Ghrelin Modulates Differential Expression of Genes Relevant to Immune Activities and Antimicrobial Peptides in Primary Head Kidney Cells of Rainbow Trout (Oncorhynchus mykiss)

Simple Summary Ghrelin is a peptide hormone that regulates food intake and energy balance, but also acts as a cytokine in many species to modulate various immune activities. Multiple ghrelin isoforms were discovered in rainbow trout, but the roles of most isoforms are undefined. This study examined the modulation of immune genes by two major ghrelin variants; the originally identified isoform and its three amino acids truncated isoform. After treating primary cells from rainbow trout head kidney with the two ghrelin analogs, the expression patterns of selected immune-related genes were profiled over time by reverse transcriptase-coupled qPCR. The results suggest that the two ghrelin isoforms exerted divergent modulatory effects on various genes at different time points. These results provide further insight into the direct actions of ghrelin as a critical component of the rainbow trout immune system. Abstract Ghrelin is a peptide hormone/cytokine that regulates metabolic processes and plays essential roles in the immune system. To evaluate the immunomodulatory actions of ghrelin isoforms in rainbow trout (RT), an in vitro model was utilized with primary cells isolated from fish head kidney (HKD). These RT-HKD cells were treated with synthetic rainbow trout ghrelin and its truncated isoform, desVRQ-ghrelin, over time (0, 2, 4, and 24 h). Reverse transcriptase-coupled qPCR was used to measure the differential expression patterns of genes relevant to various immune processes and genes of antimicrobial peptides. Ghrelin isoform treatments resulted in functional perturbations that displayed overlapping and divergent patterns of gene expression. The differing actions between the two ghrelin isoforms on various assessed genes, and at differing time points, suggested that the two analogs may activate unique pathways, thereby eliciting distinct responses in fish immunity.


Introduction
The regulation of innate and adaptive immune systems is critical to animal defense against infectious diseases. Ghrelin is an essential, and multifunctional biomolecule that regulates food intake and energy balance [1][2][3], but also modulates immune processes in many animals [4][5][6][7]. This peptide hormone was first identified in rat stomachs as an endogenous ligand that binds a 7-transmembrane G-protein coupled receptor, called growth hormone secretagogue receptor (GHSR), to mediate its biological effects [1]. The initially discovered ghrelin polypeptide consisted of 28 amino acids, with the third serine O-octanoylated. This acylation plays a critical role in receptor binding and eliciting biological activity. Ghrelin is synthesized by X/A-like enteroendocrine cells, predominantly in the gastrointestinal

Fish
The experimental rainbow trout utilized in this study were purchased from Troutlodge, Inc. (Bonney Lake, WA, USA) as bioassay grade fry (Bonney Lake, WA, USA). A total of 124 mature fish (1.5 years old, 1.6 ± 0.11 kg) were cultured, and 36 fish were used for experiment. All fish were maintained in tanks (2.44 m × 1.22 m, 3600 L). Tanks were supplied with continuous flow (50 L/min) of dechlorinated municipal water at the School of Freshwater Sciences at University of Wisconsin, Milwaukee, and photoperiod was maintained at 12 h light:12 h dark. Husbandry conditions were water temperature 11-15 • C, pH 7.6, and dissolved oxygen 9 −11 mg/L. Animals were fed 4.0 mm floating classic trout diet (Skretting, Tooele, UT, USA) to 10% body weight/day dispensed over  4 feedings/day with vibratory feeders (Sweeney, Boerne, TX, USA). Animal and food wastes were directly removed by single flow-through water system, the continuous flow rate (50 L/min) was sufficient to ensure no detectable ammonia, and the water quality was constantly monitored every two weeks by FF-1A Fish Farming Test Kit (Hach, Loveland, CO, USA). Prior to tissue collection of head kidney, fish were anesthetized by immersion in MS-222 (100 mg/L, Tricaine methanesulfonate, Syndel, Ferndale, WA, buffered with 200 mg/L of sodium bicarbonate). Fully anesthetized animals were then fully bled and euthanized by decapitation. All experiments were conducted following approval and guidelines of the Institutional Animal Care and Use Committee (IACUC) of the University of Wisconsin Milwaukee (IACUC-UWM protocol 19-20 #11&12).

