Dual Role of Insulin in Spexin Regulation: Functional Link Between Food Intake and Spexin Expression in a Fish Model

Abstract

Spexin (SPX), a neuropeptide discovered by the bioinformatics approach, has been recently identified as a satiety factor in a fish model. However, the functional link between feeding and SPX expression as well as the signal transduction for SPX regulation are totally unknown. In this study, we used goldfish as a model to examine the functional role of insulin as a postprandial signal for SPX regulation in bony fish. In goldfish, feeding could elevate plasma levels of glucose, insulin, and SPX with concurrent rises in insulin and SPX messenger RNA (mRNA) expression in the liver. Similar elevation in SPX mRNA level was also observed in the liver and brain areas involved in appetite control in goldfish after intraperitoneal injection of glucose and insulin, respectively. In parallel experiments with goldfish hepatocytes and brain cell culture, insulin signal induced by glucose was shown to exert a dual role in SPX regulation, namely (1) acting as an autocrine/paracrine signal to trigger SPX mRNA expression in the liver and (2) serving as an endocrine signal to induce SPX gene expression in the brain. Apparently, the peripheral (in the liver) and central actions of insulin (in the brain) on SPX gene expression were mediated by insulin receptor (to a lesser extent by insulin-like growth factor I receptor) coupled to mitogen-activated protein kinase kinase 3/6/p38 mitogen-activated protein kinase and phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin but not mitogen-activated protein kinase kinase 1/2/extracellular signal-regulated kinase 1/2 cascades. Our findings indicate that an insulin component inducible by glucose is present in the liver of the fish model and may serve as the postprandial signal linking food intake with SPX expression both in the central as well as at the hepatic level.

Spexin (SPX), the protein product of the C12orf39 gene in humans, is a novel peptide first identified by bioinformatics prior to its purification/functional studies (1, 2). It is a 14–amino acid peptide with C-terminal α-amidation, and its protein sequence is highly conserved from fish to mammals (3–5). In representative species, for example, in human (6), rat (7), mouse (2), and, more recently, in fish models (3–5, 8), SPX is widely expressed at the tissue level, suggesting that the peptide may have pleiotropic functions. This idea is consistent with the findings that SPX is involved/implicated in gastrointestinal motility (1, 9), feeding and energy balance (4, 5), luteinizing hormone release (8), adrenocortical cell proliferation (10), nociception and cardiovascular/renal function (11), as well as O₂ sensing in the carotid body (12). In humans, low levels of serum SPX can be associated with insulin and lipid resistance (13), especially for type 2 diabetes (6) and both adult (14) and childhood obesity (15), which has aroused the interest in developing SPX-based analogs with pharmacological/therapeutic potential (16). Recently, SPX has been shown to activate galanin type 2 and 3 receptors expressed in HEK293 cells (17), suggesting that the two receptors may act as the cognate receptors for SPX. Despite the recent progress on SPX research, there is still a lack of information regarding the endocrine control of SPX expression, and the signal transduction mechanism for SPX regulation is totally unknown.

To shed light on the comparative aspects of SPX in lower vertebrates, especially in fish species, we used goldfish as a model to establish the solution structure of fish SPX by nuclear magnetic resonance and confirm its expression at protein level within the brain by liquid chromatography–tandem mass spectrometry (4). Using a combination of in vivo and in vitro approaches, SPX expression in brain areas involved in appetite control was shown to be elevated after food intake, and SPX treatment via differential regulation of orexigenic and anorexigenic signals within these brain areas (by increasing CART, CCK, and POMC with a concurrent drop in NPY and AgRP expression) could inhibit feeding behavior and food consumption in goldfish (4), indicating that SPX can act as a novel satiety factor in the fish model. Recently, the association of SPX expression with feeding has also been reported in fish species with commercial value, for example, grouper (3) and Ya-fish (5), which raises the possibility of using SPX as a new target for manipulating/improving body growth and energy balance in cultured fish. Although our study in goldfish has clearly demonstrated a central component of SPX expression in appetite control, the possibility for peripheral input of SPX was not examined. Furthermore, the functional link between food intake and SPX expression as well as the underlying mechanisms for SPX regulation are still unclear and remain to be investigated.

Because insulin is a major component of glucose homeostasis after feeding, and by itself also acts as a satiety factor in mammals through modulation of orixegenic/anorexigenic signals (18), it raises the possibility that insulin may play a role in SPX regulation. In this study, we sought to test the hypothesis that glucose-induced insulin release/production caused by food intake might serve as a postprandial signal linking feeding with SPX expression in goldfish. Whole-animal experiments were conducted to investigate the effects of (1) food intake and (2) intraperitoneal (IP) injection of glucose and insulin, respectively, on SPX expression in the liver and/or brain areas known to be involved in appetite control in goldfish. In vitro studies were also performed in goldfish hepatocytes and brain cell culture to elucidate the receptor specificity as well as postreceptor signaling for SPX regulation by insulin using a combination of pharmacological approach and direct monitoring of activation status for signaling targets. In this study, we have provided evidence for the dual role of insulin as both the endocrine as well as autocrine/paracrine signal triggered by feeding for SPX regulation and unveiled the signal transduction mechanisms for SPX expression both in the central level (the brain) as well as at the peripheral level (the liver).

