Localization of relaxin‐like gonad‐stimulating peptide expression in starfish reveals the gonoducts as a source for its role as a regulator of spawning

Abstract Oocyte maturation and gamete release (spawning) in starfish are triggered by relaxin‐like gonad‐stimulating peptide (RGP), a neuropeptide that was first isolated from the radial nerve cords of these animals. Hitherto, it has generally been assumed that the radial nerve cords are the source of RGP that triggers spawning physiologically. To investigate other sources of RGP, here we report the first comprehensive anatomical analysis of its expression, using both in situ hybridization and immunohistochemistry to map RGP precursor transcripts and RGP, respectively, in the starfish Asterias rubens. Cells expressing RGP precursor transcripts were revealed in the ectoneural epithelium of the radial nerve cords and circumoral nerve ring, arm tips, tube feet, cardiac stomach, pyloric stomach, and, most notably, gonoducts. Using specific antibodies to A. rubens RGP, immunostaining was revealed in cells and/or fibers in the ectoneural region of the radial nerve cords and circumoral nerve ring, tube feet, terminal tentacle and other arm tip‐associated structures, body wall, peristomial membrane, esophagus, cardiac stomach, pyloric stomach, pyloric caeca, and gonoducts. Our discovery that RGP is expressed in the gonoducts of A. rubens proximal to its gonadotropic site of action in the gonads is important because it provides a new perspective on how RGP may act as a gonadotropin in starfish. Thus, we hypothesize that it is the release of RGP from the gonoducts that triggers gamete maturation and spawning in starfish, while RGP produced in other parts of the body may regulate other physiological/behavioral processes.