Tissue Collection and Cell Culture
Primary cells from rainbow trout head kidneys were isolated by modified method as described [24]. The head kidney tissue (~0.5 inches most anterior of trunk kidney) was removed from euthanized fish and placed into sterile plates with sterile-filtered Dulbecco's modified eagle medium (DMEM, ThermoFisher, Waltham, MA, USA). Head kidney tissue was minced with sterile forceps in DMEM, and the suspension was passed through 100 µm cell strainers (Greiner Bio-One, Monroe, NC, USA). Remaining tissue/residue on the cell strainer was further ground and pushed through the sterile 100 µm nylon mesh with pestles and washed twice with DMEM. The resulting cell suspension was filtered with 40 µm cell strainer (Greiner Bio-One, Monroe, NC, USA) and the cells were pelleted at 200× g for 5 min. The pellet was re-suspended in Leibovitz medium L-15, (Cytiva, Marlborough, MA, USA) supplemented with 100 IU/mL penicillin, 100 µg/mL streptomycin (Gibco, Amarillo, TX, USA) and 10% fetal bovine serum (Hyclone Characterized FBS, gamma irradiated and heat-inactivated, Logan, UT, USA). The primary cells isolated from each head kidney were evenly plated to four T25 culture flasks (Celltreat, Pepperell, MA, USA) with an average seeding density = 5 × 10 6 cells/mL and cultured at 15 • C for a total of 7 days. The nonadherent cells were removed after the first 24 h, the adherent cells were rinsed with sterile 1× phosphate-buffered saline (PBS) one time, and medium was replaced with fresh medium. Subsequently, the adherent cells were incubated for an additional 6 days with daily removal of nonadherent cells. Images of living culture were captured every day using an EVOS FL microscopy system (ThermoFisher, Waltham, MA, USA) at 100× and 400× magnification to monitor the morphological changes in immune relevant cells, and most of differentiation was achieved at 6-day post removal of non-adherent cells from RT-HKD.

Ghrelin Treatment and RNA Extraction
Prior to ghrelin treatment, the adherent RT-HKD cells were left undisturbed for 24 h in L-15 medium (without any supplement), washed twice with fresh sterile L-15, and treated for 0 h, 2 h, 4 h, and 24 h with 3 mL of L-15 containing rt-Ghrl at a final concentration of 1.25 ng/µL, and rt-desVRQ-Ghrl at final concentration of 1.11 ng/µL. Four replicates of treatments (control, rt-Ghrl, and rt-desVRQ-Ghrl) and sampling time points (0 h, 2 h, 4 h, and 24 h) were randomly assigned to T25 culture flask containing RT-HKD cells from independent fish. In total, 112 living cultures were utilized to cover all experimental conditions. At the end of each treatment, medium was removed, cells were washed once with sterile PBS, and RNAzol RT was added to each flask to lyse cells and obtain total RNA according to the manufacturer's protocol (Molecular Research Center, Cincinnati, OH, USA). Afterward, the isolated total RNA was treated with DNase, and purified by spin column chromatography (Norgen Biotek Corp., Thorold, ON, Canada). Total RNA was initially quantified by UV spectrophotometry (Nanodrop, ThermoFisher Scientific, Waltham, MA, USA) with high purity (260/280 ratios ≥ 2.0). Prior to first-strand cDNA synthesis, total RNA was further quantified by fluoroscopy with Qubit fluorimeter (Thermo-Fisher Scientific, Waltham, MA, USA). First-strand cDNA was synthesized from each RNA sample (4 µg/40 µL) with the High-Capacity cDNA reverse transcription kit (Applied Biosystems, Waltham, MA, USA) by following the manufacturer's recommended protocol. The synthetic cDNAs were stored at −80 • C until further use as template for real-time qPCR.