Materials and Methods

Animal and test substances

Goldfish (Carassius auratus) with 20 to 35g body weight were purchased from local pet stores and maintained in well-aerated 650-L aquaria at 20°C under 12-hour light/12-hour dark photoperiod for >14 days prior to in vivo experiments/harvesting tissue for cell culture. Because the fish used in our study were sexually regressed and sexual dimorphism was not apparent, mixed sexes of goldfish were used for in vivo and in vitro experiments. Drug treatment and tissue sampling were conducted as described by the protocols (CULATR no. 3890) approved by the Committee for Animal Use in Teaching and Research at the University of Hong Kong. Human insulin and its antagonist S961 was obtained from Sigma-Aldrich (St. Louis, MO). The inhibitors for insulin receptor (InsR) and/or insulin-like growth factor (IGF) I receptor (IGF1R), including picropodophyllin (PPP), hydroxy-2-naphthalenylmethylphosphonic acid (HNMPA), and N-(5-chloro-2-methoxyphenyl)-N′-(2-methyl-4-quinolinyl)urea (PQ401), and blockers for target signaling kinases, including 1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenylthio]butadiene (U0126), 2-(2-amino-3-methoxyphenyl)4H-1-benzopyranone (PD98059), 5-(2-phenylpyrazolo[1,5-a]pyridin-3-yl)-1H-pyrazolo-[3,4-c]pyridazin-3-amine (FR180204), (3R)-1-[2-oxo-2-[4-[4-(2-pyrimidinyl)phenyl]-1-piperazinyl]-ethyl]-N-[3-(4-pyridinyl)-1H-indazol-5-yl]-3-pyrrolidinecarboxamide (SCH772984), 4-(4-fluorophen-yl)-2-(4-nitrophenyl)-5-(4-pyridyl)-1H-imidazole (PD169316), 4-[5-(4-fluorophenyl)-2-[4-(methyl-sulfonyl)phenyl]-1H-imidazolyl]-pyridine (SB203580), wortmannin, 2-(4-morpholinyl)-8-phenyl-4H-1-benxopyrin-4-one (LY294002), triciribine hydrate (API-2), and rapamycin, were acquired from Calbiochem (San Diego, CA). These test substances were dissolved in phosphate-buffered saline or dimethyl sulfoxide, stored frozen in small aliquots at −80°C, and diluted with prewarmed culture medium to appropriate concentrations 10 minutes prior to drug treatment.

In vivo experiments for postprandial effect of feeding and glucose/insulin treatment

For in vivo experiments, goldfish (8 to 10 fish/group) were acclimated for ≥14 days to single housing in 25-L tanks with a one-meal-per-day feeding schedule (with daily food supply at 10:00 am). To study the postprandial effects of feeding, the fish were divided into two groups, one with a regular supply of food pellets [∼1% body weight (BW), the “fed” group] and the other without food provision (the “unfed” group), and the time for daily administration of fish feed was taken as time 0 (0 hour) for reference. For in vivo treatment with glucose and insulin, IP injection of d-(+)-glucose or human insulin dissolved in fish physiological saline was conducted in goldfish under MS222 anesthesia 5 minutes prior to the regular feeding time but with no food provision after the fish recovered from the anesthesia (which took 3 to 5 minutes for full recovery). Sampling of blood and tissues was performed according to standard procedures (4) at the time points indicated. Circulating levels of glucose, insulin, and SPX were measured using a glucose LiquiColor test (Stanbio Laboratory, Boerne, TX), fish insulin enzyme-linked immunosorbent assay kit (Cusabio Life Science, Wuhan, China), and an SPX fluorescent enzyme immunoassay kit (Phoenix Pharmaceuticals, Phoenix, AZ), respectively. Total messenger RNA (mRNA) isolation, reverse transcription (RT) sample preparation, and subsequent real-time polymerase chain reaction (PCR) for target genes were carried out as described previously (19) with quantitative PCR conditions listed in Supplemental Table 1. Authenticity of PCR products was confirmed by melting curve analysis, and parallel measurement of 18S RNA was used as the internal control. To evaluate insulin expression at tissue level and InsR and IGF1R expression both in the hepatocytes and brain cell culture, RT-PCR was also performed for tissue expression profiling of insulin as well as detection of multiple isoforms of InsR/IGF1R expressed in the respective cell culture with PCR protocols described in Supplemental Table 2. In these studies, RT-PCR for β-actin was used as the quality control for RNA preparation.

