signaling from the melanocyte‐stimulating hormones includes physiological color change mechanisms and stress response (Segura‐Noguera et al., 2000) that indirectly affect food intake. Mapping hypophysiotropic neurons in the hypothalamus by immunocytochemical studies has shown that α‐melanocyte‐stimulating hormone fibers project from NLT down to the pituitary in zebrafish (Zhang et al., 2012), but not in barfin flounder (Amano et al., 2005). The contradictory effects of pomca observed in previous studies might be explained by the end‐product of the post‐translational cleavage of pomc. While α‐melanocyte‐stimulating hormone has been shown to be a direct suppressor of appetite, β‐endorphin can antagonize the α‐melanocyte‐stimulating hormone downstream signaling pathways directly (Mercer et al., 2013). Thus, more research is needed to investigate the relationship between pomc and appetite and energy balance in vertebrates.

5. CONCLUSION

This study shows that the Atlantic salmon neuropeptides npy, cart, pomc, and npy are expressed in brain regions known to be related to feeding and energy status. This includes the hypothalamus, supporting the hypothesis that the melanocortin system and the NLT region of the hypothalamus are involved in the control of appetite in Atlantic salmon and that this function is conserved across vertebrates. In the Atlantic salmon hypothalamus, a distinct neuronal npya, agrp1, cart2b, and pomca expression was found, as well as a few neurons coexpressing cart2b/pomca. To what extent does this hypothalamic coexpression affect the physiological regulation of food intake compared to the distinct expression of these neuropeptides in Atlantic salmon is a question that needs further investigation. In addition, our data suggest that several of the neuropeptides investigated might be involved in the control of food intake and energy homeostasis through transmission and processing of sensory signals. This is based on their mRNA expression profile in the olfactory bulb, telencephalon, midbrain, and hindbrain.

AUTHOR CONTRIBUTIONS

Conceptualization: Ivar Rønnestad and Jon Vidar Helvik. Sampling: Sissel Norland, Mariann Eilertsen, Jon Vidar Helvik and Ana S. Gomes. Methodology, investigation, and analysis: Sissel Norland, Mariann Eilertsen, Jon Vidar Helvik, and Ana S. Gomes. Writing‐original draft and review and editing: Sissel Norland, Ivar Rønnestad, Mariann Eilertsen, Jon Vidar Helvik, and Ana S. Gomes.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Norland, S. , Eilertsen, M. , Rønnestad, I. , Helvik, J. V. , & Gomes, A. S. (2023). Mapping key neuropeptides involved in the melanocortin system in Atlantic salmon (Salmo salar) brain. Journal of Comparative Neurology, 531, 89–115. 10.1002/cne.25415

DATA AVAILABILITY STATEMENT

The data supporting the findings of this paper are primarily presented within the scope of this publication. Additional materials are available upon reasonable request to the corresponding author.