Accordingly, there is evidence that a combination of external stimuli triggers spawning, including changes in day length, the lunar cycle, water temperature, and pheromones released by conspecifics (Byrne et al., 1997;Lawrence, 2013;Pearse & Walker, 1986;. Sensory detection of these external stimuli then leads to activation of internal endocrine and/or neural gonadotropic mechanisms that regulate gamete maturation and expulsion of gametes from gonads into the surrounding seawater. Evidence of the existence of a gonadotropic neurohormone in starfish was first reported in 1959 with the discovery that extracts of radial nerve cords from the starfish Asterias forbesi induce spawning when injected into the coelom of reproductively mature male or female starfish (Chaet & McConnaughy, 1959). Subsequent efforts to determine the molecular identity of the active component of radial nerve cord extracts revealed it to be a peptide, which became known as gonad-stimulating substance or gamete-shedding substance (GSS).
Furthermore, purification of GSS from several starfish species enabled determination of its amino acid composition and estimation of its molecular mass (Chaet, 1966a(Chaet, , 1966bKanatani et al., 1971). However, it was not until 2009 that the molecular structure of GSS was finally determined (Mita, Yoshikuni, et al., 2009). GSS was purified from extracts of radial nerve cords from the starfish Patiria pectinifera and found to be a heterodimer comprising an A-chain and a B-chain.
These two chains are linked by three disulfide bridges, with two interchain bridges between the A-chain and B-chain, and an intrachain bridge within the A chain. Furthermore, sequencing revealed that the A-chain and B-chain have unique cysteine motifs that are a feature of the insulin/insulin-like growth factor (IGF)/relaxin superfamily, and phylogenetic analysis revealed that GSS is more closely related to the relaxin subfamily than the insulin/IGF subfamily. Therefore, GSS was classified as an invertebrate member of the insulin/IGF/relaxin superfamily and was renamed as relaxin-like gonad-stimulating peptide (RGP) (Mita, Yoshikuni, et al., 2009). RGP has subsequently been identified in other starfish species, including Asterias amurensis, (Mita, Daiya, et al., 2015), Acanthaster planci cf. A. solaris (Mita, Ikeda, et al., 2015;Smith et al., 2019), Aphelasterias japonica , Asterias rubens (Lin, Mita, et al., 2017;Semmens et al., 2016), and Astropecten scoparius (Mita, Osugi, et al., 2021), revealing evolutionary conservation of both its structure and function. Relaxin-type peptide precursors have also been identified in other echinoderms, including ophiuroids (brittle stars), holothurians (sea cucumbers), and crinoids (feather stars) (Aleotti et al., 2022;Chieu et al., 2019;Suwansa-Ard et al., 2018;Zandawala et al., 2017). Furthermore, a relaxin-type peptide triggers spawning in the sea cucumber Holothuria scabra, showing that the gonadotropic action of relaxin-type peptides also extends to other echinoderms .
Prior to the molecular identification of RGP in starfish, its mechanism of action as a regulator of oocyte maturation and spawning was investigated. This revealed that it binds to receptors expressed by follicle cells that surround oocytes and stimulates production of 1-methyladenine, which then triggers germinal vesicle breakdown and oocyte maturation (Hirai & Kanatani, 1971;Kanatani et al., 1969;Mita, 2019). RGP exerts its effects via G-protein-dependent cyclic adenosine monophosphate (cAMP) synthesis in follicle cells (Mita et al., 1989(Mita et al., , 2011(Mita et al., , 2012, but the biochemical pathway by which this leads to production of 1-methyladenine is not known (Mita, 2019).
Although GSS/RGP was first isolated from starfish radial nerve cords, use of a variety of experimental techniques has revealed that this gonadotropic peptide is also present in other tissues/organs in starfish, albeit at lower concentrations than in the radial nerve cords. Thus, use of a bioassay for gonadotropic activity revealed the presence of low concentrations of GSS in tube feet, body wall, and cardiac stomach in A. amurensis and P. pectinifera (Kanatani & Ohguri, 1966). A more recent bioassay analysis of extracts of P. pectinifera tissues/organs revealed high concentrations of GSS activity in radial nerves and circumoral nerve rings and low concentrations of GSS activity in cardiac stomach and tube feet, but no GSS activity was detected in pyloric caeca, ovaries, or testes (Mita, Ito, et al., 2009). With the identification of RGP, quantification of transcript abundance in various tissues by qPCR revealed high levels in radial nerve cords, lower levels in cardiac stomach and pyloric caeca, and trace levels in tube feet, ovaries, or testes (Haraguchi et al., 2016). More recently, specific antibodies to RGP have been generated  and used for immunoassays, enabling quantification of RGP in the radial nerve cords and circumoral nerve ring of P. pectinifera. However, using immunoassays RGP was not detected in cardiac stomach, pyloric stomach, pyloric caeca, tube feet, ovaries, or testes in P. pectinifera (Mita & Katayama, 2018;Yamamoto et al., 2017). Transcriptomic and proteomic analysis of a variety of tissues/organs from the crown-of-thorns starfish A. planci cf. A. solaris has revealed the presence of RGP precursor transcripts and/or RGP in the radial nerve cords, tube feet, terminal tentacle, body wall spines, and gonads. Furthermore, it is noteworthy that RGP precursor transcripts were also detected in coelomocytes (Jonsson et al., 2022;Smith et al., 2017).
More specific localization of RGP expression in starfish at a cellular level has been examined using mRNA in situ hybridization. This was initially restricted to analysis of the radial nerve cords, revealing a small population of cells located in the ectoneural epithelial layer in P. pectinifera (Mita, Yoshikuni, et al., 2009). Subsequently, a more comprehensive analysis of RGP expression was performed by applying mRNA in situ hybridization in the starfish A. rubens. Consistent with findings from P. pectinifera, cells expressing RGP transcripts were revealed in the ectoneural epithelium of the radial nerve cords of A.
rubens. In addition, cells expressing RGP transcripts were revealed in the circumoral nerve ring and tube feet (Lin, Mita, et al., 2017). Furthermore, an extensive population of cells expressing RGP transcripts was revealed in the arm tips of A. rubens, largely concentrated in the body wall external epithelium surrounding the sensory terminal tentacle and the associated optic cushion (simple eye). Informed by this new insight into the anatomy of RGP expression in starfish, it was proposed that the arm tips, and not the radial nerve cords, may be the physiological source of RGP that triggers spawning in starfish (Lin, Mita, et al., 2017). The rationale for this hypothesis was that RGP-expressing cells in the arm tip being located proximal to the terminal tentacle and associated sensory organs are ideally positioned to integrate sensation of changes in environmental conditions (e.g., day length, lunar cycle, pheromones) thought to trigger spawning. However, it is not known if the RGP-expressing cells in the arm tips have axonal processes that terminate at sites whereby RGP released by these cells could gain access to the coelomic cavity where the gonads are located. To address this question, immunohistochemical methods need to be employed and this has recently become feasible with development of specific antibodies to RGP Yamamoto et al., 2017). Consistent with patterns of RGP expression revealed by in situ hybridization, use of immunohistochemical methods has revealed RGP-immunoreactive cells and processes in the ectoneural region of the radial nerve cords in P. pectinifera (Yamamoto et al., 2017). However, a more extensive analysis of RGP expression in starfish using immunohistochemical methods has yet to be reported. Here, we have combined use of both mRNA in situ hybridization and immunohistochemistry to investigate comprehensively the anatomical expression pattern of RGP in the starfish A.
rubens. This has been facilitated by generation of specific antibodies to Asterias RGP, which has an identical structure in A. amurensis and A.
rubens (Katayama et al., 2019;Mita, Elphick, et al., 2021). were fed ad libitum on mussels (Mytilus edulis). After collection of starfish, animals used here for analysis of RGP expression were fixed immediately or only kept in the aquarium for less than a week prior to fixation.