In vitro experiments with goldfish hepatocytes and brain cell culture

Goldfish hepatocytes (with viability ≥94%) were prepared by the collagenase digestion method with minor modifications (19) and cultured in serum-free Dulbecco’s modified Eagle medium/F-12 medium (Invitrogen, Carlsbad, CA) at 0.8 × 10⁶ cells/mL/well in 24-well plates. Goldfish brain cells (with viability ≥92%) were prepared from brain areas covering the telencephalon, optic tectum, and hypothalamus by trypsin digestion as described previously (4) and cultured in 35-mm dishes at 3 × 10⁶ cells/2-mL/dish with Neurobasal (Invitrogen) containing B27 supplement (1:50 dilution). Cell cultures for the two target tissues were routinely prepared from tissues harvested from six goldfish and maintained at 28°C under 5% CO₂ and saturated humidity for 15 hours for carp hepatocytes (to avoid cell aggregation caused by prolonged culture) and 7 days for brain cells (to allow for neuron recovery and glial cell attachment) prior to initiation of drug treatment. After drug treatment, total RNA was extracted and subjected to real-time PCR for target gene measurement as described in the preceding section. Real-time PCR for 18S RNA and β-actin were also conducted to serve as the internal control for hepatocyte and brain cell culture, respectively. For in vitro studies on SPX secretion, goldfish hepatocytes were cultured at 15 × 10⁶ cells/2-mL/dish in a 60-mm dish and exposed to drug treatment in the presence of a 1× dilution of Complete™ protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). After that, culture medium harvested was concentrated using C18 Sep-Pak chromatography (Waters, Tokyo, Japan) and used for SPX measurement with an SPX fluorescent enzyme immunoassay kit.

Western blot of kinase activation induced by insulin treatment

To examine the postreceptor signaling for insulin-induced SPX gene expression, Western blot was performed in cell lysate prepared from goldfish hepatocytes or brain cell culture after insulin treatment as described previously (20) using antibodies specific for the phosphorylated form (P-) and total protein (T-) of mitogen-activated protein kinase kinase (MEK)_1/2 (1:1000), extracellular signal-regulated kinase (ERK)_1/2 (1:5000), mitogen-activated protein kinase kinase (MKK)_3/6 (1:2000), p₃₈ mitogen-activated protein kinase (P₃₈^MAPK) (1:1000), phosphatidylinositol 3-kinase (PI3K) (1:2000), and Akt (1:1500), respectively. (Table 1 provides details of these antibodies and their respective suppliers.) In these studies, parallel blotting of β-actin was used as the loading control. Quantitation of the results of Western blot was performed by densitometric scanning using ImageJ (https://imagej.nih.gov/ij/), and the data were transformed as a ratio of the phosphorylated form over the total protein of the respective targets.

Data transformation and statistics

For real-time PCR, standard curves with a dynamic range of ≥10⁵ and correlation coefficient of ≥0.95 were constructed for data calibration with serial dilutions of plasmids carrying the open reading frame of target genes. The raw data of mRNA expression for in vivo experiments (in terms of femtomoles of transcript detected per assay tube) were normalized against 18S RNA expression detected in the same sample and transformed as a percentage of the mean value for target gene expression in control group at time 0 (i.e., as “% Ctrl” of the reference group at 0 hour). For in vitro experiments with cell culture, the raw data of mRNA expression were measured in terms of femtomoles of transcript/10⁶ cells/well. These data were normalized against β-actin mRNA in the case of brain cell culture and transformed as a percentage of the control group without drug treatment (as % Ctrl). Because 18S RNA levels (the internal control for hepatocytes) did not exhibit notable changes in our in vitro experiments, the raw data for target gene expression in goldfish hepatocytes were simply transformed as a percentage of the mean value in the control (as % Ctrl) or fold increase vs the control group (as “fold induction”) without 18S RNA normalization . Data presented for plasma levels of glucose and insulin as well as for real-time PCR of target gene expression are expressed as mean ± standard error of the mean (SEM) (n = 8 to 10 for in vivo experiments and 4 to 8 for in vitro experiments). Data for time-course studies were analyzed with two-way analysis of variance followed by a Bonferroni post hoc test with time and treatment as the two variables for statistical analysis. The data for dose-dependence/drug treatment to elucidate postreceptor signaling were analyzed with one-way analysis of variance followed by a Newman–Keuls test. Differences between groups were considered as significant at P < 0.05.

Results

Functional role of insulin in SPX regulation in vivo

In goldfish entrained with a one-meal-per-day feeding schedule, significant increases in glucose, insulin, and SPX levels were observed in the plasma with peaks at 1 hour after initiation of feeding (at time 0) when compared with the time-matched unfed groups [Fig. 1(A–C)]. In the unfed groups, however, plasma levels of glucose, insulin, and SPX were reduced in a time-dependent manner with respect to their preprandial basal levels. Pearson analysis of the data for plasma samples also revealed that a positive correlation could be noted between plasma levels of insulin and glucose [Fig. 1(D)], SPX and glucose [Fig. 1(E)], and SPX and insulin [Fig. 1(F)]. In goldfish, as revealed by RT-PCR, insulin was found to be widely expressed in different tissues (e.g., the brain, gills, heart, gut, spleen, kidney, liver, muscle, and gonads) and brain areas (e.g., the olfactory bulb, telencephalon, optic tectum, hypothalamus, pituitary, cerebellum, medulla oblongata, and spinal cord), with the highest level of insulin signal detected in the liver [Fig. 2(A)]. In the same feeding experiment, besides the rises in plasma insulin and SPX, rapid elevation of insulin [Fig. 2(B)] and SPX mRNA levels [Fig. 2(C)] could also be observed in the liver, with peaks at 1 hour after food provision, suggesting that the liver may serve as a source of insulin and SPX in circulation.

