Relaxin-3/insulin-like peptide 7, a neuropeptide involved in the stress response and food intake


M. Tanaka, Department of Basic Geriatrics, Kyoto Prefectural University of Medicine, Kawaramachi-Hirokoji, Kamikyo-ku, Kyoto 602-8566, Japan
Fax: +81 75 251 5797
Tel: +81 75 251 5797


Relaxin-3, also known as insulin-like peptide-7, is a newly-identified peptide of the insulin superfamily. All members of this superfamily have a similar structure, which consists of two subunits (A-chain and B-chain) linked by disulfide bonds. Relaxin-3 is so named because it has a motif that can interact with the relaxin receptor. By contrast to other relaxins, relaxin-3 is mainly expressed in the brain and testis. In rodent brain, anatomical studies have revealed its predominant expression in neurons of the nucleus incertus of the dorsal pons, and a few other regions of the brainstem. On the other hand, relaxin-3-expressing nerve fibers and the relaxin-3 receptors, RXFP3 and RXFP1, are widely distributed in the forebrain, with the hypothalamus being one of the most densely-innervated regions. Therefore, relaxin-3 is considered to exert various actions through its ligand-receptor system. This minireview describes the expression of relaxin-3 in the brain, as well as its functions in the hypothalamus, including the stress response and food intake.


arcuate nucleus


corticotropin-releasing factor


CRF type 1 receptor


gonadotropin-releasing hormone






insulin-like peptide




lateral hypothalamic area


nucleus incertus


neuropeptide Y


protein kinase A


paraventricular hypothalamic nucleus


supraoptic nucleus


Relaxin-3/insulin-like peptide-7 (INSL7) has recently been identified as a new member of the insulin/relaxin family using human genomic databases [1]. The 142 amino acid human precursor polypeptide sequence is well conserved among humans, pigs, rats and mice [2]. Structurally, this precursor polypeptide consists of signal peptides, and a B-chain, C-peptide and A-chain, and contains the RXXXRXXI motif in the B chain (B12–B19 in human) for binding to the relaxin receptor [3]. Similar to insulin, a mature two-chain peptide is produced after removal of the C-peptide and the formation of three disulfide bonds between respective cysteine residues of the A-chain and B-chain [4]. An evolutionary study showed that relaxin-3 orthologs are present in fugu fish and zebrafish, but not in any invertebrate or prokaryote, and that these orthologs show high homology between different species in the mature peptide region. When compared with other insulin/relaxin superfamily members, relaxin-3 is constrained by strong purifying selection, suggesting that this protein is an ancestral form and has a highly-conserved function [5].

In the present minireview, the expression of relaxin-3 in the brain, and particularly its functions, including the stress response and food intake, are described.

Expression of relaxin-3 in the brain

Relaxin-3 neurons in the brain

Examination of relaxin-3 mRNA expression by northern blotting and reverse transcriptase-PCR revealed that relaxin-3 is abundant in the brain, but not in female reproductive tissue such as the ovary and uterus [1,6]. By contrast, the expression of two other known relaxin genes (i.e. those encoding human relaxin-1 and -2) was detected in the ovarian corpus luteum during pregnancy, and in the deciduas trophoblast [7–9]. Thus, the physiological function of relaxin-3 is considered to be different from that of other relaxin proteins involved in the growth and remodeling of reproductive and other tissues during pregnancy [10].

In the mouse and rat brain, relaxin-3 expression was reported to be localized to the central gray matter of the median dorsal pons near the fourth ventricle, termed the nucleus incertus (NI) [1,6,11]. We previously reported details of relaxin-3 expression at the cellular level using immunocytochemistry and in situ hybridization [12]. In addition to the primary site of expression (i.e. the NI), where, in the rat, approximately 2000 relaxin-3-positive neurons are found (Fig. 1A), a smaller number of these neurons are scattered in the pontine raphe nucleus, the periaqueductal gray matter, and the area dorsal to the substantia nigra in the midbrain reticular formation. By immunostaining using monoclonal antibody against the N-terminus of the human relaxin-3 A-chain [2], relaxin-3-immunoreactive fibers were observed to project densely to the septum, hippocampus, lateral hypothalamic area (LH) and intergeniculate leaflet of the thalamus (Fig. 1B). Ultrastructural examination revealed that relaxin-3 was localized to the dense-core vesicles in the perikarya, and it was also observed in the synaptic terminals of axons [12]. The NI comprises a distinct cell group in the caudoventral region of the pontine periventricular gray matter, adjacent to the ventromedial border of the caudal dorsal tegmental nucleus [13]. Studies involving neuronal tracing with anterograde and retrograde tracers have shown that the NI, together with the median raphe and interpeduncular nuclei, may form a midline behavior control network, and many targets of the NI, such as the medial septum, hippocampus, hypothalamus, mammillary complex and amygdala, are involved in arousal mechanisms, including the synchronization and desynchronization of the theta rhythm [14,15]. Recently, Ma et al. [16] reported that relaxin-3 neurons in the NI can help modulate spatial memory and the underlying hippocampal theta activity. Using immunocytochemistry studies, relaxin-3-positive neurons in the NI have been shown to be GABAergic and to co-express corticotropin-releasing factor (CRF) type 1 receptors (CRFR1) [12,17].