Real-Time qPCR Analysis
The primers utilized in this study are listed in Table 1 and were obtained from Integrated DNA Technologies (Coralville, IA, USA). Each qPCR reaction consisted of 10 µL of 2 × PowerUp SYBR Green PCR Master Mix (Applied Biosystems, Waltham, MA, USA), a 10 µL mixture containing 1 µ L of cDNA template, 1 µL forward and reverse primer (0.5 µM), and 8 µL RNase/DNase free water to a final reaction volume of 20 µL. In general, thermocycling conditions for qPCR were as follows: denaturation at 95 • C for 10 min, followed by 40 cycles consisting of 95 • C for 15 s, 60 • C for 1 min; a melting curve analysis from 65 • C to 95 • C was applied after the thermal cycle to confirm the presence of a single amplicon. The reactions were carried out with three duplicates of each sample in a calibrated QS6-Flex real-time PCR detection system (LifeTechnologies, Carlsbad, CA, USA). The standard curves of each gene were generated by a serial dilution of cDNA template from 1:1 to 1:5, 1:25, 1:125, and 1:625. Rainbow trout elongation factor-1 alpha (ef1a) was used as a housekeeping gene to normalize the RNA levels of each target gene, and beta-actin (actb) was used to confirm the reliability of normalization. Relative expression levels of target genes were calculated by Pfaffl method [25], wherein the arithmetic formula is defined as relative abundance (RA) = (1 + E target gene ) DCt /(1 + E housekeeping gene ) ∆∆Ct ; ∆ Ct = Ct control − Ct treatment , where E represents the PCR efficiency (E = (10 −1/sl•pe −1) × 100%) and Ct = the threshold cycle. Genes with RA > 1 are defined as upregulation; RA < 1 are defined as downregulation. The standard errors were determined as STDEV (value of relative abundance)/SQRT (count(number of relative abundance)).

Statistics and Bioinformatics
The significance (p < 0.05) of differentially expressed genes (DEGs) was analyzed by two-way analysis of variance (ANOVA) with time (2, 4, or 24 h) and treatment (rt-Ghrl or rt-desVRQ-Ghrl) as independent variables (main effects). Statistically significant differences between groups were defined as ANOVA F-test results having p < 0.05 for the Animals 2023, 13, 1683 5 of 17 fixed effects and interactions. Select pair-wise comparisons were performed on the group means, with Fisher's LSD (least significant difference) difference test (two-tailed), when ANOVA indicated a significant (p < 0.05) difference for fixed effects and/or interactions using Minitab statistical software package (State College, PA, USA, V 19.2.0). Graphs were prepared by Prism 9 (GraphPad Software, San Diego, CA, USA, V 9.2.1). The heatmap was generated by heatmapper web tool (http://www.heatmapper.ca/expression/ (accessed on 27 January 2023)) [26], and the hierarchical clustering analyses were conducted via using average linkage for clustering method, and absolute values of Pearson correlation coefficient (between 0 to 1) were used to compute the distance between genes. Raw data generated can be seen in supplementary material.