In a parallel experiment, IP injection of goldfish with d-(+)-glucose could elevate plasma levels of glucose [Fig. 3(A)] with concurrent rises of insulin and SPX mRNA expression in the liver [Fig. 3(B) and 3(C)]. The treatment, however, did not induce notable changes in hepatic expression of IGF-I and IGF-II transcripts (Supplemental Fig. 1A). After IP injection of d-(+)-glucose, rapid rises of SPX mRNA levels were also observed within the first 30 minutes in brain areas involved in feeding control, including the hypothalamus [Fig. 3(D)], telencephalon [Fig. 3(E)], and optic tectum [Fig. 3(F)]. To examine the role of insulin in SPX regulation, IP injection of human insulin was also performed in goldfish. Insulin of human origin was used because (1) fish insulin with bioactivity is not commercially available, and (2) human insulin is highly homologous to its goldfish counterpart (84% in the B-domain and 78% in the A-domain, which form the mature peptide of insulin). In our study, IP injection with human insulin was shown to reduce plasma levels of glucose in goldfish [Fig. 4(A)], indicating that human insulin is fully functional in the fish model. Meanwhile, plasma levels of SPX [Fig. 4(B)] as well as SPX mRNA expression in the liver [Fig. 4(C)] and brain areas, including the hypothalamus [Fig. 4(D)], telencephalon [Fig. 4(E)], and optic tectum [Fig. 4(F)], were found to be elevated shortly after insulin treatment. To evaluate the functional role of insulin in glucose-induced SPX release and gene expression, IP injection of d-(+)-glucose was tested with cotreatment of the insulin antagonist S961. Similar to our time-course study, IP injections of glucose consistently elevated plasma glucose [Fig. 5(A)] and SPX levels [Fig. 5(B)], with parallel rises in SPX mRNA expression in the liver [Fig. 5(C)] and brain areas, including the hypothalamus [Fig. 5(D)], telencephalon [Fig. 5(E)], and optic tectum [Fig. 5(F)]. These stimulatory effects, however, were found to be suppressed/totally abolished by S961 cotreatment.

SPX regulation by insulin in goldfish hepatocytes

To shed light on SPX regulation by insulin at the hepatic level, primary culture of hepatocytes was prepared from goldfish (Supplemental Fig. 2A), and validation based on growth hormone–induced IGF-I mRNA expression also confirmed that the cell culture was responsive to hormone stimulation under serum-free conditions (Supplemental Fig. 2B), which is highly desirable for subsequent investigation of insulin action. As revealed by RT-PCR, expression of the receptors for insulin (including InsR₁ and InsR₂) and IGF-I (only for IGF1R₂ but not for IGF1R₁) could be detected in the goldfish liver as well as hepatocyte culture [Fig. 6(A)], implying that the hepatocytes can act as the regulatory target for the two endocrine factors. Under serum-free conditions, static incubation of goldfish hepatocytes with d-(+)-glucose was found to elevate insulin [Fig. 6(B)] and SPX mRNA levels [Fig. 6(C)] in a dose-related manner without notable changes in IGF-I and IGF-II gene expression (Supplemental Fig. 1B). Parallel studies with insulin treatment also triggered a time-dependent [Fig. 6(D)] and dose-dependent elevation [Fig. 6(F)] in SPX mRNA expression, but similar treatment with glucagon was found to have no effects [Fig. 6(E)].

In mammals, insulin and IGF-I are known to cross-react at the receptor level (21). In our study, the receptor specificity for SPX regulation by insulin was also examined at the hepatic level. In goldfish hepatocytes, insulin induction was effective in triggering rapid phosphorylation of InsR [Fig. 7(A)] and IGF1R [Fig. 7(B)], with peaks at 15 minutes after drug treatment. At hepatocyte level, insulin-induced SPX mRNA expression was significantly suppressed by the InsR inhibitor HNMPA (by 83%), but similar treatment with the IGF1R blocker PPP could only induce a mild/marginal inhibition (by 16%) for the corresponding response [Fig. 7(C)]. However, the stimulatory effect of insulin could be totally abolished by PQ401, an inhibitor known to block InsR and IGR1R activation at the dose examined. In separate studies, cotreatment with the insulin antagonist S961 or InsR inhibitor HNMPA was also effective in blocking the stimulatory effect of d-(+)-glucose on SPX gene expression in goldfish hepatocytes [Fig. 7(D)], and again, a marginal inhibition was noted with parallel inhibition using the IGF1R blocker PPP (data not shown). To shed light on the effect of InsR activation on SPX secretion at the hepatic level, a high-density culture system supplemented with Complete™ protease inhibitor cocktail was established for goldfish hepatocytes. Based on our validation, the protease inhibitors did not alter the viability or overall morphology of our cell culture (Supplemental Fig. 3A) and had no notable effect on insulin, SPX and β-actin gene expression (Data not shown). Using this high-density cell culture, a 2-hour treatment with d-(+)-glucose (Supplemental Fig. 3B) or insulin (Supplemental Fig. 3C) was found to elevate SPX release, and the stimulatory effect by d-(+)-glucose could be negated by cotreatment with the InsR inhibitor HNMPA (Supplemental Fig. 3D).