Figure 1.

 (A) Relaxin-3 immunoreactivity in the NI. Relaxin-3 is expressed in neurons of both the pars compacta (NIc) and pars dissipata (NId) of the NI. DTg, dorsal tegmental nucleus; 4V, fourth ventricle; mlf, medial longitudinal fasciculus. Scale bars = 100 μm. (B) A schematic representation of the major projection of relaxin-3 in the forebrain. DB, diagonal band; DR, dorsal raphe nucleus; IP, interpeduncular nucleus; LS, lateral septal nucleus; MS, medial septal nucleus; PAG, periaqueductal gray matter; RSC, retrosplenial cortex.

Relaxin-3 receptor

The cognate receptor for relaxin-3 is RXFP3, formally known as GPCR135 or SALPR [6,18]. Although it can also bind and activate RXFP1 and RXFP4, relaxin-3 binds RXPF3 with higher affinity (0.31 nm) than RXFP1 (2.0 nm) or RXFP4 (1.1 nm) [6,19]. RXFP3 mRNA is abundant in the olfactory bulb, paraventricular nucleus (PVN) and supraoptic nucleus (SON) in the hypothalamus amygdaloid–hippocampal area, as well as the bed nucleus stria terminalis, paraventricular thalamus, superior colliculus and interpeduncular nucleus in the brainstem. The distribution of RXFP3 approximately overlaps with the autoradiography pattern, showing selective RXFP3 binding of the chimeric peptide, relaxin-3 B-chain/INSL5 A-chain [20]. In the brain, there is generally a close correlation between relaxin-3-positive nerve terminals and RXFP3 expression; however, the density of expression of ligand and receptor is not always equal. For example, the olfactory bulb exhibits abundant RXFP3 expression, whereas it has relatively low levels of relaxin-3-immunoreactive fibers. In the hypothalamus, relaxin-3 fibers densely innervate the lateral hypothalamic area, although RXFP3 is strongly expressed in the PVN and SON [12,17,21]. The structure and function of the relaxin family peptide receptors, including RXFP3 and RXFP4, were recently reviewed by Kong et al. [22].

Relaxin-3 expression in development and in other species

During the development of the rat, relaxin-3 mRNA expression appears at embryonic day 18 near the fourth ventricle. Relaxin-3 peptide can be detected after birth by immunocytochemistry [23]. This developmental expression pattern is comparable with that of relaxin, the rodent equivalent of human relaxin-2, whose mRNA is not detectable in the rat brain at embryonic day 15, although it is detectable at postnatal day 1 [24]. As well as rodents, the distribution of relaxin-3 in the brain has recently been reported for fish, monkeys and humans. In the zebrafish, the relaxin-3 gene is expressed in two neuron clusters in the brainstem: one is a midbrain cell cluster of the periaqueductal gray matter and the other is in a posterior region that could be homologous to the mammalian NI [25]. Two groups have described the distribution of relaxin-3 in the primate brain. In the brain of Macaca fascicularis, relaxin-3-positive cell bodies were found to be distributed within a ventromedial region of the central gray matter of the pons and medulla, which appears to correspond to the NI in lower species [26]. In the rhesus macaque and humans, relaxin-3 immunostaining was predominantly observed in the ventral and dorsal tegmental nuclei of the brainstem [27]. Thus, from fish to primates, this peptide is expressed in the dorsal tegmentum of the brain stem, corresponding to the NI in rodents.