Hierarchical Clustering Analyses of Differentially Expressed Immune Relevant Genes
High differentiation of myeloid and lymphoid cells was achieved for ghrelin analogs treatments after RT-HKD primary cell cultures were incubated without interruption for six days post-removal of non-adherent cells ( Figure 1). To profile the differential expression levels of immune-relevant genes, RT-HKD cells were left untreated or treated with ghrelin isoforms for 0 h, 2 h, 4 h, and 24 h. Expression of selective genes (Table 1) was measured by reverse transcriptase-coupled qPCR, and normalization was conducted as described in Section 2. To visualize the expression pattern of differentially expressed genes (DEGs), a heatmap was generated and the color gradient denotes the change in relative abundance (RA). As shown in Figure 2, the hierarchical clustering analysis was conducted by computed Z-scores, which represent the relationship of normalized and scaled RAs of each gene to the mean of a group of genes, revealing the distances between differentially expressed genes (rows) induced by ghrelin treatments. Three major clusters were formed for DEGs relevant to various immune processes. Cluster I was further partitioned into two To visualize the expression pattern of differentially expressed genes (DEGs), a heatmap was generated and the color gradient denotes the change in relative abundance (RA). As shown in Figure 2, the hierarchical clustering analysis was conducted by computed Zscores, which represent the relationship of normalized and scaled RAs of each gene to the mean of a group of genes, revealing the distances between differentially expressed genes (rows) induced by ghrelin treatments. Three major clusters were formed for DEGs relevant to various immune processes. Cluster I was further partitioned into two predominant subclades that had unique patterns of potential physiological significance (Figures 3 and 4). Cluster II consisted of DEGs that established the most divergent responses between two ghrelin isoforms, particularly at 24 h where rt-Ghrl, but not rt-desVRQ-Ghrl, strongly induced expression over baseline (RA = 1) ( Figures 5 and 6). Cluster III consisted of early (2 h to 4 h) response genes that showed the most robust induction with ghrelin isoforms (Figure 7). More detailed descriptions of these patterns are presented in subsequent sections.        (irf7), (B) beta-defensin (db1), (C) hepcidin antimicrobial peptide (hamp), and (D) liver-expressed antimicrobial peptide 2 (leap2) were measured by real-time qPCR. The RA of each gene of interest was quantified in responding to ghrelin analogs treatment or L15 control. The bars presented the RAs of each mRNA as means ± SEM (n = 4). The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed by using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pairwise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another.
Animals 2023, 13, x FOR PEER REVIEW 9 of 18 Figure 5. Regulatory effects of ghrelin analogs on relative abundances of genes involved in Cluster II in RT-HKD cells (Part 1). The differential expression levels of (A) tumor necrosis factor-alpha (tnfa), (B) C-X-C motif chemokine ligand 8 (cxcl8), (C) cyclo-oxygenase 2 (cox2), (D) interleukin 1 beta (il1b), (E) cathelicidin I (cath1) were measured by real-time qPCR. The RA of each gene of interest was quantified in responding to ghrelin analogs treatment or L15 control. The bars presented the RA of each mRNA as means ± SEM (n = 4). The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed by using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pairwise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another. . The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed by using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pair-wise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another.
the RA of each mRNA as means ± SEM (n = 4). The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed by using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pairwise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another.  . The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed by using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pair-wise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another.
Animals 2023, 13, x FOR PEER REVIEW 1 after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Sign differences were analyzed by using two-way ANOVA with time + treatment as main effects there was a significant ANOVA effect (p < 0.05), differences between conditions were deter by pair-wise comparisons of the group means with Fisher's LSD test. Alpha scripts represent icant differences (p < 0.05) among conditions for each specific gene; groups with the same scripts are not significantly (p > 0.05) different from one another.

Figure 7.
Regulatory effects of ghrelin analogs on relative abundances of genes involved in III in RT-HKD cells. The differential expression levels of (A) major histocompatibility comple II (mhc2) and (B) cathelicidin II (cath2) were measured by real-time qPCR. The RAs of each interest were quantified in responding to ghrelin analogs treatment or L15 control. The ba sented the RA of mRNA as means ± SEM (n = 4). The black bars indicate the RA of each mRN rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant ences were analyzed using two-way ANOVA with time + treatment as main effects. Whe was a significant ANOVA effect (p < 0.05), differences between conditions were determined b wise comparisons of the group means with Fisher's LSD test. Alpha scripts represent sign differences (p < 0.05) among conditions for each specific gene; groups with the same alpha are not significantly (p > 0.05) different from one another.

Genes in Cluster I Partition into Two Distinct Subclades
As shown in Figure 3, four immune-relevant genes, i.e., interferon-gamma (ifng Figure 7. Regulatory effects of ghrelin analogs on relative abundances of genes involved in Cluster III in RT-HKD cells. The differential expression levels of (A) major histocompatibility complex class II (mhc2) and (B) cathelicidin II (cath2) were measured by real-time qPCR. The RAs of each gene of interest were quantified in responding to ghrelin analogs treatment or L15 control. The bars presented the RA of mRNA as means ± SEM (n = 4). The black bars indicate the RA of each mRNA after rt-Ghrl treatment and white bars indicate the RA after rt-desVRQ-Ghrl treatment. Significant differences were analyzed using two-way ANOVA with time + treatment as main effects. When there was a significant ANOVA effect (p < 0.05), differences between conditions were determined by pair-wise comparisons of the group means with Fisher's LSD test. Alpha scripts represent significant differences (p < 0.05) among conditions for each specific gene; groups with the same alpha scripts are not significantly (p > 0.05) different from one another.