Localization of AruRGPP transcripts in A. rubens using in situ hybridization
Digoxigenin-labeled RNA antisense probes complementary to AruRGP precursor (AruRGPP; GenBank accession numbers KT601728 and ALJ99970) transcripts and corresponding sense probes were synthesized, as reported previously (Lin, Mita, et al., 2017). Because a second relaxin-type peptide precursor (ArRLPP2) has been identified in A. rubens (Semmens et al., 2016), we compared the sequences of the AruRGPP and ArRLPP2 transcripts to assess potential hybridization of the AruRGPP antisense probes with ArRLPP2 transcripts, and this revealed only 52% nucleotide sequence identity in the coding regions.
To investigate AruRGP expression throughout the body, small specimens (∼4 cm in diameter) were analyzed. Starfish of this size were used so that sections of arms or central disks could be collected on microscope slides, but it was not possible to determine the gender of these animals. However, to investigate AruRGP expression in the reproductive system, V-shaped regions of the body at the junction between adjacent arms that incorporated a pair of gonoducts and gonads were dissected from larger animals (∼15 cm in diameter; one male and two females). Whole starfish or body regions were then fixed in 4% paraformaldehyde (PFA; Sigma-Aldrich, Gillingham, UK) for 2 days at 4 • C and then whole starfish were dissected to separate arms from the central disc region. After decalcification with Morse's solution (10% sodium citrate and 20% formic acid in autoclaved water) for 8 h (replenished every 2 h), the decalcified body parts were washed in autoclaved water and dehydrated through an ethanol series (50%, 70%, 90%, and 100%). Body parts were cleared in xylene in two steps (5 min followed by 8 min) and then embedded in molten paraffin wax.
Sectioning of blocks at 13 μm thickness was performed using a microtome (RM2145, Leica Microsystems, Milton Keynes, UK) and then sections were mounted on SuperFrost Plus microscope slides (VWR, Lutterworth, UK).
Slides were placed in an oven at 60 • C for 1 h and then xylene was used to remove wax, followed by rehydration through a descending ethanol series (100%, 90%, 70%, 50%, and 30%; 7 min in each step). After washing in phosphate-buffered saline (PBS; pH 7.3; prepared using disodium phosphate, monosodium phosphate, and sodium chloride purchased from VWR) for 3 × 5 min and postfixation in 4% PFA/PBS for 20 min, slides were incubated with proteinase K (Novagen, an Affiliate of Merck KGaA, Darmstadt, Germany) solution at a concentration of 10 μg/mL in a buffer containing 50 mM Tris-HCl and 6.25 mM EDTA for 20 min at 37 • C. After washing in PBS, slides were acetylated for 10 min in 1.325% triethanolamine, 0.25% acetic anhydride, and 0.175% acetic acid (VWR) made up in distilled water and mixed well with stirring. Slides were then washed in PBS and 5× saline sodium citrate (SSC; prepared using sodium chloride and sodium citrate purchased from VWR) buffer (1 × 5 min) at room temperature.
Before performing probe hybridization, slides were prehybridized in hybridization buffer (50% formamide; 5× SSC; 500 μg/mL yeast total RNA; 50 μg/mL heparin; 0.1% Tween-20 in distilled water) in a humid chamber for 2 h at room temperature. A total of 1000 ng/mL DIGlabeled antisense or sense RNA probes made up in hybridization buffer were denatured by heating at 80 • C for 2 min and then were applied to slides (150 μL/slide). Slides were then covered with parafilm and incubated for 48 h at 50 • C.
Next, slides were placed in warm 5× SSC to remove the parafilm and then washed in 0.2× SSC (2 × 40 min at 50 • C; 1 × 10 min at room temperature) followed by a wash in buffer B1 (10 mM Tris-HCl, pH 7.5; 150 mM NaCl in autoclaved water) for 10 min at room temperature.
Then slides were blocked with 5% goat serum (Sigma-Aldrich) diluted in B1 buffer in a humid chamber for 2 h at room temperature. After the blocking step, slides were incubated with alkaline phosphataseconjugated anti-DIG antibody (Roche Diagnostics GmbH, Mannheim, Germany) at 1:2000 dilution in 2.5% goat serum/B1 buffer in a humid chamber overnight at 4 • C.

Antibody characterization
The generation and characterization of the AruRGP antiserum used here for immunohistochemistry, as described below in Section 2.4, have been reported previously and, using western blotting and an enzyme-linked immunosorbent assay (ELISA), the specificity and sensitivity of the antibodies to AruRGP were determined (Katayama et al., 2019;Mita, Elphick, et al., 2021). This revealed that the antibodies to AruRGP do not cross-react with Patiria pectinifera RGP (PpeRGP), which shares 64% amino acid identity with AruRGP. By way of comparison, a second relaxin-type peptide (ArRLP2) that has been identified in A. rubens (Semmens et al., 2016)