Regulation of relaxin-3 gene expression

Concerning the regulation of relaxin-3 gene expression, relaxin-3 mRNA expression in the NI is enhanced by restraint stress or forced swim stress (Fig. 2A) [12,28]. This swim stress-induced increase in relaxin-3 transcript levels is blunted by the systemic administration of CRFR1 antagonist [28]. Relaxin-3 transcript levels are also increased after treatment with p-chlorophenylalanine, a potent inhibitor of serotonin synthesis, indicating that serotonin negatively regulates relaxin-3 gene expression [23]. From these results, the expression of relaxin-3 may be observed to be dynamically altered under different physiological conditions. We found that relaxin-3 is expressed in a mouse neuroblastoma cell line, Neuro2a, and investigated the intracellular signaling that leads to activation of relaxin 3 gene transcription in vitro [29]. Using a clone stably-transfected with a relaxin-3 promoter-enhanced green fluorescent protein gene, we observed that the increase in intracellular cAMP induced by dibutyryl cAMP and forskolin treatment increased relaxin-3 promoter activity. These increases were inhibited by pretreatment with the protein kinase A (PKA) inhibitors, H89 and KT5720. Moreover, the relaxin-3 promoter activity was enhanced by CRF treatment after the expression of CRFR1 receptor in the cells. These results suggest that relaxin-3 transcription in vivo is activated via the cAMP-PKA pathway, which is downstream of CRFR1 [29] (Fig. 2B).

Figure 2.

 (A) Relaxin-3 mRNA expression in the NI after 6 h of restrained stress. The upper panel shows a representative image of in situ hybridization using the [35S]-labeled probe. The graph below indicates the calculated signal intensity of relaxin-3 mRNA. Data are shown as the mean ± SD of photostimulated luminescence (PSL) [12]. (B) A schematic representation of the intracellular signaling that regulates relaxin-3 gene expression. Downstream of CRFR1, the cAMP-PKA pathway is involved in the activation of relaxin-3 gene transcription.

The function of relaxin-3 in the brain

Because relaxin-3-producing cells showed a relatively limited distribution, predominantly in neurons of the NI, the function of this peptide has been assessed based upon anatomical studies of the NI at the neuronal level [14,15]. The NI is composed of two subdivisions, the pars compacta and pars dissipata, and relaxin-3-positive neurons are found in both regions (Fig. 1A). With reference to the distribution of relaxin-3-positive nerve fibers and RXFP3 and RXFP1 expression, several functions of relaxin-3 in the brain have been demonstrated, including those related to neuroendocrine processes, stress response, water intake and spatial memory [12,16,28,30–35]. Particularly, this peptide also regulates food intake, as well as other hypothalamic peptides described in this minireview series [36,37].

Stress response

The NI is a region showing abundant expression of CRFR1, and strong c-Fos induction was observed in the NI in response to an intracerebroventricular injection of CRF [38,39]. It is well known that CRF is expressed in parvocellular neurons of the PVN and, during the stress response, CRF activates the hypothalamic-pituitary-adrenal (HPA) axis, acting at CRFR1 on anterior pituitary corticotropes to stimulate the release of adrenocorticotropic hormone. There are also extrahypothalamic CRF-expressing neurons distributed through the brain in areas such as the neocortex and limbic regions, including the central amygdala and hippocampus [40,41]. The regulation of CRF expression may be involved in setting the ‘tone’ of stress-related behavior, including anxiety, as well as learning and memory [42,43]. CRF exerts its actions via two major receptors: CRFR1 and CRFR2. Both receptors belong to the class B subtype of G protein-coupled receptors, although they have a different distribution, suggesting that the two receptors have different functions. CRFR1 is considered to be involved in the acute phase of the stress response, whereas CRFR2 contributes to the maintenance and recovery phase that involves a gradual reduction of HPA axis activation [43,44].

In the rat NI, almost all relaxin-3-positive neurons coexpress CRFR1 and respond to CRF intracerebroventricular administration. Moreover, application of a water-restraint stress for 2–4 h induces c-Fos expression and leads to an increase in relaxin-3 mRNA levels in the NI [12]. On the other hand, relaxin-3-positive neurons project fibers to the hypothalamus, and RXFP3 is intensely expressed in the PVN where hypothalamic CRF neurons exist. These results suggest that relaxin-3-expressing neurons respond immediately to stress and modulate the HPA axis. Recently, Banerjee et al. [28] reported that exposure of rats to a repeated forced swim for 10 min each time leads to a marked increase in relaxin-3 mRNA levels in the NI at 30–60 min after the second swim. Systemic treatment with the CRFR1 antagonist alarmin 30 min before the second swim blunted the stress-induced effect on relaxin-3 transcripts in the NI [28]. This supports the idea that relaxin-3-expressing neurons in the NI (and therefore relaxin-3) play a role in the central stress regulating system by mutual interaction with CRF-expressing neurons.