Genes in Cluster I Partition into Two Distinct Subclades
As shown in Figure 3, four immune-relevant genes, i.e., interferon-gamma (ifng), suppressor of cytokine signaling 1 (socs1), integrin b2 (cd18), and immunoglobulin M (igm), were clustered in subclade I of cluster I. In general, all four genes exhibited similar expression profiles in response to the two ghrelin isoform treatments. In particular, ifng, socs1, and cd18 were suppressed in expression at 2 h but exhibited nondifferential or augmented expression levels at 4 and 24 h post-rt-Ghrl treatment. Treating with rt-desVRQ-Ghrl yielded expression profiles that were similar to rt-Ghrl, with suppression at 2 h, overexpression by 4 h, and either sustained induction or return to basal expression level (RA ∼ = 1) by 24 h (Figure 3A-C). Although clustered with these other immune-related genes, igm exhibited a slightly different pattern, with rt-Ghrl and rt-desVRQ-Ghrl inducing elevated expression levels at 2 h and 4 h but suppressed induction at 24 h ( Figure 3D). Statistical analyses revealed that four genes were significantly affected by Treatment (p < 0.000), time (p < 0.000), and a treatment × time interaction (p < 0.000) in ghrelin-treated RT-HKD cells. Additionally, the responses to both ghrelin treatments differed at various time points. The over-time effects of rt-Ghrl treatment were greater than rt-desVRQ-Ghrl on ifng (LS means 0.98 vs. 0.89), cd18 (LS means 1.15 vs. 1.02) and igm (LS means 1.74 vs. 1.50) in RT-HKD cells. However, rt-desVRQ-Ghrl treatment showed higher effects on socs1 (LS means 0.8 vs. 0.6) than rt-Ghrl treatment in RT-HKD cells.
Subclade II of cluster I contained one transcription factor, interferon regulatory factor 7 (irf7), and three antimicrobial peptide (AMP) genes, i.e., beta-defensin 1 (db1), hepcidin antimicrobial peptide (hamp) and liver-expressed antimicrobial peptide 2 (leap2). As with subclade I, highly similar over-time effects on expression patterns with three of these four genes, including irf7, db1, and hamp ( Figure 4A-C). These genes were inhibited significantly at 4 h but induced at 24 h following rt-Ghrl treatment. In contrast, rt-desVRQ-Ghrl exerted a divergent over-time regulatory profile, expression elevated significantly at 4 h and suppressed, in some cases strongly, at 24 h. Although leap2 was somewhat distinct from the other genes in its response to rt-Ghrl, it mirrored the aforementioned response to rt-desVRQ-Ghrl, justifying its inclusion in the subclade ( Figure 4D). Statistically, three of the genes (irf7, hamp, and leap2) showed significant effects of treatment (p < 0.000), time (p < 0.000), and a treatment × time interaction (p < 0.000). While there was no significant effect of treatment (p < 0.99), there were significant effects of time (p < 0.000) and a significant treatment × time interaction (p < 0.000) on RAs of db1 mRNA in RT-HKD cells. In general, the main distinguishing feature of this subclade was the inclusion of most of the AMPs tested.
Other than tgfb1, nine of ten genes in cluster II showed significant effects of treatment (p < 0.000), time (p < 0.000) and a treatment × time interaction (p < 0.000). Over time, the calculated value of LS means indicated that the effects of rt-Ghrl on these eight genes were greater than rt-desVRQ-Ghrl in RT-HKD. Although tgfb1 was not affected by treatment (p = 0.96), there were significant effects of time (p < 0.000) and treatment × time interaction (p < 0.000). In addition, tgfb1 mRNA increased through time-course (LS means, 2 h = 0.7, 4 h = 1.1, and 24 h = 1.5), further underscoring the absence of a treatment effect on this mRNA ( Figure 6C).