Localization of AruRGP expression using immunohistochemistry
To investigate AruRGP expression in A. rubens using immunohistochemistry, animals of varying size and maturity were analyzed. For general investigation of expression throughout the starfish body, whole juvenile starfish (∼1 cm in diameter) and arms and the central disc from specimens of A. rubens with a diameter of ∼4 cm were analyzed. Starfish of this size were used so that sections of arms or central disks could be collected on microscope slides, but it was not possible to determine the gender of these animals. However, to enable analysis of expression in the reproductive system, V-shaped regions of the body at the junction between adjacent arms that incorporated a pair of gonoducts and gonads were dissected from larger animals (∼15 cm in diameter; two male and two female specimens).
Whole starfish or regions of the body were fixed in Bouin's solution (75 mL saturated picric acid in seawater, 25 mL 37% formaldehyde, 5 mL acetic acid) at 4 • C for 3 days. Following decalcification in 4% ascorbic acid and 0.3 M sodium chloride (1:1 solution) for 2 weeks at 4 • C with regular changes, whole starfish or body parts were embedded in paraffin wax, sectioned at 10 μm using a microtome (RM2145, Leica Microsystems, Milton Keynes, UK), and mounted on chrome alumgelatin-coated glass slides. Wax was removed from sections using xylene (3 × 10 min at room temperature) and slides were placed in 100% ethanol (2 × 10 min). After incubating slides in 1% hydrogen peroxide in methanol for 30 min to quench endogenous peroxidase, slides were rehydrated through a graded series of ethanol (90%, 70%, and 50%; 10 min for each step) into distilled water. The slides were washed once in PBS, once in PBST, and then were incubated with 5% normal goat serum (Sigma-Aldrich; diluted in PBST) for 2 h to block nonspecific antibody-binding sites in tissue sections. Slides were then incubated overnight with the rabbit AruRGP antiserum, which was used at a dilution of 1:16,000 (whole arm or central disk sections) or 1:2000 (body region containing reproductive organs). After washing slides with PBST (5 × 10 min), slides were incubated with secondary antibodies (goat anti-rabbit horseradish peroxidase conjugated immunoglobulins [Jackson ImmunoResearch via Stratech Scientific, Newmarket, Suffolk, UK] diluted 1:500 in 2% normal goat serum/PBST) for 3 h. Immunostaining was visualized using diaminobenzidine (VWR) and after intense immunostaining was observed, the reaction was terminated by washing in distilled water (2 × 10 min). Following dehydration through an ethanol series (50%, 70%, 90%, and 2 × 100%; 10 min each), slides were cleared in xylene (2 × 10 min) and mounted with coverslips using DPX Mountant (VWR).

Imaging
An Infinity Analyse Camera (Teledyne Lumenera INFINITY5-5C) attached to a Leica DMRA2 light microscope and Infinity Analysis 7 software running on an iMac computer (27-inch with OS X Yosemite, v. 10.10) were used to capture photographs of sections. Scale bars in images were added using ImageJ Fiji (Schindelin et al., 2012) and images were compiled into montages and labeled using Adobe Photoshop CC2020 (San Jose, CA).  Tinoco et al., 2021Tinoco et al., , 2018Zhang et al., 2020Zhang et al., , 2022. Here, AruRGP was tested at a concentration of 1 μM for myoexcitatory (contraction) and myoinhibitory (relaxation) effects. To facilitate detection of potential relaxing effects, preparations were precontracted with 30 mM KCl (cardiac stomach) or 10 μM acetylcholine (tube feet). The SALM-Famide neuropeptide S2 (1 μM; custom synthesized by Peptide Protein

In vitro pharmacology
Research Ltd, Fareham, UK) was also tested on cardiac stomach preparations as a positive control for myorelaxant activity (Melarange & Elphick, 2003).

RESULTS
Expression of AruRGPP and AruRGP in A. rubens was revealed using mRNA in situ hybridization and immunohistochemistry, respectively, as described in detail below. To facilitate interpretation of the patterns of staining reported here, the anatomy of starfish is illustrated in Figure 1.

3.1.1
Radial nerve cords, tube feet, and arm tips

Digestive system
The digestive system of A. rubens includes the highly folded and evertible cardiac stomach, which is linked by a short tubular esophagus to the mouth located on the underside of the central disk. Aboral to the cardiac stomach is the much smaller pyloric stomach, which is linked via pyloric ducts to paired digestive organs (pyloric caeca) located in each arm (Anderson, 1954;Jangoux, 1982)  AruRGPP expression was observed in the pyloric caeca.
3.1.3 Reproductive system The reproductive system of A. rubens comprises 10 gonads (ovaries or testes), with a pair of gonads located in each of the five arms. Each gonad is linked via a short gonoduct to the body wall of the arm, proximal to its junction with the central disk (refer to Figure 1b for anatomy).
The gonoduct perforates the body wall and opens to the external environment via several gonopores. The luminal lining of the gonoduct comprises an epithelial layer with cross-ridges, which span the lumen of the duct and imbricate, partially blocking the lumen (Walker, 1974(Walker, , 1975.