Food intake

Relaxin-3 was first reported to stimulate food intake when administered into the third ventricle or PVN of male Wistar rats. Administration of human relaxin-3, but not human relaxin-2, either intracerebroventricularly (180 pmol) or intra-PVN (18 pmol) increased 1-h food intake both in the early light and early dark phase (Fig. 3) [31]. The doses of relaxin-3 required to elicit a significant feeding response are in the picomolar range and are similar to the effective doses of other orexigenic peptides such as ghrelin (30 pmol; intra-PVN) and neuropeptide Y (NPY) (78 pmol; intra-PVN) [45,46]. Although RXFP3 and RXFP1 are expressed in the PVN, relaxin (specifically, human relaxin-2) binds RXFP1 but not RXFP3, suggesting that this feeding-promoting action of relaxin-3 is exerted through RXFP3 because the actions of relaxin have not been reported to include hyperphagia, but do include hemodynamic effects such as increasing arterial blood pressure and vasopressin release [47], or dipsogenesis [48]. In reverse, relaxin-3 was recently reported to facilitate water intake as well as relaxin, suggesting that RXFP1 was involved in this action [35]. Concerning the chronic administration of relaxin-3, intracerebroventricular injection for 14 days (600 pmol·day−1) using osmotic minipumps led to a significant increase in food consumption and weight compared to vehicle infusion. There was no difference in locomotor activity between two groups either in the light phase or dark phase, suggesting that this effect of relaxin-3 is not a result of increased locomotor or arousal activity [34]. Chronic intra-PVN administration of human relaxin-3 (180 pmol twice a day for 7 days) also increased the cumulative food intake in ad libitum-fed rats [32]. After such chronic administration, the plasma concentration of leptin and insulin was significantly increased [32]. In addition to the PVN, relaxin-3-administration into the SON or arcuate nucleus (ARC), but not into the LH, stimulated 1-h food intake [32]. The ARC and LH are well known as feeding centers where orexigenic peptides such as NPY, melanin-concentrating hormone and orexin are distributed. Although relaxin-3-immunoreactive fibers are densely distributed, the RXFP3 level is relatively low in the ARC and LH. An electrophysiological study of neurons in these hypothalamic nuclei may help to resolve this disparity and clarify the hyperphagic mechanisms.

Figure 3.

 Effect of intracerebroventricular administration of relaxin-3 in satiated male Wistar rats. (A) Effect of human relaxin-3 (H3) (18–180 pmol) on 1-h food intake. *P < 0.05 versus vehicle in the early light phase. (B) Effect of H3 (18–180 pmol) on cumulative food intake over 4 h in the early light phase. &P < 0.05 at 18 pmol versus vehicle; *P < 0.05 at 54 pmol versus vehicle; #P < 0.05 at 180 pmol versus vehicle. Reproduced with permission [31]; © 2005, The Endocrine Society).

Recently, relaxin-3 gene knockout (KO) mice of mixed background (129S5:B6) were examined in two studies. One group reported that KO mice are smaller and leaner than congenic controls [21], although the results obtained by the second group indicated that there was no genotypic difference in body weight or motor coordination [49]. Further studies using relaxin-3 KO mice backcrossed to C57/B6 should help to clarify the role of relaxin-3 in regulating body weight and metabolism.

Actions of relaxin-3 at the hypothalamo-pituitary-gonadal (HPG) axis

Recently, a role of relaxin-3 in regulation of the HPG axis was reported in that intracerebroventricular (5 nmol) and intra-PVN (540–1620 pmol) administration of relaxin-3 in adult male rats significantly increased plasma luteinizing hormone levels. This effect was inhibited by pretreatment with a peripheral gonadotropin-releasing hormone (GnRH) antagonist. By contrast, the central administration of human relaxin-2 was not found to influence the plasma luteinizing hormone concentration. Using hypothalamic explants and GT1-7 cells that express RXFP1 and RXFP3, relaxin-3 was shown to dose-dependently stimulate GnRH release. GnRH neuronal cell bodies are found in several forebrain regions, including the medial septum, diagonal band, preoptic area and LH, where relaxin-3-positive fibers and RXFP3 are moderately-to-densely distributed [12,17,50]. These results suggest that relaxin-3 regulates the HPG axis via hypothalamic GnRH neurons. Thus, relaxin-3 is seen to belong to the group of neuropeptides that regulate energy homeostasis and reproduction (i.e. modulate both appetite and the HPG axis). This group includes NPY, orexin and galanin-like peptides [51–54].


In this minireview, relaxin-3, which is the latest member of the insulin/relaxin family, is described in terms of its gene transcript and peptide expression in the brain, as well as its functional aspects that have thus far been reported. Although relaxin-3-expressing neurons show a confined distribution in the brainstem, being particularly dense in the NI of the dorsal tegmental pons, their fibers and receptors (i.e. RXFP3 and RXFP1) are widely distributed in the forebrain. One of the target areas of relaxin-3 is the hypothalamus. Relaxin-3 is considered to have various actions mediated through receptors in the hypothalamus, including effects on the stress response, feeding and neuroendocrine function.


The present work was supported by a grant (no. 21500329) to M.T. from the Ministry of Education, Culture, Sports, Science and Technology, Japan.