Cluster III Early Response Genes Are Strongly Induced by Both Ghrelin Isoforms
Cathelicidin 2 (cath2) and major histocompatibility complex class II (mhc2) were partitioned into Cluster III ( Figure 2). As shown in Figure 7A, mRNA levels of mhc2 were massively upregulated at 2 h by treating with both ghrelin isoforms in RT-HKD cells. By 24 h, mhc2 mRNA had dropped below basal expression level with rt-Ghrl and rt-desVRQ-Ghrl treatments, characteristic of a potently transient, early response gene. For cath2, an even more transient induction was observed ( Figure 7B), with the exception of the 24 h response to rt-Ghrl, where a strong, possibly secondary response, yielded a re-induction of expression that rivaled the initial response. Statistical analyses showed that RAs of mhc2 and cath2 mRNA were significantly affected by treatment (p < 0.000), time (p < 0.000), and a treatment × time interaction (p < 0.000) in RT-HKD cells. In addition, different from genes in cluster I and II, the over-time effects of rt-desVRQ-Ghrl treatment (LS means 1.7 vs. 1.12) on mhc2 and cath2 were greater than rt-Ghrl in RT-HKD cells.

Ghrelin Modulates Host Immune Processes via Endocrine Axes and Cytokine Signaling Transduction Pathways
Ghrelin is a multifunctional hormone/cytokine that was initially discovered as a ligand for the growth hormone (GH) secretagogue receptor, which functions to influence central growth hormone release [1]. It is well known to regulate energy homeostasis, satiety, and reproduction, but it also modulates the immune system in many species [1,2,4,5]. In fish, accumulated evidence supports the notion that ghrelin regulates various immune pathways that influence the host immunity [5,10,15,27], which is consistent with the immunomodulatory effects of GH in fish [28]. Given the apparent broad immunomodulatory effects of the GHS and GH system in finfish and the apparent crosstalk between the two, it is possible that ghrelin alters the expression of genes relevant to various immune processes independently of central GH stimulation.
Preliminary studies also suggest that ghrelin generally suppresses pro-inflammation genes in various animal models [4]. However, in fish, in vivo pathogenic challenge and exogenous treatment with GHS analogs have shown divergent results. For instance, in vivo treatment of hybrid tilapia with ghrelin promotes expression level of il1b in the liver and kidney after 5 h and 10 h of infection with Aeromonas hydrophila and inhibited tnfa mRNA in the liver at 10 h and kidney at 5 h [29]. Another study reported that in vivo growth hormone-releasing peptide-6 (Ghrp6) (an agonist of the GHSR that mimics the effect of ghrelin) elevated the expression level of il1b mRNA in the head kidney but inhibited it in the spleen of tilapia challenged with Pseudomonas aeruginosa [30]. In addition to modulating pro-inflammation, ghrelin was reported to stimulate phagocytosis and superoxide production in rainbow trout leukocytes [19] as well. Other reports revealed the regulatory effects of ghrelin on multi-functional bioregulators such as ifng, tnfa, and tgfb1, etc., which may suggest that ghrelin modulates host immune responses via mediating cytokine/chemokine signaling pathways [4,29,31,32]. In the current study, we evaluated the regulatory effects of ghrelin analogs on cytokines, chemokines, transcription factors, interferon, and growth factor, and the differential expression of these bioregulators may result in functional perturbations of corresponding immune pathways. By combining our findings and the literature evidence, these results indicate ghrelin modulates host immune processes via not only the central regulation of the somatotropic axis but also through cytokine/chemokine signaling transduction pathways.

Primary Cell Culture Represents Head Kidney as an Immune Competent Organ in Rainbow Trout
Primary cells isolated from RT-HKD have been a valuable model to study fish immune processes [24,33]. For instance, stress hormones modulated the productions of il1b and il6 in the head kidney primary cell model of rainbow trout and sea bream (Sparus aurata) [34]. Additionally, RT-HKD cells have been employed to study the effects of cytokine stimulation on the differentiation of leukocytes and lymphocytes [35]. To minimize the use of experimental animals and avoid potential complexities associated with in vivo ghrelin treatment, we adapted this compatible in vitro system to explore the direct immunomodulatory effects of ghrelin on RT-HKD primary cells. Several major types of myeloid and lymphoid cells were observed and differentiated to maturation in the living culture system as presented in Figure 1. Such observations affirm that this in vitro model is sufficient to represent at least a portion of the immune-relevant cellular components of the RT-HKD and will further enable evaluation of the direct actions of ghrelin isoforms in this finfish. Expression patterns of twenty immune-relevant genes were profiled by reverse transcriptase-coupled qPCR at three time points after treatment with rainbow trout ghrelin analogs in RT-HKD. Among these DEGs, three major clusters were identified in the hierarchical analysis, differences within Cluster I were significant enough to divide this cluster into two subclades ( Figure 2). The modulatory effects will be discussed in the following sections.