Given that InsR is known to be functionally coupled with MAPK and PI3/Akt pathways (22), the possible involvement of these postreceptor signaling cascades in SPX expression induced by insulin was also examined. Similar to the rapid phosphorylation of InsR, insulin treatment was effective in triggering protein phosphorylation of MEK_1/2, ERK_1/2, MKK_3/6, P₃₈^MAPK, PI3K, and Akt in goldfish hepatocytes, with peaks at 15 minutes after drug administration [Fig. 8(A–C)]. In parallel studies with insulin treatment in the presence of inhibitors for various signaling targets, SPX mRNA expression caused by insulin stimulation was not affected by the MEK_1/2 inhibitor U1026 or PD98059 [Fig. 9(A)] or ERK_1/2 blocker FR180204 or SCH772984 [Fig. 9(B)]. However, the SPX responses triggered by insulin could be reduced by simultaneous incubation with the P₃₈^MAPK inhibitor PD169316 or SB203580 [Fig. 9(C)], PI3K inactivator LY294002 or wortmannin [Fig. 9(D)], Akt blocker API-2, or mammalian target of rapamycin (mTOR) inhibitor rapamycin [Fig. 9(E)]. Furthermore, insulin-induced SPX gene expression could be totally abolished by exposing goldfish hepatocytes to the P₃₈^MAPK inhibitor PD169316 and PI3K inactivator LY294002 at the same time [Fig. 9(F)].

SPX regulation by insulin in goldfish brain cells

Because insulin treatment, similar to glucose challenge, could elevate SPX gene expression in vivo in the telencephalon, hypothalamus, and optic tectum of the goldfish, goldfish brain cell culture was prepared from these brain areas (4) and used as a cell model to study the central actions of insulin on SPX regulation. In this cell culture, B27 supplement (with 4 mg/L insulin; Invitrogen) was used as a serum substitute to maintain a serum-free culture condition. To examine the effect of removing insulin intrinsic to the cell culture system, the serum substitute was replaced with an insulin-free preparation of B27 (Invitrogen). In this case, basal expression of SPX mRNA was reduced to a low level but with no effect on β-actin gene expression [Fig. 10(A)]. In brain cells cultured with normal preparation of B27, insulin treatment was found to have the opposite effect and elevate SPX mRNA levels in a time-related [Fig. 10(B)] and dose-related fashion [Fig. 10(C)], with peak responses observed at 12 hours after initiation of drug treatment.

As revealed by RT-PCR, transcript signals of the receptors for insulin (including InsR₁ and InsR₂) and IGF-I (including IGF1R₁ and IGF1R₂) could be detected in the brain as well as brain cell culture prepared from the goldfish [Fig. 10(D)], implying that the central nervous system (CNS) is a regulatory target for insulin and IGF-I in the fish model. In goldfish brain cells, SPX mRNA expression induced by insulin could be totally negated by the InsR inhibitor HNMPA, whereas the parallel blockade with IGF1R inactivator PPP could only yield a partial inhibition [∼50%, Fig. 10(E)]. Regarding the postreceptor signaling for insulin induction, similar to the results in hepatocytes, insulin was effective in triggering rapid phosphorylation of MEK_1/2, ERK_1/2, MKK_3/6, P₃₈^MAPK, PI3K, and Akt in brain cell culture [Fig. 11]. Parallel experiments also revealed that SPX mRNA expression induced by insulin was not sensitive to the blockade by the MEK_1/2 inhibitor PD98059 or ERK_1/2 blocker FR180204 [Fig. 12(A)]. In contrast, the SPX responses could be obliterated by cotreatment of insulin with the P₃₈^MAPK inhibitor PD169316, PI3K inactivator LY294002, Akt blocker API-2, and mTOR inhibitor rapamycin [Fig. 12].

As shown in our tissue profiling for insulin expression [Fig. 2(A)], insulin signal was also located in the brain of goldfish, especially in the hypothalamus, telencephalon, and optic tectum. These findings raise the possibility that insulin produced locally within the brain may play a role in SPX regulation by glucose in the CNS. In our in vivo experiment, however, IP injection of d-(+)-glucose consistently increased plasma insulin and insulin mRNA levels in the liver (positive control of the study) but with no effect on insulin gene expression in hypothalamus, telencephalon, and optic tectum (Supplemental Fig. 4A). In a parallel study with goldfish brain cell culture, similarly, treatment with d-(+)-glucose up to 24 hours was found to be ineffective in altering basal levels of insulin and SPX mRNA expression (Supplemental Fig. 4B). These results suggest that insulin expression within the brain of goldfish is not responsive to glucose changes in circulation. Probably, insulin from circulation rather than blood glucose represents the major signal for SPX regulation within the CNS.