In vitro pharmacology
No myoexcitatory (muscle contraction) or myoinhibitory (muscle relaxation) effects of AruRGP on cardiac stomach or tube feet preparations were observed ( Figure S2).

F I G U R E 5
Immunohistochemical localization of AruRGP in the nervous system of Asterias rubens. (a, b) AruRGP-immunoreactivity in transverse sections of radial nerve cord from the regions of an arm proximal to the arm tip (a) or proximal to the circumoral nerve ring (b).
The inset of panel a shows absence of immunostaining in a transverse section of the radial nerve cord incubated with AruRGP antiserum that had been preabsorbed with AruRGP (see also Figure S1f

DISCUSSION
The molecular identification of RGP as a gonadotropin in starfish (Mita, Yoshikuni, et al., 2009) has provided a basis for important advances in our knowledge and understanding of the reproductive neuroendocrinology of echinoderms. For example, it has been discovered that an RGP-like peptide also triggers spawning in sea cucumbers, demonstrating that this role of relaxin-type peptides is not unique to starfish . However, there is still much to be learnt about the physiological mechanisms by which RGP acts as regulator of spawning in starfish and in other echinoderms.
Because RGP was originally isolated from extracts of starfish radial nerve cords, it has generally been assumed that the radial nerve cords are the physiological source of RGP that triggers spawning in starfish (Mita, 2019). However, as discussed in more detail below, there is no direct evidence in support of this hypothesis. Furthermore, prior to the molecular identification of RGP, it was already known that extracts of other tissues/organs contain GSS/RGP-like bioactivity, including tube feet, cardiac stomach, and body wall of A. amurensis, albeit at much lower concentrations than in the radial nerve cords (Kanatani & Ohguri, 1966). These findings have been confirmed following the molecular characterization of RGP, as summarized in Table 1. Thus, using PCR and/or immunoassay methods, RGP expression has been detected in the radial nerve cords, circumoral nerve ring, cardiac stomach, pyloric caeca, pyloric stomach, and tube feet of P. pectinifera (Haraguchi et al., 2016;Mita & Katayama, 2018;Mita, Ito, et al., 2009;Mita, Yoshikuni, et al., 2009;Yamamoto et al., 2017). However, a limitation of these biochemical detection methods is their lack of resolution anatomically. Therefore, in this study, both mRNA in situ hybridization and immunohistochemistry were used to enable the first comprehensive analysis of the expression of RGP in starfish, using the common European species A. rubens as an experimental model and as summarized in Table 1.
Below, we discuss the physiological significance of our findings and, in particular, how they provide insights into the mechanisms by which RGP acts as a regulator of spawning in starfish.