Ghrelin Isoforms Exert Differential Modulation of Immune Relevant Genes in Two Distinct Subclades in Cluster I
As mentioned, the differences in RAs were significant enough for us to divide Cluster I into two distinct subclades (Figures 3 and 4). By comparing the expression patterns induced by ghrelin isoforms in subclade I (Figure 3), we observed that both isoforms moderately suppressed genes at 2 h, but that rt-desVRQ-Ghrl inhibited expression at 24 h, while levels returned to basal expression (RA ∼ = 1) at 24 h with rt-Ghrl. For DEGs in subclade II (Figure 4), with the exception of db1 at 2 h, rt-Ghrl inhibited the expression levels at 2 h or 4 h, but rt-desVRQ-Ghrl enhanced these mRNAs at 4 h and suppressed them at 24 h. This conserved divergence in the effects of the two ghrelin analogs indicates that significant physiological responses, and thus biological functions, likely exist in rainbow trout. Furthermore, while being inhibited by rt-Ghrl at 2 or 4 h, the expression levels of irf7, db1, hamp, and leap2 were promoted by rt-desVRQ-Ghrl at 4 h. This observation may suggest that the two peptides exert dynamic feedback regulation (i.e., secondary effects) at different time points, or that separate signaling pathways are engaged.
Interestingly, besides cathelicidins (cath1/2), all AMPs evaluated in this study are clustered in subclade II. Many studies indicated that GHS modulates the expressions of various antimicrobial peptides in multiple fish species [20,29]. For example, Ghrp6 (a mimic of GHS) induces the overexpression of piscidin-like AMP in tilapia (Oreochromis sp.) [30], while pituitary adenylate cyclase-activating polypeptide, a type of growth hormone releasing hormone, promotes expression of pardaxin and hepcidin in African sharptooth catfish (Clarias gariepinus) [36]. In the present study, we demonstrated that ghrelin analogs alter the expression of antimicrobial peptides at different time points and assessed different AMP genes. Many reports showed that the production of AMPs was often regulated by Toll-like receptors (TLRs) and NOD-like receptors (NLRs) mediated NF-κB signaling pathways [37,38], and ghrelin is known to regulate NF-κB signaling in various species [5,18,39]. Therefore, our results may suggest an alternative route where ghrelin isoforms modulate the host innate immune responses via TLRs/NLRs signaling, and the differential expres-sions of AMPs may result from a combination of ghrelin-mediated functional perturbations of GHS and NF-κB signaling pathways in RT-HKD cells.