Discussion

SPX, a novel peptide identified by bioinformatics (1), has recently emerged as a regulatory factor with pleiotropic functions (see the Introduction for details). Because its association with energy balance and obesity has been reported in human and animal models (14), and plasma levels of SPX can be used as a marker for childhood obesity (15) and type 2 diabetes (6), studies on SPX have aroused a lot of interest in the field. The research on SPX has become even more interesting with our recent study in goldfish revealing that central expression of SPX induced by food intake could act as a satiety signal to inhibit feeding behavior in a fish model via modulation of orexigenic/anorexigenic factors expressed in brain areas involved in appetite control (4). Given that insulin, a key component of glucose homeostasis, is responsive to feeding, and by itself also acts as an anorexigenic factor for signal integration of satiation and body adiposity (23), we postulate that it may be involved in SPX expression in the fish model. In our feeding experiment, food intake could elevate plasma levels of glucose, insulin, and SPX in goldfish with concurrent increase in insulin and SPX gene expression in the liver. Our results demonstrated that, besides the central expression of SPX, feeding could also induce a peripheral component of SPX, presumably through elevation of SPX production in the liver followed by SPX release into systemic circulation. Because (1) glucose-induced insulin release via ATP-sensitive K⁺ channels (24) and sweet taste–sensing receptors (25) coupled to mitochondrial signal amplification (26) is well documented, and (2) a positive correlation was noted for plasma levels of glucose, insulin, and SPX in our feeding experiment, it raises the possibility that glucose-induced insulin signals caused by feeding might play a role in SPX regulation. This idea is supported by our findings that (1) IP injection of glucose in goldfish could elevate plasma insulin and hepatic expression of insulin and SPX mRNA levels, and (2) parallel treatment with insulin not only could induce SPX gene expression in the liver but also increase plasma levels of SPX. In both cases, SPX mRNA levels in brain areas involved in appetite control (i.e., hypothalamus, telencephalon, and optic tectum) were also increased by glucose and insulin induction. Given that glucose-induced SPX secretion and SPX gene expression in the brain and liver were also sensitive to in vivo blockade by the insulin antagonist S961, it would be logical to conclude that the insulin signal caused by postprandial rise in blood glucose can contribute to the SPX responses within the CNS and at the hepatic level in a fish model.

In bony fish, endocrine pancreas exists as Brockman bodies scattering in the gut mesentery, for example, in tilapia (27) and wolfish (28), whereas exocrine pancreas (mainly in the form of glandular acinus) can also spread into the liver, leading to the formation of hepatopancreas in some species (29). Although hepatopancreas has been reported in goldfish (30), the presence of an endocrine component of the pancreas has not been documented at the hepatic level. In our study, distinct structure for Brockman bodies could not be identified in the goldfish, but the highest level of insulin expression was located in the liver. Furthermore, our experiments with IP injection of glucose in vivo and glucose treatment in hepatocytes in vitro have clearly shown that insulin signals not only could be detected at the hepatic level but also inducible by glucose with a parallel increase in SPX gene expression. In goldfish hepatocytes, glucose and insulin induction were both effective in elevating SPX mRNA levels, and glucose-induced SPX secretion and gene expression could be negated by insulin antagonist S961 and/or InsR inhibitor HNMPA. These findings, taken together, suggest that (1) an insulin component responsive to glucose induction is present in the goldfish liver, and (2) local production of insulin induced by glucose may act in an autocrine/paracrine manner to trigger SPX release/expression at the hepatic level. In mammals, autocrine/paracrine actions of insulin have also been reported in the brain and involved in cognitive functions related to memory, learning, and food anticipation (31). However, the local action of insulin produced at the hepatic level has not been reported. Because glucagon did not alter SPX gene expression in our hepatocyte culture, it would be logical to assume that local production of glucose (e.g., by glycogenolysis/gluconeogenesis) may not have a major role in SPX regulation. Judging from the fact that the liver represents the largest organ in the body and serves as a major target for insulin, especially for glucose (32) and lipid homeostasis (33), the hepatopancreas in goldfish with an inducible component of insulin may also act as a source of insulin for both local actions as well as endocrine functions via blood circulation. Although Brockman bodies could not be identified in goldfish, we do not exclude the possibility that postprandial rise in blood glucose could also induce insulin release from goldfish pancreas, which may also contribute to SPX regulation in the liver as well as in brain areas involved in appetite control.