Functional significance of RGP expression in the central nervous system
The distribution of cells expressing RGP precursor transcripts in the radial nerve cords has been reported previously in A. rubens (Lin, Mita, et al., 2017) and in P. pectinifera (Mita, Yoshikuni, et al., 2009). It is noteworthy that, by comparison with many other neuropeptide types, the relative abundance of cells expressing RGP is quite low. Thus, many other neuropeptide types are expressed in both the ectoneural and the hyponeural regions of the radial nerve cords in A. rubens (Cai et al., 2018(Cai et al., , 2023Lin et al., 2018;Tian et al., 2017;Tinoco et al., 2021;Zhang et al., 2020Zhang et al., , 2022, whereas expression of RGP is restricted to the ectoneural region. Furthermore, the number of cells expressing RGP in the ectoneural region is quite low, with bilaterally symmetrical clusters of only two to three stained cells revealed in transverse sections of radial nerve cords. This contrasts with many other neuropeptide types (e.g., ArCT, ArCRH, ArPPLN1b, ArPPLN2h) where larger numbers of stained cells are revealed in transverse sections of radial nerve cords (Cai et al., 2018(Cai et al., , 2023Cobb, 1985;Lin et al., 2018;. The functional significance of differences in the relative abundance of cells expressing different neuropeptides in the ectoneural region of radial nerve cords is not known. However, use of immunohistochemistry has provided new insights into RGP expression in the radial nerve cords because, unlike with mRNA in situ hybridization, the stained cells and processes Abbreviations: *, detected; BW, body wall; C, coelomocytes; CONR, circumoral nerve ring; CS, cardiac stomach; E, esophagus; ELISA, enzyme-linked immunosorbent assay; G, gonad; GD, gonoduct; IHC, immunohistochemistry; ISH, mRNA in situ hybridization; MN, marginal nerve; MS, mass spectrometry; nd, not detected; ni, not investigated; PCR, polymerase chain reaction; PM, peristomial membrane; PS, pyloric stomach; RIA, radioimmunoassay; RNA-seq, transcriptomic RNA sequencing; RNC, radial nerve cord; RT-qPCR, real-time quantitative PCR; TF, tube foot; TT, terminal tentacle. of hyponeural motoneurons project into and receive presynaptic input from the ectoneural region (Cobb, 1985;Mashanov et al., 2016;Smith, 1937;Zueva et al., 2018) and therefore it is possible that RGP released within the ectoneural neuropile influences whole-animal behavior by modulating the activity of hyponeural neurons. Accordingly, it is noteworthy that in the circumoral nerve ring, RGP-expressing fibers were found to be concentrated in the innermost layer of the ectoneural neuropile, proximal to the hyponeural region.
Prior to the molecular identification of GSS as RGP, it was proposed that GSS is synthesized in supporting cells (radial glia) in the ectoneural region of the radial nerve cord and then transported through supporting fibers to the radial and transverse hemal sinus of the hemal system, via which GSS could in theory reach the gonads (Mita, 2019;Unger, 1962). Now, with the specific visualization RGP in the ectoneural region of the radial nerve cord reported here, it is possible to evaluate this hypothesis neuroanatomically. Our findings show that RGP is not expressed in supporting cells (radial glia) but instead RGP is expressed in neuronal cells with extensive axonal processes that project longitudinally in the ectoneural region of the radial nerve cords. Furthermore, no evidence of RGP-expressing fibers projecting from the ectoneural region to the radial and transverse hemal sinus was observed. Another potential route by which RGP could in theory travel to the gonads from the radial nerve cords is via the perihemal coelomic spaces. However, it seems unlikely that RGP released by neuronal fibers in the ectoneural region could gain access to the perihemal coelom because the hyponeural region separates it from the ectoneural region. Based on these anatomical considerations, we suggest that it is unlikely that the ectoneural region of the radial nerve cords and/or circumoral nerve ring is the physiological source of RGP that triggers gamete maturation and spawning in starfish. However, spawning in starfish involves more than just the release of gametes into the surrounding seawater because it is also accompanied by changes in whole-animal behavior.
Thus, starfish typically adopt a humped posture when spawning, with animals standing on the tips of their arms and thereby bringing the sites of gamete release (gonopores) off the seabed into the water column above, which may facilitate gamete dispersal (Minchin, 1987). It is possible, therefore, that RGP-expressing neurons in the ectoneural region of the radial nerve cords participate in neural mechanisms underlying the adoption and maintenance of this spawning posture.

Functional significance of RGP expression in the tube feet, arm tips, and body wall
We have reported previously the expression of RGP precursor transcripts in the tube feet, arm tips, and associated sensory organs of A.
rubens (Lin, Mita, et al., 2017). By using immunohistochemistry, here we have obtained more detailed insights into the neuroanatomy of RGP-expressing cells in these regions of the starfish body because their axonal processes are visualized using this technique. Thus, in the tube feet, RGP-immunoreactive fibers are revealed in the subepithelial nerve plexus and basal nerve ring. Informed by this pattern of expression, we investigated if RGP affects the contractile state of tube foot preparations in vitro, but we found that it neither causes contraction nor relaxation of tube feet when tested at 10 −6 M.
An extensive system of RGP-expressing cells/fibers has been revealed in the arm tips, with the architecture of immunostained cells and their processes revealed in the terminal tentacle (a mechanosensory organ), the optic cushion (a photosensory organ), the lateral and aboral lappets (presumptive chemosensory organs), and more generally in the epithelium of the body wall cavity that surrounds the terminal tentacle. Because of the extensive expression of RGP in the arm tip and its associated sensory organs, we have proposed previously that the arm tips may be the physiological source of RGP that triggers spawning in starfish, with the rationale being that RGP-expressing cells in the arm tip and associated sensory organs could integrate sensation of changes in environmental conditions (e.g., day length, lunar cycle, pheromones) thought to trigger spawning (Lin, Mita, et al., 2017).
However, if arm tips are the source of RGP that triggers spawning, RGP released by cells in the arm tips would need to gain access to coelomic cavity of the arms where the gonads are located. Because in situ hybridization typically only enables visualization of transcripts in cell bodies, our previous analysis of RGP expression did not enable us to address this issue. Now with the availability of specific antibodies to AruRGP and use of immunohistochemistry, importantly, we did not observe any evidence of the axonal processes of RGP-expressing cells in the arm tips terminating proximal to the coelomic cavity. Therefore, as with RGP-expressing cells in the radial nerve cords and circumoral nerve ring, we conclude that the RGP-expressing cells in the arm tips are unlikely to be the physiological source of RGP that triggers spawning in starfish. However, the axonal processes of RGP-expressing cells in the arm tips may contribute to immunostaining observed in the ectoneural neuropile of the radial nerve cords because there is continuity with the neuropile regions of the terminal tentacle and optic cushion. Thus, while the RGP-expressing cells in the arm tip may not be directly involved in triggering gamete maturation and release, it is possible they are involved in neural mechanisms that initiate and maintain the humped posture that starfish adopt when they spawn.
Another indication that RGP expression in the arm tips may not be specifically associated with physiological mechanisms of RGPinduced gamete maturation and release is our discovery that RGPimmunoreactivity is detected not only in the body wall of the arm tips but also in the body wall in other regions of the arms.