Ghrelin Isoforms Modulate Divergent Late-Stage Responses on DEGs in Cluster II
All of the DEGs assigned to Cluster II via hierarchical analysis were significantly induced by rt-Ghrl treatment at 24 h, but, with some exceptions, either returned to baseline or were slightly inhibited by rt-desVRQ-Ghrl at 24 h (Figures 5 and 6). This finding affirms the notion that two ghrelin isoforms play divergent roles in modulating host immunity at various time points. Many cytokines, chemokines, and transcription factors, e.g., il1b, il6, cxcl8, cxcl11, tnfa, and tgfb1, were found in this cluster. The lateness of the induction (24 h) in response to rt-Ghrl is notable and may suggest a delayed or even a secondary induction. Because induction of these same genes by rt-desVRQ-Ghrl either was not observed or was less significant, the types of signaling cascades induced must surely differ. If the pattern is due to crosstalk among different signaling pathways, it seems that clarifying which pathways are differentially induced will assist significantly in our efforts to understand the physiological distinctions in these isoforms. Alternatively, if secondary pathways regulate cytokine/chemokine signaling, then identifying the early responders (i.e., those falling into cluster III) will help to predict which intermediates mediate cytokine/chemokine signaling cascades that may be distinct from those activated by rt-Ghrl analogs directly [22,[40][41][42]. Furthermore, multi-functional bioregulators (ifng, tnfa, and tgfb1) showed similar expression patterns, inhibited at early and enhanced at late stages, by rt-Ghrl treatment. Since these regulators mediated the production of many cytokines and chemokines [4,40,43], the perturbed downstream signaling of these bioregulators can alter the expression of cytokines and chemokines. Therefore, the combining effect from ifng, tnfa, and tgfb1 may also impact the secondary, or even tertiary responses.
As mentioned in Section 4.1, evidence indicated that ghrelin and GHS modulate inflammation via altering cytokine signaling pathways in teleost fish [29,30,36]. By combining this notion and our observations of significant secondary inductions of pro-inflammation cytokines (tnfa, il1b, and il6), and enzyme (cox2), these studies suggest that rt-Ghrl promotes expressions of pro-inflammation genes at late-stage and, thus, may exhibit proinflammation activity in RT-HKD cells, whereas rt-desVRQ-Ghrl does not exert similar regulation. In addition, a common upregulation of ifng and tgfb1 mRNAs at 24 h were induced by rt-Ghrl (rather than rt-desVRQ-Ghrl) in RT-HKD cells ( Figures 3A and 6C). Since tgfb1 induces expressions of il1b and il6 in lymphocytes [31,42] and ifng mediates the production of various types of cytokines in immune-relevant cells [22,40], our findings may suggest that ghrelin modulates inflammation via two routes: directly mediating cytokine expression and indirectly altering downstream signaling of multi-functional bioregulators (tnfa, ifng, and tgfb1), simultaneously.

Ghrelin Isoforms Induce Strong Early Responses on DEGs in Cluster III
The overexpression of mhc2 and cath2 was strongly induced by both ghrelin peptides at an early stage (2 h), and the inductions dropped over time and finally below to basal expression level at 24 h. The strong induction of mhc2 is reminiscent of an immediate-early cytokine response, such as ifng responses that elicit rapid, transient upregulation of mhc2 in most cells via JAK/STAT signaling-a pathway anticipated to modulate indirect ghrelin effects [28,40,44]. Additionally, cd86 encodes a constitutive expression co-stimulator protein (B7-2) on antigen-presenting cells (APCs) and mediates cell-cell association during antigen presenting process [45,46]. By combining the ghrelin-induced overexpression of mhc2 (antigen binding molecule) and cd86 (stimulator of APCs) at an early stage (2-4 h), this finding may suggest ghrelin modulates the enhancement of antigen-presenting process early, but negative feedback regulates this process at a late stage in RT-HKD cells.
Although cath1 and cath2 showed different patterns, some similarities of late-stage (24 h) inductions were observed in their expression profiles. As shown in Figures 5E and  7B, rt-Ghrl promoted late inductions of both cath1 and cath2 at 24 h, whereas rt-desVRQ-Ghrl inhibited the expressions of both cathelicidin paralogs at this late stage. This finding once again suggests that divergent actions of ghrelin analogs on different assessed genes may lead to differentially expressed cath1 and cath2 at 24 h, but that changes to related gene targets such as these may be subtle and change only the magnitude or timing of the response.

Conclusions
Using an in vitro system, we demonstrated that two ghrelin isoforms differentially modulate immune-related genes in rainbow trout cells. Our results support the hypothesis that ghrelin modulates host immunity via not only the somatotropic axis but perhaps also via cytokine signaling pathways-some activated directly, some responding indirectly. The divergent expression patterns of the responsive genes indicate that the two ghrelin peptides exert different functional perturbations in regulating immune-relevant genes in primary RT-HKD cells. These differing actions suggest the two analogs may activate unique pathways and, therefore, elicit distinct responses in fish immunity. Although we have directly assessed the immunomodulatory effects of ghrelin isoforms in RT-HKD cells, further studies will be required to identify the full cohorts of responsive genes and to clarify the differential roles and mechanisms by which these ghrelin analogs affect immunity in rainbow trout.