Of note, glucose treatment, both in vivo and in vitro, did not alter hepatic expression of IGF-I/IGF-II mRNA levels in our studies, the possible involvement of local production of IGFs for SPX regulation is rather unlikely. In mammals, insulin is known to cross-react with IGF1R (21), and InsR and IGF1R are functionally coupled with MAPK and PI3K/Akt pathways (34). In some tissues (e.g., placenta), InsR and IGF1R can also form hybrid receptors (35) with higher affinity for IGF-I (36). In bony fish, isoforms of InsR (37) and IGF1R (38) have been cloned and are thought to be the result of whole-genome duplication occurring during the evolution of the teleost lineage (39). In goldfish hepatocytes, two forms of InsR (InsR₁ and InsR₂) and one form of IGF1R (IGF1R₂) could be identified by RT-PCR, and insulin exposure was found to induce rapid phosphorylation of InsR and IGF1R, suggesting that the receptors for insulin and IGF-I may play a role in mediating insulin functions at the hepatic level. Given that (1) the InsR inhibitor HNMPA was notably more effective (83% blockade) than the IGF1R inhibitor PPP (16% blockade) in attenuating insulin-induced SPX mRNA expression in goldfish hepatocytes, and (2) the SPX responses caused by insulin were totally abolished by PQ401, an inhibitor known to inactivate both InsR and IGF1R at the dose tested, it is conceivable that the two receptors are both involved in SPX regulation at hepatic level, with InsR as the dominant form mediating insulin action. Regarding the postreceptor signaling mechanisms for SPX regulation, insulin was capable of triggering MEK_1/2, ERK_1/2, MKK_3/6, P₃₈^MAPK, PI3K and Akt phosphorylation in goldfish hepatocytes. These results are comparable with the recent report on InsR signaling in trout adipocytes (40) and consistent with the idea that InsR signaling via MAPK and PI3K/Akt cascades is well conserved in vertebrate evolution (41). In goldfish hepatocytes, however, insulin-induced SPX mRNA expression was not affected by the inhibitors for MEK_1/2 (U0126 and PD98059) or ERK_1/2 (FR18024 and SCH772984), but it was attenuated by the inhibitors for P₃₈^MAPK (PD169316 and SB20380), PI3K (LY294002 and wortmannin), Akt (API-2), and mTOR (rapamycin). Besides, the SPX responses caused by insulin could be totally negated by cotreatment with the inhibitors for P₃₈^MAPK (PD169316) and PI3K (LY294002), suggesting that the MKK_3/6/P₃₈^MAPK and PI3K/Akt/mTOR but not MEK_1/2/ERK_1/2 cascades are involved in SPX gene expression induced by insulin at the hepatic level in the carp model.

As mentioned in the preceding section, insulin signals induced by glucose could also trigger SPX gene expression in brain areas involved in appetite control, which might initiate the satiation response after feeding in goldfish via differential regulation of orexigenic/anorexigenic signals within the CNS. This idea is comparable to the model in mammals with central regulation of energy balance and body weight via modulation of feeding regulators (e.g., NPY and POMC) within the hypothalamus by satiety signals (e.g., leptin) (42). In our study, the effect of insulin on central expression of SPX was further substantiated by in vitro experiments with goldfish brain cells, in which (1) removing insulin from the cell culture system (using insulin-free B27) could suppress the basal level of SPX mRNA, and (2) insulin treatment, in contrast, could upregulate SPX gene expression in a time- and dose-related manner. Our findings are also in agreement with the previous reports (e.g., in rodents) on (1) insulin uptake into the CNS by transcytosis across the choroid plexus via InsR coupled to LRP2/megalin (43), and (2) the role of insulin as an anorexigenic factor via its central actions in modulating NPY, AgRP, POMC, and MCR4 expression (44). Apparently, SPX can serve as a novel component for the central actions of insulin on appetite control. Because insulin expression within the brain in goldfish was not responsive to glucose in vivo and in vitro, it is likely that insulin from circulation rather than blood glucose represents the key signal triggering SPX expression within the CNS in the fish model. Regarding the receptor specificity for insulin’s action within the CNS, the results of RT-PCR revealed that multiple isoforms of InsR (InsR₁ and InsR₂) and IGF1R (IGF1R₁ and IGF1R₂) could be detected in the whole brain as well as brain cell culture prepared from the goldfish. Similar to goldfish hepatocytes, insulin-induced SPX mRNA expression in brain cells was found to be more sensitive to the blockade by the InsR inhibitor HNMPA (∼100%) compared with similar treatment with the IGF1R inhibitor PPP (∼50% blockade). These results again suggest that the central effect of insulin on SPX expression is mediated mainly by InsR and to a lower extent by IGF1R. Unlike in the brain, IGF1R₁ was undetectable in goldfish liver/hepatocytes, and IGF1R₂ was the only IGF1R identified at the hepatic level, implying that different isoforms of IGF1R are expressed in a tissue-specific manner. Whether it can also contribute to the tissue-specific actions of IGFs as recently reported for insulin/IGF-I signaling in body metabolism and aging in rodents (45) is unclear and remains to be investigated. Although insulin was effective in triggering protein phosphorylation of MEK_1/2, ERK_1/2, MKK_3/6, P₃₈^MAPK, PI3K, and Akt in goldfish brain cells, SPX mRNA expression induced by insulin in the cell culture was not sensitive to the blockade of MEK_1/2 (e.g., by PD98059) and ERK_1/2 (e.g., by FR180204) but could be totally abolished by inhibiting P₃₈^MAPK (e.g., by PD169316), PI3K (e.g., by LY294002), Akt (e.g., by API-2), and mTOR (e.g., by rapamycin), respectively. These findings as a whole confirm that, similar to the case of goldfish hepatocytes, the central effect of insulin on SPX regulation is also mediated by MKK_3/6/P₃₈^MAPK and PI3K/Akt/mTOR but not MEK_1/2/ERK_1/2 cascades.