Functional significance of RGP expression in the digestive system
The expression of RGP in the cardiac stomach of starfish has been reported previously based on analysis of tissue extracts for GSS/RGP bioactivity, PCR-based detection of RGP transcripts, or immunoassays (Haraguchi et al., 2016;Mita & Katayama, 2018Mita, Ito, et al., 2009). However, this is the first study to visualize RGP expression in the starfish digestive system using mRNA in situ hybridization and immunohistochemistry. Cells expressing AruRGP precursor transcripts were visualized in the cardiac stomach and pyloric stomach of A. rubens, but these were sparse in number by comparison with many other neuropeptides we have analyzed previously (Cai et al., 2018(Cai et al., , 2023Lin et al., 2018;Tian et al., 2017;Tinoco et al., 2021;Zhang et al., 2020Zhang et al., , 2022. Nevertheless, use of AruRGP antibodies enabled visualization of the axonal processes of these cells, revealing that they contribute extensive immunostaining in the basiepithelial nerve plexus of the cardiac stomach. Furthermore, RGP-immunoreactivity was also observed in other regions of the digestive system in A. rubens, including the peristomial membrane that surrounds mouth and the esophagus that links the mouth and the cardiac stomach. Additionally, although cells expressing RGP were not revealed in the pyloric caeca by mRNA in situ hybridization, immunostaining was observed in the basiepithelial nerve plexus of these digestive organs. Collectively, these findings indicate that RGP may be involved in regulation of feeding and/or digestive physiology in starfish. Our previous investigations have revealed that several neuropeptides expressed in the cardiac stomach of A. rubens are myoactive, causing relaxation or contraction of in vitro preparations of this organ. This is of interest with respect to the extraoral feeding behavior of starfish, where in species such as A. rubens the cardiac stomach is everted out of the mouth over the soft tissues of prey (e.g., mussels) and then is retracted when partial digestion and uptake of prey tissues are completed (Anderson, 1954;Lawrence, 2013). For example, the vasopressin/oxytocin-type neuropeptide asterotocin causes cardiac stomach relaxation in vitro and eversion in vivo (Odekunle et al., 2019) and the neuropeptide NGFFYamide causes cardiac stomach contraction in vitro and retraction in vivo (Semmens et al., 2013;Tinoco et al., 2018). Because AruRGP is expressed in the cardiac stomach of A.
rubens, here we investigated if it has in vitro pharmacological effects on this organ and found that it causes neither relaxation nor contraction of cardiac stomach preparations. Therefore, AruRGP is presumably involved in regulation of other physiological processes in the digestive system of A. rubens.

Functional significance of RGP expression in the gonoducts
Previous studies, using bioassays, radioimmunoassays, or ELISAs, have investigated if RGP is expressed in starfish gonads and these have typically reported that RGP expression is not detected (Table 1). Accordingly, our analysis of A. rubens of using mRNA in situ hybridization and immunohistochemistry revealed no evidence of RGP expression in gonads. However, we have for the first time discovered that RGP is expressed in the gonoducts, tubular shaped structures that link the gonads to the gonopores. The small size and inaccessibility of the gonoducts probably explain why previous investigations using biochemical assays have not investigated or detected RGP expression in these organs.
Detailed descriptions of the anatomy of the gonoducts in Asterias have been reported previously (Walker, 1975). A distinctive feature of the gonoducts is the presence of cross-ridges in the luminal epithelium that occlude the lumen and are thought to act as a barrier to prevent gametes from exiting the gonad to the gonopore prior to spawning.
Interestingly, use of in situ hybridization revealed AruRGP precursor transcripts in cells located basal to the luminal epithelium at a position proximal to the junction of the gonoduct with the gonopore. Furthermore, immunohistochemical analysis of AruRGP expression revealed extensive immunoreactivity in the basiepithelial nerve plexus located beneath the luminal epithelium of the gonoducts, in both male and female animals. These observations indicate that there is a population of RGP-expressing neurons located in the wall of the gonoducts with axonal processes that account for immunostaining observed in the basiepithelial nerve plexus of the gonoduct. Importantly, these findings provide a new perspective on the mechanisms by which RGP acts as a regulator of spawning in starfish. Thus, because the gonoducts are located proximal to the gonads and the lumen of the gonoducts is continuous with the lumen of the gonads where immature gametes are located, the RGP-expressing neurons in the gonoducts are excellent candidates as physiological sources of the RGP that triggers gamete maturation and release in starfish. As illustrated in Figure 9, we hypothesize that prior to spawning, RGP-expressing neurons in the gonoducts release RGP and this diffuses as a neurohormone into the lumen of the gonoduct and then into the lumen of the gonads. Then, informed by findings from previous studies, in female animals the RGP binds to G-protein-coupled receptors expressed by follicle cells that surround immature oocytes, which then results in cAMP-dependent production of 1-methyladenine. 1-Methyladenine then binds to 1-methyladenine receptors located in the membrane of immature oocytes and this then triggers a downstream signaling cascade that causes germinal vesicle breakdown and oocyte maturation (Mita, 2019). A parallel pathway is thought to occur in male starfish where RGP binds to receptors on interstitial cells, which produce 1-methyladenine that then triggers spermatocyte maturation and spawning (Kubota et al., 1977). Following gamete maturation, waves of contraction caused by the muscular wall of the gonads then expel gametes into the gonoducts, where breakdown of the cross-ridges of the luminal lining facilitates rapid and effective expulsion of gametes to the external environment via the gonopores.