In summary, using goldfish as an animal model, we have demonstrated that insulin signals induced by a postprandial rise in blood glucose could serve as the functional link between food intake and SPX expression both in the central (i.e., within CNS) and peripheral (i.e., in the liver) levels. In our study, we also found that the liver in goldfish represents a major site of insulin expression, which is also inducible by changes in glucose levels. In our working model (Fig. 13), food intake in goldfish can induce insulin expression in the liver via elevation in blood glucose (probably with a parallel rise in insulin signal from the pancreas). Local production of insulin acts in an autocrine/paracrine manner to increase SPX expression at the hepatic level via InsR (to a lesser extent via IGF1R) coupled to the MKK_3/6/P₃₈^MAPK and PI3K/Akt pathways. Hepatic output of SPX also elevates SPX level in circulation and constitutes a peripheral SPX signal acting on the CNS. Meanwhile, insulin released from the liver (together with pancreatic insulin) can elevate insulin level in blood and serve as an endocrine signal to induce SPX expression in brain areas involved in feeding behavior. Apparently, the central effect of insulin on SPX expression is also mediated through InsR coupled with MKK_3/6/P₃₈^MAPK and PI3K/Akt signaling. Because our previous study with intracerebroventricular injection of SPX in goldfish confirmed that (1) SPX by acting centrally could inhibit feeding behavior and food intake, and (2) these inhibitory effects were mediated by SPX inhibition of NPY and AgRP, with concurrent stimulation of CART, CCK, and POMC expression in brain areas involved in appetite control (4), local production of SPX within the CNS together with the SPX input from the periphery induced by insulin presumably can inhibit feeding behavior in goldfish via differential actions on orexigenic/anorexigenic signals within the feeding circuitry in the brain. Our study, as a whole, has provided new insights on the mechanisms and signal transduction for the role of SPX as a satiety factor in a fish model. Of note, plasma SPX is closely associated with glucose homeostasis (e.g., in diabetic patients) (6) and fatty acid uptake/body adiposity (e.g., in obese mice) (14). The role of circulating SPX induced by feeding in energy balance and glucose/lipid metabolism will certainly be an interesting topic for further study. Recently, type 2 and 3 galanin receptors have been proposed to be the cognate receptor for SPX, and, interestingly, the type 3 galanin receptor could not be found in the genome of fish species(e.g., zebrafish) (17). Research is now under way in our laboratory to study the receptor specificity as well as the postreceptor signaling mechanisms for SPX-induced satiation and regulation of orexigenic/anorexigenic factors expressed within the brain in goldfish.

Abbreviations:

 
• API-2

  triciribine hydrate

 
• BW

  body weight

 
• CNS

  central nervous system

 
• ERK

  extracellular signal-regulated kinase

 
• FR180204

  5-(2-phenylpyrazolo[1,5-a]pyridin-3-yl)-1H-pyrazolo-[3,4-c]pyridazin-3-amine

 
• HNMPA

  hydroxy-2-naphthalenylmethylphosphonic acid

 
• IGF

  insulin-like growth factor

 
• IGF1R

  insulin-like growth factor I receptor

 
• InsR

  insulin receptor

 
• IP

  intraperitoneal(ly)

 
• LY294002

  2-(4-morpholinyl)-8-phenyl-4H-1-benxopyrin-4-one

 
• MAPK

  mitogen-activated protein kinase

 
• MKK

  mitogen-activated protein kinase kinase

 
• mRNA

  messenger RNA

 
• mTOR

  mammalian target of rapamycin

 
• PCR

  polymerase chain reaction

 
• PD169316

  4-(4-fluorophen-yl)-2-(4-nitrophenyl)-5-(4-pyridyl)-1H-imidazole

 
• PD98059

  2-(2-amino-3-methoxyphenyl)4H-1-benzopyranone

 
• PI3K

  phosphatidylinositol 3-kinase

 
• P₃₈^MAPK

  p38 mitogen-activated protein kinase

 
• PPP

  picropodophyllin

 
• PQ401

  N-(5-chloro-2-methoxyphenyl)-N′-(2-methyl-4-quinolinyl)urea

 
• RT

  reverse transcription

 
• RT-PCR

  reverse transcription–polymerase chain reaction

 
• SB203580

  4-[5-(4-fluorophenyl)-2-[4-(methyl-sulfonyl)phenyl]-1H-imidazolyl]-pyridine

 
• SCH772984

  (3R)-1-[2-oxo-2-[4-[4-(2-pyrimidinyl)phenyl]-1-piperazinyl]-ethyl]-N-[3-(4-pyridinyl)-1H-indazol-5-yl]-3-pyrrolidinecarboxamide

 
• SEM

  standard error of the mean

 
• SPX

  spexin

 
• U0126

  1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenylthio]butadiene

Acknowledgments

This article is dedicated to Prof. J. P. Chang (University of Alberta, Edmonton, Alberta, Canada) for his genuine interest in training young scientists in the field of comparative endocrinology. We also thank Dr. Chen Ting for his help in setting up the goldfish hepatocyte culture and tissue sampling and processing for the feeding experiment.

This work was supported by General Research Fund Grants 17117716, 17128215, 781113, and 780312 and National Science Foundation of China/Research Grants Council of Hong Kong Joint Grant N_HKU 732/12 from the Research Grant Council (Hong Kong) and by Health and Medical Research Fund Grant 13142591 from the Food and Health Bureau (Hong Kong Special Administrative Region). This work was also supported in part by School of Biological Sciences, University of Hong Kong, in the form of postgraduate studentships to A.M., J.B., and M.K.H.W.

Disclosure Summary: The authors have nothing to disclose.

References

Author notes