Evolution and comparative physiology of relaxin-type neuropeptides as regulators of reproductive physiology
Lastly, it is of interest to reflect on the findings of this study from a comparative and evolutionary perspective, taking into account what is known about the physiological roles of relaxin-type peptides in mammals. The hormone relaxin was first discovered as a substance present in the serum of pregnant guinea pigs or rabbits that causes relaxation of the pubic ligament of virgin guinea pigs. Subsequently, relaxin was found to be derived from corpus luteum in pigs, the placenta in rabbits, and the uterus in guinea pigs. Acting as an endocrine regulator, relaxin-type peptides cause softening and hypertrophy of the pubic symphysis, cervix, uterus, and vagina during the second half of F I G U R E 9 Schematic diagrams showing the anatomy of RGP signaling in starfish, with the gonoduct as a source of RGP and the gonad as a site of action. (a) Diagram showing a model of our hypothesis that prior to spawning in starfish, RGP molecules released from nerve processes in the gonoduct diffuse into lumen of the gonoduct and then into the lumen of the adjoining gonad (ovary), where they can bind to cell surface receptor proteins on follicle cells surrounding immature oocytes. (b) Diagram showing how binding of RGP to RGP receptors on follicle cells stimulates formation of 1-methyladenine, which then binds to receptors on oocytes to trigger oocyte maturation (germinal vesicle breakdown). (c) Detail of signal transduction pathway in follicle cells, in which binding of RGP to a G-protein-coupled receptor triggers G-protein (G s )-mediated stimulation of adenylyl cyclase activity and elevation of intracellular cyclic adenosine monophosphate, which then triggers formation of 1-methyladenine via as yet unknown molecular mechanisms. α, alpha subunit of G-protein; ATP, adenosine triphosphate; β, beta subunit of G-protein; BW, body wall; Cr, cross-ridged epithelium of gonoduct; γ, gamma subunit of G-protein; Gd, gonoduct; GDP, guanosine diphosphate; GTP, guanosine triphosphate; 1-MeAde, 1-methyladenine; O, ovary; RGP, relaxin-like gonad-stimulating peptide; RGPR, RGP receptor. The diagram in panel a is adapted from figure 1 in Walker (1975) and was drawn in Photoshop CC2020 (San Jose, CA) using the digital pen of a creative drawing tablet (Wacom, Tokyo, Japan), which was connected to a laptop (MacBook Pro, Apple, USA). The diagrams in panels b and c are based upon figure 2 in Mita (2019) and were created using BioRender (https://www.biorender.com/). pregnancy, which prepares the maternal body to be ready for parturition. In male humans and rats, the prostate gland is a source of relaxin-type peptides, and relaxin gene-knockout mice exhibit poor growth of the reproductive tract, which is associated with reduced fertility (Bathgate et al., 2013(Bathgate et al., , 2006. In the context of this wellestablished evidence that the reproductive system is both a source and a site of action of relaxin-type peptides in mammals, our discovery that the reproductive system is both a source (gonoduct) and a site of action (gonads) for RGP in starfish is noteworthy. It suggests that relaxin-type peptide production and action in the context of reproductive physiology is an evolutionarily ancient phenomenon that can be traced back to the deuterostome common ancestor of vertebrates and echinoderms. Furthermore, with the recent discovery of a gene encoding a relaxin-type peptide in crinoids (e.g., feather stars), which are a sister group to eleutherozoan echinoderms (e.g., starfish, brittle stars, sea urchins, sea cucumbers) (Aleotti et al., 2022), there may be opportunities to obtain further insights into the comparative physiology of relaxin-type peptides as regulators of reproductive processes.

AUTHOR CONTRIBUTIONS
All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

ACKNOWLEDGMENTS
We are grateful to Phil Edwards for his help in obtaining starfish and to Ian Sanders for maintaining our seawater aquarium in the School of Biological & Behavioural Sciences at QMUL.